Search is not available for this dataset
article stringlengths 4.36k 149k | summary stringlengths 32 3.35k | section_headings listlengths 1 91 | keywords listlengths 0 141 | year stringclasses 13
values | title stringlengths 20 281 |
|---|---|---|---|---|---|
Ligand binding to death receptors activates apoptosis in cancer cells . Stimulation of death receptors results in the formation of intracellular multiprotein platforms that either activate the apoptotic initiator Caspase-8 to trigger cell death , or signal through kinases to initiate inflammatory and cell survival signalling . Two of these platforms , the Death-Inducing Signalling Complex ( DISC ) and the RIPoptosome , also initiate necroptosis by building filamentous scaffolds that lead to the activation of mixed lineage kinase domain-like pseudokinase . To explain cell decision making downstream of death receptor activation , we developed a semi-stochastic model of DISC/RIPoptosome formation . The model is a hybrid of a direct Gillespie stochastic simulation algorithm for slow assembly of the RIPoptosome and a deterministic model of downstream caspase activation . The model explains how alterations in the level of death receptor-ligand complexes , their clustering properties and intrinsic molecular fluctuations in RIPoptosome assembly drive heterogeneous dynamics of Caspase-8 activation . The model highlights how kinetic proofreading leads to heterogeneous cell responses and results in fractional cell killing at low levels of receptor stimulation . It reveals that the noise in Caspase-8 activation—exclusively caused by the stochastic molecular assembly of the DISC/RIPoptosome platform—has a key function in extrinsic apoptotic stimuli recognition .
Apoptotic signalling cascades are designed to irreversibly lead to cell death once specific death thresholds are overcome [1 , 2] . Activation of caspases plays a central role in this process . In certain scenarios , apoptotic cell death signalling is interrupted . This may lead to the activation of other forms of cell death or escape from cell death altogether . Death ligands ( DL ) bind to death receptors ( DR ) at the plasma membrane and have been developed as novel cancer therapeutics . However , many cells in our body are exposed from time to time to endogenous DLs , such as TNF-α and TRAIL , without induction of cell death . Several studies have shown that while binding of DLs to DRs can induce apoptosis , not all cells will respond to DR stimulation with cell death , and only a fraction of the cell population will undergo apoptosis even if DLs bind at death-inducing concentrations [2–6] ( Fig 1A ) . Interestingly , in vivo studies have shown that fractional death resistance has no direct association with the amount of DRs expressed on the plasma membrane [7 , 8] . Therefore , cell signalling activated by extrinsic ‘death’ signals is rather encoded downstream of receptor binding . Binding of DLs to dedicated DRs triggers either the formation of receptor-associated Death-Inducing Signalling Complexes ( DISC ) ( ‘Complex I’ ) in proximity to the plasma membrane , or RIPoptosome complexes ( ‘Complex II’ ) in the cytosol [5 , 9–17] . Both complexes provide a platform for the activation of the initiator Caspase-8 ( Casp8 ) . For the activation of Casp8 , the inactive pro-form of Casp8 ( ProCasp8 ) must undergo autocatalytic activation . This is achieved through ProCasp8 dimerization and sequential inter- and intradimer cleavage , a process which results in the release of active Casp8 [18–20] . The dimeric ProCasp8 association-dissociation balance has been suggested to play a crucial role in the molecular control of apoptotic responses after DR activation [21] . However , as demonstrated by mutagenesis studies , ProCasp8 dimerization alone is not sufficient to enhance apoptotic responses in vivo [22] . Instead , formation of the DISC or RIPoptosome platforms are necessary for effective ProCasp8 dimerization and Casp8 activation [10 , 23 , 24] . Apart from apoptosis initiation , DR-induced complexes also initiate necroptosis by accumulating heterodimers of receptor-interacting proteins ( RIPs ) , RIP1 and RIP3 ( RIP1/3 ) , and the formation of filamentous scaffolds [25–28] . Formation of such ‘Necrosome’ platforms activates the mixed lineage kinase domain-like ( MLKL ) pseudokinase . MLKL activation triggers necroptosis , a cell death distinct from apoptosis [29–31] . In theory , activation of DRs in individual cells could lead to both apoptosis and necroptosis signalling through the formation of different platforms . However , if RIP1/3 proteins are close to the site of Casp8 activation , RIP1/3 is cleaved by Casp8 [32] . This cleavage eliminates the kinase activity of RIP1/3 , and consequently necroptosis activation is suppressed [9 , 33–36] ( Fig 2B ) . This suggests that if one type of cell death is triggered in a given cell , the other type of cell death is suppressed , i . e . , that the two types of cell death are mutually exclusive . Previous studies of the apoptotic signalling network activated by DRs have identified that variability in death signalling arises from the process preceding the mitochondrial outer membrane permeabilization ( MOMP ) . This process triggers Casp8-mediated cleavage of the pro-apoptotic Bid protein [2 , 4 , 37] , which mediates MOMP and leads to cytochrome-C release , apoptosome formation and executioner caspase activation [38] . To understand cell death decision making in more detail , we created a mathematical model which incorporates the central events prior to Bid cleavage . The model was constructed to estimate apoptotic and necroptotic pathway initiation through the random assembly of the DISC/RIPoptosome platform . As a multiprotein platform with diverse functionality , we hypothesised that the random and stochastic process of its assembly may lead to the heterogeneous cellular responses ( Fig 1A and 1B ) . Combining this model with experimentally derived sets of quantitative protein profiles and literature-based catalytic and binding rates , we simulated the heterogeneous responses of HeLa cells to DR activation . By modelling different conditions of DR stimulation and clustering , we investigated in particular how heterogeneous apoptotic responses arise , which role the random assembly of DR-induced platforms play in determining death delay at the single cell level , and how DR clustering facilitates death signalling . Our analysis reveals that the noise in Casp8 activation exclusively caused by the stochastic molecular assembly of the DISC/RIPoptosome platform has a key function in the low level extrinsic apoptotic stimuli recognition .
Apoptosis inducing DRs such as Tumour Necrosis Factor Receptor 1 ( TNFR1 ) and Death Receptors 4 and 5 ( DR4/5 ) are expressed at comparable protein levels in HeLa cells [39] . Additionally , it is known that their protein expression level is correlated with the receptor abundance on the cell surface [8] . High variation in TNFR1 surface abundance were estimated in previous studies ranging from 300 to 3000 molecules per single HeLa cell [40 , 41] . To get more accurate estimates , we performed the single cell quantification of TNFR1 membrane expression in HeLa cells employing the QuantiBRITE phycoerythrin beads based assay ( see S1 File ) . We determined that the average number of TNFR1 does not exceed 905 receptors per cell . We further used this quantity as the reference in our comparative quantification of DR4/5 receptors based on MS data set ( S1 File ) . Thus , we calculated that DR4 and DR5 receptors are present on HeLa cell surface in an average amount of 769 and 926 monomeric receptors , respectively ( Table A in S1 File ) . Next , we estimated the amount of the DR complexes associated with DL at the single cell level . Due to the fact that the DR-DL association is generally much quicker [42] than the downstream processes such as ProCasp8 dimerization and subsequent Casp8 activation [43] , we applied the rapid equilibria approximation to calculate the amount of DL bound receptors . According to the law of mass action the time evolution of the amount of DR-DL complexes is d[RL]dt=kon[R][L]-koff[RL] , where[R]=[Rtotal]-[RL] ( 1 ) Where [Rtotal] is the total number of receptors per cell ( Table A in S1 File ) , [RL] is the number of DR-DL complexes and [L] is death ligand concentration ( Table B in S1 File ) . Setting the RL to the rapid equilibrium d[RL]dt=0 ( 2 ) From ( 1 ) we calculated the average number of DR-DL complexes per cell as a function of L , Rtotal and the DL dissociation constant Kd [RL]=[Rtotal] ( Kd[L]+1 ) ( 3 ) The minimal unit of the active DR-DL complex is the trimer [44] . The trimeric DR-DL complex gives birth to a single DISC platform which internalizes within the subsequent 10–15 minutes [45 , 46] . If the DISC has not bound to cellular Inhibitor of Apoptosis Proteins ( cIAPs ) , cIAP1 or cIAP2 , then it either releases active RIP1 protein into the cytosol [47] where it can form RIPoptosome or Necrosome platforms ( Fig 2B ) ( as in case of TNFR1 ) , or it makes active RIP1 protein accessible for further RIP1/3 and ProCasp8 proteins accumulation on the DISC itself ( as in case of DR4/5 activation ) [5 , 17] . Therefore , in the modelling routine each activated DISC was translated into a single RIP1 protein molecule which is available immediately after DL introduction to the cell culture . Trimeric DR-DL complexes tend to organise high order clusters in cellular membranes [44 , 48] and bring several associated DISC/RIPoptosomes into close proximity . Such clustering stimulates more efficient signalling [49] and enables ProCasp8 activation not only by dimerization on the single DISC/RIPoptosome but also by synchronised binding of two ProCasp8 monomers with two independent DISC/RIPoptosomes within one cluster . To introduce DISC/RIPoptosomes clustering processes in the model , we estimated the number and the size of the DR-DL clusters based on the experimentally derived DR-DL probability distribution from a study published earlier by Fricke and co-workers [44] . We calibrated probability redistribution from the total pool of activated trimeric DR-DL complexes , calculated in the previous step , to the clusters of different size ( see S1 File ) . Using these probabilities , we assigned for each random DISC/RIPoptosomes complex formed its associated cluster . The final algorithm assumes that DISC/RIPoptosomes complexes within one cluster are able to first encourage the activation of ProCasp8 by direct dimerization ( cis-activation ) and subsequently activate ProCasp8 via simultaneous binding within closed proximity ( trans-activation ) ( Fig 2A ) . Thus , this information about the amount of the activated DR-DL complexes and their clustering conformation served as an important input for the model . The scenario of non-clustering DR signalling was studied as well by setting the probability of trans-activation of Casp8 within DISC/RIPoptosomes complexes cluster to zero . This scenario is hereafter referred to as disrupted clustering . We have developed a core model capturing the cascade of intracellular reactions that are essential for the initiation of the apoptosis . The model reactions are partitioned into two modules: a stochastic and a deterministic module ( Fig 2 ) . The first stochastic module represents the process of stochastic assembly of DR-induced DISC/RIPoptosome multiprotein platform which facilitates initiation of ProCasp8 dimerization and self-activation by cleavage ( Casp8*; activated Casp8 dimer in Fig 2E ) . We implemented this module with the direct Gillespie stochastic simulation algorithm [50 , 51] which accounts for molecular fluctuations and slow association and dissociation rates following each component of the platform individually . It assigns the reaction propensities in probabilistic terms . The binding propensities of ProCasp8 together with its binding partner protein , Fas-associated death domain protein ( FADD ) , and competitor protein RIP1/RIP3 that comprise the core scaffold of RIPoptosome are calculated from the concentrations that we quantified experimentally in HeLa cell culture ( Table D in S1 File ) . FADD protein is crucial for apoptotic initiation [35 , 52] . This protein consists of both a Death Effector Domain ( DED ) and Death Domain ( DD ) which are specific motifs for ProCasp8 [53] and RIP1 [16 , 54] self-oligomerization respectively . Through these domains , ProCasp8 and RIP1 are bridged via FADD ( as shown in grey in Fig 2B ) . RIP3 protein can form homo-oligomers , but can also associate with RIP1 scaffolds through the RIP homotypic interaction motif ( RHIM ) , forming amyloid structures [27 , 28] ( Fig 2B ) . Intensive recruitment of RIP3 molecules to the amyloid triggers transphosphorylation of RIP3 by RIP1 with consequent transmission of phosphate groups to the MLKL pseudokinase . Phosphorylated MLKL executes necroptosis [25 , 30] . Therefore , in the absence of FADD and joint Casp8 activation platforms these structures spontaneously trigger necroptosis [35 , 55 , 56] ( necrosome complex; purple in Fig 2B ) . Additionally , we quantified the concentrations of the cellular FLICE ( FADD-like IL-1β-converting enzyme ) -inhibitory protein ( c-FLIP ) . As a DED-containing protein , cFLIP in its short ( cFLIPs ) and long ( cFLIPl ) form , can be recruited to the ProCasp8 platform abrogating or restricting activation of Casp8 [53 , 57 , 58] ( cFLIP molecules; light and dark blue in Fig 2B ) . In addition to this suppression , Casp8 activation can be disrupted by binding its own processed DEDs which may remain in the cytosol ( DED1-DED2; white in Fig 2B ) . The second deterministic module mimics the activation of two effector caspases , Caspase 3 ( Casp3 ) and Caspase 6 ( Casp6 ) which is triggered by stochastically activated Casp8 . Pro-forms of both caspases form stable dimers at physiological concentrations [59] . By cleavage , Casp8 activates Casp3 ( Casp3*; activated dimer of Casp3 in Fig 2E ) [60] . Casp3* activates Casp6 ( Casp6*; activated dimer of Casp6 in Fig 2E ) [61 , 62] and has autocatalytic function cleaving ProCasp3 [63 , 64] . Finally , Casp6* can cleave free ProCasp8 ( Casp8*; cleaved monomer of Casp8 in Fig 2E ) [64–67] however Casp8 becomes active only after a very slow dimerization ( Casp8* ) [19 , 21] . Previous models suggest that this effector caspase feedback upon weak DR stimulation probably can accelerate Casp8 activation which was initially started at the DISC or RIPoptosome platform [68] . However , the feedback can be inhibited by X-linked IAP ( XIAP ) which tightly binds Casp3 and , further , marks Casp3 with ubiquitin that leads to its proteasomal degradation [69 , 70] . The overall dynamics of Casp8 activation can be tracked quantitatively with a Casp8-specific FRET cleavage probe ( FRET , Fig 2E ) . The fixed threshold rate of this FRET probe cleavage accurately determines the moment of MOMP in HeLa cells [2] . Based on the mass action and conservation laws , the time evolution of the variables that comprise this module were modelled by a deterministic system of ordinary differential equations ( ODE ) ( details in S1 File ) . All protein concentrations and parameters used in the model are provided in Tables D and E of Materials and Methods file ( S1 File ) . The estimated weight of the RIPoptosome after short DR-targeted stimulation may exceed 2MDa [10 , 24 , 29] . To reproduce the RIPoptosome growth and composition we first employed the stochastic modelling module simulating the assembly of the individual RIPoptosomes at the single cell level . RIP1 on its own forms unlimited filaments in vitro [28] , however , in the cell the long-term RIPoptosomal filament growth is limited by the cell volume and the stiffness of the cellular components . We followed unlimited filament growths without implementation of these physical limits , focusing on the initial dynamics of RIPoptosome progression . Fig 3 and S2 Fig illustrates the simulated molecular composition of RIPoptosome in the single HeLa cell treated with a dose of 5 ng/mL of the DL ( rhTRAIL ) . The composition and the time evolution of individual RIPoptosomes within single cell differed from one to another . Consequently , the size and , therefore , molecular weight of those RIPoptosomes varies as well . As an example , we display the composition change in a few randomly chosen RIPoptosomes over the first 20 min with 1 min step interval ( Fig 3A , S2 Fig ) . Next , we calculated the progression of the molecular weight of a complete cellular pool of RIPoptosomes as simulated by the model . Interestingly , we found a high degree of variation between the RIPoptosomes formed within the same cell ( Fig 3B ) . Our simulations confirmed that in HeLa cells , the most populated protein within each RIPoptosome is RIP1 through its highly stable association mechanism . This is explained by the RHIM domain binding property that shares homology with β-amyloids assembly domains . Simulation of the model revealed that the RIP1 filaments formation is triggered immediately after DR stimulation ( Fig 3A ) . The model also predicts that it would be possible to observe RIPoptosomes of size 2 MDa only 5 minutes after DR stimulation ( Fig 3B ) . FADD recruitment to the fraction of the high molecular weight complexes is persistently increasing with post treatment time [29] . Our simulations show as well that the abundance of FADD within a single RIPoptosome increases linearly with time progression ( Fig 3C ) in conjunction with the filament growth . As a result , the abundance of FADD on average will not exceed the amount of 10 molecules per origin within the first two hours . Moreover , this abundance is independent of DL dose . Thus , a low dose of 5 ng/mL of the DL and a high dose of 50 ng/mL will result in similar FADD abundance ( S4A and S4B Fig ) . On the contrary , ProCasp8 recruitment in the single cell is most abundant in the RIPoptosome of the lower molecular weight ( Fig 3B ) . The binding of the ProCasp8 or its DEDs domain to the end of the filament blocks the RIP1 recruitment and therefore also blocks intensive filament growth by competition . The population average over 600 cells shows that ProCasp8 abundance per RIPoptosome ( origin ) saturates after 2 hours of stimulation ( S4C and S4D Fig ) . This relative abundance does not vary significantly for doses of 5 or 50 ng/mL of the DL and is unaffected by the clustering or non-clustering assumption in the model . These rapid saturation dynamics of ProCasp8 compared to linear FADD translocation has been observed earlier in experiments where no co-binding of FADD and Casp8 has been observed after 1 hour of stimulation but has become apparent at the second hour [29] . Molecular fluctuations in the RIPoptosome composition within single cells cause the fluctuations in the active Casp8 abundance ( Fig 3D ) . Stochastic single cell Casp8 activation traces for 5 ng/mL dose simulation with the corresponding per cell accumulation of Casp8 Pro domain ( DED1-DED2 ) are shown in Fig 3D and 3E . Interestingly , limited expression of RIP3 [28] protein in HeLa cell gives rise to very low and therefore heterogeneous distribution of RIP1-RIP3 heterodimers among the cells ( Fig 3F ) making the spontaneous event of the necroptosis less probable to overtake the apoptotic course of the cell death . Averaged over the population the Casp8 activation time course demonstrated high dependence on the dose of the DL as well as the clustering capacity ( S3 Fig ) . Thus , even low doses of the DL with enhanced clustering property can activate Casp8 . This result confirms the established success in the application of combinational therapeutics where the DL has been combined with the ligand specific cross-linking antibodies that enhance receptor clustering [49] . As expected , the overall variability in the Casp8 activation is a function of the treatment dose ( Fig 3G ) . Despite the coefficient of variation being within the limits of low-variance ( less than 1 ) , the early Casp8 initiation dynamics can bring significant stochasticity into triggering the downstream death pathway . Interestingly , the enhanced receptor clustering did not reduce the variability in the individual HeLa cell Casp8 activation dynamics significantly . We observed only a minor decrease in the coefficient of variation over all tested conditions ( Fig 3G ) . Next we studied the downstream caspase cleavage cascade , the second deterministic modelling module ( Fig 2E ) , which feedbacks to the DISC/RIPoptosome based Casp8 production and is potentially capable of boosting cell apoptotic capacity especially following treatment of low doses of DL [68] . As an input we used the population average of the stochastic traces ( Fig 2D , S3 Fig ) we simulated for the first module of the DISC/RIPoptosome based network initiation assuming DR clustering ( Fig 2B ) . Thus , we merged two modules into one complete deterministic system ( Fig 2C ) which enabled us to adjust undetermined parameters and estimate parameter sensitivity , hence avoiding computationally expensive parameter scans of the full stochastic formalism ( see Materials and methods , S1 File ) . The first undetermined parameter is the rate constant of Casp3 ubiquitin dependent degradation ( kcat ) . Ubiquitination of active Casp3 , which is set by XIAP , will attract proteasomal complex leading to Casp3 degradation . However , application of proteasome inhibitors does not stabilise the pool of active Casp3 and consequently does not result in reduced Casp3 proteasomal degradation . Instead , Casp3 catalytic activity is absolutely required for its own proteasomal degradation [71 , 72] . Therefore , dynamics of Casp3 degradation triggered by XIAP will not match the general degradation dynamics triggered by ubiquitin ligases for other types of proteins and this specific rate constant needed to be identified . We estimated that kcat needs to be significantly higher ( 1 . 75 min-1 ) from the general ( basal ) ubiquitin-dependent degradation rate ( 0 . 04 min-1 ) [73] ( Table E in S1 File ) . Again , low doses of the DL bring into play a switch-like sensitive response to the change in kcat value ( Fig 4A ) . In this case the cell death delay can be initiated in a spontaneous fashion if the Casp3 degradation mechanism is perturbed . Furthermore , the similar steep ultra-sensitive response can be also initiated by the mild fluctuations in the XIAP concentration . We found that slight deviations from the mean XIAP level , 63 nM , quantified earlier for HeLa [1] could speed up the cell death by more than 3-fold in the case of low DL doses ( Fig 4B ) . This decrease could be very sudden through this switch-like type of response . Indeed , XIAP specific inhibitors such as Embelin , Mithramycin A are able to overcome the DL resistance in different cancer types [74 , 75] . Finally , with the fully identified parameter set we formulated the new semi-stochastic hybrid model of apoptotic pathway initiation in a single cell with the fixed partitioning of the whole network into discrete ( Fig 2B ) and continuous reactions ( Fig 2E ) . The slow discrete reactions are the DISC/RIPoptosome assembly . The fast continuous reactions capture the caspase cleavage cascade . The simulation results for a single cell response on the addition of low and high amounts of the DL are demonstrated in Fig 5 . We observed a prolonged ramp effect for all variables of the network before the system switched to the rapid response . The ramp duration for the displayed example exceeded 10 hours after treatment with the low dose of the DL ( Fig 5A ) . Whereas the high dose treatment stimulates the ramp for shorter times , around one hour for a shown example ( Fig 5C ) . In a similar manner to the simulations with our entirely deterministic model , the delay for the switch in the single cell response is a function of DL dose ( Fig 4 ) . However , for both high and low doses we also observed very high dynamic noise in the ramp ( Fig 5A and 5C ) . This noise characterises the time course of dimeric Casp8 and active Casp3 accumulation . In experiments both proteins are very unstable and hardly detectable in the pre MOMP period of apoptotic initiation [19 , 71] . As we have shown earlier initial formation of new Casp8 dimer species can be limited by the vulnerability in molecular assembly of the DISC/RIPoptosome platform ( Fig 3 ) . Moreover , active Casp8 dimer is unstable due to high dissociation rate in the cytoplasm [19 , 20] . Indeed , Casp8 under physiological concentrations is found mainly in monomeric form [18 , 20 , 59] ( Fig 5B and 5D ) . Therefore , this process prevents accumulation of the excess catalytically active pool of Casp8 for further downstream apoptotic signalling in the pre MOMP period . Casp3 , as the main Casp8-dependent effector caspase [60] , follows the noise in the dynamic course of Casp8 dimer during the ramp . Besides , Casp3 is sacrificed in the pre MOMP period due to the excess amount of XIAP which effectively [1 , 76] blocks Casp3 activity by binding and subsequent ubiquitination which leads to Casp3 degradation . To study how the ramp noise property in individual cells influences the cell death delay we have performed 600 independent simulations of the semi-stochastic model mimicking the overall cell culture response . These simulations were repeated for four different scenarios: low and high dose treatment scenarios with or without receptor clustering order . The coefficient of variation in Casp8 dependent FRET probe cleavage calculated over ramp period for each cell was considered as the measure of the noise strength . As earlier , the moment of the individual cell death was recorded once the rate of FRET probe cleavage exceeded the expected experimental threshold rate [2] ( Fig 5E and 5H ) . For the individual cells treated with low dose the cell death delay varied from 1 to 10 hours if we integrated the receptor clustering order . Even higher variability was observed when the clustering was absent . In this case the cell death time could vary from 1 to 22 hours . Examples of FRET time traces for five individual cells are shown in Fig 5E . By visualising the relationship between the single cell death delay and dynamic ramp noise strength over a population , we found out that noise was an important determinant of the delay . For both clustering and non-clustering scenarios this relationship follows the same trend ( Fig 5F and 5G ) . Moreover , this trend was independent of the treatment dose ( Fig 5I and 5J ) . Furthermore , for all tested scenarios coefficient of variation higher than 0 . 5 strictly characterised early dying cells which commit apoptosis within the first 2 hours . Interestingly , receptor clustering enhanced ramp noise resulting in higher values of coefficient of variation ( Fig 5F and 5I ) . Fractional cell killing was observed in DR-targeted treatments especially when applied in low amounts [2] . As we have shown , the high dispersion of the death delays was the main reason for fractional cell killing . What we found more interesting is that dispersion of the delays could exhibit strong bimodality clearly distinguishing between the fraction of early and late dying cells . Clear bimodality was predicted by our model particularly for the low ligand dose upon receptor clustering order ( Figs 6 and 5A ) . Taking this fact together with the ramp noise analysis ( Fig 5F ) we can conclude that the high noise in the ramp sensitises cells for early death which will take place within the first five hours at the latest . This fluctuation-enhanced sensitivity has been called ‘stochastic focusing’ and allows quicker system relaxation to the stationary state when the noise is high . The bimodality breaks when receptor clustering is interrupted ( Fig 6C , see S1 File ) and most of the cells would die only after 10 hours . On the population average cell dynamics receptor clustering provides slightly quicker Casp8 activation for the low dosage of the DL ( Fig 6E ) . This may enable better coupling of this stochastic process with the continuous positive caspase feedback loop . Thus , stochastic focusing coupled with the positive feedback facilitates a more robust bimodal response without the need of multi-stability encoded in the system itself . Finally , the overall cell survival can be dramatically reduced by enhancing the receptor clustering mechanisms ( Fig 6G ) .
The roles of multiprotein signalling platforms assembled upon DR stimulation have been broadly discussed in the context of the programmed cell death initiation [11 , 17 , 29 , 31] as well as proliferation and proinflammatory signalling [5 , 77] over the last decades . The effect of DR and apoptotic inhibitors targeting on the structure and function of these platforms were investigated in different experimental models . However , the mechanism through which these platforms give rise to distinct functions is still poorly understood . Particularly , the mechanism through which the heterogeneous apoptotic response to the DR targeted therapeutics is initiated and how it can explain fractional cell death remains unclear . Our study shows that the noise exclusively caused by the stochastic molecular assembly of the DISC/RIPoptosome platform is able to explain fractional cell killing at low receptor level engagement . Furthermore , this noise in conjunction with receptor clustering facilitates a more rapid apoptotic response . Most of the variability in cell death delay raised upon DR stimulation originates from the pre-MOMP phase . Individually , none of the proteins involved in the apoptosis activation prior to MOMP can explain variation in cell death delays . Casp8 activation rate and consequently the rate of Casp8-dependent BID cleavage are the only determinants of the process [2 , 4 , 78] . Casp8 activation is entirely dependent on the assembly of the multiprotein signalling platform such as RIPoptosome . Though there have been a few models developed none have explicitly accounted for the stochastic nature of the signalling platform assembly [79] . Hence in this study , we developed a novel mathematical model of the stochastic assembly of the RIPoptosome in the single cell together with downstream effector-caspases cascade . Two of these processes are paired together in the pre-MOMP phase of apoptotic pathway initiation . By incorporating the absolute protein concentrations that we have measured in HeLa cells experimentally , and using kinetic parameters derived from the literature we have simulated the Casp8 activation dynamics in the single cell for various conditions: different DL doses , full and disrupted DR clustering propensity . Our modelling simulations have shown that the random and competitive multiprotein assembly of RIPoptosome allows prolonged and slow activation of Casp8 in a ramp-like fashion which is prone to high stochastic fluctuations . Such fluctuations in conjunction with downstream positive feedback loop of effector caspases after certain delay can lead to the spontaneous acceleration of Casp8 accumulation . Because of these fluctuations each cell behaves differently . We have found that the time the single cell will commit to apoptosis depends on the amount of intrinsic noise level in the initial ramp Casp8 activation . The higher ramp noise favours quicker cell death . By that we provide the evidence that the random assembly of RIPoptosome on its own , without any contribution of extrinsic noise in protein expression may explain the heterogeneous cell death response . Our modelling predictions confirm that the receptor clustering process is critical in the extrinsic apoptotic response initiation [80] . Furthermore , a lower DL treatment dose will benefit the most from the enhanced clustering capacity over all . However , the significant fraction of the cell population will remain in the delayed apoptotic state . This new finding is clearly reflected in the bimodality of the distribution of death delays initiated by low DL dose where we demonstrated the clear split of the cell population into early and late responders ( Fig 6A ) . Despite the high affinity of XIAP to Casp3 , their concentration balance in HeLa cell does not ensure robust Casp3 inhibition prior to MOMP [76] . Additionally , XIAP stimulated Casp3 ubiquitination that leads to Casp3 degradation is critical to keeping the downstream executioner caspases cascade shut till the MOMP is set . We have shown that for the fixed XIAP level in HeLa , Casp3 will play an important role in determination of cell death delay . Thus , suppression of the Casp3 ubiquitination/degradation rate at some point can trigger an ultra-sensitive switch from late to early cell death ( Fig 4A ) . This response is characteristic for the low doses of DL and has been suggested in previous modelling studies [68] . However , experimentally Casp3 proteasomal degradation is hard to inhibit unless the catalytic activity of Casp3 is suppressed [71 , 72] . Instead , XIAP inhibition can initiate the same effect ( Fig 4B ) and as we showed very minor suppression is needed to return rapid cell death response initiated by subminimal DL doses . We believe that this ultra-sensitivity serves the best explanation for established success in the application of XIAP specific inhibitors for DL dependant cell death amplification [74 , 75 , 81 , 82] . Strikingly , we found that at the low DL doses an increase in XIAP level exclusively would cause a tremendous linear increase in the time of cell death delay . Indeed , exceptionally only XIAP overexpression , not cIAP1/2 or Smac up and down regulation respectively , is the apoptosis resistance mechanism which can be developed in cancer cells in response to the chemotherapeutics [83] . The content and dynamics of the RIPoptosome assembly predicted by model conform the general knowledge that RIP1 is the most abundant protein among all that are comprising the core RIPoptosome scaffold [10 , 23 , 24 , 54 , 84–86] . The engraftment of ProCasp8 molecules into RIP1 oligomer can happen when the RIP1 filament growth is interrupted by binding of single FADD molecule that occasionally can lead to the sequential binding of ProCasp8 . Our simulations have showed that this event is very rare for a given level of RIPoptosome proteins in HeLa cell and we do not see strong oligomerization of ProCasp8 or its DEDs in HeLa cell . Despite , overexpressed truncated form of ProCasp8 which includes only DED1-DED2 domain is prone to form filamentous structure by oligomerisation [12 , 53 , 87 , 88] , the full length protein do not oligomerize [87–89] . Overall , the quantitative balance between the components may dictate the structure of the RIPoptosome that vary between different cell types [12 , 84 , 85] . Therefore , we can conclude that the RIPoptosome formation in HeLa is a competitive process of RIP1 , FADD and ProCasp8 assembly and the structure and function of this assembly varies due to the noisy nature of the core protein binding and dissociation events . In this context , the slow and probabilistic nature of Casp8 activation explained in our current study by the random RIPoptosome assembly serves as the basis for caution mechanism of kinetic proofreading . This mechanism needs to be in place to verify weak or temporal apoptotic stimuli . The cells which succeeded to assemble the pool of RIPoptosomes that can sustain efficient Casp8 activation will proceed further down the apoptotic pathway triggering MOMP . The high noise in the ramp of Casp8 activation , in this case , will signify the high RIPoptosome efficiency showing that each moment the Casp8 activity is sacrificed the next moment it can be reconstituted or even amplified ( Fig 7 ) . The vulnerability of the apoptotic pathway and its susceptibility to adaptation are currently the key limitation of therapeutics designed to kill cancer cells through the DR targeting therapeutics . In this paper , we have uncovered the original mechanism that explains inefficient cell death stimulation through stochastic activation of apoptosis initiating caspase signalling , leading to heterogeneous responses . We believe that detailed understanding of basic principles of early events of cell death initiation may also stimulate more rationalised approaches in the development of combinational treatments against cancer .
We quantified TNFR1 in HeLa cells by QuantiBRITE phycoerythrin beads based assay . The amount of DR4/5 was calculated from TNFR1 level by comparative MS data analysis ( Table A in S1 File ) . Receptor clustering conformation was calculated from experimentally derived cluster size probability distributions ( S1 Fig ) . Initial protein concentrations were taken from the literature ( Table D in S1 File ) . Except FADD and RIP1 , which we quantified with recombinant protein comparative Western Blot and ProCasp6 concertation that we adjusted using the complete deterministic model ( S1 File ) . Most binding kinetics and catalytic enzymes activity parameters were retrieved from the literature ( Table E in S1 File ) . Hence FRET probe cleavage rate and Casp3 degradation rate were adjusted in the simulations . Modelling formalism of Gillespie stochastic simulation algorithm ( SSA ) and ODE integration as well as semi-stochastic hybrid model was implemented in the MATLAB 2017b environment ( see also S1 File ) . | Death receptors are targets of novel cancer therapeutics . Most of them signal through flexible multiprotein platforms to either activate apoptotic or necroptotic cell death , or propagate cell survival and pro-inflammatory signals . We focused our study on the role of dynamic assembly and composition of these platforms in the initiation of cell death at the single cell level . Since the assembly is slow through the competitive nature of protein binding within the platforms core we developed a stochastic mathematical model of the death inducing signalling platform . Our model provided an explanation for delayed cell death and fractional killing upon the death receptor stimulation . Additionally , we found that the variability in the cell death response arises through the random assembly initiates a slow noise-prone ramp activation of initiator Caspase-8 spontaneously triggering the apoptotic cascade . Our computational simulations predicted high variation in the time required for cell death induction at the single cell level and highlighted a significant role of death receptor clustering in effective Caspase-8 activation . Our knowledge and data driven model captures detailed processes governing the early events of cell death initiation and can be used to guide the development of more rational combinational treatments against cancer . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"death",
"hela",
"cells",
"cell",
"processes",
"biological",
"cultures",
"simulation",
"and",
"modeling",
"fluorophotometry",
"cell",
"cultures",
"research",
"and",
"analysis",
"methods",
"chemical",
"properties",
"physical",
"chemistry",
"fluorescence",
"resona... | 2019 | Heterogeneous responses to low level death receptor activation are explained by random molecular assembly of the Caspase-8 activation platform |
The NF-κB signaling pathway is central to the innate and adaptive immune responses . Upon their detection of pathogen-associated molecular patterns , Toll-like receptors on the cell surface initiate signal transduction and activate the NF-κB pathway , leading to the production of a wide array of inflammatory cytokines , in attempt to eradicate the invaders . As a countermeasure , pathogens have evolved ways to subvert and manipulate this system to their advantage . Enteropathogenic and enterohemorrhagic Escherichia coli ( EPEC and EHEC ) are closely related bacteria responsible for major food-borne diseases worldwide . Via a needle-like protein complex called the type three secretion system ( T3SS ) , these pathogens deliver virulence factors directly to host cells and modify cellular functions , including by suppressing the inflammatory response . Using gain- and loss-of-function screenings , we identified two bacterial effectors , NleC and NleE , that down-regulate the NF-κB signal upon being injected into a host cell via the T3SS . A recent report showed that NleE inhibits NF-κB activation , although an NleE-deficient pathogen was still immune-suppressive , indicating that other anti-inflammatory effectors are involved . In agreement , our present results showed that NleC was also required to inhibit inflammation . We found that NleC is a zinc protease that disrupts NF-κB activation by the direct cleavage of NF-κB's p65 subunit in the cytoplasm , thereby decreasing the available p65 and reducing the total nuclear entry of active p65 . More importantly , we showed that a mutant EPEC/EHEC lacking both NleC and NleE ( ΔnleC ΔnleE ) caused greater inflammatory response than bacteria carrying ΔnleC or ΔnleE alone . This effect was similar to that of a T3SS-defective mutant . In conclusion , we found that NleC is an anti-inflammatory bacterial zinc protease , and that the cooperative function of NleE and NleC disrupts the NF-κB pathway and accounts for most of the immune suppression caused by EHEC/EPEC .
Enteropathogenic Escherichia coli ( EPEC ) and enterohemorrhagic E . coli ( EHEC ) are worldwide causative agents of illness and death [1] . EPEC causes infantile diarrhea , which is often lethal in developing countries , and EHEC is a frequent cause of bloody diarrhea and hemolytic uremic syndrome ( HUS ) even in developed countries [2] . These pathogens are often transmitted in contaminated food . Once they reach the human intestine , the bacteria multiply and colonize on the mucosal surface . These bacteria are also known as “attaching and effacing ( A/E ) ” pathogens due to the histopathological lesions caused by the intestinal colonization [3] , [4] . A/E lesions are characterized by localized damage to the intestinal microvilli and the rearrangement of host cytoskeletal proteins beneath the intimately attached bacterial colonies [5] , [6] . The key virulence factors in A/E pathogens are encoded at the locus of enterocyte effacement ( LEE ) . The LEE , which is required for the formation of A/E lesions during infection , encodes regulators , an adhesin ( intimin ) , chaperones , a translocator , effector proteins , and type III secretion system ( T3SS ) components [6] . In particular , the T3SS , an organelle common to the A/E pathogens , is responsible for delivering bacterial effector proteins directly from the bacterial cytoplasm into the host cytoplasm , where they modify and disrupt host cell functions [6] . An isogenic mutant defective in the T3SS loses the ability to establish successful colonization on host cells , indicating that the T3SS is a major determinant of pathogenicity [7] , [8] . In addition to the seven effector proteins encoded by the LEE , EPEC and EHEC possess a variety of effector proteins that are encoded elsewhere on the genome . The EHEC genome encodes more than 40 effector proteins [9] . and EPEC encodes at least 21 [10] , [11] . Although the functions of several effector proteins have been reported , those of many others have yet to be determined [12] . The intestinal epithelium plays an important role in generating signals in response to pathogen infection to activate cells of the innate and acquired immune systems present in the underlying intestinal mucosa [13] . Mucosal inflammation is characterized by the coordinated expression and up-regulation of a specific set of gene products , including the cytokine and chemoattractant interleukin ( IL ) -8 , and macrophage inflammatory protein 1α [14] . The NF-κB family proteins are key regulators of inflammatory genes; they are structurally similar transcription factors , and include c-Rel , RelB , p65 ( RelA ) , p50 ( p105 processed form ) , and p52 ( p100 processed form ) [15] . In unstimulated cells , homo- or heterodimers of NF-κBs are maintained in an inhibited state in the cytoplasm by their binding with IκB [16] . In an immune-challenged state or during infection , stimulated Toll-like receptors ( TLRs ) initiate a signaling cascade that results in activation of the IκB kinase ( IKK ) complex [17] , [18] . The IKK complex phosphorylates and ubiquitinates IκB , marking it for degradation , which frees the NF-κB dimmers [15] , [19] . The dimers then enter the nucleus and promote the transcription of genes for inflammatory proteins such as IL-8 , IL-1β , and TNF-α [20] . Over the course of EPEC or EHEC infection , the balance between the pro-inflammatory bacterial extracellular components and the proposed anti-inflammatory effector proteins shapes the outcome of the host immune responses [21] , [22] , [23] . At the early stage of infection , the NF-κB-mediated inflammatory response is activated by T3SS-independent mechanisms; at later stages , the response is repressed in a T3SS-dependent manner [18] , [19] . Several studies have indicated that flagellin is a major pro-inflammatory mediator [20] , [21] , [22]; the detection of flagellin by TLR-5 signals the activation of NF-κB to cause inflammation . On the other hand , in differentiated Caco-2 cells infected by EPEC , subsequent stimulation with TNF-α does not activate or cause the nuclear translocation of NF-κB [18] , and this inhibition is independent of the LEE-encoded effectors [19] . Together , these observations suggest that anti-inflammatory activity may be mediated by non-LEE-encoded effector ( s ) . Recently , NleH and NleE have been shown to suppress host NF-κB activation independently , through different mechanisms [24] , [25] , [26] . NleH inhibits the translocation of ribosomal protein S3 ( RPS3 ) into the nucleus , by binding to it in the cytoplasm , thereby reducing the activation of NF-κB/RPS3-dependent promoters [24] . A report on NleE elegantly demonstrated that NleE acts by interfering with the activation of the IKK complex , thereby maintaining the NF-κB dimers in an inhibited state [25] . However , unlike the T3SS-defective EPEC , the host immune response shows residual repression when infected with the ΔnleH or ΔnleE mutant of EPEC . This suggests the existence of an as-yet-unidentified non-LEE effector protein ( s ) . In the present study , by using an artificially created , reconstituted TOB02 strain ( EPI300/LEE+BFP ) that possesses the LEE , bfp operon , and perABC regulators of EPEC , and a series of EPEC deletion mutants , we have identified NleC as a novel effector that suppresses NF-κB activation . We determined that NleC acts by directly cleaving the NF-κB subunit p65 to a form that is degraded by proteasomes . We also showed that an EPEC compound mutant lacking both NleC and NleE was less effective at suppressing NF-κB activation than EPEC deficient in either gene alone ( ΔnleC or ΔnleE ) . Furthermore , infection with the double mutant of EHEC or EPEC leaves the host response toward inflammatory stimulants nearly intact , at the level seen with a T3SS-defective strain .
To evaluate the effect of individual non-LEE effectors on host-cell immune responses , we employed a reconstituted Escherichia coil K12 strain that carries two plasmids , one harboring the full LEE locus and the other bearing a bfp operon for the bundle-forming pilli and the perABC regulatory genes of EPEC B171-8 ( hereinafter referred to as TOB02 ) ( Fig . S1 ) . The LEE locus encodes the injectisome of the type III secretion ( T3S ) apparatus and essential virulence factors such as Tir and Intimin , which are necessary for intimate attachment , a hallmark of the EPEC/EHEC adherence [2] . To confirm that TOB02 was capable of colonization and establishing intimate attachment , we infected HeLa cells with it and looked for the reorganization of F-actin beneath the attached bacteria , a characteristic result of the intimate attachment that is dependent on virulence factors encoded by LEE . Following the infection , the cells were stained with DAPI and Rhodamine-phalloidin . Similar to infections of EPEC , microcolonies formed within 30 min of the infection , and an intense accumulation of F-actin was observed beneath the bacterial colonies ( Fig . 1A ) . To confirm that the reconstituted TOB02 bacterium expressed a functional T3S system ( T3SS ) , we assayed the secretion of a LEE-encoded protein , EspB , which is secreted through the T3SS [25] . We probed the culture supernatant of the TOB02 cells with an anti-EspB antibody , and detected EspB in an immunoblot ( Fig . 1B ) . In addition , we also compared the amount of Tir translocated into host cells between the native EPEC and TOB02 strains ( Fig . S2 ) . Taken together , these results indicate that the T3SS of TOB02 is functional and that this strain can establish an intimate attachment to epithelial cells . We next analyzed the inflammatory response of host cells upon TOB02 infection . As reported previously , the detection of PAMPs ( pathogen-associated molecular patterns ) by TLRs on the cell surface triggers the inflammatory response , which is suppressed by EPEC/EHEC during the early stage of colonization . Because T3SS-defective EPEC/EHEC cannot suppress the host immune response , the suppression of the host inflammatory response has been proposed as a T3SS-dependent phenomenon [22] , [23] , and therefore effectors on the LEE locus or non-LEE loci have been predicted to downregulate the host immune responses following the T3SS injection [22] , [23] . To compare the response between cells infected with TOB02 and wild-type EPEC , we infected HeLa cells with TOB02 or with wild-type EPEC strains E2348/69 and B171-8 and their ΔT3SS mutants ( E2348/69 ΔescF and B171-8 ΔescC ) . Following the infection , the cells were stimulated with heat-killed E . coli ( HKE ) , and the amount of secreted IL-8 was measured . We found that cells infected with TOB02 released a large amount of IL-8 ( Fig . 1C ) . Although TOB02 possesses the LEE-encoded effectors , the lack of a marked reduction in the inflammatory response of the TOB02-infected cells indicated that effectors residing on the LEE locus may not be very important in the host inflammatory suppression . Therefore , we used the TOB02 strain to determine which effectors on non-LEE loci could cause reduced immune responses upon infection . Since EPEC and EHEC are closely related pathogens that suppress host immune responses [22] , [23] , we postulated that the T3SS-dependent anti-inflammatory effector proteins might be well conserved between these pathogens . We therefore compared the repertoires of the T3SS-dependent non-LEE effector proteins of EPEC ( E2348/69 and B171-8 ) [10] , [11] and EHEC ( O157:H7 Sakai , O26 , O111 , O103 ) [9] , [27] , [28] , and established strains of TOB02 carrying each of these shared non-LEE-effector genes ( espF , espG , espJ , espK , espO , espZ , Map , nleA , nleB , nleF , nleH1 , nleG , nleC , nleE , and espH ) . HeLa cells were infected with each strain followed by stimulation with HKE , and the IL-8 released into the medium was measured . Strains that caused a significantly lower secretion of IL-8 in comparison with the empty-vector control ( TOB02/HA ) were considered effective . Using this method , we identified NleC and NleE as effectors that could greatly decrease IL-8 secretion by the host cells ( Fig . 2A ) . Recently , Nadler et al . and Newton et al . reported that NleE blocks NF-κB signaling by inhibiting TAK1/IKK activation [25] , [26] . Thus , the identification of NleE using the TOB02 system supported the validity of our screening method . We therefore focused on analyzing NleC . We first confirmed the secretion of NleC-HA by TOB02/nleC-HA ( Fig . S3 ) . Because IL-8 is a downstream target of the activated NF-κB pathway , we used an NF-κB reporter assay to verify that TOB02/nleC-HA could interfere with the NF-κB signaling cascade . HeLa cells transfected with reporter plasmids that harbored a NF-κB-regulated SEAP gene and a constitutively active luciferase gene , were infected with either TOB02/HA or TOB02/nleC-HA . After 3 hours of infection , the cells were further stimulated with HKE and assayed for the NF-κB reporter activity . As shown in Figure 2B , the NF-κB activity in the cells infected with TOB02/nleC-HA was significantly lower than in the cells infected with TOB02/HA . These results suggested that NleC suppressed the NF-κB activation , which led to the reduced secretion of IL-8 by the host cells . In our effort to screen for potential immune-suppressive non-LEE effectors , we devised another system using a series of EPEC mutants . These mutants , designated TOE-A1 to -A7 , were designed to lose a cluster of effector genes in each specific horizontally transferred element ( IE: integrative element , and PP: prophage ) in a step-wise and additive fashion ( Table S1 ) . We used this approach to identify effectors with similar anti-inflammatory effects . We infected HeLa cells with these mutants , and the activity of the NF-κB reporter was determined after stimulating the cells with HKE . In this experiment , TOE-A4 lost most of the host immune suppression activity ( Fig . 2C ) . TOE-A4 was derived from TOE-A3 by the deletion of a cluster of effector genes in IE6 , which contains three effector genes , nleE , nleB1 , and espL . Thus , TOE-A4 lacked nine effector genes ( nleB2 , nleH1 , espJ , nleG , nleC , nleD , nleE , nleB1 , and espL: Table S1 ) . Since NleE is required but insufficient for the full repression of NF-κB activation by EPEC , additional effectors were predicted to be involved [25] . Furthermore , host cells infected with TOE-A5 to -A7 did not show a significantly greater recovery of their NF-κB activity than cells infected with TOE-A4 or ΔescF mutant , suggesting that none of the non-LEE effector proteins besides the ones missing from TOE-A4 contributed significantly to the disruption of the NF-κB signaling cascade . In another words , the full de-repression of the NF-κB activity by EPEC can be attributed to the combined loss of NleE and one or more of the protein ( s ) missing from the TOE-A4 mutant . From our primary screening with the TOB02 system , we ruled out nleH1 , espJ , nleG , nleD , espL , or nleB , because they were unable to suppress the host inflammatory response to HKE ( data not shown ) . Therefore , we examined EPEC TOE-A4 harboring an nleC-expressing plasmid for the ability to suppress NF-κB activity . We found that TOE-A4/nleC restored the suppression of the host NF-κB activity to a level near that achieved with TOE-A4/nleE or wild type ( Fig . 2D ) . Taken together , our results strongly indicated that NleC is an immune suppressor with activity comparable to NleE in a complementary setting that may act in parallel to or be redundant with NleE . To determine whether the function of NleC required other bacterial co-factors , we constructed a mammalian expression vector bearing a fusion protein , eGFP-NleC , and introduced it into HeLa cells together with the NF-κB reporter plasmid . After stimulating the cells with HKE , the reporter activities were determined . As shown in Figure 3A , whereas the non-transfected ( NF-κB reporter plasmid only ) and empty-vector controls showed high NF-κB activity in response to the HKE stimulation , the eGFP-NleC-expressing HeLa cells showed markedly lower NF-κB activation . Taken together , these results suggested that NleC is a negative regulator of the host NF-κB signaling pathway , and that it does not require bacterial co-factors for its function . The NF-κB signaling cascade is divided mainly into cytoplasmic and nuclear portions [15] . To elucidate where NleC functions to interfere with this cascade , we examined its localization after being injected into host cells via T3SS . After 3 hours of infection with TOB02/nleC-HA , the cells were fixed and stained with anti-HA/anti-mouse IgG-FITC , Rhodamine-Phalloidin , and DAPI . Most of the NleC-HA was apparently in the cytoplasm , although the strongest signals were observed beneath the microcolonies ( Fig . 3B ) . We also found that NleC-HA showed the same distribution pattern in host cells infected with wild-type EHEC , via the endogenous T3SS ( Fig . S4 ) . These results suggested that NleC targets one or more cytoplasmic factors in the host NF-κB pathway . Since NleC was determined to be predominantly cytoplasmic , and since various upstream signals converge to modify the IκBα/p65/p50 complex , which serves as a signaling hub , we decided to first examine changes that might occur at this complex . To do this , HeLa cells were infected with either TOB02/HA or TOB02/nleC-HA bacteria . After 3 hours of infection , the HeLa cells were further stimulated with HKE for 20 and 40 min , before being collected for western blot analysis . Using antibodies specific to the N-terminal region of p65 ( p65 N-term ) , phospho-p65 , and IκBα , we detected a significant down-regulation of p65 in cells infected with TOB02/nleC-HA ( Fig . 3C ) . Compared to the level in cells infected with TOB02/HA , the amount of total p65 as well as the phospho- ( active ) p65 was lower in the TOB02/nleC-HA-infected cells even before the HKE stimulation . These results suggested that NleC may either directly or indirectly target p65 to decrease NF-κB signaling . Based on this result , we speculated that NleC promotes the degradation of p65 in the cytoplasm . The ubiquitination-proteasome pathway is involved in selective protein decomposition , and several studies have shown that p65 is poly-ubiquitinated and targeted for proteasomal degradation [29] , [30] , [31] , [32] . We therefore tested whether MG132 , a reversible inhibitor of proteasomes , could prevent the p65 degradation in the presence of NleC . HeLa cells pretreated with either DMSO or MG132 ( 5 µM ) were infected with TOB02/HA or TOB02/nleC-HA for 3 hours . As shown in Figure 3D , using the p65 N-terminal specific antibody , we found that the total p65 decreased substantially regardless of DMSO or MG132 treatment in the TOB02/nleC-HA infected cells . Next , we used a C-terminal-specific antibody to detect p65 , which at first glance confirmed the reduction of the protein . However , upon closer examination , we noticed a smaller fragment in addition to the full-length p65 that was specific to the TOB02/nleC-HA-infected cells ( Fig . 3D; middle panel ) . Moreover , treatment with MG132 or lactacystin resulted in an accumulation of this smaller fragment ( Fig . S5 ) . This observation suggested that NleC may be directly or indirectly involved in generating this cleaved product , which is targeted for proteasomal degradation . Thus , our findings strongly suggest that NleC injection could trigger the reduction of NF-κB activity by decreasing the amount of p65 in infected cells . Since the smaller p65 ( hereafter referred to as p65-C ) could only be detected by the C-terminal-specific antibody , and the size difference between the full-length and cleaved form differed by only several kiloDaltons , we predicted that the cleavage site would be located at N-terminus . To determine the actual cleavage site , we extracted p65-C and analyzed its amino acid sequence from the N-terminal end . We found that the first four amino acids matched the 11th to 14th residues of the full length p65 . Therefore , we concluded that the cleavage occurs between the 10th and 11th amino acid of p65 . Coiras et al . reported that activated caspase-3 can mediate the cleavage of p65 at its N-terminal region to generate a C-fragment in non-apoptotic T-lymphocytes [33] . To examine the possible involvement of caspase-3 in the cleavage of p65 in cells infected with nleC-expressing bacteria , we pretreated HeLa cells with either DMSO or z-VAD-fmk ( a pan-inhibitor of caspase ) . However , the inhibition of caspase-3 activity with z-VAD-fmk did not affect the generation of p65-C ( Fig . S6 ) . To elucidate the relationship between NleC and p65-C , we carried out in vitro cleavage assays to recapitulate the observation made in TOB02/nleC-HA-infected HeLa cells . We constructed and purified a fusion protein of NleC with GST attached to its N-terminal end ( GST-NleC ) . To confirm the activity of the purified protein , cell lysates prepared from unstimulated HeLa cells were incubated with the purified GST-NleC . Using N- and C-terminal-specific antibodies to p65 , we detected a decrease in p65 and the appearance of p65-C only in the mixture containing GST-NleC ( Fig . 4A ) . Next , to determine the dependency of this cleavage on host co-factors , purified p65 was mixed with GST-NleC . While both the p65 only and GST-added control samples showed no change in total p65 nor the detection of p65-C , the GST-NleC-containing mixture had less p65 , and p65-C was readily detected ( Fig . 4B ) . Based on these observations , we concluded that NleC could mediate the direct digestion of p65 . We next performed a bioinformatic search for any known domains in NleC . The search predicted a zinc protease domain located between amino acids 183 and 187 . This amino acid sequence , HEIIH , corresponds to the consensus sequence HExxH , where x is any amino acid . To test if this domain was responsible for mediating the cleavage of p65 , we performed site-directed mutagenesis in which we replaced the second histidine with tyrosine ( H187Y ) ( Fig . 4C ) . Since both histidines of HExxH come in contact with zinc ions , the mutation of either residue should disrupt the domain and render it non-functional [34] . We mixed purified p65 with GST , GST-NleC wild type ( GST-NleC ) , or GST-NleC mutant ( GST-NleCmut ) in vitro . As shown in Figure 4D , although GST-NleC produced p65-C , GST-NleCmut did not . Moreover , 10 mM EDTA greatly impeded the function of GST-NleC , suggesting that divalent metal cations are necessary for its activity ( Fig . 4D ) . Finally , we used the NF-κB reporter to assay the necessity of the zinc protease domain and the protease activity of NleC in the NF-κB suppression . HeLa cells were infected with TOB02/HA , TOB02/nleC-HA , or TOB02/nleCmut-HA for 3 hours , stimulated , and then reporter activities measured . As shown in Figure 4E , whereas TOB02/nleC-HA significantly suppressed the NF-κB activation , TOB02/nleCmut-HA failed to do so , indicating that the mutant had lost the anti-inflammatory effect of intact NleC . Taken together , these results showed that the HEIIH region of NleC is a zinc protease domain necessary for cleaving p65 and suppressing the host NF-κB activity . Using the reconstituted TOB02/nleC-HA strain , we discovered that NleC functions in host cells as a p65-targeting enzyme . Since the expression of nleC in the TOB02 strain is activated through an IPTG-inducible promoter , we wanted to know if we could observe a similar enzymatic activity when NleC was at its physio-pathologically relevant level , in infections with native strains of EPEC or EHEC . Therefore , we infected HeLa cells with wild type ( WT ) , ΔnleC , or ΔnleC/nleC-HA EPEC for 2 h , and analyzed the amount of cytoplasmic and nuclear p65 in the cells using an anti-p65 ( N-terminus ) antibody after treatment of cells with HKE . We observed a reduction in cytoplasmic p65 in the WT-infected cells compared to cells infected with the ΔnleC mutant strain ( Fig . 5A; left panel ) . As for nuclear p65 , the amount of p65 was significantly reduced in both wild type and ΔnleC mutant infected cells compared to that of ΔescF ( Fig . 5A; right panel ) . Next , we also analyzed the appearance of p65-C fragment in wild type-infected cells . When p65 in total lysate was detected using anti-p65 ( C-term ) , we identified the generation of p65-C fragment in wild type infected cells , albeit in a lesser amount of that by ΔnleC/nleC-HA strains ( Fig . S7 ) . Taken together , we showed that NleC also functions as a protease targeting p65 in native strains . These results suggested that , even when NIeC was at physio-pathological levels , it could cleave p65 , although the effect was less profound than that of the nleC complemented strain , presumably due to the difference in the amount of NleC present in the host cells . After determining that NleC was an anti-inflammatory effector and identifying its functional zinc protease domain , we next examined whether NleC contributes to the immune suppression by EPEC/EHEC . For this , we generated an equivalent set of mutant strains in EPEC and EHEC , and measured the IL-8 secretion of cells infected with the EPEC/EHEC mutant . As shown in Figure 5B , the ΔnleC mutant of EPEC showed a comparable suppressor activity to the wild type EPEC . On the other hand , we observed statistically significant elevation of IL-8 secretion from the ΔnleC mutant of EHEC-infected cells compared to the WT EHEC-infected ones . We also infected HeLa cells with the ΔnleE mutant of EPEC and EHEC and observed increased IL-8 secretion , which is in agreement with the reported results of EPEC∶ nleE by Nadler et al and Newton et al [25] , [26] . Although the IL-8 response by the host cells after the ΔnleC or ΔnleE mutant challenges was better , the inflammatory response was still only partial compared to that permitted by the ΔescF or ΔescD mutant . Based on the proposed actions of NleE and NleC as negative regulators of IKK activation and homeostasis of p65 , respectively , and the observation that full host NF-κB activity was preserved during infection with TOE-A4 , we predicted that NleE and NleC contributed the major portion of the suppression of pathogen-induced activation of NF-κB pathway . We therefore analyzed the total IL-8 secreted from HeLa cells infected with the ΔnleC ΔnleE ( double knock-out ) mutant . As shown in Figure 5B and 5C , whereas the wild type substantially inhibited the IL-8 secretion in infected cells , the infection of isogenic ΔnleC ΔnleE compound mutant resulted in a significantly diminished inhibition of IL-8 secretion compared to WT . Furthermore , the level of the IL-8 response exhibited by host cells infected with the compound mutant was comparable to that of cells infected by the T3SS defective mutant ( ΔescF of EPEC and ΔescD of EHEC ) . The introduction of the nleCwt but not nleCmut expression plasmid into the ΔnleC ΔnleE mutant effectively restored the inhibition of IL-8 secretion by the infected cells . This major de-repression of the host NF-κB activation in the ΔnleC ΔnleE mutant suggested that these two T3SS-dependent , non-LEE effector proteins function in concert for the suppression of the NF-κB-signaling cascade and that the zinc protease domain is necessary for the anti-inflammatory function of NleC ( Fig . 6 ) .
Bacterial antigens of infecting pathogens can trigger intense local inflammation and the mobilization of immune cells . To prevent their early elimination , the EPEC and EHEC A/E pathogens control the host response by delivering anti-inflammatory bacterial proteins into the host cells . Although the recent identifications of NleH1 NleE and NleB showed that EPEC/EHEC use different mechanisms to modify the host immune response [24] , [25] , [26] , those mechanisms did not completely account for the full immune-suppressive potential of these pathogens . Here , we found that NleC of EHEC and EPEC negatively modulates the host inflammatory response by reducing the amount of the mammalian p65 subunit of NF-κB . Specifically , this reduction is due to the direct cleavage of p65 by NleC , followed by proteasomal degradation of the cleaved p65 . We also demonstrated that two type III secretion effectors , NleC and NleE , are necessary for and function in cooperation to achieve most of the immune suppression caused by these pathogens . By combining these effectors , EPEC and EHEC efficiently inhibit the activation of the NF-κB signaling pathway in infected host cells , thereby weakening the inflammatory response . With the aim of discovering putative anti-inflammatory T3SS-dependent , non-LEE-encoded effector proteins , we employed both gain-of-function and loss-of-function approaches . In the gain-of-function analysis , we used a reconstituted bacterial system to study the effect of individual effector proteins on the host innate immune response . This method is preferable to transfection , because these artificial A/E bacteria bearing selected non-LEE effector genes faithfully recapitulated the infection and colonization of host cells in a T3SS-dependent manner . Moreover , because they expressed BFP , the reconstituted strain established colonies as quickly and efficiently as wild-type EPEC . As a result , the pathological action of the effectors could be compared over a similar time course with that of the native A/E pathogens . Another reason we avoided using transfection is that the expression levels may vary from cell to cell , and the transfection efficiency is difficult to maintain at a high level for all candidate genes , leading to possible false read-outs . Finally , the reconstitution method is also appropriate for isolating effectors with redundant functions . However , since amount and stability of translocated effector can not be controlled to be the same level , we can not rule out the presence of other effector ( s ) involving in modulation of inflammatory response by only this screening . Nevertheless , the use of this reconstituted system revealed not only NleC as a novel anti-inflammatory factor , but also the already-reported NleE [25] , [26] . In the loss-of-function analysis , we used a set of EPEC mutant constructs with serial deletions of the pathogenic islands . This approach allowed the identification of multiple cooperative effectors that may have similar functions . Using this series of deletion mutants , we identified the EPEC mutant TOE-A4 , which lacked four horizontally transferred elements that included nine non-LEE effector genes , and was as incapable of inhibiting NF-κB activation as the T3SS-deficient mutant . This complete loss of immune suppressing activity was also seen when NleE-NleB were deleted from TOE-A3 . In contrast , Nadler et al . showed only a partial alleviation of the suppressed host inflammatory response using an EPEC ΔnleE ΔnleB double mutant [25] . The inconsistency between our findings in TOE-A4 and theirs may lie with the genes that had already been deleted from TOE-A3 . Because we ruled out espJ , nleF , nleG , nleH1 , nleD , espL , and nleB , based on our primary screening using the TOB02 strains , we speculate that the effectors responsible for most of the immune-suppressive effect exerted by EPEC or EHEC in response to HKE are NleC and NleE . Indeed , we showed that compound EPEC and EHEC mutants lacking both NleC and NleE ( ΔnleC ΔnleE ) were incapable of suppressing the host IL-8 secretion to a similar degree as the ΔT3SS ( ΔescF ) mutant after being stimulated with HKE , which engaged and activated mainly TLR-/IL-1β -NF-κB signaling pathways . The introduction into this compound mutant of either an nleC- or nleE- expressing plasmid restored the mutant's ability to suppress NF-κB activation . On the other hand , NleH1 , another reported anti-inflammatory effector was not identified in these screenings . For verification , we generated an EHEC mutant deficient in NleH1 ( ΔnleH1 ) and tested its ability to suppress the host IL-8 secretion . We did not detect any apparent difference in effect between the wild type and ΔnleH1 ( see Fig . S8 ) . Therefore , the use of the EPEC mutant series not only helped us narrow down the list of candidate effectors substantially , but also , in conjugation with the TOB02 system , allowed us to identify NleC as a novel anti-inflammatory non-LEE effector . Since IL-8 production is dependent on NF-κB activation , we examined the NF-κB signaling pathway to understand the mechanism of the NleC-dependent anti-inflammatory effect . We found that cells infected with NleC-expressing bacteria had markedly reduced active ( phosphorylated ) p65 as well as total p65 , and that this reduction was due to the cleavage of p65 by NleC and the degradation of the remaining p65 fragments by proteasome . Thus far , caspase-3 is the only reported host factor that can cleave p65 similarly at the N-terminal region [33] . However , we showed that this cleavage event was not due to caspase-3 , because the reduction of p65 was not prevented when the cells were pre-treated with a pan-caspase inhibitor . On the other hand , our bioinformatic search for known domains in NleC suggested the presence of a region ( 183HEIIH187 ) corresponding to a known zinc protease domain . By in vitro cleavage assays , we showed that this zinc protease domain of NleC is crucial for the processing of p65 and that the addition of a metal ion chelating agent , such as EDTA , abrogated the wild-type NleC function . Furthermore , it appears that the protease activity of NleC is selective as it could not digest p50 , another common NF-κB subunit nor IκB ( see Fig . S10 ) . This inability to digest IκB suggests that the reduced IκB in cells infected with NleC-expressing bacteria may be of indirect consequences . These results clearly indicated that NleC is a metalloproteinase with an HEIIH domain that is necessary for its activity . Furthermore , the H187Y mutant of NleC lost its cleavage activity and the ability to suppress NF-κB activation . Therefore , we conclude that the p65 inhibitory function of NleC relies on this zinc protease domain . As the NF-κB signaling pathway consists of multiple components , pathogens have evolved multiple strategies to subvert this system . For example , OspG and OspF of Shigella spp . respectively inhibit the ubiquitination of IκBα and modify the epigenetic information on the promoters of NF-κB-associated transcriptions [35] , [36] . In Chlamydia , CT441 is a T3SS-secreted protease that can cleave p65 at its C-terminal region , generating p40 and p22 fragments [37] . Whereas p22 is degraded by proteasomes , p40 appears stable and inhibits the NF-κB activity when over-expressed [37] . Although CT441 and NleC share the same host target , these two effectors exhibit completely different pathological kinetics , possibly owing to differences in protease structures and their target sequences . We are currently attempting to elucidate the site of cleavage by NleC . Whereas NleC functions relatively early during bacterial colonization , CT441 acts at the mid-to-late stage of Chlamydia infection . However , since it is degraded by proteasomes upon being generated , it is unlikely that p65-C could exist long enough in the cell to inhibit NF-κB activity . Thus , NleC seems to be a novel type of virulence factor/effector which is evolved independently on these other NF-kB-targeting factors . In our study , we showed that EHEC ΔnleC but not EPEC ΔnleC has statistically significant relief on host IL-8 response; this discrepancy between these two close-related pathogen may due to differences in their infection efficiency and amount of other translocated anti-inflammatory effectors , such as NleE or NleB . Nevertheless , the partial preservation of IL-8 secretion in cells infected with either the ΔnleC or ΔnleE mutant compared to its full preservation in cells infected with a ΔT3SS mutant shows that multiple bacterial effector proteins are needed for full inflammatory suppression . Based on our current understanding that NleE and NleC target different components of the NF-κB pathway , we believed that the contributions of NleE and NleC to the immune suppressiveness of EHEC/EPEC are additive . Indeed , when EPEC/EHEC mutants lacking both NleC and NleE ( ΔnleC ΔnleE ) were tested in the infection assay , the cells responded with a level of IL-8 secretion that was as high as in cells infected with a ΔT3SS ( ΔescF or ΔescD ) mutant . This near-total loss of inflammation suppression was also seen with EPEC TOE-A4 , which is deficient in both nleC and nleE genes . These results demonstrated that EPEC/EHEC utilizes mainly these two effectors to suppress NF-κB activation triggered by TLR-/IL-1β pathways . The mode of actions of NleE and NleC in NF-κB interference is becoming clear . NleE was shown to stabilize IκB and retain p65 in the IκB-p65 complex even after stimulation with TNF-α [25] , [26]; however , the precise role of NleE as a direct or indirect regulator remains to be determined . On the other hand , NleC cleaves the N-terminus of p65 , and thereby triggers p65's proteasomal degradation . Not only can NleC digest free p65 , but when NleC is present in large amounts , it can also digest IκB-bound p65 ( see Fig . S9 ) , strongly suggesting that NleC can function whether p65 is in an inhibited or free state . However , considering NleE's ability to cause the retention of IκB ( i . e . increasing the IκB-bound p65 ) , we speculate that NleC cleaves released p65 more efficiently than the p65 bound by IκB . This is supported by our finding that little p65-C was observed after inhibiting the proteasome activity of a wild-type-infected cells . Since the expression of NleC in the reconstituted or complemented TOE-A4 strains showed a similar competency to reduce NF-κB activity as the strain expressing NleE , NleC is a potentially efficient inhibitor of NF-κB activation . Thus , in the wild-type pathogen , NleC , functioning downstream of NleE , may act by preferentially targeting and decreasing the number of active p65 molecules released by activated IKK that escapes the NleE inhibition ( Fig . 6 ) . In our study , we did not identify NleB . As proposed by Nadler et al . , NleB functions as an accessory factor to enhance inflammatory suppression of NleE in infected cells that were challenged by TNF-α [25]; and Newton et al . further demonstrated that NleB could also act independently to suppress TNF-α but not IL-1β induced activation of NF-κB [26] . Based on their studies , it has been suggested that NleE works either directly or indirectly targeting the activation of IKK complex and IκB degradation and that NleB interferes the upstream components of TNF-α signaling pathway . The likely explanation for our inability to identify NleB may due to the use of different stimulant than theirs , i . e . the heat-killed bacteria , to provoke the second inflammatory response following bacterial infection of HeLa cells . As HKE contain bacterial compounds , such as LPS , flagellin , unmethylated bacterial DNA , and RNA , these mainly trigger the activation of NF-κB via TLR-/IL-1β associated pathways . Therefore , the use of TNF-α in our screenings will likely yield similar results as presented by Nadler et . al and Newton et al . [25] , [26] . Nevertheless , as epithelial cells first come in contact with antigens of bacterial components at the onset of infection , the anti-inflammatory NleE and NleC at the level of IKK/IκB/NF-κB cascade are important for executing successful infection by EPEC/EHEC . Several studies have examined the in vivo role of nleC using animal models and found no clear evidence showing the necessity of NleC in the bacterial colonization [38] , [39] . As shown in our in vitro experiments , deletion of nleC gene alone shows only little or no difference in suppression of inflammatory response . This explains the reason of no apparent effect of nleC deletion in colonization in animal models . Although we are yet to provide in vivo evidence of ΔnleC ΔnleE double mutant , it is speculated that this mutant would be highly attenuated and causes early elimination by the host due to the increase of recruited immune cells to the sites of infection . In conclusion , we showed that NleC negatively modulates the host NF-κB activity by the direct enzymatic digestion of p65 . We also demonstrated that NleE and NleC function in concert to interfere with the NF-κB pathway , and that these two molecules are responsible for anti-inflammatory effect of EPEC/EHEC . The presence of alternative factors for modulating the NF-κB activation pathway in EPEC/EHEC indicates that the manipulation of cell signaling must be important for successful infection . The outcome of the inflammatory response to infection depends on multiple factors contributed by both the host and the pathogen . Our study showing how NleE and NleC interfere with the host innate immune response not only illustrates the importance of the NF-κB pathway , which functions at the center of the confrontation , but also broadens our understanding of the intricate interplay between pathogen and host .
The bacterial strains and plasmids used in this study are described in Table S1 . Primers used for cloning are described in Table S2 . DNA fragments containing nleC ( ECs0847 ) , nleE ( ECs3858 ) , espL2 ( ECs3855 ) , and nleB1 ( ECs3857 ) were amplified directly by PCR from EHEC Sakai chromosomal DNA ( Accession No . NC_002695 ) . The PCR products were subcloned into the pHA-CTC plasmid ( Table S1 ) , resulting in pHA-NleC , pHA-NleE , pHA-EspL , and pHA-NleB . The NleC ( ECs0847 ) PCR fragment was also subcloned into pGEM-T ( Promega ) , pEGFP-C1 ( Clontech ) , and pGEX-6P ( GE Healthcare ) to yield pGEM-T-NleC , pEGFP-NleC , and pGEX-6P-NleC . For the site-directed mutagenesis generating H187Y in NleC , pGEM-T-NleC was used as the template and amplified with primers designed to change the second histidine to tyrosine , yielding pGEM-T-NleCmut . The NleCmut fragment was excised and cloned into pGEX-6P to obtain pGEX-6P-NleCmut . The same fragment was also subcloned into pHA-CTC to generate pHA-NleCmut . To construct pET21A-RelA , RelA was excised from pEv3s-T7-RelA ( obtained through Addgene , plasmid # 21984 ) and cloned into the pET21A plasmid . The DNA sequences were verified by sequencing . The primers used are listed in Table S2 . The human cervical cancer cell line HeLa was maintained in MEM ( Sigma ) supplemented with 10% FCS ( Sigma ) and 0 . 1 mM non-essential amino acids ( Invitrogen ) . Request for EPEC strains with a series of deletion and recombinant E . coli K12 strain harboring LEE and bfp ( TOB01 and TOB02 ) should be sent to Tetsuya Hayashi of University of Miyazaki . We constructed a series of mutants from EPEC strain E2348/69 ( Table S1 ) as described previously by Sekiya et al . ( 2001 ) [40] with some modifications . For example , to construct an nleC deletion mutant ( strain TOE-S1 ) , we amplified upstream and downstream regions of the nleC genes ( about 700-bp each and including a short sequence encoding the N- or C-terminal part of the protein ) by PCR using the nleC_B1F ( 5′-aaaaagcaggctTCTATCGGGAAGATGTTGA-3′ ) /nleC_B1R ( 5′-TGCAAAGACGAATCATCGCATGTTTATATCTAATACCCT-3′ ) and the nleC_B2F ( 5′-CGATGATTCGTCTTTGCA-3′ ) /nleC_B2R ( 5′-agaaagctgggtGATTCAATAGCATTCAGGAG-3′ ) primer pairs , respectively . As the nleC_B1R primer contained an 18-base sequence complementary to the nleC_B2F primer sequence ( underlined ) , the resulting two PCR products shared an identical 18-base sequence at their right and left ends . By the joint-PCR method [41] using the two PCR products as a template and the adapt-F ( 5′-GGGGACAAGTTTGTACAaaaaagcaggct-3′ ) /adapt-R ( 5′-GGGGACCACTTTGTACAagaaagctgggt-3′ ) primer pair , we obtained a chimeric PCR product ( referred to as ΔnleC fragment ) consisting of the upstream and downstream sequences of the nleC gene . At this stage , an in-frame deletion was introduced into the target gene . Using the adaptor sequences in the adapt-F and -R primers ( indicated by lower-case letters ) , we cloned the ΔnleC fragment into the pDONR201 entry vector by BP clonase II ( Gateway cloning system: Invitrogen ) . The ΔnleC fragment-containing pDONR201 derivative and the Not1/NcoI double digested pABB-CRS2 vector were incubated with LR clonase II ( Invitrogen ) to transfer the ΔnleC fragment from pDONR201 to pABB-CRS2 ( a R6K-derived positive suicide vector ) . The ΔnleC fragment-containing pABB-CRS2 derivative was introduced into E . coli SM10λpir and then transferred to strain E2348/69 by conjugation . The transductants were first screened on M9 minimum plates containing 0 . 8% glucose ( host marker ) and ampicillin ( Ap; 50 µg/ml ) to obtain clones in which the pABB-CRS2 derivative was integrated into the targeted chromosomal region of E2348/69 . These clones were grown on LB plates containing 5% sucrose to obtain clones in which the pABB-CRS2 plasmid was eliminated by homologous recombination . Finally , among the Ap-sensitive and sucrose-resistant clones , we screened for clones that contained an in-frame deletion in the nleC gene by PCR and sequencing analysis using the nleC_ckF/nleC_ckR primers . Other mutants were also constructed from E2348/69 or its derivatives by the same method using the primers listed in Table S2 . Since multiple T3SS effector genes are present in a cluster on many of the prophages ( PPs ) and integrative elements ( IEs ) in E2348/69 , we deleted these gene clusters en bloc ( strains TOEA1 to TOEA7 ) . A fosmid library of EPEC strain B171-8 was constructed using the CopyControl Fosmid Library Production Kit ( Epicentre , Madison , WI ) , as described previously [10] . Fosmid clones containing part or all of the LEE element were screened by PCR using the eae gene-specific primers ( SK1 and SK2 in Table S2 ) . The end sequences of all the eae-positive clones were determined using the FosF and FosR primers ( Table S2 ) , to select a clone containing the entire LEE element of B171-8 ( referred to as pTOK-02 ) . We then introduced a plasmid named pTOK-01 , which contained the bfp operon and the perA-C locus , encoding the genes for bundle forming pilus biosynthesis and a positive regulator ( perC ) for LEE gene expression , respectively , into an EPI300-T1-derivative containing pTOK-02 , to obtain strain TOB02 . To construct plasmid pTOK-01 , we purified the pB171 plasmid from B171-8 using the Qiagen Plasmid Midi kit ( Qiagen , Tokyo , Japan ) and digested it with SmaI . A 22-kb fragment containing the bfp operon and the perA-C locus was separated by Pulsed-field gel electrophoresis ( PFGE ) , extracted from the PFGE gel using beta-agarase I ( Nippon Gene Co . , LTD , Tokyo , Japan ) , and cloned into the EcoRV site of pWKS130 [42] using the BKL kit ( Takara Bio Inc . , Shiga , Japan ) . The recombinant plasmid ( pTOK-01 ) was introduced into E . coli strain DH10B ( Electrocomp GeneHogs E . coli DH10B; Invitrogen ) by electroporation . Finally , the pTOK-01 plasmid was purified using the QIAprep Spin Miniprep kit ( Qiagen ) and used to construct strain TOB02 , described above . Strain TOB01 was constructed by introducing a pCC1FOS fosmid vector ( constructed by self-ligation ) and the pTOK-01 plasmid into E . coli strain EPI300-T1 by electroporation . More details of the genetic structures of the inserted fragments in the pTOK-01 and pTOK-02 plasmids are shown in Figure S1 . TOB02/HA and TOB02/nleC-HA were cultured in Luria-Bertani broth overnight with constant agitation at 37°C . Bacterial cultures at stationary phase were diluted 100-fold in serum-free DMEM ( Sigma ) and cultured with 120× rpm agitation at 37°C until the O . D600 reached 1 . 0 . The bacterial cultures were separated into the cell pellet ( whole cell ) and the culture supernatant ( supernatant ) fractions by centrifugation at 8 , 000 rpm . The supernatant was filtered ( pore size , 0 . 22 µM ) and concentrated by the addition of trichloroacetic acid ( Sigma ) and deoxycholic acid ( Wako ) to a final concentration of 6% and 0 . 05% , respectively . The precipitates were re-dissolved in acetone ( Wako ) and centrifuged . Finally , after removing the acetone , the residual precipitates were dissolved in 2× SDS sampler buffer . The whole cells were also lysed with 2× SDS sampler buffer ( 100 µl per O . D600 unit of original culture ) . HeLa cells were seeded at a density of 2×105 cells/well in 24-well plates . The next day , overnight-grown bacteria were inoculated into serum-free DMEM and shaken for 2 hrs at 37°C ( Pre-activation ) . During the pre-activation period , the cell medium was changed to serum-free DMEM , and the cells were cultured at 37°C , 5%CO2 until the beginning of infection . The cells were subjected to bacterial infection at moi ( multiplicity of infection ) of 100 for 3–4 hours . Infection was terminated by washing the cells with PBS to remove non-adherent bacteria , and the medium was replaced with fresh DMEM containing gentamicin ( 0 . 1 mg/ml ) and HKE ( heat-killed E . coli ) at concentration of 108 bacteria/ml . The cells were further cultured for 8 to 12 hours , and the medium was collected for the analysis of IL-8 by ELISA , performed according to the manufacturer's protocol ( Thermo Scientific ) . All the experiments were performed in triplicate and repeated three times . Student's t-test was used to calculate the significance . HeLa cells were first seeded at 1×105/well in 24-well plates one day prior to the transfection . The pNFKB-SEAP reporter plasmid ( Clonetech; containing the κB-binding element and SEAP reporter gene ) and pGL4-13 control luciferase plasmid ( Promega ) were transfected using Lipofectamine 2000 ( invitrogen ) , according to the manufacturer's protocol . Forty-eight hours after transfection , the cells were infected as described above for the IL-8 ELISA , except that at the end of stimulation , the cell medium and lysates were collected and assayed for reporter activity . SEAP ( for NF-κB activity ) and Luciferase ( transfection control ) were analyzed using the Great EscApe SEAP Fluorescence Detection kit ( Clontech ) and ONE-Glo luciferase assay system ( Promega ) , respectively . The NF-κB activities were normalized using the luciferase reporter . For the EGFP-nleC SEAP/Luciferase assay , pNFKB-SEAP and pGL4-13 , pEGFP-C1 , or pEGFP-NleC were co-transfected into HeLa cells . Forty-eight hours later , the cell medium was changed to serum-free DMEM for 2 hours . The cells were then stimulated with HKE for 7 hours . The SEAP and Luciferase activity were measured as described above . All experiments were performed in triplicate and repeated three times . The student's t-test was used to calculate the significance . p65 , GST , GST-NleC , and GST-NleCmut were expressed by bacteria . p65 was purified by Dynabeads ( Invitrogen ) conjugated to anti-p65 antibodies ( ab7970 , Abcam ) ; GST , GST-NleC , and GST-NleCmut were purified by GST sepharose beads ( GE Healthcare ) . For the in vitro cleavage assay , p65 ( 50 nM ) was mixed with GST ( 0 . 3 nM ) , GST-NleC ( 0 . 3 nM ) , or GST-NleCmut ( 0 . 3 nM ) in a final volume of 20 µl of reaction buffer ( 10 mM Tris-HCl pH 7 . 4 , 150 mM NaCl , 0 . 5 mM DTT , 2 . 5 mM CaCl2 , and 0 . 5 mM MgCl2 , 0 . 5 nM ZnCl2 ) . The reaction mixtures were incubated at 25°C for 8 hours . In some samples , EDTA was added to the reaction buffer at final concentration of 10 mM . Unless otherwise specified , the infected or transfected HeLa cells were washed twice with ice-cold PBS , then directly lysed with 2× SDS sampler buffer . The lysates were sonicated briefly to reduce the viscosity of the genomic DNA . For cell fractionation , PBS washed cells were pelleted , resuspended in Buffer A ( 10 mM HEPES pH 7 . 9 , 1 . 5 mM MgCl2 , 10 mM KCl , 0 . 5 mM DTT , 0 . 05% Triton X-100 and 1× protease inhibitor cocktail ) , and centrifuged . The supernatant was collected as the cytoplasmic fraction . Then the remaining pellet was washed twice with buffer A and re-suspended in Buffer B ( 300 mM NaCl , 15 mM HEPES pH 7 . 9 , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 0 . 5 mM DTT , 26% glycerol ( v/v ) ) . Samples were sonicated and left on ice for 30 min before being centrifuged . The resulting supernatant containing the nuclear content was collected . For western blotting , a standard protocol was used . Anti-p65 N-term ( # 4764 , Cell Signaling Technology ) , anti-phospho-p65 ( # 3033 , Cell Signaling Technology ) , anti-IκBα ( # 4814 , Cell Signaling Technology ) , anti-Lamin A/C ( # 2032 , Cell Signaling Technology ) , anti-p65 C-term ( ab7970 , Abcam ) , and anti-α-tubulin ( Clone B-5-1-2 , Sigma ) antibodies were used . HeLa cells were seeded at density of 1×105 cells/cm2 on cover glasses one day prior to the experiment . On the day of experiment , culture media were exchanged with serum free DMEM and cells were subjected to pre-activated bacteria at m . o . i of 100 . Infection was allowed to proceed for 3 hours and was terminated by washing cells with PBS to remove non-adherent bacteria . Cells were fixed with 4% paraformaldehyde and permeabilized with 0 . 5% Triton X-100 in PBS . Following the blocking with 5% BSA in PBS , cells were stained with anti-HA ( Bethyl Laboratories , Inc . , Montgomery , TX , USA ) , Rhodamin-Phallodin ( Wako , Japan ) , and DAPI ( Sigma-Aldrich , Japan ) . Anti-rabbit-FITC ( against anti-HA antibody ) was used as the secondary antibody . Immunofluorescence images were taken using BioRad Radiance2100 confocal microscope . | Enteropathogenic Escherichia coli ( EPEC ) and enterohemorrhagic E . coli ( EHEC ) cause food-borne diseases , including watery diarrhea or severe bloody diarrhea and life-threatening kidney disease ( hemolytic uremic syndrome ) . Upon ingestion , EPEC/EHEC colonize the cells of the epithelial lining in the intestinal tract . In response , the affected cells initiate an immune response by secreting cytokines that attract immune cells . To prevent their early elimination by the host , these bacteria have developed strategies to prevent the host immune response . They do this by injecting bacterial effectors into the host cells to disrupt the NF-κB pathway , an essential effector of the host cell immune response . In the current study , we report the discovery of an NF-κB suppressive effector in EPEC/EHEC called NleC , and its novel mechanism . We found that NleC is a zinc protease that can digest p65 , a critical component of the NF-κB pathway , thus dampening the host inflammatory response . NleE is another recently identified anti-inflammatory effector . We show here that an EPEC/EHEC mutant deficient in both NleC and NleE loses most of its ability to suppress the host inflammatory response . Our findings show how two different bacterial effectors can function in cooperation to modify the host immune response . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] | 2010 | NleC, a Type III Secretion Protease, Compromises NF-κB Activation by Targeting p65/RelA |
Mammarenaviruses are associated with human hemorrhagic fever diseases in Africa and America . Recently , a rodent mammarenavirus , Wēnzhōu virus ( WENV ) and related viruses , have been reported in China , Cambodia , and Thailand . Moreover , in Cambodia , these viruses were suspected to be associated with human disease . In China , Yunnan Province is famous for its abundant animal and plant diversity and is adjacent to several South-eastern Asia countries . Therefore , it is necessary to know whether WENV-related viruses , or other mammarenaviruses , are prevalent in this province . Small mammals were trapped , euthanized , and sampled . Mammarenavirus RNA was detected using a nested reverse transcription polymerase chain reaction ( RT-PCR ) and quantified by real-time RT-PCR . A total of 1040 small mammals belonging to 13 genera and 26 species were trapped in Yunnan Province . WENV-related mammarenaviruses were detected in 41 rodent liver samples , mainly in brown rats ( Rattus norvegicus ) and oriental house rats ( R . tanezumi ) . Viral nucleocapsid protein was detected in liver sections by indirect immunofluorescence assay . Full-length-genomes were amplified by RT-PCR and used for phylogenetic analysis with the MEGA package . Recombination analysis was performed using the SimPlot and Recombination Detection Program . WENV related viruses circulated in small mammals in Yunnan Province . Whole genome sequence analysis of five selected viral strains showed that these viruses are closely related to WENVs discovered in Asia and form an independent branch in the phylogenetic tree in the WENV clade . Paying attention to investigate the influence of these viruses to public health is essential in the epidemic regions .
Mammarenaviruses , belonging to the genus Mammarenavirus , family Arenaviridae , are associated with several hemorrhagic fever diseases around the world [1] . Mammarenavirus particles are spherical to pleomorphic in shape and , range from 40 to 300 nm in diameter [2] . The virus particle carries a variable number of ribosomes ( 20–25 nm ) from host cells that are responsible for their sandy , or arena-like , appearance under electron microscopy[3] . Mammarenavirus possess a bi-segmented , ambisense , single-stranded RNA genome . The L and S segments share conserved sequences at both ends and are around 7 . 2 and 3 . 5 kb long , respectively [4] . Each genomic RNA encodes two different proteins in opposite orientations [5] . The two open reading frames ( ORFs ) in each segment are separated by an intergenic noncoding region that forms one or more energetically stable stem-loop ( hairpin ) structures [6] . The L segment encodes a viral RNA-dependent RNA polymerase ( RdRp , L ) and a zinc binding matrix protein ( Z ) . The S segment encodes a nucleoprotein ( NP ) and an envelope glycoprotein precursor ( GPC ) [2] . Rodents are the primary reservoir of mammarenaviruses , with the exception of the Tacaribe virus that was isolated from Jamaican fruit-eating bats ( Artibeus jamaicensis ) and Wēnzhōu virus ( WENV ) also detected in shrews [7–9] . Most mammarenaviruses induce a persistent and frequently asymptomatic infection in their reservoir hosts and are associated with a specific host species or group of species [10] . Mammarenaviruses can be transmitted through rodent urine or blood [11] . Humans usually become infected through contact with infected rodents or inhalation of infectious rodent excreta or secretions [2] . Animal models , such as hamsters , mice , guinea pigs , and non-human primates , have been used to study the pathogenicity of mammarenaviruses and to measure vaccine or drug efficacy [12] . Viral infections have been identified in many organs , including the spleen , liver , adrenal gland , placenta , and lungs [10 , 13–17] . The Mammarenavirus genus is divided into the two main groups corresponding to Old and New World viruses . Mammarenavirus infections contribute significantly to the human disease burden in both Africa and the Americas , but little is known about the disease burden in Asia [18] . Lassa virus ( LASV ) causes Lassa hemorrhagic fever ( LHF ) with a mortality of between 5 , 000 and 10 , 000 each year in West Africa[19 , 20] , whilst Junín virus ( JUNV ) , Machupo virus , Sabía virus , and Guanarito virus cause fatal hemorrhagic fever diseases and had a mortality rate between 20 and 30% in Latin America[18 , 21] . In 2014 , a rodent arenavirus , WENV was identified in Wēnzhōu , Zhejiang Province , China [8] . In 2015 , a genetic variant of WENV , associated with influenza-like illness , was found in Cambodia and Thailand [22] . Recently , similar virus sequences were reported in Shandong , Guangdong , and Hainan provinces , China [9 , 23–26] . Yunnan Province is notable for its plant and animal diversity and is geographically adjacent to South-eastern Asian countries . Diverse hantaviruses , coronaviruses , hepatitis viruses , paramyxoviruses , and astroviruses had been reported in rodents in Yunnan Province[27–33] . There are more than 88 rodent species belonging to eight families in Yunnan Province [34] . Therefore , it is essential to study the distribution of mammarenaviruses in Yunnan Province to understand its genetic evolution and geographic distribution .
The collection of small wild animals was performed by veterinarians with approval from the Animal Ethics Committee of Yunnan Institute of Endemic Diseases Control and Prevention ( Animal ethics approval number: 201302 ) . From April 2015 to September 2017 , small mammals were trapped in 12 counties in urban or residential districts of eastern and western Yunnan Province ( Fig 1 ) . Small animals were captured through mouse traps using fried foods as bait . Animals were assigned to species through morphologic observation and confirmed by mitochondrial cytochrome B DNA amplification and sequencing . Captured animals were humanely euthanized and blood and/or organs ( heart , liver , spleen , lung , kidney , and intestine ) were collected , stored temporarily in liquid nitrogen , and transferred to the laboratory where they were stored at −80ºC until use . Animal tissues were thawed and viral RNA extracted from approximately 0 . 1 g of tissue using the High Pure Viral RNA Kit ( Roche , Germany ) . A nested reverse transcription polymerase chain reaction ( RT-PCR ) was performed using degenerate primers targeting the conserved RdRp regions , as previously described [8 , 9] . In brief , the first round of PCR was performed in a 25 μl reaction mix using the SuperScript III/ Platinum Taq Enzyme Mix , and 3 μl of RNA as a template . The second round of PCR was performed using the Platinum Taq Enzyme ( Invitrogen , USA ) , and 1 μl of the first-round PCR product as a template . The resultant amplicons were separated by agarose gel electrophoresis and bands of the expected sizes were gel purified and directly sequenced or cloned into the pGEM-T Easy Vector ( Promega , USA ) if direct sequencing was unsuccessful . A qRT-PCR using the HiScript II One Step qRT-PCR SYBR Green kit ( Vazyme Biotech , Nanjing , China ) was developed and optimized according to the manufacturer’s instructions using 20 μl reaction volumes . Specific primers ( 5′-AGAAGGAAGATGCTCTTGTT-3′ and 5′-AAGACCTGATTGAGTGTTGG-3′ ) were designed based on the sequence obtained in this study . A plasmid containing the target sequence was used for viral RNA transcription in vitro and for generating the standard curves for viral RNA quantification . PCR conditions were as follows: 3 min at 50°C for reverse transcription; 5 min at 95°C for activation of Taq DNA polymerase; and 40 cycles at 95°C for 10 sec and 60°C for 30 sec . Positive samples were characterized by a well-defined exponential fluorescence curve that crossed the cycle threshold ( Ct ) within 36 cycles . Specimens with a Ct >36 were considered negative . Each run included three viral positive template controls and two negative controls to monitor performance . Viral genome copies were calculated in each sample using the standard curves of the template RNA . Viral genomes of the detected rodent mammarenavirus were amplified by nested-PCR and rapid amplification of cDNA ends ( RACE ) . In brief , liver sample RNA was reverse transcribed to cDNA using MLV reverse transcriptase ( Promega , USA ) then eight short fragments were amplified using eight primer pairs ( sequences provided upon request ) based on the conserved domains of the L and S fragment sequences . The remaining gaps were filled by specific primers that were designed based on the obtained sequences . Genomic 5′ and 3′ ends were amplified using the SMARTer RACE cDNA Amplification Kit ( Clontech , USA ) . Each PCR product was cloned into the pGEM-T Easy Vector ( Promega , USA ) and at least three clones for each PCR fragment were sequenced to obtain a consensus sequence . Sequences were assembled by Geneious R10 software ( Biomatters , New Zealand ) . Genomic nucleotide ( nt ) sequences and the deduced ORF amino acid ( aa ) sequences were compared to those of representative mammarenaviruses using Geneious R10 software ( Biomatters , New Zealand ) . Recombination events were scanned for the sequences of Yunnan isolates using the Recombination Detection Program [35] . Suggested recombination events with strong p value were further confirmed in Simplot 3 . 5 . 1 [36] . Phylogenetic trees were constructed using neighbor-joining methods from MEGA 7 [37] . Virus isolation was conducted using canine macrophage cell DH82 and VeroE6 as described previously [8] . Cells were cultured and inoculated with viral RNA positive samples after a 10-fold dilution . The inoculated cells were incubated in the Dulbecco's Modified Eagle's Medium ( DMEM ) with 2% fetal bovine serum ( FBS ) . After three blind passages , the cell culture supernatant was tested for the presence of virus using the nested RT-PCR . The enzyme linked immunosorbent assay ( ELISA ) for mammarenavirus antibody detection was developed based on a procedure of a previous study [38] . In brief , the NP gene was amplified and cloned into a pET expression plasmid fused with a C-terminal His tag . The NP-His fusion was expressed in E . coli BL21 ( DE3 ) cells and purified by Profinity IMAC Nickel Charged Resin ( BIO-RAD , US ) according to the manufacturer’s instructions . Purified NP was used to immunize rabbits to obtain rabbit polyclonal antibodies . Rodent serum samples were diluted 1:20 and added to the ELISA plate coated with NP . The Peroxidase conjugated Affinipure Goat Anti Mouse antibodies ( Proteintech , Wuhan , China ) were used as the secondary antibodies . Absorbance values that were three times greater than those of the negative control were considered positive . For western blotting , 200 ng of NP was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane . Rodent serum samples were diluted ( 1:100 ) and the peroxidase conjugated Affinipure Goat Anti Mouse ( H+L ) ( Proteintech , Wuhan China ) used as secondary antibody . The His-tagged human coronavirus NL63 N was used as parallel control . Liver tissues from rodents positive or negative for mammarenavirus RNA were fixed in 10% formalin for 36 h . Samples were then dehydrated , paraffin embedded , sectioned , and mounted onto glass slides . After staining with HE , the slides were examined using microscopy ( Olympus , Japan ) . Sections ( 5 mm ) of formalin-fixed paraffin-embedded tissues were placed onto positively charged glass slides and air dried for 30 min . The tissue sections were deparaffinized , rinsed , and incubated with target retrieval solution ( Sigma-Aldrich , USA ) . After the sections were blocked with 5% FBS ( Gibco , Australia ) , they were incubated with anti-NP rabbit serum for 30 min at 37°C followed by incubation with goat anti-rabbit IgG cyanine 3 ( Cy3 ) -conjugated secondary antibody ( TransGen Biotech , China ) . The sections were checked using an immunofluorescence laser scanning confocal microscope ( Zeiss , Germany ) .
A total of 1040 small mammals , belonging to 13 genera and 26 species , were trapped in 12 counties ( Fig 1 and Table 1 ) . A total of 41/1040 ( 3 . 94% ) liver tissue samples were positive for mammarenavirus RNA , including 34/195 ( 17 . 44% ) brown rat samples , 5/328 ( 1 . 53% ) oriental house rat samples , 1/12 ( 8 . 33% ) Himalayan field rat ( R . nitidus ) samples , and 1/10 ( 10% ) Northern tree shrew ( Tupaia belangeri ) sample . Among the 41 positive samples , 16 brown rat samples were from Luxi County , three brown rat samples were from Yiliang County , and all other samples were collected from Yuanmou County . The 41 partial viral sequences ( 600 bp ) share 85–100% similarity with each other . Furthermore , the isolated viral sequences share 86–91% similarity with the closest strain discovered in Wēnzhōu , which is higher than the similarity observed with Cambodian isolates ( 86–88% ) and Thai isolates ( 70–73% ) . These results suggest that viral sequences detected at the same location are closely related irrespective of host differences . The phylogenetic tree shows that the viruses detected in Yunnan Province form an independent lineage that is divided into two branches , one containing the sequences from Yuanmou and Luxi counties , and the other containing the sequences from Yiliang County ( Fig 2 ) . Because the failure of virus isolation , five liver samples were selected for full length virus genome , including two each from oriental house rats ( WENV/Rt2015YM16 and WENV/Rt2015YM51 ) and brown rats in Yuanmou ( WENV/Rn2016YM03 and WENV/Rn2016YM51 ) and one from brown rat in Yiliang ( WENV/Rn2016YL04 ) . All virus and partial host cytochrome b gene sequences obtained in this study have been submitted to GenBank ( MG736216-MG736236; MK192269-MK192293; MK177225-MK177229 ) . The lengths of the L and S segments ranged from 7140–7153 bp and 3318–3377 bp , respectively . Highly conserved non-coding regions were identified in both segments ( 25 and 55 for the L and S segments , respectively ) . All L segments contained two ORFs which encoded a RdRp of 2 , 223 amino acids and a Z of 92 amino acids , either side of an 124 nucleotide non-coding sequence . All S segments contained two ORFs which encoded an NP of 568 amino acids and a GPC of 493 amino acids , either side of a 62 nucleotide non-coding region . Nucleotide and deduced ORF amino acid sequences from the five isolates were compared to those of representative mammarenaviruses ( Table 2 ) . Similar to results obtained from the analysis of the partial sequences , viral strains detected in the same Yunnan Province location are most closely related , irrespective of their host species . The most closely related strains are the Zhejiang and Cambodian viriants of WENV , with L and the S segment nucleotide identities of 85 . 4–87 . 0% and 83 . 0–87 . 7% , respectively . Based on the International Taxonomy Committee on Viruses ( ICTV ) Arenaviridae Study Group guidelines for new mammarenavirus species , the Yunnan strains examined here are strain of WENV ( species Wenzhou mammarenavirus ) [1 , 39] Phylogenetic analysis of complete L segments showed that mammarenaviruses detected in South-eastern Asia form a distinct lineage in which the viruses are geographically separated ( Fig 3A ) . The topology of the tree changed for the S segments of Yunnan isolates , indicating potential genetic reassortments between the detected viruses in this study ( Fig 3B ) . After recombination scanning , we found an obvious recombination event in the S segment of WENV/Rn2016YM03 with strong p value ( < 10−38 ) , indicating that WENV/Rn2016YM03 is likely a result of recombination between WENV/Rn2016YL04 and WENV/Rt2015YM16 ( Fig 4 ) . To determine the tropism of these mammarenaviruses in their natural hosts , viral genomic RNA was quantified in tissue samples ( heart , liver , spleen , lung , spleen , and intestine ) by qRT-PCR . The results showed that the viral load in the positive samples ranged from 5 . 1 ×108 to 1 . 14 × 1011 copies per gram of tissue ( Fig 5 ) . Viral RNA was detected in liver tissues of all five positive samples . Some viruses showed wider tissue tropism including the heart , spleen , lung , kidney , and intestine . The highest viral load was found in samples collected in Yuanmou in 2016 . The histopathology of the virus was examined in liver sections in both viral RNA-positive and viral RNA-negative animals ( M . pahari ) . Although indirect immunofluorescence , using rabbit polyclonal antibodies against viral NP detected virus infection , no obvious lesions were observed in viral positive livers ( Fig 6 ) . The cutoff absorbance value for positive serum samples was set at 0 . 036 , a value three times greater than that of the negative control . Overall , 24 of 118 serum samples were seropositive to NP ( 24/118 , 20 . 34% ) ( S1 and S2 Figs and S1 Table ) . All positive serum samples were from rodent species in which the viral nucleic acid had been detected .
In this study , we conducted an investigation of WENV infections in small mammals in a wide geographical area of Yunnan Province , China . Our results demonstrated , for the first time , that WENV is prevalent in three of five Rattus species and one Tupaia species in three geographically related counties . Together with previous reports , we have shown that WENV or related viruses have a host range that includes rats ( R . norvegicus , R . rattus , R . losea , R . exulans , R . tanezumi , and R . nitidus ) , White-bellied Rat ( Niviventer niviventer ) , Asian house shrews ( Suncus murinus ) , and Tree shrews ( T . belangeri ) [8 , 9 , 22 , 24–26] . No mammarenavirus was detected in other locations , suggesting that the distribution of these viruses may currently be limited to these three regions . However , considering the movement of rodents due to ecological changes , we cannot exclude the possibility of these viruses spreading in the future . Therefore , it is necessary to continue surveillance for these viruses in their natural reservoirs . Sequences analysis of five full-length genomes showed that these discovered viruses are in fact closely related to WENVs , which indicated a common ancestor . But the diversity within that clade also revealed an independent lineage of Asian arenaviruses[8 , 9 , 22 , 24 , 25] . The phylogenetic tree based on partial L sequences had different shape with the tree based on complete length of L segment suggesting more complexity evolution of WENV . Meanwhile , the tree topologies for the L and S segments of the Yunnan isolates are also incongruent . In the L tree , the Yunnan strains are separated according to the two sampling locations irrespective of host species , while two strains from different locations are closely related in the S tree ( Fig 3 ) . Recombination analysis suggested that one of the Yunnan strains , WENV/Rn2016YM03 , results from an S segment recombination event between WENV/Rn2016YL04 and WENV/Rt2015YM16 . We hypothesize that the divergence of the viruses between the two locations is a recent event caused by host movement . More evolution models and more viruses complete genome sequences were needed to improve the understanding of WENV evolution . The pathogenic potential of WENV to humans was the first reported in 2015 and remains an issue of further studies [8] . In Cambodia , the WENV was associated with human respiratory diseases . In this study , we have gained both molecular and serological evidence to demonstrate that brown rats ( R . norvegicus ) and oriental house rats ( R . tanezumi ) are the major natural host of WENVs . In Yunnan , these two rodent species are the dominant species and commensal with humans , suggesting a high risk of spillover of WENVs and occurrence of similar respiratory illness cases caused by WENV infection . Thus the ELISA method developed in this study can serve as a tool for screening WENV antibodies in undiagnosed fever patients . In Yunnan Province , the Haemorrhagic Fever with Renal Syndrome ( HRFS ) and Dengue fever are also occurring , which may lead to the misdiagnosis of the disease caused by WENVs or other mammarenaviruses . The ELISA test for WENV antibodies will reduce the likelihood of misdiagnosis , though the results need to be further verified by virus neutralization assay . The recent report of WENV association with human illness reminds us there is still high risk of spillover of these rodent arenaviruses across species[22] . Moreover , we found that some rodent mammarenaviruses have a wider tissue tropism in their natural hosts . In addition to the liver , these viruses were also detected in the heart , spleen , lung , kidney , and intestine and may increase the spillover potential of some viral strains . Deep borderline , many remote areas , and abundant animal sources increase the burden to discover and control zoonotic diseases in Yunnan Province . Taken together , our results emphasize the importance of extensive surveillance of mammarenavirus in both their natural reservoirs and in humans . | Rodents are natural reservoirs of mammarenavirus . Lymphocytic choriomeningitis virus ( LCMV ) , isolated in Asian countries during the 1990s , has a worldwide distribution and was the first mammarenavirus isolated . In 2014 , a second mammarenavirus , Wēnzhōu virus ( WENV ) , was identified in rodents in Zhejiang Province of China and later in Guangdong , Shandong , and Hainan Provinces . Most importantly , WENV or related viruses were reported in Thailand and Cambodia . In Cambodia , the isolated virus was associated with human respiratory diseases . In this study , we detected WENV or related viruses in Yunnan Province and found a high prevalence in rats of two species ( Rattus norvegicus and R . tanezumi ) . Phylogenetic analysis of the complete L and S segments of five strains showed that these viruses form an independent phylogenetic branch in WENV clade most closely related to WENVs found in China and Cambodia . Considering the wide spread distribution of rats and altered distribution patterns due to ecological changes , we propose that these viruses may have a wider prevalence and be found in countries from South-eastern Asia to China . Given that WENV may be associated with human diseases , it is necessary to improve surveillances of these viruses in their natural reservoirs and in humans . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"taxonomy",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"vertebrates",
"animals",
"mammals",
"viruses",
"phylogenetics",
"data",
... | 2019 | Prevalence of Wēnzhōu virus in small mammals in Yunnan Province, China |
Dietary fatty acids can be incorporated directly into phospholipids . This poses a specific challenge to cellular membranes since their composition , hence properties , could greatly vary with different diets . That vast variations in diets are tolerated therefore implies the existence of regulatory mechanisms that monitor and regulate membrane compositions . Here we show that the adiponectin receptor AdipoR2 , and its C . elegans homolog PAQR-2 , are essential to counter the membrane rigidifying effects of exogenously provided saturated fatty acids . In particular , we use dietary supplements or mutated E . coli as food , together with direct measurements of membrane fluidity and composition , to show that diets containing a high ratio of saturated to monounsaturated fatty acids cause membrane rigidity and lethality in the paqr-2 mutant . We also show that mammalian cells in which AdipoR2 has been knocked-down by siRNA are unable to prevent the membrane-rigidifying effects of palmitic acid . We conclude that the PAQR-2 and AdipoR2 proteins share an evolutionarily conserved function that maintains membrane fluidity in the presence of exogenous saturated fatty acids .
The adiponectin receptors AdipoR1 and AdipoR2 are members of the PAQR family of proteins characterized by seven transmembrane domains with their N-terminus facing the cytoplasm [1] . PAQR proteins likely act as hydrolases with different specificities [2] , and the crystal structure of the AdipoRs suggests that they act as ceramidases [3 , 4] , which is confirmed by enzymatic assays [3] and is conserved in yeast homologs [5 , 6] . The ceramidase activity , which would produce a signaling sphingosine 1-phosphate as well as a putative fatty acid that may also serve as a signal , has been proposed to mediate the well-established anti-diabetic effects of the AdipoRs [7 , 8] . Several important questions still remain regarding the mechanisms by which the AdipoRs exert their effects . In particular , we do not know what conditions trigger AdipoR signaling nor the precise cellular consequences of this signaling . Our previous studies of PAQR-2 , an AdipoR homolog in the nematode C . elegans , showed that it acts as a sensor of plasma membrane rigidity that can restore membrane fluidity by promoting fatty acid desaturation [9–11] . Mutant worms lacking a functional PAQR-2 protein are intolerant of conditions that promote membrane rigidification , such as cold or diets that increase saturated fatty acid ( SFA ) content in the worms . For example , including small amounts of glucose in the culture plate results in increased SFA content in membrane phospholipids , membrane rigidification and death of the paqr-2 mutant [11] . The growth and membrane fluidity of wild-type worms is unaffected by glucose . The extreme sensitivity of the paqr-2 mutant to glucose is particularly interesting given the anti-diabetic properties of the mammalian AdipoRs since it suggests that these proteins may mitigate glucose toxicity specifically by regulating membrane composition [12] . In C . elegans , the paqr-2 mutant defects can be suppressed genetically by mutations that result in increased production of unsaturated fatty acids ( UFAs ) accompanied by normalization of plasma membrane fluidity or , alternatively , by cultivation in the presence of fluidizing amounts of non-ionic detergents [10 , 11] . The high degree of sequence homology between PAQR-2 and the AdipoRs ( 53 . 7% amino acid identity with AdipoR2 over a 283 aa region ) suggests that these proteins have the same cellular functions , i . e . act as sensors/regulators of membrane properties . However , this has not yet been demonstrated for the mammalian AdipoRs . The present study produced two important advances in our understanding of PAQR-2 and AdipoR2 . Firstly , we show that PAQR-2 is essential to prevent membrane rigidification by diets containing a high SFA/monounsaturated fatty acid ( MUFA ) ratio . Secondly , we demonstrate that the mammalian AdipoR2 protein also acts as a regulator of membrane fluidity that counters the rigidifying effects of exogenous SFAs . Regulation of membrane fluidity is therefore an evolutionary conserved function of the PAQR-2/AdipoR proteins that is essential to cellular health in the presence of exogenous SFAs .
We previously showed that the C . elegans paqr-2 mutant is sensitive to glucose supplementation and that this sensitivity is accompanied by an increase in phospholipid SFAs and membrane rigidity [11] . More recently we discovered that other glycolysis-related metabolites , such as glycerol , dihydroxyacetone , pyruvate and lactate , are also toxic to paqr-2 ( Fig 1 ) . This sensitivity is specific to glycolysis-related metabolites: the paqr-2 mutant is not more sensitive to other types of stressors such as osmotic or oxidative stress ( paraquat ) , inhibition of the respiratory chain or mevalonate pathway , or toxic doses of DMSO ( S1 Fig ) . The glucose-related stressors , like glucose itself , also cause rigidification of membranes in the paqr-2 mutant ( Fig 2 ) . Others have shown that glucose shortens the lifespan of C . elegans by acting directly on the worms [13 , 14] . We were therefore surprised to discover that glucose is not toxic to paqr-2 mutants grown on an E . coli strain carrying a ΔPTS mutation that prevents glucose uptake ( Fig 3A ) . This mutation specifically abolishes glucose toxicity , and has no effect on glycerol or pyruvate toxicity ( S2A Fig ) . To better understand the role of the bacteria in mediating metabolite toxicity , we took advantage of the Keio collection of E . coli deletion mutants [15] . This collection is derived from the BW25113 E . coli strain , a food source that does not attenuate the cold sensitivity or tail tip defects of the paqr-2 mutant grown on NGM plates ( S2B and S2C Fig ) . Of five metabolites tested ( glucose , glycerol , dihydroxyacetone , lactate and pyruvate ) that are toxic to the paqr-2 mutant grown on a standard diet of OP50 E . coli , only two ( glucose and glycerol ) are toxic to paqr-2 mutants grown on a diet of BW25113 ( S2D Fig ) . Additionally , mutations in several E . coli genes important for the metabolism of glucose or glycerol abrogated the toxicity of these dietary supplements on paqr-2 mutants fed BW25113 ( Fig 3B–3D; S3 Fig ) . E . coli metabolism is therefore responsible for the toxicity of several dietary metabolites in the paqr-2 mutant . Conversely , a normal function of PAQR-2 must be to protect against such dietary toxicity . The metabolites toxic to paqr-2 mutants can all be readily converted into acetyl-CoA , a precursor for fatty acid synthesis ( Fig 3B ) . To better understand the nature of the dietary toxicity , we analyzed the fatty acid composition of E . coli grown with different supplements or mutations . We found that dietary E . coli containing elevated SFA/MUFA ratios are toxic to the paqr-2 mutant ( Fig 4A; complete lipidomics values are provided as a separate supplementary dataset file ) . Specifically , a dietary SFA/MUFA ratio higher than ~1 . 8 results in an excess SFA in the worm phospholipids that is almost always lethal to paqr-2 mutants ( Fig 4A and 4B and S4A Fig ) . Interestingly , evaluating the simpler palmitic acid ( PA; 16:0 ) /cis-vaccenic acid ( CVC; 18:1 delta 11 ) ratio was equally predictive of lethality ( Fig 4A ) . The only exception was the pfkA E . coli mutant grown on glucose that supported the growth of paqr-2 mutant worms in spite of an elevated PA/CVC ratio ( S4A Fig ) ; this pfkA E . coli mutant had a very unusual glossy appearance indicative of complex metabolic reprogramming ( see Fig 3C ) . To test directly the effect of SFAs , we developed a protocol to pre-load E . coli with the SFA PA ( Fig 5A ) , resulting in a doubling of dietary PA content ( Fig 5B ) . Using this method , we found that ≥1 mM PA during the cultivation of the E . coli produces a diet that is extremely toxic to the paqr-2 mutant but harmless to wild-type worms ( Fig 5C ) . This toxicity is accompanied by an increase in PA among phospholipids in the worms that is more pronounced in paqr-2 mutants ( Fig 5D ) , and confers a dramatic decrease in membrane fluidity in the paqr-2 mutant ( Fig 5E and 5F ) . Consistently , wild-type worms respond to PA-loaded E . coli by strongly upregulating a GFP-reporter of the Δ9-desaturase FAT-7 , which the paqr-2 mutant fails to do ( S5 Fig ) . paqr-2 is therefore required for the upregulation of desaturases that prevent membrane rigidification by dietary SFAs . Importantly , preloading the bacterial diet with a combination of PA and oleic acid ( OA; 18:1 delta 9 ) , and thus re-setting the 16:0/18:1 ratio ( S4B Fig ) , greatly reduces toxicity and rigidification of the membrane in paqr-2 mutants ( Fig 5G–5I ) . Others have shown that the addition of PA to mammalian cells causes membrane rigidification [16–18] . We were able to verify this in human HEK293 cells using the FRAP assay , and also found that the inclusion of OA counters the rigidifying effects of PA ( Fig 6A and 6B and Table 1 ) . To test whether the AdipoRs provide protection against the rigidifying effects of PA , we optimized conditions to knockdown the levels of the AdipoRs and other genes using siRNA ( Fig 6C ) . AdipoR2 knockdown ( but not AdipoR1 knockdown ) caused a clear disorganization of the cellular appearance when HEK293 cells are treated with PA , suggesting a toxic effect ( Fig 6D–6G and S6A–S6D Fig ) . Inhibition of AdipoR1 or AdipoR2 using two independent sets of siRNA oligonucleotides had no effects on the membrane fluidity of HEK293 cells grown under normal conditions ( Fig 6H and 6I; S6E and S6F Fig ) . In contrast , AdipoR2 knockdown ( but again not AdipoR1 knockdown ) caused a dramatic increase in membrane rigidification of PA-treated HEK293 cells ( Fig 6J and 6K; S6G and S6H Fig ) . As controls , inhibition of the non-essential gene GAPDH had no effect on membrane fluidity , while inhibition of SCD , encoding a stearyl-CoA desaturase , caused the expected reduction of membrane fluidity when HEK293 cells are challenged with PA ( S6I and S6L Fig ) . Additionally , FCCP , an uncoupler of mitochondrial oxidative phosphorylation , did not affect membrane fluidity; this demonstrates that toxicity in itself is not sufficient to lower membrane fluidity ( Table 1 ) , as others have also shown [19] . To summarize , among all our FRAP experiments only AdipoR2 and SCD knockdown enhanced the rigidifying effects of PA ( Table 1 ) . As mentioned earlier , lipidomics analysis in C . elegans revealed that PAQR-2 is essential to keep phospholipid SFAs below a critical threshold when the worms are fed a SFA-rich diet . We found that a similar phenomenon occurs in HEK293 cells: there is a dramatic increase in the SFA content among PCs , PEs and TAGs when HEK293 cells are incubated in the presence of PA , and this effect is exacerbated when AdipoR2 is knocked-down using siRNA ( Fig 6L–6N ) . Note that AdipoR2 siRNA causes an increased SFA content in PCs , PEs and TAGs even when PA is not added but that the effect is then more modest . Given the proposed ceramidase activity of AdipoR2 [3 , 8] , we also examined the levels of ceramides in our experiments . PA is a precursor for the synthesis of ceramides and we were therefore not surprised that their relative levels were increased in PA-treated cells ( Fig 6O ) . However , the ceramide levels in PA-treated cells did not increase as much when AdipoR2 had been knocked-down by siRNA , which is somewhat surprising if AdipoR2 acts as a ceramidase ( Fig 6O ) . Up to this point , our experiments using PA employed concentrations of 400 μM , which is the concentration at which a detectable rigidifying effect occurs on normal cells . However , we reasoned that if AdipoR2 acts by preventing membrane rigidification , then cells where AdipoR2 has been knocked-down using siRNA should be sensitive to lower amounts of PA . This is indeed the case: 200 μM PA causes a dramatic decrease in membrane fluidity in AdipoR2 siRNA-treated cells but not in control cells ( Fig 6P and 6Q ) , and this rigidification is accompanied by an equally dramatic increase in the SFA content in both PCs and PEs ( Fig 6R and 6S ) . Altogether , these results demonstrate that AdipoR2 , like its C . elegans homolog PAQR-2 , is required to prevent membrane rigidification by exogenous SFAs .
The fatty acid composition of cellular membranes reflects the composition of the dietary fats . This is especially evident for complex dietary PUFAs that become incorporated into membrane phospholipids in C . elegans [20] and in mammals [21–23] . However , regulatory mechanisms must exist to adjust membrane composition , hence properties , in response to diets with a wide range of SFA/UFA ratios . The present study shows that PAQR-2 in C . elegans , and its homolog AdipoR2 in mammals , are essential to maintain membrane homeostasis in the presence of exogenous SFAs ( see the model in Fig 7 ) . De novo lipogenesis inside the worm is likely adjusted to cellular needs , whereby desaturation is coordinated with lipogenesis to produce a healthy mixture of SFAs and UFAs [24] , which is also the case in mammalian cells [25] . However , the worm cannot control the nature of external factors that could insult the membranes , such as temperature or dietary fat composition . PAQR-2 seems especially required in response to such external factors . There is a remarkably high turnover of membranes in C . elegans: nearly 70% of phospholipids are renewed daily in post-reproductive adults , with a majority of recruited FAs being of dietary origin [20] . Similarly , C . elegans fat stores also reflects the fatty acid composition of dietary bacteria [26] . Such high reliance on dietary FAs necessitates mechanisms to sense and adjust lipid composition to fit cellular needs , and PAQR-2 seems to act as such a sensor/regulator . The desaturases fat-5 , -6 , and -7 are key regulators of membrane turnover and composition in C . elegans [20 , 27] , and their regulation by PAQR-2 is therefore a potent mechanism to regulate membrane homeostasis [10 , 11] . PAQR-2 is not the only regulator of membrane homeostasis identified in eukaryotes . In particular , changes in the lipid composition of the endoplasmic reticulum ( ER ) , where lipids are synthesized , can cause the induction of the unfolded protein response in yeast and mammals , which regulates several genes involved in lipid biosynthesis [28 , 29] . In the ER , it is the protein Ire1 that senses aberrant properties of the membrane via an amphipathic helix [30] . The mechanism of membrane stress sensing by PAQR-2 has not been identified , but it is possible that its conformation or interaction with IGLR-2 , an obligate partner for PAQR-2 activity , is sensitive to the plasma membrane environment where PAQR-2 is localized [11] . Besides influencing membrane properties , lipid composition is important for many other aspects of C . elegans biology . For example , PC levels regulate SBP-1 ( a homolog of mammalian SREBP ) by altering membrane properties important for vesicular trafficking [31 , 32] , and specific UFAs regulate germline and other developmental processes [33 , 34] . It is therefore not surprising that lipid disequilibrium causes activation of stress responses such as the endoplasmic reticulum unfolded protein response [35 , 36] . By acting as a regulator of lipid metabolism , PAQR-2 is therefore likely to impinge on many other processes besides membrane homeostasis . Our study of membrane composition and fluidity in human HEK293 cells led to several intriguing observations . We noted for example that while AdipoR2 siRNA strongly aggravated the effects of PA on membrane composition and fluidity , AdipoR1 siRNA had no significant effect . This suggests that , at least in this cell line , AdipoR2 has a greater role in membrane homeostasis than AdipoR1 , a situation reminiscent of C . elegans where PAQR-2 is more important than PAQR-1 [9] . Our work also suggests that AdipoR2 can regulate membrane fluidity in the absence of supplemented adiponectin , its proposed ligand secreted nearly exclusively by adipocytes [37–39] , since the siRNA-treated HEK293 cells were cultivated for 24 hours in serum-free media prior to FRAP analysis . Again , this is analogous to the situation in C . elegans , where no adiponectin homolog has been identified either by sequence homology searches or in a screen for paqr-2 genocopiers [11] . In this context , it is important to note that Vasiliauskaité-Brooks and co-workers have shown that the AdipoRs ceramidase activity is not strictly adiponectin dependent [3] . Finally , we found that AdipoR2 siRNA led to a decrease in the abundance of ceramides , which is counter-intuitive if AdipoR2 is a ceramidase as has been proposed [3 , 6 , 8] . Further studies will hopefully clarify this and other observations . Mammalian mutant phenotypes suggest that the AdipoRs have an evolutionary conserved role in preventing dietary FA toxicity . In particular , mouse mutants lacking adiponectin , the ligand for AdipoR1 and AdipoR2 , develop metabolic complications ( glucose intolerance , insulin resistance ) only when fed a high fat diet [40] , which is analogous to the C . elegans paqr-2 mutants that show severe phenotypes when fed a SFA-rich diet . Also like the C . elegans paqr-2 mutant , AdipoR1 and AdipoR2 knockout mice show several defects in lipid homeostasis [41–44] . Conversely , overexpression of the AdipoRs leads to improved lipid homeostasis and can even rescue the diabetic phenotype in mice lacking a functional leptin receptor [7] . In human , a rare loss-of-function mutation in AdipoR1 causes autosomal dominant retinitis pigmentosa [45] , which is likely due to AdipoR1 regulating fatty acid composition of the retina [46] . In the future , it will be interesting to investigate further the relevance of our findings for human physiology . In particular , several studies have documented an increase in both membrane SFAs and rigidity in diabetics [12 , 47–55] , and these changes may result from the conversion of glucose into SFAs via de novo lipogenesis . Membrane rigidification could therefore be a component of glucose toxicity , and the oft-documented anti-diabetic activities of the AdipoRs could lie in their ability to counter such a rigidifying effect .
The wild-type C . elegans reference strain N2 and the mutant alleles studied are available from the C . elegans Genetics Center ( CGC; MN; USA ) . The pfat-7::GFP ( rtIs30 ) carrying strain HA1842 was a kind gift from Amy Walker [31] , and its quantification was performed as previously described [10] . C . elegans strains maintenance and experiments were performed at 20°C using the E . coli strain OP50 as food source , which was maintained on LB plates kept at 4°C ( re-streaked every 6–8 weeks ) and single colonies were picked for overnight cultivation at 37°C in LB medium then used to seed NGM plates [56]; new LB plates were streaked every 3–4 months from OP50 stocks kept frozen at -80°C . Stock solutions of supplements ( 1M glucose , dihydroxyacetone , pyruvate , and lactate; 5 M NaCl and KCl; 50 mM fluvastatin ) were filter sterilized then added to cooled NGM after autoclaving; 100 mM paraquat , 30 mM FCCP ( in ethanol ) , pure DMSO and pure glycerol were used without sterilization . The Keio collection of E . coli K-12 in-frame , single-gene knockouts was used as source of E . coli mutants and kept in the presence of 50 μg/ml kanamycin [15] . Mutants were picked as single colonies from LB plates and cultivated overnight at 37°C in LB then seeded onto NGM plates with or without additives . All mutants were confirmed by PCR . E . coli strains were streaked onto MacConkey agar plates with 0 . 4% glucose , grown at 37°C over night and scored for colony color . For length measurement studies , synchronized L1s were plated onto test plates seeded with E . coli , and worms were mounted then photographed 96 hour ( oleic acid rescue experiments ) or 72 hours ( all other experiments ) later . The length of >20 worms was measured using ImageJ [57] . Quantification of the withered tail tip phenotype was done on synchronous 1-day old adult populations , i . e . 72 h post L1 ( n≥100 ) [10] . FRAP experiments in C . elegans were carried out using a membrane-associated prenylated GFP reporter expressed in intestinal cells , as previously described and using a Zeiss LSM700inv laser scanning confocal microscope with a 40X water immersion objective [11 , 58] . Briefly , the GFP-positive membranes were photobleached over a circular ( 7 pixel radius ) using 20 iterations of the 488 nm laser with 50% laser power transmission . Images were collected at a 12-bit intensity resolution over 256x256 pixels ( digital zoom 4X ) using a pixel dwell time of 1 . 58 μsec , and were all acquired under identical settings . For FRAP in mammalian cells , HEK293 cells were stained with BODIPY 500/510 C1 , C12 ( 4 , 4-Difluoro-5-Methyl-4-Bora-3a , 4a-Diaza-s-Indacene-3-Dodecanoic Acid ) ( Invitrogen ) at 2 μg/ml in PBS for 10 min at 37°C . FRAP images were acquired with an LSM880 confocal microscope equipped with a live cell chamber ( set at 37°C and 5% CO2 ) and ZEN software ( Zeiss ) with a 40X water immersion objective . Cells were excited with a 488 nm laser and the emission between 493 and 589 nm recorded . Images were acquired with 16 bits image depth and 256x256 resolution using a pixel dwell of ~ 1 . 34 μs . Ten pre-bleaching images were collected and then the region of interest was beached with 50% of laser power . The recovery of fluorescence was traced for 25 seconds . Fluorescence recovery and Thalf were calculated as previously described [11] . Stocks of 0 . 1 M palmitic acid or 0 . 5 M oleic acid dissolved in ethanol were diluted in LB media to final concentrations of 0 . 25–2 mM , inoculated with OP50 bacteria , then shaken overnight at 37°C . The bacteria were then washed twice with M9 to remove fatty acids and growth media , diluted to equal OD600 , concentrated 10X by centrifugation , dissolved in M9 and seeded onto NGM plates lacking peptone ( 200μl/plate ) . Worms were added the following day . For worm lipidomics , samples were composed of synchronized L4 larvae ( one 9 cm diameter plate/sample ) grown overnight on OP50-seeded NGM , NGM containing 20mM glucose , 0 . 5% glycerol , 20 mM pyruvate , or plates lacking peptone but seeded with fatty acid-supplemented bacteria . Worms were washed 3 times with M9 , pelleted and stored at -80°C until analysis . For bacterial lipidomics , E . coli liquid cultures grown overnight at 37°C were seeded on NGM , NGM containing 20 mM glucose , 0 . 5% glycerol , 20 mM pyruvate or concentrated ( as described above ) and seeded on plates lacking peptone for lipid-supplemented cultures , kept upside-down at 20°C then washed off 96 h later using pure water , pelleted then frozen at -80°C until analysis . For HEK293 lipidomics , cells were cultivated in serum-free media with or without fatty acids for 24 h prior to harvesting using TrypLE Express ( Gibco ) . For lipid extraction , the pellet was sonicated for 10 minutes in methanol and then extracted according to published methods [59] . Internal standards were added during the extraction . Lipid extracts were evaporated and reconstituted in chloroform:methanol [1:2] with 5 mM ammonium acetate . This solution was infused directly ( shotgun approach ) into a QTRAP 5500 mass spectrophotometer ( Sciex , Toronto , Canada ) equipped with a Nanomate Triversa ( Advion Bioscience , Ithaca , NY ) as described previously [60] . Phospholipids were measured using multiple precursor ion scanning [61 , 62] . Ceramides from HEK293 cells were measured using ultra performance liquid chromatography coupled to tandem mass spectrometry according to previous publication [63] . The data was evaluated using the LipidView software ( Sciex , Toronto , Canada ) . The complete lipidomics dataset is provided in the supplementary S1 File . HEK293 were grown in DMEM containing glucose 1 g/l , pyruvate and GlutaMAX and supplemented with 10% fetal bovine serum , 1% non-essential amino acids , HEPES 10 mM and 1% penicillin and streptomycin ( all from Life Technologies ) at 37°C in a water humidified 5% CO2 incubator . Cells were sub-cultured twice a week at 90% confluence . Cells were cultivated on treated plastic flask and multi-dish plates ( Nunc ) . For FRAP experiments , HEK293 were seeded in glass bottom dishes ( Ibidi ) pre-coated with 0 . 1% porcine gelatin ( Sigma ) . The following pre-designed siRNAs were purchased from Dharmacon: AdipoR1 J-007800-10-005 ( set 1 ) and J-007800-09-0002 ( set 2 ) , AdipoR2 J-007801-10-0005 ( set 1 ) and J-007801-09-0002 ( set 2 ) , GAPDH D-001830-10-05 , Non-target D-001810-01-05 or D-001810-10-05 , and SCD J-005061-07-0005 . Transfection of 25 nM siRNA was performed in complete media using Viromer Blue according to the manufacturer’s instructions 1X ( Lipocalyx ) . Knockdown gene expression was verified 48 h after transfection . Total cellular RNA was isolated using RNeasy Kit according to the manufacturer’s instructions ( Qiagen ) and quantified using a NanoDrop spectrophotometer ( ND-1000; Thermo Scientific ) . cDNA was obtained using a High Capacity cDNA Reverse Transcription Kit ( Applied Biosystem ) with random hexamers . qPCR were performed with a CFX Connect thermal cycler ( Bio Rad ) using Hot FIREpol EvaGreen qPCR SuperMix ( Solis Biodyne ) and standard primers . Samples were measured as triplicates . The relative expression of each gene was calculated according to the ΔΔCT method [64] . Expression of the housekeeping gene PPIA was used to normalize for variations in RNA input . Primers used were: AdipoR1-For ( CCATCTGCTTGGTTTCGTGC ) and -Rev ( AGACGGTGTGAAAGAGCCAG ) , AdipoR2-For ( TCATCTGTGTGCTGGGCATT ) and -Rev ( CTATCTGCCCTATGGTGGCG ) , GAPDH-For ( GAGAAGGCTGGGGCTCATTT ) and -Rev ( TAAGCAGTTGGTGGTGCAGG ) , PPIA-For ( GTCTCCTTTGAGCTGTTTGCAG ) and -Rev ( GGACAAGATGCCAGGACCC ) , and SCD-For ( TTCGTTGCCACTTTCTTGCG ) and -Rev ( TGGTGGTAGTTGTGGAAGCC ) . PA and OA were dissolved in sterile DMSO ( Sigma ) then mixed with fatty acid-free BSA ( Sigma ) in serum-free medium for 20 min at room temperature . The molecular ratio of BSA to fatty acid was 1 to 5 . 3 ( except in the experiment using using 200 μM PA in which case the ratio was 1 to 2 . 65 ) . Cells were then cultivated in this serum-free media containing the fatty acids for 24 h prior to analysis , and with 400 μM PA or OA being used unless stated otherwise . FCCP was dissolved in ethanol to produce a 30 mM stock and used at a concentration of 10 μM in serum-free-media for 4 hours prior to analysis . Error bars for worm length measurements show the standard error of the mean , and t-tests were used to identify significant differences between worm lengths . SFA/MUFA ratios were normalized using a logN conversion prior to test for significance using a t-test . Error bars for the frequency of the tail tip defect show the 95% confidence interval and significant differences determined using Z-tests . t-tests were also used to determine significance in FRAP experiments , and the lipidomics data in HEK293 cells was analyzed using ANOVA and a Dunnett's multiple comparisons test . All experiments were repeated several times with similar results . Asterisks are used in the figures to indicate various degrees of significance , where *: p<0 . 05; **: p<0 . 01; and ***: p<0 . 001 . | Our cells and their internal organelles are bound by membranes composed primarily of phospholipids , i . e . polar molecules containing two fatty acids attached to a hydrophilic head group . The types of fatty acids in phospholipids greatly influence membrane properties: saturated fatty acids make the membranes rigid while unsaturated fatty acids promote fluidity . The fact that dietary fats can be incorporated into cellular membranes poses a serious challenge to the cells: how to regulate membrane composition to compensate for dietary variations ? For the present study we used bacteria mutants with different fat compositions as food sources for the nematode C . elegans , together with assays to determine membrane rigidity and composition , and discovered that the C . elegans membrane protein PAQR-2 is responsible for detecting membrane rigidification by dietary saturated fatty acids and to promote fatty acid desaturation to restore membrane fluidity . We also studied the human homolog of PAQR-2 , a protein called AdipoR2 , and showed that it too is essential to prevent membrane rigidification by saturated fatty acids . AdipoR2 and PAQR-2 therefore serve a critical function in cells by acting as evolutionarily conserved regulators of membrane properties . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"carbohydrate",
"metabolism",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"caenorhabditis",
"gene",
"regulation",
"light",
"microscopy",
"glucose",
"metabolism",
"animals",
"toxicology",
"animal",
"models",
"toxi... | 2017 | The adiponectin receptor AdipoR2 and its Caenorhabditis elegans homolog PAQR-2 prevent membrane rigidification by exogenous saturated fatty acids |
There are no effective therapeutics that antagonize or reverse the protein-misfolding events underpinning polyglutamine ( PolyQ ) disorders , including Spinocerebellar Ataxia Type-3 ( SCA3 ) . Here , we augment the proteostasis network of Drosophila SCA3 models with Hsp104 , a powerful protein disaggregase from yeast , which is bafflingly absent from metazoa . Hsp104 suppressed eye degeneration caused by a C-terminal ataxin-3 ( MJD ) fragment containing the pathogenic expanded PolyQ tract , but unexpectedly enhanced aggregation and toxicity of full-length pathogenic MJD . Hsp104 suppressed toxicity of MJD variants lacking a portion of the N-terminal deubiquitylase domain and full-length MJD variants unable to engage polyubiquitin , indicating that MJD-ubiquitin interactions hinder protective Hsp104 modalities . Importantly , in staging experiments , Hsp104 suppressed toxicity of a C-terminal MJD fragment when expressed after the onset of PolyQ-induced degeneration , whereas Hsp70 was ineffective . Thus , we establish the first disaggregase or chaperone treatment administered after the onset of pathogenic protein-induced degeneration that mitigates disease progression .
Many neurodegenerative diseases , such as Alzheimer's Disease , Parkinson's Disease ( PD ) , prion disease , and the collection of polyglutamine ( PolyQ ) disorders , including Huntington's Disease ( HD ) and the Spinal Cerebellar Ataxias ( SCAs ) , are characterized by the formation of protein inclusions in the nervous system [1]–[3] . Moreover , despite vastly different primary sequences , many of the proteins implicated in these diseases adopt the stereotypical amyloid conformation in the aggregated state [1] . Amyloid is defined by a highly stable cross-β conformation , in which proteins polymerize via intermolecular contacts of β-strands that align orthogonal to the fiber axis . Amyloid is typically a stable structure that is resistant to denaturation by heat , detergents ( up to 2% sodium dodecyl sulfate ( SDS ) ) , and proteases [2] , [4] . Despite the extraordinary structural stability of amyloid , a protein disaggregase from yeast , Hsp104 , can rapidly solubilize amyloid . Hsp104 is a hexameric AAA+ ( ATPases Associated with diverse cellular Activities ) protein that couples ATP hydrolysis to translocation of substrate through a central pore , thus prying individual monomers from the amyloid fiber [5]–[8] . In yeast , Hsp104 is a heat shock protein ( HSP ) , promoting survival following stresses by resolubilizing denatured protein aggregates and restoring proteins to native form and function [9] , [10] . Hsp104 also maintains beneficial prion states by controlling the disassembly and dissemination of amyloid aggregates [11]–[13] . Curiously , Hsp104 has no homologue in metazoa . Indeed , until recently it was unclear whether the metazoan proteostasis network possessed any coupled protein disaggregase and reactivation machinery . It is now clear that Hsp110 , Hsp70 , and Hsp40 collaborate to promote the dissolution and reactivation of disordered aggregates [14] , [15] , and can even slowly depolymerize amyloid fibrils from their ends [16] . However , these disaggregase activities are slow and ineffective compared to Hsp104 [15] , [16] . In particular , amyloid depolymerization by Hsp110 , Hsp70 , and Hsp40 is many orders of magnitude slower ( weeks versus minutes ) than amyloid dissolution by Hsp104 [16] . Importantly , Hsp104 can synergize with metazoan Hsp110 , Hsp70 , and Hsp40 to promote dissolution of amyloid and nonamyloid aggregates [15] , [16] . Thus , introduction of Hsp104 into an animal system may provide an unprecedented opportunity to directly and rapidly target the intractable protein aggregates that underlie amyloid diseases [17] , [18] . Spinocerebellar Ataxia Type 3 or Machado-Joseph Disease ( MJD/SCA3 ) is the most prevalent dominantly inherited ataxia [19] , [20] . The genetic basis of MJD/SCA3 is an expansion of the polyglutamine ( PolyQ ) tract of ataxin-3 ( also known as Machado-Joseph Disease protein; MJD ) . When the PolyQ tract surpasses 50 consecutive Qs it is associated with the formation of amyloid aggregates and development of disease [21]–[23] . The normal physiological function of MJD is as a deubiquitylase ( DUB ) that catalyzes the cleavage of polyubiquitin ( poly-ub ) chains to promote proteostasis . It has a chain-editing function , preferentially cleaving certain poly-ub linkages to increase the presence of poly-ub chains that signal for degradation via the ubiquitin-proteasome system ( UPS ) [24] , [25] . MJD has DUB activity in the N-terminal Josephin domain , plus two ubiquitin-interacting motifs ( UIMs ) that present poly-ub chains to the Josephin domain , as well as the C-terminal PolyQ tract that is associated with disease [26] . The PolyQ domain is known to form amyloid fibers , and interestingly , MJD aggregation occurs in a two-step process in vitro , with the Josephin domain forming SDS-soluble linear polymers that then convert into SDS-insoluble PolyQ-driven amyloid fibers [27]–[29] . As such , Hsp104 may be well suited to combating MJD protein aggregation because it antagonizes non-amyloid aggregates , pre-amyloid conformers , and amyloid fibers [5] , [13] , [30] , [31] . Hsp104 has been introduced to combat protein-aggregation disease in metazoan systems with various levels of success [31]–[35] . In C . elegans , Hsp104 prevented aggregation and toxicity of GFP-tagged PolyQ [33] . In a lentiviral rat model , co-expression of Hsp104 with a PolyQ fragment implicated in HD resulted in the accumulation of more but smaller aggregates and rescue of striatal dysfunction [32] . In mouse , animals transgenic for both an HD fragment and Hsp104 showed limited suppression of PolyQ inclusion formation and a lifespan prolonged by ∼20% [34] . While these studies suggest promise for Hsp104 as a therapeutic against disease-associated protein aggregation , none has provided mechanistic insight into how Hsp104 interacts with amyloidogenic proteins in an animal system . Further , studies to date have looked only at antagonism of aggregation by concomitant co-expression of Hsp104 . There has not been an evaluation of the potential of Hsp104 to modulate disease phenotypes in vivo after aggregates have already formed and degeneration has begun; a situation likely to mimic an actual therapy . Therefore , we created novel Hsp104 Drosophila lines to exploit well-characterized models of disease in combination with powerful genetic tools to temporally control the expression of Hsp104 after disease-associated aggregation and degeneration has begun . Our studies reveal surprisingly distinct interactions of Hsp104 with the full-length versus a truncated version of the MJD protein . Importantly , we establish that Hsp104 possesses the ability to suppress the progression of degeneration when activated subsequent to onset of expression of the disease protein . These data indicate that protein context is central in Hsp104 interactions , and that Hsp104 displays the ability to halt the progression of pre-established disease in vivo .
The disaggregase Hsp104 efficiently antagonizes protein aggregates in yeast , and while homologues are present in bacteria , plants , fungi , chromista , and protozoa , no functional homologue has been found in metazoa [17] , [36] . We stably introduced Hsp104 into Drosophila to evaluate its ability to prevent and potentially reverse aggregation of disease-associated human proteins , readily available in various fly models of disease . To achieve strong expression of the Hsp104 protein in the fruit fly , we codon-optimized the transgene for Drosophila ( see Materials and Methods ) , and added a fly-optimal Kozak sequence ( ACAAA ) before the start codon [37] . The Hsp104 transgene was then expressed in Drosophila using the GAL4/UAS system [38] . Because we achieved high expression of Hsp104 , expression by the gmr-GAL4 driver in the eye had a mild disruptive effect ( Fig . 1A ) , which has also been observed for another AAA+ protein , p97 [39] . As the gmr-GAL4 driver line has multiple copies of the glass gene element for driving GAL4 expression , we instead used a driver line with reduced expression ( Fig . 1B ) bearing only a single glass element , 1×gr-GAL4 . Using this driver , the effect of Hsp104 was minimized ( Fig . 1A and Fig . 2A ) . Thus , we used the 1×gr-GAL4 driver line for our experiments to evaluate the impact of Hsp104 on protein-aggregation disease in vivo . Hsp104 dissolves PolyQ amyloid in vitro [5] , [16] and has been expressed in various PolyQ animal models , with results ranging from minimal beneficial effect to strong abrogation of PolyQ aggregation [32]–[34] . However , a detailed analysis of the underlying protein interactions is lacking in vivo . We sought to dissect the ability of Hsp104 to antagonize PolyQ aggregation and toxicity using MJD as a model protein . Pathogenic MJD with expanded PolyQ has been previously established in fly models of MJD/SCA3 , and induces progressive neurodegeneration with the formation of nuclear inclusions [40] , [41] . We also examined a truncated C-terminal fragment of MJD that is predominantly comprised of the PolyQ tract because fragmentation of the protein may be associated with MJD/SCA3 pathogenesis [40] , [42] , [43] . We examined interactions of Hsp104 with the pathogenic , full-length MJD containing an expanded glutamine tract ( MJDnQ78 ) and the truncated C-terminal region of the protein containing the expanded glutamine tract ( MJDtrQ78 ) [40] . Hsp104 had no effect on non-pathogenic forms of MJD containing non-expanded PolyQ tracts ( Fig . 2B ) , confirming that the interaction is PolyQ length-dependent . With expanded PolyQ domains , the pathogenic MJDtrQ78 and MJDnQ78 both caused degeneration of the external eye and disruption to internal retinal structure ( Fig . 2C , D ) . Unexpectedly , we found that Hsp104 had opposite effects on these two forms of MJD that have an identical PolyQ expansion: Hsp104 mitigated MJDtrQ78 degeneration ( Fig . 2C ) , yet enhanced degeneration associated with the full-length MJDnQ78 ( Fig . 2D ) . This effect is in contrast to human Hsp70 , a molecular chaperone that suppresses PolyQ disease in multiple systems [44]–[47] . Despite more severe degeneration due to stronger expression by the gmr-GAL4 driver , Hsp70 suppressed the toxicity of both MJDtrQ78 and MJDnQ78 ( Fig . 2E ) . To probe the mechanism underlying the dichotomous results found for the Hsp104 interaction with MJDtrQ78 and MJDnQ78 , an in-depth investigation of the protein aggregates was performed . To slow protein aggregation such that we could analyze underlying protein accumulations in detail , we expressed the transgenes in the eye with an adult-onset driver rhodopsin1 ( rh1 ) -GAL4 . Analysis of the PolyQ protein accumulations showed that Hsp104 altered the kinetics of inclusion formation for both MJD protein isoforms . By cryosectioning and subsequent immunohistochemistry ( IHC ) , MJDtrQ78 formed compact inclusions that increased in size over time ( Fig . 3A , top row ) . Quantification of inclusion size over time ( Fig . 3A , gray bars in graph ) reveals distinct inclusion size populations ( small <2 . 5 µm , medium 2 . 5–5 µm , large >5 µm in optical diameter ) , demonstrating that inclusions became larger and more numerous with time . Consistent with previous studies , co-expression of Hsp70 delayed the kinetics and significantly reduced MJDtrQ78 protein aggregation ( Fig . 3A , bottom row; green bars in graph , p = 0 . 03 ) . By contrast , Hsp104 initially delayed inclusion formation ( p = 0 . 004 , but then significantly enhanced the formation of small inclusions ( p = 0 . 04 ) , eventually reaching accumulation levels similar to that with MJDtrQ78 alone ( Fig . 3A , center row; red bars in graph , n . s . p = 0 . 5 ) . To examine protein accumulation by biochemical methods , we used SDD-AGE ( Semi-Denaturing Detergent–Agarose Gel Electrophoresis ) , a protein agarose gel technique that can resolve amyloid aggregates [48] . This technique is useful for resolving high molecular weight polymer assemblies that maintain stable contacts in 2% SDS ( a feature of highly stable amyloid ) . SDD-AGE revealed that the truncated MJDtrQ78 protein formed SDS-resistant amyloid structures that accrue with time ( Fig . 3C ) . Unlike Hsp70 , which significantly suppressed amyloid formation ( p = 0 . 003 ) , Hsp104 did not change the overall kinetics of MJDtrQ78 amyloid formation or the overall level of aggregation ( Fig . 3C ) . Confirming that insoluble amyloid material was increased , the reduction of SDS-soluble levels of protein by immunoblot matched the concomitant increase in amyloid formation observed by SDD-AGE ( Fig . 3C ) . Thus , Hsp104 rescues MJDtrQ78 toxicity , but the relationship to MJDtrQ78 aggregation is complex . IHC revealed that Hsp104 initially delays MJDtrQ78 inclusion formation , but then significantly enhances the formation of small inclusions ( Fig . 3A ) . However , when amyloidogenesis was tracked by SDD-AGE , Hsp104 affected neither the rate nor the extent of amyloid formation ( Fig . 3C ) . This finding indicates that to rescue toxicity Hsp104 might reduce formation of soluble and toxic oligomeric MJDtrQ78 species that are populated during amyloidogenesis , just as it does with the yeast prion proteins Sup35 and Ure2 [12] , [13] . Next , we assessed MJDnQ78 misfolding . In contrast to the truncated MJDtrQ78 isoform , the pathogenic full-length MJDnQ78 initially formed amorphous inclusions that did not become more numerous after day 1 ( Fig . 3B , top row; gray bars in graph ) . These MJDnQ78 amorphous aggregates appeared early by IHC , and insoluble amyloid aggregates developed later as observed by SDD-AGE ( Fig . 3B and D ) . Thus , the early MJDnQ78 aggregates are non-amyloid in nature but later convert into the insoluble amyloid structure , closely resembling the two-step aggregation kinetics observed in vitro [27]–[29] . As with MJDtrQ78 , co-expression of Hsp70 delayed the kinetics of aggregation and significantly suppressed inclusion formation ( Fig . 3B , bottom row; green bars in graph , p = 0 . 02 ) . However , in marked contrast , co-expression of Hsp104 significantly increased the formation of large aggregates at early time points ( p = 0 . 002 ) , and then significantly increased the number of small inclusions over time ( Fig . 3B , center row; red bars in graph , p<0 . 001 ) . Consistent with IHC results , SDD-AGE analysis demonstrated that Hsp104 significantly promoted the early formation of insoluble MJDnQ78 amyloid aggregates ( Fig . 3D , p<0 . 001 ) , whereas Hsp70 delayed kinetics of amyloid formation ( Fig . 3D , p = 0 . 01 ) . Hsp70 , but not Hsp104 , stably colocalized with both MJDtrQ78 and MJDnQ78 inclusions ( Fig . 3A and B; see Fig . 4 for channel breakdown ) . The striking contrast between the effects of Hsp104 and Hsp70 on inclusion formation reinforces their functional differences [30] , [49] . While it is known that in select conditions , Hsp104 promotes amyloid formation of specific yeast prions [12] , [13] , [30] , we did not anticipate that Hsp104 would have opposite actions on two constructs of the same PolyQ protein . Thus , we assessed which domains of the full-length MJD protein prevented rescue by Hsp104 by employing a series of expression-matched MJD variants with disruptions to specific motifs ( Fig . 5 , 6 , and see summary in Fig . 7A ) . Because the PolyQ domains are pure CAG repeats , they are subject to instability . Given this instability , the repeat lengths have been matched as closely as possible with matching protein expression levels ( see [41] ) . Because of the reduced expression level by the 1×gr-GAL4 driver used for these experiments , MJDnQ84 ( the pathogenic protein for this set of expression matched proteins ) now showed mild degeneration . However , as before , analysis of retinal integrity demonstrated that the toxicity of MJDnQ84 was enhanced upon co-expression of Hsp104 ( Fig . 5A ) . In contrast , an MJD variant in which both UIMs were mutated and unable to engage poly-ub , MJD-Q80-UIM* ( Fig . 7A ) , exhibited mild toxicity that was suppressed by Hsp104 ( Fig . 5A ) . Thus , the ability of the UIM domains to engage poly-ub hinders protective Hsp104 activity . We also examined variants lacking DUB activity through mutation of the active site in the Josephin domain , MJD-Q88-C14A ( Fig . 7A ) , which causes more severe toxicity than MJDnQ84 due to the loss of the physiological UPS function [41] . This occurs because DUB activity of MJD can suppress its own PolyQ toxicity; the C14A mutation is innocuous when MJD has a normal length , non-expanded Q repeat ( Fig . 6A ) [41] . Co-expression of Hsp104 did not affect the severe MJD-Q88-C14A toxicity ( Fig . 5A ) . However , when the active site mutation was combined with the UIM mutations , in MJD-Q80-C14A-UIM* ( Fig . 7A ) , Hsp104 now suppressed toxicity ( Fig . 5A ) . This result reiterates that functional UIMs hinder rescue by Hsp104 . We further examined a separate splice variant lacking DUB activity through an exon deletion that includes the active site , MJD-Q79-Δexon2 ( missing amino acids 9–63 ) ( Fig . 6B , 7A ) [50] , [51] , which , like MJD-Q88-C14A , conferred severe toxicity ( Fig . 5A ) . Co-expression of Hsp104 with MJD-Q79-Δexon2 strongly suppressed degeneration ( Fig . 5A ) . Thus , Hsp104 mitigated MJD toxicity when the exon containing the active site was deleted ( MJD-Q79-Δexon2 ) but had no effect when the active site was inactivated by a single point mutation ( MJD-Q88-C14A ) . Taken together , these data indicate that functional UIMs and an intact Josephin domain both prohibit Hsp104 from rescuing full-length MJDnQ84 toxicity . To uncover additional mechanistic insight into the interactions with Hsp104 , we examined inclusion formation and kinetics with adult-onset rh1-GAL4 expression . By IHC , the MJD variants formed accumulations in a manner roughly consistent with severity of eye degeneration ( Fig . 5B , top row; gray bars in graph ) . Those variants with mutated UIMs , MJD-Q80-UIM* and MJD-Q80-C14A-UIM* , showed significantly reduced levels of aggregate formation with Hsp104 ( Fig . 5B , bottom row; red bars in graph , p = 0 . 003 for both ) . SDD-AGE analysis revealed that Hsp104 significantly enhanced the conversion of soluble protein to SDS-resistant polymers for variants with intact UIMs: MJDnQ84 and MJD-Q88-C14A ( Fig . 5C , p = 0 . 002 and p = 0 . 005 , respectively ) . Moreover , Hsp104 significantly reduced the formation of amyloid material by MJD-Q80-C14A-UIM* ( Fig . 5C , p = 0 . 01 ) . The MJD-Q79-Δexon2 protein was not detectable by immunoblot ( but was confirmed by genotyping , Fig . 6C , D ) , precluding aggregate analysis of this variant . These findings underscore the role of active ubiquitin binding in obstructing productive remodeling by Hsp104 . A summary of the effect of Hsp104 on MJD variants is presented in Fig . 7B . In the two cases in which Hsp104 enhanced aggregation ( MJDnQ84 , MJD-Q88-C14A ) , the MJD protein has both functional UIMs and an intact Josephin domain . By contrast , protein variants whose toxicity and underlying protein accumulations were suppressed by Hsp104 ( MJD-Q80-UIM* , MJD-Q80-C14A-UIM* , MJD-Q70-Δexon2 ) each lack UIM binding or a portion of the Josephin domain . This supports a model in which an inflexible or “closed” loop is formed , possibly through associations with a poly-ub chain , between the functional UIMs and the intact Josephin domain ( Fig . 7B , red circle ) . Our hypothesis is that Hsp104 is able to effectively remodel a more flexible or “open” conformation of select protein variants ( e . g . , MJD-Q80-UIM* , MJD-Q80-C14A-UIM* , MJD-Q70-Δexon2 , or the truncated MJDtrQ78 ) , but that the inflexible/closed conformation of other proteins ( e . g . , MJDnQ78 , MJDnQ84 , or MJD-Q88-C14A ) obstructs protective Hsp104 activities . To verify the critical role of active remodeling by Hsp104 , we created an ATPase-Dead and substrate-binding defective Hsp104 transgenic fly . We introduced four mutations ( Y257A:E285Q:Y662A:E687Q ) into Hsp104 to ensure that Hsp104 could not engage substrate or hydrolyze ATP , creating the mutant known as Double Pore Loop Double Walker B ( Hsp104DPLDWB ) , which is structurally identical to wild-type but functionally inactive [5] . Unlike wild-type Hsp104 , which caused mild retinal disruption by the gmr-GAL4 driver , similar expression levels of Hsp104DPLDWB were innocuous ( Fig . 8A , B ) . Moreover , Hsp104DPLDWB did not modulate the toxicity of either MJDtrQ78 or MJDnQ78 ( Fig . 8A ) , underscoring the importance of substrate translocation for Hsp104 to mitigate MJDtrQ78 or worsen MJDnQ78-associated degeneration . Thus , ATPase activity and substrate binding are required in vivo for modulatory effects of Hsp104 . Hsp104 is unique in its capacity to reverse pre-existing amyloids in yeast and in vitro [5] , [52] . However , the potential of Hsp104 to affect pre-existing protein-aggregation disease in a metazoan , i . e . , a genuine in vivo treatment situation , has never been addressed . To address this deficit , we constructed fly lines containing three elements: ( 1 ) the toxic MJDtrQ78 protein driven directly by a gmr element such that the disease-associated protein was constitutively expressed in the eye; ( 2 ) a drug-inducible gmr-GAL4 driver known as “GeneSwitch” ( gmr-GS ) to activate GAL4 expression only in the presence of the drug RU486 ( mifepristone ) [53] , [54]; and ( 3 ) the UAS-HSP treatment molecule ( here , Hsp104 or Hsp70 ) , such that the HSP will be expressed conditionally only when RU486 is present in the fly food ( Fig . 9A ) . This system allows the activation of HSP expression sequential to disease-associated protein onset . In this manner , we could test the ability of exogenous HSPs to mitigate the toxicity of the pathogenic PolyQ protein after the pathogenic protein was already accumulating in aggregated forms and causing degeneration . We hypothesized that , due to its disaggregation rather than chaperone activity , Hsp104 may have the potential to markedly mitigate degeneration associated with pre-existing PolyQ protein aggregates . We established that retinal degeneration associated with gmr-MJDtrQ78 had begun at the time of adult fly emergence ( d0 ) and progressed in severity to d7 ( Fig . 9B , Fig . 10 ) . We then activated Hsp104 or Hsp70 expression at an early time point ( d1 ) , or a later time point ( d3 ) , and examined the pathogenic impact of the MJDtrQ78 protein at d7 by retinal section . When activated at d1 , Hsp104 was able to significantly mitigate retinal degeneration associated with MJDtrQ78 ( p = 0 . 001 ) , while Hsp70 did not have a significant effect ( Fig . 10A , n . s . p = 0 . 06 ) . These data show that Hsp104 is significantly more effective than Hsp70 at mitigating toxicity once disease progression has begun ( Fig . 10A , p = 0 . 01 ) . Importantly , inactive Hsp104DPLDWB had no effect ( Fig . 10A ) . Hsp104 significantly improved tissue structure even when expression was induced at a later time point of d3 when degeneration was even more severe ( Fig . 10B , p = 0 . 003 ) . Moreover , while MJDtrQ78 treated with Hsp70 continued to degenerate with time , induction of Hsp104 arrested disease progression ( Fig . 10A , d7 vs d1; Fig . 10B , d7 vs d3 ) . Thus , Hsp104 mitigates pathogenesis even when administered after the onset of pathogenic protein-induced degeneration . Next , we examined the underlying protein aggregates by SDD-AGE and Western immunoblot . We observed that gmr-MJDtrQ78 had high levels of amyloid , and this was lessened with time ( potentially due to tissue loss ) ( Fig . 11 ) . When turned on at d1 , Hsp104 did not reverse MJDtrQ78 amyloid formation , but rather significantly increased the amyloid present by d7 ( p = 0 . 02 ) , while Hsp70 had no effect ( Fig . 11A ) . Neither molecule significantly altered amyloid load when turned on at d3 ( Fig . 11B ) . These results imply that Hsp104 is not acting as a MJDtrQ78-amyloid disaggregase , but rather is mitigating toxicity in a distinct manner , which is also consistent with our earlier results where MJDtrQ78 and Hsp104 are expressed at the same time ( see Fig . 2 , 3 ) . Intriguingly , representative densitometry traces for the amyloid smears for each treatment condition turned on at d1 suggest that the peak of MJDtrQ78 amyloid species shifts downward to indicate smaller amyloid accumulations by d3 following Hsp104 activation ( Fig . 11C ) . No shift in the densitometry trace is observed for the control or Hsp70 treatment ( Fig . 11C ) , suggesting that Hsp104 is indeed altering the character of amyloid species although not eliminating these fibrils completely . In summary , this novel system of temporally controlled HSP expression demonstrates that although concomitant expression of Hsp70 is more successful than Hsp104 at preventing degeneration ( see Fig . 2 ) , inducible expression of Hsp104 is more effective than Hsp70 at suppressing disease progression once protein aggregation and degeneration are already established ( Fig . 10 ) .
Our detailed investigations of the effects of Hsp104 on the MJD protein led to the unexpected result that Hsp104 has opposite effects on the toxicity of different versions of the MJD protein ( MJDtrQ78 and MJDnQ78 ) , despite the fact that these proteins contain the same pathogenic PolyQ stretch . These disparate actions indicate that Hsp104 might be a useful probe to understand the nature of aggregates and the toxicity imposed by them . Hsp70 suppressed both MJDtrQ78 aggregation and toxicity ( Fig . 2E , 3C ) . By contrast , Hsp104 mitigated toxicity of MJDtrQ78 without suppressing the extent of MJDtrQ78 aggregation ( Fig . 2C , 3C ) . This result suggests that MJDtrQ78 aggregation per se need not be deleterious . Uncoupling of aggregation and toxicity has also been observed in other settings . For example , numerous genetic suppressors of FUS and TDP-43 toxicity , which are connected with amyotrophic lateral sclerosis and frontotemporal dementia , rescue toxicity without affecting FUS or TDP-43 aggregation in yeast [55]–[57] . We suggest that Hsp104 likely mitigates toxicity of MJDtrQ78 accumulations via subtle biochemical changes rather than gross changes in aggregation levels . What might these biochemical changes be ? Hsp104 can disrupt toxic soluble oligomers of various proteins , including Sup35 , which may help explain why Sup35 prion formation is not intrinsically toxic to yeast [5] , [12] , [13] , [31] , [58] . Thus , Hsp104 might eliminate toxic soluble oligomers formed by MJDtrQ78 just as it does for Sup35 and alpha-synuclein [5] , [12] , [13] , [31] . Furthermore , the amyloid-remodeling activity of Hsp104 can selectively amplify some amyloid strains ( i . e . different cross-β structures formed by the same polypeptide ) at the expense of others [59] . Given that PolyQ can access both toxic and benign amyloid strains [60] , it is plausible that the presence of Hsp104 might amplify benign amyloid strains of MJDtrQ78 at the expense of toxic strains [59] , [60] . Indeed , we observed that Hsp104 activity visibly altered the distribution of MJDtrQ78 amyloid species ( Fig . 11C ) suggesting that Hsp104 may be selectively eliminating certain fibril strains . Finally , to promote toxicity amyloid structures typically sequester large metastable proteins with unstructured regions , which occupy key nodes in functional networks linked to transcription , translation , chromatin organization , cell structure , and proteostasis [61] , [62] . Hsp104 might disaggregate and rescue these proteins sequestered by MJDtrQ78 amyloid or promote the formation of MJDtrQ78 amyloid strains that do not deplete such an essential constellation of proteins . Further studies are needed to distinguish these non-mutually exclusive possibilities . In other settings , it has been suggested that chaperone-initiated formation of large , insoluble amyloid aggregates can actually be protective by sequestering potentially toxic pre-fibrillar conformers [63]–[65] . Our results , however , indicate that , at least for Hsp104-driven enhancement of MJDnQ78 , increased and accelerated aggregation is more toxic than MJDnQ78 aggregation that occurs in the absence of Hsp104 . Our findings illustrate the complex relationship between aggregation and toxicity , which likely extends to other neurodegenerative disease models [64] . Moreover , our studies suggest that cautious interpretation is required when translating findings from cell culture experiments to neurodegeneration in animal models [34] , [66] . Although Hsp104 is not found in the metazoan proteostasis network , our observations could help inform how to manipulate existing components of the metazoan proteostasis network for therapeutic purposes . Thus , components that suppress MJDnQ78 aggregation are likely beneficial , whereas MJDtrQ78 toxicity can be mitigated without having to suppress MJDtrQ78 aggregation . Moreover , our findings demonstrate that an agent with a mitigating effect on the truncated version of the MJD protein may act in a different manner against the full-length MJD protein . Thus , what is good for one may not be beneficial for the other . In MJD/SCA3 , as well as other neurodegenerative diseases , fragmentation of the disease protein may initiate aggregation and this process is critical for disease progression [42] , [43] , [67] , [68] . Therefore , agents that effectively eliminate one specific sub-population of toxic protein accumulation but enhances another toxic sub-population may not be therapeutically viable in the complicated mixed populations that occur in disease . Our results highlight the complexity in developing therapeutic agents for neurodegenerative disorders . Although Hsp104 enhances MJDnQ78 amyloidogenesis and toxicity , we found that elimination of functional domains not implicated in PolyQ aggregation facilitated the ability of Hsp104 to suppress MJD-associated degeneration . Elimination of UIM functionality or removal of a component of the Josephin domain ( exon 2 ) restored the remodeling capacity of Hsp104 . This suggests that MJDnQ78 pathogenicity is not intrinsically intractable , but is capable of being suppressed by Hsp104 if other domains of the protein are inactivated ( e . g . , the UIMs ) . Alternatively , potentiated or MJDnQ78-optimized Hsp104 variants might be developed that are able to overcome these hindrances via increased unfolding power [5] , [17] . Our studies underscore the importance of protein context in studying protein-misfolding diseases . Within the protein itself , neighboring domains not thought to be involved in aggregation may be impacting accumulation kinetics and the biochemical properties of inclusions [69] , as well as accessibility of the aggregation domain to potential disaggregase therapeutics . That Hsp104 efficiently mitigates toxicity of MJD variants with ubiquitin-binding defects also demonstrates that , in addition to protein context , the cellular context of the protein is critical to consider; for example , the interaction between poly-ub chains and MJD may hinder protective Hsp104 modalities . Previous studies in vitro have characterized aggregation of the full-length , pathogenic MJD protein as a two-step process in which the protein assembles first into SDS-soluble fibrillar polymers associating via the Josephin domain , and then converts to SDS-insoluble amyloid fibers driven by the PolyQ domain [27]–[29] , [70] . We propose that this two-step process occurs in vivo as well . Indeed , it is consistent with our observation that full-length MJDnQ78 forms amorphous accumulations that appear visually by IHC before they can be observed as SDS-insoluble amyloid aggregates by SDD-AGE ( see Fig . 3B and D ) . We suspect that full-length MJD initially forms SDS-soluble , Josephin-driven non-amyloid accumulations that initiate Hsp104 remodeling . The initial formation of non-amyloid polymers is also compatible with our model of Hsp104 interacting differentially with the “open” and “closed” conformations of the protein discussed above ( see Fig . 7B ) . We hypothesize that a poly-ub chain creates the closed loop by mutually interacting with the UIMs and the flexible helical hairpin encoded by exon 2 in the Josephin domain ( Fig . 6B ) , as this arm is thought to be important for interacting with substrates [71] . Further experiments are required to confirm a poly-ub-mediated interaction between the two domains . According to our model , Hsp104 is able to efficiently translocate and release proteins that are more flexible ( e . g . , MJD-Q80-UIM* ) , resulting in fewer aggregates ( see Fig . 5B ) . But due to an inflexible conformation imposed by the UIMs and the Josephin domain , Hsp104 is unable to efficiently remodel proteins containing the closed loop ( e . g . , MJDnQ84 ) . This incomplete or slow translocation may expose or “prime” the PolyQ region to drive the formation SDS-insoluble amyloid inclusions ( see Fig . 3D , Fig . 5C ) . Our model suggests that the UIMs and the Josephin domain act together to obstruct Hsp104 remodeling , but we cannot rule out a separate function of the Josephin domain outside of poly-ub interactions . For example , removal of 55 amino acids might destabilize the Josephin domain such that it gets proteolytically cleaved and Hsp104 would then encounter a protein similar to MJDtrQ78 and rescue toxicity . Alternatively , because the MJD-Q79-Δexon2 protein could not be detected by biochemical methods , we cannot exclude the possibility that the deletion within the Josephin domain disrupts the proposed process of Josephin-domain-driven polymerization . In this case , toxicity of this variant may be dependent on highly soluble , possibly oligomeric species , which are effectively targeted by Hsp104 . These findings indicate that there is an opportunity to tailor therapies that are optimized for a specific disease scenario . In the case of full-length MJD , if inefficient translocation by Hsp104 does indeed drive the switch from less toxic SDS-soluble aggregates to highly toxic SDS-insoluble amyloid inclusions , then the development of substrate-optimized Hsp104 mutants ( or Hsp104 mutants with altered ATPase rates or unfolding power ) may increase efficiency of such interactions and enable Hsp104 to rescue disease phenotypes . Moreover , if UIM binding to poly-ub chains is impairing access of Hsp104 to MJD , this suggests that co-administering an agent to modulate function of a neighboring domain may affect the access of a treatment to the aggregation-prone domain . Indeed , increasing global DUB activity coupled with Hsp104 induction could overcome antagonism due to poly-ub chains . Chaperone treatment , and examination of Hsp70 in particular , has been an exciting avenue of research in the battle to combat and contain neurodegenerative disease [72] , [73] . However , all studies investigating Hsp70 as a modulator of disease have looked only at the chaperone transgenically co-expressed or activated prior to the disease insult . In a study seeking to activate existing Hsp70 rather than introducing exogenic expression , researchers sought to evaluate the chaperone in a mouse model of PD by boosting endogenous Hsp70 expression by treating animals with geldanamycin [74] . While this procedure has the potential to test true reversal of disease course , the authors found that beneficial effects were only observed if upregulation of Hsp70 was initiated prior to pharmacological induction of the PD phenotype [74] . In fact , in a cell culture PD model , geldanamycin was required at least 24 hours prior to disease protein transfection to provide any protection against inclusion formation [75] . In addition , pharmacologic activation of Hsp70 has been shown to suppress both PD and PolyQ disease in Drosophila [76] , [77] , but again , these manipulations were performed prior to disease onset . Despite the existing pharmacological paradigms , and other genetic tools available , such as the tetracycline-inducible system in mouse , no group has evaluated specific chaperone or disaggregase expression induced subsequent to expression of a disease-associated protein . An inducible system is particularly well suited for Hsp104 because of its unique ability to rapidly dismantle pre-existing amyloid aggregates . Since metazoan chaperones can only very slowly depolymerize amyloid [16] , Hsp104 may be more effective in an environment with pre-existing aggregation than a chaperone such as Hsp70 , which is more adapted to prevent the initial aggregation . Here , we address the value of temporally controlled induction of disaggregase function after the initiation of PolyQ protein aggregation and the beginning of disease progression . To our knowledge , previous research has been performed with concomitant expression of a therapeutic gene , and thus does not distinguish prevention of disease from halting the progression of the disease state . Neurodegenerative diseases are not detected until later in life , and symptoms may not be apparent until pathological damage has accumulated beyond a tolerable point [78]–[80] . Thus , an agent that can rapidly impact the existing trajectory would be more valuable than one that can only prevent the development of the disease . Our experimental paradigm offers the exciting possibility to address the efficacy of Hsp104 ( or other molecules ) in a more genuine therapeutic setting . Indeed , we found that turning on Hsp104 was able to significantly suppress disease-associated degeneration . Interestingly , however , Hsp104 did not disaggregate MJDtrQ78 amyloid in these experiments ( Fig . 11A , B ) . Thus , Hsp104 might mitigate disease progression in this setting by: ( a ) eradicating toxic soluble MJDtrQ78 oligomers , ( b ) amplifying benign amyloid forms of MJDtrQ78 at the expense of toxic MJDtrQ78 amyloid , ( c ) by disaggregating and rescuing essential metastable proteins sequestered by MJDtrQ78 aggregates . Our observation that activation of Hsp104 shifted the MJDtrQ78 amyloid smears resolved by SDD-AGE toward smaller species without eliminating the total amyloid population ( Fig . 11C ) suggests that Hsp104 may indeed have strain selectivity . Further studies are required to define precisely how Hsp104 mitigates disease progression . Our data show that Hsp70 induction after MJD-associated degeneration has already initiated was unable to significantly mitigate disease progression . Optimization of Hsp70 expression , or administration of the suite of chaperones ( for example , Hsp70 with Hsp110 and Hsp40 ) , may improve the outcome , but our findings are consistent with other reports that Hsp70 must be administered before disease initiation to have a positive effect [74] , [75] . Our observation that even a later-onset induced expression of Hsp104 is able to significantly suppress progressive PolyQ degeneration suggests that it is possible to mitigate disease phenotypes even after aggregates have begun accumulating and marked pathological degeneration is underway . Naturally , several barriers must be surmounted to translate Hsp104 into a therapeutic agent for human neurodegenerative disease [17] , [18] . Not least is the issue that gene therapy might be required to introduce Hsp104 ( or any other genetic modifier ) as a therapeutic agent . Gene therapy has yielded encouraging preclinical results for several disorders including congenital blindness [81]–[83] . However , technical and safety issues restrict facile translation to the clinic . Indeed , gene therapy for neurodegenerative diseases remains in early developmental stages and considerable caution is essential at this time . However , initial studies have generated cautious optimism that gene therapy in the adult brain might be safe for various neurodegenerative disorders , including Parkinson's disease [84]–[88] . Thus , even though we await several key advances before any Hsp104 gene therapy ( or any other gene therapy ) becomes truly viable it is , nonetheless , important to develop solutions to protein misfolding and to test these solutions both in vitro and in the most appropriate animal models . Moreover , the fact that Hsp104 is well tolerated by mammalian systems is encouraging [31] , [32] , [34]–[36] , [66] . Ultimately , we envision that only transient expression of Hsp104 ( or a substrate-optimized variant ) would be required to provide therapeutic benefit . In this way , long-term expression of an exogenous agent and potential off-target side effects would be minimized . Alternatively , methods could be developed to deliver pure Hsp104 ( or a substrate-optimized variant ) to targeted areas in a single or multiple doses , and thereby avoid issues connected with long-term expression . These various issues and others highlight the complexities of designing therapeutics to treat human neurodegenerative disease . Finally , the concept of using a yeast protein as the basis for a therapeutic agent might at first glance seem implausible . However , it must also have seemed equally implausible to use a lethal protein toxin from the bacterium , Clostridium botulinum , as a therapeutic agent . Despite being a deadly toxin , botulinum toxin variants have found key clinical applications due to their highly potent and selective ability to cleave SNARE proteins and prevent secretion [89] . Importantly , they are used to treat a variety of neuromuscular disorders including: blepharospasm , strabismus , muscle spasms , upper motor neuron syndrome , cervical dystonia and chronic migraine [90]–[95] . Indeed , the massive clinical success of botulinum toxin variants suggests it is critical to identify potentially therapeutic biological activities that originate in the microbial world and utilize and develop them to treat human disease .
Transgenic flies expressing UAS-Hsp104 and UAS-Hsp104DPLDWB were generated by standard techniques using the pUAST vector [38] . In order to boost expression of the transgene , pUAST-Hsp104 was codon-optimized ( via gene synthesis; GenScript ) for expression in Drosophila and a Kozak sequence ( ACAAA ) was added prior to the start codon [37] . The full sequence of codon-optimized Hsp104 is: ATGAACGATCAGACCCAGTTCACCGAGCGCGCCCTGACCATCCTGACCCTGGCCCAGAAGCTGGCCAGCGATCACCAGCACCCCCAGCTGCAGCCCATCCACATCCTGGCCGCCTTCATCGAGACCCCCGAGGATGGCAGCGTGCCCTACCTGCAGAACCTGATCGAGAAGGGCCGCTACGATTACGATCTGTTCAAGAAGGTGGTGAACCGCAACCTGGTGCGCATCCCCCAGCAGCAGCCAGCCCCAGCCGAGATCACCCCAAGCTACGCCCTGGGCAAGGTGCTGCAGGATGCCGCCAAGATCCAGAAGCAGCAGAAGGATAGCTTCATCGCCCAGGATCACATCCTGTTCGCCCTGTTCAACGATAGCAGCATCCAGCAAATCTTCAAGGAGGCCCAGGTGGATATCGAGGCCATCAAGCAGCAGGCCCTGGAGCTGCGCGGAAACACCCGCATCGATAGCCGCGGAGCCGATACCAACACCCCCCTGGAGTACCTGAGCAAGTACGCCATCGATATGACCGAGCAGGCCCGCCAGGGAAAGCTGGACCCAGTGATCGGACGCGAGGAGGAGATCCGCAGCACCATCCGCGTGCTGGCCCGCCGCATCAAGAGCAACCCATGCCTGATCGGAGAGCCAGGAATCGGCAAGACCGCCATCATCGAGGGAGTGGCCCAGCGCATCATCGATGATGATGTGCCAACCATCCTGCAGGGAGCCAAGCTGTTCAGCCTGGATCTGGCCGCCCTGACCGCCGGCGCCAAGTACAAGGGCGATTTCGAGGAGCGCTTCAAGGGCGTGCTGAAGGAGATCGAGGAGAGCAAGACCCTGATCGTGCTGTTCATCGATGAGATCCACATGCTGATGGGCAACGGCAAGGATGATGCCGCCAACATCCTGAAGCCAGCCCTGAGCCGCGGACAGCTGAAGGTCATCGGAGCCACCACCAACAACGAGTACCGCAGCATCGTGGAGAAGGATGGAGCCTTCGAGCGCCGCTTCCAGAAGATCGAGGTGGCCGAGCCAAGCGTGCGCCAGACCGTGGCCATCCTGCGCGGACTGCAGCCCAAGTACGAGATCCACCACGGCGTGCGCATCCTGGATAGCGCCCTGGTGACCGCCGCCCAGCTGGCCAAGCGCTACCTGCCATACCGCCGCCTGCCAGATAGCGCCCTGGATCTGGTGGATATCAGCTGCGCCGGAGTGGCCGTGGCCCGCGATAGCAAGCCAGAGGAGCTGGATAGCAAGGAGCGCCAGCTGCAGCTGATCCAGGTGGAGATCAAGGCCCTGGAGCGCGATGAGGATGCCGATAGCACCACCAAGGATCGCCTGAAGCTGGCCCGCCAGAAGGAGGCCAGCCTGCAGGAGGAGCTGGAGCCACTGCGCCAGCGCTACAACGAGGAGAAGCACGGCCACGAGGAGCTGACCCAGGCTAAGAAAAAGCTGGATGAGCTGGAGAACAAGGCCCTGGATGCCGAGCGCCGCTACGATACCGCCACCGCCGCCGATCTGCGCTACTTCGCCATCCCCGATATCAAGAAGCAGATCGAGAAGCTGGAGGATCAGGTGGCCGAGGAGGAGCGCCGCGCCGGCGCCAACAGCATGATCCAGAACGTGGTGGATAGCGATACCATCAGCGAGACCGCCGCCCGCCTGACCGGCATCCCCGTGAAGAAGCTGAGCGAGAGCGAGAACGAGAAGCTGATCCACATGGAGCGCGATCTGAGCAGCGAGGTGGTGGGCCAGATGGATGCCATCAAGGCCGTGAGCAACGCCGTGCGCCTGAGCCGCAGCGGACTGGCCAACCCACGCCAGCCAGCCAGCTTCCTGTTCCTGGGCCTGAGCGGCAGCGGCAAGACCGAGCTGGCCAAGAAGGTGGCCGGCTTCCTGTTCAACGATGAGGATATGATGATCCGCGTGGATTGCAGCGAGCTGAGCGAGAAGTACGCCGTGAGCAAGCTGCTGGGCACCACCGCCGGCTACGTGGGCTACGATGAGGGCGGCTTCCTGACCAACCAGCTGCAGTACAAGCCCTACAGCGTGCTGCTGTTCGATGAGGTGGAGAAGGCCCACCCCGATGTGCTGACCGTGATGCTGCAGATGCTGGATGATGGCCGCATCACCAGCGGCCAGGGCAAGACCATCGATTGCAGCAACTGCATCGTGATCATGACCAGCAACCTGGGCGCCGAGTTCATCAACAGCCAGCAGGGCAGCAAGATCCAGGAGAGCACCAAGAACCTGGTCATGGGCGCCGTGCGCCAGCACTTCCGCCCCGAGTTCCTGAACCGCATCAGCAGCATCGTGATCTTCAACAAGCTGAGCCGCAAGGCCATCCACAAGATCGTGGATATCCGCCTGAAGGAGATTGAGGAGCGCTTCGAGCAGAACGATAAGCACTACAAGCTGAACCTGACCCAGGAGGCCAAGGATTTCCTGGCCAAGTACGGCTACAGCGATGATATGGGCGCCCGCCCCCTGAACCGCCTGATCCAGAACGAGATCCTGAACAAGCTGGCCCTGCGCATCCTGAAGAACGAGATCAAGGATAAGGAGACCGTGAACGTGGTGCTGAAGAAGGGCAAGAGCCGCGATGAGAACGTGCCAGAGGAGGCCGAGGAGTGCCTGGAGGTGCTGCCAAACCACGAGGCCACCATCGGAGCCGATACCCTGGGCGATGATGATAACGAGGATAGCATGGAGATCGATGATGATCTGGATTAA Multiple insertion lines were characterized for each transgene . To create the 1×gr-GAL4 driver line , the pGMR ( glass multimer reporter ) vector [96] was digested by XhoI/Acc651 to remove the insert containing five glass-binding sites . Complementary oligonucleotides ( 5′-TCGAACCCAGTGGAAACCCTTGAAATGCCTTTAACTCGAGACGG-3′ and 5′-GTACCCGTCTCGAGTTAAAGGCATTTCAAGGGTTTCCACTGGGT-3′ ) , with a single copy of the 31 bp glass-binding site from the Rh1 proximal enhancer , were duplexed and ligated into the vector , producing p1×GR ( 1 copy of glass reporter ) . This plasmid was then modified to introduce the GAL4 coding sequence , excised from pGaTN [38] with HindIII , to create p1×gr-GAL4 . MJD lines are from [41] . Experiments were performed at 25°C except for a select few conducted at 29°C as indicated . All results were confirmed with multiple UAS-Hsp104 insertion lines . Eye images were obtained on day 7 of adulthood using a Leica Z-16 apo zoom microscope . To view internal retinal structure , heads were embedded in paraffin according to standard protocols , sectioned at 8 µm , and autofluorescence was viewed with a Leica fluorescence microscope . To quantify tissue loss , a standard area ( 3×15 rectangle; 7 , 000 µm2 ) ( see also Fig . 9 ) was selected within the paraffin retinal section and the percentage of area covered by tissue was measured in ImageJ . Statistical analysis was performed using one-way ANOVA and unpaired t-test . Heads were frozen in Tissue Freezing Medium ( Electron Microscopy Sciences ) and sectioned at 12 µm by cryotome , and the tissue sections were then fixed with 4% paraformaldehyde . Immunohistochemistry was performed according to standard procedures using primary antibodies anti-HA 5B1D10 ( 1∶100 , Invitrogen 32-6700 ) or anti-myc 9E10 ( 1∶100 , Santa Cruz sc-40 ) ( both mouse ) alongside either anti-Hsp104 ( 1∶100 , Enzo Life Sciences ADI-SPA-1040 ) or anti-Hsp70 ( 1∶100 , Enzo Life Sciences ADI-SPA-812 ) ( both rabbit ) . Hsp70 staining was confirmed with human-specific anti-Hsp70 ( 1∶100 , Santa Cruz sc-24 ) ( mouse ) alongside anti-HA Y11 ( 1∶100 , Santa Cruz sc-805 ) or anti-myc A14 ( 1∶100 , Santa Cruz sc-789 ) ( both rabbit ) . Rabbit primary antibodies were preadsorbed at 1∶25 with fixed , dissected wild-type larvae . Secondary antibodies were Alexa Fluor 594 Goat-anti-Mouse IgG ( 1∶100 , Life Technologies A-11032 ) , Alexa Fluor 488 Goat-anti-Rabbit IgG ( 1∶100 , Life Technologies A-11008 ) , Alexa Fluor 594 Goat-anti-Rabbit IgG ( 1∶100 , Life Technologies A-11037 ) , and Alexa Fluor 488 Goat-anti-Mouse IgG ( 1∶100 , Life Technologies A-11029 ) . Sections were co-stained with Hoechst nuclear dye ( 1∶1000 , Molecular Probes 33342 ) and viewed with a Leica fluorescence microscope . A 75 µm×75 µm square ( 5625 µm2 ) area was selected; particle analysis was performed with ImageJ and statistics performed with one-way ANOVA and unpaired t-test . For Hsp104 expression level characterization , heads were ground with a pestle in NuPage LDS Sample Buffer , boiled for 3 min , run on NuPage 4–12% Bis-Tris gel , and semi-dry transferred onto nitrocellulose membrane . Antibodies used were anti-Hsp104 ( 1∶2000 , Enzo Life Sciences ADI-SPA-1040 ) and anti-actin ( 1∶2000 , Abcam ab8227 ) with secondary antibody Goat-anti-Rabbit-HRP ( 1∶5000 , Chemicon AP307P ) . For MJD aggregation analysis through SDD-AGE ( Semi-Denaturing Detergent Agarose Gel Electrophoresis ) and accompanying Western immunoblots , heads were ground in lysis buffer ( 100 mM Tris pH 7 . 5 , 50 mM NaCl , 10 mM β-Mercaptoethanol , and Roche complete mini EDTA-free protease inhibitor cocktail tablets ) [48] and an aliquot was taken for evaluation of soluble material by Western immunoblot , as above . To the remaining sample , 4× Sample Buffer ( 2× TAE , 20% glycerol , 8% SDS , bromophenol blue ) was added to final concentration 1× . The samples were run on a 1 . 5% agarose gel with 0 . 1% SDS in a running buffer of 1× TAE ( 40 mM Tris , 20 mM Acetic acid , 1 mM EDTA , pH 8 . 3 ) containing 0 . 1% SDS , and then transferred overnight onto nitrocellulose membrane using downward capillary transfer [48] . Antibodies used were anti-HA-conj-HRP 3F10 ( 1∶500 , Roche 12013819001 ) , anti-myc 9E10 ( 1∶500 , Santa Cruz sc-40 ) followed by Goat-anti-Mouse-HRP ( 1∶2000 , Jackson ImmunoResearch 115-035-146 ) , and anti-tubulin-conj-HRP ( 1∶1000 , Cell Signaling 11H10 ) . All immunoblots were imaged using a FujiFilm LAS-3000 imaging system and quantification was performed in ImageJ and statistically analyzed by one-way ANOVA and unpaired t-test . A 4 . 0 mg/ml stock solution of RU486 ( Sigma M8046 ) was prepared in 100% ethanol , and then 50 µl ( 200 µg ) was added to pre-prepared food vials containing ∼12 ml of food and gently shaken overnight [97] . For control conditions , 50 µl of 100% ethanol was added to vials . Adult flies were aged in food treated with either RU486 or ethanol for the time periods indicated . | There are no effective therapeutics for any of the neurodegenerative disorders caused by expanded polyglutamine ( PolyQ ) tracts including Spinocerebellar Ataxia Type-3 ( SCA3 ) . These disorders are connected with the misfolding and aggregation of proteins bearing expanded PolyQ tracts in the neurons of affected individuals . In SCA3 , ataxin-3 ( MJD ) is the protein that bears the PolyQ expansion and forms insoluble aggregates . Here , as a therapeutic strategy we introduce Hsp104 , a powerful protein disaggregase from yeast , into Drosophila models of SCA3 . Hsp104 has no homologue in animals , but has an unusual ability to dissolve PolyQ aggregates in vitro , an activity that could be harnessed therapeutically . Indeed , Hsp104 suppressed degeneration caused by a C-terminal ataxin-3 ( MJD ) fragment containing the pathogenic expanded PolyQ tract , which accumulates in disease . However , Hsp104 enhanced aggregation and toxicity of full-length pathogenic MJD . Hsp104 rescued forms of MJD unable to engage polyubiquitin or with a deletion in the deubiquitylase domain indicating that MJD-ubiquitin interactions hinder protective Hsp104 activities . Importantly , Hsp104 suppressed toxicity of a C-terminal MJD fragment when expressed after the onset of PolyQ-induced degeneration , whereas Hsp70 was ineffective . Thus , we establish the first disaggregase or chaperone treatment administered after the onset of pathogenic protein-induced degeneration that mitigates disease progression . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Hsp104 Suppresses Polyglutamine-Induced Degeneration Post Onset in a Drosophila MJD/SCA3 Model |
The category B agent of bioterrorism , Entamoeba histolytica has a two-stage life cycle: an infective cyst stage , and an invasive trophozoite stage . Due to our inability to effectively induce encystation in vitro , our knowledge about the cyst form remains limited . This also hampers our ability to develop cyst-specific diagnostic tools . Three main aims were ( i ) to identify E . histolytica proteins in cyst samples , ( ii ) to enrich our knowledge about the cyst stage , and ( iii ) to identify candidate proteins to develop cyst-specific diagnostic tools . Cysts were purified from the stool of infected individuals using Percoll ( gradient ) purification . A highly sensitive LC-MS/MS mass spectrometer ( Orbitrap ) was used to identify cyst proteins . A total of 417 non-redundant E . histolytica proteins were identified including 195 proteins that were never detected in trophozoite-derived proteomes or expressed sequence tag ( EST ) datasets , consistent with cyst specificity . Cyst-wall specific glycoproteins Jacob , Jessie and chitinase were positively identified . Antibodies produced against Jacob identified cysts in fecal specimens and have potential utility as a diagnostic reagent . Several protein kinases , small GTPase signaling molecules , DNA repair proteins , epigenetic regulators , and surface associated proteins were also identified . Proteins we identified are likely to be among the most abundant in excreted cysts , and therefore show promise as diagnostic targets . The proteome data generated here are a first for naturally-occurring E . histolytica cysts , and they provide important insights into the infectious cyst form . Additionally , numerous unique candidate proteins were identified which will aid the development of new diagnostic tools for identification of E . histolytica cysts .
The parasitic protozoan Entamoeba histolytica is the causative agent of amebic colitis and amebic liver abscesses in humans [1] , [2] . The World Health Organization estimates up to 50 million invasive infections world-wide annually [3] . E . histolytica has a simple , two-stage life cycle , consisting of the infective cyst and colon-invasive trophozoite forms . E . histolytica infections occur when cysts are ingested through contaminated food or water . In the lower intestine trophozoites emerge from cysts ( a process known as excystation ) . As a result of unknown stimuli in the intestine , trophozoites again can differentiate into cysts ( a process known as encystation ) , which may be excreted in feces to infect other humans . Although the cyst is the only form to transmit infections , most studies on E . histolytica have focused on the trophozoite form , which is the only form that can be readily cultured . The inability to encyst trophozoites in vitro has severely impaired our knowledge on the infectious stage of E . histolytica . There is an increasing recognition of the burden of infection due to this protozoan parasite . Carefully conducted serologic studies in Mexico , where amebiasis is endemic , demonstrated antibody to E . histolytica in 8 . 4% of the population [4] . In the urban slum of Fortaleza , Brazil , 25% of the people tested carried antibody to E . histolytica; the prevalence of anti-amebic antibodies in children aged six to fourteen years was 40% [5] . A prospective study of preschool children in a slum of Dhaka , Bangladesh demonstrated new E . histolytica infection in 39% of children over a one year period of observation , with 10% of the children having an E . histolytica infection associated with diarrhea and 3% with dysentery [6] . The diagnosis of E . histolytica infection in endemic areas still relies on microscopy , which is neither sensitive nor specific [7] . PCR-based diagnostic methods have not replaced microscopy in endemic areas , as they require skilled people and sophisticated laboratory settings which are absent in these areas . Although there are simple ( ELISA-based ) diagnostic tools available to detect the trophozoite form of E . histolytica , none are designed to detect the cyst form of parasite . There are two reasons to produce tests that detect the cyst form of E . histolytica . First , the cyst is the infectious form of the parasite , and therefore of greatest importance for detection in potential food- or water-borne outbreaks , both natural and man-made . Second , cyst antigens are expected to be more abundant , stable , and readily detectable in stool samples , including formalin-preserved samples . Lack of such stability is the major limitation of the current E . histolytica antigen-detection test by TechLab [8] . However our understanding of E . histolytica cyst proteins remains the major factor limiting our ability to develop cyst specific diagnostic reagents . Relatively more is known about the cyst stage of the reptilian parasite Entamoeba invadens [9] . E . invadens can be induced to encyst in vitro [10]–[12] , a phenomenon that has made it a model system to study encystation . Studies have identified osmotic shock [13] , low glucose [12] or carbon source [14] , and levels of serum and mucin [15] as potential triggers of encystation in E . invadens in vitro . Additionally , cyst- and cyst-wall specific genes have been identified including chitin synthetase [16] , chitinase [17] , chitosan [18] , and the lectins Jessie and Jacobs [19] . A 70-kDa heat shock protein [20] , and proteins involved in proteasome function [21] were also found to be linked to encystation in E . invadens . E . histolytica strains can undergo spontaneous encystation , although very inefficiently , when grown in presence of bacteria [22] . A pioneering microarray analysis of this process identified about 15% of all genes in the genome as developmentally regulated based on their mRNA transcript levels ( >3-fold change , p-value<0 . 01 ) including 672 genes referred to as cyst-specific and 767 genes referred to as trophozoite-specific . The cyst-specific genes included cysteine proteases , putative DNA-binding or transcription factor-related proteins ( such as Myb domain proteins ) and signal transduction-related transmembrane protein kinases . The promoter motif for one of the Myb domain proteins was later characterized for and this motif appeared to function in the regulation of a subset of cyst-specific genes [23] . In contrast to transcriptomic data , no proteomic data are currently available for E . histolytica cysts . Proteomic analysis of cysts and trophozoites of E . invadens using high resolution 2D PAGE and digitized video image analysis of silver stained gels identified a total of 155 proteins unique to trophozoites and a total of 72 proteins unique to cysts [24] . Lack of knowledge of the E . histolytica cyst proteome hampers our ability to develop cyst-targeted diagnostic tools and to investigate stage conversion in this parasite . In this study , we applied mass spectrometry based whole genome shotgun sequencing approaches to 5 purified cyst samples in order to identify proteins expressed in E . histolytica cysts in vivo . Here we present for the first time the identification of 417 proteins ( representing ∼5 . 1% of all 8201 predicted proteins ) in the purified cyst samples . Out of the 417 proteins identified , 195 have never been detected in the E . histolytica trophozoite specific proteome or EST databases . This is the first proteomic analysis of naturally-occurring E . histolytica cysts .
The collection of clinical samples used in this study has been approved by the Ethical Review Committee ( ERC ) of ICDDR , B . Since the study participants were minors ( aged 2–6 years ) , parents or legal guardians provided written informed consent on behalf of all child participants . All clinical investigation has been conducted according to the principles expressed in the Declaration of Helsinki . Fecal specimens were collected from children enrolled in an ongoing field study of human immunity to amebiasis in an urban slum community at Mirpur , Dhaka , Bangladesh . They were screened for the presence of E . histolytica cysts in their stools . Initially , fresh stool samples were checked by microscopy for cysts of E . histolytica/E . dispar/E . moshkovskii . Positive stools were then subject to E . histolytica-specific antigen detection by E . histolytica-II ELISA ( TechLab , Blackburg , VA ) as described previously [25] . Since this ELISA was incapable of distinguishing between a mixed infection of E . histolytica/E . dispar and a single infection of E . histolytica , ELISA positive stool samples were further subject to DNA purification and diagnostic PCR for E . histolytica and E . dispar . PCR was also performed for E . moshkovskii-specific DNA . Cysts were only purified from stool samples that were PCR positive solely for E . histolytica . Cysts were purified from stool samples as described previously with some modifications [26]–[28] . In brief , about 4–5 g of stool was dissolved in PBS , filtered through two layers of gauges , centrifuged at 3220×g for 20 minutes , and the supernatant discarded . After washing 3 times with PBS at 3220×g for 20 minutes , the pellet was suspended in 5 mL of ethyl acetate ( to separate the cyst and the fecal debris ) and centrifuged at 3220×g for 20 minutes . The supernatant was discarded and the pellet was washed 3 more times with PBS ( at 3220×g for 20 minutes ) , and the pellet was resuspended in 2 mL of PBS . This was then carefully layered using a Pasteur pipette on top of a previously prepared 10–80% Percoll gradient in 17×120 mm , 15 mL high-clarity polypropylene conical tubes by layering 2 mLs each of 80% Percoll ( in PBS ) , followed by 50% , 40% , 30% , 20% and 10% Percoll solutions . This was centrifuged at 3220×g for 20 minutes . The content between 80% and 40% of the Percoll gradient was transferred to a new tube , washed 3 times with PBS and checked by microscopy for purified cysts . The rabbit sera developed against one member of E . histolytica Jacob ( EHI_044500 ) , which shows 83 . 1% identity and 88 . 1% similarity with the E . dispar Jacob at amino acid level , were purchased from the Cocalico Biologicals , Inc . ( project number 2010-0158 ) . This protocol has been approved by the Animal Care and Use Committee of Cocalico Biologicals , Inc . who follows the USDA and NIH guidelines . The Office of the Laboratory of Animal Welfare , Division of Assurance , National Institute of Health has approved the Animal Welfare Assurance ( #A3669-01 ) of Cocalico Biologicals , Inc . About 50 µl of the E . histolytica cysts purified using the Percoll gradient centrifugation were fixed on a microscopic slide by adding 4% formaldehyde solution for 1 hour at 37°C . After washing with 1× PBS containing 0 . 1% Triton X-100 , cysts were treated with either 1∶200 diluted post-immune or pre-immune antisera against E . histolytica cyst-wall specific Jacob protein raised in rabbits . They were then stained with 1∶200 diluted Fluorescein isothiocyanate ( FITC ) -conjugated goat anti-rabbit secondary antibody and examined by immunofluorescence microscopy . Crude proteins from purified cyst samples were isolated in either of two ways . In the first protocol , cysts were subject to 5 freeze-and-thaw cycles to facilitate the breakage of cyst wall , followed by 20–25 minutes of sonication ( 30-second pulse followed by 30-second rest ) in presence of protease inhibitor ( Sigma-Aldrich ) . This procedure was followed for only one sample , AM951 . In the second protocol , cysts were subject to 20–25 minutes of sonication ( 30-second pulse followed by 30-second rest ) in presence of protease inhibitor ( Sigma-Aldrich ) eliminating the first freeze-and-thaw cycles as above . This procedure was followed for all 5 samples , including AM951 . Protein concentration was measured using the Bio-Rad Protein Assay ( Bio-Rad , USA ) according to the manufacturer's instructions . All the cyst samples ( after extraction of crude proteins ) were divided into two parts ( except for one , sample 4268 , which had too little protein to divide into two parts ) – supernatant ( which should contain soluble proteins ) and pellet ( which should contain insoluble proteins ) . The sample supernatant was removed , dried , and suspended in SDS-PAGE loading buffer . The remaining pellet was soaked in SDS-PAGE loading buffer . For sample 4268 , the whole sample ( containing both the supernatant and the beads ) was dried , and suspended in SDS-PAGE loading buffer . They were loaded on a gel and run until the dye front was ∼1 cm into the gel . This 1 cm×1 cm area of gel was cut for processing ( 2 gel pieces for each sample , except sample 4268 ) . For sample 4268 , the entire sample ( containing both the supernatant and the beads ) was dried , suspended in SDS-PAGE loading buffer , and processed as above as single gel piece . The gel piece for each sample was cubed and transferred to a siliconized tube and washed and destained in 200 µL 50% methanol for 4 h . The gel pieces were dehydrated in acetonitrile , rehydrated in 30 µL of 10 mM dithiolthreitol in 0 . 1 M ammonium bicarbonate and reduced at room temperature for 0 . 5 h . The DTT solution was removed and the sample alkylated in 30 µL 50 mM iodoacetamide in 0 . 1 M ammonium bicarbonate at room temperature for 0 . 5 h . The reagent was removed and the gel pieces dehydrated in 100 µL acetonitrile . The acetonitrile was removed and the gel pieces rehydrated in 100 µL 0 . 1 M ammonium bicarbonate . The pieces were dehydrated in 100 µL acetonitrile , the acetonitrile removed and the pieces completely dried by vacuum centrifugation . The gel pieces were rehydrated in 20 ng/µL trypsin in 50 mM ammonium bicarbonate on ice for 10 min . Any excess enzyme solution was removed and 20 µL 50 mM ammonium bicarbonate added . The sample was digested overnight at 37°C and the peptides formed extracted from the polyacrylamide in two 50 µL aliquots of 50% acetonitrile/5% formic acid . These extracts were combined and evaporated to 15 µL for MS analysis . The LC-MS/MS system consisted of a Thermo Electron Orbitrap Velos ETD mass spectrometer system with a Protana nanospray ion source interfaced to a self-packed 8 cm×75 um id Phenomenex Jupiter 10 um C18 reversed-phase capillary column . Seven microliters of the extract was injected and the peptides eluted from the column by an acetonitrile/0 . 1 M acetic acid gradient at a flow rate of 0 . 5 µL/min over 1 hr . The nanospray ion source was operated at 2 . 5 kV . The digest was analyzed using the double play capability of the instrument acquiring a full scan mass spectrum ( MS – Orbitrap 60K resolution ) to determine peptide molecular weights and 20 product ion spectra ( MS/MS – Ion trap ) to determine amino acid sequence over the gradient elution . The mass spectrometry data were analyzed by database searching using the Sequest search algorithm ( Bioworks 3 . 3 . 1 ) against E . histolytica ( downloaded from NCBI Oct 2010 ) , IPI Human ( Ver 3 . 78 ) , and NCBI NR ( downloaded Jan 2011 ) . The data were loaded in Scaffold v3 and minimal filters were set ( Peptide score >60% , Protein score >90% , Xcorr >1 . 8 ( +1 ) , 2 . 2 ( +2 ) , 2 . 5 ( +3 ) and 3 . 5 ( +4 or greater ) ) . The p-values were determined using the two-tailed Fisher's exact test .
In order to identify E . histolytica cyst positive samples , we performed microscopical examination of stool specimens from children enrolled in an ongoing study on amebiasis . However , since the cyst of E . histolytica is morphologically indistinguishable from that of E . dispar and E . moshkovskii , microscopy-positive samples were then subject to species-specific E . histolytica II ELISA ( TechLab , Blacksburg , VA ) . One limitation of E . histolytica II ELISA is that it cannot differentiate between a co-infection of E . histolytica with E . dispar or E . moshkovskii and a single infection of E . histolytica . Therefore , microscopy/ELISA positive samples were then subject to E . histolytica- , E . dispar- and E . moshkovskii-specific PCRs . Using these methods , we identified 5 stool specimens that were positive only for E . histolytica ( data not shown ) . Microscopy revealed that 4 of these samples had only the cyst form of the parasite , while the 5th sample ( 4268 ) had both cyst and trophozoite forms in the original stool sample , although cysts were the predominant form . Five samples were derived from children aged 2–6 years , three of them were from males ( 4268 , AM951 , and AM797 ) , and two of them were from females ( 8076 and CMS33-7132 ) . Following a positive microscopical identification , Percoll purification was used to significantly enrich the cyst form of the parasite ( data not shown ) . In order to verify that the cyst-like structures in the cyst sample after Percoll purification were indeed E . histolytica cysts , we performed immunofluorescence assay using a polyclonal antibody developed in rabbits against a highly abundant cyst protein Jacob ( EHI_044500 ) . While the rabbit preimmune sera could not recognize cyst structures , the immune sera clearly stained the round-shaped cyst wall of E . histolytica in the Percoll-purified sample as expected ( Figure 1 ) . In order to investigate whether the Percoll gradient cyst purification protocol may have also co-purified the trophozoite form of the parasite , we added 0 , 100 , and 1000 E . histolytica trophozoites into 1 g each of three E . histolytica-negative stool samples ( by microscopy , culture , ELISA and PCR ) , and they were subject to identical Percoll gradient cyst purification as performed with the five test samples . We then performed E . histolytica II ELISA using the contents from the 40–80% fractions of Percoll gradient to detect lectin antigen ( which expresses most abundantly in the trophozoite form ) , and found that all three samples were negative for E . histolytica lectin ( data not shown ) . This suggested that the Percoll gradient purification of stool specimens selectively enriched cysts but not trophozoites . Since only 1 out of 5 cyst samples ( ID: 4268 ) had microscopically positive trophozoite in the initial stool examination , and the number of trophozoites in this sample was much smaller than that of cysts ( data not shown ) , we conclude that our proteomic data was comprised predominantly ( if not entirely ) of proteins expressed in cysts . In order to determine which method of lysate preparation worked best for mass spectrometry analysis , we applied two methods on 1 of the 5 cyst samples ( ID: AM951 ) . The first method was sonication only , while the second method employed freeze-thaw followed by sonication . Mass spectrometry results suggested that a simple sonication-only approach was able to identify more proteins than an alternative approach involving a prior freeze-thaw step . Samples processed by sonication yielded about 20% more identifiable E . histolytica proteins ( total 101 ) than samples processed by the freeze-thaw-sonication approach , which yielded only 81 identifiable E . histolytica proteins . Fifty-two proteins overlapped between the two approaches ( Tables S1 and S2 ) . We investigated if there were any qualitative differences in proteins identified by the two different methods of lysate preparation . Two calcium binding or signal related proteins , and two cyst wall specific proteins were detected in both methods of lysate preparations ( Table S2 ) . In contrary , out of 3 DNA repair proteins 2 were identified in the sonication only method and the third one was identified by freeze-thaw-sonication method only . The only surface associated protein ( EHI_065330 ) was isolated by the sonication only method . Proportionately more proteins with putative enzymatic function were identified by freeze-thaw-sonication ( 19/81 or 23 . 5% ) compared to sonication only ( 18/101 or 17 . 8% ) . In contrast , slightly more actin-related proteins were identified in sonication only method ( 7/101 or 6 . 9% ) compared to freeze-thaw-sonication method ( 3/81 or 3 . 7% ) . However , these differences were not statistically significant . So , we conclude that there was no particular trend that could be detected in differentiating protein categories identified by sonication only or freeze-thaw-sonication methods . Mass spectrometry was carried out for 4 out of 5 cyst samples in such a way as to detect both soluble and insoluble proteins ( for details , see the materials and methods section ) . For the 5th sample ( 4268 ) , there was not enough cyst material to separate soluble and insoluble proteins and the two fractions were analyzed together . In most cases more proteins were identified in insoluble fraction than soluble fraction ( 59/48 for AM951 , 99/42 for AM797 , and 57/35 for CMS33-7132; Figure S1 ) . A total of 417 unique E . histolytica proteins were identified in the 5 cyst preparations . Human , bacterial , and rice-related proteins comprised about 90% of all proteins identified in the purified cyst samples ( Figure 2 ) . Human and rice-related proteins were abundant because of the source of cyst material ( human feces ) and nutritional habit of people in that region of the world ( Dhaka , Bangladesh ) . Bacteria attached to the surfaces of cysts or inside the cysts might be the predominant source of bacterial proteins in the proteomic data . The proportions of human or rice related proteins may be reduced by using more stringent washing steps ( from large amount of initial stool volume ) . However , cyst protein yields would be reduced and it would be very difficult , if not impossible , to get rid of bacteria that are attached to or ingested by the cysts . The 5 cyst samples came from 5 children aged 2–6 years , as a result , the amount of original stools that could be collected was very low , which was a limitation of this work . Nevertheless , E . histolytica proteins were identified in each of the 5 cyst samples as expected . The percentage of total proteins identified that were E . histolytica proteins in various cyst samples ranged from 6 . 9 to 10 . 2% ( Figure 2 ) . Since the total number of proteins identified in the naturally-occurring cyst samples was small , we asked how unique these proteins are in comparison with the publicly available protein datasets in AmoebaDB ( http://amoebadb . org/amoeba/ ) . Two LC-MS/MS proteomic datasets were available for the trophozoite stage of E . histolytica comprising a total of 1825 proteins ( as of November 30 , 2011 ) [29] , [30] . There was also available an EST database for E . histolytica trophozoite comprising a total of 1223 proteins ( as of October 1 , 2011; from a total of 20 , 812 public entries at http://www . ncbi . nlm . nih . gov/dbEST/dbEST_summary . html ) . A three-way comparison of protein overlaps between the three datasets ( i . e . cyst proteome , trophozoite proteome , and trophozoite EST ) was done using the freely available software in the internet ( http://www . cmbi . ru . nl/cdd/biovenn/index . php ) [31] . The Venn diagram shows that a total of 191 proteins overlapped between the cyst proteomic and trophozoite proteomic ( LC-MS/MS ) datasets , 109 proteins overlapped between the cyst proteomic and trophozoite EST datasets , and 78 proteins overlapped between all three datasets ( Figure 3 ) . It was surprising to see that only 515 proteins overlapped between the trophozoite proteome ( 515/1825 or 28 . 2% ) and the EST datasets ( 515/1223 or 42 . 1% ) . This disparity is probably because only limited numbers of trophozoite-derived proteome data are available for E . histolytica . The two proteome works available at the AmoebaDB public database when this manuscript was being written ( November 2011 ) were highly specific in nature - one attempted to identify trophozoite proteins from phagosome [29] , while the other attempted to identify Concavalin A-enriched glycoproteins from the E . histolytica trophozoite [30] . So , we think many of the trophozoite-specific proteins are not represented in them , resulted in showing poor overlap with the trophozoite-derive EST datasets from E . histolytica . Nevertheless , a unique set of 195 proteins in the cyst proteomic dataset did not overlap with the other two datasets . This set might include E . histolytica cyst-specific proteins . The cyst proteomic data was also compared with the only available mRNA transcriptome data for encysting cultures grown in vitro [22] . The encysting transcriptome was performed on recently isolated E . histolytica strains that can spontaneously encyst in vitro when grown in complex diphasic Robinson's medium . Six hundred and seventy two genes were identified to be up-regulated in E . histolytica cysts ( based on ≥3-fold expression change and p-value of <0 . 05 ) . Out of 417 cyst proteins detected in this study , 23 had no mRNA expression data available . Forty-seven of the remaining 394 proteins identified here overlapped with the 672 cyst-specific genes identified by transcriptomics ( p-value 0 . 0058 ) ( Figure S2 ) . We then asked whether the unique set of 195 proteins that did not overlap with the trophozoite proteomic and the EST datasets exhibited a stronger overlap with the 672 cyst-specific genes . Out of 195 unique cyst proteins , 10 had no mRNA expression data available . However , 28 of the remaining 185 proteins identified by cyst proteomic analysis overlapped with the 672 cyst-specific genes identified by transcriptomic analysis . As expected , this overlap between cyst-specific transcripts and unique cyst proteins was stronger ( p-value 0 . 0014 ) than the overlap between cyst-specific transcripts and cyst proteins in general ( p-value 0 . 0058 ) ( Figure S3 ) . Despite the fact that there was a statistically significant overlap between the cyst-specific transcripts [22] and cyst proteins of this study , we did notice a large discrepancy between the two datasets . Several factors could explain this discrepancy: ( i ) our cyst proteomic data was based on naturally occurring cyst samples , while the Ehrenkaufer and colleague's cyst transcriptome data was based on cyst-like cultures obtained in vitro; ( ii ) the number of proteins identified in the cyst proteome was very low , and they represent only ∼5 . 1% ( 417/8201 ) of all proteins; ( iii ) cyst-specific genes were categorized based on >3-fold higher mRNA levels compared to that of trophozoites , as a result , there will be some proteins that are expressed in lower ( >3-fold or more ) levels in cysts compared to trophozoites , but will still be regarded as cyst-specific by Ehrenkaufer et al; and ( iv ) some mRNA transcript levels may not correspond to protein levels due to post-transcriptional regulation or protein degradation . The cyst proteomic data was also compared with available data from developmental studies on the reptilian species E . invadens , whose encystation can be induced easily in vitro . Two E . histolytica chitinases , EHI_109890 and EHI_152170 , show 79% and 47% identities with the E . invadens chitinases , EiChit1 and EiChit4 , respectively [32] . In E . invadens , the mRNA expression of 4 chitinases was studied by Makioka and colleagues [32] during in vitro encystation and excystation . In the early phase of encystation , mRNA expression of all 4 chitinases increased although the greatest increase was seen for EiChit1 and EiChit4 . However , following 5 hours of excystation , mRNA levels of these two chitinases dropped sharply compared to pre-induction stage , suggesting that these are highly cyst-specific . Both of these chitinase homologues ( EHI_109890 and EHI_152170 ) were detected in our cyst samples as expected . Three actin depolymerizing factor ( ADF ) family proteins have been identified in E . invadens [33] . These proteins are thought to be important in actin cytoskeleton reorganization during development of E . invadens , and can be detected in both trophozoite and cyst stages of this parasite . The single member of ADF family protein in E . histolytica ( actophorin/EHI_197480 ) , which shows virtually no mRNA differences between trophozoite and cyst stages ( Table S3; [22] ) was identified in our cyst proteome , consistent with a similar role in the development of E . histolytica . An E . histolytica glycolytic enzyme enolase ( EHI_130700 ) shows 85% identity with the E . invadens enolase ( EIN_093390 ) . This protein was present in the cytoplasmic vesicles as well as in the cyst wall of E . invadens as revealed by immunofluorescence microscopy [34] . However , this protein was present only in the cytoplasmic vesicles of E . histolytica trophozoites recovered from amebic liver lesions of experimental animals [34] . Additionally , enolase was detected in the mature cyst wall of E . histolytica in samples derived from human infections . Consistent with this finding , we could also detect enolase in 3 out of 5 cyst samples in our proteomic study . An earlier study has shown that heat shock treatment of E . invadens can result in strong induction of cyst-specific chitinase and Jacob mRNAs , and moderate induction of heat shock protein mRNAs , including a 70-kDa heat shock protein known as BiP ( AF252299 ) [20] . Although heat shock alone cannot produce matured , chitin-walled cysts , these authors suggest that amebic heat shock proteins are involved in degradation of cytoskeletal proteins during encystation . In our cyst proteome study , 5 putative heat shock proteins were detected ( Table S3 ) , and one of these ( EHI_199590 ) is a homoologue of E . invadens BiP ( with 88% identity ) , suggesting a similar role of heat shock proteins in E . histolytica encystation . However , other proteins that appeared to be important in E . invadens development were not detected in our study . For example , the E . histolytica homologue of one of the cysteine proteases EiCP-B9 [35] , or profilins ( such as EiPFN1 and EiPFN4; [36] ) that are expressed in the cyst stage of E . invadens , were not detected in our proteome study . These discrepancies could be due to the fact that we were only able to identify a fraction of all cyst-specific proteins . Alternatively , they might reflect differences between parasite species and/or between cysts generated in hosts and in vitro . The 417 E . histolytica proteins ( Table S3 ) fell into several broad functional groups . About 28% ( 117 proteins ) did not have a known function and are referred to as ‘hypothetical proteins’ ( Figure 4 ) . About 12 . 9% ( 54 proteins ) contained putative transmembrane ( TM ) domains ( Table S3 ) . Perhaps the most important were the categories of 195 unique proteins that showed no overlaps with the trophozoite specific proteome or EST datasets ( Figure 5 ) . About 39 . 5% of these proteins ( 77 proteins ) did not have a known function . The E . invadens cyst wall is composed of chitin ( a homopolymer of ß-1 , 4-linked GlcNAc ) , and chitin-binding Jacob and Jessie lectins and chitinases [37]–[39] . The prior transcriptome data from encysting cultures in vitro supported that these proteins were components of the E . histolytica cyst [22] . In this study , among the 195 unique cyst proteins , we identified one member of the Jacob gene family ( EHI_028930 ) , two chitinases ( EHI_109890 and EHI_152170 ) , and one chitin synthetase 2 ( EHI_044840 ) supporting prior suppositions as to the composition of the cyst wall [22] , [39] . Besides cyst wall proteins , 4 putative surface associated proteins were identified in the 195 cyst proteins . Three of these belonged to the BspA-like leucine rich repeat ( LRR ) protein family . An important aspect of BspA family proteins is to mediate protein-protein interactions [40] . In E . histolytica , more than 70 proteins belong to BspA-like LRR protein family , although only 4 of these were previously identified as being cyst-specific based on mRNA transcriptome data [22] . One BspA family member ( EAL42510/849 . m00008 ) was previously reported to be primarily located on the plasma membrane of E . histolytica trophozoite [41] . Consistent with the trophozoite-specificity of this protein , we could not identify this protein in the present cyst proteome . Another BspA family protein has been implicated in trifluoromethionine resistance [42] . Immunofluorescence studies with the episomal overexpression of this protein suggest that it is expressed in the cytoplasm of the E . histolytica trophozoites . Again , this protein was not identified in our cyst proteome , consistent with trophozoite-specificity . The other potential surface protein identified was a putative member of lectin family protein ( EHI_110830 ) . Actin is one of the most conserved and ubiquitous proteins in eukaryotes . During encystation , morphological changes occur in the trophozoite structure of E . histolytica , and elongated trophozoites become spherical cysts . We anticipate a reorganization of the actin cytoskeleton during this process . One study shows that jasplakinolide , an actin-polymerizing and filament-stabilizing drug , inhibits both growth and encystation of E . histolytica and E . invadens through perturbation of the actin cytoskeleton [43] . Four actin or actin-related proteins were identified in the cyst proteome that might be involved in these morphological and structural changes , and deserve further studies . Transmembrane domain containing protein kinases ( TMKs ) are involved in signal transduction in higher eukaryotes . Over 100 TMKs have been identified in the E . histolytica genome [44] . Since these proteins have extracellular domains coupled to cytoplasmic kinase domains , they have the potential to sense environmental cues . Fourteen of these TMKs showed significantly differential mRNA expression in encystating E . histolytica cultures compared to that of trophozoites [22] . Additionally , a number of cytoplasmic protein kinases have also been identified to be cyst-specific [22] . The E . histolytica genome encodes for about 343 protein kinases ( including >100 TMKs ) , and 19 of these were identified in the 195 unique cyst proteins . Protein kinases regulate cellular pathways , especially those involved in signal transduction . More than half of these protein kinases ( 11/19 ) were predicted to have transmembrane domains , suggesting their potential roles in sensing outside clues for stage conversion ( Table S3 ) [44] . GTPases are a large family of GTP-binding hydrolase enzymes involved in various biological functions including signal transduction , protein biosynthesis , cell division , translocation of proteins through membranes and transport of vesicles within the cell . The GTP-binding proteins are classified into five families – Ras , Rho , Rab , Arf and Ran [45] . Recent data suggest that the E . histolytica genome codes for 17 , 22 , 73 , 12 and 1 member ( s ) of Ras , Rho , Rab , Arf and Ran proteins , respectively [45] . Additionally there are 41 unclassified small G proteins in the E . histolytica genome . The Ras proteins are mainly involved in cell proliferation . The Rho proteins are implicated in cytoskeleton regulation . Both Rab and Arf proteins are involved in membrane trafficking in the cytosol . The Ran proteins are implicated in nuclear-cytosol transport . It is anticipated that a dramatic change of cellular components occurs in a rapid , yet highly regulated fashion during encystation , involving de novo synthesis of new proteins and degradation of unwanted proteins . In order to achieve this , the amoeba is expected to have a developmental–specific system of membrane trafficking . The transcriptomic changes of Rab genes have been studied during the encystation of E . invadens , and 23 cyst-specific , 36-trophozoite-specific and 31 constitutively expressed Rab genes were identified [46] . In the 195 unique cyst proteins , we identified 10 GTPases ( including members of Rab or related proteins ) that may have stage-specific functions in E . histolytica ( Figure 5; Table S3 ) . Four DNA repair proteins , including two DNA double-strand break repair proteins ( EHI_053200 and EHI_125910 ) , one DNA mismatch repair protein ( EHI_126120 ) and a type A flavoproteins ( EHI_129890 ) were identified in the 195 cyst proteins . We also found 8 vesicle coating ( or trafficking ) related proteins including 3 putative clathrin proteins which are generally involved in shaping rounded vesicles in the cytoplasm for intracellular trafficking [47] . In addition , a SET domain containing histone lysine methyltransferase ( EHI_0319060 ) that adds methylmarks on histone tails , and two chromodomain-containing proteins ( EHI_000780 and EHI_031370 ) that bind methylmarks on histone tails and facilitate the recruitment of transcriptional effector molecules were also identified , suggesting a role for epigenetic machineries in the development of E . histolytica . Among various species of Entamoeba that can infect humans , only E . histolytica can cause intestinal and extraintestinal diseases in humans . At least two other human-infecting Entamoeba species , E . dispar and E . moshkovskii , are morphologically identical to E . histolytica . These commensal species create diagnostic challenges . Although E . dispar has never been documented to cause diseases , some emerging data suggest E . moshkovskii may cause diseases in humans [48] , [49] . It is recommended that individuals with E . histolytica infection regardless of clinical status be treated due to the risk of the development of invasive amebiasis even in asymptomatic individuals . However , no simple E . histolytica cyst-specific diagnostic test is currently available to detect carriers . Here , we developed an immunofluorescence assay ( IFA ) using rabbit anti-Jacob antiserum that could detect E . histolytica cysts ( Figure 1 ) ; however this polyclonal antibody might not be specific to E . histolytica . We conclude that proteins identified in this study are likely to be at least partially specific to E . histolytica , and therefore warrant further investigation as to their diagnostic potential . In order to identify candidate diagnostic targets , we focused initially on the list of 195 unique proteins that were not detected in any trophozoite specific proteome or EST datasets . Twenty-five of these proteins showed 80% or less identity on the amino acid level compared to homologues in E . dispar ( Table 1 ) . These proteins displayed even lower identities with other proteins in the database including E . moshkovskii and human proteins . Over half ( 13/25 ) of the 25 proteins have no known functions and are annotated as hypothetical proteins . Five of these had previously shown 141- to 1012-fold higher mRNA transcript levels in cyst-like cultures compared to cultured trophozoite of E . histolytica [22] . Three out of 25 proteins in the list have no homologues in E . dispar including two that are putative reverse transcriptases ( 416 . m00035 and 453 . m00043 ) and one hypothetical protein ( 112 . m00115 ) . For diagnostic purposes , proteins that were identified in a majority of the cyst samples are hypothesized to be consistently expressed in excreted cysts at detectable levels , and therefore potentially interesting as candidate diagnostic targets . Eight proteins were identified ( from the list of 195 unique proteins ) in 3 or more ( out of 5 ) cyst samples ( Table 2 ) . Four of the 8 proteins displayed 90% or less amino acid level homologies with proteins from E . dispar . Two of these were putative cyst-wall specific proteins: chitinase ( EHI_109890 , which showed 87-fold higher mRNA transcript in encysting culture ) and chitinase Jessie 3 ( EHI_152170 , which showed 44-fold higher mRNA transcript in encysting culture ) . Proteins detected in our study have variable number of spectrum counts as detected by the LC-MS/MS experiments . For diagnostic purposes , proteins with greater numbers of spectrum counts may be better candidates for diagnostic assay development , as they may be more stable and more abundant in samples . Seven proteins were identified ( from the list of 195 unique proteins ) with more than 10 spectrum counts ( Table 3 ) . Four of the 7 proteins displayed 90% or less amino acid level identities with proteins from E . dispar . Two of these are chitinase ( EHI_109890 ) and chitinase Jessie 3 ( EHI_152170 ) with 52 and 24 spectrums detected , respectively , while the remaining two are hypothetical proteins ( EHI_146120 , 36 spectrums; and EHI_019630 , 15 spectrums ) . Overall , Tables 1 , 2 , and 3 describe a total of 32 unique ( non-redundant ) proteins with diagnostic potential . In addition to candidate proteins for cyst specific diagnosis , we also identified proteins in the remaining 222 proteins out of the 417 ( that were also present in the trophozoite specific proteome and EST datasets ) . These may also be useful to develop diagnostic tools to detect both cysts and trophozoites of E . histolytica . These proteins are listed in Tables S4 , S5 , S6 . Thirty-one human proteins were identified in at least 3 out of 5 cyst samples including 7 that were identified across all 5 samples , and 10 that were identified in 4 samples ( Table S7 ) . Multiple lectin- or sugar-binding proteins such as galectin-3 ( IPI00023673 ) , galectin-4 ( IPI00009750 ) , intelectin-1 ( IPI00291737 ) , intelectin-2 ( IPI00103436 ) , glycoprotein 2 ( IPI01014468 ) , and proteoglycan 3 ( IPI00005778 ) were detected . The possible interaction of some of these proteins with the cyst or trophozoite surface and their potential roles in stage-conversion may warrant further investigation . Additional proteins detected in the cyst samples derived from bacterial origins ( Table S8 ) . The E . histolytica gene annotation is limited at present , although it is improving through the combined efforts of biostatisticians and researchers [50] . More than half of the genes ( 53 . 8%; 4413 genes out of a total predicted 8201 genes ) are designated as “hypothetical” proteins of unknown function [50] . This is mainly because the E . histolytica genes are highly diverse in sequence compared to other organisms , which makes prediction of gene function difficult . One hundred and seventeen ( or 28% ) of 417 proteins identified in this study were predicted “hypothetical” proteins including 40 that were previously identified in trophozoite derived proteome or EST datasets . For the remaining 77 cyst specific “hypothetical” genes we now have evidence of protein level expression . We also found that 9 out of 417 protein genes identified in this study were missing in the most recent genome annotation ( http://amoebadb . org/amoeba/ ) . Three of these are putative clathrin heavy chain containing proteins ( gi|103484580 , found in sample AM951; 200 . m00090 , found in samples 8076 , AM797 , CMS33-7132; and 141 . m00078 , found in sample AM797 ) that are involved in vesicle trafficking in other organisms . Additional non-annotated genes encode 2 putative reverse transcriptases ( 416 . m00035 and 453 . m00043 , both found in sample 4268 ) , a putative Sec61 alpha subunit ( gi|52352493 , found in sample AM951 ) , an actin ( 10 . m00319 , found in samples AM797 and CMS33-7132 ) , and two hypothetical proteins ( 112 . m00115 , found in samples 4268 and AM797; and 270 . m00054 , found in sample CMS33-7132 ) . We now have evidence that these genes are expressed as proteins in E . histolytica cysts .
E . histolytica: EHI_044500 , EHI_065330 , EHI_109890 , EHI_152170 , EHI_197480 , EHI_130700 , EHI_028930 , EHI_044840 , EAL42510 , EHI_110830 , EHI_199590 , EHI_053200 , EHI_125910 , EHI_126120 , EHI_129890 , 416 . m00035 , 453 . m00043 , 112 . m00115 , EHI_146120 , EHI_019630 , gi|103484580 , 200 . m00090 , 141 . m00078 , gi|52352493 , 10 . m00319 , 270 . m00054 , EHI_0319060 , EHI_000780 and EHI_031370 . E . invadens: EiChit1 , EiChit4 , AF252299/BiP , EIN_093390 , EiCP-B9 , EiPFN1 , and EiPFN4 . Human: galectin-3/IPI00023673 , galectine-4/IPI00009750 , intelectin-1/IPI00291737 , intelectin-2/IPI00103436 , glycoprotein 2/IPI01014468 , and proteoglycan 3/IPI00005778 . | We used tandem mass spectrometry to identify E . histolytica cyst proteins in 5 cyst positive stool samples . We report the identification of 417 non-redundant E . histolytica proteins including 195 proteins that were not identified in existing trophozoite derived proteome or EST datasets , consistent with cyst specificity . Because the cysts were derived directly from patient samples with incomplete purification , a limited number of proteins were identified ( N = 417 ) that probably represent only a partial proteome . Nevertheless , the study succeeded in identifying proteins that are likely to be abundant in the cyst stage of the parasite . Several of these proteins may play roles in E . histolytica stage conversion or cyst function . Proteins identified in this study may be useful markers for diagnostic detection of E . histolytica cysts . Overall , the data generated in this study promises to aid the understanding of the cyst stage of the parasite which is vital for disease transmission and pathogenesis in E . histolytica . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"biochemistry",
"infectious",
"diseases",
"diagnostic",
"medicine",
"global",
"health",
"biology",
"microbiology",
"proteomics"
] | 2012 | Proteomic Analysis of the Cyst Stage of Entamoeba histolytica |
Interactions between genes and proteins are crucial for efficient processing of internal or external signals , but this connectivity also amplifies stochastic fluctuations by propagating noise between components . Linear ( unbranched ) cascades were shown to exhibit an interplay between the sensitivity to changes in input signals and the ability to buffer noise . We searched for biological circuits that can maintain signaling sensitivity while minimizing noise propagation , focusing on cases where the noise is characterized by rapid fluctuations . Negative feedback can buffer this type of noise , but this buffering comes at the expense of an even greater reduction in signaling sensitivity . By systematically analyzing three-component circuits , we identify positive feedback as a central motif allowing for the buffering of propagated noise while maintaining sensitivity to long-term changes in input signals . We show analytically that noise reduction in the presence of positive feedback results from improved averaging of rapid fluctuations over time , and discuss in detail a particular implementation in the control of nutrient homeostasis in yeast . As the design of biological networks optimizes for multiple constraints , positive feedback can be used to improve sensitivity without a compromise in the ability to buffer propagated noise .
Cells sense and process information using biochemical networks of interacting genes and proteins . Typically , a signal is sensed at a specific point of the network ( input ) and is propagated to modulate the activity or abundance of other network components ( output ) . Reliable information processing requires high sensitivity to changes in the input signal but low sensitivity to random fluctuations in the transmitted signal . Since the detection of signal is inherently stochastic [1] , and the microenvirnment of the cell is also fluctuating randomly , understanding the principles of noise propagation in biochemical and genetic networks is of interest [2–5] . Linear ( unbranched ) cascades present the simplest instance of biochemical networks . Recent studies have shown that such cascades display an interplay between sensitivity to changes in input signal and the ability to buffer stochastic fluctuations [6–9] . Indeed , an increase in the sensitivity toward input signals results also in elevated sensitivity to noise in the input . A key question is whether network connectivity , e . g . , the presence of positive or negative feedbacks , can modulate this interplay , reducing propagated noise while maintaining high sensitivity . Previous studies argued that negative feedbacks can buffer noise relative to linear cascades [10–12] . These studies , however , did not consider the associated changes in signaling sensitivity . In general , the fine line that separates “noise” from “signal” is established functionally . Nevertheless , in many systems such as the sensing of temperature , nutrient levels , ligand concentration , etc . , the signal is interpreted as a long-term change in the input , whereas noise is characterized by rapid stochastic fluctuations . In this study , we focus on this particular class of systems . We explore for gene circuits that can buffer propagated noise while maintaining signaling sensitivity . We consider a large set of networks that are differentially designed but are equally sensitive to long-term changes in the input , and compare their ability to buffer propagated noise . Systematic analysis of all three-gene circuits revealed that negative feedback amplifies propagated noise . In contrast , positive feedback appears to be a necessary element for buffering such noise . Analytical analysis demonstrated that positive feedback contributes to noise buffering by slowing down the dynamics , thus providing a longer averaging time . A detailed analysis of a recurrent network design , found in systems controlling nutrient homeostasis , suggests that it functions as a noise-reduction device based on the principles identified in our analysis .
To begin analyzing the effect of network architecture on the interplay between sensitivity and noise buffering , we considered a two-component cascade with a negative feedback loop . This cascade is composed of an input node , n0 , which activates an output node , n1 . The output node feeds back to repress its own expression ( Figure 1A ) . Formally , this system is described by where β1 denotes the maximal rate of n1 production , denotes the rate of n1 degradation , and h0 , h1 are Hill coefficients . Note that n0 and n1 are normalized by their respective dissociation constants from the gene promoter . We consider an input signal n0 ( t ) = 〈n0〉 + σ0 ( t ) which fluctuates around some mean level 〈n0〉 . The fluctuating component σ0 ( t ) has a zero mean and some autocorrelation time τ0 . Figure 1B depicts the temporal fluctuations of n1 for a system with a strong negative feedback ( Hill coefficient h1 = 4 ) . The analogous dynamics for a system that lacks such feedback ( h1 = 0 ) is also shown . Consistent with previous studies [10–12] , output noise is lower in the presence of negative feedback . Nevertheless , negative feedback also significantly lowers the sensitivity of the system to a two-fold change in the level of the mean input ( Figure 1C ) . To rigorously quantify the interplay between the sensitivity of the input–output relation and the buffering of propagated noise , we define two measures for the sensitivity and noise-amplification of the system . The steady-state sensitivity is captured by the susceptibility , s , [3 , 13 , 14] ( also termed gain [9] ) defined as the relative change in output following a change in the input: with all quantities measured at steady state . The measure for noise amplification is defined as the ratio between the output and input noise: As before , all quantities are measured at steady state . Both s and depend on the different parameters of the system , including the Hill coefficients and mean input levels . Figure 1D depicts the noise amplification versus susceptibility for different levels of mean input . The case of no feedback ( h1 = 0 ) is compared to that of increasing feedback cooperativity ( h1 = 1 and 2 ) . Again , a clear interplay between susceptibility and noise buffering is observed , with systems that are more sensitive to changes in the input level being also more vulnerable to noise . Notably , this interplay seems to be more severe in the presence of negative feedback . Thus , for a given level of susceptibility , propagated noise is amplified to a greater extent in the presence of a negative feedback . This result is consistent with the theoretical arguments for a two-node model [3]: with negligible intrinsic noise , and when controlling for sensitivity , negative feedback enhances , rather then represses , propagated noise . Our analysis implies that negative feedback cannot be used to buffer against rapidly varying propagated noise in systems which require a sensitive response to long-term changes in their input . To identify network architectures that can buffer noise while maintaining sensitivity , we characterized systematically the relation between susceptibility and noise-buffering of all three-node circuits ( Figure 2A ) . A three-node circuit is composed of an input node ( n0 ) , an output node ( n2 ) , and an intermediate node ( n1 ) , connected via activating or repressing interactions ( arrows ) . We allowed for all incoming or outgoing arrows , with the exception of the input n0 which could affect both n1 and n2 ( outgoing arrows ) , but was not subject to feedback regulation ( no incoming arrows ) . Each arrow was assigned a positive sign ( activation ) or a negative sign ( repression ) , thus leading to a total of 324 networks ( Text S1 , Section IV ) . To ensure controlled comparison between networks , we assume that degradation is not regulated and that all proteins degrade at the same rate ( e . g . , by dilution ) . Each specific circuit supports a range of dynamic behaviors , depending on the precise value of the interaction parameters . Following the formalism presented by Paulsson [3 , 13] , we used the Fluctuation Dissipation Theorem ( FDT [15] ) to derive an analytical formula for susceptibility and noise-amplification ( Methods ) . Briefly , the system of equations that describe the dynamics of the three-node network was The degradation terms were assumed to be first-order , and we considered τi = 1 to maintain a mathematically controlled comparison . This system of equations was linearized around the steady state . The linearization process excludes from the analysis possible noise-filtering mechanisms that show very sharp functions such as AND or OR gates [16] as well as oscillations or transitions between multiple steady states ( such as the study on positive feedback loops and noise in [17] ) . Note also that the steady state n0 = n1 = n2 = 0 is not formally part of our analysis , because it renders the relative fluctuation η infinite . Following linearization , the combined effect of all interaction parameters ( i . e . , Hill coefficients and saturation levels ) is captured by the elasticity [3 , 13]: where is the rate of generating ni ( e . g . , via transcription ) and is its degradation rate . With these definitions , the susceptibility of the output was ( Text S1 , Section I ) The absolute value facilitates a comparison between systems that increase or decrease their output when the input goes up . Noise amplification was found by solving the matrix equation [3 , 13] where the matrix η is composed of the normalized noise terms , the matrix M is related to the elasticities and time scales τi , and the matrix D contains a single term corresponding to noise input from n0 . In the construction of D , we assume that the sole noise source is fluctuations in n0 , and that these fluctuations die out exponentially with a time scale τ0 = 1 ( autocorrelation time of one unit ) . The exact terms of Equation 7 are defined in Section II of Text S1 and in [3 , 13] . A particular choice of elasticity values for all arrows of the network defines a single point in the plane , and the general interplay between sensitivity and noise buffering was derived by considering a large number of different elasticity values ( Figure 2A , Methods ) . As expected , noise amplification in linear ( unbranched ) cascades is precisely proportional to the susceptibility ( Figure 2B ) . This case of no feedback provides the reference for comparison for other network architectures . Consistent with the analysis above , in the case of a negative feedback , all the points appear above the reference line ( Figure 2C ) implying an increase in noise for a given level of susceptibility . Noise amplification at constant susceptibilities is observed also for the coherent ( Figure 2D ) and incoherent ( Figure 2E ) feed-forward loops ( FFLs ) [16 , 18] , probably reflecting the addition of nonsynchronous noise components mediated through the intermediate node . In contrast , for positive feedback the points in the plot appear below the reference line ( Figure 2F ) . Thus , for a given level of sensitivity , positive feedback buffers propagated noise . To further characterize the properties of all three-node circuits , we calculated for each network the fraction of parameter sets that produce a stable steady state ( Text S1 , Section III ) , following the paradigm of [19] . We then calculated the fraction of stable parameter sets ( Methods , Section IV of Text S1 ) that display high susceptibility and low noise ( Figure 2G ) . Notably , networks that were stable throughout the entire parameter range provided poor noise buffering for a given susceptibility . None of these circuits contained positive feedback loops . In sharp contrast , the circuits that enhanced noise buffering were amenable to instability and were all composed of positive feedback loops . Taken together , among the networks tested , positive feedback appears to be required for buffering propagated noise while maintaining sensitivity . To better understand the reason underlying the ability of positive feedback to buffer propagated noise for a given susceptibility , we used the analytical description of a two-component system with an input n0 and an output n1 , as was derived by Paulsson in [3 , 13] using the FDT approach [15] . In this framework ( and while neglecting intrinsic noise ) , noise amplification was shown to be given by [3 , 13]: where s = −H10/H11 denotes the susceptibility; τ0 , τ1 denote the degradation time scales of n0 and n1 , respectively; and H10 , H11 are the elasticities ( as defined in Equation 5 above ) . In the absence of feedback , H11 = 1 . Negative feedback of n1 on itself implies H11 > 1 , whereas positive feedback implies 0 < H11 < 1 ( if H11 falls below zero , instability arises ) . As was shown by Paulsson [3] , negative feedback impairs noise buffering at a given susceptibility by effectively accelerating the dynamics and reducing averaging of fluctuations over time . Similarly , positive feedback enhances noise buffering for constant susceptibility by slowing down the dynamics and allowing for better time-averaging of fluctuating components . Notably , Equation 8 suggests that positive feedback is not the only way to reduce H11 [3 , 13] . Moving beyond the first-order degradation assumed in our study , H11 can also be decreased if the degradation is independent of n1 ( in Equation 5 ) . This , in fact , is likely to be the case for nondividing microorganisms . Hence , time averaging would also be improved if the degradation is close to zero-order and the synthesis is not influenced by n1 . While positive feedback appears to be important for buffering propagated noise ( when sensitivity is controlled for ) , such a mechanism needs to comply with several requirements . First , the feedback loop itself should produce low internal noise because intrinsic noise is not buffered . Second , the effective elasticity H11 ( Equation 8 ) should be of intermediate magnitude: when it is too high ( H11 → 1 ) the effect of the positive feedback is negligible , but when it is too low ( H11 → 0 ) the system is on the verge of instability , and the steady state will no longer resist small fluctuations . Finally , to avoid decrease in susceptibility due to saturation effects , H11 must be maintained constant over a large range of parameters . A class of mechanisms that complies with the above requirements is based on a combination of positive and negative feedbacks . Fast-acting negative feedback functions to ensure stability , while positive feedback provides the required noise buffering . A specific example for such a network is involved in nitrogen homeostasis in yeast [20–22] ( Figure 3A ) . Here , a transcription factor ( Gat1p ) , which is activated by nuclear Gln3p , feeds back to enhance its own transcription , and in addition induces a transcriptional repressor ( DAL80 ) that competes with Gat1p for the same DNA binding sites . This competition effectively weakens the positive feedback and ensures stability . Denoting the input signal to the system by n0 , the output Gat1p by n1 and the repressor Dal80p by n2 , the system can be modeled by the following two differential equations: Here αi and βi denote the degradation and transcription rate constants , respectively , and l is a low rate of basal transcription required to prevent the shutdown of the system , n1 = n2 = 0 . We will neglect this factor in subsequent analysis . The Kij coefficients in the protein production terms are dissociation constants , with n2/Ki2 describing the competitive inhibition of Dal80p . The Hill coefficient of n2 binding to its own promoter is 2 because Dal80p binds as a dimer [21 , 22] . The Hill coefficient of n2 binding to the n1 promoter is set to 1 to enhance noise buffering and susceptibility ( although a value of 2 would still increase noise averaging ) . For the system described by Equations 9 and 10 to operate as a sensitive noise buffer , it must work in a regime where all interactions are unsaturated . Hence , all the binding constants of the repressor , Ki2 , must be small , and all binding constants of the activator , Ki1 , must be large . In this regime , Equation 9 and 10 reduce to and Finally , if n2 responds more rapidly than n1 and n0 ( H22α2 ≫ H11α1 , 1/τ0 ) , then it can be assumed to be at quasi-steady state , and Equations 11 and 12 are combined to The power law dependence of the transcription rate on n1 results in an almost-constant elasticity = 1/3 ( + in Equation 5 ) . Hence , this network can buffer noise and maintain susceptibility for a large range of concentrations at which it remains unsaturated . A more rigorous analysis of the system is presented in Section VI of Text S1 . Detailed simulations confirm that this system can indeed buffer propagated noise , as compared to a loop-free system with the same levels of susceptibility ( Figure 3B and 3C ) . Furthermore , the noise buffering capacity and the susceptibility of this system are maintained over a large range of input levels ( Text S1 , Section VI ) .
The ability to distinguish input signals from stochastic fluctuations is crucial for reliable information processing . Yet , being processed by the same computation device , signal and noise are inherently coupled . It thus comes as no surprise that increasing the ability to buffer propagated noise comes typically at the expense of reducing the sensitivity toward the input signal . We study this interplay in the context of a special class of systems where the signal is retained for long time periods , whereas the noise fluctuates rapidly . Such systems are ubiquitous in the adjustment of cells to aspects of their extracellular environment . Previous studies reported that negative feedback buffers gene expression noise [10–12] . Nonetheless , when considering propagated noise that originates upstream of the feedback loop , this noise filtering merely reflects the reduction in the ability of the system to respond to changes in its input . Moreover , when parameters are chosen to preserve system sensitivity , negative feedback in fact amplifies , rather than reduces , propagated noise . By the same token , positive feedback , which appears to both increase the sensitivity of the system to changes in its input and to amplify intrinsic noise , reduces propagated noise when susceptibility ( steady-state sensitivity ) is controlled for . Analytical analysis [3 , 13] revealed that noise propagation depends on two factors: the sensitivity to changes in input ( susceptibility ) on the one hand , and the averaging time [3 , 13] on the other hand . In the absence of feedback , this averaging time depends only on degradation rate . However , both negative and positive feedbacks impact this averaging time: negative feedback accelerates the dynamics [23] and consequently it reduces time averaging and does not buffer against noise . In contrast , positive feedback delays the kinetics leading to attenuation of propagated noise . If we view the feedback modules as low-pass frequency filters [12 , 24 , 25] and define a critical frequency [25] above which fluctuations are eliminated , then negative feedback increases this critical frequency , allowing more propagated noise to pass , whereas positive feedback decreases this frequency , thus reducing the amount of noise . Whereas our study illustrates the effect of positive feedbacks , additional mechanisms could be used for reducing propagated noise by similarly increasing the averaging time . Such mechanisms include long linear cascades; cascades with an intermediate component that has a relatively large half-life [3 , 26]; or scenarios where both synthesis and degradation are essentially zero-order ( from the definition of H11 in Equation 5 ) . Finally , we note that systems that exhibit time delays together with bistability were not included in our screen but could also attenuate noise [17] . Positive feedbacks did not emerge as a recurrent network motif in several of the transcriptional networks analyzed [19] . One possibility is that designing the proper feedback that will maintain stability while providing noise buffering is evolutionarily difficult for the small size networks considered in these studies [19] , due to the requirement it imposes on the extent of nonlinearities ( Hill-coefficients ) of the interactions . A simple realization of this concept , however , can easily be implemented by somewhat larger networks , as exemplified by the coupled positive–negative feedback we described . This and similar implementations function over a broad range of parameters and do not require strict tuning . Further analysis will be required to assess the abundance of this positive feedback–based noise-reduction scheme in different biological systems .
All simulations were based on the Gibson-Bruck [27] modification of Gillespie [28] algorithm . Input noise was implemented via transcription from a low copy mRNA with a short half life . No other mRNAs were explicitly considered . Simulation parameters are detailed in Section VII of Text S1 . Simulations were carried out using Dizzy [29] . The interaction parameters of each arrow in each network are captured by the interaction elasticities Hij . The susceptibility ( Equation 2 ) for each network is calculated from the elasticities using Equation 6 ( for a general derivation of the susceptibility , see Text S1 , Section I ) . The noise amplification is connected to the elasticities through the solution of Equation 7 . Definitions for the terms in Equation 7 appear in Section II of Text S1 . Equation 7 was solved symbolically for all three node networks using Maple ( MapleSoft , Waterloo Maple ) . Solutions for specific network architectures are shown in Table S1 . The values for the elasticities were randomly assigned to each network . To control for similar distribution of positive and negative interactions , we defined the synthesis elasticity Sij by Hii = 1 − Sii and Hij = −Sij . When i enhances the synthesis of j , then Sij > 0 , and vice versa . Positive feedback of i on itself implies Sii > 0 and vice-versa . The synthesis elasticity values were sampled from a uniform distribution between zero and four and assigned to the arrows in each network . Different sampling ranges did not have a significant effect on the conclusions ( Text S1 , Section IV ) . We sampled 20 , 000 random sets of parameters for each circuit . The time constants were held fixed at a value of one , but different values did not change the results ( Text S1 , Section IV ) . Stability criteria were established via the sign of the eigenvalues of the interaction matrix ( Text S1 , Section III ) . | Biological circuits need to be sensitive to changes in environmental signals but at the same time buffer rapid fluctuations ( noise ) that might be imposed on this input . In this paper , we analyze the interplay between sensitivity to signals and the ability to buffer noise . Previous studies reported that negative feedback attenuates noise . We show , however , that this ability comes at the expense of an even more dramatic reduction in sensitivity . In fact , when comparing systems of the same sensitivity , a system with negative feedback is more amenable to noise than a system without such feedback . We searched for small biological circuits that can buffer noise while maintaining high sensitivity , and found that positive feedback exhibits this property . This ability of positive feedback to buffer noise reflects its slowed-down dynamics . We discuss general requirements for the function of positive feedback as a noise-filtering device and describe a particular implementation that appears to function in yeast nutrient homeostasis . Our study emphasizes the need to consider multiple constraints when analyzing the design logic of biological networks . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"saccharomyces",
"biophysics",
"none",
"computational",
"biology"
] | 2008 | Noise Propagation and Signaling Sensitivity in Biological Networks: A Role for Positive Feedback |
Comparative analyses of pathogen genomes provide new insights into how pathogens have evolved common and divergent virulence strategies to invade related plant species . Fusarium crown and root rots are important diseases of wheat and barley world-wide . In Australia , these diseases are primarily caused by the fungal pathogen Fusarium pseudograminearum . Comparative genomic analyses showed that the F . pseudograminearum genome encodes proteins that are present in other fungal pathogens of cereals but absent in non-cereal pathogens . In some cases , these cereal pathogen specific genes were also found in bacteria associated with plants . Phylogenetic analysis of selected F . pseudograminearum genes supported the hypothesis of horizontal gene transfer into diverse cereal pathogens . Two horizontally acquired genes with no previously known role in fungal pathogenesis were studied functionally via gene knockout methods and shown to significantly affect virulence of F . pseudograminearum on the cereal hosts wheat and barley . Our results indicate using comparative genomics to identify genes specific to pathogens of related hosts reveals novel virulence genes and illustrates the importance of horizontal gene transfer in the evolution of plant infecting fungal pathogens .
Crop losses due to fungal pathogens represent one of the most serious threats to global food production . Staple cereal crops such as wheat , barley , rice and maize are subject to attack from a diverse array of fungal pathogens including biotrophs such as rust fungi that feed on living cells and necrotrophs such as Fusarium pathogens that kill host cells to obtain nutrients . Many Fusarium pathogens not only reduce crop yields but also produce mycotoxins that are harmful to humans and livestock when consumed in food and feed . A better understanding of the infection strategies used by these pathogens would help develop novel plant protection strategies . Comparative analysis of pathogen genomes offers a new and powerful approach to identify common and divergent virulence strategies as well as evolutionary history of pathogen lineages . Shared virulence strategies may be used by different fungi to invade specific plant hosts . Presumably in many cases , the existence of common virulence strategies in different pathogen species may be explained by conservation of virulence gene function through vertical inheritance and/or exposure to common host defensive selection forces during pathogenesis on the same or related hosts . However , in some instances , horizontal gene transfer events have been identified in fungal pathogens and subsequently shown to have roles in pathogenicity [1]–[3] . A striking example of a locus-specific horizontal gene transfer event emerged from the sequencing of the wheat pathogen Phaeosphaeria nodorum ( anamorph Stagonospora nodorum ) genome , where a gene encoding a host-specific protein toxin ( ToxA ) was identified by homology to a known toxin from another wheat pathogen Pyrenophora tritici-repentis . Functional analyses demonstrated that ToxA was necessary for virulence in both pathogens [1] . It was proposed that transfer of ToxA from P . nodorum to P . tritici-repentis resulted in the emergence of the tan spot disease of wheat caused by P . tritici-repentis in the 1930s [1] , [4] . In another example , genome analysis of the tomato vascular wilt pathogen F . oxysporum f . sp . lycopersici revealed the presence of several supernumerary chromosomes . Non-sexual transfer of one of these chromosomes to a non-virulent and genetically diverged recipient strain was shown to be sufficient to confer virulence on tomato [2] . Recently , association genomics has been used to identify the fungal effector Ave1 ( for Avirulence on Ve1 tomato ) in Verticillium dahliae . Ave1 homologs were shown to be present in diverse plant pathogenic fungi and important for virulence in at least one fungal species and one plant pathogenic bacterium [5] . In addition , Ave1 had strong homologies to plant proteins , suggesting that a cross-kingdom gene transfer event from plant to fungi may have occurred [5] . Ancient horizontally acquired virulence genes that have been retained because of their selective advantage may have more subtle sequence homologies and therefore are harder to detect [6] , [7] . Such genes are best identified by gene phylogenies where a gene associates with a clade of sequences from unrelated species and relationships are incongruent with established ancestries [7] . For example , Klosterman et al [3] found the presence of a glucosyltransferase encoding gene in the genomes of three different wilt causing fungi , F . oxysporum f . sp . lycopersici , V . dahliae , and V . albo-atrum as well as an insect pathogen Metarhizium anisopliae that was otherwise found only in bacteria . These authors proposed that the acquisition of this gene predates the Verticillium spp . split and probably occurred independently in F . oxysporum f . sp . lycopersici [3] . Using targeted gene disruption approaches , the glucosyltransferase was shown to be important for virulence in V . dahliae toward tobacco ( Nicotiana benthamiana ) but not lettuce ( Lactuca sativa ) . Similarly , it was recently suggested that the ability to synthesize auxin in members of both the Fusarium and Colletotrichum genera has probably resulted from horizontal transfer of auxin biosynthetic genes from bacteria [8] . Recent advances in genomics are revolutionizing the analysis of fungal species , which are particularly suited to analysis by the current generation of DNA sequencing technologies due to their relatively small genomes and in many cases minimal repetitive DNA contents . The de novo detection of shared virulence strategies without a priori information on the roles of shared genes in pathogen virulence offers an exciting possibility of uncovering new insights into the pathogenesis-related processes . The work of Klosterman et al [3] is one of the few examples where nothing was known about the role of the identified genes in pathogen virulence prior to being identified through comparative genomics . Therefore , it is reasonable to expect that further unbiased comparative genomic analyses will uncover examples of shared virulence genes or niche specialization genes in these pathogens , and ultimately provide insights into the co-evolution of virulence/niche specialization functions and the mechanisms of plant defense . In this paper , we report the sequencing , assembly and annotation of the genome of Fusarium pseudograminearum ( Aoki and O'Donnell ) using Illumina sequencing and a comparative genomics analysis of its gene content . In many parts of the world , Fusarium pseudograminearum is the principal cause of Fusarium crown rot ( FCR ) of wheat and barley . FCR is significant in arid cereal-growing regions worldwide including South Africa [9] , Northern Africa [10] , the Middle East [11] , Europe [12] and Australia [13] as well as the northwest of the United States of America [14] . In Australia , FCR is a chronic problem and among the most economically significant diseases of both wheat and barley [15] , [16] . The recent increase in incidence of FCR is ascribed to the increased use of conservation farming practices such as zero tillage and stubble retention . These practices permit the survival of fungal inoculum on residual plant matter across planting seasons . Fusarium root-rot is a related cereal disease caused by F . pseudograminearum as well as other fusaria [17] , [18] . F . pseudograminearum was initially distinguished from F . graminearum by host tissue preferences [19] and on the basis of molecular data was formally recognized as a separate species [20] , [21] . Multi-locus sequence analysis of diverse isolates has shown that F . pseudograminearum is a single phylogenetic species globally [22] in contrast to the F . graminearum species complex , which shows geographical structure [23] , [24] . In addition F . pseudograminearum is heterothallic whilst F . graminearum has a homothallic mating system . Whilst F . pseudograminearum , F . graminearum and F . culmorum are present throughout the Australian wheat growing regions and can all cause FCR , F . pseudograminearum is the species most commonly recovered from plants showing FCR symptoms [25] , [26] . Field surveys in Australia have revealed that F . graminearum is the most frequently isolated species from wheat plants showing Fusarium head blight ( FHB ) symptoms [26] , although F . pseudograminearum can also cause FHB disease . These observations suggest that F . pseudograminearum , while a broadly adapted cereal pathogen , may have evolved adaptations for niche specialization or infection processes that favor stem and/or crown infection . We hypothesized that at least some genes in the F . pseudograminearum genome that were either uniquely or predominantly present in other cereal fungal pathogens may have specialized functions related to cereal pathogenesis and niche specialization . Throughout the manuscript the term ‘niche specialization’ is used to encompass other potential aspects of the biology of these fungi for which are poorly understood such saprophytic colonization of dead plant material , non-pathogenic interactions with other plants or potential endophytism of other hosts . A broad comparative genomic analysis indicated that the F . pseudograminearum genome contains genes that have strong homology to genes that are unevenly distributed across cereal pathogens while being apparently absent in other fungal genomes . Some of these genes shared by cereal pathogens also encode proteins with significant similarity to those from plant-associated bacteria . This finding is consistent with multiple horizontal acquisition events and indeed phylogenetic analysis of selected genes supported the hypothesis of horizontal gene transfer into diverse cereal pathogens . Functional analysis of two potentially horizontally acquired genes revealed important roles in the virulence of F . pseudograminearum on cereal hosts . Our results illustrate an important role for horizontal gene transfer in the evolution of cereal associated fungi .
The genome of F . pseudograminearum isolate CS3096 was sequenced using a combination of paired and single-end Illumina reads . De novo assembly of these reads resulted in a nuclear genome size of 36 . 8 Mbp assembled in 656 contigs with 50% of all nucleotides in contigs of 189 Kbp in length or greater ( AFNW00000000 ) . The average sequence coverage across these contigs was 179-fold . Compared to many other fungal genome assemblies using next generation sequencing technologies , the F . pseudograminearum genome sequence assembly has a relatively high N50 ( Table 1 ) . Gene model predictions from three programs were combined to identify 12 , 448 protein coding gene ( see materials and methods ) . The repeated sequence content of the F . pseudograminearum genome , assessed using the RepeatMasker program , is only 1 . 6% , which is slightly higher than that of the F . graminearum genome ( 0 . 7% ) assessed using identical parameters , albeit that the sequencing and assembly methodology differed . RepeatMasker recognizes both simple sequence repeats and transposable elements present in RepBase [27] . Although approximately four times as many base pairs were flagged as being derived from Gypsy type Long Terminal Repeat ( LTR ) elements in the F . pseudograminearum genome ( 26 Kbp ) compared to F . graminearum ( 6 Kbp ) , the difference in repetitive DNA content could mostly be attributed to a higher level of simple repeats and low complexity DNA ( 1 . 5% versus 0 . 41% of the genome , respectively ) . One high coverage contig in F . pseudograminearum encodes a LTR-type retrotransposon with best match to an F . oxysporum transposon , probably present in 9–10 copies based on an average coverage of 1689 fold . This transposon matched to sequences in a number of different fungi and also to transposons in both monocot and dicot plants . The contig showed no polymorphism in the sequence read assembly across 5 . 5 Kbp , suggesting all copies are identical and thus could not be placed in other assembled contigs . Also not included in the overall repeat counts are the contigs that represent rRNA encoding genes . The Illumina sequencing approach used here was unable to resolve these repeats in F . pseudograminearum and these currently appear in the assembly as high-coverage contigs . The F . pseudograminearum genome sequence assembly was also compared globally to that of F . graminearum [28] by aligning the two genomes after masking simple repeat sequences and known fungal repetitive elements . In total , 89 . 8% of the F . pseudograminearum genomic sequence could be aligned to the F . graminearum genome at >70% nucleotide identity ( Figure 1 ) . An alignment with increased sensitivity was performed using a six frame translations of both genomes enabling an alignment of 94% of the F . pseudograminearum genome to that of F . graminearum . Thus , at least 6% of the low copy region of the genome ( approximately 2 . 2 Mbp ) appears to be completely unique to F . pseudograminearum . Very few rearrangements between the F . pseudograminearum and F . graminearum genomes in the aligned regions were found ( Figure 1A ) . The amount of aligned sequence between the two species decreases towards the ends of the F . graminearum chromosomes ( Figure 1B ) and also in regions previously reported to be undergoing higher rates of genome innovation [28] . Genes showing a distribution of homologues limited to cereal pathogens would be both candidates for strong selection , because of a possible involvement in pathogenicity , and for horizontal acquisition . To determine whether the F . pseudograminearum predicted gene set contained such candidates , a BLASTmatrix analysis pipeline based on the identification of gene homologues by reciprocal best BLASTp hits was developed using predicted proteomes from the genome sequence of 16 different cereal pathogens and 11 non-pathogens of cereals . The entire BLASTmatrix analysis is presented in Dataset S1 along with the filtered protein sets described below . A numerical summary of the matches is shown in Table S1 . In these analyses , 156 predicted F . pseudograminearum proteins did not have reciprocal best BLAST hits of any strength in any of the 27 organisms , while 239 predicted proteins had reciprocal best BLASTp matches only in cereal pathogens , with a score of at least 200 bits . In contrast , applying the same filtering criteria to identify proteins that are found in F . pseudograminearum and only non-cereal pathogens in this BLASTmatrix analysis identified only 32 proteins , 24 of which were present in the Fusarium lineage . Amongst the 239 cereal pathogen proteins , 214 had matches within the two other Fusarium cereal pathogens ( F . graminearum and F . verticillioides ) and not in other species . Amongst the remaining 25 F . pseudograminearum proteins , only five had no match in other fusaria while matching other cereal pathogen proteins . Since the genome sampling described herein was limited to 27 fungal genomes , the sets of 239 and 156 proteins were combined and additionally curated using BLASTp against the NCBI nr ( non-redundant ) database . This process identified a total of 17 proteins with strong BLASTp matches ( >200 bits ) to bacterial proteins , and six of these 17 proteins were present only in a small number of other fungi ( Table 2 ) . The corresponding genes for these proteins are therefore good candidates for having been ancestrally acquired from bacteria . The GC content for each of these six genes was not obviously different to that of surrounding genes , possibly with the exception of FPSE_07765 ( see below ) . Other candidates for cereal pathogen specificity , which to varying degrees showed distributions limited to cereal pathogens , were also identified through the BLASTmatrix filtering . These are therefore also of interest with respect to putative functions in niche specialisation and virulence towards plants . Examples of predicted genes from this latter category suggesting conserved roles in virulence towards cereals , and/or horizontal gene transfer are shown in Table 3 . These candidate genes were subjected to more detailed analysis of sequence relationships as described below . Amongst the F . pseudograminearum genes of potential bacterial origin ( Table 2 ) , the intronless gene FPSE_07765 and its orthologues in F . graminearum and F . verticillioides encode proteins highly similar to a bacterial protein with both aminotransferase and homoserine kinase domains . FPSE_07765 is 75% identical at the amino acid level across the entire length to a protein from Microbacterium testaceum , which is a bacterial endophyte of a variety of plants including the cereals sorghum and maize [29] , [30] . Other fungal hits to this sequence in GenBank only align across less than half of the protein and at much lower identities , with the best fungal match in Trichophyton verrucosum at only 31% amino acid identity and with only partial query coverage of 44% . Phylogenetic analysis of FPSE_07765 identifies a Fusarium sequence clade embedded amongst a range of related bacterial sequences and strongly supports an origin of this gene via horizontal acquisition from bacteria , with retention in the cereal infecting fusaria ( Figure S1 ) . The GC content of FPSE_07765 was 58% compared to the genome average of 51 . 8±3 . 9% ( ±SD ) for coding sequences . FPSE_11233 , another F . pseudograminearum gene with possible horizontal acquisition , encodes a putative secreted hydrolase that has reciprocal best BLAST hits only in cereal pathogens among the 27 selected genomes as well as hits in GenBank to putative pectin- and xylan-hydrolases in bacteria ( best match is 38% identical with an e-value of 1e−125 to a Streptomyces hygroscopicus protein ) . Further manual database queries showed FPSE_11233-like sequences in the cereal endophyte Chaetomium globosum and the Brassica pathogen Leptosphaeria maculans . This gene sequence was analyzed further because , although L . maculans is not a cereal pathogen , it causes blackleg disease of canola which is a common rotation crop used in cereal farming systems [31] , and L . maculans is related to other cereal pathogens in the Pleosporales family [32] , [33] . Again , phylogenetic analysis supported the relatedness of these fungal genes to bacterial sequences ( Figure S2 ) . However , any potential acquisition from bacteria is clearly extremely ancient as the sequence seems to have undergone considerable vertical diversification within the fungi following acquisition from bacteria . This hypothesis is not only supported by the phylogenetic analysis but also computational predictions , suggesting that the position and number of introns in FPSE_11233-like sequences are not conserved between fungal species ( data not shown ) . The placement of the L . maculans sequence outside of the Dothideomycete clade in the phylogenetic analysis ( Figure S2 ) may also indicate a more complex mode of inheritance within the fungi . A number of other enzyme-encoding genes of putative bacterial origin were also identified in the F . pseudograminearum genome , including genes encoding a hydrolase of unknown specificity ( FPSE_07775 ) and a NAD+ dependent dehydratase/epimerase ( FPSE_11221 ) . Two putative amidohydrolase encoding genes ( FPSE_00725 and FPSE_05717 ) that have clear bacterial homologues were also identified in this analysis . One of the amidohydrolase genes , FPSE_05717 hereafter termed FpAH1 , also appeared to have a clear homologue in P . nodorum but not in any other fungus , suggesting a potential role for this gene in cereal pathogenesis and an unusual evolutionary history . The genomic organization , sequence diversity and virulence function of FpAH1 will be described in more detail later in the manuscript . Many F . pseudograminearum genes also had homologues in other fungal cereal pathogens but no clear bacterial matches , suggesting that these genes are either laterally inherited or rapidly diversifying and therefore have been selectively retained only in a limited number of pathogenic fungi . Examples of these ( Table 3 ) include three genes ( FPSE_06956 , FPSE_10646 and FPSE_02381 ) encoding small secreted proteins ( candidate effector molecules ) that were present in F . pseudograminearum and F . graminearum as well as other cereal pathogens . FPSE_06956 had orthologous matches in the Magnaporthe and Fusarium lineages but no other matches in the BLASTmatrix or GenBank . FPSE_10646 is a member of the killer protein 4 family ( PFAM09044 ) of toxins that were shown to be extensively laterally shared between multiple fungal lineages including non-pathogens [34] . FPSE_02381 is a member of a two-gene family encoding small secreted cysteine rich proteins in F . pseudograminearum and F . graminearum and has strong , albeit non-reciprocal BLASTp matches in M . oryzae . The other member of this family in F . pseudograminearum , FPSE_05488 , had reciprocal best BLASTp hits only in cereal pathogens , with the exception of a single match in the canola pathogen Colletotrichum higginsianum in the BLASTmatrix analysis ( Dataset S1 ) . Three other gene products shown in Table 3 ( FPSE_05718 , FPSE_05719 and FPSE_05720 ) were encoded in a gene cluster in the F . pseudograminearum genome . Interestingly , FpAH1 , which is predicted to be of bacterial origin ( see above and Table 2 ) , also seems to be part of this cluster . The function of FpAH1 will be described in more detail later in the manuscript . Also shown in Table 3 is a gene encoding a putative dienelactone hydrolase ( FPSE_08136 ) , hereafter termed FpDLH1 , with very strong orthologues in F . verticillioides and C . graminicola with 98% and 88% identities , respectively at the amino acid level . The next best BLASTp match ( 69% identical ) of FpDLH1 in GenBank , which is not a reciprocal best hit , is to FGSG_00089 from F . graminearum . The absence of strong reciprocal matches in all other fungi suggests that this gene may have been either shared between , or independently acquired from another donor species , by these Fusarium and Colletotrichum species through horizontal transfer mechanisms . The genomic organization and function of FpDLH1 will be described in more detail later in the manuscript . FpAH1 encodes a putative amidohydrolase ( Pfam domain PF07969 ) . The best BLASTp match ( e-value 2e−141 , 88% identical , 94% similar at the amino acid level ) of FpAH1 was to a predicted protein from the wheat glume blotch pathogen P . nodorum ( encoded by SNOG_04819 hereafter termed PnAH1 ) [35] . The lack of FpAH1 homologues in other fusaria and presence in P . nodorum was confirmed via hybridization of a FpAH1 probe to genomic DNA of various Fusarium and a P . nodorum isolates ( Figure S3 ) . The overwhelming majority of subsequent BLASTp matches of FpAH1 in GenBank were to predicted proteins from bacteria ( Table 4 ) . The next fungal match after P . nodorum was to the F . graminearum protein FGSG_10599 ( 25% identical 2e−25 ) but this was much weaker than the bacterial matches ( Table 4 ) . FGSG_10599 also had another much stronger , indeed nearly identical ( 97% ) protein ( FPSE_00725 ) encoded in the F . pseudograminearum genome . The FPSE_00725/FGSG_10599 orthologs also had next best matches in bacteria . Phylogenetic analysis of these two F . pseudograminearum amidohydrolases strongly supports the hypothesis that they are of bacterial origin ( Figure 2 ) . The F . graminearum genome contains six entries in this class of amidohydrolases ( PF07969 ) , all of which contain predicted orthologous matches with other genes in the F . pseudograminearum genome ( Table S2 ) . In other fusaria , clear orthologous relationships exist between the remaining five amidohydrolases with Pfam domain PF07969 found in F . pseudograminearum , although both F . oxysporum and F . verticillioides contain additional members in this class of amidohydrolases ( Table S2 ) . In P . nodorum , only four proteins fall into this class of amidohydrolases . Thus FPSE_00725 and FpAH1 encode two amidohydrolases with extremely restricted distribution in fungi and are most likely of bacterial origin . BLASTp analysis of the other F . pseudograminearum amidohydrolases identified a much wider distribution in fungi with FPSE_00474 , FPSE_02365 , FPSE_03227 and FPSE_11444 having more than fungal 30 hits of greater strength than the best bacterial hit ( Table S2 ) . FPSE_05738 was less widely distributed with seven hits of greater score than the best bacterial hit ( Table S2 ) and may also be a candidate for acquisition from bacteria . The predicted FpAH1 protein of 570 amino acid residues was encoded by an uninterrupted open reading frame of 1710 bp that was confirmed by RNAseq analysis of cDNA ( data not shown ) . The P . nodorum genome annotation for PnAH1 contained a single intron , but it is likely that this prediction was not correct as the PnAH1 genomic region has a single uninterrupted open reading frame . In the coding region , FpAH1 was conserved between F . pseudograminearum and P . nodorum with 89% identity at the nucleotide level and 174 bp upstream of the predicted start codon and 85 bp downstream of the predicted stop codon could also be readily aligned . The chromosomal complement of F . pseudograminearum has not been characterized and therefore the location of FpAH1 in the F . pseudograminearum genome is unknown . FpAH1 is present near the end of a ∼90 kb sequence contig , the first 70 kb of which aligns with the end of F . graminearum chromosome 1 , as shown in Figure 3 . Interestingly , the genes surrounding FpAH1 did not have clear orthologues in the F . graminearum genome . On the contig containing FpAH1 , 28 genes ( FPSE_05686 through to FPSE_05714 , excluding FPSE_05712 ) had clear orthologues in F . graminearum . However , FPSE_ 5715 to FPSE_05720 did not . Furthermore , three of the gene products ( FPSE_05718 , FPSE_05719 and FPSE_05720 ) were also identified in the BLASTmatrix analysis to be cereal pathogen-specific , suggesting that parallel selection or coordinated acquisition may have affected this region ( Table 3 ) . The putative function and position of the genes in this region in F . pseudograminearum are shown in Figure 3 . Of the remaining 40 genes on the end of chromosome 1 in F . graminearum , only eight encoded proteins with reciprocal best BLAST hits in F . pseudograminearum , and these were distributed across six different contigs in the F . pseudograminearum assembly ( data not shown ) . In contrast , PnAH1 is ∼100 kb from the end of supercontig seven of the P . nodorum genome and appears to be surrounded by genes encoding proteins conserved in other fungi . In both P . nodorum and F . pseudograminearum , the GC content in the region of the PnAH1 and FpAH1 was similar to that of other gene rich regions of the respective genomes and some of the surrounding genes in the regions had introns . In summary , the genomic region containing FpAH1 has no equivalent sequence at the syntenic location in the F . graminearum genome nor the region containing PnAH1 in the P . nodorum genome . The lack of close orthologues of FpAH1 and PnAH1 in other fungi and the presence of similar genes in bacteria ( Table 4 ) based on BLASTp searches suggested an origin for these fungal genes via horizontal acquisition from bacterial species . fusaria belong to the order/class Hypocreales/Sordariomycetes while Phaeosphaeria is in the distantly related Pleosporeales/Dothidiomycetes . Acquisition of the gene may have occurred independently in both species or alternatively was horizontally transferred between the fungal species . To differentiate these possibilities , the sequence diversity of each gene was assessed in several globally sourced Fusarium and Phaeosphaeria isolates ( Table S3 ) . Isolates identified as F . pseudograminearum by multi-locus DNA sequencing [22] or by sequencing the elongation factor 1 alpha ( EF1α ) gene , were the only Fusarium sp . out of six tested that produced positive amplicons . PCR analyses of the distribution of FpAH1 in fusaria were confirmed for a limited number of isolates by hybridization analysis ( Figure S3 ) . PnAH1 sequences were PCR amplifiable from all P . nodorum isolates tested and also from two of four sister species , P . avenaria f . sp . tritici ( Pat ) 1 and Pat3 , suggesting PnAH1 was present in a common ancestor in this lineage . In order to test the hypothesis of horizontal transfer between F . pseudograminearum and P . nodorum , a ∼500 bp region of the AH1 gene was sequenced in isolate collections . A haplotype network showing the sequence relationships across both genera is shown in Figure 4 . In F . pseudograminearum , seven haplotypes observed did not correspond to the geographic origin of isolates , consistent with global gene flow within the species as previously described [22] . Diversity in the Phaeosphaeria spp . sequences was more limited with only one haplotype observed in a global sample of P . nodorum and two haplotypes detected within each of the two Phaeosphaeria sister species containing PnAH1 orthologues . There were no shared sequence haplotypes between Phaeosphaeria spp . and F . pseudograminearum . The presence of AH1 orthologues in up to four related Phaeosphaeria spp . suggests that the acquisition of this gene occurred before the divergence of these species . Divergence of PnAH1 between the Phaeosphaeria spp . was comparable to divergence observed at neutral sequence loci ( M . C . McDonald , unpublished data ) . Furthermore , the haplotypes observed in the Phaeosphaeria species complex and F . pseudograminearum were distinct , suggesting independent acquisitions of AH1-like sequences by each lineage . The likely independent acquisition and retention of AH1 orthologues by F . pseudograminearum and P . nodorum suggests that this gene may encode a protein that is necessary for virulence , at least in some hosts . To test this hypothesis , a functional analysis of the FpAH1 and PnAH1 genes was undertaken . FpAH1 was expressed in both infected barley and wheat leaves and roots ( Figure S4 ) . The role of the FpAH1 amidohydrolase in FCR of barley was also assessed by generating gene deletion mutants in F . pseudograminearum by replacing 635 nucleotides of the FpAH1 locus with a geneticin resistance gene cassette ( Figure S5A ) . There were no obvious defects in appearance or sporulation of the mutants . A culture time-course was used to compare the growth rates of one mutant to that of the parental strain and these were indistinguishable ( Figure S5C ) . As shown in Figure 5 , two independently-derived knock-out mutants consistently showed reduced virulence towards barley ( cv . Golden Promise ) seedlings across multiple independent experiments using a previously established FCR inoculation assay [36] . A second barley cultivar ( cv Gairdner ) also showed similar reduced virulence ( data not shown ) . FCR disease severity caused by the mutant strains was significantly ( P-value<0 . 01 ) reduced compared to those caused by the wild type strain . Genetic complementation of an FpAH1 mutant with a cassette containing FpAH1 under the control of the Aspergillus nidulans TrpC promoter restored virulence towards barley ( Figure 5C and 5D ) , providing further evidence that FpAH1 is required for virulence against barley . F . pseudograminearum has a wide host range within cereals with no evidence for race specialization [37] , [38] for FCR disease on wheat . Unlike the experiments conducted on barley , highly replicated FCR assays on wheat using the FpAH1 knock-out mutants failed to reveal reduction of virulence on two unrelated varieties of wheat ( Figure S6 ) . However , reduced virulence was detectable when inoculated directly on roots toward both wheat and barley and again complementation restored virulence ( Figure 6 ) . P . nodorum is not pathogenic on barley and therefore was not tested on this host but mutant strains of this pathogen with deletions of the PnAH1 gene were also generated ( Figure S5D ) and tested on wheat plants in replicated leaf infection assays . No significant differences in virulence were observed between mutant and wild type strains on wheat ( Figure S7 ) . These experiments indicate that FpAH1 is required for full virulence on both wheat and barley but the importance of PnAH1 in pathogenesis remains unknown . The F . pseudograminearum locus FpDLH1 encoding a dienelactone hydrolase was identified in the BLASTmatrix analysis as a cereal-pathogen associated gene with homologues detected only in F . verticillioides and C . graminicola . Strikingly , the gene ( FPSE_08135 hereafter termed FpAMD1 ) immediately adjacent to FpDLH1 in the F . pseudograminearum genome encodes an amidase that is also present in the same divergently transcribed arrangement in the genomes of F . verticillioides and C . graminicola ( Figure 7 ) , indicating this is likely to be a two-gene cluster . The amidase family is much larger in ascomycetes than the dienelactone hydrolase family , making one-to-one orthologous relationships between species difficult to detect . However , for both FpAMD1 and FpDLH1 , strong homology is detected between all three species at the nucleotide level ( >80% ) across the coding sequences of these genes . Furthermore phylogenetic analysis also supports the grouping of both of these genes in the three species in clades incongruent with expected evolutionary relationships ( Figure S8 ) . Thus , their genomic arrangement , discontinuous distribution in the fungi and strong homology to each other all suggest these genes represent a two-gene cluster that may have a common origin . The FpDLH1 gene has the same intron-exon structure as another dienelactone hydrolase that is present in both F . pseudograminearum and F . graminearum ( FPSE_08131 and FGSG_00089 , Figure 7 ) . This intron arrangement is not a generic feature of dienelactone hydrolase encoding genes in F . pseudograminearum , indicating a possible gene duplication event in the Fusarium lineage . Furthermore , FPSE_08131 had reciprocal best BLAST hits in all fusaria included in the analysis , with the exception of F . solani , as well as in two Magnaporthe spp . , and the intron-exon structure was maintained in all species . The synteny in this region is also well conserved in the F . pseudograminearum-F . graminearum comparison ( albeit split into two regions that align near the ends of different chromosomes in F . graminearum ) but outside of this comparison the synteny weakens with the orthologues of the genes flanking this two-gene cluster in F . pseudograminearum appearing on multiple different contigs in F . verticillioides ( Figure 7 ) . In the F . verticillioides-F . pseudograminearum comparison , the conservation of FpDLH1/FvDLH1 ( 98% identical at the amino acid level ) is much greater than the more widespread dienelactone hydrolase ( FPSE_08131 and its F . verticillioides orthologue FVEG_12625 ) at 75% identity , suggestive of either an intra-fusaria transfer or strong conservation and selection . In C . graminicola the DLH1-AMD1 gene cluster is located on chromosome 2 in a region where only two of 11 genes flanking the cluster have clear orthologous relationships between C . graminicola and the closely related C . higginsianum ( data not shown ) . However , three of the nine genes on one flank of the C . graminicola DLH1-AMD1 gene cluster have orthologues present in the flank adjacent to FpDLH1-FpAMD1 ( Figure 7 ) . These genomic associations suggest an ancient relationship in these regions . Both FpDLH1 and FpAMD1 are expressed during infection of root by F . pseudograminearum and leaf tissue with higher expression in wheat than in barley ( Figure S4 ) . The role of FpDLH1 in fungal pathogenesis was assessed by creating knockout strains of two different F . pseudograminearum strains ( CS3096 and CS3427 ) , where the FpDLH1 gene was replaced by the geneticin resistance cassette ( Figure S9 ) . No obvious differences in sporulation were observed in the mutants nor were there differences in growth rate compared to their respective parents ( Figure S9C ) . However , FpDLH1 mutants in both strain backgrounds showed significantly reduced virulence towards both wheat and barley in both root-rot and FCR assays ( Figure 8 and Figure S10 ) , suggesting that FpDLH1 contributes to fungal virulence against cereal plants .
The new DNA sequencing technologies are well suited to characterizing low copy , gene-rich regions of fungal genomes that are relatively small in size . Using Illumina technology it was possible to assemble de novo an almost complete sequence of the F . pseudograminearum genome . A large proportion ( ∼94% ) of the F . pseudograminearum genome showed high similarity to F . graminearum and a great deal of co-linearity was observed . Alignment of the F . pseudograminearum genome sequence to the chromosomes of F . graminearum revealed that regions of poor sequence match were concentrated in specific regions , such as the ends of chromosomes and what are thought to be regions of ancient chromosome fusion in this Fusarium lineage and probable regions of genome innovation [2] , [28] . Although more strains will need to be sequenced to confirm the species-specificity of these regions of the F . pseudograminearum genome , it is possible that the genes contained in these regions may be responsible for various phenotypes that distinguish F . pseudograminearum from F . graminearum , such as its propensity to cause FCR rather than FHB , its broad geographical adaptation in arid cereal production areas [39] . We hypothesized that genes in the F . pseudograminearum genome that were either uniquely or predominantly present in other cereal fungal pathogens may have a specialized function related to cereal pathogenesis and niche specialization . To identify these genes , we undertook a BLASTmatrix analysis and found that many genes present in the F . pseudograminearum genome were also present exclusively in cereal pathogens . 214 of these genes appeared to be conserved in the three cereal-infecting fusaria but had no equivalent matches in the genomes of three fusaria that infect dicotyledonous plants . These genes may have undergone specialized selection in these cereal pathogen Fusarium lineages but been diversified or lost in other fusaria . Several F . pseudograminearum genes also appeared to have equivalents in cereal pathogens outside the fusaria , and these were present mostly in other necrotrophic or hemibiotrophic Ascomycete fungi . Two of these genes , encoding an amidohydrolase ( FpAH1 ) and a dienelactone hydrolase ( FpDLH1 ) were selected for functional analysis in F . pseudograminearum . In both cases , we demonstrated roles in virulence on cereal hosts . This suggests that the comparative genomic analysis that we undertook is a powerful approach for the identification of genes with specialized pathogenesis on related hosts . Our analyses also identified many other candidate F . pseudograminearum genes showing similar distributions and a systematic analysis of their functions in fungal virulence is now warranted . Our analysis identified genes in the genome of F . pseudograminearum with matches in bacterial genomes , and a number of genes were restricted to other fungal pathogens of cereals . These findings are consistent with possible acquisition of these genes by horizontal transfer . These observations also suggest considerable genome plasticity in F . pseudograminearum and provide a number of candidates for further study of potential horizontal acquisition . Compelling evidence for acquisition of a gene of bacterial origin and retention in cereal-infecting fusaria is illustrated by FPSE_07765 encoding a putative aminotransferase . Most matches to this aminotransferase were from bacteria , with the closest match showing a remarkable 75% amino acid identity to a predicted protein in the genome of Microbacterium testaceum , a common bacterial endophyte of cereals . Horizontal transfer from co-habiting endophytic bacteria into the Fusarium lineage with selective retention in cereal pathogens is a simple explanation for these strong , but restricted gene homologies and organismal relationships . Another significant bacterial match was the putative cell wall hydrolase encoded by FPSE_11233 , with equivalent proteins present in many cereal pathogens . In this case , phylogenetic analysis indicated that all identified proteins from diverse fungal species clustered into a single clade . The biased occurrence of this gene in plant pathogens is likely due to selective retention after an ancient acquisition event during fungal evolution . FpAH1 represents a striking example of likely horizontal movement of a gene from a bacterium into the F . pseudograminearum genome . The only closely related gene to FpAH1 in the fungi examined was PnAH1 found in the genome sequence of P . nodorum , a pathogen of wheat . Sequencing of several globally distributed isolates of F . pseudograminearum revealed several distinct haplotypes for this gene . This suggests that acquisition of FpAH1 was not recent or that there has been significant selection driving the creation of new alleles . There was no evidence of horizontal transfer directly between F . pseudograminearum and a Phaeosphaeria spp . The limited divergence of PnAH1 between the Phaeosphaeria spp . orthologues , suggests that a common ancestor of these Phaeosphaeria species acquired the gene . The presence of the gene in at least two lineages of cereal pathogens suggests that the gene may play an important role in wheat pathogenesis . Both FpAH1 and PnAH1 were clearly dispensable for growth and their retention in two otherwise unrelated fungal pathogens suggests a specialized function for these genes in fungal virulence . Indeed , a virulence function for FpAH1 against two different cereal hosts was confirmed . However , PnAH1 knockout strains did not show altered virulence against wheat . Pathogenesis in any one species is the sum of many different components and the relative contribution of these genes to pathogenesis in the different species may be substantially different . The genomic context of the genes in these two fungal pathogens was also very different . PnAH1 is located in a region of the P . nodorum genome adjacent to several conserved genes . In contrast , FpAH1 occurs at the end of a long contig in a cluster of genes found in other cereal pathogens , but not in F . graminearum , and it could be that FpAH1 is functioning in F . pseudograminearum as part of this group of genes . It appears that most of the genes on the end of chromosome 1 in F . graminearum are absent from F . pseudograminearum . This observation provides further support to the notion that chromosome ends or ancient chromosome fusion sites are regions of genome innovation in the Fusarium lineage [28] and regions like this may have played a role in niche separation between F . pseudograminearum and F . graminearum . The role and reason for retention of PnAH1 in P . nodorum however , remains unknown . The two-gene cluster represented by FpDLH1-FpAMD1 is likely to be of fungal origin in the genomes of F . pseudograminearum , F . verticillioides and C . graminicola . A comparison of the genomic context of the DLH1-AMD1 genes in F . pseudograminearum and C . graminicola also supports a common origin with some related genes being present in flanking regions of this cluster in both pathogens . The conservation ( ∼80% ) of the nucleotide sequence across the coding regions of both genes between C . graminicola and the two Fusarium spp . suggests that any possible genetic exchange event between these lineages is ancient and has thereby allowed accumulation of considerable sequence divergence . The close physical proximity in the F . pseudograminearum genome of FpDLH1 to another dienelactone hydrolase encoding gene , FPSE_08131 , with identical intron-exon structure and with orthologues in other fusaria , suggests FpDLH1 may have arisen vertically by gene duplication and subsequent diversification within the Fusarium lineage . The FpDLH1/FvDLH1 and FpAMD1/FvAMD1 orthologues of F . pseudograminearum and F . verticillioides show remarkable identity ( 98% ) at the amino acid level but very different genomic context . However , currently , it is not possible to resolve whether an inter-species transfer had occurred between the Fusarium lineages , or alternatively whether strong DLH1 and AMD1 gene conservation with regional genomic rearrangements had occurred in these two cereal infecting fusaria . Further diversity surveys will be required to resolve the origin of FpDLH1- and FpAMD1- related genes in these cereal pathogens . Both FpAH1 and FpDLH1 are thought to be catabolic enzymes based on conserved domain matches . Amidohydrolases are a diverse superfamily of enzymes that catalyze the hydrolysis of C-N bonds in small molecules . This family includes enzymes functioning in central metabolism ( eg urease ) , enzymes that degrade xenobiotics ( eg atrazine ) as well as those that are known to catalyze reactions other than C-N cleavage , including P-O cleavage and also isomerisation [40] . Likewise , the dienelactone hydrolase family of enzymes appears to be large , with members involved in the degradation of chloroaromatic compounds [41] , [42] by bacteria and activation of prodrugs containing lactone-like side chains in humans [43] . The specific biochemical roles of FpAH1 and FpDLH1 during this particular host-microbe interaction remain elusive . However , the predicted catabolic activity of these enzymes suggests they could be either targeting specific plant defense compound ( s ) or involved in production of fungal toxin ( s ) . Although the defensive compounds important in the response to F . pseudograminearum are currently unknown in barley and wheat , candidates may include hordatine , hordenine and gramine in barley and the benzoxazolinones in wheat [44]–[47] . All these compounds contain C-N bonds and are known to have antifungal properties [44]–[47] . However , fungal growth inhibition assays conducted in liquid culture medium with synthetic hordatine and gramine showed that while F . pseudograminearum is moderately sensitive to both compounds , FpAH1 knockout strains were equally as sensitive as the wild type ( data not shown ) . Likewise , the FpDLH1 mutants were as sensitive as the wild-type to the benzoxazolinones , BOA ( 2-3H-Benzoxazolinone ) and MBOA ( 6-methoxy-2 ( 3 ) -benzoxazolinone ) . However , in F . verticillioides BOA detoxification is a two-step process with only one of the two responsible genes having been cloned [48] . A more detailed understanding of the barley and wheat metabolites involved in defense against F . pseudograminearum would be an important starting point to identify molecular mechanism ( s ) of FpAH1- and FpDLH1- mediated virulence . Genomics is allowing increased discovery of examples of likely horizontal gene transfer events [6] , [7] . Our work provides additional evidence for this hypothesis although detection of horizontal gene transfers , particularly of ancient events , requires robust phylogenetic analyses [3] , [7] , [49] . The extent of the horizontal gene transfer phenomenon has not yet been fully ascertained amongst the sequenced fungal genomes but there will presumably be many more examples . A robust methodology needs to be developed to enable identification of horizontally acquired genes on a scale of kingdoms and beyond . The BLASTmatrix analysis presented here is one possible method for identifying these genes , although it is limited in its ability to cope with post-acquisition family expansion . It is also important to note that the analysis reported here was centered on F . pseudograminearum and limited by the species included in the comparative analyses . A broader systematic analysis for each fungal pathogen genome is warranted to obtain a more complete perspective of potential gene sharing and its relation to host range and virulence . In summary , in this paper we report the first genome sequence for F . pseudograminearum and demonstrate that a broad comparative genomic analysis can identify genes that show a biased distribution in fungal pathogens of plants with putative roles in virulence processes or niche specialisation . Functional analysis in F . pseudograminearum of two such genes demonstrated novel virulence functions on cereals .
The F . pseudograminearum isolate ( CS3096 ) chosen for genome sequencing was originally isolated from a wheat crown collected in 2001 near Moree in Northern New South Wales , Australia [26] . Isolates for phylogenetic analysis of the FpAH1 gene were selected from a collection housed at CSIRO Plant Industry , Brisbane , Australia , consisting of isolates from Australia , New Zealand , Canada , United States of America and Turkey ( Table S3 ) . Phaeosphaeria isolates were selected from a collection housed at ETH Zurich and the Australian isolate SN15 . Two lanes of single-end 75 bp and one lane of 100 bp paired-end sequencing was performed on an Illumina GAIIx genome sequencer by the Australian Genome Research Facility , Melbourne Australia . Reads were imported into CLCBio Genomics Workbench 3 and quality trimmed ( default parameters ) prior to assembly . A total of 60 million paired reads and 33 million single end reads were used in de novo assembly plug in ( version 3 . 03 ) , again using default parameters with a minimum contig size of 500 bp . Low coverage ( <40× ) contigs ( compared to the genome average of 179× ) were excluded from further analyses . BLASTn comparison to the F . graminearum mitochondrial sequence and coverage information ( 3 , 500–6 , 000× ) was used to separate nuclear and mitochondrial sequence . 29 contigs with total length 107 Kbp were identified in this process and excluded from the genome annotation but are included in the F . pseudograminearum genome submission to GenBank . This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession AFNW00000000 . The version described in this paper is the first version , AFNW01000000 . Repeat masking was performed using RepeatMasker version open-3 . 2 . 8 [50] , run in sensitive mode with cross_match version 0 . 990329 as the search engine and RepBase version 20090604 [27] . The species descriptor “fungi” was used . Whole genome alignments were performed with the NUCmer and PROmer algorithm in the MUMmer package [51] . The minimum cluster length was set to 100 bp . The total length of the genome alignment between F . graminearum and F . pseudograminearum was calculated by summing the length of all non-redundant alignments extracted by using the show-coords program ( part of the MUMmer package ) . Protein coding genes were ab initio predicted in the F . pseudograminearum genome using FGENESH [52] based on the F . graminearum gene models as part of the MolQuest2 package from Softberry , AUGUSTUS [53] and GeneMark-ES version 2 [54] . AUGUSTUS was run with a training set of F . graminearum genes that had 100% nucleotide alignment and identity to the F . pseudograminearum genome . When the three programs disagreed in the splicing for the same gene , the following strategy was used to prioritize the predictions . When two of the three programs agreed , this prediction was taken . When all three programs differed for what appeared to be the same gene , priority was given in the order FGENESH , GeneMark and AUGUSTUS . Predicted transcripts encoding <20 amino acids were removed manually from the main prediction set . In total 17 , 503 unique transcripts were predicted , which after removing 4833 alternative predictions for the same gene left a putative set of 12448 unigenes . The proteins coded by 27 different fungal genomes were downloaded from the Broad Institute ( www . broadinstitute . org ) , Joint Genome Institute [55] or NCBI for the F . oxysporum strain 5176 genome [56] . Reciprocal best BLASTp hits were used to determine the orthologous protein relationships between F . pseudograminearum and each of the 27 fungal protein sets . This was performed using a pair-wise all versus all BLASTp analysis was conducted on the National Computational Infrastructure Specialized Facility in Bioinformatics Cluster located at the University of Queensland , Australia . Perl scripts obtained from http://sysbio . harvard . edu/csb/resources/computational/scriptome/unix/Protocols/Sequences . html were used to extract best reciprocal hits and the associated blast score from each species comparison BLASTp outputs . The F . pseudograminearum–F . graminearum relationship was treated as a special case and used to curate the dataset . In total , 11481 genes had clear one-to-one relationships between the two species , but differences in gene prediction ( eg splice sites or predictions split/fused with respect to each other in each of the genomes ) resulted in false identification of genes without orthologous matches in this analysis . These were accounted for by performing BLASTn analysis using F . pseudograminearum gene sequences containing introns as the query against the F . graminearum genomic sequence . Transcripts showing strong matches to the F . graminearum genome ( match length greater than 50% ) were classified as having different annotation between F . graminearum and F . pseudograminearum . 456 genes were removed from the analysis using this process . Logical formulae and filtering functions in Microsoft Excel were used to identify potential cereal pathogen specific proteins via the presence of reciprocal best BLAST hits in one or more of the 16 cereal pathogens at a BLAST bit score of greater than 200 and in the complete absence of a reciprocal best BLAST hit in any of the 11 non-cereal pathogens . Manual curation of putative cereal pathogen specific genes was conducted to account for the possibility that genes were predicted in F . pseudograminearum that were missed in the annotation of other genomes . This was performed by querying the intergenic sequences of the nearest non-cereal pathogen , F . oxysporum f . sp . lycopersici in a tBLASTn search using the putative cereal pathogen specific proteins . Through this process , 112 proteins were removed due to potential missed predictions in the F . oxysporum f . sp . lycopersici genome . Phylogenetic analyses were implemented using phylogeny . fr [57] . Briefly , multiple alignments were generated using MUSCLE [58] with default parameters , and curated using Gblocks [59] . Phylogeny was performed using PhyML [60] with the WAG amino acid substitution matrix [61] using an approximate likelihood-ratio test for branch support [62] . Trees were drawn using TreeDyn [63] . Trees were exported to Adobe Illustrator to allow shading of fungal branches . A slot blot membrane was prepared using a Hoefer PR 648 apparatus using Amersham Hybond-XL membrane ( GE Healthcare ) prewetted with 0 . 2 M NaOH . Approximately 500 ng of DNA was prepared in 0 . 2 M NaOH to a final volume of 60 µL and incubated at 37°C for 15 min prior to application to the membrane . Following application of the DNA the membrane was cross-linked using a GS GeneLinker UV cross-linker ( BioRad ) with 150 mJ of energy . Two identical membranes were prepared and probed with either a FpAH1 or rRNA probe . Probes were generated using the PCR DIG Probe synthesis kit ( Roche ) according to the manufacturer's instructions using primers FpAHDiversityF and FpAHDiversityR for the FpAH1 probe and ITS1-F and ITS4 primers for the rRNA probe ( Table S4 ) . Hybridization and washing was performed with the DIG Easy Hyb solution and DIG wash and block buffer set ( Roche ) . Detection was performed with the DIG Luminescent detection kit ( Roche ) . For Phaeosphaeria spp . PCR amplification was performed in 20 µL reactions containing 0 . 05 µM of PnAHDiversityF and FpAHDiversityR ( supplied by Microsynth ) , 1× Dream Taq Buffer ( Fermentas ) , 0 . 4 µM dNTPs ( Fermentas ) and 0 . 5 units of Dream Taq DNA polymerase ( Fermentas ) . For F . pseudograminearum isolates , the forward primer was replaced with FpAHDiversityF and PCR was conducted using Invitrogen Taq DNA polymerase as per the manufacturer's instructions . Primer sequences are listed in Table S4 . The PCR cycle parameters were: 2 min initial denaturation at 96°C followed by 35 cycles of 96°C for 30 s , 58°C for 45 s and 72°C for 1 min and a final 5 min extension at 72°C . The same cycling conditions were used for F . pseudograminearum isolates except the denaturation was performed at 94°C . Fusarium-derived PCR products were purified using the QIAgen PCR purification kit and sequencing was carried out by the Australian Genome Research Facility using both the forward and reverse primers . Sequencing reactions for Phaeosphaeria-derived products were conducted in 10 µL volume using the BigDye Terminator v3 . 1 Sequencing Standard Kit ( Applied Biosystems , Foster City , CA ) with both the forward and the reverse primer . The cycling parameters were 96°C for 2 min followed by 55 or 99 cycles of 96°C for 10 s , 50°C for 5 s and 60°C for 4 min . The products were cleaned with the illustra Sephadex G-50 fine DNA Grade column ( GE Healthcare ) according to the manufacturer's recommendations and then sequenced with a 3730×/Genetic Analyzer ( Applied Biosystems ) . Alignment of forward and reverse sequences for each isolate was performed in SeqScape software V2 . 5 ( Applied Biosystems ) . Translation and identification of protein haplotypes was also performed using this software . The program TCS v1 . 2 was used to visualize the most-parsimonious haplotype network [64] . The vectors for deletion of the FpAH1 and FpDLH1 genes were constructed using lambda phage mediated recombination as described previously [65] using primers listed in Table S4 . For FpAH1 the targeting construct consisted of 1858 bp of sequence immediately upstream of the start codon and 1132 bp covering the 3′ end of the gene and downstream sequence . 635 bp were deleted . For FpDLH1 the targeting construct consisted of 1523 and 1898 bp flanks , deleting 891 bp of the gene , leaving 125 bp of the 5′ coding sequence and 22 bp of the 3′ coding sequence . F . pseudograminearum protoplasting and transformation was performed as previously described for F . graminearum [66] except the STC ( sorbitol , Tris and calcium chloride solution ) was made with 0 . 8 M sorbitol instead of sucrose and the 40% PEG8000 solution also contained 0 . 8 M sorbitol . Transformants were selected on 50 mg L−1 geneticin ( Sigma , St . Louis , MO , USA ) , subcultured onto geneticin and DNA was prepared using the REDextract and Amp kit as per the manufacturer's instructions ( Sigma ) . Transformants were screened for gene deletion using AHKOscr1 , AHKOscr2 and gpdAr primers . In total , 42 transformants were screened and six were identified as carrying a deletion of FpAH1 . Mutant strains were single-spored and then stored as water cultures from ½ PDA plates . The PnAH1 deletion construct was made using overlap PCR as previously described [67] . PnAH1KO5'F and PnAH1KO5'R were used to amplify a 1456 bp region upstream of PnAH1 whilst PnAH1KO3'F and PnAHKO3'R amplified a 1519 bp region downstream . These flanks were fused using overlap PCR with a hygromycin resistance cassette amplified from pAN7 . 1 , resulting in a deletion construct of 5 . 9 kb . The deletion cassette was transformed into P . nodorum isolate SN15 protoplasts as previously described [68] and screened by using primers designed outside the flanking DNA ( PnAH1KOscr-F and PnAH1KOscr-R ) . Copy number of the transformed construct was determined as previously described [69] . The ΔFpAH1 mutant was complemented by co-transformation of ΔFpAH1 mutant 96T492 with pAN7 . 1 [70] and a construct designed to expresses FpAH1 under the control of the Aspergillus nidulans TrpC promoter . This construct was created by PCR amplification of the coding sequence and 341 bp of terminator of FpAH1 from genomic DNA of isolate CS3096 with forward and reverse primers ( AH-OXf and AH-OXr ) containing ClaI and EcoRI sites respectively and replacement of the hygromycin cassette in pUChph [71] using the same restriction sites . The TrpC-FpAH1 section of the construct was fully sequenced prior to transformation into ΔFpAH1 mutant 492 . Selection was with hygromycin at 200 mg L−1 . To verify that transformants were over-expressing the FpAH1 gene after successive rounds of sub-culture and single-sporing , they were grown in 6-well tissue culture plates in liquid basal media with 10 mM ( NH4 ) 2HPO4 as the nitrogen source [72] . Mycelia were harvested , freeze-dried and RNA extracted using TRIzol ( Invitrogen ) according to the manufacturer's instructions after four days of growth at 28°C in the dark . Relative FpAH1 gene expression was measured using quantitative reverse transcriptase PCR as previously described [73] . β-tubulin and FpAH1 ( FpAH1f and FpAH1r ) primers are shown in Table S4 . Expression of FpAH1 in the two transformants used was ∼3–5-fold higher than the wild type strain during in vitro culture and undetectable in the mutant as expected . A microtitre plate assay to monitor growth was performed based on the method described by Schmelz [74] . Briefly , each well contained 200 µL of basal media [72] with 5 mM glutamine as the nitrogen source with a final spore concentration of 1×104 spores mL−1 . Absorbance at 405 nm was measured using an iEMS microplate reader . For the FpAH1 mutant test the plate remained in the instrument for the duration of the assay at ambient temperature . In the case of the FpDLH1 mutant assay after an initial reading the plate was incubated at room temperature and readings commenced 22 hours post inoculation . Assessment of fungal virulence during FCR was carried out using the method described previously [36] . For virulence testing of FpAH1 mutant and complemented strains , isolates were inoculated onto 14 cm Synthetischer Nährstoffarmer Agar plates from potato dextrose agar plugs stored as water cultures and allowed to grow for 14 days under 12/12 hour day/night cycle under white and black fluorescent light at room temperature ( ∼22°C ) . Spores were harvested by flooding the plates with water and adjusted to the same concentration . To test the role of FpDLH1 in virulence , spores were produced in shaking 1 L flasks containing 100 mL of 25% Campbell's V8 juice inoculated with a single plug taken from a 20% V8 juice plate . Flasks were incubated at room temperature for 8 days and spores harvested by filtration through miracloth followed by centrifugation ( 5 , 500×g ) to remove media . Spores were resuspended in sterile water , counted and adjusted to the same concentration . Seeds were plated out two days prior to inoculation on wet paper towel in a 14 cm Petri dish and allowed to germinate on the laboratory bench . Germinated seeds were transferred to a 50 mL falcon tube containing 2 mL of the spore suspension and rolled gently to coat the seeds . Six to ten seeds were placed in a single paper towel , rolled up and tapped closed and placed in a jar containing water . Paper towel rolls were kept moist throughout the experiment . For the analysis of FpAH1 mutants a concentration of 1×105 spores mL−1 was used and plants were maintained on a laboratory bench without supplementary lighting . To test complemented FpAH1 mutants and FpDLH1 mutants , 4×105 spores mL−1 were used and plants were maintained in a closed bench top biohazard cabinet with a cool white fluorescent light suspended from the roof of the cabinet ( approximately 15 cm from the plants ) providing 16 hours of light per day . Assays were scored by recording the number of plants that were alive as a proportion of the total plants . Fungal strains were plated from PDA plug water storage onto 20% Campbell's V8 juice 1 . 2% agar plate and incubated for seven days at room temperature under white and black fluorescent light on a 12/12 hour day/night cycle . Seed for the assay were surface sterilized by soaking for 5 minutes in a 0 . 64% sodium hypochlorite-10% ethanol solution , followed by several rinses in sterile distilled water . Seed were plated onto three sheets of pre-wetted Whatmann no 3 12 . 5 cm filter disks in a 14 cm Petri dish . Seeds were incubated in the dark for 5 days at 4°C prior to germination at 20°C also in the dark . Germinated seedlings were distributed to Petri plates assembled with three filter papers wetted with 20 mL of sterile water prior to inoculation . Each plate contained 15–16 seedlings and was used for inoculation with one isolate . Inoculum consisted of agar plugs taken using either an inverted 1-mL pipette tip or a number 4 ( 6 mm ) cork borer from the edge of the growing V8 agar plate . Plugs were placed on a single root per seedling about 1 cm below the seed with the mycelia in direct contact with the root . Plates were sealed with sealing film ( PhytoTechnology Laboratories ) , incubated in a Thermoline illuminated incubator for 6 days at 20°C with 12 hours of light provided by fluorescent lamps . Assays were scored by measuring the length of the shoot . Parental and mutant strains of P . nodorum were tested on the susceptible wheat cultivar Amery as previously described [75] . Latent period sporulation was assessed using detached leaves as described previously [76] . Gene expression during infection of wheat and barley by F . pseudograminearum was performed using both a detached leaf infection assay and a root infection time course . For the latter , plants were inoculated as described for the root-rot virulence assay and root segments were harvested into liquid nitrogen at 24 , 48 and 96 hours post inoculation . For expression in detached leaves , fourth leaf segments ( 7–8 cm long ) of both hosts were taken from glasshouse-grown plants and each end of the cut leaf was sandwiched between water agar in a 14-cm Petri plate . Leaf segments were pierced at two points in a central region using a 200 µL pipette tip . A spore suspension of the wild type isolate ( 1×106 sp mL−1 ) was placed on the wound sites and the plate was sealed with sealing film . Plates were maintained in a fluorescent bulb lit growth chamber ( Thermoline ) at 20°C with 12 hours of light . The whole leaf segment was harvested for RNA extraction . RNA extraction was performed using TRIzol reagent ( Invitrogen ) according to the manufacturer's instructions . Relative expression was compared to β-tubulin . Primers used for gene expression analysis are shown in Table S4 . qRT-PCR was performed as previously described [73] . | Cereals are our most important staple crops and are subject to attack from a diverse range of fungal pathogens . A major goal of molecular plant pathology research is to understand how pathogens infect plants to allow the development of durable plant protection measures . Comparing the genomes of different pathogens of cereals and contrasting them to non-cereal pathogen genomes allows for the identification of genes important for pathogenicity toward these important crops . In this study , we sequenced the genome of the wheat and barley pathogen F . pseudograminearum responsible for crown and root-rot diseases , and compared it to those from a broad range of previously sequenced fungal genomes from cereal and non-cereal pathogens . These analyses revealed that the F . pseudograminearum genome contains a number of genes only found in fungi pathogenic on cereals . Some of these genes appear to have been horizontally acquired from other fungi and , in some cases , from plant associated bacteria . The functions of two of these genes were tested by creating strains that lacked the genes . Both genes had important roles in causing disease on cereals . This work has important implications for our understanding of pathogen specialization during the evolution of fungal pathogens infecting cereal crops . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"functional",
"genomics",
"plant",
"biology",
"genome",
"sequencing",
"plant",
"science",
"cereals",
"crops",
"plant",
"pathology",
"comparative",
"genomics",
"crop",
"diseases",
"biology",
"agriculture",
"plant",
"pathogens",
"genomics",
"genetics",
"and",
"genomics"
] | 2012 | Comparative Pathogenomics Reveals Horizontally Acquired Novel Virulence Genes in Fungi Infecting Cereal Hosts |
Parasitic helminths establish chronic infections in mammalian hosts . Helminth/Plasmodium co-infections occur frequently in endemic areas . However , it is unclear whether Plasmodium infections compromise anti-helminth immunity , contributing to the chronicity of infection . Immunity to Plasmodium or helminths requires divergent CD4+ T cell-driven responses , dominated by IFNγ or IL-4 , respectively . Recent literature has indicated that Th cells , including Th2 cells , have phenotypic plasticity with the ability to produce non-lineage associated cytokines . Whether such plasticity occurs during co-infection is unclear . In this study , we observed reduced anti-helminth Th2 cell responses and compromised anti-helminth immunity during Heligmosomoides polygyrus and Plasmodium chabaudi co-infection . Using newly established triple cytokine reporter mice ( Il4gfpIfngyfpIl17aFP635 ) , we demonstrated that Il4gfp+ Th2 cells purified from in vitro cultures or isolated ex vivo from helminth-infected mice up-regulated IFNγ following adoptive transfer into Rag1–/– mice infected with P . chabaudi . Functionally , Th2 cells that up-regulated IFNγ were transcriptionally re-wired and protected recipient mice from high parasitemia . Mechanistically , TCR stimulation and responsiveness to IL-12 and IFNγ , but not type I IFN , was required for optimal IFNγ production by Th2 cells . Finally , blockade of IL-12 and IFNγ during co-infection partially preserved anti-helminth Th2 responses . In summary , this study demonstrates that Th2 cells retain substantial plasticity with the ability to produce IFNγ during Plasmodium infection . Consequently , co-infection with Plasmodium spp . may contribute to the chronicity of helminth infection by reducing anti-helminth Th2 cells and converting them into IFNγ-secreting cells .
Infections with Plasmodium and helminths are extremely common , each contributing to substantial morbidity in affected populations [1–3] . Additionally , co-infections with Plasmodium species and intestinal helminths occur frequently in co-endemic areas [4 , 5] . The impact of co-infection on disease burden , pathogenesis , resistance to infection and immunity is complex and poorly understood . The vast majority of reported co-infection studies have focused on the impact of helminth infection on Plasmodium-associated responses , identifying altered anti-malarial immune responses or malaria-associated pathology during helminth co-infection [6–11] . However , the specific impact of Plasmodium infection on anti-helminth immunity has not been well characterized . Experimental murine models of helminth and Plasmodium co-infections have been established , however these have also mainly focused on how concomitant helminth infection affects Plasmodium immunity and pathology [11–16] , with much less focus on how Plasmodium infection impacts helminth-associated type 2 responses . Murine models of intestinal helminth infections have delineated a clear role for Th2-directed immune responses for proficient immunity . In particular , infection with the natural murine helminth , Heligmosomoides polygyrus , results in a chronic infection with the induction of a polarized type 2 response , characterized by IL-4-producing Th2 cells , alternative activation of macrophages and elevated IgE , closely mimicking human helminthiasis . Following anthelmintic treatment , Th2 cell-dependent immunity protects mice from re-infection ( reviewed in [17 , 18] ) . In contrast , acute blood-stage infection with the rodent malaria parasite , Plasmodium chabaudi chabaudi ( AS ) , results in polyclonal lymphocyte activation with a strongly polarized Th1 response [19] . Disease is associated with a spectrum of immunopathologies including splenomegaly and anemia [20–22] with peak parasitemia occurring 7–9 days post-infection [23] . These well-studied experimental systems , modeling human disease , provide appropriate tools to dissect the immune responses during co-infection . There is a large body of literature describing the antagonistic relationship between Th1 and Th2 cell differentiation . In vitro-based studies have clearly established that under Th1 and Th2 polarizing conditions , differentiated cells become more fixed in their phenotype with increasing rounds of cell division , losing their ability to convert to alternative phenotypes [24 , 25] . Mechanistically , T-bet and GATA-3 , transcription factors required to promote Th1 and Th2 differentiation , respectively , inhibit differentiation of the opposing phenotype [26 , 27] . Despite this clear antagonistic relationship , IL-4+IFNγ+ and T-bet+GATA-3+ Th cells are readily observed in vivo [28 , 29] , and several studies have established that Th subsets retain flexibility in their ability to produce non-lineage-specific cytokines [30–32] . Indeed , recent studies challenging the fate-lineage dogma demonstrated that antigen-restricted TCR transgenic Th2 cells co-produced IFNγ and IL-4 following LCMV infection [33 , 34] . In light of these new data , it is possible that Th cell conversion occurs during co-infection , altering immunity to one or both pathogens or contributing to the chronicity of helminth infection . In this study , we observed that Plasmodium and helminth co-infection led to a reduction of helminth-elicited Il4gfp+ Th2 cells and compromised anti-helminth immunity . We hypothesized that helminth-elicited Th2 cells were being converted into IFNγ-secreting Th1 cells during Plasmodium co-infection , as pressure to control both pathogens was placed on the Th cell population . To test this hypothesis , we generated triple cytokine reporter mice to accurately purify and identify Il4gfp , Ifngyfp and Il17aFP635-expressing cells to determine whether Th2 cells had the ability to change their phenotype . We observed that Il4-expressing Th2 cells could readily produce IFNγ following adoptive transfer in Rag–/–recipients , and these cells reduced severe parasitemia during acute P . chabaudi infection . Conversion of Th2 cells was dependent upon IL-12 and IFNγ-signaling , and blockade of these cytokines during co-infection preserved the Th2 response . Overall , this study provides fresh insight into the functional relationship between IFNγ- and IL-4-producing Th cells during co-infection and indicates that limiting acute Th1 responses may preserve Th2-mediated anti-helminth immunity .
To assess the impact of concomitant Plasmodium infection on the development of Th2 responses , we infected mice with H . polygyrus and 6 days later with 105 P . chabaudi-infected red blood cells ( Fig 1A ) . To accurately identify simultaneous transcription of Th1 ( Ifng ) , Th2 ( Il4 ) and Th17 ( Il17a ) lineage-defining genes , we generated a triple cytokine reporter mouse ( Il4gfpIfngyfpIl17aCreR26FP635 ) using existing and new fluorescent cytokine reporter mouse strains [35–37] ( S1 Fig ) . Following infection with L3 larvae of the intestinal helminth , H . polygyrus , we observed a significant expansion of Il4gfp+ CD4+ Th2 cells in the mesenteric lymph nodes 14 days post-infection . Co-infected mice had significantly reduced numbers of Il4gfp+ CD4+ Th2 cells in the mesenteric lymph nodes ( Fig 1B ) as well as a reduction in serum IgE ( Fig 1C ) and decreased expression of the alternative macrophage activation marker , Retnla ( Relmα ) in the gut ( S2 Fig ) . These data indicated that helminth-elicited Th2 cells and Th2-driven immune responses were compromised during Plasmodium co-infection . The reduced Il4gfp+ cells in the mesenteric lymph nodes correlated with an increase in Ifngyfp+ cells in the spleen during co-infection . Very few Il17aFP635+ cells were induced in this model ( S2 Fig ) . Following the resolution of acute malarial parasitemia , Th2 cell numbers in the mesenteric lymph nodes and serum IgE returned to levels observed in mice infected with H . polygyrus only ( S2 Fig ) . H . polygyrus establishes a chronic infection in wild type C57BL/6 mice . However , treating mice with anthelmintics kills adult parasites and allows a protective memory Th2 response to develop . Upon re-infection , mice expel worms in a CD4+ T cell- and IL-4-dependent manner [38 , 39] . Following the observation that P . chabaudi infection compromised Th2 cell responses ( Fig 1B ) , we tested whether P . chabaudi infection would impact Th2-dependent anti-helminth immunity . We infected wild type mice with H . polygyrus , treated mice with the anthelmintic , pyrantel pamoate , and then infected mice with P . chabaudi 7-days prior to re-infection with H . polygyrus ( Fig 1D ) . Although H . polygyrus-specific IgG1 levels were comparable between groups of mice ( S2 Fig ) , P . chabaudi-infected mice that had been given a secondary H . polygyrus challenge infection had significantly more adult worms in the intestinal lumen ( Fig 1E ) , indicating that Plasmodium infection compromised proficient anti-helminth immunity . It has become clear in recent years that lineage-committed CD4+ T cells retain a degree of plasticity , with the ability to convert between phenotypes [30] . Plasmodium infection elicits a polyclonal expansion of lymphocytes and IFNγ-secreting T cells [21 , 22] . We therefore hypothesized that the loss of Il4gfp+ Th2 cells in the mesenteric lymph nodes and the increase in Ifngyfp+ cells in the spleen during H . polygyrus and P . chabaudi co-infection was due to conversion of Th2 cells to an IFNγ-producing Th1-like phenotype . To test whether Th2 cells could produce IFNγ during P . chabaudi infection , we FACS-purified CD4+TCRβ+Il4gfp+Ifngyfp–Il17aFP365– Th2 cells from 2-week in vitro cultures ( S1 Fig ) , adoptively transferred them into Rag1–/–mice and infected the recipient mice with P . chabaudi . Cytokine expression in the transferred cells was analyzed in the spleen at day 8 post-infection ( Fig 2A ) . Transferred Th2 cells ( Il4gfp+Ifngyfp–Il17aFP365– ) almost completely lost expression of Il4gfp and , comparable to naïve T cells , expanded with approximately 80% of cells expressing Ifngyfp ( Fig 2B ) . Il17aFP635+ cells were barely detectable ( <1% ) following Plasmodium infection , in line with previous data [21 , 22 , 40] . IFNγ protein was also detectable in the serum of mice that received either naive CD4+ T cells or purified Th2 cells , but not in P . chabaudi-infected Rag1–/–mice that received no T cells , indicating that serum IFNγ was T cell-dependent ( Fig 2C ) . Thin blood smears from recipient mice identified that following infection of Rag1–/–mice , very high parasitemia is observed ( Fig 2D ) . The adoptive transfer of naïve T cells to Rag1–/–mice significantly reduced the high parasitemia , confirming an important T cell-dependent role in the control of high parasitemia during acute infection . This system permitted us to test whether Th2 cells , which had converted into IFNγ+ cells , could also control high parasitemia following acute infection . Indeed , adoptive transfer of Th2 cells also significantly reduced parasitemia ( Fig 2D ) , suggesting a functional loss of hemoglobin and severe anemia were also prevented in Rag1–/–mice given Th2 cells ( Fig 2E and 2F ) . Although Th2 cells up-regulated IFNγ in uninfected recipient Rag1–/–mice , significantly greater expansion of these converted cells occurred in P . chabaudi infected recipient mice ( S3 Fig ) . These data demonstrate that purified Il4-expressing Th2 cells were capable of producing IFNγ and could protect mice during acute P . chabaudi infection , similar to naive CD4+ T cells . Finally , to determine whether Th2 cells had the capacity to produce non-lineage cytokines in another model system , we infected Rag1–/–recipient mice with Candida albicans ( S4 Fig ) . At day 6 post C . albicans infection , transferred Il4gfp+ Th2 cells had lost Il4 expression and up-regulated IFNγ , similar to P . chabaudi infection . Interestingly , transferred Th2 cells did not up-regulate IL-17a , unlike naïve controls ( S4 Fig ) . We next asked whether Th2 cells that had down-regulated Il4gfp and expressed Ifngyfp retained the ability to re-express Th2-associated cytokines . We transferred Il4gfp+ Th2 cells into Rag1–/–mice and infected recipient mice with P . chabaudi , as in Fig 2A . At day 8 post-infection with P . chabaudi , we sorted CD4+TCRβ+ Ifngyfp+Il4gfp-Il17aFP635– cells from the spleens of recipient mice ( Fig 2G ) . Converted cells were then cultured in vitro with IL-4 and TCR stimulation . As expected , Ifngyfp+ cells that were previously either naïve or Il4gfp+ secreted IFNγ protein ( Fig 2H ) , validating the fidelity of the transcriptional reporter system . However , only Ifngyfp+ cells that were previously Il4gfp+ secreted the Th2-associated cytokines IL-13 and IL-5 ( Fig 2I ) , indicating that converted cells were indeed plastic , retaining the ability to produce Th2 cytokines . To identify the degree of transcriptional re-wiring of the converted cells in this model , we performed RNA sequencing on Th2 cells ( Il4gfp+ ) , converted Th2 cells ( Il4gfp+ → Ifngyfp+Il4gfp- ) , naïve CD4+ T cells , and Th1 cells ( naïve → Ifngyfp+Il4gfp– ) , using the same sorting strategy as in Fig 2G . Comparing the transcriptome of all significantly differentially regulated genes ( p<0 . 05 , >2-fold relative to naive T cells ) between the populations , we identified that converted cells had adopted a transcriptional profile very similar to Th1 cells ( Fig 3A and 3B , S1 Table ) with the majority of differentially regulated genes common with Th1 cells , while retaining some transcriptional similarity with their Th2 origin . Converted cells expressed Ifng , Tnf , Il2 and Il10 and largely lost expression of Il4 and Il6 , in comparison to the Th2 controls ( Fig 3C ) . Similarly , the transcriptional machinery in converted cells resembled Th1 cells with elevated Tbx21 ( Tbet ) and Eomes and low expression of Th2-associated transcription factors Gata3 and Nfil3 ( Fig 3D ) . To identify putative mechanistic pathways responsible for Th2 cell conversion , we used an upstream pathways algorithm to predict factors that may contribute to the observed transcriptional profile ( Ingenuity Pathways Analysis ) . This analysis identified canonical Th1 differentiation factors including IL-12 , IFNγ and type 1 IFN as potential upstream factors contributing to the observed transcriptional profile in converted cells ( Fig 3E ) . Furthermore , converted cells expressed Il12rb1 , Il12rb2 , Ifngr1 and Ifnar1 ( Fig 3F ) . In summary , converted Th2 cells had undergone significant re-wiring , closely resembling Th1 cells . When T cells undergo expansion in lymphopenic environments a population of rapidly dividing cells up-regulate CD44 and IFNγ [41–43] . To test whether conversion of Th2 cells into IFNγ-expressing cells could occur in a CD4+ T cell replete mouse , we transferred purified Th2 cells or naïve CD4+ T cells into OTII Rag1–/–mice [44] , which have CD4+ T cells specific only for OVA peptide . We infected recipient mice with P . chabaudi and analyzed donor and host cells at day 8 post-infection ( Fig 4A ) . Purified Th2 cells transferred into CD4+ OTII Rag1–/–mice , similar to Th2 cells transferred into Rag1–/–mice , produced IFNγ and down-regulated IL-4 ( Fig 4B and 4C ) , contributing to elevated levels of serum IFNγ ( Fig 4D ) . In contrast , host OVA-specific CD4+ T cells did not produce IFNγ following Plasmodium infection ( Fig 4C ) . Thus , Th2 cell conversion was not dependent on lymphopenia . Given that Th cells require both TCR stimulation and cytokine-mediated signaling for differentiation , it was conceivable that pre-activated Th2 cells in this system would only require a second cytokine receptor-mediated signal to up-regulate IFNγ , without the need for any additional TCR stimulation . We took two independent approaches to test whether TCR engagement was required for Th2 cells to produce IFNγ . First , we generated and FACS-purified TCR-restricted Th2 cells from OTII Rag1–/–mice crossed with Il4gfp reporter mice . We then transferred these OVA-specific Il4gfp+ Th2 cells into Rag1–/–recipients ( devoid of OVA ) and infected recipient mice with P . chabaudi ( Fig 5A ) . Unlike polyclonal Il4gfp+ Th2 cells that lost expression of Il4gfp and produced IFNγ , antigen-restricted OTII Il4gfp+ Th2 cells retained expression of Il4gfp and failed to produce IFNγ ( Fig 5B ) . Furthermore , IFNγ was not detectable in the serum of mice that received OVA-specific Il4gfp+ Th2 cells ( Fig 5C ) . Functionally , the failure to produce IFNγ correlated with significantly higher parasitemia , comparable to mice that received no T cells ( Fig 5D ) . These data indicate that TCR signaling was required for the functional conversion of Th2 cells into IFNγ-secreting cells . To verify the requirement of TCR-signaling for conversion , we transferred purified Il4gfp+Ifngyfp–Il17aFP365– Th2 cells into Rag1–/–recipient mice which were also deficient in MHC Class II and therefore unable to present antigens to Il4gfp+ Th2 cells . Recipient mice were infected with P . chabaudi , and transferred cells were analyzed at day 8 post-infection ( Fig 5E ) . As before , Il4gfp+ Th2 cells transferred into MHC Class II-sufficient Rag1–/–recipient mice down-regulated Il4gfp and up-regulated Ifngyfp . However , Il4gfp+ Th2 cells transferred to MHC Class II-deficient Rag1–/–recipient mice remained Il4gfp+ , did not express Ifngyfp ( Fig 5F ) and failed to reduce severe parasitemia ( Fig 5H ) . IFNγ was also undetectable in the serum ( Fig 5G ) . Taken together , these two experimental systems demonstrate that conversion of Th2 cells in this model requires TCR engagement . It has been shown previously that type I IFN signaling was required for IFNγ production from LCMV-specific TCR transgenic Th2 cells [34] . We had also observed that type 1 IFN was a candidate cytokine that could contribute to the transcriptional profile of converted Th2 cells ( Fig 3E ) . We therefore tested the requirement for type 1 IFN signaling by crossing Ifnar–/–mice with Il4gfp reporter mice . FACS purified Il4gfp+Ifnar–/–or Il4gfp+Ifnar+/+ Th2 cells were transferred to Rag1–/–recipient mice , subsequently infected with P . chabaudi and analyzed at day 8 post-infection ( Fig 6A ) . Both type I IFN responsive and unresponsive Th2 cells were capable of up-regulating IFNγ ( Fig 6B and 6C ) , contributing to serum IFNγ levels ( Fig 6D ) . Furthermore , type I IFN responsive and unresponsive Th2 cells afforded similar protection from high parasitemia ( Fig 6E ) , and prevented a loss in hemoglobin and red blood cells ( Fig 6F ) . Thus , type I IFN signaling was dispensable for IFNγ production from ex-Th2 cells and for controlling high parasitemia . From our RNA-Seq analysis we also identified that the canonical Th1 differentiating cytokines , IL-12 and IFNγ , may be responsible for the transcriptional profile observed in our converted cells ( Fig 3E ) . We first tested whether Th2 cells were responsive to IL-12 by measuring the phosphorylation of STAT4 following exposure to IL-12 . Supporting previous studies [45–47] , neither naïve CD4+ T cells nor sorted Il4gfp+ Th2 cells phosphorylated STAT4 in response to IL-12 ( Fig 7A and 7B; Pre- transfer ) . We then sorted transferred cells from naïve CD4+ T cell or Il4gfp+ Th2 cell recipient Rag1–/–mice 2 weeks post-transfer and found that both populations were responsive to IL-12 ( Fig 7A and 7B; Post-transfer ) . Thus , it was possible that IL-12 was promoting IFNγ expression in Th2 cells following P . chabaudi infection . We tested the role of IL-12 by transferring naïve or Il4gfp+ Th2 cells to Rag1–/–mice and blocking IL-12 prior to and after P . chabaudi infection ( Fig 7C ) . Blocking IL-12 reduced expression of Ifngyfp in naïve T cells ( reduced from 78 . 9% to 52 . 61% ) ; however , IL-12 blockade did not substantially alter the frequency of Ifngyfp+ cells derived from Th2 cells . Instead , IL-12 blockade maintained expression of Il4gfp+ in the Th2 population , with significantly larger Il4gfp+ and Il4gfp+Ifngyfp+ populations ( Fig 7D–7F ) . These data indicate that in this system IL-12 down-regulated Il4gfp expression , but was not required for IFNγ from Th2 cells . Furthermore , neutralization of IL-12 did not impact parasitemia ( Fig 7G ) . We next tested whether IFNγ , which contributes to Th1 differentiation [48] , was required for IFNγ expression by Th2 cells . To do this , we blocked IFNγ , IL-12 , or both IFNγ and IL-12 throughout the experiment ( Fig 8A ) . Blockade of IFNγ or IL-12 alone did not have a major impact on IFNγ production by Th2 cells ( Fig 8B ) . As above , IL-12 blockade preserved Il4gfp expression in a population of Th2 cells ( Fig 8C ) . However , blockade of both IFNγ and IL-12 led to a >50% reduction in IFNγ-expressing cells deriving from Th2 cells ( from 66 . 7%±1 . 5% IFNγ+ cells to 31 . 6%±3 . 4% IFNγ+ cells , Fig 8B ) , indicating that both IL-12 and IFNγ were required for optimal conversion of Th2 cells into IFNγ-secreting cells during Plasmodium infection . Despite a 50% reduction in IFNγ-secreting cells following IL-12 and IFNγ blockade , the remaining ~30% of IFNγ+ cells were sufficient to prevent high parasitemia ( S5 Fig ) . Finally , we translated these new observations back into a co-infection scenario , as presented in Fig 1 , and tested whether helminth-induced Th2 cells had the capacity to up-regulate IFNγ in a co-infection scenario . First , we purified ex vivo Il4gfp+Ifngyfp–Il17aFP635– Th2 cells from d14 H . polygyrus-infected mice and transferred them into day 14 H . polygyrus-infected Rag1–/–mice . Recipient mice were then co-infected with P . chabaudi and the transferred cells were analyzed at day 8 post P . chabaudi infection ( Fig 9A ) . Similar to in vitro-derived Th2 cells , H . polygyrus-derived Th2 cells down-regulated Il4gfp and up-regulated Ifngyfp , albeit to a slightly lesser extent than naïve T cells ( Fig 9B ) . Re-stimulation of lymph node cells with H . polygyrus antigen and IL-4 led to the secretion of IL-5 and IL-13 from mice given H . polygyrus Th2 cells , but not from mice given naïve T cells ( Fig 9C ) . These data suggested that despite a high degree of conversion to IFNγ-secreting cells , cells retained antigen-associated cytokine secretion . To more accurately determine whether converted cells retained the capacity to produce Th2 cytokines in an antigen-specific manner , we sorted Th2 cells , or naïve cells , that had converted into Ifngyfp+ cells from recipient mice and restimulated them in vitro with H . polygyrus antigen or P . chabaudi infected red blood cells ( iRBC ) . Ifngyfp+ cells , which were previously naïve or Il4gfp+ Th2 cells , produced IFNγ when co-cultured with irradiated APCs , supporting the cytokine reporter expression ( Fig 9D ) . iRBCs further stimulated more IFNγ from naive T cells , but not from Th2 cells , suggesting that either ex vivo Th2 cells were not responding to malarial antigens , or that they were already secreting IFNγ at capacity . In addition , ex vivo H . polygyrus elicited Th2 cells which had down-regulated Il4gfp and up-regulated Ifngyfp produced IL-5 in response to H . polygyrus antigen , suggesting that converted cells retained antigen specificity and plasticity in this model ( Fig 9D ) . Finally , we tested whether the factors promoting IFNγ in the adoptive transfer model , IL-12 and IFNγ ( Fig 8B ) , were responsible for the loss of Th2 cells and type-2 immunity during H . polygyrus and P . chabaudi co-infection . To do this , we infected wild type mice with H . polygyrus and at six days post-infection , mice were co-infected with P . chabaudi with or without blocking antibodies to IL-12 and IFNγ ( Fig 10A ) . Blockade of IL-12 and IFNγ preserved Il4gfp+ Th2 cells in co-infected mice ( Fig 10B ) and maintained elevated levels of helminth-induced type-2-associated IgE ( Fig 10C ) . However , despite preserving Th2 cells and IgE , proficient anti-helminth immunity was not fully restored in mice given blocking antibodies ( S6 Fig ) . Thus , IL-12 and IFNγ play a major role compromising Th2 responses during helminth/ Plasmodium co-infection , but additional factors also contribute to compromised anti-helminth immunity during co-infection .
In this study , we identified that Plasmodium infection significantly reduced CD4+ Th2 cells during co-infection with H . polygyrus and that anti-helminth immunity was compromised during co-infection . Mechanistically , we found that Il4gfp+Ifngyfp–Il17aFP635– Th2 cells , purified from novel triple cytokine reporter mice , converted to IFNγ-secreting cells , contributing significantly to anti-Plasmodium immunity . IFNγ production by Th2 cells was dependent on TCR , IL-12 , and IFNγ signaling , all of which contributed to the transcriptional re-programming of Th2 cells . Finally , we found that blockade of IL-12 and IFNγ during Plasmodium and helminth co-infection preserved Th2 responses and IgE production , but was insufficient to fully restore anti-helminth immunity . There is a large body of literature describing the prevalence of helminth and Plasmodium co-infection in human populations [4 , 5 , 8 , 11 , 49 , 50] , and mouse models [16 , 51] , with the majority of studies focusing on the impact of helminth infections on anti-Plasmodium responses . Relatively few have focused on how parasite-elicited Th2 responses are affected during Plasmodium co-infection . Our data show that IL-4-expressing Th2 cells , serum IgE , and functional parasite expulsion are reduced during co-infection ( Fig 1 ) . This is in line with previous reports , including reduced schistosome-specific IL-4 and IL-5 in Plasmodium and schistosome co-infected individuals [52] and suppressed IL-4 responses during H . polygyrus and Plasmodium yeolii co-infection [53] . Reduced type-2 responses [54] and Th2-mediated immunopathology have also been observed in schistosome and Plasmodium co-infected mice [55] , consistent with the notion that anti-helminth associated Th2 responses are compromised during Plasmodium co-infection . However , these studies did not offer mechanistic insight as to how this reduction in type-2 immunity might occur and importantly how type-2 immunity might be preserved during co-infection . In this study , we focused on the impact of co-infection on CD4+ T cells , which are a critical cell type for immunity to H . polygyrus and contribute significantly to anti-malarial immunity [56] . For our studies , we developed a triple cytokine reporter mouse ( Il4gfpIfngyfpIl17aFP635 , S1 Fig ) , which had several important advantages . These mice allowed the determination of T cell phenotype ex vivo without the need for re-stimulation , as well as the ability to obtain highly purified populations of Il4gfp+Ifngyfp–Il17aFP635– Th2 cells , which were not expressing other lineage-associated cytokines[29] . Adoptive transfer of these cells allowed us to accurately determine whether purified Th2 cells changed their phenotype , and finally , simultaneous cytokine reporters allowed us to test whether any conversion was reversible and truly plastic . To this end , we observed that highly-purified Il4gfp+Ifngyfp–Il17aFP635– Th2 cells , either generated in vitro for two weeks ( Fig 2 ) or isolated ex vivo from H . polygyrus-infected mice ( Fig 9 ) , were able to produce IFNγ during Plasmodium infection in Rag1–/–mice . This phenomenon is in line with several previous observations 1 ) identifying that in vitro generated LCMV-specific TCR transgenic Th2 cells could express both IFNγ and IL-4 [34] , 2 ) a ‘bi-functional’ population of Tbet+ GATA3+ cells are generated following H . polygyrus infection [29] and 3 ) the Tbx21 locus ( encoding T-bet ) has bivalent epigenetic histone modifications in Th2 cells [57] suggesting Th2 cells retain some flexibility . We observed expression of Ifng , Tbx21 , Klrg1 , Gzmb , Gzmc in converted Th2 cells , while maintaining low levels of Il4 transcription ( Fig 3 , S1 Table ) and the ability to produce IL-5 and IL-13 ( Fig 2 ) . This suggested that converted cells were possibly poly-functional . Whether they are similar to ‘bi-functional’ cells [29] is unclear . Helmby observed exacerbated liver pathology with significantly increased IFNγ and mortality during H . polygyrus and Plasmodium co-infection [58] . Whether Th2 cells converted to IFNγ-secreting cells , contributing to aggravated liver pathology in their study was unclear . Similarly , Th2 cells that up-regulate IL-17 during airway allergen challenge in mice contribute to more severe airway pathology [59] , and allergic patients have a greater frequency of IFNγ-secreting cells [60] . Indeed , polyfunctional T cells , which secrete multiple cytokines , correlate with greater protection following vaccination [61] , contribute to severe inflammatory syndromes in humans [62] and mice [37] and have greater anti-tumor activity [63] . Thus , understanding the mechanisms of Th cell conversion and the generation of polyfunctional T cells may provide important insight into immunity and immunopathology . Interestingly , in our model of C . albicans , in vitro polarized Th2 cells were unable to produce IL-17a , unlike naïve cells ( S4 Fig ) , suggesting that there is either an important relationship between Th2 and Th1 cells , or that the transcriptional machinery required for IL-17 production is more tightly regulated than for IFNγ . To identify mechanistic pathways contributing to Th2 cell conversion , we employed RNA-Seq analysis of Th1 cells ( Ifngyfp+ ) , Th2 cells ( Il4gfp+ ) and Th2 cells that had up-regulated IFNγ ( Ifngyfp+Il4gfp– ) . We identified a high degree of transcriptional similarity between Th1 cells and converted cells , extending significantly beyond cytokine expression . For example , Th1 and converted Ifngyfp+Il4gfp–cells , but not Th2 cells , had similar transcript abundance encoding for several enzymes ( Bace2 , Cdc25c , Cd38 Chst11 , Dusp5 , Gzmb , Gzmc and Gzmk , Gstt1 , Pdcd1 , Ptpn5 , Spag5 , Troap ) , chemokine receptors ( Cxcr3 , Cmklr1 , Cx3cr1 and Ccr5 ) , ion channels ( Cacna1l and Ttyh2 ) , kinases ( Stk32c , Ttbk1 , Ttk , Ltk , Cdk1 , Pbk , Ccnb1 in addition to many other kinases ) , nuclear receptors ( Nr4a2 , Ahr ) , miRNAs ( miR-142 , miR-155 and miR-Let7d ) and transcriptional regulators ( Rai14 , E2f7 , Gas7 , Cdkn2b , E2f8 , Klf12 , Runx2 and Eomes ) . Significantly , Th1 cells use a feed-forward regulatory circuit involving Tbx21 ( Tbet ) and Runx3 for maximal IFNγ production and silencing of Il4 [64] . In our study , both Th1 cells and converted Th2 cells which had lost Il4 and up-regulated Ifng , had elevated Runx3 and Tbet , suggesting that this feed-forward loop was transcriptionally active , supporting optimal IFNγ production in converted cells . Whether the epigenetic landscape of converted cells matched that of their Th1 counterparts is of great interest , as converted Th2 cells retained the capacity to produce Th2-associated IL-5 and IL-13 ( Fig 2 ) in an antigen-specific manner ( Fig 9D ) . Previous studies have indicated that Th1 cells have the capacity to up-regulate Th2-associated features in vivo following helminth infection [65] . In our hands , naïve T cells which had up-regulated IFNγ+ in vivo following Plasmodium infection did not have the capacity to secrete IL-5 or IL-13 when re-stimulated in vitro with anti-CD3/28 and IL-4 . Whether there are specific in vivo factors which more readily support T cell plasticity is currently unclear . We would hypothesize that in vitro generated or ex vivo H . polygyrus Th2 cells had bivalent methylation marks in the Il5 and Il13 locus allowing re-expression of these genes following the appropriate activating signal . Supporting this , converted Th2 cells retained some Th2-associated features , including elevated expression of Gfi1 , Il4 and Il33r , which may provide the appropriate machinery to re-activate Th2-associated genes , reminiscent of their Th2 past ( Fig 3 and S1 Table ) . Using an upstream analysis algorithm ( Ingenuity Pathways Analysis ) with our transcriptional data sets we identified IL-12 , IFNγ and to a lesser extent type 1 IFN , as putative factors that could contribute to the observed transcriptional profile of converted cells . This supports a recent study that identified the requirement of Tbet and Stat4 for IFNγ expression in memory Th2 cells [66] . In our study , unlike previous studies , type I IFN signalling in Th2 cells was dispensable for IFNγ production from converted Th2 cells in vivo ( Fig 6 ) [34] . Blocking IL-12 or IFNγ alone did not impact the frequency of converted IFNγ+ cells from transferred Th2 cells ( Figs 7 and 8 ) . These data are in agreement with a previous study that found restoring IL-12 responsiveness in Th2 cells , through ectopic expression of IL-12Rβ2 , was insufficient to convert Th2 cells into IFNγ-secreting cells [67] . However , in our model , anti-IL-12 treatment alone preserved IL-4 expression in a sub-population of transferred cells ( Figs 7 and 8 ) . Blockade of both IFNγ and IL-12 substantially reduced IFNγ+ cells deriving from Th2 cells , suggesting that an IL-12-STAT4 signaling pathway down-regulated IL-4 , while an IFNγ / STAT-1 / T-bet pathway was required for optimal IFNγ expression , in accordance with canonical Th1-inducing conditions for naive T cells [68] . While we found that blockade of these cytokines reduced IFNγ+ cells , there was no change in control of parasitemia ( S5 Fig ) . We speculate that this is due to the incomplete loss of conversion , with the remaining IFNγ being sufficient to control levels of parasitemia . TCR stimulation was essential for in vitro-derived Th2 cells to produce IFNγ ( Fig 5 ) and ex vivo H . polygyrus-elicited Th2 cells required H . polygyrus-infected recipient mice to survive and up-regulate IFNγ . Thus , with sufficient TCR signaling , a change in the local cytokine milieu may be sufficient to re-program Th cells . During helminth and Plasmodium co-infection , either cross-reactive antigens or microflora-derived signals may provide the necessary first TCR signal [69–71] . Alternatively the broad polyclonal activation of non-specific T cells during Plasmodium infection may be sufficient [21 , 22 , 72] . Although TCR engagement , IL-12 and IFNγ were required for optimal conversion of Th2 cells into IFNγ-secreting cells , it is possible that other factors also contribute to conversion , including IL-27 , which can induce expression of Tbet , and IL-18 , which can induce IFNγ production [73 , 74] . In conclusion , we have shown that IL-12 and IFNγ suppressed Th2 responses during H . polygyrus and P . chabaudi co-infection . Mechanistically , we identified that TCR engagement with IL-12 and IFNγ signaling converted in vitro-generated Th2 cells into IFNγ-producing cells during P . chabaudi infection . Importantly , although blocking IL-12 and IFNγ during co-infection did not retain fulminant anti-helminth immunity , it did preserve Th2 cell numbers and serum IgE , highlighting a novel mechanistic pathway of how Plasmodium infection negatively impacts anti-helminth Th2 responses . Overall , our studies indicate that Plasmodium infection can negatively impact anti-helminth responses , that Th2 cells retain substantial plasticity in the context of Plasmodium infection , and that this plasticity may play a role in the reduced Th2 response during co-infection .
All mice were bred and maintained under specific pathogen-free conditions at the National Institute for Medical Research . Strains used included: C57BL/6 , Ifngyfp [36] , Il4gfp[35] , C57BL/6 Rag1–/–[75] , MhcII–/– ( B6 . 129-H-2<dlAb1-Ea ) [76] crossed with Rag1–/–at NIMR [77] , OTII Rag1–/– ( B6 . Cg ( Tcrαβ ) 425Cbn/J ) [78] , OTII Il4gfp Rag1–/– ( OTII Rag1–/–crossed with Il4gfp at NIMR ) , and Ifnar–/–Il4gfp ( Ifnar–/–[79] crossed with Il4gfp at NIMR ) . Triple cytokine reporter mice ( Il4gfpIl17CreIfngyfpR26FP635 ) were established by crossing Il4gfp/gfpIl17Cre/Cre[37] mice with Ifngyfp/+R26FP635/FP635 mice , producing Il4gfp/+Il17Cre/+Ifngyfp/+R26FP635/+ . The generation of R26FP635 reporter mice will be presented in detail elsewhere ( JB and AP , manuscript in preparation ) . Briefly , R26FP635 mice were generated by inserting the coding sequences of the red fluorescent protein FP635 [80] into the pROSA26 targeting vector downstream of a loxP-flanked neomycin resistance cassette containing three transcriptional stop signals by homologous recombination . R26FP635 reporter mice in this study were backcrossed to C57BL/6 for more than 8 generations . Mice were infected by oral gavage with 200 infective stage 3 ( L3 ) Heligmosomoides polygyrus larvae , diluted in water . The anthelmintic drug pyrantel pamoate ( Sigma , 5mg/dose in water ) was given orally on two consecutive days . Infections with Plasmodium chabaudi chabaudi ( AS ) were performed by i . p . injection of 105 parasitized red blood cells . Parasitemia was measured by blinded counting of Giemsa-stained blood smears . Anemia and hemoglobin were measured by diluting blood in Krebs buffered saline with 0 . 2% glucose and with 100 IU/mL heparin and measured using Vetscan ( Abaxis-VetScan HM5 Hematology ) . Infections with Candida albicans were performed by i . v . injection of 105 yeast forms . Cell sorting was performed using a FACS Aria II ( BD Biosciences ) , MoFlo XDP ( Beckman Coulter ) , or Influx ( BD Biosciences ) cell sorter . To prepare cells for sorting , CD4+ cells were first positively selected using MACS CD4 beads and magnetic columns ( Miltenyi Biotec ) . Cell suspensions were then stained for 25 minutes with antibodies in PBS with 1% FCS . To prepare for sorting , stained cells were diluted in phenol-red free IMDM ( Gibco ) ( with 1% FCS , 2mM EDTA ( Invitrogen ) , 100 U/mL Penicillin and 100 μg/mL Streptomycin ( Gibco ) , 8 mM L-glutamine ( Gibco ) , and 0 . 05 mM 2-mercaptoethanol ( Gibco ) ) . Propidium iodide ( PI ) was used to determine cell viability in sorting experiments . Intracellular cytokine staining ( ICS ) was performed following 6 hours of re-stimulation with 50ng/mL phorbol 12-myristate 13-acetate ( PMA , Promega ) and 1 μg/mL ionomycin ( Sigma ) and BD Golgi Stop and BD Golgi Plug ( diluted 1:1000 , BD Biosciences ) . Following surface stain , cells were incubated with eBioscience Fixation/Permeabilization buffer for 25 minutes followed by 25 minutes in Permeabilization buffer ( eBioscience ) , and incubation with antibodies in Permeabilization buffer for a further 30 minutes . For flow cytometry analysis , cells were analyzed using a BD LSRII ( BD Biosciences ) and data were analyzed using FlowJo software ( Version 7 . 6 . 5 , Treestar Inc ) . In all cases using triple cytokine reporter mice , cells from wild type , Ifngyfp or Il4gfp single cytokine reporter mice were used as controls to set gates to differentiate yfp and gfp . Antibodies used include: CD4 ( efluor450 and PE-Cy7 , RM4-5 , eBioscience ) , CD25 ( Fitc , 7D4 , BD Pharmingen ) , CD44 ( Fitc , Percpcy5 . 5 , and APC , IM7 , eBioscience ) , CD45 . 1 ( PE-Cy7 and APC , A20 , eBioscience ) , IFNγ ( Pacific Blue , XMG1 . 2 , Biolegend ) , IL4 ( PE , 11B11 , eBioscience ) , pSTAT4 ( Alexa Fluor 647 , BDPhosflow ) , TCRβ ( APC , H57-597 , eBioscience ) and GFP ( Alexafluor647 , FM264G , BioLegend ) . Staining was performed in presence of FcR Blocking Reagent ( Miltenyi Biotec ) . In analysis experiments , viability was determined using the Molecular Probes Live/Dead Fixable Blue Dead Cell Stain Kit ( Life Technologies ) . For phospho-STAT staining , sorted cells were resuspended into serum-free media and incubated at 37 degrees for 20 minutes , followed by incubation with 10 ng/mL IL-12 ( R&D ) for 15 minutes . Cells were then fixed for 10 minutes at 37 degrees with prewarmed BD Phosflow Lyse/Fix Buffer , washed , permeabilized with BD Phosflow Perm Buffer III for 30 minutes on ice , washed , and stained for 1 hour with antibodies in PBS for FACS analysis . Naive CD4+ T cells were sorted from spleens as CD4+TCRβ+CD44–CD25–Il4gfp–PI− ( Il4gfp reporter ) or CD4+TCRβ+CD44–CD25–Il4gfp-Ifngyfp–Il17aFP635–PI− ( triple reporter ) . Th2 cells were cultured for 2 weeks from splenic CD4+ cells in vitro with 10 ng/mL IL-4 ( R&D ) , 5 ng/mL IL-2 ( R&D ) , 10 μg/mL anti-IFNγ ( XMG1 . 2 , BioXcell ) , and Mouse T-Activator CD3/CD28 Dynabeads ( Life Technologies ) in IMDM with 10% FCS . Th2 cells were sorted as CD4+TCRβ+Il4gfp+PI− ( Il4gfp reporter ) or CD4+TCRβ+Il4gfp+Ifngyfp–Il17aFP635–PI− ( triple reporter ) . For each experiment , 0 . 2x106 to 1x106 cells were adoptively transferred i . v . into recipient C57BL/6 Rag1–/- mice . Blocking antibodies diluted in PBS ( anti-IFNγ , XMG1 . 2 , anti-IL12p40 C17 . 8 , BioXcell ) were used at 0 . 4 or 0 . 5 mg/ dose . Sorted cells were cultured in 96 well round bottom plates in various conditions . Where indicated , antigen presenting cells were spleens depleted of CD4+ cells by MACS magnetic separation ( Miltenyi Biotec ) and irradiated ( 3000 rads ) . H . polygyrus antigen was isolated by homogenization of cleaned adult worms in PBS . IFNγ , IL-5 , and IL-13 were measured using DuoSet ELISA kits , according to the manufacturer’s instructions ( R&D ) . Total IgE ELISA was performed by coating with Purified Rat Anti-Mouse IgE ( R35-72 , BD Pharmingen ) at 2 μg/mL overnight , followed by overnight incubation with serum and standard ( Purified Mouse IgE , k isotype Standard , BD Pharmingen ) , and detection with Biotin Rat Anti-Mouse IgE at 1 μg/mL ( R35-118 , BD Pharmingen ) , Streptavidin HRP at 1:000 ( BD Pharmingen ) and ABTS One Component HRP Microwell Substrate ( SurModics ) . H . polygyrus-specific IgG1 was detected by coating plates with 5 μg/mL H . polygyrus antigen overnight , followed by overnight incubation with serially diluted serum and detection with Biotin Rat Anti-Mouse IgG1 ( Invitrogen ) and streptavidin and ABTS , as above . RNA was isolated from cells or tissue using RNeasy Mini Kit according to manufacturer’s instructions ( Qiagen ) . For qRT-PCR of small intestine-derived RNA , 1 cm sections of tissue were harvested and stored in RNAlater ( Sigma ) before homogenisation and RNA extraction using RNeasy Mini Kit ( Qiagen ) . cDNA was reverse transcribed from RNA using QuantiTect Reverse Transcription Kit ( Qiagen ) according to the manufacturer’s instructions . qRT-PCR analysis was performed using Power SYBR Green PCR master mix ( Applied Biosystems ) on an ABI Prism 7900HT Sequence Detection System ( Applied Biosystems ) . Relative quantities of mRNA were determined by the comparative threshold cycle method as described by Applied Biosystems for the ABI Prism 7700/7900HT Sequence Detection Systems using the following primers; Hprt Fwd: 5’-GCCCTTGACTATAATGAGTACTTCAGG-3’ and Rvs: 5’-TTCAACTTGCGCTCATCTTAGG-3’; Retnla Fwd: 5’-CCCTCCACTGTAACGAAGACTC-3’ and Rvs: 5’-CACACCCAGTAGCAGTCATCC-3’; Chil3: Fwd: 5’- CATGAGCAAGACTTGCGTGAC-3’ and Rvs: 5’-GGTCCAAACTTCCATCCTCCA-3’; Arg1 Fwd: 5’- GGAAAGCCAATGAAGAGCTG -3’ and Rvs: 5’- GCTTCCAACTGCCAGACTGT -3’ . RNA-seq libraries were constructed using the TruSeq RNA Sample Preparation Kit V2 according to manufacturer’s instructions ( Illumina ) . Libraries were sequenced using the HiSeq 2500 System ( Illumina ) . The raw Illumina reads were analyzed as follows . First , the data quality was analyzed using FastQC ( www . bioinformatics . babraham . ac . uk/projects/fastqc ) . Low quality bases were trimmed using Trimmomatic [81] , and the read pairs which passed the trimming quality filters were aligned to mm10 ( Ensembl version 75 ) using Tophat2 [82] . Counts were determined using htseq_count [83] . Normalisation and statistical analysis was performed using edgeR [84] . Statistically significant genes with FDR < 0 . 05 are reported . Significantly differentially expressed genes were uploaded into Ingenuity Pathways Analysis ( IPA ) and subjected to upstream analysis to identify factors that could have contributed to the transcriptional profile observed in converted Th2 cells . Data sets were compared by Mann Whitney test using GraphPad Prism ( V . 5 . 0 ) . Differences were considered significant at *P ≤ 0 . 05 . All animal experiments were carried out following United Kingdom Home Office regulations ( project license 80/2506 ) and were approved by UK National Institute for Medical Research Ethical Review Panel . | Approximately a third of the world’s population is burdened with chronic intestinal parasitic helminth infections , causing significant morbidities . Identifying the factors that contribute to the chronicity of infection is therefore essential . Co-infection with other pathogens , which is extremely common in helminth endemic areas , may contribute to the chronicity of helminth infections . In this study , we used a mouse model to test whether the immune responses to an intestinal helminth were impaired following malaria co-infection . These two pathogens induce very different immune responses , which , until recently , were thought to be opposing and non-interchangeable . This study identified that the immune cells required for anti-helminth responses are capable of changing their phenotype and providing protection against malaria . By identifying and blocking the factors that drive this change in phenotype , we can preserve anti-helminth immune responses during co-infection . Our studies provide fresh insight into how immune responses are altered during helminth and malaria co-infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | IFNγ and IL-12 Restrict Th2 Responses during Helminth/Plasmodium Co-Infection and Promote IFNγ from Th2 Cells |
The nematode C . elegans is an important model for the study of social behaviors . Recent investigations have shown that a family of small molecule signals , the ascarosides , controls population density sensing and mating behavior . However , despite extensive studies of C . elegans aggregation behaviors , no intraspecific signals promoting attraction or aggregation of wild-type hermaphrodites have been identified . Using comparative metabolomics , we show that the known ascarosides are accompanied by a series of derivatives featuring a tryptophan-derived indole moiety . Behavioral assays demonstrate that these indole ascarosides serve as potent intraspecific attraction and aggregation signals for hermaphrodites , in contrast to ascarosides lacking the indole group , which are repulsive . Hermaphrodite attraction to indole ascarosides depends on the ASK amphid sensory neurons . Downstream of the ASK sensory neuron , the interneuron AIA is required for mediating attraction to indole ascarosides instead of the RMG interneurons , which previous studies have shown to integrate attraction and aggregation signals from ASK and other sensory neurons . The role of the RMG interneuron in mediating aggregation and attraction is thought to depend on the neuropeptide Y-like receptor NPR-1 , because solitary and social C . elegans strains are distinguished by different npr-1 variants . We show that indole ascarosides promote attraction and aggregation in both solitary and social C . elegans strains . The identification of indole ascarosides as aggregation signals reveals unexpected complexity of social signaling in C . elegans , which appears to be based on a modular library of ascarosides integrating building blocks derived from lipid β-oxidation and amino-acid metabolism . Variation of modules results in strongly altered signaling content , as addition of a tryptophan-derived indole unit to repellent ascarosides produces strongly attractive indole ascarosides . Our findings show that the library of ascarosides represents a highly developed chemical language integrating different neurophysiological pathways to mediate social communication in C . elegans .
Communication among individuals of a species relies on a number of different sensory inputs including chemical , mechanical , auditory , or visual cues [1] . Chemical signaling is perhaps the most ancient form of interorganismal communication [1] , [2] , and analysis of the chemical signals and the behaviors they mediate is of great significance for understanding the ecological and evolutionary dynamics of intra- and inter-specific interactions . The free-living nematode C . elegans is used extensively as a model system for social behaviors such as foraging , population density sensing , mating , and aggregation ( http://www . wormbook . org; [3] ) . Recent investigations have shown that a family of small molecules , the ascarosides , play important roles as chemical signals regulating several different aspects of C . elegans behavior ( Figure 1A ) [4]–[8] . The ascarosides ascr#1 , ascr#2 , and ascr#3 were originally identified as major components of the dauer pheromone , a population-density signal that promotes entry into an alternate larval stage , the non-feeding and highly persistent dauer stage [4]–[7] . Additional work showed that at concentrations far below those required for dauer formation , synergistic mixtures of ascarosides act as strong male-specific attractants , and that male attraction to ascarosides requires the amphid sensory neurons ASK and the cephalic sensory neurons CEM [6] , [7] . Wild-type ( N2 ) hermaphrodites do not respond to low concentrations of ascarosides and show repulsion at dauer-inducing concentrations [4] . However , a recent study showed that mutation of the neuropeptide-Y receptor homolog NPR-1 strongly affects hermaphrodite response to ascarosides [9] . The strong loss-of-function mutant npr-1 ( ad609 ) showed attraction or reduced repulsion to specific combinations of ascarosides , in contrast to wild type ( N2 ) worms that express a high-activity variant of NPR-1 [10] , [11] . The interneuron RMG , the central site of action of NPR-1 , is proposed to serve as a central hub computing aggregation and attraction signals originating from several different sensory neurons , including the ascaroside-sensing ASK neurons [9] . These findings suggested that mutual attraction and aggregation in C . elegans are mediated primarily by signaling via NPR-1 , and that strains carrying the high-activity form of NPR-1 including wild-type ( N2 ) hermaphrodites may not rely on small molecule signaling to promote aggregation . Nonetheless , wild-type ( N2 ) hermaphrodites also display aggregation behaviors , for example , in response to environmental cues such as limited food availability [12] or perturbations of transforming growth factor–β ( TGF–β ) signaling [13]–[15] . Given the existence of small molecules that serve as social cues for population density sensing and mate-finding , and the complicated neural circuitry implicated in aggregation behavior , we hypothesized that structurally distinct small molecules might exist that serve as aggregation signals in C . elegans . Here we show that C . elegans aggregation behavior is regulated by a dedicated set of highly potent signaling molecules , the indole ascarosides , which form part of a modular chemical language that elicits structure-specific behaviors via several distinct neurophysiological pathways . Our findings provide evidence for multi-layered social signaling in C . elegans .
All currently known small-molecule pheromones in C . elegans are derived from peroxisomal β-oxidation of long-chained fatty acids via DAF-22 , a protein with strong homology to human sterol carrier protein SCPx [6] , [16] . We hypothesized that putative aggregation pheromones may be derived from the same pathway , suggesting that daf-22 mutants would not produce them . In this case , a spectroscopic comparison of the wild-type metabolome with that obtained from daf-22 mutant worms should reveal candidate compounds for attraction or aggregation signals . In a previous study , we had used an NMR spectroscopy-based technique termed Differential Analysis of NMR spectra ( “DANS” ) to compare the wild-type metabolome with that of daf-22 mutant worms [6] . This comparison had led to identification of ascr#6–8 , of which ascr#8 is a major component of the male-attracting signal [6] . Based on NMR spectra with improved signal-to-noise ratio , we conducted a more detailed comparison of wild type and daf-22-mutant metabolomes , which revealed several indole-containing compounds in the wild-type metabolome that were not produced by daf-22 worms ( Figure 1B , C ) . The established role of DAF-22 in pheromone biosynthesis [6] , [16] , [17] suggested that these indole derivatives may represent a previously unrecognized family of signaling molecules . To clarify the structures and biological roles of the daf-22-dependent indole derivatives , we pursued their complete identification via NMR spectroscopy-guided fractionation of the wild-type metabolome . Reverse-phase chromatography produced eight metabolite fractions , which were analyzed by two-dimensional NMR spectroscopy . The NMR spectra revealed the presence of daf-22-dependent indole-derivatives in two fractions , which were selected for additional NMR-spectroscopic and mass spectrometric studies . These analyses indicated that the most abundant daf-22-dependent indole derivative consists of an indole carboxy unit linked to ascarylose bearing a 9-carbon unsaturated side-chain identical to that found in the known ascr#3 ( see Supporting Information for NMR and MS data ) [7] . Based on its structural relationship to the known ascr#3 , we named the newly identified metabolite indole carboxy ascaroside “icas#3” ( Figure 1E ) . Next we asked whether some of the other daf-22-dependent indole compounds we had detected by DANS also represent indole ascarosides . For this purpose , we employed a mass spectrometric ( MS ) approach , because analysis of the mass spectra of icas#3 had revealed a characteristic MS fragmentation pattern ( loss of the indole-3-carboxy moiety , Figure S1 ) that enabled a screen for related compounds . MS screening for compounds with similar fragmentation profiles indicated that icas#3 is a member of a larger series of indole ascarosides featuring side chains with five to nine carbons ( Figure 1D , E ) . The most abundant components of this family of indole ascarosides are icas#3 , icas#9 , and icas#10 , which are accompanied by smaller amounts of icas#1 and icas#7 ( Figure 1E ) . All of these compounds represent new metabolites , except for icas#9 , which recently has been reported to possess moderate dauer-inducing activity and is unique among known dauer pheromones in producing a bell-shaped response curve [18] . We also detected two new non-indole ascarosides: ascr#9 , which features a saturated 5-carbon side chain , and ascr#10 , which features a saturated 9-carbon side chain , thus representing the saturated analog of the known ascr#3 ( Figure 1F ) . The MS analyses further revealed that the indole ascarosides' quantitative distribution is distinctly different from that of the corresponding non-indole ascarosides , suggesting that incorporation of the indole unit is strongly regulated . Notably , the most abundant indole ascaroside , icas#3 , is accompanied by 10–40-fold larger amounts of the corresponding non-indole ascaroside , ascr#3 , whereas icas#9 is more abundant than the corresponding ascr#9 ( Figure 1G ) . To determine the biosynthetic origin of the indole ascarosides and to exclude the possibility that they are produced by the E . coli food source , we established axenic ( bacteria-free ) in vitro cultures of C . elegans ( N2 ) using the chemically defined CeMM medium [19] , [20] . HPLC-MS analysis of the axenic cultures revealed the presence of icas#1 , icas#3 , icas#9 , and icas#10 , thus indicating that indole ascarosides are produced by C . elegans without participation of dietary bacteria . Use of a 1∶1 mixture of L-[2 , 4 , 5 , 6 , 7-D5]-tryptophan and L-tryptophan in the axenic medium resulted in production of [D5]-icas#1 , [D5]-icas#3 , [D5]-icas#9 , and [D5]-icas#10 , along with equivalent amounts of the unlabelled compounds ( Figure S2 ) . In conclusion , our biochemical studies established the tryptophan origin of the indole-3-carboxy moiety in the indole ascarosides and indicate that these compounds are products of a strongly regulated biosynthetic pathway . The addition of an indole-3-carboxy moiety to the ascarosides represents a significant structural change , and we hypothesized that this chemical difference may indicate signaling functions for these compounds distinct from those of their non-indole cognates . Using synthetic samples ( see Supporting Information ) , we tested three indole ascarosides of varying side-chain lengths , icas#1 , icas#3 , and icas#9 , in the spot attraction assay we had used previously to demonstrate social functions of small molecules ( Figure 2A ) [6] , [7] . We found that all three tested indole ascarosides , icas#1 , icas#3 , and icas#9 , attract both males and hermaphrodites at high concentrations ( Figure 2C ) . Testing the most abundant indole ascaroside , icas#3 , over a broader range of concentrations , we observed that at low concentrations icas#3 was strongly attractive to hermaphrodites , whereas males were no longer attracted ( Figure 2D , Movies S1 , S2 , S3 ) . Similarly , hermaphrodites , but not males , are strongly attracted to low concentrations of icas#9 ( ) . We further investigated hermaphrodite attraction to icas#3 using a quadrant chemotaxis bioassay as described previously ( Figure 2B ) [9] , [21] . In contrast to the spot attraction assay , which measures attraction to a point source of compounds , the quadrant chemotaxis assay measures aggregation of hermaphrodites on plate sections with well-defined compound concentration [9] , [21] . We found that concentrations as low as 1 pM icas#3 result in strong attraction of hermaphrodites ( Figure 2E ) , both in the presence and absence of food ( Figure S3B ) . The biological role of icas#3 thus starkly differs from that of the corresponding non-indole ascaroside ascr#3 , which strongly attracts males but repels hermaphrodites [6] , [7] . Our results show that simply by attaching an indole-3-carboxy group to the 4-position of the ascarylose , the strongly male-attracting ascr#3 is converted into a signal that primarily attracts hermaphrodites . The difference in the amounts at which ascr#3 and icas#3 are produced by the worms corresponds to their relative potency: the male-attracting ascr#3 , which is of much lower potency than icas#3 , is produced in much higher concentrations than the highly potent hermaphrodite attractant icas#3 ( Figure 1G ) . The results from the spot attraction and quadrant chemotaxis assays indicate that hermaphrodites are strongly attracted to indole ascarosides , suggesting that these compounds regulate C . elegans aggregation behavior . C . elegans exhibits natural variation in its foraging behavior with some strains ( e . g . , the common laboratory strain N2 ) dispersing individually on a bacterial lawn , whereas most wild-type strains ( e . g . , RC301 and CB4856 ( Hawaii ) ) accumulate and aggregate where bacteria are the most abundant [10] , [22] . These variants are referred to as “solitary” and “social , ” respectively [10] , [11] . These differences in foraging and aggregation behavior are associated with two different alleles of the neuropeptide Y-like receptor NPR-1 [10] , [11] , which differ at a single amino acid position: solitary strains such as N2 express a high-activity variant of NPR-1 ( 215-valine ) , whereas aggregating strains such as CB4856 express a low-activity variant of NPR-1 ( 215-phenylalanine ) [10] , [11] . The strong loss-of-function mutants npr-1 ( ad609 ) and npr-1 ( ky13 ) , which were generated in the N2 background , also show a high tendency to aggregate [10] , [22] . A previous study showed that loss of function of npr-1 affects hermaphrodite response to non-indole ascarosides [9] . Whereas wild-type ( N2 ) worms expressing the high-activity variant of NPR-1 are repulsed by non-indole ascarosides , npr-1 ( ad609 ) mutants showed attraction to a near-physiological mixture of the most abundant non-indole ascarosides , ascr#2 , ascr#3 , and ascr#5 [9] . We confirmed attraction of npr-1 ( ad609 ) hermaphrodites to ascr#2/3/5 mixtures using both the quadrant chemotaxis and spot attraction assays , but found that hermaphrodites of the two tested social wild-type strains ( RC301 and CB4856 ) show no attraction in either assay ( Figure 3A–B ) . In contrast , both social wild-type strains ( RC301 and CB4856 ) as well as npr-1 ( ad609 ) hermaphrodites were strongly attracted to icas#3 , in both the quadrant chemotaxis and spot-attraction assays ( Figures 3B–D , S3B–C ) . These results indicate that icas#3 functions as a hermaphrodite attractant in both solitary and social C . elegans strains . We next tested how a constant background concentration of indole ascarosides affects hermaphrodite behavior . We measured aggregation of solitary N2 worms and several social strains ( including the social wild-type strain CB4856 and two npr-1 loss-of-function mutants ) in response to icas#3 using two different conditions: “high worm density , ” with 120 worms per 5 cm plate , and “low worm density , ” with 20 worms per 5 cm plate . At low worm density , we observed a very strong increase in aggregation at concentrations as low as 10 fM ( femtomolar ) icas#3 for both solitary and social hermaphrodites ( Figure 4A , 4E ) . Aggregation of N2 hermaphrodites increased as much as 4-fold at 1 pM icas#3 , with higher icas#3 concentrations producing less aggregation . Similarly , the naturally occurring social strain CB4856 displayed a bell-shaped response curve with maximal aggregation at 1 pM of icas#3 and lower aggregation not significantly different from control at 1 nM of icas#3 ( Figure 4A ) . In contrast , icas#3 increased aggregation of npr-1 ( ad609 ) hermaphrodites over the entire tested concentration range , without a drop-off at higher concentrations ( Figure 4A ) . At high worm density , we observed up to a 3-fold increase in aggregation of N2 hermaphrodites on icas#3 plates ( Figure 4B , F ) , whereas hermaphrodites from all three tested social strains showed nearly complete aggregation even in the absence of icas#3 , which precluded detection of any additional aggregation-promoting effect of icas#3 ( Figure 4B ) . These results show that icas#3 increases hermaphrodite aggregation even in the absence of a concentration gradient of this compound , and that solitary and social strains are similarly affected . Similarly , the second-most abundant indole ascaroside , icas#9 , increased aggregation of both solitary and social hermaphrodites ( Figure S4A ) . We also investigated the effect of icas#3 on aggregation of males , which generally tend to aggregate in the absence of hermaphrodites [23] . We found that aggregation of him-5 males on icas#3 plates was significantly increased ( Figure S4B ) . These results show that indole ascarosides promote aggregation behavior even in the absence of a concentration gradient , suggesting that sensing of icas#3 and icas#9 affects response to other aggregation-promoting ( chemical or other ) signals or conditions . For example , secretion of additional indole ascarosides by the worms on plates containing exogenous icas#3 could contribute to the observed increase in aggregation . To investigate this possibility , we tested daf-22 hermaphrodites in the aggregation assay . daf-22 hermaphrodites do not produce indole ascarosides but respond to icas#3 in both the spot attraction and quadrant chemotaxis assay as strongly as N2 worms ( Figures 3B , S3C ) . We found that daf-22 hermaphrodites show less aggregation than N2 worms at 1 pM icas#3 but not at 10 pM icas#3 ( Figure 4C ) . These results suggest that secretion of additional indole ascarosides or other daf-22-dependent compounds by the worms may contribute to aggregation on icas#3 plates , but that other factors , for example low oxygen levels or contact with other worms [12] , [13] , [24] , are more important . Furthermore , changes in locomotory behavior on icas#3 plates could affect the level of aggregation [12] . Using an automated machine-vision system to track worm movement [25] , we found that aggregation-inducing concentrations of icas#3 strongly increase mean stopped duration and affect other locomotory parameters ( Figures 4D , S4D , S4E ) . These changes in worm locomotion , in conjunction with other aggregation-mediating factors , may contribute to the observed increase in aggregation on icas plates . The amphid single-ciliated sensory neurons type K ( ASK ) play an important role in mediating C . elegans behaviors , and previous work has shown that the ASK neurons are required for behavioral responses of males and hermaphrodites to non-indole ascarosides [7] , [9] . ASK sensory neurons are connected via anatomical gap-junctions to the RMG interneuron , which has been shown to act as a central hub regulating aggregation and related behaviors based on input from ASK and other sensory neurons ( Figure 5A ) [9] , [26] . To investigate the neural circuitry required for icas#3-mediated hermaphrodite attraction and aggregation , we first tested whether the ASK neurons are required for these behaviors . For this purpose , we used worms lacking the ASK neurons due to cell-specific expression of mammalian caspase in the developing neurons ( Tokumitsu Wakabayashi , Iwate University Japan , personal communication ) . We found that ablation of ASK sensory neurons resulted in a near complete loss of attraction to icas#3 ( Figure 5B ) . In contrast , ablation of the ASI neurons , which like the ASK neurons partake in dauer pheromone sensing , had no significant effect on icas#3 mediated attraction in hermaphrodites ( Figure 5B ) . Further , ablation of both ASI and ASK neurons did not result in a more significant loss of attraction compared to ASK ablations alone , suggesting that the ASK sensory neurons are required for sensing icas#3 ( Figure 5B ) . Next we tested whether the ASK neurons are required for icas#3 mediated aggregation . We found that hermaphrodites lacking the ASK neurons do not aggregate in response to icas#3 at any of the tested concentrations ( Figure 5C ) . Locomotory analysis of ASK-ablated hermaphrodites on icas#3 plates showed neither increased reversal frequency nor decreased velocity , as we had observed for wild-type worms ( Figure S5A , B ) . Next we tested whether the RMG interneuron is required for icas#3-mediated behaviors . We identified the cell position of the RMG interneuron in wild-type worms using DIC microscopy [27] and in a transgenic strain expressing ncs-1::gfp ( a gift from the Bargmann Lab ) . This transgene expresses GFP in the RMG interneuron and a few other sensory neurons [9] . We found that ablation of the RMG interneuron in both wild-type and ncs-1::gfp worms did not affect icas#3-response in the spot attraction assay ( Figure 5B ) . These results indicate that the RMG interneuron is not required for transduction of icas#3-derived attraction signals from the ASK sensory neurons , in contrast to the behavioral effects of non-indole ascarosides , which require both the ASK sensory neurons and the RMG interneuron [9] . Given this observation , we sought to understand which interneuron downstream of ASK is required for response to icas#3 . According to the wiring diagram of C . elegans , the primary synaptic output of the ASK neuron is the AIA interneuron [26] . To test whether this neuron is required for sensing icas#3 , we used a transgenic line expressing a hyperactive form of MEC-4 in the AIA interneuron ( a kind gift from the Ishihara lab , Japan ) [28] . Expression of MEC-4 , a DEG/ENaC sodium channel , causes neuronal toxicity in C . elegans , thereby resulting in the loss of the AIA neuron [29] . These AIA-deficient worms did not show any attraction to icas#3 , suggesting that the AIA interneurons are required for icas#3 response . Hence the neural circuitry required for attraction to icas#3 is different from that of the non-indole ascarosides . Since behavioral assays showed that the ASK and AIA neurons participate in sensing icas#3 , we asked whether icas#3 elicits calcium flux in these neurons . To measure Ca2+ flux , we used transgenic lines expressing the genetically encoded calcium sensors ( GCaMP ) in these neurons [9] . We used the “Olfactory chip” to restrain the worms and applied ON and OFF steps of icas#3 while imaging from these neurons [30] . We were not able to detect Ca2+ transients in ASK neurons even when applying a wide range of concentration ranging from 1 pM to 1 µM . We then monitored calcium responses in the AIA interneuron , which is the primary synaptic target of the ASK neuron [26] . We found that icas#3 elicited significantly increased G-CaMP fluorescence in the AIA neurons ( Figure 5D , E , Movie S4 ) , similar to the results reported by Macosko et al . for stimulation of AIA interneurons with a mixture of three non-indole ascarosides [9] . These results show that the ASK sensory neurons are required for icas response and that this response is transduced via the AIA interneuron . Previous studies have shown that high , dauer-inducing concentrations of ascr#3 strongly deter both social and solitary hermaphrodites [7] , [9] . To investigate whether addition of ascr#3 would affect icas#3-mediated attraction of hermaphrodites , we tested mixtures containing these two compounds in a near-physiological ratio of 12∶1 ( ascr#3∶icas#3 ) in a modified spot attraction assay , in which we scored N2 hermaphrodite attraction to three concentric zones A–C ( Figure 2A ) . We found that at the lower of the two concentrations tested ( 120 fmol ascr#3 and 10 fmol icas#3 ) , the presence of ascr#3 did not interfere with icas#3-mediated attraction , whereas higher concentrations of ascr#3 resulted in strong repulsion , even in the presence of proportionally increased icas#3 concentrations ( 12 pmol ascr#3 and 1 pmol icas#3 , Figure 6A ) . Following retreat from the outer edge of zone A , many worms remained “trapped” in a circular zone B surrounding central zone A , repulsed by the high concentration of icas#3/ascr#3-blend inside zone A , but attracted by the lower concentrations of the icas#3/ascr#3 blend that diffused into zone B ( see Movie S5 for a visual record of this behavior ) . These results show that at high concentrations of physiological icas#3/ascr#3 mixtures the repulsive effect of ascr#3 prevails , whereas at lower concentrations attraction by icas#3 dominates .
Indole ascarosides are the first C . elegans pheromones that strongly attract wild-type hermaphrodites and promote aggregation . The indole ascarosides fit the broad definition of aggregation pheromones in that they attract and/or arrest conspecifics to the region of release irrespective of sex [1] , [31] , [32] . In promoting these behaviors , the indole ascarosides are active at such low ( femtomolar ) concentrations that the worm's behavioral response must result from sensing of only a few molecules . For example , at an icas#3 concentration of 10 fM there are only about 20 icas#3 molecules contained in a cylinder corresponding to length and diameter of an adult hermaphrodite . Given their high specific activity , it is not surprising that indole ascarosides ( icas' ) are of much lower abundance than non-indole ascarosides ( ascr's ) . The indole ascarosides' strongly attractive properties suggest that these compounds serve to attract conspecifics to desirable environments such as food sources . However , the results from our competition experiments indicate that attraction of hermaphrodites by icas#3 can be counteracted by high concentrations of ascr#3 , which are repulsive to hermaphrodites [7] . The competition experiments further showed that at low concentrations of a physiological blend of icas#3 and ascr#3 , the attractive properties of icas#3 dominate , whereas at high concentrations of the same blend the repulsion by ascr#3 becomes dominant ( Figure 6A , Movie S5 ) . These findings suggest that under dauer-inducing conditions with high population density , the associated high concentrations of ascr#3 promote dispersal [7] , whereas low population density and correspondingly lower concentrations of ascr#3 result in attraction mediated by icas#3 . Therefore , icas' and ascr's could represent opposing stimuli regulating population density and level of aggregation . In turn , population density , food availability , and other environmental factors may affect relative rates of the biosyntheses of ascr's and icas' as part of a regulatory circuit . Indole ascarosides affect aggregation behavior even in the absence of a concentration gradient: very low background concentrations ( fM-pM ) of icas#3 and icas#9 strongly increase the propensity of hermaphrodites ( and males ) to aggregate . This finding suggests that sensing of icas#3 and icas#9 increases susceptibility for aggregation-promoting ( chemical or other ) signals or conditions , for example additional quantities of icas' secreted by the worms on the plate . Aggregation in C . elegans is known to depend on a diverse set of genetic factors and environmental conditions , including food availability and oxygen concentration , suggesting the existence of neuronal circuitry that integrates inputs from different sources [10] , [33]–[36] . Aggregation and attraction signals originating from several different sensory neurons , including the oxygen-sensing URX-neurons and the ascr-sensing ASK neurons , have recently been shown to converge on the RMG interneuron , which is proposed to act as a central hub coordinating these behaviors [9] . The RMG interneuron is the central site of action of the neuropeptide-Y receptor homolog NPR-1 , which distinguishes solitary strains ( high NPR-1 activity ) from social strains ( low NPR-1 activity ) [10] , [11] . In social npr-1 ( lf ) mutant hermaphrodites , oxygen-sensing URX neurons promote aggregation at the edges of the bacterial lawn , whereas solitary N2 hermaphrodites do not respond to oxygen gradients . Similarly , repulsion by ascr's depends on NPR-1 , as solitary hermaphrodites are repelled by ascr's , whereas social npr-1 ( lf ) hermaphrodites display either greatly diminished repulsion or weak attraction [9] . In contrast , we show that icas#3 promotes hermaphrodite attraction and aggregation in both social and solitary strains . Icas#3 attracts solitary N2 as well as social npr-1 ( lf ) hermaphrodites and increases hermaphrodite aggregation in the solitary strain N2 , the social wild-type strains RC301 and CB4856 ( Hawaii ) carrying a low-activity variant of NPR-1 , and the two tested npr-1 null alleles . The finding that icas#3-mediated attraction and aggregation is not reduced by high NPR-1 activity suggests that these icas#3-mediated behaviors rely on signaling pathways distinct from those controlling aggregation responses to other types of stimuli , for example low oxygen levels . This hypothesis is supported by our observation that hermaphrodites lacking the RMG interneuron , which coordinates other aggregation responses via NPR-1 , are still attracted to icas#3 . Furthermore , icas#3-mediated aggregation differs from NPR-1-dependent aggregation behavior in that aggregation of worms on icas#3 plates is more dynamic and not restricted to the edge of the bacterial lawn where oxygen is limited ( Animations S1 , S2 ) . Worm velocity is not significantly reduced at the icas#3 concentrations that induce maximal aggregation ( 1–10 pM , Figure S4D ) , and icas#3-mediated aggregation is associated with less clumping ( average clump size 3-5 worms ) than found for aggregating NPR-1 mutant worms ( average clump size 6–16 worms ) [12] . These observations show that icas#3-mediated aggregation is phenotypically distinct from aggregation behaviors controlled by NPR-1 and the RMG interneuron . Icas#3-mediated attraction and aggregation depend on the ASK neurons , similar to hermaphrodite repulsion and male attraction by ascr's [7] , confirming the central role of this pair of sensory neurons for perception of different types of pheromones in C . elegans ( Figure 5 ) . We further show that icas#3 responses are dependent on the AIA interneurons and do not require the RMG interneuron . Therefore , it appears that the sensory neuron ASK participates in perception of two different types of pheromones , ascr's and icas' , and that these signals are transduced via two different neurophysiological pathways , as part of a complex neural and genetic circuitry integrating a structurally diverse array of pheromone signals . Calcium transients have been recorded from amphid sensory neurons in response to non-indole ascarosides; however , the reported changes in G-CaMP fluorescence were relatively small ( on the order of about 20% ) [9] , [37] . Recently , it was reported that the non-indole ascaroside ascr#5 does not elicit calcium transients in the ASI sensory neurons , although the ASI neurons function as sensors of ascr#5 and express the ascr#5-receptors srg-36 and srg-37 [38] . Similarly , we were unable to detect significant Ca2+ transients in the ASK neurons in response to a wide range of concentrations of icas#3 ( unpublished data ) . It is possible that any icas#3-elicited Ca2+ signals in this neuron are even weaker than those of non-indole ascarosides , as icas#3 is active at extremely low concentrations ( femtomolar to low picomolar ) . Additionally , we cannot rule out involvement of additional neurons in icas#3 signaling , given that the ASK neurons are postsynaptic to a number of other sensory neurons [26] . Notably , icas#3 elicited significant changes in G-CaMP fluorescence in the AIA interneurons , which are the primary postsynaptic targets of the ASK sensory neurons ( Figure 5D , E , Movie S4 ) . The identification of indole ascarosides as aggregation signals reveals unexpected complexity of social signaling in C . elegans . Our results indicate that ascarylose-derived small molecules ( icas' and ascr's ) serve at least three distinct functions in C . elegans: dauer induction , male attraction , and hermaphrodite social signaling ( Figure 6B ) . Previous studies have shown that ascr's often have more than one function; ascr#3 , for example , plays significant roles for both dauer signaling and male attraction [4] , [7] . Our study demonstrates that specific structural variants of ascarylose-derived small molecules are associated with specific functions ( Figure 6C ) . We show that addition of an indole carboxy group to ascr's changes the signaling properties such that the indole-modified compounds can have signaling effects very different from those of the unmodified compounds: icas#3 strongly attracts hermaphrodites and promotes aggregation , whereas ascr#3 repulses hermaphrodites and attracts males . In addition to structural variation , distinct signaling functions are associated with different concentration windows: whereas for dauer formation , high nanomolar concentrations of ascr's are required , low nanomolar to high picomolar concentrations of ascr's promote male attraction , and picomolar to femtomolar concentrations of icas' promote hermaphrodite attraction and aggregation ( Figure 6B ) . Social signaling in C . elegans thus appears to be based on a modular language of small molecules , derived from combinatorial assembly of several structurally distinct building blocks ( Figure 6C ) . Different combinations of these building blocks serve different , occasionally overlapping signaling functions . Our results for the relative abundances of ascr's and icas' with identical side chains ( Figure 1G ) indicate that integration of the different building blocks is carefully controlled . Biochemically , the building blocks are derived from three basic metabolic pathways: carbohydrate metabolism , peroxisomal fatty-acid β-oxidation , and amino acid metabolism . These structural observations raise the possibility that social signaling via small molecules transduces input from the overall metabolic state of the organism . Food availability and nutrient content in conjunction with other environmental factors may control ascr and icas biosynthesis pathways to generate specific pheromone blends that differentially regulate aggregation , mate attraction , and developmental timing . The expansive vocabulary of a modular chemical language would make it possible for different nematodes to signal conspecifically as well as interspecifically , but it is not known whether nematode species other than C . elegans rely on ascarylose-based small molecules for chemical communication . However , lipid-derived glycosides of ascarylose have been identified from several other nematode species [39] , suggesting that many nematodes have the ability to produce ascr- or icas-like compounds . The identification of indole ascarosides as attraction and aggregation signals demonstrates that C . elegans aggregation behavior depends on dedicated chemical signals produced by conspecifics and not just shared preference for specific environmental conditions . C . elegans social signaling thus appears to be significantly more highly evolved than previously suspected .
NMR spectra were recorded on a Varian INOVA 600 NMR ( 600 MHz for 1H , 151 MHz for 13C ) . NMR-spectra were processed using Varian VNMR and MestreLabs MestReC software packages . Additional processing of bitmaps derived from NMR spectra was performed using Adobe Photoshop CS3 as described [6] . HPLC–MS was performed using an Agilent 1100 Series HPLC system equipped with a diode array detector and connected to a Quattro II spectrometer ( Micromass/Waters ) . Data acquisition and processing was controlled by MassLynx software . Flash chromatography was performed using a Teledyne ISCO CombiFlash system . All strains were maintained at 20°C unless mentioned otherwise on NGM agar plates , made with Bacto agar ( BD Biosciences ) , and seeded with OP50 bacteria grown overnight . For the attraction bioassays and the automated tracker experiments , we used C . elegans var . N2 Bristol and males from the him-5 ( e1490 ) strain CB1490 . The him-5 ( e1490 ) mutant segregates XO male progeny by X chromosome nondisjunction during meiosis [40] . For genetic ablation of the ASK neuron , we used the transgenic strain PS6025 qrIs2[sra-9::mCasp1] , which expresses mammalian caspase in the ASK neuron under the influence of the sra-9 promoter ( this strain is a kind gift of Tokumitsu Wakabayashi , Iwate University ) . Other strains used are as follows: CB4856 , C . elegans Hawaii isolate [22]; RC301 , C . elegans Freiburg isolate [10] , [22]; DA609 npr-1 ( ad609 ) ; CX4148 npr-1 ( ky13 ) [10]; CX9740 C . elegans ( N2 ) ; kyEx2144 [ncs-1::GFP] [9]; N2;Ex ( gcy-28::dp::mec-4D ) [28]; CX10981 kyEx2866 [“ASK::GCaMP2 . 2b” sra-9::GCaMP2 . 2b SL2 GFP , ofm-1::GFP] ( ASK imaging line ) ; CX11073 kyEx2916 [“AIA::GCaMP2 . 2b” T01A4 . 1::GCaMP2 . 2b SL2 GFP , ofm-1::GFP] ( AIA imaging line ) [9]; DR476 daf-22 ( m130 ) [17]; and daf-22 ( ok693 ) [16] . All newly identified ascarosides are named with four letter “SMID”s ( Small Molecule IDentifiers ) —e . g . , “icas#3” or “ascr#10 . ” The SMID database ( www . smid-db . org ) is an electronic resource maintained by Frank Schroeder and Lukas Mueller at the Boyce Thompson Institute in collaboration with Paul Sternberg and WormBase ( www . wormbase . org ) . This database catalogues newly identified C . elegans small molecules , assigns a unique four-letter SMID ( a searchable , gene-style Small Molecule IDentifier ) , and for each compound includes a list of other names and abbreviations used in the literature . Metabolite extracts were prepared according to a previously described method [6] , which was modified as follows . Worms ( N2 or daf-22 ) from three 10 cm NGM plates were washed using M9-medium into a 100 mL S-medium pre-culture where they were grown for 5 d at 22°C on a rotary shaker . Concentrated OP50 derived from 1 L of bacterial culture ( grown for 16 h in LB media ) was added as food at days 1 and 3 . Subsequently , the pre-culture was divided equally into four 1 L Erlenmeyer flask containing 400 mL of S-medium for a combined volume of 425 mL of S-medium , which was then grown for an additional 10 d at 22°C on a rotary shaker . Concentrated OP50 derived from 1 L of bacterial culture was added as food every day from days 1 to 9 . Subsequently , the cultures were centrifuged and the supernatant media and worm pellet were lyophilized separately . The lyophilized materials were extracted with 95% ethanol ( 250 mL 2 times ) at room temperature for 12 h . The resulting yellow suspensions were filtered and the filtrate evaporated in vacuo at room temperature , producing media and worm pellet metabolite extracts . The media metabolite extract from two cultures was adsorbed on 6 g of octadecyl-functionalized silica gel and dry loaded into an empty 25 g RediSep Rf sample loading cartridge . The adsorbed material was then fractionated via a reversed-phase RediSep Rf GOLD 30 g HP C18 column using a water-methanol solvent system , starting with 100% water for 4 min , followed by a linear increase of methanol content up to 100% methanol at 42 min , which was continued up until 55 min . The eight fractions generated from this fractionation were evaporated in vacuo . The residue was analyzed by HPLC-MS and 2D-NMR spectroscopy . Worm media extracts or metabolite fractions derived from the chromatographic fractionation were resuspended in 1 . 5 ml methanol , centrifuged at 2 , 000 g for 5 min , and the supernatant submitted to HPLC-MS analyses . HPLC was performed using an Agilent 1100 Series HPLC system equipped with an Agilent Eclipse XDB-C18 column ( 9 . 4×250 mm , 5 µm particle diameter ) . A 0 . 1% acetic acid–acetonitrile solvent gradient was used , starting with an acetonitrile content of 5% for 5 min , which was increased to 100% over a period of 40 min . Mass spectrometry was performed with a Quattro II spectrometer ( Micromass/Waters ) using electrospray ionization in either negative or positive ion mode . Axenic in vitro cultures of C . elegans ( N2 , Bristol ) were established as described by Nass & Hamza [20] , using the C . elegans Maintenance Medium ( CeMM , [19] ) with cholesterol ( 5 mg/l ) instead of sitosterol and nucleoside-5-phosphates . After 21 d the cultures were centrifuged and the supernatant media and worm pellet were lyophilized separately . The lyophilized worm pellets ( 1 . 2–2 . 0 mg ) were extracted with 2 ml methanol , filtered , and concentrated in vacuo . The lyophilized worm media were extracted with ethyl acetate–methanol ( 95∶5 , 100 mL 2 times ) , filtered , and concentrated in vacuo . Residues were taken up in 150 µl methanol and investigated by HPLC-ESI-MS . For the application experiment 50 ml CeMM medium was supplemented with 9 . 2 mg L-[2 , 4 , 5 , 6 , 7-D5]-tryptophan ( from Cambridge Isotope Laboratories ) . These assays were done as previously described [6] , [7] . For both C . elegans hermaphrodites and males , we harvested 50–60 worms daily at the fourth larval stage ( L4 ) and stored them segregated by sex at 20°C overnight to be used as young adults the following day . For the competition experiments we used 120 nM ascr#3 and 10 nM icas#3 ( Condition 1 ) , or 12 µM ascr#3 and 1 µM icas#3 ( Condition 2 ) in water containing 10% ethanol . Aliquots were stored at −20°C in 20 µL tubes . 10% ethanol in water was used as control . Chemotaxis to both non-indole and indole ascarosides was assessed on 10 cm four-quadrant Petri plates [21] . Each quadrant was separated from adjacent ones by plastic spacers ( Figure 2B ) . Pairs of opposite quadrants were filled with nematode growth medium ( NGM ) agar containing either indole ascarosides or non-indole ascarosides at different concentrations . Animals were washed gently in a S-basal buffer and placed in the center of a four-quadrant plate with ascarosides in alternating quadrants , and scored after 15 min and 30 min . A chemotaxis index was calculated as ( the number of animals on ascaroside quadrants minus the number of animals on buffer quadrants ) / ( total number of animals ) . Reversal frequency and velocity were measured using an automated worm-tracking system as previously described [6] , [7] . We measured aggregation behavior of worms using assays described previously [10] . Aggregation assays were conducted on standard NGM plates . Plates containing indole ascarosides were made by adding the indole ascaroside stock solution to the NGM media before they were poured onto the plates . These plates were dried at room temperature for 2–3 d . Control plates were treated similarly except that instead of icas solutions ethanol solutions were added to the plates , corresponding to the amount of ethanol introduced via the icas solutions . Final ethanol concentrations of the plates were below 0 . 1% for all conditions . After drying , both control plates and plates containing indole ascarosides were seeded with 150 µl of an overnight culture of E . coli OP50 using a micropipette and allowed to dry for 2 d at room temperature . For “low worm density” experiments , we placed 20 worms onto the lawn and left them at 20°C for 3 h . For “high worm density” experiments we placed approximately 120 worms onto the bacterial lawn and left them at 20°C for 3 h . Aggregation behavior was quantified as the number of animals that were in touch with two or more animals at >50% of their body length . For calcium imaging we used transgenic lines that express the genetically encoded Ca2+ sensor in ASK ( kyEx2866 ) and AIA ( kyEx2916 ) [9] . Young adults or adult worms were inserted into an “Olfactory chip” microfluidic device . [30] . Dilutions of icas#3 were done with S-basal buffer ( with no cholesterol ) . As stock solutions of icas#3 contained small amounts of ethanol , equivalent amounts of ethanol were added to the S-basal control flow . Imaging was done using an inverted Zeiss microscope equipped with an Andor camera . Exposure time for image acquisition was 300 ms . Before imaging the ASK neuron , the worm was exposed to blue light for 3 min since ASK responds to the blue light itself . This step is necessary so that the neuron adapts to the blue light that is used for Ca2+ measurements . The movies were analyzed using custom-made Matlab scripts . For calculating the average change in fluorescence upon exposure to either buffer or icas#3 , we chose the first peak of fluorescence immediately after exposure to buffer or icas#3 . The value for this maximum was then subtracted from the mean fluorescence during the 5 s before the delivery of icas#3/buffer ( corresponding to the region between 5 s to 10 s in Figure 5D ) . Figures 2C , D , 3A , D , 6A , S3A , and S4C: We used unpaired Student's t tests with Welch's correction for comparing attraction of hermaphrodites and males on indole ascarosides *p<0 . 01 , **p<0 . 001 , ***p<0 . 0001 . Figures 2E , 3B , C: For comparing the quadrant chemotaxis indices of the various strains , we used one-factor ANOVA followed by Dunnett's post-test , *p<0 . 05 , **p<0 . 01 . Figures 4A–C , S3C , S4A , B: For comparing aggregation of solitary , social worms and Cel-daf-22 on plates containing indole ascarosides , we used one-factor ANOVA followed by Dunnett's post-test , *p<0 . 05 , **p<0 . 01 . Figure 4D: To compare stopped duration of worms on plates with indole ascarosides , we used one-factor ANOVA followed by Dunnett's post-test , *p<0 . 05 , **p<0 . 01 . Figure S4D , E: To compare velocities and reversal frequencies on plates with indole ascarosides , we used one-factor ANOVA followed by Dunnett's post-test , *p<0 . 05 , **p<0 . 01 . Figure S5A , B: To compare reversals between unablated and ASK ablated lines , we used Student's t tests with Welch's correction , *p<0 . 01 , **p<0 . 001 . Figure 5B: To compare the attraction of wild-type worms to the genetically ablated lines for ASK and AIA as well as the ASI and RMG neuron ablations , we used unpaired Student's t test with Welch's correction , ***p<0 . 0001 . Figure 5E: For comparing G-CaMP fluorescence changes to buffer and icas#3 , we used unpaired Student's t test with Welch's correction , **p<0 . 001 . All error bars indicate standard error of mean ( S . E . M ) . Samples of indole ascarosides icas#1 , icas#7 , icas#3 , and icas#9 for use in biological assays and as standards for HPLC-MS were prepared via chemical synthesis . Detailed procedures and NMR-spectroscopic data are contained in Text S1 . | Chemical signaling is an ancient form of inter-organismal communication . The nematode Caenorhabditis elegans exhibits a wide range of social behaviors , including mutual attraction and aggregation , and has served as a useful model towards investigating the signaling pathways that regulate these behaviors . Recent investigations showed that other C . elegans behaviors , like population density sensing and mating , are regulated by small molecule signals called ascarosides . However , it remained unclear whether C . elegans uses small molecules to promote intraspecific attraction and aggregation , despite the presence of extensive neural circuitry regulating these behaviors . In this study , we show that C . elegans uses a specifically modified variant of the ascarosides including an indole unit as a highly potent aggregation pheromone . These indole ascarosides integrate input from two major metabolic pathways , amino acid catabolism and lipid beta-oxidation , suggesting that C . elegans communicates metabolic status via a modular code of small-molecule signals . Our study thus provides evidence for use of a multilayered chemical language for inter-organismal signaling by a model organism . Understanding of chemical signaling in nematodes may aid the development of new treatment approaches for parasitic nematodes , which remain among the most prevalent human disease agents . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"behavioral",
"neuroscience",
"model",
"organisms",
"ecology",
"chemical",
"biology",
"chemistry",
"biology",
"neuroscience",
"behavioral",
"ecology"
] | 2012 | A Modular Library of Small Molecule Signals Regulates Social Behaviors in Caenorhabditis elegans |
Cryptococcus neoformans is an encapsulated pathogenic yeast that can change the size of the cells during infection . In particular , this process can occur by enlarging the size of the capsule without modifying the size of the cell body , or by increasing the diameter of the cell body , which is normally accompanied by an increase of the capsule too . This last process leads to the formation of cells of an abnormal enlarged size denominated titan cells . Previous works characterized titan cell formation during pulmonary infection but research on this topic has been hampered due to the difficulty to obtain them in vitro . In this work , we describe in vitro conditions ( low nutrient , serum supplemented medium at neutral pH ) that promote the transition from regular to titan-like cells . Moreover , addition of azide and static incubation of the cultures in a CO2 enriched atmosphere favored cellular enlargement . This transition occurred at low cell densities , suggesting that the process was regulated by quorum sensing molecules and it was independent of the cryptococcal serotype/species . Transition to titan-like cell was impaired by pharmacological inhibition of PKC signaling pathway . Analysis of the gene expression profile during the transition to titan-like cells showed overexpression of enzymes involved in carbohydrate metabolism , as well as proteins from the coatomer complex , and related to iron metabolism . Indeed , we observed that iron limitation also induced the formation of titan cells . Our gene expression analysis also revealed other elements involved in titan cell formation , such as calnexin , whose absence resulted in appearance of abnormal large cells even in regular rich media . In summary , our work provides a new alternative method to investigate titan cell formation devoid the bioethical problems that involve animal experimentation .
Cryptococcus neoformans is a basidiomycetes yeast widely distributed in the environment that can behave as a pathogen in susceptible patients [1 , 2] . Cryptococcus neoformans can survive in the lung , but in immunosuppressed patients it can also spread to the central nervous system and cause meningoencephalitis [2] . Cryptococcal infections are major causes of death in HIV patients . Although the incidence has significantly decreased in developed countries due to the introduction of antiretroviral therapy ( ART ) , associated mortality remains high [3 , 4] . Moreover , infections by this yeast still present a high incidence in developing areas , such as the sub-Saharan Africa and Southeast Asia [5 , 6] . One of the most characteristic features of Cryptococcus is its ability to adapt to the lung environment and to evade the host immune response . Several factors contributing to cryptococcal adaptation to the lung have been described . The most important is the presence of a polysaccharide capsule [7–9] , which is antiphagocytic and protects the yeasts from stress conditions [2 , 10] . The size of the capsule is not constant , and it increases during the first hours of interaction with the host [11] , which indicates that this process is a response that contributes to immune evasion . Furthermore , the capsular polysaccharide is also secreted to the extracellular media where it induces immunological paralysis through multiple mechanisms [9 , 12–15] . Cryptococcus neoformans is also a facultative intracellular pathogen in phagocytic cells [16–19] , which is another important factor that contributes to fungal survival in the host . Cryptococcus has also developed other adaptation mechanisms that contribute to the evasion of the immune response . One of them involves the formation of titan cells , which have an abnormal large size . The average diameter of Cryptococcus cells grown in vitro ranges between 4–7 microns . In contrast , the fungal population in the lungs is very heterogeneous , and cells of even 100 microns have been described [11 , 20–22] . Titan cells have been arbitrarily defined as those with a cell body diameter above 15 microns or with a total size ( capsule included ) over 30 microns [23] . Because of their size , titan cells cannot be phagocytosed , and they can persist in the host for long periods [24 , 25] . Titan cells also contribute to virulence through other mechanisms . For example , they can divide and produce a progeny of regular size that has increased resistance to stress factors [26] as well as the ability to inhibit the phagocytosis of cells of regular size [25] . Some signals and pathways involved in titan cell formation have been characterized , and it is known that the cAMP pathway is required for this phenomenon [20 , 21 , 27] . Titan cell formation has been associated in vivo with anti-inflammatory Th2 type immune responses [23] , but the host's factors that trigger this morphological transition remain unknown . The investigation and characterization of titan cells has been limited by the lack of media allowing the transition in vitro , so most of the data about these cells has been obtained using animal models . Although this approach has been shown to be useful for some purposes , it does not allow obtaining a large population of titan cells . In addition , the use of mice for these purposes presents associated significant bioethical issues . In this work , we have defined in vitro conditions that induce cell enlargement in C . neoformans , which lead to the appearance of cells similar to those found in vivo , here denominated titan-like cells . We found that incubation of this fungus in low nutrient media supplemented with serum in a CO2 enriched atmosphere induced cryptococcal cell size increase . Moreover , other factors , such as oxygen limitation , or low cell density enhanced the cell growth . We have used this medium as a first step in the characterization of this transition . Our findings open future research lines that will help to define the molecular mechanisms that trigger titan cell formation and their role during infection .
In the last years , we have characterized the phenomenon of capsule growth in vitro using a medium that contains 10% Sabouraud buffered at pH 7 . 3 with 50 mM MOPS . In a set of experiments , we observed that addition of serum and of the respiration inhibitor sodium azide to this medium produced not only growth of the capsule but also of the cell body ( Fig 1A–1C ) . These cells resembled the titan cells observed in vivo , although they did not reach the same size . Despite this difference , we argued that this phenomenon might reflect the first steps of titan cell formation , so we decided to characterize the factors that induced in vitro cell growth . First , the morphology of C . neoformans was analyzed in regular growth conditions ( liquid Sabouraud ) , in capsule inducing medium ( 10% Sabouraud buffered with 50 mM MOPS ) and in 5% Sabouraud buffered with MOPS and 5% FBS + azide ( 15 μM ) . Titan cells have been defined as those with total diameter of 30 μm or those with a cell body diameter larger than 15 μm [22] . As shown in Fig 1D , we observed a significant increase of both the cell size and the capsule in this last medium , almost reaching the threshold for titan cells definition after three days of culture . For this reason , we decided to name the medium as TCM ( Titan Cell Medium ) . Serum was essential for cellular enlargement formation because in its absence the increase in cell size was significantly lower ( p<0 . 05 , Fig 2A ) . We characterized in detail the factors and conditions that favor cryptococcal cell size increase in vitro . In our initial experiments , the medium in which we first observed cells of enlarged size contained subinhibitory concentrations of the mitochondrial inhibitor sodium azide to prevent contamination . As shown in Fig 2B , the process of cell enlargement was enhanced in the presence of sodium azide . This compound is an inhibitor of complex IV of the respiratory chain , so we argued that other factors that alter respiration could induce increases in cell size . For this reason , we investigated whether cryptococcal cell growth was influenced by oxygen limitation . To assess that , we compared the yeast morphology under shaking or static conditions . As shown in Fig 2C , static incubation of the cultures resulted in a greater proportion of cells with enlarged size . Cryptococcal cells sense and respond to environmental levels of CO2 and it is known that this molecule induces capsule growth . For this reason , we investigated if incubation of the cultures in a CO2-enriched environment altered cryptococcal cell size . We found that yeast size was significantly larger when the plates were placed in a 5% CO2 atmosphere in comparison to growth without CO2 ( Fig 2D ) . To visualize the phenomenon of cell growth , we carried out in vivo imaging by placing the cells in a 96-wells plate in TCM in 5% CO2 at 37°C under a microscope overnight and obtained videos of the cellular enlargement . As shown in S1 Video , cells actively grew and replicated in Sabouraud medium . However , in TCM , after 5–8 h of incubation the cells started to enlarge during 8–10 h ( S2 Video ) . After this time , the cells stopped enlarging and continued budding . We also observed that cellular enlargement was associated with some intracellular phenotypic changes . For instance , a significant proportion of the cells displayed an intracellular compartment that started to divide by fission , but then fused again to render a large vesicle ( S3 Video ) . To evaluate to which extent the cells obtained in vitro resembled titan cells found in animal models , we infected mice and isolated cryptococcal cells after 14 days of infection ( see Material & Methods ) . We first determined the total size , capsule size , and cell body size of titan cells . As shown in Fig 3 , the cells obtained in vitro in TCM did not reach the size found in the cells in vivo . The main difference between both types of cells was at the capsule , being it size significantly larger in titan cells isolated from mice . For this reason , we decided to denominate the cells obtained in vitro as “titan-like” cells . Titan cells also present differences in capsular features , such as the density , compared to cells obtained in vitro [21] . To investigate if the density of titan-like cells was similar to that observed in vivo , we measured the permeability index using fluorescently labeled dextrans of different molecular weights ( 70 and 40 kDa ) . As shown in Fig 3D and 3E , penetration of the dextrans in the capsule of titan-like cells was reduced compared to cells of regular size and similar to that of titan cells isolated from the lung of infected mice . This result indicated that titan-like cells had a capsule of similar density of that of titan cells generated in vivo . Although serum was required to induce cellular enlargement in vitro , it was not sufficient for this process . Serum did not induce cellular growth in rich media ( Fig 4A and 4B ) , in contrast to the situation in the diluted medium nutrients ( Fig 4C ) , indicating that nutrient limitation was important for cell size enlargement . Phospholipids , in particular phosphatidylcholine , can trigger the appearance of titan cells in vitro [28] . For this reason , we performed a lipid extraction of fetal calf serum , present in TCM and we incubated the cells with different amounts of these lipids ( 1/40 , 1/100 and 1/200 dilution of the original lipid solution ) . As shown in Fig 4D , the lipids present in the serum induced titan-like cell formation . Phosphatidylcholine ( PC ) is one of the major phospholipids contained in serum and mammalian membranes . For this reason , we investigated if this molecule had any effect on titan-like cell formation . We performed experiments in which serum was replaced by different concentrations of PC , and compared its effect with control samples containing the same concentration of solvent ( ethanol ) . As shown in Fig 4E , PC induced an increase in cryptococcal cell size . However , this increase was lower compared to the one observed with serum , indicating that although PC seemed to induce titan-like cells , there are other serum components responsible for the cell increase . We found that titan-like cell formation depended on the cell density of the cultures . We inoculated 96-wells plates with different concentrations of cells from H99 strain ( 106 , 105 , 104 and 103 cells/mL ) in TCM and Sabouraud as a control of growth . After overnight incubation at 37°C with CO2 , titan-like cells were observed in the wells inoculated with 103 , 104 and 105 cells/mL but were almost absent in the wells that were inoculated with the higher cell density ( 106 cells/mL , Fig 5A ) . We found that titan-like cells were observed more frequently when the cultures were inoculated with cellular concentrations around 104 cells/mL . The fact that the formation of titan-like cells depends on cell density suggests that this process could be regulated by quorum sensing ( QS ) . QS is a cell-cell communication mechanism mediated by molecules that are released directly into the medium by microorganisms . These molecules are released as a function of growth and replication rate [29 , 30] . In this way , we evaluated the influence on titan-like cell formation of cell-free media obtained from titan-like and regular C . neoformans cultures . We inoculated TCM with the H99 strain at 106 cells/mL and 104 cells/mL and incubated the cultures for 18 h at 37°C in 5% CO2 to obtain cells of regular size and titan-like cells , respectively . We then collected the supernatants ( named RCS and TCS , respectively ) . These conditioned media were added to wells that contained fresh TCM inoculated at 104 cells/mL . We found that the conditioned medium RCS significantly inhibited the formation of titan-like cells ( Fig 5B ) even when added to fresh TCM ( TCM + RCS ) ( p<0 . 001 ) . In contrast , the supernatant from titan-like cells cultures ( TCS ) did not block the formation of the titan-like cells , demonstrating a negative effect of the supernatant obtained from cells of regular sizes on titan-like cell formation ( Fig 5B ) . The effect of the TCS conditioned medium was not explained by the dilution of the nutrients of the fresh TCM , since titan-like cells were still formed in TCM diluted with distilled water ( Fig 5B ) . In C . neoformans , the main QS molecule described is a short peptide ( 11-mer ) called Qsp1 that is required for fungal virulence , replication , cell wall synthesis and protease activities [31 , 32] . To investigate the influence of Qsp1 in the formation of titan-like cells , different concentrations of the peptide were added to the TCM medium and the formation of titan-like cells was evaluated . We observed that Qsp1 significantly inhibited formation of titan-like cells in a dose-dependent manner ( Fig 5C ) . As control , we used both inactive and scrambled versions of Qsp1 , and observed that none of them had any effect on titan-like cell development ( Fig 5C ) . It could be argued that the production of Qsp1 in TCM cultures inoculated at high cell densities was responsible for the inhibition of titan-like cell formation . To test this idea , we used a qsp1 mutant that does not produce Qsp1 [32] . Our results showed that the mutant produced titan-like cells in a similar way as the wild type strain KN99 ( Fig 5D ) [32] , even at high cell densities ( 106 cells/mL ) . This result indicates that absence of titan-like cells in TCM cultures inoculated at high densities was not only due to Qsp1 , and that most probably , other QS molecules secreted by C . neoformans might influence cellular enlargement . Titan cells formed in the lungs are polyploid and single-nucleated . So we investigated the morphology of the nucleus and the DNA content after staining with DAPI . As shown in Fig 6A and 6B , titan-like cells contained one nucleus . This result was confirmed using a strain that expresses a fluorescent nucleolar protein ( NOP1-mCherry , [33] , S1 Fig ) . We also quantified the fluorescence intensity of the DAPI staining by flow cytometry . As shown in Fig 6C–6E , titan-like cells emitted more fluorescence than cells of regular size . The fluorescence intensity was more heterogeneous in titan-like cells , ranging from 2 to 5-fold increase compared to cells of normal size ( Fig 6E ) . PKC is a family of protein kinases , activated by Ca2+ , diacylglycerol ( DAG ) and phospholipids , and involved in different virulence-related aspects in C . neoformans such as melanin production [34] , temperature tolerance , cell integrity [35] and fluconazole tolerance [36] . We argued that serum phospholipids could activate the PKC signaling pathway , so we tested the effect of three PKC inhibitors , calphostin C , staurosporine and bisindolylmaleimide I on titan-like cell formation . We observed that all PKC inhibitors ( in particular , staurosporine and Calphostine C ) , impaired the formation of titan-like cells in a dose-dependent manner ( Fig 7A–7C ) . We also included the tyrosine kinase inhibitor genistein and found that this compound had no visible effect on cellular enlargement ( Fig 7D ) . Cryptococcus neoformans is divided in different serotypes and varieties: variety grubii ( serotype A ) , variety neoformans ( serotype D ) , and A/D hybrids . It has been proposed that these groups should be divided into different species [37] , although recent reports argue against this differentiation and suggest the term of species complexes [38] . We investigated if there was a correlation between the main species complex of the strain and the formation of titan-like cells . As shown in Fig 8A–8C , for each serotype/species complex , there were strains with high and low capacity to induce titan-like cells , which indicates that there was not a direct association between the serotype and cell growth . We also investigated the behavior of strains from the related species C . gattii , which can infect immunocompetent patients . As shown in Fig 8D , there were strains that had high and low ability to induce titan-like cells . We next examined if the hyper ( CBS10514 , R265 ) and hypovirulent ( CBS10865 , R272 ) strains isolated at the Vancouver outbreak [39 , 40] formed titan-like cells in a different way . Interestingly , the hypervirulent strain tended to produce more titan-like cells compared to the strain with reduced virulence ( Fig 8D ) . We also tested other C . gattii strains ( NIH 191 and NIH198 ) , which also presented differences in their capacity in producing titan-like cells . In summary , the many inter-strain differences in the capability to form titan-like cells did not allow associating this ability to the serotype/genotype of the isolates . Coinfection with a and α strains results in a higher proportion of titan cells in the lungs [20] , so we investigated whether strains from different mating type had different ability to form titan-like cells . We studied three pair of strains with different mating type JEC20/JEC21 , NE822/NE824 and 3259/3260 ( KN99 ) , and we found that the pair 3259/3260 increased the size of the cell body in the TCM medium compared to the Sabouraud significantly ( p<0 . 05 , Fig 8E ) . In NE822/NE824 and JEC20/JEC21 pairs , there was a small increase in cell size in TCM and we observed a small amount of titan-like cells in this medium although this difference was not statistically significant . These results indicate that titan-like cell formation can occur in vitro independently of the mating type of the strains . However , co-incubation of both mating types did not result in a higher proportion of titan cells in TCM ( S2 Fig ) . To identify genes that play an important role in the formation of titan-like cells , we studied the phenotype of mutants with problems to induce capsule growth , such as gat201 or ada2 [41 , 42] or acapsular strains ( cap59 and cap60 ) . As shown in Fig 9A , none of these mutants induced cellular enlargement in TCM . Titan cell formation is regulated by the cAMP pathway [21 , 27] . Moreover , the CO2 activates adenylate cyclase [43 , 44] . For this reason , we investigated the formation of these cells in cac1 mutant ( which lack the enzyme adenylate cyclase ) and in the reconstituted strain cac1/CAC1 . As shown in Fig 9B , the cac1 mutant was defective to produce cellular enlargement , whereas the reconstituted strain produced titan-like cells as the wild type . CO2 is transformed into HCO3- by the action of carbonic anhydrases ( Can ) . In C . neoformans , there are two genes encoding these enzymes ( CAN1 and CAN2 ) [44 , 45] , being Can2 the most abundant and physiologically active . Since can2 mutant can only grow in a CO2 enriched environment , these strains were maintained always in 5% CO2 . Deletion of CAN2 did not have any effect of titan-like cell formation . Strikingly , in the absence of CAN1 , titan-like cell size was even larger than that observed for the WT strain in TCM ( Fig 9C ) . Titan cells are not phagocytosed in vivo [21 , 25] . For this reason , we examined the interaction between macrophage-like cell lines and titan-like cells obtained in vitro . We compared the phagocytosis of titan-like cells ( incubated in TCM inoculated at 3x104 cells/mL ) , and of cells of regular size obtained in TCM inoculated at 106 cells/mL or in Sabouraud . Phagocytosis was quantified both by microscopic observation and by flow cytometry using a cryptococcal strain that expressed GFP ( see Material and Methods ) . In all cases , we found that titan-like cells obtained in vitro were not phagocytosed ( Fig 10A and S4 Video ) . Interestingly , cells of regular size were not equally internalized , since cells cultivated in Sabouraud medium inoculated at high cellular density were more efficiently phagocytosed than the same cells inoculum grown in TCM ( S5 and S6 Videos , Fig 10A ) . This difference was not due to difference in binding to the antibody used as opsonin in this experiment ( 18B7 , S3 Fig ) , but correlated with the increase in cell size ( in particular , due to capsule enlargement ) found in cells inoculated in TCM at high density compared to the cells cultivated in Sabouraud ( S3 Fig ) . It has been shown that titan cells can also prevent phagocytosis in vivo of cryptococcal cells of regular size [25] , so we tested if titan-like cells induced a similar phenomenon in vitro . For this purpose , RAW264 . 7 macrophages were preincubated with titan-like cells for 1 h with mAb 18B7 . As control , parallel samples were incubated only with titan-like cells without mAb or with medium without yeast cells ( with or without mAb ) . After this time , the plate was washed to remove titan-like cells , and H99-GFP cells of normal size cultivated in liquid Sabouraud were added to the macrophages . As shown in Fig 10B , preincubation of the macrophages with titan-like cells did not affect the phagocytosis of regular size yeasts ( Fig 10B ) . To gain insights about the molecular mechanisms involved in titan-like cell formation , we compared their gene expression profile with that of cells of regular size . To this end , cryptococcal cells were inoculated in TCM at low ( 104 cells/mL , titan-like cells ) and high densities ( 106 cells/mL , regular cells ) at 37°C with CO2 , and total RNA was isolated after 7 and 18 h of incubation . Differences in gene expression were investigated by RNAseq ( see material and methods ) . After mapping the reads and subsequent analysis of differentially expressed genes using the DESeq2 algorithm , we found 42 genes induced at least 1 . 6-fold in titan-like cells after 7 h and 400 genes induced at least 2-fold after 18 h ( Fig 11 ) . A lower expression threshold was set for the 7h data to carry out a comparative GO analysis . Interestingly , the number of repressed genes ( after 7h ) was relatively high ( 327 ) and did not increase after 18 h ( 312 ) . Gene Ontology analysis of genes induced at 7 h revealed a moderate increase in genes involved in the tricarboxylic acid ( TCA ) /glyoxylate cycle and in response to stress and protein folding . The effect of the transition was more evident after 18 h ( Fig 11B and Table 1 ) , with the induction of numerous genes involved in carbohydrate metabolism . This included diverse genes related to glycolysis , such as CNAG_03769 and CNAG_05480 , encoding two isoforms of hexokinase , CNAG_03916 ( glucose-6-phosphate isomerase ) , CNAG_04676 ( 6-phosphofructokinase ) , CNAG_06770 ( fructose-bisphosphate aldolase 1 ) , CNAG_02035 ( triose-phosphate isomerase ) and two isoforms of enolase ( CNAG_03072 and CNAG_06868 ) . The TCA cycle was also affected , with induction of mitochondrial citrate synthase ( CNAG_00061 ) , E1 and E2 components of ketoglutarate deshydrogenase ( CNAG_03674 and CNAG_03596 ) , the two subunits of isocitrate dehydrogenase ( CNAG_07363 and CNAG_07851 ) , and CNAG_05653 , encoding malate synthase . In addition , genes belonging to the oxidative segment of the pentose-phosphate shunt were also induced , such as glucose-6-phosphate dehydrogenase ( CNAG_03245 ) , 6-phosphogluconolactonase ( CNAG_02133 ) , and two isoforms of 6-phosphogluconate dehydrogenase ( CNAG_04099 and CNAG_07561 ) . Finally , genes involved in trehalose metabolism , such as CNAG_03113 , CNAG_03765 , and CNAG_05292 , encoding a trehalose phosphatase and two trehalose synthases , respectively , were upregulated . Diverse genes encoding closely related enzymes involved in amino sugar metabolism were also induced , including glutamine-fructose-6-phosphate transaminase ( CNAG_01164 ) , phosphoacetylglucosamine mutase ( CNAG_02445 ) , glucosamine-6-phosphate deaminase ( CNAG_06098 ) , and N-acetylglucosamine-6-phosphate deacetylase ( CNAG_06098 ) , plus CNAG_02355 , coding for a UDP-xylose/UDP-N-acetylglucosamine transporter . The expression of two additional genes related to chitin degradation ( CNAG_01239 , encoding chitin deacetylase , and CNAG_06659 , hexosaminidase ) , was also increased . Gene Ontology analysis did not report any significantly enriched category for the set of gene repressed at 7h when the Benjamini-Hochberg adjustment method was used . However , when a non-adjusted search was conducted , we found several genes related to ribosome biogenesis , DNA repair and DNA replication . In particular , three genes ( CNAG_00681 , CNAG_01959 , and CNAG_03148 ) involved in the nuclear condensin complex , required for establishment and maintenance of chromosome condensation , and 3 out of 4 components of the GINS complex ( CNAG_03374 , CNAG_02884 , and CNAG_04682 ) which participates in both initiation and elongation of DNA replication . Similarly , the expression of two cyclin-encoding genes ( CNAG_03385 and CNAG_00442 ) belonging to the G1-phase Pcl1 , 2 family , was clearly repressed ( Table 2 ) . These assignations were confirmed by exploring the FunCat classification ontology . Analysis of the repressed genes after 18 h incubation yielded two main GO categories: genes related to transmembrane transport and to the regulation of RNA polymerase II promoters . In the former category there were included diverse sugar transporters ( CNAG_01862 , CNAG_02733 , CNAG_05662 , CNAG_06292 , and CNAG_06932 ) , plus a likely glycerol proton symporter similar to budding yeast Stl1 ( CNAG_04784 ) . The latter category contained 14 putative transcription factors , most of them belonging to the zinc finger transcriptional activator family . However , in most cases their sequence similarity with previously characterized transcription factors was rather low , thus precluding a tentative functional assignment . Interestingly , several genes encoding proteins involved in Golgi-related protein traffic were induced , specifically five out of the seven components of the COPI-coated vesicles , also known as coatomer ( α , β , β' , γ , and ε , encoded by CNAG_03554 , CNAG_03299 , CNAG_04074 , CNAG_01274 , and CNAG_01211 , respectively ) . COPI-coated vesicles are found associated with Golgi membranes , are involved in Golgi to endoplasmic reticulum ( retrograde ) vesicle transport , and also likely in intra-Golgi transport . The expression of components of the COPII coat , such as CNAG_06773 ( SEC24 ) and CNAG_04803 ( SEC31 ) , as well as that of the interacting protein CNAG_00148 ( SEC16 ) also increased . In addition , genes encoding proteins required for cargo-selective , clathrin-dependent transport from the TNG to endosomes in yeast were induced . These are CNAG_00977 , encoding a likely homolog of the budding yeast monomeric clathrin adaptor proteins GGA1 or GGA2 , CNAG_07318 , encoding the AP-1 complex subunit gamma-1 , CNAG_06564 , coding for an AP-1 interacting protein , or the genes coding for clathrin light and heavy chains ( CNAG_04499 and CNAG_04904 ) . One of the genes clearly overexpressed at both 7 and 24 h encoded Calnexin ( CNAG_02500 ) , which is a chaperone located at the ER that contributes to the proper folding of glycoproteins . To test if this protein had any effect on titan cell formation , we obtained the mutant and investigated its phenotype . As shown in Fig 12 , in the absence of calnexin , the cells had significantly larger size in regular rich medium compared to the wild type strain ( Fig 12 ) , and no change was detected when the cells were transferred to TCM . This result confirmed that calnexin was involved in processes that regulate cell size in C . neoformans . We also noticed that some genes related to iron metabolism were also induced in titan-like cells compared to cells of regular size . One of them ( CNAG_01653 ) encoded Cig1 , ( initially annotated as cytokine inducing glycoprotein ) , which is involved in iron uptake from heme groups and is overexpressed during iron limitation conditions [46 , 47] . Iron is an important element required by yeasts , and in C . neoformans , its limitation also induces capsular enlargement . For this reason , we investigated whether iron concentration had any effect on titan-like cell formation . As shown in Fig 13 , depletion of iron using the chelator BPS induced the appearance of cells of large size when the cultures were inoculated at high cell densities ( 106 cells/mL ) , suggesting that iron limitation also contributes to cellular enlargement .
Cryptococcus neoformans is an exceptional model to understand mechanisms induced by pathogenic fungi to adapt to the host and cause disease . Some of the most important changes are related to changes in cell size , which can occur by growth of the capsule , or growth of both the cell body and capsule . In this last case , C . neoformans induces a specific cell type that has been denominated titan cells [20 , 21] . At the moment , only the cAMP-dependent and the mating pathways have been described as relevant in titan cell formation [20 , 21 , 27] . It has been shown that increased expression of PKA induces cell enlargement in C . neoformans and the ploidy of the cells [48 , 49] . One of the limitations to study titan cells is the difficulty of reproducing this phenomenon in vitro . Originally , we described that a small proportion of titan cells could be observed in minimal media [21] . Then , it was described that these cells appear in the presence of phospholipids [28] . In this work , we have defined a medium in which we have consistently replicated this phenomenon in vitro . This medium has several characteristics: limitation of nutrients at neutral pH , inclusion of mammalian serum and a CO2 enriched atmosphere . In these conditions , we could obtain a high proportion of cells of a size around 30 μm ( capsule included ) . This diameter is smaller than the average size measured in cells isolated from mouse lungs [11 , 20 , 22] , where average size is around 40–50 μm , so we argue that in our TCM we could be examining the initial steps of this morphological transition . It is possible that in vitro the cells do not reach the same size than cells obtained from in vivo infections because the nutrients of the medium or the factors that favor this morphological change are consumed . Instead , in vivo infections are maintained for a longer time , with a stable nutrient concentration allowing cells to reach significantly larger sizes . Despite this limitation , the availability of in vitro conditions that , at least in part mimic titan cell formation is a key contribution to understand the biology of these cells . In general , our results indicate that titan-like cell formation is induced by multiple factors , being some of them necessary , but not sufficient , to trigger the transition . This is the case of serum which only induced cell growth under nutrient limiting conditions . These results indicate that titan-like cells are formed in response to some elements of the host present in the serum in the context of the stress produced by the limitation of nutrients in the medium . These findings allow the dissection of the intracellular pathways that are triggered during cellular growth . It has been described that serum and the fraction of polar lipids induce cell enlargement in C . neoformans . Our results agree with previous findings [28] , because polar lipids were also able to induce the formation of titan cells in TCM . The most plausible mechanism for this effect is that phospholipids are degraded by phospholipase C , which produces diacylglycerol ( DG ) , or by phospholipase B , which produces arachidonic acid ( AA ) [50] . DG activates human as well as C . neoformans PKC [34] . To investigate the role of this signaling pathway , we blocked the activity of this kinase with several pharmacological inhibitors , and found that both staurosporine and calphostine C had a dramatic effect on titan-like cell formation . This protein participates in multiples processes , such as maintenance of the integrity of the cell wall [51–53] and polarized growth [54] . Mutants lacking the PKC1 gene present many cellular alterations , such as osmotic instability and susceptibility to temperature [35] . We tried to evaluate the effect of these inhibitors at a lower temperature and in the presence of sorbitol in an attempt to overcome the cellular defects associated with the absence of Pkc1 . For this reason , further studies using mutants affected in signaling pathway components are required to understand in detail how this pathway participates in titan-like cell formation . We also observed that other important factor for titan-like cell formation in vitro is CO2 . The capsule is induced in response to CO2 concentrations present in the host [55] , which is consistent with this factor also favoring the growth of the cell body . Carbonic anhydrase converts CO2 into HCO3- , which activates adenylate cyclase [44 , 56] , so our results are consistent with the hypothesis that CO2 induces cell growth through the cAMP pathway . In C . neoformans , there are two carbonic anhydrases encoding genes , CAN1 and CAN2 , with CAN2 being the most important [44] . Interestingly , deletion of CAN1 resulted in an enhancement of the production of titan-like cells . Although we do not know the molecular mechanism for this phenomenon , we postulate that in the absence of Can1 , there might be a compensatory overexpression of Can2 that could induce the hyperactivation of the cAMP pathway . We also observed that subinhibitory concentrations of azide have a modest , but reproducible positive effect on titan-like cell development . We argue that a partial inhibition of the respiratory chain can trigger a stress signal that results in a stop of the cell cycle and allows cellular size increase . In this regard , it could be assumed that a limitation in the respiratory capacity might lead the organism to generate energy , at least in part , by fermentative metabolism , a situation that has been associated to increased PKA activity in many fungi [57] . Although C . neoformans is mainly a respiratory yeast , it has been shown that it can produce both ethanol and acetate in vitro and in vivo [58 , 59] , so a similar activation of the PKA activity could also occur in these conditions . However , the effect of the mitochondrial inhibitor was not significant in the presence other factors ( in particular , serum and CO2 ) . Since CO2 is an activator of the adenylate cyclase , it is reasonable to argue that in these conditions the effect of azide would be not significant . One of the most striking results of this work is the effect of cell density on the formation of titan-like cells , suggesting that this transition is regulated by quorum sensing ( QS ) phenomena . In this way , a higher cell density results in a higher concentration of released molecules and greater effect on the cells [29 , 60 , 61] . A major QS signal for C . neoformans is a 11-mer peptide ( Qsp1 ) [31 , 32] . We tested the effect of this peptide on titan-like cell formation , and found that it inhibited the transition . Qsp1 promotes cell division and replication . Although further studies are required to understand the molecular mechanism by which Qsp1 regulates this transition it is plausible to propose that , since titan-like cells are formed in the absence of budding , Qsp1 blocks this development due to its positive effect of cell division and replication . The biological role of QS phenomena on titan-like cells is not known . However , it is worth noting that a cell-dose relationship similar to that found here has been described in vivo , since a low number of cryptococcal CFUs in the lung correlates with a higher proportion of titan cells [21] . Therefore , it will be necessary to elucidate the role of Qsp1 in the context of lung colonization in the future . However , this peptide does not seem to be the only factor that represses cellular growth at high densities , since qsp1 mutants form titan-like cells similarly to the wild type strain . In agreement with this notion , C . neoformans produces QS molecules that are not susceptible to high temperature , proteinase , trypsin , pronase , DNAse , RNAse and glucosidase [30] . These authors also described that farnesol , tyrosol and Qsp1 did not replicate the effects observed with their conditioned media , and demonstrated that pantothenic acid could in part reproduce QS phenomena . In summary , we hypothesize that several QS molecules negatively regulate titan-like cell formation . The formation of titan cells has been observed in clinical samples [62–64] , and despite being a mechanism that confers advantages to C . neoformans against the host during infection , we observed that this process is not a universal phenotype . Our results showed that there is great variability among different isolates , and not all the strains did form titan-like cells in vitro . Our work provides a new way to investigate the genetic differences between strains with high and low capacity to form titan-like cells using genomic approaches . However , the strains of C . neoformans var . grubii ( serotype A ) exhibited the highest proportion of titan-like cells , suggesting that this serotype has a greater capacity of adaptation to the lung . This result is in agreement with the literature , since this serotype is the one most frequently isolated in infected patients [10] , suggesting that there is a correlation between the ability to form titan cells and development of the disease . We believe that this correlation should be confirmed in future clinical studies . Previous reports demonstrate that during capsule growth , the size of this structure correlates with cell body size [65] . In addition , capsule growth mainly occurs in G1 [66] , which is also the cell cycle phase in which the growth of the cell body occurs . However , although titan-like cells increased capsule thickness , the size of this structure was significantly smaller than the one found in titan cells isolated from animals , indicating that the process that occurs in vitro is , at least quantitatively , not fully equivalent to the one observed in vivo . We argue that the phenomenon that we observed in vitro in this work describes the first steps of titan cell formation . In addition , it is reasonable to argue that when the cell body enlarges significantly , there could be a metabolic limitation affecting the growth of the capsule , since its volume increases with the cube value of the radius . Titan cells contribute to virulence through different mechanisms , such as polarization of Th2-type immune responses [23 , 24] , replication [26 , 67] , resistance to oxidative damage [20 , 21] and phagocytosis avoidance [21 , 25] . In our case , titan-like cells obtained in vitro were not phagocytosed , most probably due to their large size . In our case , these cells did not prevent the phagocytosis of cryptococcal cells of regular size , as it is the case in experiments performed in vivo [25] . We believe that activation of the macrophages in vivo is different from in vitro , so in the lungs there could be multiple factors that affect the phagocytic activity of macrophages , explaining why the effect of titan cells in vivo on macrophages are not identical from the effects induced in vitro . Furthermore , it is also possible that titan cells in vivo express virulence factors or epitopes that are not produced in vitro . Further studies are required to fully understand the response of immune cells to cryptococcal titan-like cells . The isolation of titan-like cells in vitro opens new perspectives and research lines . For example , we investigated gene expression changes associated with titan cells development . In our approach , we compared cells of different size incubated in the same medium ( TCM ) , so the cells were exposed to the same nutritional environment and inducing factors ( CO2 and serum ) . The difference in cell size was obtained due to the different initial cellular density of the cultures ( 104 vs 106 cells/mL ) . In this way , our approach would also detect those changes due to quorum sensing phenomena . Despite the cells were incubated in the same medium , we detected changes in many genes involved in metabolism at both times ( 7 and 18 h ) . This suggests that the increase in cell size requires a metabolic adaptation that allows the efficient use of extracellular and intracellular sources of nutrients . The induction of many genes involved glycolysis and TCA cycle suggest a higher demand of energy , whereas the increase in the expression most of the components of the oxidative branch of the pentose phosphate shunt ( including glucose-6P dehydrogenase , which catalyzes the rate-limiting step ) suggest an increased need for NADPH , a requirement for anabolic biosynthetic pathways . Such requirement could be explained by a higher demand for fatty acid synthesis to provide structural components for an increasing membrane surface . Alternatively , it could be considered that induction of a large number of metabolic enzymes would be simply required to maintain its normal concentration in response to a growing cellular volume . Our results also highlight other processes that might be involved in titan cell formation . In particular , we found that several genes encoding components of COPI and COPII vesicles , which belong to the coatomer complex . These proteins cover the surface of vesicles involved in intracellular protein sorting and early secretory pathway . Interestingly , we found that clathrin- encoding genes were also induced . Although we do not know how these proteins contribute to cryptococcal cellular enlargement , we argue that titan cells might recycle , import and export proteins at a higher rate than regular cells . Interestingly , we also found that the expression of genes encoding some chaperones , such as calnexin , involved in folding of glycosylated proteins increased in titan-like cells , and mutants lacking this protein have an abnormal large size , even in regular rich medium . This result was somehow unexpected , because the expression of this gene increased in titan-like cells . In consequence , we argue that calnexin plays a role , not only in proper folding of glycosylated proteins , but also regulates cell size in conditions of active protein synthesis and replication . In addition , it is possible that it absence results in an stress signal that results in cell cycle changes and G1 arrest , that would produce an increase in cell size too . In this sense , mutants lacking Atg7 ( which is part of the ubiquitin-activating enzyme ( E1 ) family and is required for proper authophagy [68] ) have also increased cell size and higher DNA content , supporting the idea that titan formation is affected by the trafficking and recycling of intracellular proteins . Our results indicated that iron depletion from the medium can also regulate the induction of titan-like cells . Interestingly , iron limitation also induces the growth of the cryptococcal capsule [69] . Iron is an important cofactor required for multiple enzymatic activities , so its absence is an important stress signal that results in the expression different acquisition mechanisms , involving siderophores , Cig1 and transporters [70] . Our findings are in agreement with the idea that titan-like cell formation is a response to nutritional stress . Titan cells isolated from mice are polyploid , so presumably the factors and signaling pathways must alter the normal progression of cell cycle . Interestingly , we found that in titan-like cells , genes related to DNA replication were repressed . In agreement , G1/S cyclins , like Pcl1 , were repressed , which suggests that the process requires an elongation of the G1 phase . This result is in agreement with previous findings that demonstrated that pcl1 mutants have increased capacity to form titan-like cells in vivo [27] . In our titan-like cells , we found that they have a higher content of DNA compared to cells of regular size . However , the DNA content of titan-like cells is lower than the content in titan cells isolated from mice , indicating that cell cycle changes that occur in vitro are different from in vivo . Finally , we would like to acknowledge that other groups ( Dr . Alanio , Pasteur Institute , France , and Dr . Ballou , University of Birmingham , UK ) have identified in parallel to our work other conditions in which titan-like cells are formed . We believe that these works together with our findings provide significant advances to the scientific community because they will allow the design of multiple research lines that will facilitate the characterization of the factors and signaling pathways that are involved in titan cell development . The data presented by different groups also indicate that C . neoformans may induce titan-like cells in vitro in response to multiple factors . In addition , the ability to obtain titan cells in vitro will also have a positive bioethical impact because we will be able to significantly reduce the number of experimental animals compared to what was required previously to characterize these cells .
In most of the experiments , we used C . neoformans var . grubii ( serotype A ) H99 strain [71] , but we also included C . neoformans var . neoformans ( C . deneoformans , serotype D ) , A/D hybrids and C . gattii , and different mutants obtained from the library described by Liu and coworkers [41] and from the Fungal Genetic Stock Centre . All strains are described in Table 3 . Strains were preserved in Sabouraud medium containing 30% glycerol at -80% and were recovered at 30°C in Sabouraud solid medium ( Oxoid LTD , UK ) . In order to delete the CNE1 ( CNAG_2500 ) , a gene a disruption cassette was constructed by overlapping PCR [72] using PCR fragments amplified from either the plasmid pHYG ( kindly given by Prof . Jennifer Lodge , Washington University in St Louis , USA ) or genomic DNA extracted from the strain KN99α [73] . The C . neoformans strain KN99α was then transformed by biolistic delivery [74] and transformants were screened on YPD medium containing 200 U ml-1 of hygromycin ( Calbiochem ) . The correct deletion of the gene was tested using primers within and outside the cassette . Finally , the absence of additional ectopic integration of the cassette was checked by Southern blot experiment . Two independent cne1Δ::HYG strains ( NE375 and NE376 ) were selected for further studies . The primers used in these PCR experiments are listed below CNE1seroAex CCATCTCTTCTTCGGAATCCG CNE1seroA-5’5 TAGCACTGTGAATCGATCCCG CNE1seroA-5’3 GTCATAGCTGTTTCCTGGGATGGGATGAATGGAAGACG MKRrCNE1seroA CGTCTTCCATTCATCCCATCCCAGGAAACAGCTATGAC CNE1seroA-3’5 TACAACGTCGTGACTGGGGAGATTCCTGCTGAAGGCTCG MKRfCNE1seroA CGAGCCTTCAGCAGGAATCTCCCCAGTCACGACGTTGTA CNE1seroA-3’3 TGTTACGTTCGACTTGACGCTG CNE1seroAex2 ACAACGCTTCGACATCTGCAG The yeasts were routinely grown in liquid Sabouraud medium at 30°C or 37°C with shaking ( 150 r . p . m . ) . To induce titan-like cells , strains were grown in a medium that we have defined as Titan Cells Medium ( TCM ) , which is based in the medium described to induce capsule growth ( 10% Sabouraud buffered at pH 7 . 3 with 50 mM MOPS , [75] ) . TCM contains 5% Sabouraud and 5% inactivated fetal calf serum ( FCS , Biological Industries ) diluted in MOPS 50 mM at pH 7 . 3 plus 15 μM sodium azide ( Sigma Aldrich ) . Cultures were grown in tissue culture flasks or 96-wells plates at 37°C in an atmosphere enriched with CO2 for 18 hours . For iron depletion different concentrations of bathphenanthroline disulfonate ( BPS , Sigma Aldrich ) were prepared directly in TCM . Cells were then inoculated at 104 or 106 per mL and incubated at 37°C with 5% CO2 for 18 hours . To observe and measure the size of the cells , 10 μL of a cell suspension were mixed with a drop of India Ink drop ( Remel Bactidrop , Lenexa , Kansas ) and observed under a Leica DMI 3000B microscope . Pictures were taken with a Leica DFC 300FX camera using the Leica Application Suite ( Leica Microsystems ) and processed with Adobe Photoshop 7 . 0 ( San Jose , CA ) or ImageJ ( https://imagej . nih . gov/ij/ ) [75] . In some experiments , the permeability index was measured using fluorescently-labeled dextrans as described in [67] . Six to eight weeks-old male mice from C57BL/6J ( in house bred at the National Centre for Microbiology ) were used in all experiments . The animals were kept in ventilated racks at 22–24°C with proper environmental enrichment ( cupboard houses and hollow cylinders ) . Yeast cells were incubated overnight in liquid Sabouraud medium at 30°C , centrifuged at 2830 g , washed and suspended in sterile PBS . The cell density was determined using a TC20 cell counter ( BioRad ) and a suspension of 3 . 3x107 cells/mL was prepared in sterile PBS . The animals were anesthetized with a mixture of ketamine ( Imalgene 1000 , 50 mg/Kg ) and xylazine ( Xilagesic 2% , 20 mg/Kg ) and infected intranasally with 30 μL of the yeast suspension ( 106 cells per mouse ) as previously described [23] . Animals were sacrificed after 14 days of infection by exposure to a high CO2 enriched environment . We excised and homogenized the lungs in 10 mL of sterile water using cell strainers ( 100 μm size pore , BD Falcon ) and a 5 mL syringe plunge in Petri plates . This process disrupts the mammalian cells without significantly affecting the integrity of the cryptococcal cells . Then the cell suspension was centrifuged at 2 , 830 g and washed with sterile water three times to fully break and remove the mouse cells . Finally , the yeasts were suspended in PBS . Polar lipids from serum were obtained as described in [28] . Briefly , aliquots of 1 mL of FBS were shaken with a mixture of chloroform and methanol ( 2:1 ) ( v:v ) for 3 hours at room temperature . The samples were centrifuged for 10 minutes at 2 , 830 gs for phase partitioning . The upper phase was collected in a 1 . 5 mL tube and dried during 1 hour in a SpeedVac concentrator . The pellet was suspended in 200 μL of PBS and conserved at 4°C . To evaluate the effect of serum polar lipids on titan-like cell formation , we performed experiments with different amounts of the extraction solution described above ( 1/40; 1/100 and 1/200 dilution ) in 5% Sabouraud buffered at pH 7 . 3 with 50 mM MOPS and 15 μM sodium azide . As a control , the same medium with PBS was used . In parallel , cells were grown in Sabouraud and TCM as growth control and titan-like cell formation respectively . To evaluate the role of phosphatidylcholine ( Sigma Aldrich ) , a 20 mM stock of this phospholipid was prepared in EtOH 100% . Then , 1/10 dilutions were done in distilled water to yield 0 . 1 , 0 . 01 and 0 . 001 mM ( final concentration ) in TCM without serum . Same dilutions of EtOH 100% were added to the medium as control . Yeast cells were inoculated in TCM at 104 cells/mL as detailed above . One hundred seventy μL from the yeast suspension were placed in a 96 well plate and incubated at 37°C with 5% CO2 under a Leica DMI 4000B microscope . Photographs were taken every 3 min using the 20x objective . The videos generated by the Leica software were exported as AVI documents and processed with ImageJ software . In all cases , the videos were assembled with 12 frames per second , so one second of video corresponds to 36 minutes of real time . Time was included in each frame using the Time Stamper plugin from ImageJ . Cell suspensions were prepared at a concentration 106 cells/mL in parallel in Sabouraud with 15 μM sodium azide and TCM . Serial 1/10 dilutions were made up to 103 cells/mL . A volume of 170 μL of these suspensions was incubated in a 96 well plate at 37°C overnight with 5% CO2 without shaking . The plates were observed with a Leica DMI 3000B microscope . Pictures were taken with a Leica DFC 300FX camera using Leica Mycrosystems software . The cell body diameter of thirty to fifty cells was measured with Adobe Photoshop 7 . 0 . Cryptococcus neoformans was inoculated at 104 and 106 cells/mL in TCM as described above and incubated at 37°C in a 5% CO2 enriched atmosphere . After 18 h of incubation , the cultures were centrifuged and the supernatants collected to yield Titan-like Cell Supernatant ( TCS ) and regular cells supernatant ( RCS ) . To evaluate the influence of these supernatants on the titan-like cell formation , C . neoformans cultures inoculated at 104 cells/mL in 96 wells plates were prepared in different growth conditions: 1 ) Fresh TCM medium ( TCM ) , 2 ) Supernatant from cultures of titan-like cells ( TCS ) , 3 ) Supernatant from cultures of cells of regular size ( RCS ) . These conditioned media were mixed with fresh TCM ( 1:1 proportion v/v ) . As control , we carried out a culture in which fresh TCM was diluted with the same volume of distilled sterile H2O . After the inoculation of the different media and mixtures with C . neoformans at 104 and 106 cells/mL and incubation at 37°C in a CO2 incubator for 18 h , the cell size was measured by microscopy as described above . Chemical synthesis of the peptides was done by the proteomic facility of the National Centre for Biotechnology ( CSIC , Spain ) using an Multipep automatic synthesizer ( Intavis , Köln , Germany ) and Fmoc-Amino Acid Wang resins ( Merck , Darmstadt , Germany ) . After release from the resin , the peptides were purified by reverse-phase chromatography in a semipreparative HPLC system ( Jasco , Tokio , Japan ) with a C18 Kromaphase column ( Scharlab , Barcelona , Spain ) . The fractions were analyzed by mass spectrometry and lyophilized until their use . We synthesized peptides described in [32]: Qsp1 ( NFGAPGGAYPW ) , an inactive version of this peptide ( NFGAPGAAYPW ) and a scrambled Qsp1 peptide ( AWAGYFPGPNG ) . The peptides were dissolved in sterile PBS at 1 mM , and their effect on titan-like cell formation was tested at 30 μM and 15 μM in TCM . The samples were incubated for 18 hours at 37°C with 5% of CO2 in 96 wells plates . After the incubation period , the cells were observed by optical microscopy and the body cell sizes were measured . Cryptococcus neoformans cells from H99 strain were cultured overnight in TCM at 104 and 106 cells/mL at 37°C with CO2 without shaking . After the confirmation of titan-like cell formations by optical microscopy , cells were washed with dH2O and fixed with 70% ethanol at 24°C for one hour followed by an overnight incubation at 4°C . After fixing , the cells were washed twice with RNAse A buffer ( 0 . 2 M Tris , pH 7 . 5 , 20 mM EDTA ) and treated with 10 μg/mL of RNAse A for 4 hours at 37°C . After the incubation , cells were washed twice with PBS , suspended in PBS and incubated overnight at 4°C . Next day , the cells were centrifuged and suspended in a 200 ng/mL solution of 4' , 6-diamidino-2-phenylindole ( DAPI ) and incubated in the dark for 10 minutes at room temperature to stain the nucleus . Then , the cells were washed , suspended in PBS and the fluorescence intensity of the nucleus were analyzed by flow cytometry . Cells were examined for cell size by forward scatter parameter ( FCS ) and granularity by side scatter parameter ( SSC ) using the BD LSRFortessa X-20 cytometer ( BD , Bioscience ) . Two populations of titan-like and regular cells were delimited and , in each population , the fluorescence intensity of the DAPI staining in 10 , 000 cells was measured . Data obtained were analyzed with the BD FACSDiva ( BD , Bioscience ) and FlowJo 10 . 4 . 2 ( Tree Star Inc , Ashland , Oregon ) softwares . The nuclei of the C . neoformans H99 titan-like and regular cells stained with DAPI were also observed by conventional fluorescence in a Leica DMI3000B microscope and confocal microscopy using a Leica SP5 confocal microscope . The influence of the PKC pathway in the formation of the titan-like cells was evaluated by the addition of four different inhibitors: calphostin C , staurosporine and bisindolylmaleimide I , and genistein ( all from Calbiochem ) that inhibits tyrosine kinase as a control . For this purpose , 10 μM , 5 μM and 1 μM of calphostin C , bisindolylmaleimide I and Genistein and 0 . 01 μM and 0 . 001 μM of staurosporine . In all cases , the final concentration of DMSO was 0 . 1% , which did not inhibit titan-like cell formation . RAW 264 . 7 macrophages were maintained in DMEM medium supplemented with heat-inactivated 10% fetal bovine serum ( FBS , Hyclone-Perbi ) , 10% NCTC medium and 1% non-essential amino acids ( Sigma-Aldrich , Steinheim , Germany ) . The day before the experiment , the macrophage monolayer was separated from the plate by pipetting and the cells were centrifuged at 1 , 265 g . Macrophage suspensions were prepared at 2 . 5x105 cells/mL . Two hundred μL per well were inoculated into 96 well plates and incubated overnight at 37° C and 5% CO2 . The next day , different types of cells ( titan-like cells obtained in vitro , and cells of regular size ) at a final concentration of 5x105 cells/mL with 5 μg/mL of monoclonal antibody 18B7 [76] were added to the macrophages for 2 hours . Phagocytosis was quantified by two different methods . First , the plates were observed with a Leica DMI 3000B microscope , and the percentage of infected macrophages was determined visually . In some cases , the plate was visualized with a Leica 4000B with a chamber that allowed incubation at 37°C and 5% CO2 , and videos were taken as explained above . Alternatively , we quantified the phagocytosis percentage by flow cytometry . In this approach , we performed phagocytosis assays as above , but using larger volumes in 24-well plates containing 1 mL of medium . To differentiate yeast cells , we used a H99 strain that expresses the green fluorescence protein ( H99-GFP ) [77] . In some experiments , macrophages were exposed for 1 h to titan-like cells ( H99 strain , 1:1 ratio ) with mAb 18B7 . As control , macrophages were also preincubated with the same cells , but without mAb , and also with medium alone ( with and without mAb ) . Then , the wells were washed with fresh medium , and cryptococcal cells of regular size ( H99-GFP , grown overnight in Sabouraud at 30°C ) were added at 1:1 ratio for 2 h . After the incubation , we washed the plates , and separated the macrophages by continuous pipetting . The cell suspensions were washed and suspended in PBS containing 1% FCS . Macrophages were blocked with anti-Fc mAb ( 2 . 4G2 , BD Biosciences , 5 μg/mL ) for 10 min at 4°C . After that , the macrophages were washed with PBS/FCS , and then incubated with an anti-Mac1 mAb ( anti-CD11b/Mac1-PE/Cy7 , BioLegend , 1 μg/mL ) for 20 min at 4°C in the dark . Finally , the cells were washed and suspended in 4% p-formaldehyde prepared in PBS . The cells were analyzed in a FACS Canto cytometer ( Biosciences , California , EEUU ) using FASCDiva software ( versión 6 . 1 ) . The phagocytosis percentage was calculated as followed: ( number of PE-Cy7+/GFP+ ) / ( total number of the PE/ PE-Cy7+ cells ) *100 . Experiments were performed in triplicate in three times on different days . In some experiments , cryptococcal cells were labeled with mAb 18B7 labeled with Alexa-488 [78] at 1 μg/mL to label the capsule , and calcofluor ( 10 μg/mL , Sigma Aldrich ) to label the cell wall . The cells were incubated for 1 h at 37 , and then , they were washed and observed in a confocal SP5 Leica microscope . RNA extraction was performed using Trizol ( TRI Reagent , Sigma Aldrich ) with some modifications . After incubation of the cells , Trizol was added to the samples immediately and maintained in ice . Cells were broken during 5 minutes with FastPrep -24 ( MPTM ) , alternating 20 seconds beating with 1 minute on ice . The RNAs concentration and quality was determined with a Nanodrop 8000 Spectrophotometer ( Thermo scientific ) . RNA samples ( 0 . 1 μg/μL ) were treated with DNase using the DNA-free kit ( Thermo Fisher Scientific ) . Then , the RNA samples were purified using RNeasy Mini Kit ( Qiagen ) . Total RNA samples ( 0 . 5–1 μg ) were treated to remove rRNA using Ribo-Zero Magnetic Kit ( Epicentre , Illumina , San Diego , CA ) according to the manufacturer's instructions . Then , mRNA was processed for library preparation using ScriptSeq v2 RNA-Seq Library Preparation kit ( Epicentre , Illumina , San Diego , CA ) . Libraries were quantified using the QuantiFluor RNA System ( Promega ) and the quality and average size was determined using an Agilent 2100 Bioanalyzer . An Illumina NextSeq 500 High Output ( 400 M reads , 1x75 cycles ) was using for sequencing . The FastQ files generated by the equipment were analyzed for low quality reads and those with Q<30 were removed . Original FastQ documents were uploaded to GEO repository ( series record GSE111400 ) . Mapping was carried out with the Bowtie2 software [79] using the C . neoformans var . grubii H99 genome as reference . The SAM files generated by Bowtie2 were analyzed using the software SeqMonk v0 . 39 . 0 ( Babraham Bioinformatics ) . Three biological replicates for cells treated for 7 h ( comprising 8 . 3 to 9 . 5E+06 reads each ) and two for 18 h ( 6 . 9 and 7 . 9E+06 reads each ) were combined , subjected to the RNA-Seq pipeline and evaluated for differential expression using the DESeq2 algorithm [80] with a FDR of p<0 . 05 . The products of this test were recovered as Annotated Probe Report , which included the coding sequences ( overlapping option ) , and processed in Excel format . To identify functional families significantly overrepresented among the differentially expressed genes , the lists of genes was submitted to Gene Ontology ( GO ) and FunCat analyses at the FungiFun server ( https://elbe . hki-jena . de/fungifun/fungifun . php ) [81] . To this end , unless otherwise stated , a Fisher's exact test with significance level set to 0 . 05 and the Benjamini-Hochberg adjustment method were used . Due to the large number of cryptococcal gene products lacking functional annotation in the databases , a Blastp search was conducted against the proteomes of several different , relatively well-annotated fungi , in order to expand the available functional information on the relevant genes . All the animal procedures were approved by the Bioethical Committee and Animal Welfare of the Instituto de Salud Carlos III ( CBA2014_PA51 ) and of the Comunidad de Madrid ( PROEX 330/14 ) and followed the current Spanish legislation ( Real Decreto 53/2013 ) . Statistical analysis was performed with GraphPad Prism 5 . Before comparison among groups , the normality of each simple was assessed using the Kolmogorov-Smirnov test ( non-normal distribution when p<0 . 1 ) . When normal distribution was assumed , differences were estimated using ANOVA and T-Student . For non-parametric distributions , the Kruskal-Wallis and Mann-Whitney tests were used . Statistical significant is highlighted with asterisks in the figures as follows: p>0 . 05 , not significant ( ns ) ; p<0 . 05 and >0 . 01 ( * ) ; p<0 . 01 and p>0 . 001 ( ** ) ; p<0 . 001 and p>0 . 0001 ( *** ) ; p<0 . 0001 ( **** ) . | Cryptococcus neoformans is a fungal pathogen that has a significant incidence in HIV+ patients in particular , in Sub-saharan Africa , Asia and South America . This yeast poses an excellent model to investigate fungal virulence because it develops many strategies to adapt to the host and evade the immune response . One of the adaptation mechanisms involves the formation of Titan Cells , which are yeast of an abnormal large size . However , research on these cells has been limited to in vivo studies ( mainly in mice ) because they were not reproducibly found in vitro . In this work , we describe several conditions that induce the appearance of cells that mimic titan cells , and that we denominated as titan-like cells . The main factor that induced titan-like cells was the addition of serum to nutrient limited media . This has allowed to easily performing new approaches to characterize several signaling pathways involved in their development . We found that the formation of these cells is regulated by quorum sensing molecules , and that pathways such as cAMP and PKC regulate the process of cellular enlargement . We have also performed transcriptomic analysis , which led to the identification of new genes that could be involved in the process . This work will open different research lines that will contribute to the elucidation of the role of these cells during infection and on the development of cryptococcal disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] | [
"blood",
"cells",
"cell",
"physiology",
"cryptococcus",
"neoformans",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"cryptococcus",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"atmos... | 2018 | Cryptococcus neoformans can form titan-like cells in vitro in response to multiple signals |
Genome-wide association studies ( GWAS ) are now used routinely to identify SNPs associated with complex human phenotypes . In several cases , multiple variants within a gene contribute independently to disease risk . Here we introduce a novel Gene-Wide Significance ( GWiS ) test that uses greedy Bayesian model selection to identify the independent effects within a gene , which are combined to generate a stronger statistical signal . Permutation tests provide p-values that correct for the number of independent tests genome-wide and within each genetic locus . When applied to a dataset comprising 2 . 5 million SNPs in up to 8 , 000 individuals measured for various electrocardiography ( ECG ) parameters , this method identifies more validated associations than conventional GWAS approaches . The method also provides , for the first time , systematic assessments of the number of independent effects within a gene and the fraction of disease-associated genes housing multiple independent effects , observed at 35%–50% of loci in our study . This method can be generalized to other study designs , retains power for low-frequency alleles , and provides gene-based p-values that are directly compatible for pathway-based meta-analysis .
Traditional single-SNP GWAS methods have been remarkably successful in identifying genetic associations , including those for various ECG parameters in recent studies of PR interval ( the beginning of the P wave to the beginning of the QRS interval ) [1] , QRS interval ( depolarization of both ventricles ) [2] and QT interval ( the start of the Q wave to the end of the T wave ) [3]–[5] . Much of this success has relied upon increasing sample size through meta-analyses across multiple cohorts , rather than through the use of novel analytical methods to increase power . One analytical approach , gene-based tests proposed during the initial development of GWAS [6] , has natural appeal . First , variations in protein-coding and adjacent regulatory regions are more likely to have functional relevance . Second , gene-based tests allow for direct comparison between different populations , despite the potential for different linkage disequilibrium ( LD ) patterns and/or functional alleles . Third , these analyses can account for multiple independent functional variants within a gene , with the potential to greatly increase the power to identify disease/trait-associated genes . Despite these appealing properties , gene-based and related multi-marker association tests have generally under-performed single-locus tests when assessed with real data [7] , [8] . A general drawback of methods that attempt to exploit the structure of LD to reduce the number of tests , for example through principal component analysis , is the loss of power to detect low-frequency alleles . Methods that consider multiple independent effects often require that the number of effects be pre-specified [9] , which loses power when the tested and true model are different . Multi-locus tests often have the additional practical drawback of being highly CPU and memory intensive . Several methods use Bayesian statistics to drive a brute-force sum or Monte Carlo sample over models [10] , [11] , but again often restrict the search to one or two-marker associations . In general , the computational costs have made these approaches infeasible for genome-wide applications . The Gene-Wide Significance ( GWiS ) test addresses these problems by performing model selection simultaneously with parameter estimation and significance testing in a computational framework that is feasible for genome-wide SNP data ( see Methods ) . Model selection , defined as identifying the best tagging SNP for each independent effect within a gene , uses the Bayesian model likelihood as the test statistic [12]–[14] . Our innovation is to use gene regions to impose a structured search through locally optimal models , which is computationally efficient and matches the biological intuition that the presence of one causal variant within a gene increases the likelihood of additional causal effects . Models are penalized based on the effective number of independent SNPs within a gene and the number of SNPs in the model , akin to a multiple-testing correction . The Schwarzian Bayesian Information Criterion corrects for the difference between the full model likelihood and the easily computed maximum likelihood estimate [15] . This method has greater power than current methods for genome-wide association studies and provides a principled alternative to ad hoc follow-up analyses to identify additional independent association signals in loci with genome-wide significant primary associations .
The ECG parameters PR interval , QRS interval and QT interval are ideal test cases because recent large-scale GWAS studies have established known positive associations . These traits are all clinically relevant , with increased PR interval associated with increased risk of atrial fibrillation and stroke [16] , and both increased QRS and QT intervals associated with mortality and sudden cardiac death [17]–[20] . We assessed the ability of standard methods and GWiS to rediscover these known positives using data from only the Atherosclerosis Risk in Communities ( ARIC ) cohort , which contributes 15% of the total sample size for QRS , 25% for PR , and 50% for QT ( Table 1 ) . The SNPs were assigned to genes based on the NCBI Homo sapiens genome build 35 . 1 reference assembly [21] . Gene boundaries were defined by the most transcriptional start site and transcriptional end position for any transcript annotated to a gene , yielding 25 , 251 non-redundant transcribed gene regions . Incorporating additional flanking sequence increases coverage of more distant regulatory elements , which increases power , but also increases the number of SNPs tested , which decreases power . Expression quantitative trait loci ( eQTL ) mapping in humans has shown that most cis-regulatory SNPs are within 100 kb of the transcribed region [22] , [23] , with quantitative estimates that of large effect eQTNs ( functional nucleotides that create eQTLs ) are within 20 kb of the transcribed region [24] . We report results for 20 kb flanking regions; the performance ranking is robust to flanking by up to 100 kb ( Table S1 ) . SNPs within these regions are then assigned to one or more genes . Of the approximately 2 . 5 million genotyped and imputed SNPs , about 1 . 4 million are assigned to at least one gene . The median number of SNPs per gene is 43 and the mean is 72 ( Table 1 ) , reflecting a skewed distribution with many small genes having few SNPs . The “gold standard” known positives rely on previously published meta-analyses of PR interval [1] , QRS interval [2] and QT interval [4] , [5] . We first identify gold-standard SNPs having . Any gene within 200 kb of a gold-standard SNP is classified as a known positive , and known positives within a 200 kb window are merged into a single locus , yielding 38 known positive gene-based loci . This procedure was followed to ensure that each association signal results in a single locus as opposed to being split between adjacent loci , which could result in over-counting . The minSNP test uses the p-value for the best single SNP within a gene . The minSNP-P test converts this SNP-based p-value to a gene-based p-value by performing permutation tests within each gene . BIMBAM averages the Bayes Factors for subsets of SNPs within a gene , with restriction to single-SNP models recommended for genome-wide applications [10] . Because the Bayes Factor sum is dominated by the single best term , results for BIMBAM are very similar to minSNP-P . The Versatile Gene-Based Test for Genome-wide Association ( VEGAS ) [25] is a recent multivariate method that sums the association signal from all the SNPs within a gene and corrects the sum for LD to generate a test statistic . The terms summed by VEGAS are asymptotically equivalent to the negative logarithms of the Bayes Factors summed by BIMBAM . LASSO regression , or L1 regularized regression , is a multivariate method that combines sparse model selection and parameter optimzation [26]–[28] , with promising recent applications to GWAS [29] . See Methods for more details . Power calculations used genotypes from the ARIC population to ensure realistic LD . Phenotypes were then simulated for genetic models with one or more causal variants within a gene . GWiS was the best-performing method , with an advantage growing as more independent effects are present ( Figure 1a ) . Theoretically , GWiS should have lower power than single-SNP tests when the true model is a single effect; according to the “no free lunch theorem” , this loss of power cannot be avoided [30] . The performance of GWiS therefore depends on the genetic architecture of a disease or trait: higher power if genes house multiple independent causal variants , and lower power if each gene has only a single causal variant . In practice , the loss of power was so slight as to be virtually undetectable . Of the other methods , minSNP-P and BIMBAM had similar performance that degraded as the true model included more SNPs . The VEGAS test did not perform well , presumably because the sum over all SNPs creates a bias to find causal variants in LD blocks represented by many SNPs and to miss variants in LD blocks with few SNPs . In the absence of LD , with genotypes and phenotype simulated using PLINK [31] , VEGAS performs better ( Figure S1 ) . The LASSO method performed worst . The advantage of GWiS arises in part from better power to detect associations with low-frequency alleles ( Figure 1b ) . GWiS , minSNP-P , and BIMBAM have roughly constant power for a given variance explained , regardless of minor allele frequency . In contrast , both VEGAS and LASSO suffer from a two-fold loss of power when minor allele frequencies drop from 50% to 5% . VEGAS may lose power because these low-frequency SNPs lack correlation with other SNPs , reducing the contribution to the VEGAS sum statistic . The LASSO penalty shrinks the regression coefficient , which may adversely affect SNPs with large regression coefficients that balance low minor allele frequencies . The model size selected by GWiS and LASSO was evaluated by simulation ( Figure 2 ) . These simulations also used the ARIC population to supply realistic LD , with genes selected at random with replacement from chromosome 1 . In chromosome 1 , the number of SNPs in a gene ranges from 1 to over 1000 , and the number of independent effects ranges from 1 to over 100 , similar to the distributions in the genome as a whole ( Figure S2 ) . A subset of SNPs within a gene had causal effects assigned ( “True ” ) , phenotypes were simulated to mimic weak and strong gene-based signals , and then models were selected by GWiS and LASSO . Model selection to retain a subset of SNPs ( “Estimated ” ) was performed both for the full genotype data and for the genotype data with the causal SNPs all removed . GWiS provides a better estimate of the true model size than LASSO , assessed from the of estimated versus true . With causal SNPs kept , for GWiS is substantially higher , 0 . 65 versus 0 . 47 at low power ( Figure 2a , 2c ) and 0 . 81 versus 0 . 60 at high power ( Figure 2b , 2d ) . GWiS also performs better when causal SNPs are removed , 0 . 55 versus 0 . 33 at low power and 0 . 60 versus 0 . 39 at high power . GWiS also provides a conservative estimate of the model size , with the ratio of estimated to true size ranging from a worst-case of 44% ( low power , causal SNPs removed ) to a best-case of 81% ( high power , causal SNPs kept ) over the four scenarios examined . In contrast , LASSO is prone to over-predict the size of the model , with a worst-case of models that are on average 33% too large ( high power , causal SNPs kept , Figure 2d ) . Removing a causal SNP results in GWiS predicting a smaller model , with the ratio of estimated to true dropping from 0 . 55 to 0 . 44 for low power and from 0 . 85 to 0 . 81 for high power . These reductions in model size are highly significant ( p for both , Wilcoxon pair test ) and counter a concern that the absence of a causal variant from a marker set will inflate the model size by introducing multiple markers that are partially correlated with the untyped causal variant . These results demonstrate that the model size returned by GWiS is conservative for causal variants with small effects , and approaches the true model size for causal variants with large effects . We then obtained p-values from GWiS , minSNP , minSNP-P , BIMBAM , VEGAS , and LASSO for the ARIC data . Permutations of phenotype data holding genotypes fixed [32] provided thresholds for genome-wide significance for each method ( Table S2 ) . Due to LD across genes , a strong signal in one gene can lead to a neighboring gene reaching genome-wide significance . This effect is well known , and scoring these as false positives would unduly penalize traditional univariate tests . Instead , neighboring genes reaching genome-wide significance were merged , and overlap ( even partial ) with a known positive was scored as a true positive . GWiS out-performed all other methods in the comparison ( Figure 3 and Table 2 ) . GWiS identifies 6 of 38 known genes or loci as genome-wide significant . In contrast , BIMBAM identifies 5 known positives; minSNP , minSNP-P and VEGAS identify 4; and LASSO identifies 2 . Loci identified by the other methods are all subsets of the 6 found by GWiS . None of the methods produced any false positives at genome-wide significance . Due to the limited size of the ARIC cohort relative to the studies that generated the known positives , no method was expected to find all 38 known loci to be genome-wide significant . Nevertheless , known positives should still rank high among the top predictions of each method , assessed by the ranks of the known positives at 40% recall ( Figure S3 ) . We found that GWiS , minSNP , minSNP-P , BIMBAM , and VEGAS were equally effective in ranking known positives ( Mann-Whitney rank sum p-values for any pairwise comparison ) . LASSO performed below the other methods ( p-value for a pairwise comparison of LASSO to any other method ) . Top associations ( up to 100 false positives ) from each method are provided for PR interval , QRS interval , and QT interval ( Tables S3 , S4 , S5 ) . While our conclusions are based on cardiovascular phenotypes , the results suggest that GWiS will have an advantage when causal genes have multiple effects . When an association is sufficiently strong to be found by a univariate test , GWiS is generally able to identify it . Beyond these association , GWiS is also able to detect genes that are genome-wide significant , but where no single effect is large enough to be significant by univariate tests . The association of QRS interval with SCN5A-SCN10A is a striking example: 4 independent effects are found by GWiS ( p-value = ) but the association is not genome-wide significant by univariate methods ( p-value = for minSNP-P ) ( Figure 4 ) . A common follow-up strategy for single-SNP methods is to search for secondary associations in the same locus as a strong primary association . These results for ARIC together with results above for simulated data ( Figure 2 ) demonstrate that GWiS performs this task well . While this feature is present in previous follow-up methods for candidate loci [11] , [33] , [34] , it is absent from methods generally used for primary analysis of GWAS data . Of the 38 known positives , 20 have GWiS models with at least one SNP ( regardless of genome-wide significance ) , and 7 of these are predicted to have multiple independent effects ( Figure 5 ) . These results suggest that the genetic architecture of ECG traits supports the hypothesis underlying GWiS . Moreover , for QT interval where the power is greatest to identify known positives ( the ARIC sample size is 50% of the GWAS discovery cohorts ) , 5 of the 10 loci identified by GWiS are predicted to have multiple independent effects .
In summary , we describe a new method for gene-based tests of association . By gathering multiple independent effects into a single test , GWiS has greater power than conventional tests to identify genes with multiple causal variants . GWiS also retains power for low-frequency minor alleles that are increasingly important for personal genetics , a feature not shared by other multi-SNP tests . Furthermore , GWiS provides an accurate , conservative estimate for the number of independent effects within a gene or region . Currently there are no standard criteria for establishing the genome-wide significance of a weak second association in a gene whose strongest effect is genome-wide significant . While the number of effects can be provided by existing Bayesian methods [34] , their computational expense has limited their applicability to candidate regions , and they are not routinely used . By providing a computationally efficient alternative to existing methods , GWiS provides a new capability to estimate the number of effects as part of primary GWAS data analysis . Demonstrated effectiveness on real data may lead to more widespread use of this type of analysis . Applied to cardiovascular phenotypes relevant to sudden cardiac death and atrial fibrillation , GWiS indicates that 35 to 50% of all known loci contain multiple independent genetic effects . The test we describe includes a prior on models designed to be unaffected by SNP density , in particular by the number of SNPs that are well-correlated with a causal variant . The priors on regression parameters are essentially uniform , with the benefit of eliminating any user-adjustable parameters . A theoretical drawback is that the priors are improper [35] , [36] . Theoretical concerns are mitigated , however , because improper priors pose no challenge for model selection , and our permutation procedure ensures uniform p-values under the null . Bayesian methods can be computationally expensive . GWiS minimizes computation by evaluating only the locally optimal models of increasing size in a greedy forward search . This appears to be an approximation compared to previous Bayesian methods that sum over all models . Previous Bayesian methods entail their own approximations , however , because the search space must either be truncated at 1 or 2 SNPs , heavily pruned , or lightly sampled using Monte Carlo . Our results demonstrate that the approximations used by GWiS provide greater computational efficiency than approximations used in previous Bayesian frameworks , with no loss of statistical power . GWiS currently calculates p-values , rather than Bayesian evidence provided by other Bayesian methods . If Bayesian evidence is desired , an intriguing alternative to Bayesian post-processing of candidate loci might be to use the Bayes Factor from the most likely alternative model identified by GWiS as a proxy for the sum over all alternatives to the null model . This may be an accurate approximation because , in practice , the Bayes Factor for the most likely model from GWiS dominates all other Bayes Factors in the sum . The GWiS framework , using gene annotations to structure Bayesian model selection , may be applied to case-control data by encoding phenotypes as 1 ( case ) versus 0 ( control ) , a reasonable approach when effects are small . More fundamental extensions to logistic regression , Transmission Disequilibrium Tests ( TDTs ) , and other tests and designs should be possible and may yield further improvements . Moreover , similar gene-based structured searches can be applied to genetic models to include explicit interaction terms [14] . The Bayesian format also permits incorporation of prior information about the possible functional effects of SNPs [37] , [38] , and disease linkage [39] , [40] . Finally , the gene-based p-values provide a natural entry to gene annotations and pathway-based gene set enrichment analysis [41]–[43] .
This research involves only the study of existing data with information recorded in such a manner that the subjects cannot be identified directly or through identifiers linked to the subjects . Known positive associations are taken from published genome-wide significant SNP associations ( p-value ) [1] , [2] , [4] , [5] . Genes within 200 kb of any genome-wide significant SNP are scored as known positives . Finally , genes within 200 kb that are both positive are merged into a single known positive locus to avoid over-counting . The ARIC study includes 15 , 792 men and women from four communities in the US ( Jackson , Mississippi; Forsyth County , North Carolina; Washington County , Maryland; suburbs of Minneapolis , Minnesota ) enrolled in 1987-89 and prospectively followed [44] . ECGs were recorded using MAC PC ECG machines ( Marquette Electronics , Milwaukee , Wisconsin ) and initially processed by the Dalhousie ECG program in a central laboratory at the EPICORE Center ( University of Alberta , Edmonton , Alberta , Canada ) but during later phases of the study using the GE Marquette 12-SL program ( 2001 version ) ( GE Marquette , Milwaukee , Wisconsin ) at the EPICARE Center ( Wake Forest University , Winston-Salem , North Carolina ) . All ECGs were visually inspected for technical errors and inadequate quality . Genotype data sets were cleaned initially by discarding SNPs with Hardy-Weinberg equilibrium violations at p , minor allele frequencies , or call rates . Imputation with HapMap CEU reference panel version 22 was then performed , and all imputed SNPs were retained for analysis , included imputed SNPs with minor-allele frequencies as low as 0 . 001 . These cleaned data sets contributed to the meta-analysis to yield the known positives , and full descriptions of phenotype and sample data cleaning are available elsewhere [1] , [2] , [4] . Regional association plots were generated using a modified version of “make . fancy . locus . plot” [45] . The phenotype vector Y for N individuals is an vector of trait values . The genotype matrix X has N rows and P columns , one for each of P genotyped markers assumed to be biallelic SNPs . For simplicity , the vector Y and each column of X are standardized to have zero mean . A standard regression model estimates the phenotype vector as , where b is a vector of regression coefficients and e is a vector of residuals assumed to be independent and normally distributed with mean 0 and variance . The log probability of the phenotypes given these parameters is ( 1 ) The maximum likelihood estimators ( MLEs ) are and , where denotes the transpose of . The total sum-of-squares ( SST ) is , and the sum-of-squares of the model ( SSM ) is . The sum-of-squares of the errors or residuals ( SSE ) is ( 2 ) A conventional multiple regression approach uses the F-statistic to decide whether adding a new SNP improves the model significantly , ( 3 ) for a model with K SNPs , distributed as under the null . This approach fails , however , when the best SNPs are selected from the much larger number of M total SNPs , because the statistic does not account for the selection process . A model M is defined as the subset of SNPs in a gene with P total SNPs that are permitted to have non-zero regression coefficients . For each gene , GWiS attempts to find the subset that maximizes the model probability , where each of the P columns of X corresponds to a SNP assigned to the gene . In the absence of association , the null model with = 0 usually maximizes the probability , indicating no association . When a model with maximizes the probability , an association is possible , and permutation tests provide a p-value . According to Bayes rule , ( 4 ) The factor is model-independent and can be ignored . The prior probability of the model , , assumes that each of the P SNPs within the gene has an identical probability of being associated with the trait . This probability , denoted f , is unknown , and is integrated out with a uniform prior . The prior is also designed to make the model probability insensitive to SNP density: it should be unaffected if an existing SNP is replicated to create a new SNP marker with identical genotypes . We do this by replacing the number of SNPs within a gene with an effective number of tests , , calculated from the local LD within a gene . Correlations between SNPs make the effective number of tests smaller than the number of SNPs . The model prior based on the effective number of tests is ( 5 ) or for integer values . As the effective number of tests , , whose calculation is described below , is generally non-integer , we use the standard Beta function rather than factorials . The remaining factor in Eq . 4 is ( 6 ) The integration limits and prefactor ensure normalization . We assume that these limits are sufficiently large to permit a steepest descents approximation as in Schwarzian BIC model selection [15] . First , assuming that the genotypes are centered , the genotype covariance matrix is , where indicates matrix transpose as before , and diagonal elements for SNP with allele frequency . Provided that is much greater than each component of , the integral over is approximately ( 7 ) where the sum-squared-error SSE is . Provided that the limit is much greater than the maximum likelihood value , the integral over can be approximated as ( 8 ) where is the standard Gamma function . To avoid the cost of Gamma function evaluations , we instead use the steepest descents approximation , ( 9 ) The log-likelihood is then ( 10 ) As in the BIC approximation , we retain only terms that depend on the model and are of order or greater . Thus we replace by , and . For historical reasons , we also included a factor of in the prior for model size , to yield the asymptotic approximation ( 11 ) The strategy of GWiS is therefore to find the model that maximizes the objective function ( 12 ) The terms involving provide a Bayesian penalty for model performance , but also make this an NP-hard optimization problem . We have adopted two efficient deterministic heuristics for approximate optimization . First is a greedy forward search , essentially Bayesian regularized forward regression , in which the SNP giving the maximal increase to the model likelihood is added to the model sequentially until all remaining SNPs decrease the likelihood . The second is a similar heuristic , except that the initial model searches through all subsets of 2 SNPs or 3 SNPs . We adopted this subset search to permit the possibility that all = 1 models are worse than the = 0 null , whereas a more complex model with or 3 has higher score . In practice , all associations identified by subset selection were also identified by greedy forward search . We therefore used the greedy forward search for computational efficiency . GWiS is designed to select a single model for each gene . An alternative related approach would be to test for the posterior probability of the null model , , against all other models , + + + , using our model selection procedure either to choose the locally best model of each size or to include multiple models ( which could suffer from a systematic bias favoring SNPs in large LD blocks ) . This is in fact the strategy of BIMBAM , which attempts to systematically evaluate all terms up to a given model size . Unfortunately , the number of terms increases exponentially fast with model size , and the brute-force approach does not scale to genome-wide applications . Monte Carlo searches over models have also been difficult to apply genome-wide . Our work suggests that approximations that limit the search for fixed model size can be accurate , and further that the probabilities of models that are too large are expected to decrease exponentially fast , permitting the sum to be pruned and truncated . We have observed in practice that the model with the most likely value of dominates the sum , and similarly for BIMBAM that the single SNP with the best Bayes Factor dominates the sum-of-Bayes-Factors test statistic . These results suggest that the results of a more computationally expensive sum over all models would be largely consistent with the results of GWiS method . Furthermore , the Bayes Factor for the most likely model could provide a proxy for the Bayesian evidence . The effective number of tests is an established concept in GWAS to provide a multiple-testing correction for correlated markers . While the exact correction can be established by permutation tests , faster approximate methods can perform well [46]–[49] . While we use a fast procedure , a final permutation test ensures that p-values are uniform under the null . The method we adopt is based on multiple linear regression of SNPs on SNPs . The genotype vector for each SNP i is standardized to have zero mean . Correlations between all pairs of SNPs i and j are initialized as . Each SNPs weight is initialized to 1 , and the number of effective tests T is initialized to 0 . The SNP i with maximum weight is identified , and the following updates are executed: ( 13 ) This process continues until all weights are equal to zero . When SNPs with maximum weight are tied ( as occurs for the first SNP processed ) , the SNP with lowest genomic coordinate is selected to ensure reproducibility; we have ensured that this method is robust to other methods for breaking ties , including random selection . For simplicity , the correlations are not updated ( the update rule would be ) , which may lead to an overestimate for T . Model selection may therefore have a conservative bias . The p-values are not affected , however , because they are calculated by permutation tests as described below . The effective number of tests implies a trivial renormalization of the model prior , ( Eq . 5 ) , that does not affect the test statistic . Letting be the total number of markers , be the effective number tests , and be the size of the model , our prior gives each model of size the weight . If and are identical , there are models of this size , and the total weight of all models of size is . Since can range from 0 to , the sum is normalized . But when is larger than , the sum of all models of size is , which is . The sum from to is therefore . A normalization of 1 can be recovered by including an overall normalization factor , . The explicit prior for models of size is , which is normalized to 1 . Since is model-independent , it does not contribute to the test statistic . We use two stages of permutation tests: the first stage converts the GWiS test statistic into a p-value that is uniform under the null; the second stage establishes the p-value threshold for genome-wide significance . The first stage is conducted gene-by-gene . We permute the trait array using the Fisher-Yates shuffle algorithm [50] , [51] and use the permuted trait to calculate the test statistics using the same procedure as for the original trait . Specifically , the model size is optimized independently for each permutation , with most permutations correctly choosing = 0 . For S successes ( log-likelihoods greater than or equal to the unpermuted phenotype data ) out of Q permutations , the empirical p-value is S/Q . To save computation , permutations are ended when . Furthermore , once a finding is genome-wide significant , there is no practical need for additional permutations . For gene-based tests ( GWiS , minSNP-P , BIMBAM , and VEGAS ) , the p-value for genome-wide significance depends on the number of genes tested ( rather than the number of SNPs ) , for humans . We therefore also terminate permutations after Q = 1 million trials , regardless of S . In these cases , for purposes of ranking , a parametric p-value is estimated for GWiS as ( 14 ) The first factor is the parametric p-value for the F statistic from the MLE fit , and the second term is the combinatorial factor for the number of possible models of the same size . While these p-values are uniform under the null , the threshold for genome-wide significance requires a second set of permutations . To establish genome-wide significance thresholds , in the second stage we permuted the ARIC phenotype for each trait 100 times , ran GWiS for the permuted phenotypes on the entire genome , and recorded the best genome-wide p-value from each of the 100 permutations . We then combined the results from each trait to obtain an empirical distribution of the best genome-wide p-value under the null . We then estimated the p = 0 . 05 genome-wide significance threshold as the 15th best p-value of the 300 . This procedure was performed for GWiS , minSNP , minSNP-P , LASSO , and VEGAS to obtain genome-wide significance thresholds for each . Since minSNP-P and BIMBAM are both uniform under the null , we used the genome-wide significance threshold calculated for minSNP-P , , for BIMBAM to avoid additional computional cost ( Table S2 ) . The threshold for GWiS is more stringent , , presumably because of the locus merging procedure described below . Changes in the genome-wide significance thresholds of up to 50% would not affect any of the reported results . In a region with a strong association and LD , GWiS can generate significant p-values for multiple genes in a region . A hierarchical version of GWiS is used to distinguish between two possibilities . First , through LD , a strong association in one gene may cause a weaker association signal in a second gene . In this case , only the strong association should be reported . Second , the causal variant may not be localized in a single gene; for example , the best SNP tags are assigned to multiple genes . In this case , the individual genes should be merged into a single associated locus . The hierarchical procedure is as follows . The trimming in step 2 handles the first possibility , a strong association in one gene that causes a weaker association in a neighbor . The rationale for accepting the smallest p-value in step 3 is the case of a single SNP assigned to multiple genes . The merged region will have a less significant p-value than any single gene , and it does not seem reasonable to incur such a drastic penalty for gene overlap . For these tests , SNPs are assigned to gene regions as before . The p-value for each SNP is then calculated using the F-statistic as the test statistic , with empirical p-values from permutation to ensure correct p-values for SNPs with low minor allele frequencies . The minSNP method assigns a gene the p-value of its best SNP . Selection of the best p-value out of many leads to non-uniform p-values under the null . It is standard to reduce this bias by scaling p-values by a Bonferroni correction based on the number of SNPs or number of estimated tests . Instead , we perform gene-by-gene permutation tests using the best F statistic for SNPs within the gene as the test statistic . As with GWiS , if 1 million permutations do not lead to one success , the association is clearly genome-wide significant and we use the Bonferroni-corrected p-value for ranking purposes . The Bayesian Imputation-based Association Mapping ( BIMBAM ) is a Bayesian gene-based method [10] . BIMBAM calculates the Bayes Factor for a model and then averages the Bayes Factors for all models within a gene to obtain a test statistic . Because 1-SNP models were found to have as much power as 2-SNP models , and because 2-SNP models are not computationally feasible for genome-wide analysis , BIMBAM by default restricts its sum to all 1-SNP models within a gene [10] . The Bayes Factor for a single SNP is ( 15 ) The design matrix has first column s and second column equal to the dosages of SNP in the individuals; is the phenotypic mean; ; the matrix is diagonal with diagonal terms ; and contains the regression coefficients . We used the recommended value relative to the phenotypic standard deviation . The test statistic for a gene with SNPs is . As with other methods , we used gene-by-gene permutations to convert this statistic into a p-value that is uniform under the null . Up to 1 million permutations were used , stopping after 10 successes . The sufficient statistics used by BIMBAM are identical to minSNP and minSNP-P , yet we found that the runtime of the public implementation was much slower , taking 270 sec for 1000 permutations of a gene with 135 SNPs across 8000 individuals . By improving memory management and optimizing computations , we improved the timing to 14 sec per 1000 permutations , a 19-fold speed-up . This implementation is included in our Supplementary Materials . The Versatile Gene-Based Test for Genome-wide Association ( VEGAS ) [25] is a recently proposed method that considers the SNPs within a gene as candidates for association study . VEGAS assigns SNPs to each of the autosomal genes using the UCSC genome browser hg18 assembly . The gene boundaries are defined as of the and UTRs . Single SNP p-values are used to compute a gene-based test statistic for each gene and significance of each gene is evaluated using simulations from a multivariate normal distribution with mean 0 and covariance matrix being the pairwise LD values between the SNPs from HapMap Phase 2 . As a result the method avoids permutations in calculating per gene p-values , although permutations are required to obtain the genome-wide significance threshold . LASSO regression is a recent method for combined model selection and parameter estimation that maps L1 regularized regression onto a computationally tractable quadratic optimization problem [26]–[28] . Applications to GWAS are attractive because it is possible to perform model selection on an entire chromosome . We therefore implemented a recent LASSO procedure developed specifically for GWAS [29] . To reduce computational cost , univariate p-values are estimated from parametric tests , and gene-based SNPs with are retained ( we have confirmed that this computational constraint does not lose any known positive associations ) . Incremental model selection was performed by Least Angle Regression [27] using the R lars package [52] . The LASSO parameter was determined using 5-fold cross validation . All genes with at least one SNP selected were identified , and selected genes overlapping other selected genes ( including flanking regions ) were merged into single loci . As suggested previously , we used the Selection Index to rank genes and as the test statistic for a permutation p-value [29] . To obtain the Selection Index , the MLE log-likelihood is calculated for the full model and for a reduced model with a subset of SNPs removed . Twice the log-likelihood difference is interpreted as a statistic , and the Selection Index is defined as the corresponding p-value for a distribution with the number of removed SNPs as the degrees of freedom . Due to the LASSO model selection procedure , the Selection Index is not distributed as a under the null , and permutation tests are used to establish genome-wide significance levels . For each true model size of to 8 , we performed a series of simulations by picking 1000 genes from chromosome 1 randomly with replacement , using genotype data from the ARIC population of approximately 8000 individuals . For each gene , we selected “causal” SNPs that have from regression with other “causal” SNPs within the gene . A gene had to have at least SNPs to be picked for models of size to ensure enough remaining SNPs after the removal of the causal SNPs to permit a model of the true size . We attempted to distribute the total population variance explained , , equally across the SNPs . The covariance matrix for the SNPs calculated from the population is denoted , with understood to be . The coefficient for SNP in the model was set to ( 16 ) which ensures that . The phenotype for an individual with genotype row-vector was then calculated as , with again the population average of and drawn from a standard normal distribution . The power was calculated as ( number of genes that are genome-wide significant ) /1000 , and the error of the estimate was calculated using 95% exact binomial confidence intervals . The p-value thresholds were taken directly from genome-wide permutations ( Table 2 ) . Phenotypes that were used to estimate the model size were generated by assigning each “causal” SNP the same power of 0 . 1 and 0 . 8 . The population variance explained for each SNP was calculated as , in which is the quantile of the standard normal for upper-tail cumulative probability of , and is the quantile for lower-tail probability power . We chose to be , the commonly used genome-wide significance threshold for univariate tests . The effect of SNP is then , in which is the genotype covariance matrix . The simulated phenotypes are then , with drawn from a standard normal distribution . In this test we control for the variance explained by the SNP , not by the gene , and therefore do not rescale the regression coefficients to account for LD . For each ranging from 0 to 10 , we repeated these steps using ARIC genotype data for 100 genes chosen at random from chromosome 1 . Only GWiS and LASSO give model size estimates . GWiS directly reports the model size as the number of independent effects within a gene and LASSO reports the model size as the number of selected SNPs within a gene . We ran both methods using the simulated data with LD . We also tested both scenarios when the causal SNPs were kept or removed from gene . Gene associations were scored as true positives if the gene ( or merged locus ) overlapped with a known association , and as false positives if no overlap exists . Only the first hit to a known association spanning several genes was counted . The primary evaluation criterion is the ability to identify known positive associations at genome-wide significance . The genome-wide significance threshold was determined separately for each method ( see above ) , and no method gave any false positives at its appropriate threshold . A secondary criterion was the ability to enrich highly ranked loci for known associations , regardless of genome-wide significance . This criterion was assessed through precision-recall curves , with precision = TP/ ( TP+FP ) , recall = TP/ ( TP+FN ) , and true positives ( TP ) , false positives ( FP ) , and false negatives ( FN ) defined as a function of the number of predictions considered . Small differences in precision and recall may not be statistically significant . To estimate statistical significance , we performed a Mann-Whitney rank sum test for the ranks of the known associations at 40% recall for GWiS , minSNP , minSNP-P , and LASSO . GWiS runs efficiently in memory and CPU time , roughly equivalent to other genome-wide tests that require permutations ( Table 3 ) . Computational times are greater for real data because real associations with small p-values require more permutations . LASSO required far less computational resources , but also pre-filtered the SNPs and had the worst performance . Genome-wide studies can be finished within around 100 hours . Low memory requirements allow GWiS to run in parallel on multiple CPUs . The GWiS source code implementing GWiS , minSNP , minSNP-P , and BIMBAM is available under an open source GNU General Public License as Supplementary Material , also from the authors' website ( www . baderzone . org ) , and is being incorporated into PLINK [31] . | Genome-wide association studies ( GWAS ) have successfully identified genetic variants associated with complex human phenotypes . Despite a proliferation of analysis methods , most studies rely on simple , robust SNP–by–SNP univariate tests with ever-larger population sizes . Here we introduce a new test motivated by the biological hypothesis that a single gene may contain multiple variants that contribute independently to a trait . Applied to simulated phenotypes with real genotypes , our new method , Gene-Wide Significance ( GWiS ) , has better power to identify true associations than traditional univariate methods , previous Bayesian methods , popular L1 regularized ( LASSO ) multivariate regression , and other approaches . GWiS retains power for low-frequency alleles that are increasingly important for personal genetics , and it is the only method tested that accurately estimates the number of independent effects within a gene . When applied to human data for multiple ECG traits , GWiS identifies more genome-wide significant loci ( verified by meta-analyses of much larger populations ) than any other method . We estimate that 35%–50% of ECG trait loci are likely to have multiple independent effects , suggesting that our method will reveal previously unidentified associations when applied to existing data and will improve power for future association studies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"statistics",
"bayes",
"theorem",
"algorithms",
"genome",
"analysis",
"tools",
"mathematics",
"biostatistics",
"statistical",
"methods",
"biology",
"probability",
"theory",
"computer",
"science",
"genetics",
"genomics",
"genetics",
... | 2011 | Gene-Based Tests of Association |
Heritability of bone mineral density ( BMD ) varies across skeletal sites , reflecting different relative contributions of genetic and environmental influences . To quantify the degree to which common genetic variants tag and environmental factors influence BMD , at different sites , we estimated the genetic ( rg ) and residual ( re ) correlations between BMD measured at the upper limbs ( UL-BMD ) , lower limbs ( LL-BMD ) and skull ( SK-BMD ) , using total-body DXA scans of ∼4 , 890 participants recruited by the Avon Longitudinal Study of Parents and their Children ( ALSPAC ) . Point estimates of rg indicated that appendicular sites have a greater proportion of shared genetic architecture ( LL-/UL-BMD rg = 0 . 78 ) between them , than with the skull ( UL-/SK-BMD rg = 0 . 58 and LL-/SK-BMD rg = 0 . 43 ) . Likewise , the residual correlation between BMD at appendicular sites ( re = 0 . 55 ) was higher than the residual correlation between SK-BMD and BMD at appendicular sites ( re = 0 . 20–0 . 24 ) . To explore the basis for the observed differences in rg and re , genome-wide association meta-analyses were performed ( n∼9 , 395 ) , combining data from ALSPAC and the Generation R Study identifying 15 independent signals from 13 loci associated at genome-wide significant level across different skeletal regions . Results suggested that previously identified BMD-associated variants may exert site-specific effects ( i . e . differ in the strength of their association and magnitude of effect across different skeletal sites ) . In particular , variants at CPED1 exerted a larger influence on SK-BMD and UL-BMD when compared to LL-BMD ( P = 2 . 01×10−37 ) , whilst variants at WNT16 influenced UL-BMD to a greater degree when compared to SK- and LL-BMD ( P = 2 . 31×10−14 ) . In addition , we report a novel association between RIN3 ( previously associated with Paget's disease ) and LL-BMD ( rs754388: β = 0 . 13 , SE = 0 . 02 , P = 1 . 4×10−10 ) . Our results suggest that BMD at different skeletal sites is under a mixture of shared and specific genetic and environmental influences . Allowing for these differences by performing genome-wide association at different skeletal sites may help uncover new genetic influences on BMD .
Bone mineral density ( BMD ) at the femoral neck and lumbar spine [as measured by dual-energy X-ray absorptiometry , ( DXA ) ] , represents the primary diagnostic marker for osteoporosis as it serves as a good predictor of bone strength and fracture risk in adults [1] . Bone strength and fracture risk are influenced by: i ) bone acquisition in childhood , adolescence and young adulthood ii ) the subsequent maintenance of bone mass over the life course and iii ) the progressive loss of bone in later life [2] , [3] . Large-scale genome-wide association studies ( GWAS ) using adult-BMD measured at the femoral neck ( FN ) and lumbar spine ( LS ) have successfully identified variants in 56 loci explaining 4–5% of the phenotypic variance in adult-BMD [4]–[6] . However , it is possible that the genetic variants influencing bone acquisition are different from the ones involved in bone maintenance and bone loss across the life course . Consequently , GWAS using paediatric-BMD measurements have recently been performed with the goal of identifying novel genetic variants primarily associated with bone acquisition , whilst limiting the noise introduced by bone maintenance and bone loss [7] . This approach has resulted in the successful identification of novel BMD associated variants in the WNT16 [7] and Osterix ( SP7 ) loci [8] and it is highly likely that more variants will be discovered as the sample size of these paediatric studies increases . In growing children , changes in bone area create artefacts influencing the reproducibility , comparability and interpretation of DXA measurements . For this reason , regions of interest ( ROI ) containing larger bone areas [i . e . total-body , ( TB ) ] , which are less prone to these artefacts , are preferred for paediatric evaluations of bone health [9] . The skull region is generally excluded from TB-DXA scans as its relative contribution to bone mass is proportionally larger with respect to the rest of the body in children , and its inclusion has been shown to make diagnostic interpretation difficult [10] . However , from a locus discovery perspective , it may be advantageous to partition TB-DXA further into different regions , such as the upper and lower limbs and the skull . This is important if genetic heterogeneity exists in terms of loci differentially affecting BMD at different skeletal sites , or whose effect is greater at some locations than in others . Considering that environmental factors ( i . e . mechanical loading ) influence skeletal sites differently , analysis of skull-BMD may be particularly informative and even provide greater power to identify genetic variants . This is the case given that the skull is less influenced by mechanical loading than appendicular and other axial sites . Further , the skull is frequently affected in monogenic conditions involving the skeleton . For example , craniofacial abnormalities such as thickening of the cranium and skull base are cardinal features of van Buchems disease , Sclerosteosis and other sclerosing bone dysplasias [11] , [12] . In the current study we examined whether genetic factors influence bone mass accrual in a site-specific manner , by performing regional analysis of TB-DXA scans , focussing on the total-body less head ( TBLH ) , lower limb ( LL ) , upper limb ( UL ) , and skull ( SK ) regions . Using genome-wide complex trait analysis ( GCTA ) on participants from the Avon Longitudinal Study of Parents and their Children ( ALSPAC ) , we assessed the proportion of BMD variance explained by common genetic variants , across each sub-region and additionally determined the shared genetic and residual correlation between each sub-region . Subsequently , we performed a genome-wide association ( GWA ) meta-analysis of BMD at each skeletal site in the ALSPAC and Generation R studies and went on to identify factors , which preferentially influence one or more skeletal regions .
Univariate GCTA analysis revealed that common genotyped variants explained a greater proportion of the variance in SK-BMD ( vg = 0 . 51 , SE = 0 . 07 , P = 2 . 0×10−13 ) than LL- ( vg = 0 . 40 , SE = 0 . 07 , P = 8 . 0×10−9 ) or UL- ( vg = 0 . 39 , SE = 0 . 07 , P = 2 . 0×10−8 ) BMD . Higher phenotypic correlations were observed when comparing LL- and UL-BMD than with SK-BMD ( Table 1 ) . Similarly , bivariate GCTA analysis indicated that the strongest genetic correlation was between BMD at the two appendicular sites , whereas the genetic correlations involving SK-BMD were more moderate . The residual correlation between the different sites was in general smaller than the genetic correlation , and was higher for BMD between the appendicular sites than for comparisons involving the skull ( Table 1 ) . Highly similar magnitudes and patterns of residual correlations were obtained for a sensitivity analysis in which BMD at all skeletal sites was corrected for age , gender , weight and height ( Table S1 ) . Genome-wide association meta-analyses were performed on TBLH- , LL- , UL- and SK-BMD , using regional BMD data derived from ∼9 , 395 TB-DXA scans . Detailed population characteristics of the ALSPAC and Generation R cohorts are summarised in Tables S2 and S3 . Summary statistics from each GWAS ( after meta-analysis ) indicated that negligible systematic inflation of test statistics was observed ( META λGC = 1 . 01–1 . 03 ) . In contrast , a marked deviation from the null was observed in the tail of the distribution amongst the lowest observed P-values of the meta-association analyses ( Figure S1 ) . SNPs in thirteen published BMD-associated loci exceeded the genome-wide significance ( GWS ) threshold for association ( P≤5×10−8 , Table 2 ) . They included variants which mapped close to , or within: WNT4 ( 1p36 . 12 ) , WNT16/FAM3C/CPED1 ( 7q31 . 31 ) for all skeletal sites measured , EYA4 ( 6q23 . 2 ) , COLEC10/TNFRS11B ( 8q24 . 12 ) , LIN7C/LGR4 ( 11p14 . 1 ) , PPP6R3/LRP5 ( 11q13 . 2 ) and TNFRSF11A ( 18q21 . 33 ) for SK-BMD , CENPW/RSPO3 ( 6q22 . 32 ) for UL- and SK-BMD , TNFSF11 ( 13q14 . 11 ) and GALNT3 ( 2q24 . 3 ) for UL- and TBLH-BMD . In addition , variants proximal to or within FUBP3 ( 9q34 . 11 ) and KLHDC5/PTHLH ( 12p11 . 22 ) were associated with TBLH- and LL-BMD . Furthermore , a novel signal ( top SNP rs754388 , 14q32 . 12 ) , located within Ras and Rab interactor 3 ( RIN3 ) achieved genome-wide significance after meta-analysis of LL-BMD ( β = 0 . 13 , SE = 0 . 02 , P = 1 . 4×10−10 , Figure 1-I , Table 2 ) and TBLH-BMD ( β = 0 . 12 , SE = 0 . 02 , P = 3 . 0×10−9 , Table 2 and Figure S2 ) . The full list of all genome-wide significant SNPs and regional association plots for each locus and skeletal site are presented in Supplementary Tables S4 , S5 , S6 , S7 and Figures S2 , S3 , S4 , S5 . A followup of 66 independent SNPs at 58 loci , previously associated with BMD [4] , [13] , indicated that 31 loci showed nominal evidence of association ( P<0 . 05 ) with TBLH-BMD , 28 with LL-BMD , 26 with UL-BMD and 26 with SK-BMD ( versus an expectation of 3 . 3 . per phenotype ) ( Table S8 ) . A similar distribution of associations was also observed when a more conservative threshold considering multiple hypothesis testing was adopted that took into account the fact that 66 variants and four phenotypes had been tested ( i . e . α<1 . 9×10−4 ) . Using this threshold nine variants showed evidence of association with TBLH-BMD , seven with LL-BMD , six with UL-BMD and 10 with SK-BMD ( versus an expectation of 0 . 1 per phenotype ) . We note that in all cases where nominal significance was reached , the direction of effect was consistent with previous studies . To ensure that our results were robust to the possible effects of population stratification and our choice of covariates , we performed sensitivity analyses where we either restricted our analysis to white European individuals only , or adjusted for the same set of covariates across all analyses ( i . e . age , gender , height , and weight ) . Similar effect sizes and patterns of association were observed for the top SNPs when adjusting BMD measures of all four regions for age , gender , height and weight ( Model 1a , Table S9 ) and when limiting the GWAS meta-analysis to individuals of European ancestry ( Model 1b , Table S9 ) . In both sensitivity analyses , no additional loci reached the threshold of genome-wide association ( Figure S6 and S7 ) . We assessed the presence of novel secondary association signals at loci that contained genome-wide associated variants . Meta-analysis of conditional association analyses resulted in the attenuation of the majority of our top association signals ( Table S10 , Figures S2 , S3 , S4 , S5 ) , indicating that these loci were not independent from signals previously reported by other BMD GWAS . However , the top signal for SK-BMD ( rs2130604 , β = 0 . 11 , SE = 0 . 02 , P = 3 . 3×10−11 ) , mapping near RSPO3 , but closest to CENPW ( previously known as C6orf173 , 6q22 . 32 , Figure 2A-I , Table 2 ) was only marginally attenuated after conditional analysis ( rs2130604 , β = 0 . 10 , SE = 0 . 02 , P = 7 . 1×10−9 , Figure 2A-II , Table S10 ) . This suggests that rs2130604 is largely independent from the previously reported signal at RSPO3 ( rs13204965 , 6q22 . 32 ) , which was identified in a GWAS of individuals with extremely high or low BMD at the hip [12] and later replicated in the second GEnetic Factors for OSteoporosis Consortium ( GEFOS-II ) BMD meta-analysis [4] . This observation is further supported by low estimates of LD ( r2 = 0 . 14 ) between rs2130604 and rs13204965 . Furthermore , the secondary signal ( after conditional analysis ) reached the estimated significance threshold of association after multiple testing correction ( i . e . P≤7 . 2×10−5 ) . Interestingly , after conditioning rs4418209 ( another SNP in the same locus ) on the published BMD-associated SNP rs13204965 , we observed a marked increase in its evidence of association [ ( β = 0 . 07 , SE = 0 . 02 , P = 1 . 1×10−6 ) before and ( β = 0 . 09 , SE = 0 . 01 , P = 7 . 9×10−10 ) after] , ( Figure 2A-II , Table S10 ) . The rs4418209 variant maps closest to CENPW ( 6q22 . 32 ) and is in moderate LD with the secondary independent signal ( rs2130604 , r2 = 0 . 43 ) and in low LD with the published RSPO3 SNP ( rs13204965 , r2 = 0 . 12 ) . Whilst no other SNPs reached the threshold for declaring genome-wide significance ( P<5×10−8 ) , variants from three loci still yielded suggestive evidence for association ( P<1×10−5 ) after conditional analyses ( Table S10 and Figures S2 , S3 , S4 , S5 ) . They included: i ) KLHDC5/PTHLH ( rs4420311 , 12p11 . 22 ) associated with TBLH- ( β = 0 . 08 , SE = 0 . 016 , P = 7 . 6×10−7 ) and LL-BMD ( β = 0 . 08 , SE = 0 . 016 , P = 1 . 9×10−6 ) , ii ) TNFSF11 ( rs17536328 and rs2148072 , 13q14 . 11 ) associated with TBLH- ( β = 0 . 08 , SE = 0 . 015 , P = 5 . 6×10−7 ) and UL-BMD ( β = 0 . 07 , SE = 0 . 015 , P = 2 . 1×10−6 ) respectively and iii ) LIN7C/LGR4 [rs10160456 , 11p14 . 1 , ( β = 0 . 07 , SE = 0 . 015 , P = 7 . 8×10−6 ) ] with SK-BMD . After conditional analysis , the secondary signal at LIN7C/LGR4 ( i . e . rs10160456 ) mapped closest to CCDC34 and not to LIN7C , the gene closest to the primary signal . All these three loci might represent novel secondary signals as the residual signal reached the predicted locus specific threshold of association after multiple testing correction ( Table S10 ) . However , we cannot exclude that both associations ( i . e . the primary and secondary signals ) could potentially arise from their association with one or more causal variants , which could occur , on the same haplotype background . For example , one such BMD-associated rare variant has recently been identified in LGR4 in Icelandic populations although this mutation appears specific to this population and therefore is unlikely to account for the LIN7C/LGR4 signal we observe [14] . The standardized per allele effect sizes ( β ) of all the top BMD-associated SNPs were compared across three ( SK- , UL- , and LL ) BMD regions to determine if they preferentially influenced one or more skeletal sites ( Table 3 , Figure S8 , S9 , S10 , S11 ) . Effect sizes of the following variants: rs2130604 ( CENPW/RSPO3 , 6q22 . 32 ) , rs3012465 ( EYA4 , 6q23 . 2 ) , rs2450083 ( COLEC10/TNFRS11B , 8q24 . 12 ) , rs10835187 ( LIN7C/LGR4 , 11p14 . 1 ) and rs884205 ( TNFRSF11A , 18q21 . 33 ) appeared to be largest for SK-BMD when compared to UL- and LL-BMD ( Figure S8 ) . Furthermore , differences in the magnitude of the effect were evident when comparing independent genetic variants that occurred in close proximity within a locus , as shown at the CENPW/RSPO3 ( 6q22 . 32 ) and WNT16/FAM3C/CPED1 ( 7q31 . 31 ) loci . Specifically , the independent signal ( rs2130604 , CENPW/RSPO3 , 6q22 . 32 ) associated with SK-BMD [β = 0 . 11 ( CI95: 0 . 08 , 0 . 15 ) P = 3 . 3×10−11] , was not strongly related to LL-BMD [β = 0 . 02 ( CI95: −0 . 02 , 0 . 05 ) , P = 0 . 28] , or UL-BMD [β = 0 . 04 , ( CI95: 0 . 01 , 0 . 07 ) , P = 0 . 02] ( Table 3 , Figure 2A-III ) . In contrast , a neighbouring SNP ( rs1262476 ) primarily associated with UL-BMD appeared to influence BMD across all skeletal sites ( Table 3 , Figure 2B-III ) . Differential patterns of association between SNPs at neighbouring positions were also observed at the WNT16 locus ( Table 3 , Figure 3A–C ) . Effect sizes were largest for UL-BMD at rs2908004 ( WNT16 , 7q31 . 31 , Table 3 , Figure 3A-II ) when compared to SK- and LL-BMD . Interestingly , as compared to LL-BMD , we observed consistently larger effect sizes for rs13223036 and rs798943 ( CPED1 , previously known as C7orf58 ) for SK- and UL-BMD , ( Table 3 , Figure 3B-II and 3C-II ) . To formally determine whether the standardized regression coefficients of each of the above-mentioned variants truly differed across the skeletal sites , we fitted a multivariate normal likelihood model to the raw data in ALSPAC and Generation R ( see Methods ) , and then meta-analysed the results using Fisher's method . Using a conservative threshold ( i . e . α = 5×10−8 ) , we observed robust evidence indicating that i . e . rs13223036 and rs798943 , located at CPED1 exerted strong effects on UL and SK-BMD , when compared to LL-BMD [P = 2 . 01×10−37 and P = 4 . 44×10−36 ( Table3 ) ] , whereas the variant rs2908004 ( WNT16 ) was strongly related to UL-BMD in comparison to BMD at the other sites ( P = 2 . 31×10−14 ) . Several variants at other loci were also suggestive of some degree of skeletal site specificity including EYA4 and LIN7C , although they did not formally meet the criteria for statistical significance ( Table 3 , Figures S8 , S9 , S10 , S11 ) . To elucidate if any of the novel primary and/or secondary signals , identified during the course of this study , were nominally associated with BMD in adults , we performed a lookup of these variants in the publicly available results of the GEFOS-II meta-analysis of hip and spine BMD ( Table S11 ) [4] . The novel RIN3 variant ( rs754388 ) was not associated with femoral neck ( PFN = 0 . 87 ) and lumbar spine BMD ( PLS = 0 . 42 ) . The G allele of the EYA4 variant ( rs3012465 ) , associated with increased SK-BMD ( β = 0 . 13 , SE = 0 . 02 , P = 8 . 3×10−17 ) , but surprisingly showed nominal association with decreased BMD at both the hip ( P = 7 . 1×10−3 ) and spine ( P = 0 . 04 ) . A followup of this variant in a recent published GWAS of 4061 premenopausal women aged 20 to 45 revealed no evidence of association with FN-BMD ( P = 0 . 73 ) [15] . A lookup of the secondary independent SNPs revealed no evidence of a relationship between the TBLH- and LL-BMD-associated KLHDC5/PTHLH variant ( rs4420311 ) and associations with hip or spine BMD ( PFN = 0 . 33 and PLS = 0 . 45 ) in GEFOS-II . Similarly no evidence of association was detected for the SK-BMD-associated variant at CENPW/RSPO3 ( rs2130604: PFN = 0 . 98 and PLS = 0 . 40 ) . Interestingly , the T allele of the CENPW/RSPO3 variant ( rs4418209 ) , which was associated with increased SK-BMD ( β = 0 . 07 , SE = 0 . 02 , P = 1 . 1×10−6 ) , appeared to be nominally associated with decreased hip BMD ( P = 5 . 0×10−3 ) but not spine BMD ( P = 0 . 34 ) . Further inspection revealed that the T allele of rs4418209 was nominally associated with decreased BMD at the TBLH ( β = −0 . 03 , SE = 0 . 02 , P = 1 . 7×10−2 ) , LL ( β = −0 . 04 , SE = 0 . 02 , P = 6 . 5×10−3 ) and UL ( β = −0 . 04 , SE = 0 . 02 , P = 8 . 2×10−3 ) . The T allele of rs17536328 located within TNFSF11 , associated with increased TBLH-BMD , showed nominal evidence of association with increased hip ( P = 0 . 04 ) but not spine BMD ( P = 0 . 87 ) . In contrast , the G allele of an independent TNFSF11 variant ( rs2148072 ) associated with increased UL-BMD was associated with decreased spine BMD ( P = 0 . 05 ) . In addition , the C allele LIN7C/LGR4 variant ( rs10160456 ) associated with increased SK-BMD showed weak evidence of association with increased hip ( P = 0 . 06 ) and spine ( P = 0 . 03 ) BMD . We fine mapped the RIN3 region by imputing common and rare variants using a reference panel from the 1000 Genomes Project and identified a missense variant ( rs117068593 ) that was in strong linkage disequilibrium ( r2 = 0 . 96 ) with the top LL-and TBLH-BMD associated RIN3 variant ( rs754388 ) . The C allele of rs117068593 ( EAF = 0 . 82 ) was associated with increased BMD of the lower limbs ( β = 0 . 13 , SE = 0 . 020 , P = 5 . 97×10−11 ) and total-body less head ( β = 0 . 12 , SE = 0 . 020 , P = 1 . 87×10−9 ) . A search of the SIFT database [16] revealed that the missense variant could negatively affect RIN3 functioning . This prediction was further supported by a search of the Regulome database [17] , which suggested that the missense variant alters the binding of the following transcription factors: EBF1 , EGR1 , SP1 , NFKB1 and POLR2A , in lymphoblastic cell lines . Evaluation of cis-expression quantitative trait loci ( eQTLs ) from primary human osteoblasts using array-based gene expression suggested that variants located within 1MB of RIN3 ( i . e . including variants tagging SLC24A4 , LGMN , GOLGA5 , CHGA and ITPK1 ) were nominally associated with ITPK1 expression ( P = 0 . 04 ) . This observation failed to meet the level of significance after correction for multiple testing . Examination of the temporal pattern of gene expression across osteoblastogenesis , using mouse calvarial derived cells , starting with the pre-osteoblast stage , through to mature osteoblasts revealed that Rin3 , Golga5 and Lgmn , and Iptk1 were expressed in this cell type ( Figure S12 ) . In contrast , Slc24a4 and Chga were not expressed at all in the pre- or mature osteoblast , as determined by RNAseq . A further investigation of the expression profiles of the aforementioned genes in human mesenchymal stem cells [ ( hMSCs ) , differentiated into adipocytes and osteoblasts] and peripheral blood monocytes [ ( PBMCs ) differentiated into osteoclasts] indicated the following: SLC24A4 was not expressed in any of these cell lines when differentiated , GOLGA5 had an intermediate expression level in both differentiating hMSCs and PBMCs and LGMN was immediately upregulated upon differentiation into adipocytes ( 8 fold ) , osteoblasts ( 5 fold ) and osteoclasts [ ( 5 fold ) , Figure S13 and S14] . Moreover , we found that the expression of RIN3 was 2-fold downregulated during the proliferative phase of differentiating PBMCs into osteoclasts ( Figure S13 ) . Finally a comparison of expression profiles across the RIN3 region of illiac bone biopsies derived from 39 osteoporotic and 27 healthy postmenopausal donors revealed one transcript ( i . e . 220439_at , originating from RIN3 ) , that demonstrated reduced expression in the osteoporotic group relative to the control group [P = 2 . 7×10−3 , ( Table S13 ) ] .
This study assessed whether regional analysis of skeletal sites from TB-DXA could be used to estimate the extent to which genetic and environmental factors influence bone mass accrual of differentially loaded skeletal sites ( skull , lower limbs , and upper limbs ) . Point estimates indicated that common SNPs on a commercially available genotyping array , explained a larger proportion of the overall variance of SK-BMD , when compared to BMD measured at the appendicular sites ( i . e . lower and upper limbs ) . These differences potentially reflect differential exposure of each skeletal site to varying environmental stimuli that influence BMD . Specifically the skull , as opposed to appendicular sites , is less influenced by environmental factors , particularly those acting through mechanical loading . To explore this result further , we estimated the residual correlation ( i . e . the proportion of environmental and other sources of variation not tagged by SNPs on the Illumina platform ) across the different skeletal sites and found that whilst the environmental ( and other residual ) factors influencing the appendicular sites were moderately similar to each other , they appeared to be appreciably different from the factors influencing SK-BMD . Taken together , lower vg estimates , coupled with a high residual correlation between the two appendicular sites , may reflect the greater exposure of these sites to loading and muscular stimulation , when compared to the skull . Likewise , estimates of the genetic correlations indicated that the appendicular limbs shared a more similar genetic architecture when compared to the skull , possibly reflecting the composition of bone at each skeletal site and the biological processes that govern their growth and maintenance . For example , appendicular sites consist of broadly equivalent proportions of cortical and trabecular bone . The skull on the other hand is mainly comprised of flat bones , which consist primarily of cortical bone [18] . The developmental processes also differ between long and flat bones , with dermal bones such as the skull vault arising exclusively through intramembranous bone formation , in contrast to long bones , which form through endochondral bone formation involving intermediary formation of cartilage [19] . To further explore the basis for the above-mentioned differences in underlying genetic architecture , GWA meta-analyses of sub-regional TB-DXA data were performed . These analyses helped identifying genetic signals that were associated with one or more skeletal region ( s ) . When comparing the evidence of association for all SNPs ( identified in this effort ) across each skeletal site , our GWA meta-analyses echoed the findings of our GCTA results , supporting the notion that although the underlying genetic architecture influencing BMD appears to be largely similar , it does vary according to skeletal site . The majority of the top SNPs were nominally associated ( P≤0 . 05 ) with BMD across all skeletal sites ( i . e . SNPs at WNT4 , WNT16 , FAM3C , GALNT3 , FUBP3 , KLHDC5/PTHLH , TNSF11 , LIN7C/LGR4 and PPP6R3/LRP5 ) . In contrast , variants near or within CPED1 , COLEC10/TNFRS11B and EYA4 were strongly associated with UL- and SK-BMD , but not LL-BMD . A further variant was identified within TNFRSF11A that appeared to be solely related to SK-BMD . Most notably we observed a novel association between rs754388 ( located within RIN3 ) and LL-/UL-BMD , but not SK-BMD . To the best of our knowledge this is the first GWAS to report an association between RIN3 and BMD . It seems likely that this association reflects a true relationship with BMD as the same RIN3 signal ( as determined by conditional analysis ) has previously been associated with an increased risk of Paget's Disease [i . e . rs10498635-C OR: 1 . 44 , 95%-CI ( 1 . 29–1 . 60 ) P = 3 . ×10−11] [20] . In an attempt to further understand how the genetic variation surrounding RIN3 may influence BMD , we fine mapped RIN3 and identified a missense variant ( rs117068593 ) that was in high LD with our LL-BMD associated SNP . Data mining of SIFT and ENCODE databases suggested a functional role of the missense variant that putatively affects binding of several transcription factors in lymphoblastic cell lines . We further evaluated expression quantitative trait locus ( eQTL ) data from primary human osteoblasts using SNP data from HapMap ( i . e . not including rs117068593 ) and found no substantial evidence that our LL-BMD associated SNPs located at 14q32 . 12 regulated the expression of RIN3 or any of the genes located nearby . However , differential patterns gene expression were detected when comparing RIN3 expression profiles of osteoporotic and healthy individuals . Further , we also observed differential expression during osteoclast differentiation that was not present in osteoblast and adipocyte differentiation processes . Collectively , the aforementioned observations appear to be in line with previous findings that suggest that RIN3 could influence osteoclast activity , especially when considering the prior association of RIN3 with Paget's Disease , a disease driven by osteoclast dysfunction and molecular studies that indicate that RIN3 is involved in vesicular trafficking , a process critical for bone resorption [20] , [21] . Further study is however needed to elucidate the precise role of role of RIN3 in bone metabolism . To further understand the preferential associations of some variants with different skeletal sites , we compared the standardized effect sizes of all the genome-wide significant BMD-associated variants , across each skeletal site using a formal multivariate normal likelihood model . Variants at the CPED1 locus were strongly associated with BMD at the skull and upper limb sites , but not with LL-BMD . Similarly variants at WNT16 were more strongly related to UL-BMD , than to BMD at the other sites . Several other SNPs showed evidence for site specificity including variants at the EYA4 and LIN7C loci that were very strongly related to SK-BMD , although these variants did not surpass our conservative criterion for declaring significant heterogeneity , corroboration is needed from independent studies . Conceivably , differences in the pattern of results across SNPs may have arisen from an artefact of the measurement ( i . e . where sub-regional-specific associations reflect how accurately BMD is measured at each skeletal site ) . However , if the latter were the case , one would expect to observe a consistent pattern of results across all loci ( i . e . the strength of association should be greatest at those sites measured more accurately ) . From our results , this is clearly not the case as evidence of association is sometimes greatest for the skull , whilst for other SNPs evidence is greatest for lower and/or upper limbs . In terms of biological explanations , larger effect sizes of genetic variants that influence SK-BMD possibly reflect their preferential involvement in cortical as opposed to trabecular bone metabolism and/or the involvement of intramembranous ossification vs . endochondral ossification [13] . Certain genetic factors also appeared to influence UL-BMD more strongly than LL-BMD , or vice versa . Since the composition and developmental origin of these two sites is broadly similar , presumably , other explanations are responsible . It is reasonable to think that genetic factors , which we identified , could be acting to alter responses to stimuli that are themselves site-specific . For example , adipose tissue has previously been reported to influence cortical bone of the tibia in preference to the radius [22] . Quantitative SK-BMD measurements have traditionally been ignored by genetic and epidemiological studies as they are thought to be prone to errors such as dental augmentation . Despite these concerns , a study conducted in premenopausal woman found a high correlation between the upper half of the skull ( i . e . cranial vault ) and total skull-BMD ( r2 = 0 . 991 , n = 91 , Age range 19–30 years ) , with a mean difference of −0 . 004 g/cm2 , suggesting that these two measurements of bone mass are similar [23] . We found that paediatric SK-BMD measures are well suited to GWAS , as indicated by the very low P-values obtained at some of the known BMD associated loci ( 10−17 to 10−28 ) despite our relatively small sample size . This observation may reflect the fact that SK-BMD is considerably less subject to environmental influences , such as those acting through mechanical loading . In addition , genetic variants associated with SK-BMD identified in this study may primarily reflect molecular pathways involved in bone mass accrual and growth , in contrast to variants identified from previous adult scans which may be more strongly related to mechanisms involved in bone maintenance and/or loss . Almost all the loci we have identified in this study ( i . e . with the exception of SNPs in RIN3 and EYA4 ) have been associated with BMD at either the hip or the lumbar spine previously . Variants mapping to RIN3 have been implicated in Paget's disease but this is the first time the locus is associated with BMD , and interestingly , the alleles associated with increased BMD are associated with increased risk for the condition . This shows that performing GWAS of BMD at sites other than at the femoral neck ( FN-BMD ) or lumbar spine ( LS-BMD ) can be used to identify loci that exert pleiotropic effects on bone . Potential advantages of examining these additional sites from a locus discovery perspective are that ( i ) genetic variants may exert stronger effects at these sites than at FN-BMD/LS-BMD , and/or ( ii ) the genetic effects may be more apparent at these sites because the effect of environmental noise is minimized . For example , the P-values for skull BMD at several loci ( e . g . variants around CPED1 , EYA4 and LIN7C ) are many orders of magnitude stronger than the corresponding P-values for TBLH-BMD ( see Table S8 ) . Likewise variants in LIN7C were first discovered using a GWAS meta-analysis of lumbar spine that was over five times the size of the present study , and even then only just exceeded the threshold for genome-wide significance [4] , whereas in our study a variant at this locus has P<1×10−16 with SK-BMD . Hence , GWAS of BMD at sites such as the skull could be used to efficiently detect clinically relevant loci that might be more difficult to discover in GWAS of the femoral neck and/or lumbar spine . To further illustrate the value of SK-BMD , we draw attention to rs3012465 , a variant proximal to the eyes absent ( EYA4 ) gene and associated with increased SK-BMD . We show that the signal is analogous to that previously associated with increased volumetric cortical BMD of the tibia ( i . e . C allele of rs271170: β = 0 . 11 , P = 2 . 7×10−12 ) , based on a GWAS in ALSPAC and other young adult cohorts [13] , suggesting that both findings reflect the relationship of the EYA4 locus with cortical bone . However , a look-up in a separate cortical bone site ( i . e . the femoral neck of the hip ) , from a GWAS in older adults , revealed that the BMD-increasing allele at the EYA4 locus was in fact associated with lower BMD for both rs3012465 and rs271170 [4] . Taken together , these findings may reflect an age dependent effect of EYA4 whereby EYA4 contributes to bone accrual in early life , yet maybe influences bone loss in older adults . To test this hypothesis , we followed up these EYA4 variants in a recent GWAS meta-analysis of FN-BMD in 4061 pre-menopausal women aged 20–45 ( as described in Koller and colleagues [15] ) and failed to find any evidence of association with FN-BMD ( P = 0 . 73 ) . These results suggest that the discrepancy in results between GEFOS and the present study is unlikely to be solely due to age , but rather is likely to represent a real difference between skeletal sites . In summary , our strategy of analysing regional paediatric DXA measures of TB-BMD represents a novel approach to dissecting the genetic architecture influencing bone mass accrual and growth at different skeletal sites . Specifically , variants at 13 loci reached genome-wide significance with BMD and several displayed different degrees of association according to skeletal site . Furthermore , we report a novel association between a variant within RIN3 and LL-BMD and note its previous association with risk of Paget's disease . We additionally provide suggestive evidence of allelic heterogeneity at the CENPW/RSPO3 , KLHDC5/PTHLH and LIN7C/LGR4 loci . In conclusion our results provide evidence that different skeletal sites as measured by TB-DXA are to a certain extent under distinct environmental and genetic influences . Allowing for these differences may help to uncover new genetic influences on BMD , particularly those examined in children as involved in bone growth and accrual .
In an attempt to identify a potential functional or regulatory mechanism underlying the association between RIN3 and BMD , a range of bio-informatic and functional analyses were performed . These included: fine mapping the RIN3 locus , data mining Regulome [17] and SIFT [16] databases and performing eQTL analysis on primary human osteoblasts . The expression profiles of RIN3 and neighboring genes: SLC2484 , LGMN , GOLGA5 , CHGA and ITPK1 were also investigated in bone biopsies of healthy and osteoporotic women , in addition to murine and human cell lines that were differentiated into osteoblasts and/or osteoclasts . Methods specific to each analysis are described below . | The heritability of bone mineral density ( BMD ) varies across skeletal sites , reflecting different relative contributions of genetic and environmental influences . To investigate whether the genes underlying bone acquisition act in a site-specific manner , we quantified the shared genetic influences across axial and appendicular skeletal sites by estimating the genetic and residual correlation of BMD at the upper limb , lower limb and the skull . Our results suggest that different skeletal sites as measured by total-body Dual-Energy X-Ray Absorptiometry are to a certain extent under distinct genetic and environmental influences . To further explore the basis for these differences , genome-wide association meta-analyses were performed to identify genetic loci that are preferentially associated with one or more skeletal regions . Variants at 13 loci ( including RIN3 , a novel BMD associated locus ) reached genome-wide significance and several displayed evidence of differential association with BMD across the different skeletal sites in particular CPED1 and WNT16 . Our results suggest that it may be advantageous to decompose the total-body BMD measures and perform GWAS at separate skeletal regions . By allowing for site-specific differences , new genetic variants affecting BMD and future risk of osteoporosis may be uncovered . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"biochemistry",
"rheumatology",
"medicine",
"and",
"health",
"sciences",
"genome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"metabolism",
"bone",
"and",
"mineral",
"metabolism",
"h... | 2014 | Phenotypic Dissection of Bone Mineral Density Reveals Skeletal Site Specificity and Facilitates the Identification of Novel Loci in the Genetic Regulation of Bone Mass Attainment |
Stable rhythmic neural activity depends on the well-coordinated interplay of synaptic and cell-intrinsic conductances . Since all biophysical processes are temperature dependent , this interplay is challenged during temperature fluctuations . How the nervous system remains functional during temperature perturbations remains mostly unknown . We present a hitherto unknown mechanism of how temperature-induced changes in neural networks are compensated by changing their neuromodulatory state: activation of neuromodulatory pathways establishes a dynamic coregulation of synaptic and intrinsic conductances with opposing effects on neuronal activity when temperature changes , hence rescuing neuronal activity . Using the well-studied gastric mill pattern generator of the crab , we show that modest temperature increase can abolish rhythmic activity in isolated neural circuits due to increased leak currents in rhythm-generating neurons . Dynamic clamp-mediated addition of leak currents was sufficient to stop neuronal oscillations at low temperatures , and subtraction of additional leak currents at elevated temperatures was sufficient to rescue the rhythm . Despite the apparent sensitivity of the isolated nervous system to temperature fluctuations , the rhythm could be stabilized by activating extrinsic neuromodulatory inputs from descending projection neurons , a strategy that we indeed found to be implemented in intact animals . In the isolated nervous system , temperature compensation was achieved by stronger extrinsic neuromodulatory input from projection neurons or by augmenting projection neuron influence via bath application of the peptide cotransmitter Cancer borealis tachykinin-related peptide Ia ( CabTRP Ia ) . CabTRP Ia activates the modulator-induced current IMI ( a nonlinear voltage-gated inward current ) that effectively acted as a negative leak current and counterbalanced the temperature-induced leak to rescue neuronal oscillations . Computational modelling revealed the ability of IMI to reduce detrimental leak-current influences on neuronal networks over a broad conductance range and indicated that leak and IMI are closely coregulated in the biological system to enable stable motor patterns . In conclusion , these results show that temperature compensation does not need to be implemented within the network itself but can be conditionally provided by extrinsic neuromodulatory input that counterbalances temperature-induced modifications of circuit-intrinsic properties .
Maintaining neural function at different temperatures is a particularly difficult challenge for the nervous system since all biophysical processes are temperature dependent . General knowledge about how this task is achieved remains rather limited , in particular since all biological processes , including those that govern signal transduction and neuronal excitability , vary substantially in their response to temperature [1] . Central pattern generators ( CPGs ) are a class of neural networks that generate rhythmic activity patterns . CPG activity has to be particularly resilient against perturbations because many CPGs drive vital behaviors such as respiration , swallowing , and locomotion . CPG activity depends on the coordinated interplay between synaptic and cell-intrinsic ionic conductances [2] . Many conductance combinations can give rise to rhythmicity , allowing networks to individually vary in conductance levels while remaining within the permissive conductance space for rhythmic activity . It has been suggested that compensation of temperature perturbations may be achieved by keeping conductance levels within this permissive range via a balanced coregulation of cellular and synaptic properties that result in opposing effects on network output . For example , phase constancy in the pyloric rhythm of crabs over a wide temperature range is accompanied by a balanced change of two opposing conductances ( Ih and IA; [3 , 4] ) . In Aplysia , release of a neuromodulator that modulates muscle contraction drops 20-fold at higher temperatures , but this drop is partially counterbalanced by an increase in modulator efficacy [5] . More recent studies indicate that neuromodulators may contribute to temperature compensation: Thuma et al . [6] show that dopamine modulation can restore muscle force after temperature-induced loss of muscle contractions . This study tests the hypothesis that temperature compensation is conditional and under control of extrinsic neuromodulatory input fibers eliciting compensatory changes that oppose temperature-induced changes in intrinsic conductance levels . For this , we use the well-characterized pyloric ( filtering of food ) and gastric mill ( chewing ) CPGs in the crustacean stomatogastric ganglion ( STG; Fig 1A; [7 , 8] ) , which , like most CPGs , are modulated by well-regulated extrinsic neuromodulatory pathways [9] . The triphasic pyloric motor pattern is driven by a three-neuron pacemaker ensemble ( the single anterior burster [AB] and two pyloric dilator [PD] neurons ) that allows it to be continuously active with and without modulatory input [10] . The phase relationship of the pyloric rhythm is maintained constant over a broad temperature range ( 7°C to 31°C; [3 , 4] ) . In contrast , the gastric mill rhythm is two-phasic , episodic , and driven by half-center oscillations of Interneuron 1 ( Int1 ) and the lateral gastric ( LG ) neuron ( Fig 1A; [11 , 12] ) . Rhythmic gastric mill activity requires modulatory input from descending projection neurons in the commissural ganglia ( CoG; [8 , 13] ) . Modulatory commissural neuron 1 ( MCN1 ) , for example , mediates various sensory responses and elicits a robust gastric mill rhythm [7 , 12] . We provide direct evidence that temperature compensation in the gastric mill network depends on a balanced change of circuit-intrinsic properties with opposing function that are regulated by descending modulatory input from MCN1 . Modest temperature increase by 3°C led to a cessation of CPG activity that was caused by a concomitant increase in leak currents in LG . Dynamic clamp-mediated subtraction of the leak was sufficient to rescue the rhythm and achieve temperature compensation . Compensation was also achieved by stronger extrinsic neuromodulatory input from MCN1 or by augmenting MCN1’s influence with bath application of MCN1’s peptide cotransmitter C . borealis tachykinin-related peptide Ia ( CabTRP Ia ) [15 , 16] . CabTRP Ia release from MCN1 activates a modulator-induced current ( IMI [17] ) in LG , and we show that this current effectively acts as a negative leak current to counterbalance the detrimental effects of the leak increase . Thus , temperature compensation is under extrinsic neuromodulatory control , allowing conditional compensation of rapid temperature influences .
To test the role of circuit extrinsic neuromodulatory inputs for counterbalancing temperature-induced changes on CPG activity , we altered the temperature of the STG motor circuits but kept the CoGs at a constant temperature . The CoGs contain descending projection neurons that provide extrinsic modulatory input to the STG circuits . This approach is fundamentally different from previous studies [3 , 4] in which extrinsic neuromodulatory input from CoG projection neurons as well as STG motor circuits were affected by temperature changes . Here , we thermally isolated the STG circuits from the rest of the nervous system by building a petroleum jelly well around the STG . Extrinsic input fibers such as the descending CoG projection neurons remained mostly unaffected by these temperature changes . However , the axon terminals of some projection neurons have local synaptic interactions within the STG [11] and may show ectopic spike initiation . Temperature effects on these interactions were not investigated in this study . CPG activity in the STG was recorded extracellularly on three different motor nerves containing the axons of pyloric and gastric neurons ( pyloric: PD , LP , and PY on the lvn; gastric: LG on the lgn and DG on the dgn ) . We found that a moderate temperature increase from 10°C to 13°C had distinct effects on the two rhythms ( Fig 1B ) : the pyloric rhythm was resilient to temperature changes and continued its regular activity , while the spontaneous gastric mill rhythm terminated , as can be seen by the sporadic and nonrhythmic activity of the gastric mill neurons LG and DG on the lgn and dgn ( Fig 1B ) . The pyloric rhythm had previously been shown to be “resistant” against temperature perturbations in vivo and in vitro ( although for much larger temperature ranges , up to 26°C or more; [3 , 18] ) in that the phase relationship of the pyloric neurons remains constant while cycle period decreases . In all previous studies , however , temperature affected CPG as well as its input fibers , making it unclear whether extrinsic inputs from other parts of the nervous system are necessary to maintain the rhythm or not . Despite the fact that in our experiments the temperature perturbation exclusively affected the STG , we found similar results for the pyloric rhythm as described previously . In none of our experiments did the pyloric rhythm cease or show any obvious change from its canonical pattern ( see S1 Fig ) . In fact , this was true even when we increased the temperature up to 19°C ( N = 6 ) . To conclude , our results predict that the broad temperature range of the pyloric rhythm is likely to be intrinsic to the STG circuit and that the permissive temperature range of this rhythm is independent of temperature effects on other areas of the nervous system , such as the CoGs . The gastric mill rhythm depends on the activity of upstream modulatory projection neurons in the CoGs [7 , 8] . About 20 CoG projection neurons innervate the STG via the unilateral stomatogastric nerve ( stn , see Fig 1A , [19] ) . Sensory input like olfactory or mechanosensory stimuli [20–22] as well as sensory feedback from proprioceptors [20 , 23 , 24] activates CoG projection neurons and elicits gastric mill activity . Even individual projection neurons can start the gastric mill rhythm in vitro [12] and in vivo [25] . One particularly well-characterized projection neuron is MCN1 [11 , 12 , 25] , a bilaterally symmetric neuron in each CoG with axonal projections to the STG ( Fig 1A ) . To study the mechanism of the temperature-induced breakdown of the gastric mill rhythm , we first decentralized the nervous system to remove the influence of all CoG projection neurons . To initiate a gastric mill rhythm , we then stimulated MCN1 tonically with the lowest frequency eliciting a gastric mill rhythm at 10°C ( = threshold frequency , see Materials and Methods ) . MCN1 was activated by extracellular stimulation of the inferior oesophageal nerve ( ion ) , which contains the axons of only two projection neurons , MCN1 and MCN5 . MCN1 has the lower stimulation threshold of the two and can thus be activated selectively [11 , 12] . For the analysis , we focused on LG since this neuron is part of the core pattern generator of the gastric mill rhythm and has a strong influence on all other gastric mill neurons: if spiking is prevented in LG , the gastric mill rhythm stops [11] . Similarly to the spontaneous gastric mill rhythm , LG activity was rhythmic at 10°C during MCN1 stimulation but became substantially reduced and irregular at 13°C ( Fig 2A ) . In four of ten experiments , LG spiking ceased completely ( Fig 2B ) . This effect was reversible , i . e . , the rhythm returned to its original regularity and strength when temperature was decreased back to 10°C . To quantify temperature effects on LG firing rate , we counted the number of LG bursts in 100 s and the number of LG spikes per burst ( see Materials and Methods ) . We found that at 10°C , LG showed rather regularly spaced action potentials during the bursts followed by relatively long interburst intervals ( Fig 2B ) . In contrast , at 13°C LG was either not active or its firing was erratic or tonic . LG activity was never rhythmic at this temperature ( Fig 2B ) . Concurrently , the number of LG bursts , as well as the number of LG spikes per burst , dropped significantly ( Fig 2C ) . When temperature was returned to 10°C , in all preparations rhythmicity was recovered , and the number of LG bursts and spikes per burst returned to control values ( Fig 2C ) . In these experiments , temperature was changed by ~1°C/min , and measurements were taken at 10°C and 13°C . Given that physiological temperature changes might occur over a longer time period than in our experiments , we performed an additional set of experiments in which we slowly changed the temperature ( 1°C/h ) . The goal of this set of experiments was to examine if homeostatic processes exist that compensate ( slow ) temperature perturbations . Specifically , we kept the temperature at 10°C for 1 h , then slowly increased temperature by ~1°C/h , and recorded LG activity at 10°C and 13°C . However , we found no obvious difference as compared to the faster temperature ramps used earlier: again , there was a significant decrease in the number of LG bursts and spikes per burst ( Fig 2D ) , demonstrating that the cessation of LG rhythmicity is not counterbalanced by homeostatic processes in vitro . Next , we tested if the termination of the rhythm occurs abruptly or in a graded fashion . Like in previous experiments , MCN1 stimulation frequency was determined at 10°C . Stimulation was then stopped , and the temperature was lowered to 8°C . Stimulation was then restarted to elicit rhythmicity , and temperature was continuously increased by 1°C/min until LG firing ceased . We found that the number of LG spikes per burst continuously decreased linearly as temperature was increased ( Fig 2E , green trace ) . In the example shown , all burst activity stopped as temperature reached 12 . 5°C . The linear decrease in LG spike activity was consistent across preparations , and on average LG bursting stopped at 12 . 7 ± 0 . 2°C ( N = 5 ) . Additionally , we tested the permissive temperature range for normal operation of this system by performing experiments with a broader temperature range ( 8°C to 16°C , N = 5 ) . In these experiments , we increased the MCN1 stimulation frequency ( 175% threshold frequency ) to facilitate LG rhythmicity at temperatures above 13°C . Again , we found that the number of LG spikes per burst continuously decreased in a highly linear fashion as temperature was increased ( Fig 2E , purple trace ) . In the example shown , all burst activity stopped as temperature reached 15 . 8°C . On average , this happened at 15 . 9 ± 0 . 4°C ( N = 5 ) . The linear response to temperature changes indicates that the gastric mill CPG is unable to compensate moderate temperature influences . The permissive temperature range , however , increased with higher MCN1 stimulation frequencies . In summary , even a moderate temperature increase led to a consistent disruption of the gastric mill rhythm , which was in stark contrast to the robust behavior of the pyloric rhythm . The pyloric and gastric mill rhythms share the same main function: digestion of food . Proper digestion is vital for the animal’s survival , and it is intuitive to assume that both rhythms are equally important and that mechanisms exist to prevent cessation of both rhythms . For the pyloric rhythm , it has been suggested that physiological temperature compensation is achieved by opposing temperature dependencies of membrane currents ( Ih and IA; [3] ) . The gastric mill rhythm , in contrast , apparently lacks adequate compensation despite the fact that pyloric and gastric mill neurons are located in the same ganglion and comprise comparable ion channels and membrane currents ( e . g . , Ih and IA can be found in pyloric and gastric mill neurons; [26 , 27] ) . To determine what provoked the termination of gastric mill activity , we asked whether intrinsic factors contributed to the observed temperature-induced changes in LG activity . We first compared the intracellular response of LG to temperature changes: we found that LG’s resting potential hyperpolarized at 13°C ( Fig 3A ) , with an average drop of 2 . 67 ± 1 . 34 mV ( 10°C: −67 . 01 ± 3 . 00 mV , 13°C: −69 . 68 ± 3 . 50 mV , N = 13 ) . Hyperpolarization was continuous and linear with temperature increase ( Fig 3A , right ) . Also , LG spike amplitude decreased significantly by 4 . 73 ± 1 . 22 mV ( Fig 3B , 10°C: 17 . 02 ± 5 . 58 mV , 13°C: 12 . 28 ± 4 . 36 mV , N = 13 ) . Next , we looked at the electrical postsynaptic potential ( ePSP ) , which LG receives from MCN1 [12] . For this , MCN1 was stimulated with frequencies that did not elicit gastric mill rhythms , but rather only individual ePSPs ( typically 1 Hz or below ) . Fig 3C shows an example of the change in ePSP amplitude when temperature was increased . On average , ePSP amplitude was reduced by 2 . 47 ± 0 . 54 mV at 13°C ( 10°C: 8 . 73 ± 3 . 06 mV , 13°C: 6 . 26 ± 2 . 52 mV , N = 13 ) . All effects were reversible when temperature was decreased back to 10°C . The MCN1 to LG gap junction has been shown to be voltage sensitive such that more hyperpolarized LG membrane potentials lead to smaller ePSPs . This could have possibly contributed to the diminishment of the ePSP amplitudes ( since the resting potential hyperpolarized ) . However , this effect only leads to an average ePSP amplitude change of 0 . 14 mV/1 mV membrane potential change [12] . The temperature-induced change in ePSP amplitude in our experiments was almost six times larger ( 0 . 82 mV/1 mV ) . Thus , the observed change in LG membrane potential was not sufficient to explain the diminished ePSP amplitudes . The changes in resting potential , spike , and ePSP amplitude indicated that the input resistance of LG might have changed , causing a shunt of all of LG’s responses . We found that input resistance decreased significantly by 4 . 12 ± 1 . 4 MΩ ( 34 . 34 ± 10% , 10°C: 12 . 03 ± 2 . 38 MΩ , 13°C: 7 . 92 ± 2 . 12 MΩ , N = 13 ) when temperature was increased ( Fig 3D ) . Changes in input resistance can be due to changes in leak currents , voltage-gated currents , or synaptic input . The latter appears unlikely to have contributed , since the STG was decentralized . Decentralization removes most spontaneous activity of descending projection neurons that may cause synaptic input to LG , and it stops the gastric mill rhythm and silences or strongly diminishes the pyloric rhythm . Consequently , because of the lack of STG activity in this condition , synaptic input from other STG neurons was also unlikely to have contributed to the observed change in input resistance . To further reduce STG and projection neuron input to LG , we blocked action potentials with tetrodotoxin ( TTX ) ( 0 . 1 μM; N = 4 ) . The result was the same: we obtained a decrease in input resistance when temperature was increased . Hence , the decrease in input resistance was independent of synaptic input . We also tested a broader range of current amplitudes by injecting 10 s long current pulses into LG , ranging from 1 to 3 nA in both the depolarizing and the hyperpolarizing direction . Fig 3E shows that the voltage deflections of all current steps were smaller at 13°C than at 10°C . This was true for all preparations tested ( N = 5 ) . Fig 3F shows the change in LG membrane potential as a function of the injected current . We noted a difference between LG’s voltage response to current injections in the positive and negative direction . While this has not been reported directly before , it is most likely a result of the aforementioned voltage dependence of the gap junction between LG and MCN1 [12] . Importantly , the resulting skew of LG’s voltage response in the tested current range was small in comparison to the shunting effect of temperature increase . Since many , if not all , processes in the nervous system are temperature dependent , a causal connection between temperature effects on a specific process and the output of a motor circuit is difficult to show . Our data so far show that when temperature increases ( 1 ) leak conductance of LG increases , associated with ( 2 ) a hyperpolarization of LG’s resting potential . To test whether either of these two effects or both could contribute to the termination of the gastric mill rhythm at higher temperature , we first tested the effects of a change in membrane potential . For this , we recorded LG intracellularly and measured the membrane potential at 10°C and 13°C . We then elicited a gastric mill rhythm via MCN1 stimulation at 10°C and hyperpolarized LG to resting potential values obtained at 13°C ( ΔVm = 2 . 88 ± 1 . 55 mV , N = 4 ) . The gastric mill rhythm was not affected by this manipulation . Thus , the observed change in membrane potential at 13°C was not a significant contributor to the termination of the gastric mill rhythm . We next tested whether an increase in leak conductance is sufficient to explain the termination of the gastric mill rhythm by using the dynamic clamp technique [28] . We either added an artificial leak conductance at 10°C or subtracted leak conductance at 13°C . First , we measured LG input resistance and resting potential at 10°C and 13°C and used the difference to calculate the leak conductance increase ( = Δleak , see Materials and Methods ) . We then elicited a gastric mill rhythm at 10°C , and after several gastric mill cycles , we turned the dynamic clamp on and injected the appropriate amount of additional leak ( +Δleak ) . Immediately after the onset of the artificial leak conductance , LG bursting ceased ( Fig 4A ) . Thus , an increase in leak conductance as caused by a temperature increase of 3°C was sufficient to terminate the gastric mill rhythm . Consequently , a reduction of leak conductance at high temperature should also be sufficient to restore the rhythm . We tested this prediction by carrying out the reverse experiment ( Fig 4B ) . We stimulated MCN1 at 13°C with the threshold frequency that was sufficient to elicit a rhythm at 10°C . As seen in our previous experiments , no gastric mill rhythm was elicited at 13°C despite the continuous MCN1 stimulation . We then turned on the dynamic clamp and subtracted the appropriate leak ( −Δleak ) . Immediately , LG regained its spiking ability , and rhythmicity was restored . In two out of the four experiments , LG firing stopped completely when artificial leak was added at 10°C ( Fig 4C ) . In the other half , LG either generated sporadic action potentials or infrequent bursts of a few action potentials with varying interburst intervals . Thus , in all experiments MCN1 stimulation elicited a gastric mill rhythm at 10°C but failed to do so at 13°C ( similar to our previous findings; see Fig 2 ) . When leak was subtracted ( 13°C − Δleak ) , all preparations recovered the rhythm ( Fig 4C , right ) . Consistent with the previous experiments ( Fig 3 ) , adding leak diminished action potential ( AP ) and ePSP amplitudes , while subtracting leak increased them . In summary , our results demonstrate that a temperature-induced increase in leak conductance was sufficient to terminate the rhythm . Accordingly , bursting in LG could be restored at elevated temperatures by adding a negative leak . C . borealis , the animal used for this study , experiences substantial temperature fluctuations in its habitat [29 , 30] . One would thus assume that the nervous system should be able to cope with the small temperature fluctuations we applied in our experiments . To test if there may be mechanisms to compensate for the temperature-induced termination of the gastric mill rhythm in vivo , we implanted extracellular electrodes in intact animals and recorded the main motor nerve ( lvn ) . Recordings typically lasted for several days . We investigated temperature effects on the gastric mill rhythm using two approaches: first , the temperature of the water was changed at a rate comparable to the in vitro experiments ( 1°C/min ) —a velocity that has previously been shown to be sufficiently slow to cause similar changes at the STG somata [18] . Since we were interested in spontaneous gastric mill rhythms , i . e . , rhythms that were independent of artificial stimulation , temperature was only increased after a gastric mill rhythm was present at 10°C . The rhythm was then continuously monitored during the temperature change . Fig 5A shows the rhythm obtained at 10°C and 13°C . In contrast to the in vitro condition , the rhythm persisted in vivo and showed no signs of irregularity . The number of LG spikes per burst declined slightly with increasing temperature ( Fig 5B ) , but the number of LG bursts increased at the same time , which was in stark contrast to the isolated nervous system . Yet , physiological temperature changes might occur over longer periods . Thus , in a second set of experiments , the temperature was kept at 10°C for at least 1 h and then slowly increased at a rate of 1°C/h to 13°C ( similar to Fig 2D ) . We found that in these conditions spontaneous gastric mill rhythms occurred at 13°C ( N = 7 , Fig 5C ) , implying the existence of mechanisms that compensate the temperature-induced changes in the gastric mill circuit in vivo . To mechanistically understand the adaptations rescuing the gastric mill rhythm , we went back to the in vitro preparation . Since there was no apparent compensation within the STG circuit , we focused on one of LG’s modulatory input , namely MCN1 . MCN1 had been the only projection neuron providing input to LG in our experimental setup ( see also [11] ) , and our initial experiments had already indicated that increasing MCN1 stimulation frequency increased the dynamic range of the gastric mill rhythm ( Fig 2E ) . To scrutinize this idea and to test if a temperature-dependent up-regulation of MCN1 projection neuron activity could counterbalance the termination of the gastric mill rhythm , we first determined MCN1 activity at different temperatures . We recorded spontaneous MCN1 spike activity in preparations in which feedback from the pyloric and gastric mill CPGs in the STG was severed to exclude ascending influences on the activity of MCN1 [31] . In contrast to the previous experiments , we now altered the temperature of the CoG . MCN1 activity was recorded extracellularly from the ion stump connected to the CoG . We found that MCN1 activity increased at 13°C . In the example in Fig 6A , MCN1 firing frequency increased by 57 . 59% . Note that the activity of both MCN1 neurons in a given nervous system preparation was analyzed to determine if temperature affects both MCN1 copies similarly . Although MCN1 firing frequency at 10°C was quite variable between the two MCN1 neurons within a given preparation and across animals , in seven of eight preparations , firing frequency of both MCN1 neurons increased at 13°C ( by 51 . 29 ± 26 . 59% , N = 8 , n = 16 , Fig 6B ) . Next , we asked whether the temperature-induced up-regulation in MCN1 firing frequency is sufficient to counterbalance the increase in LG leak conductance and to prevent the termination of the gastric mill rhythm at elevated temperatures . To test this , we went back to the original experimental setup in which we decentralized the STG from all CoG inputs and stimulated the ion on the STG side of the nerve transection to elicit gastric mill rhythms ( Fig 1A ) . Specifically , we stimulated MCN1 at 10°C with threshold frequency , observed the rhythm , and monitored its cessation after increasing the STG temperature to 13°C . We then raised the MCN1 stimulation frequency in 1 Hz steps to mimic the increase in MCN1 firing frequency observed at 13°C . Fig 6C shows that an increase in MCN1 firing frequency from 7 to 10 Hz ( 42 . 85% ) was sufficient to restore the rhythm in this particular example . On average , a 56 . 07 ± 11 . 99% ( N = 10 ) increase in MCN1 stimulation frequency rescued the gastric mill rhythm . This was true although threshold MCN1 stimulation frequencies varied considerably between preparations ( 5–9 Hz at 10°C , 8–15 Hz rescue frequency at 13°C ) . We found a significant decrease in the number of LG bursts and LG spikes per burst at 13°C , but both returned to control values when MCN1 stimulation frequency was increased ( Fig 6D ) . In fact , when we measured the minimum MCN stimulation frequency at different temperatures ( Fig 6E ) , we found that a linear increase in MCN1 frequency of 0 . 96 Hz/1°C was sufficient to rescue the rhythm at increasing temperatures . What is the mechanism that allows MCN1 to rescue the rhythm at 13°C ? Our previous results indicate that subtracting a leak conductance is sufficient to achieve this goal ( Fig 4B ) . Bursting in LG is mainly driven by the release of MCN1’s peptide cotransmitter CabTRP Ia [11] . In the STG , CabTRP Ia is exclusively found in the MCN1 terminals and thus is specific to MCN1 . Like many other modulators in the STG , CabTRP Ia activates a well-characterized voltage-gated cation conductance ( IMI , modulator-induced current; [17] ) . IMI supports membrane potential oscillations because of its inverted bell-shaped voltage-current relationship [32] . Importantly , IMI has recently been suggested to act as a negative leak conductance because of the linear falling edge of its voltage-current relationship [33] . Could the CabTRP Ia-activated IMI be sufficient to rescue the rhythm by counterbalancing the temperature-induced leak increase in LG ? To test this , we bath applied CabTRP Ia ( 1 μM; [11] ) as a means to increase IMI and measured the response of LG during MCN1 stimulation at 13°C . The release concentration and dynamics of CabTRP Ia are unknown , and the effective concentrations of peptide transmitters on STG neurons differ greatly between neuron types [34] . As current responses to peptide modulators also vary substantially from animal to animal [35] , we made no attempt to determine the CabTRP Ia threshold concentration . Rather , and most importantly for our purposes , we used a concentration shown to be effective in activating IMI [15] . We first elicited a rhythm at 10°C , then increased the temperature to 13°C to elicit the termination , and finally applied CabTRP Ia . Fig 7A shows that CabTRP Ia application indeed can restore the rhythm . In the example shown , 7 Hz MCN1 stimulation elicited a gastric mill rhythm at 10°C ( Fig 7A , i ) , but not at 13°C ( ii ) . We then stopped the MCN1 stimulation and applied CabTRP Ia . CabTRP Ia alone never elicited a gastric mill rhythm nor did it cause LG action potentials ( iii ) . CabTRP acts specifically on IMI , a G-protein coupled voltage-dependent inward current [14 , 34] . Hence , to cause sustained LG activity , an additional depolarization of the membrane potential would be required . We noted a consistent small depolarization of the membrane potential ( 2 . 01 ± 0 . 97 mV , N = 8 ) , which is consistent with earlier findings [15] and neuronal release of CabTRP Ia by MCN1 [11] , indicating that the concentration used was within the physiological range used by MCN1 . We also observed subthreshold oscillations in the LG membrane potential as a result of rhythmic disinhibitions from Int1 that were triggered by increased pyloric activity in the presence of CabTRP Ia [17] . When MCN1 stimulation was turned on ( iv ) , however , LG responded immediately to the threshold stimulation frequency and generated rhythmic bursts of action potentials at 13°C . The effects of CabTRP Ia on LG were reversible ( v ) , i . e . , MCN1 stimulation at 13°C with the threshold frequency after CabTRP Ia washout was neither sufficient to elicit LG spikes nor to start a gastric mill rhythm . Across animals ( N = 4 ) , we found that CabTRP Ia application always restored rhythmic LG activity at 13°C when MCN1 was activated with the threshold frequency . Correspondingly , the number of LG bursts and the number of LG spikes per burst first decreased significantly at 13°C ( Fig 7B ) and then increased in the presence of CabTRP Ia . We noted similar but weaker effects in experiments with lower CabTRP Ia concentrations ( N = 4 ) . In two preparations , rhythmic LG activity recovered at 100 nM , and LG firing frequency was not significantly different from the 10°C control . Lower concentrations did not elicit spiking in LG at elevated temperatures in those two experiments . In the other two experiments , the rhythm recovered at 10 nM , but the firing frequency of LG was lower when compared to the 10°C control condition , indicating that 10 nM did not fully recover LG rhythmicity at elevated temperatures . Only with a simultaneous increase in MCN1 stimulation frequency were control values reached . MCN1 activation in all experiments was necessary to elicit the rhythm , independently of whether CabTRP Ia was present or not . Thus , MCN1's additional transmitter release and network effects ( such as activating the LG half-center antagonist Int1 ) were necessary to start the rhythm . In summary , the CabTRP Ia-induced IMI broadened the permissible temperature range of the gastric mill rhythm and allowed LG to generate rhythmic bursts of activity . Our results show that CabTRP Ia is sufficient to counterbalance the temperature-induced leak current in LG and to rescue the rhythm at 13°C , presumably by its known effect on IMI . We tested the dynamics and range of this compensation by using computational models of the gastric mill network with the known connectivity of the circuit ( Fig 1A; [36] ) . Pyloric influences on the gastric mill network were modeled by driving the pyloric pacemaker neuron AB with constant sinusoidal currents . Initially , using values within the physiological range , we set leak and IMI conductances such that the model produced oscillations that were similar to the biological network ( Fig 8A , i ) . To mimic the effects of a temperature increase , we then added an additional leak conductance to the model LG . With twice the amount of leak , rhythmicity was absent , and LG was completely silent ( ii ) . All other parameters were intentionally kept constant . No temperature-dependent changes other than the increase in the leak current ( as indicated by Fig 3F and Fig 4 ) were added since the goal was to test the interplay between leak and IMI conductances rather than the effect of temperature on synaptic and membrane properties other than leak . When we increased the IMI maximum conductance by 50% , rhythmicity was restored ( iii ) , indicating that IMI indeed can counterbalance the leak . To determine whether this was true for a greater range of parameter combinations and to identify the borders of stable network oscillations , we carried out an exhaustive search of 1 , 100 LG models . Model neurons varied in terms of the maximal conductances of leak and IMI currents but were otherwise identical . This allowed us to test the effects of various combinations of leak and IMI conductances on the behavior of the network . Leak and IMI conductance levels were independently varied in steps of 5 nS , starting at 1 nS . To scrutinize model activity , we analyzed rhythmicity and the number of LG spikes and bursts in 100 s simulation time . Models were classified as rhythmic when at least three clearly identifiable bursts were present within 100 s . Bursts were defined as series of action potentials followed by interburst intervals of more than 2 s . We plotted the number of bursts for all 1 , 100 models as a function of leak and IMI conductances ( Fig 8B ) . In blue areas , the model LG neuron was either not spiking or not rhythmically active ( area highlighted with * ) , whereas in yellow and red areas the model output was rhythmically active . For each leak conductance tested , an increase in IMI was able to restore rhythmicity . Increasing leak conductance required larger IMI . Transitions in model activities were abrupt , i . e . , within small changes in conductance levels , models either produced regular bursting or they were silent . This bimodal behavior most likely reflected the properties of the network that elicited oscillations in the first place . Increasing IMI beyond the minimum level for rhythmicity elicited more bursts and sped up the rhythm . The number of LG spikes increased with higher IMI levels and decreased with higher leak ( Fig 8C ) , indicating that the parameter combination of leak and IMI determined not only the rhythmicity of the model but also burst and spike frequency . Only very few models ( five out of 1 , 100 ) with high IMI and low leak produced nonrhythmic activities with a single , high frequency burst ( * in Fig 8B and 8C ) . Here , the spike number within the one burst dropped with higher IMI , indicating that action potentials could no longer be generated because of sodium channel inactivation . The model also indicates how closely tuned conductances in the biological system are . In our experiments , we saw a 52% average increase in LG leak conductance ( = 34% decrease in input resistance , see Fig 3D ) at 13°C . This increase was sufficient to stop the rhythm . At 10°C , IMI must have been sufficiently high to elicit oscillations but low enough that a 52% increase in leak conductance was sufficient to terminate them . This fact allowed us to estimate the maximum IMI conductance for each leak value to identify parameter combinations sufficient to explain the temperature-induced termination observed in the biological system . Although higher IMI values would allow oscillations ( with faster rhythms and more action potentials ) , they would not lead to a termination of the rhythm with the average amount of leak increase that was observed during temperature elevation . The white line in Fig 8B depicts the maximum IMI for all tested leak conductances . Maximum IMI was always close to the threshold for oscillations , indicating that leak and IMI may be closely balanced in the biological system . The analysis of the model also gives an interpretation of the CabTRP Ia experiments: bath application of the modulator led to faster and stronger bursting than in the 10°C control situation ( Fig 7 ) , indicating that IMI increased beyond control values ( a vertical shift from yellow to red areas in Fig 8B ) . The almost linear progression of the threshold between bursting models and silent ones ( transition between blue and yellow in Fig 8B ) indicates that IMI indeed acted mostly in a linear fashion to counterbalance the leak . This suggests that the two conductances may be coregulated when temperature changes in the biological system to enable stable motor patterns . In conclusion , our analysis shows that increases in leak conductance can be counterbalanced by higher IMI to restore network oscillations . This effect , although independent of temperature by itself , seems to be a crucial mechanism to compensate temperature-induced leak conductance changes for the rescue of the gastric mill rhythm . IMI has been suggested to support membrane potential oscillations by acting as a negative leak conductance [33] . It appears that this property can facilitate compensation of the temperature-induced leak increase in LG . In summary , IMI acts as a negative leak to counterbalance changes in leak conductance and to facilitate rhythmicity in neuronal oscillators . In the case of the gastric mill rhythm , IMI is most likely increased by the actions of CabTRP Ia released from MCN1 during higher temperature-induced neuronal activity . This rescues oscillations in the gastric mill CPG and might be responsible for the broader temperature tolerance in vivo .
When body temperature varies , neuronal compensatory mechanisms are crucial to maintain nervous system function . Even in homeotherms , body temperature fluctuates in a daily and monthly fashion , albeit within a small range . This small range of experienced temperature may lead to a small permissible range in which temperature effects can be compensated . This is especially interesting for circuits that drive vital body functions: hyperthermia associated with pathological conditions such as fever and heat stroke can , for example , cause dysfunction of the breathing CPG and induce apnea [49] . In infants , in whom 85% of the body heat loss is accommodated by the head , apnea and sudden infant death syndrome occur more frequently when the head is heavily wrapped during sleep [50] . In fact , pacemaker neurons in the respiratory neural network are temperature sensitive [49] , but only when synaptically isolated , suggesting that synaptic influences may help to counterbalance temperature effects . As with most rhythmic motor systems , the respiratory network is synaptically innervated by descending modulatory pathways [51] . Neuromodulation modifies network and synaptic properties on various time scales and has been shown to be involved in motor pattern selection and sensory functions that underlie behavioral performance . Often , the global presence or absence of a neuromodulator is equivalent to a specific behavioral state . Our study is the first to show that descending neuromodulation can also be crucial in temperature compensation of an oscillatory neural network . Descending modulation activated a cellular property of opposite sign to the temperature-induced intrinsic changes , making the compensation conditional on the activity of extrinsic input . The pyloric rhythm appears to be intrinsically compensated for larger temperature ranges because of network properties whose response properties change similarly with temperature but have opposing effects . Such mechanisms may be optimal for networks that are required to continuously produce rhythmic activity . For episodic pattern generators such as the gastric mill circuit , temperature compensation may rather be achieved by using the already present descending extrinsic innervation of the pattern generators that exert sophisticated modulatory control over the generated activity . Thus , two quite distinct mechanisms , one depending on the characteristics of the individual components of the network and the other emerging from the effects of descending modulatory fibers , can either individually or in combination compensate for temperature changes to maintain the output of a physiological system . Involving the neuromodulatory system may also allow more flexibility in response to temperature challenges: most descending neuromodulatory pathways integrate information from multiple sensory modalities , rendering temperature compensation conditional on sensory and behavioral conditions . This provides the opportunity to adjust type and strength of the compensation , much more so than with the inherent and rather inflexible compensation provided by intrinsic characteristics of the network .
Adult C . borealis were purchased from Ocean Resources ( Sedgwick , Maine ) or Fresh Lobster Company ( Boston , Massachusetts ) and maintained in filtered , aerated artificial seawater at 11°C before use . Animals were anesthetized on ice for 20–40 min . For in vitro experiments , the stomatogastric nervous system was isolated from the animal according to [52] , pinned out in a silicone-lined ( Wacker ) petri dish , and continuously superfused with physiological saline ( 11°C ) . We worked with fully intact and decentralized STNS preparations . In the latter , the STG was separated from the CoGs by transecting the paired ion and son . For in vivo electrode implantation , anesthetized crabs were immobilized in a custom-built holder . Surgery was performed according to previously published protocols [18] . In short , animals were surrounded by ice to maintain anesthesia during surgery . A 3 x 3 cm window was cut into the dorsal carapace to expose the lateral ventricular nerve ( lvn ) . A hook electrode was placed around the lvn , and the surgery site was sealed with Parafilm . For recovery , animals were placed back into the tank for at least 1 day . Neuronal activity was continuously recorded for several days in unrestrained animals . C . borealis , the animal used in this study , is not subject to ethics approval at Illinois State University . While C . borealis is not a protected species , we still adhered to general animal welfare considerations regarding humane care and use of animals while conducting our research . Crabs were delivered from Massachusetts or Maine via UPS Express Over Night Shipping . For transportation , the animals were covered with wet sea grass and cooled on ice as appropriate for the species . After arriving , the crabs were housed for a maximum of 16 d in 12 tanks ( each with a holding capacity of 100 gallons ) at 10°C to 12°C as appropriate for the species . Water quality , salinity , and temperature were monitored daily . We never kept more than six animals in one tank . The crabs were then euthanized using ice , which is a method recognized as acceptable under the AVMA guidelines for euthanasia of aquatic invertebrates . All animals were confirmed dead before use . C . borealis saline was composed of ( in mM ) 440 NaCl , 26 MgCl2 , 13 CaCl2 , 11 KCl , 10 Trisma base , and 5 maleic acid , pH 7 . 4–7 . 6 ( Sigma Aldrich ) . In some experiments , 0 . 1 μm TTX ( Alomone Labs ) or 10 nM to 1 μM CabTRP Ia ( GenScript ) was added to the saline . Solutions were prepared from concentrated stock solutions immediately before the experiment . Stock solutions were stored at −20°C in small quantities . Measurements were taken after 45 min wash in/out . In vitro recordings were performed using standard methods [53–55] . Extracellular signals were recorded , filtered , and amplified with an AM Systems amplifier ( Model 1700 ) . Intracellular recordings were obtained from STG cell bodies using 10–30 MΩ glass microelectrodes ( Sutter 1000 puller , 0 . 6 M K2SO4 + 20 mM KCl solution ) and an Axoclamp 900A amplifier ( Molecular Devices ) in bridge or two electrode current clamp mode . Files were recorded , saved , and analyzed using Spike2 Software at 10 kHz ( version 7 . 11; CED ) . Input resistance was measured using hyperpolarizing current pulses ( 1 nA , 500 ms duration ) . Membrane potential voltage deflections were measured in steady state ( after 500 ms ) . In some experiments , current injections ranging from −3 nA to +3 nA in 0 . 5 nA steps were used . In this case , current pulses were 10 s long , and interpulse intervals were 30 s . To elicit gastric mill rhythms in decentralized nervous system preparations , we extracellularly stimulated the axon of MCN1 in the part of the transected ion that remained connected to the STG ( Fig 1A ) . This reliably and specifically elicited a specific version of the gastric mill rhythm at 10°C . This version of the rhythm has been characterized in detail before [11 , 12 , 25] . Both ions were tonically stimulated for 200 s with the same frequency ( Master-8 stimulator [AMPI] , 1 ms pulse duration ) . The ion contains the axons of two projection neurons ( MCN1 and MCN5 ) . The extracellular activation threshold is lower for MCN1 than for MCN5 [12 , 53] . In all experiments , we confirmed that MCN1 was selectively activated by adjusting MCN1 activation threshold separately for each ion by slowly increasing stimulation voltage until a PSP in LG was obtained ( for details , see [12] ) . In addition , we monitored the activity of the pyloric LP neuron to confirm whether MCN5 had been activated or not . LP is strongly inhibited by MCN5 [56] , and activation of this projection neuron results in a prominent decrease in LP firing frequency . Preparations without discrete activation thresholds for MCN1 and MCN5 were discarded and not used for experiments . The presence of LP PSPs during ion stimulation was also used to confirm that MCN1 activation threshold was still effective at elevated temperatures . Stimulation frequency was increased in 1 Hz steps until a gastric mill rhythm could be observed at 10°C ( = “threshold frequency” ) . In vitro preparations were continuously superfused with physiological saline . To manipulate temperature of the STG , we built a petroleum jelly well around the STG to thermally isolate it from the rest of the nervous system . Temperature inside and outside of the well was controlled independently with two saline superfusion lines , cooled by separate Peltier devices . Temperature was continuously measured close to the STG and CoGs with separate temperature probes ( Voltcraft 300K ) . We selectively altered temperature at the STG between 10°C and 13°C , while the surrounding nervous system was kept constantly at 10°C . In experiments in which the STG was decentralized by transecting ions and sons , the temperature of the whole bath was changed . To evaluate temperature effects on MCN1 activity ( Fig 6A ) , the temperature at the saline surrounding the CoGs was varied . In this case , the CoGs were isolated from the STG by transecting all connecting nerves . MCN1 action potentials were recorded extracellularly from the ion stump still connected to the CoG . The temperature was changed by ~1°C/min unless otherwise mentioned . With intracellular recordings , the temperature was changed by ~1°C/6–7 min to prevent swelling of the neurons . Measurements were taken after 10 min at the target temperature . Dynamic clamp [28] was used to inject artificial leak currents into LG using Spike2 software and two electrode current clamp mode . In each preparation , input resistance and resting potential were measured at 10°C and 13°C to determine temperature-induced changes . Leak conductance was calculated from these measurements and computed in dynamic clamp according to the following: Idyn=Δleak* ( V−E ) , where Idyn is the injected current , Δleak represents the difference in leak conductance between 10°C and 13°C , and V the membrane potential . E was taken as the resting potential at 13°C . E and Δleak were calculated separately for each preparation . The effects of leak and IMI on pattern generating networks were modeled with MadSim [36 , 57] ( freely available for download at http://www . neurobiologie . de and added as supplemental S1 Data ) using standard morphology and passive properties according to [58] . Active membrane properties were implemented according to modified Hodgkin-Huxley equations [59 , 60] . Leak was implemented as instantaneous linear current in the form of I=g* ( V−E ) , with g being the maximum conductance , V the membrane potential , and E the reversal potential . The reversal potential was set to resting potential in all simulations , and g was varied . IMI was implemented as noninactivating current using I=g¯*ap* ( V−E ) , with p = 1 , E = 0 , and varying g¯ . Activation a was calculated using a=11+eV−V0s , with V0 = −40 mV and s = −10 mV . Maximum conductance g¯ was varied . The time constant of activation was set to 50 ms in all models . The model contained the core gastric mill and pyloric networks and was built with MCN1 , LG , Int1 , and AB according to the real network configuration [7] . The MCN1 terminal was modeled as a separated compartment as it receives synaptic feedback from LG [12] . MCN1 was activated with 15 Hz current pulses . AB membrane potential oscillations were elicited with a sinusoidal current of 1 Hz to provide pyloric feedback to Int1 . We used this model to run an exhaustive search with 1 , 100 simulations by altering leak conductance ( gleak ) and IMI conductance ( g¯IMI ) of LG . All other neurons and parameters were left unchanged . gleak and g¯IMI were increased linearly within physiologically realistic values ( gleak= 1 to 106 nS , 22 steps , 5 nS step size; g¯IMI = 1 to 246 nS , 50 steps , 5 nS step size ) . For all models , simulations produced 100 s-long voltage waveforms . Kinetic parameters for the ionic conductances were set to physiologically realistic values . Maximum conductances g¯ of ionic conductances in the model neurons were chosen to achieve a functional ( network ) output . The model and the simulation are provided as S1 Data and can be found at ModelDB database ( Accession Number 184404 ) . The gastric mill rhythm was considered active when LG produced regular bursts in alternation with the DG neuron . Cycle period was defined as the time between the onset of an LG burst and the onset of the next burst . Rhythmicity was defined as a series of busts with interburst intervals of at least 2 s . Bursts were defined as a series of at least four action potentials with interspike intervals ( ISIs ) below 1 s . We refer to termination of the rhythm if either one or all of those criteria were not fulfilled . Pyloric cycle period was determined using PD bursts . Intraburst firing frequency was defined as the number of spikes within a burst minus one divided by burst duration . Mean values were determined from at least ten consecutive cycles of pyloric activity and phase as normalized time during a cycle . Figures were prepared with CorelDraw X3 ( Corel Cooperation ) , Excel 2010 , SigmaStat , and SigmaPlot ( version 11 , Jandel Scientific ) . Color maps were generated with Spike2 . Data used for analyses and figure generation are given as S2 Data . Unless stated otherwise , data are presented as mean ± SD when normally distributed or as box plot ( 25% and 75% quartiles plus fifth/95th percentile , lines = median , diamonds = mean ) for nonparametric data . Alternatively , individual data points for each animal are given . Significant differences are stated as * p < 0 . 05 , ** p < 0 . 01 , *** p < 0 . 001 . | All physiological processes are influenced by temperature . This is a particular problem for the nervous system , as temperature changes can disrupt the well-balanced flow of ions across the cell membrane necessary for maintaining nerve cell function . Possessing compensatory mechanisms that counterbalance detrimental temperature effects and maintain vital behaviors is especially important for poikilothermic animals , because they do not actively maintain their body temperature and can experience substantial temperature fluctuations . In this study , we analyze the mechanisms that allow the nervous system to maintain rhythmic activity over a range of different temperatures . To do so , we use the well-characterized central pattern generator of the stomatogastric nervous system of the crab that controls the motion of the gut . In this system , when experimentally isolated from the rest of the nervous system , even a small temperature increase can lead to termination of rhythmic activity due to a change in the balance of ionic conductances at elevated temperatures . However , the intact animal can compensate for these detrimental temperature effects . We demonstrate that such compensation can be achieved by restoring the balance of ionic conductance via an increase in neuromodulator release from projection neurons that control the motor circuits . We conclude that temperature compensation via neuromodulation may be a widespread phenomenon since it allows quick and flexible compensation of temperature influences on the nervous system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Neuromodulation to the Rescue: Compensation of Temperature-Induced Breakdown of Rhythmic Motor Patterns via Extrinsic Neuromodulatory Input |
Epidemic dengue fever ( DF ) and dengue hemorrhagic fever/dengue shock syndrome ( DHF/DSS ) are overwhelming public health capacity for diagnosis and clinical care of dengue patients throughout the tropical and subtropical world . The ability to predict severe dengue disease outcomes ( DHF/DSS ) using acute phase clinical specimens would be of enormous value to physicians and health care workers for appropriate triaging of patients for clinical management . Advances in the field of metabolomics and analytic software provide new opportunities to identify host small molecule biomarkers ( SMBs ) in acute phase clinical specimens that differentiate dengue disease outcomes . Exploratory metabolomic studies were conducted to characterize the serum metabolome of patients who experienced different dengue disease outcomes . Serum samples from dengue patients from Nicaragua and Mexico were retrospectively obtained , and hydrophilic interaction liquid chromatography ( HILIC ) -mass spectrometry ( MS ) identified small molecule metabolites that were associated with and statistically differentiated DHF/DSS , DF , and non-dengue ( ND ) diagnosis groups . In the Nicaraguan samples , 191 metabolites differentiated DF from ND outcomes and 83 differentiated DHF/DSS and DF outcomes . In the Mexican samples , 306 metabolites differentiated DF from ND and 37 differentiated DHF/DSS and DF outcomes . The structural identities of 13 metabolites were confirmed using tandem mass spectrometry ( MS/MS ) . Metabolomic analysis of serum samples from patients diagnosed as DF who progressed to DHF/DSS identified 65 metabolites that predicted dengue disease outcomes . Differential perturbation of the serum metabolome was demonstrated following infection with different DENV serotypes and following primary and secondary DENV infections . These results provide proof-of-concept that a metabolomics approach can be used to identify metabolites or SMBs in serum specimens that are associated with distinct DENV infections and disease outcomes . The differentiating metabolites also provide insights into metabolic pathways and pathogenic and immunologic mechanisms associated with dengue disease severity .
Epidemic dengue fever ( DF ) and dengue hemorrhagic fever/dengue shock syndrome ( DHF/DSS ) have emerged throughout the tropical and subtropical world with devastating consequences and are overwhelming public health capacity for diagnosis and patient care [1 , 2] . Upon presentation early after disease onset , it is clinically impossible to differentiate dengue virus ( DENV ) -infected patients who will have an unremarkable DF disease episode from those who will progress to potentially fatal DHF/DSS [3–7] . Viral biomarkers that correlate with dengue severity include viremia titer and nonstructural protein 1 ( NS1 ) concentration in the blood , secondary DENV infection , and infection with specific virus genotypes [8–11] . Host biomarkers associated with disease severity include multiple immune molecules , biochemical and physiological response indicators , and genetic polymorphisms [3 , 4 , 12–21] . Algorithms based upon clinical signs and laboratory test results have been proposed to predict dengue severity [22–28] . However , currently there are no standardized biomarkers or algorithms for prognosis of severe disease outcomes . Current diagnostic tests and approaches are not meeting the challenges posed by dengue [29 , 30] . A paradigm shift in diagnosis/prognosis is essential to address the increasing threat of severe dengue disease . Advances in mass spectrometry , metabolite databases , and analytical software provide exciting new opportunities to identify small molecule biomarkers ( SMBs ) of dengue disease outcome in acute-phase serum specimens . Mass spectrometry-based metabolomics techniques are being applied with increasing frequency for diagnosis , investigation of pathogenic mechanisms , and monitoring the effects of treatments and interventions of infectious diseases [31–36] . Metabolomics is the analysis of low molecular weight biological molecules that result from metabolic processes . Disease states result in changes in metabolism in cells and systems that affect the profile of metabolites [34] . Analysis of metabolite profiles in disease conditions and comparison with the profiles of non-diseased individuals can be used in diagnosis . Metabolites that differentiate DF and DHF/DSS outcomes could potentially be exploited as SMBs for diagnosis of DENV infections and prognosis of disease severity . Liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) metabolomics approaches have been used to detect and characterize changing metabolite levels in humans and mosquito vectors that are directly attributable to DENV infection and pathogenesis [32 , 37] . Primary DENV infection in humans was shown to cause temporally distinct changes in the serum metabolome , particularly in the lipidome , reflecting the pathogenic mechanisms and metabolic pathways perturbed during the time course of DF [32] . Here , hydrophilic interaction liquid chromatography ( HILIC ) -MS [38 , 39] was used to characterize retrospectively the serum metabolome of patients who were diagnosed as DHF/DSS , DF , or non-dengue ( ND ) febrile disease , as well as for preliminary characterization of the serum metabolome following infection with two different DENV serotypes and after primary or secondary DENV infection . In this exploratory , proof-of-concept study , metabolites that were associated with and differentiated DF and DHF/DSS , that predicted progression to DHF/DSS in serum of DF patients , and that differentiated infecting DENV serotypes and primary and secondary infections were identified . The differentiating metabolites reflect host responses to the pathogen , including tissue damage , inflammation , and other virus-induced pathology and thus provide insights into fundamental metabolic pathways associated with DENV pathogenesis and potentially novel targets for therapeutic intervention [32 , 35] . We have identified candidate SMBs to be evaluated in prospective clinical studies for their diagnostic and prognostic efficacy for DENV infections .
Serum samples were obtained from collections of sera from patients who had presented with dengue-like febrile disease in Managua , Nicaragua , and Mérida , México . Nicaraguan serum samples had been collected as part of two ongoing pediatric studies being conducted by the University of California , Berkeley , the Nicaraguan Ministry of Health , and the Sustainable Sciences Institute: the Pediatric Dengue Cohort Study , which is focused upon studying transmission of DENV and identifying immune correlates of protection , and the Hospital-based Dengue Study , which is focused on studying clinical , immunological , and viral risk factors for severe DENV infections . Parents or legal guardians of participants provided written informed consent , participants 6 years of age and older provided assent , and participants in the Hospital-based Dengue Study 12 years of age and older provided written assent . These studies were approved by the UC Berkeley Committee for the Protection of Human Subjects ( Protocols # 2010-06-1649 and 2010-09-2245 ) and the IRB of the Nicaraguan Ministry of Health . Mexican samples had been collected in the Laboratorio de Arbovirología , Centro de Investigaciones Regionales Dr . Hideyo Noguchi or the Unidad Universitaria de Inserción Social ( UUIS ) San José Tecoh , both from the Universidad Autónoma de Yucatán ( UADY ) , Mérida , Yucatán , México , from patients who were referred by a primary-care physician for diagnostic testing . These samples were procured as part of the normal dengue diagnosis mission of the laboratory and not as part of an experimental protocol . These samples provided an opportunity to determine if metabolites detected in Nicaraguan patients that differentiated dengue disease outcomes could also be detected in dengue patients with different genetic , environmental , and demographic backgrounds . This research was approved by the Bioethics Committee of the Centro de Investigaciones Regionales “Dr . Hideyo Noguchi” ( CIR ) of the Universidad Autónoma de Yucatán and reviewed by the CSU Institutional Review Board and considered to be an exempt project ( Category 4 ) . A portion of serum samples from Nicaraguan and Mexican patients who were diagnosed as DHF/DSS , DF , or ND were de-identified and sent to CSU for metabolomics analysis . In Nicaragua , 88 serum samples were retrospectively obtained from patients who had been diagnosed as DHF/DSS , DF , or ND ( Table 1 ) . The serum sample collection dates for the DF patients ranged from days 1 to 6 of illness and for the DHF/DSS patients from days 3 to 6 of illness . These patients had presented to the study clinic Centro de Salud Sócrates Flores Vivas and met the 1997 WHO case definition for dengue [30] or presented with undifferentiated fever , or to the Hospital Infantil Manuel de Jesús Rivera , the National Pediatric Reference Hospital , with a fever or history of fever <7 days and one or more of the following signs and symptoms: headache , arthralgia , myalgia , retro-orbital pain , positive tourniquet test , petechiae , or signs of bleeding . In Nicaragua , all samples were from pediatric patients <15 years of age , and 50% of the samples were from male and 50% from female patients . Fifty-nine positive samples were included in the analysis , and all were from patients infected with DENV-2 . Of the 59 positive samples , 18 ( 30% ) were primary infections and 41 ( 69% ) were secondary infections ( Table 1 ) . The majority of both DF and DHF/DSS patients experienced secondary infections . In México , 101 serum samples were retrospectively obtained from patients who were referred to the UADY clinics and had been diagnosed as DHF/DSS , DF , or ND ( Table 2 ) . The sample collection dates for the DF patients ranged from days 2 to 5 of illness , and for the DHF/DSS patients , from days 2 to 6 of illness . Approximately 10% ( 11 ) of these samples were from pediatric patients ( 1–15 years old ) and 90% ( 90 ) from adult patients ( ages 16–71 years ) ; 53% were from female and 47% from male patients . Sixty-eight DENV-positive samples were included in the analysis . In 47 samples ( 69% ) , the infecting DENV serotype was determined: 25 patients were infected with DENV-1 ( 37% ) , and 22 ( 32% ) were infected with DENV-2 . For patients infected with DENV-1 , 15 patients were diagnosed as DF and 10 as DHF/DSS . For patients infected with DENV-2 , 15 were diagnosed as DF and 7 as DHF/DSS ( Table 2 ) . In 21 patients ( 31% ) , the infecting virus serotype was not determined ( Table 2 ) . For most of the Mexican patients , information on whether the patients experienced a primary or secondary infection was not available , as these clinical specimens were not collected as part of an experimental protocol . Serum samples from patients in Nicaragua and Mexico were frozen at -80°C until thawed prior to preparation for LC-MS . In Nicaragua , a case was considered laboratory-confirmed dengue when acute DENV infection was demonstrated by detection of DENV RNA by RT-PCR , isolation of DENV , seroconversion of DENV-specific IgM antibody titers observed by MAC-ELISA in paired acute- and convalescent-phase samples , and/or a ≥4-fold increase in anti-DENV antibody titer measured using inhibition ELISA in paired acute and convalescent samples . In Nicaragua , computerized algorithms based on the 1997 WHO schema were used to classify cases according to disease severity [7 , 30 , 40] . In Mexico , a case was considered laboratory-confirmed dengue by detection of DENV RNA by RT-PCR , isolation of DENV , or detection of DENV-specific IgM antibodies . Classification of dengue severity was based on the traditional 1997 WHO definitions [7 , 30 , 41] . The final diagnosis of DHF/DSS , DF , or ND febrile illness based upon clinical and laboratory test results was available for each patient . In the Nicaraguan hospital study , a medical history was taken upon enrollment , and a complete physical exam was performed . Clinical data were collected every 12 hours for inpatients and every 24 hours for outpatients on Case Report Forms ( CRFs ) to follow patients’ disease evolution , with vital signs and fluid intake/output recorded more often as appropriate . In the Pediatric Dengue Cohort Study , febrile illnesses that met the WHO criteria for suspected dengue and undifferentiated febrile illnesses were treated as possible dengue cases and followed during the acute phase of illness by study physicians at the clinic . Cases were monitored closely for severe manifestations and were transferred by study personnel to the Infectious Disease Ward of the Hospital Infantil Manuel de Jesús Rivera when they presented with any sign of alarm [42 , 43] . For LC-MS analysis , serum samples from Nicaraguan and Mexican patients were thawed on ice , and 25 μl of serum was added to cold LC-MS grade methanol ( final concentration 75% ) [33] . The extract was dried using a speed vacuum at room temperature , suspended in 25 μl of 100% LC-MS grade acetonitrile ( ACN ) , and incubated at room temperature for 10 minutes ( min ) . Following vortexing for 1 min and centrifuging for 5 min at 4°C at 14 , 000 rpm , 15 μl of the supernatant was transferred to a glass vial for LC-MS analysis [38 , 39] . Biological samples were randomized , and the clinical diagnosis was not considered during sample preparation and data collection . Abundance measures of MFs produced by LC-MS metabolomics analyses should be considered semi-quantitative in discovery-phase studies [36] and are influenced by instrument and technical variation . To address this , quality control ( QC ) for SMB measurement followed the recommendations of Dunn et al . [44] for LC-MS analysis of human biofluids . Specifically , human serum ( Sigma ) was purchased and aliquoted . Each aliquot was processed using the protocol for preparation of human serum samples . After drying , the aliquot was frozen at -80°C until analyzed . For each experimental analysis , the QC sample and the dried experimental samples were reconstituted at the same time . The QC sample was analyzed first by LC-MS and the results compared to QC results obtained in previous analyses . Comparisons included the number of MFs detected , abundance of the MFs , and the baseline of the total ion chromatogram ( TIC ) among previously analyzed runs . In addition , the QC sample was analyzed after every 15 clinical samples . If differences were detected between QC control results ( either previous results or within the analysis ) , the analysis was stopped , the ionization source and the column ( see below ) were cleaned , and the mass spectrometer recalibrated . Additional samples were not analyzed until the QC analyses were satisfactory . A reference solution containing ions with m/z ( mass-to-charge ratio ) values 121 . 050873 and 922 . 009798 was infused directly with a capillary pump to ensure mass accuracy; the mass spectrometer continually was normalized to the intensity of these two ions . To evaluate the reproducibility of the LC-MS analysis , the retention time ( RT ) , and the area under the peaks of ten randomly selected representative metabolites were determined using the Nicaraguan serum specimens ( N = 88 ) . All differences in RTs and m/z values were ≤0 . 25 min and 15 ppm , respectively . All relative standard deviations ( RSD ) of the peak areas was below 25% ( Table 3 ) , confirming acceptable reproducibility of the chromatographic separation and accuracy of the mass measurements . Analyses of the prepared serum samples were performed using an Agilent 1200 series high performance liquid chromatography ( HPLC ) system connected to an Agilent 6520 Quadrupole Time-of-Flight ( Q-TOF ) MS fitted with a dual electrospray ionization ( ESI ) source ( Agilent Technologies , Palo Alto , CA ) . Metabolites were separated using a Cogent hydrophilic type-C silica diamond-hydride column ( particle size 4μm , pore size 100 Å , 2 . 1 mm x 150 mm ) with a Cogent diamond hydride guard column ( size 2 . 0 mm x 20 mm ) ( Microsolv Technology Corporation , NJ ) [45 , 46] . A 5-μl aliquot of each processed serum sample was applied to the column that had been equilibrated with 5% solvent A ( 0 . 1% formic acid in H2O ) and 95% solvent B ( 0 . 1% formic acid in ACN ) . Metabolites were eluted with the following nonlinear gradient formed with solvents A and B at a flow rate of 0 . 4 ml/min: 0 . 2 to 30 min , 95–50% B; 30 to 35 min , hold at 50% B; 35 to 40 min 50–20% B; 40 to 45 min 20–95% B . HPLC column eluent was directly introduced into the Q-TOF instrument for metabolite detection . The MS parameters used were as follows: scan rate: 1 . 4 spectra/sec; Vcap: 4000V; drying gas ( N2 ) : 325°C at 10 l/min; nebulizer pressure: 45 psi fragmentor: 150 V; skimmer: 65V; octopole RF peak: 750 V and 2 GHz extended dynamic range mode; mass range: 100–1700 Da . Reference solution containing ions with m/z ( mass-to-charge ratio ) values of 121 . 050873 and 922 . 009798 was infused directly with a capillary pump to ensure mass accuracy . Mass spectra data were collected in both centroid and profile modes . HILIC-MS data were analyzed using Agilent’s MassHunter Qualitative Analysis version B . 05 software to detect molecular features ( MFs ) ( compounds with defined accurate mass and RT ) present in each sample with a minimum abundance of 600 counts , ion species H+ , charge state maximum 1 , compound ion count threshold 2 or more ions , and all other parameters left at default values . MFs from the dengue diagnosis groups ( DHF/DSS , DF , and ND ) were compared using Agilent’s Mass Profiler Professional ( MPP ) , version B . 12 . 01 . MFs were aligned with 0 . 2 min retention time and 15 ppm mass tolerance and filtered based on their presence in at least 50% of samples in at least one diagnosis group . Subsequently , MFs were baselined to the median of all samples and normalized to the 75th percentile shift . The relative abundance of each filtered MF was then compared pairwise between diagnosis groups using ANOVA and Student's t-test . For all comparisons , the false discovery rate was calculated using the Benjamini-Hochberg algorithm , and the fold change ( FC ) was calculated for metabolites with corrected p-values of <0 . 05 . MFs with a corrected p-value of <0 . 05 and FC of >2 ( positive or negative ) were identified in silico when possible by interrogating the neutral mass of each in online databases [36] . The metabolites were putatively identified using Metlin [47] , HMDB [48] , or the Omics discovery pipeline [49] . Metlin parameters used for identification were neutral charge and mass tolerance of ± 10 ppm . HMDB parameters used were ion mode: positive , adduct type M+H , and molecular weight tolerance ± 0 . 01 Da . The Omics pipeline parameters used for identification included charge 0 and mass tolerance 0 . 01 Da . The number of database hits or other possible identities presented in S1 and S2 Tables were obtained using Metlin . All of the MFs that statistically differentiated the DHF/DSS , DF and ND disease outcomes are listed in S1 and S2 Tables ( Nicaraguan and Mexican serum differentiating metabolites , respectively ) . The chemical formulas were calculated using ChemCalc ( Institute of Chemical Sciences and Engineering ) [50] . The metabolites are listed by the Metabolomics Standard Initiative ( MSI ) level of identification [51 , 52] . MSI level 1: Identified metabolites ( experimental data matched chemical reference standards acquired on the same analytical platform ) . MSI level 2: Identified metabolites ( without chemical reference standards , based on physicochemical properties and spectrum similarity with public/commercial spectrum libraries ) . MSI level 3: Putatively identified metabolites ( based on physicochemical characteristics of a chemical class of compounds or by spectrum similarity to known compounds of a chemical class ) . MSI level 4: Unidentified metabolite ( unidentified or unclassified MFs that still could be differentiated or quantified based on spectrum data ) . These MFs could not be identified using the databases and Omics discovery pipeline [52] . Based upon their potential biological relevance , selected in silico-identified metabolites were further analyzed by targeted liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) to corroborate their identities . When available , commercial standards were purchased , and the MS/MS spectra of the standard and the candidate SMB in the native sample were compared . If a commercial standard was not available for the in silico-identified compound , the spectrum obtained from LC-MS/MS analysis of the native sample was compared to spectra available in the NIST commercial library [53] . Many metabolites that differentiated the disease outcomes with strong p-values and FCs remain to be identified at MSI levels 1 and 2 ( S2 and S3 Tables ) . This is due in part to the lack of commercially available standards and to lack of appropriate spectra ( e . g . , spectra obtained using same platforms and parameters ) in the commercially available libraries . A targeted analysis of a subset of 15 samples was utilized to validate the identification of vitamin D3 isotypes [54 , 55] . To each sample , 10 ng of [2H]3-25-hydroxyvitamin D3 , 10 ng of [2H]6−1 , 25-dihydroxyvitamin D3 , and 100 ng of [2H]6-vitamin D3 internal standard were added , followed by 1 ml of cold ( -20°C ) acetone . Each sample was vortexed and centrifuged to precipitate protein . The supernatant was dried using a rotary evaporation device . Just before analysis , each sample was derivatized by adding 50 μl 4-phenyl-1 , 2 , 4-triazoline-3 , 5-dione ( PTAD ) solution ( 1 mg/ml of ACN ) to each dry sample and reacting for 1 hour at room temperature . Samples were immediately analyzed by LC-MS/MS ( Agilent 6460 QQQ coupled to Rapid Resolution 1200 LC system; Agilent Technologies , Santa Clara , CA ) . Vitamin D3 concentrations were determined with [2H]6-vitamin D3 internal standard; 25-hydroxyvitamin D3 concentrations were determined with [2H]3-25-hydroxyvitamin D3- internal standard; 1 , 25-dihydroxyvitamin D3 concentrations were determined with [2H]6−1 , 25-dihydroxyvitamin D3 internal standard . An Agilent Zorbax C18 2 . 1x50 mm column was used for analysis . Buffers A and B consisted of 0 . 1% formic acid + 0 . 1% methylamine and ACN + 0 . 1% formic acid + 0 . 1% methylamine . All data were acquired in MRM mode by monitoring the methylamine adducts [54 , 55] . The transitions that were monitored by MRM for the identification of vitamin D3 isotypes and the collision energy used for fragmenting each MF are shown in Table 4 . Two transitions ( a deuterated and a non-deuterated ) were monitored for each compound [55] .
Characterization of metabolites in sera by HILIC-MS revealed 15 , 930 MFs in Nicaraguan specimens and 17 , 665 MFs in Mexican specimens ( Fig 1 ) . These were further analyzed using Mass Profiler Professional ( MPP ) to select MFs present in at least 50% of samples of at least one diagnosis group . This yielded 744 MFs in Nicaraguan serum specimens and 861 in Mexican samples ( Fig 1 ) . PCA plots demonstrated clustering of specimens by dengue diagnosis group in Nicaraguan samples , notably of the DHF/DSS patients ( Fig 2A ) . In contrast , there was little evidence of clustering by diagnosis group in the Mexican samples ( Fig 2B ) . Potential factors , such as age , gender , or infecting serotype that could contribute to the lack of clustering in the Mexican samples will be addressed below . Many factors are known to condition dengue disease severity , including primary versus secondary infection and infecting virus serotype and genotype . PCA was used to investigate the role of these potentially confounding factors on the serum metabolome of dengue patients . Pairwise comparisons of abundances revealed MFs in acute phase serum specimens that statistically ( corrected p-value <0 . 05 , FC >2 ) differentiated the DHF/DSS , DF , and ND diagnosis groups . In Nicaraguan specimens , 83 MFs differentiated DHF/DSS from DF patients , 191 MFs differentiated DHF/DSS from ND patients , and 191 MFs differentiated DF from ND patients ( Fig 1 ) . In Mexican serum specimens , 36 MFs differentiated DHF/DSS from DF patients , 313 MFs differentiated DHF/DSS from ND patients , and 309 MFs differentiated DF from ND patients ( Fig 1 ) . MFs that statistically differentiated the dengue diagnosis groups were ( when possible ) given putative structural identification by interrogation of the Metlin and HMDB databases and the Omics discovery pipeline [47–49 , 56 , 57] . The metabolites are listed by MSI Level of identification in S1 and S2 Tables . In the Nicaraguan specimens , 13 identified metabolites ( MSI Levels 1 and 2 ) , 103 putatively identified metabolites ( MSI Level 3 ) , and 101 unidentified metabolites ( MSI Level 4 ) differentiated the dengue diagnosis groups ( S1 Table ) . In the Mexican specimens , 12 identified metabolites ( MSI Levels 1 and 2 ) , 120 putatively identified metabolites ( MSI Level 3 ) , and 182 unidentified metabolites ( MSI Level 4 ) differentiated the diagnosis groups . Sixty-two of the differentiating metabolites were detected in both Nicaraguan and Mexican serum specimens ( S1 and S2 Tables ) . Thirty-eight of the differentiating metabolites ( denoted by ** ) exhibited a similar FC trend in the two groups; 24 metabolites exhibited an opposite FC trend ( denoted by *** ) in the two groups . Thus far , the structural identities of 13 metabolites that statistically differentiate DHF/DSS , DF , and ND disease outcomes in at least one of the pairwise comparisons of the diagnosis groups have been confirmed using MS/MS ( Table 4 ) . These metabolites were grouped into biochemical classes including amino acids and lipids such as fatty acids and phospholipids , as well as vitamins . The identities of six metabolites were confirmed ( MSI level 1 ) by comparing the HILIC-LC-MS/MS spectrum of the candidate metabolite in the native serum with that of a commercial standard ( Table 5 ) . The spectra of MSI level 1 compounds that identified proline , α-linolenic acid ( ALA ) , docosahexaenoic acid ( DHA ) , lysophosphatidylcholine ( lysoPC ) ( 16:1 ) , lysoPC ( 18:1 ) and arachidonic acid ( AA ) , and the International Chemical Identifier ( InChl ) [58] for each of these metabolites are presented in S1–S6 Figs . The presence of the three vitamin D3 metabolites detected by HILIC-MS was validated ( MSI level 1 ) by comparing the MRM spectrum of the candidate metabolite with that of a deuterated commercial standard using MRM LC-MS/MS [54 , 55] . The spectra identifying endogenous vitamin D3 , 25-hydroxyvitamin D3 , 1 , 25-dihydroxyvitamin D3 , and the InCHl identifiers are shown in S7–S9 Figs . The identities of four additional metabolites , myristoleic acid and three phosphatidylcholines ( PCs ) ( 34:1 , 34:0 , and 36:1 ) were confirmed ( MSI level 2 ) by spectrum similarity with spectra in the NIST library [59] . The differentiating metabolites that were identified at MSI level 3 in Nicaraguan and Mexican samples are listed in S1 and S2 Tables , respectively . These metabolites ( 102 in Nicaraguan specimens and 121 in Mexican specimens ) were identified in silico by interrogating online databases and libraries [47–49] and were assigned potential identities . These remain to be structurally confirmed . MFs that could not be identified in silico ( 101 in Nicaraguan specimens and 185 in Mexican specimens ) but that were differentiated and quantified based on LC-MS spectrum data ( MSI level 4 ) are also listed in S1 and S2 Tables . Although the day of defervescence was unavailable for either Nicaraguan or Mexican patients , information regarding progression to DHF/DSS of patients initially diagnosed as DF was available for 31 Nicaraguan patients . These specimens were collected ≤ 4 days post onset of symptoms , presumably before the time of defervescence . Of these , 16 were collected from patients initially diagnosed as DF who later on progressed to DHF/DSS and 15 from patients who did not progress to DHF/DSS . The PCA plot revealed clustering of the patients who experienced unremarkable DF and those who progressed to severe dengue disease ( Fig 4 ) . Statistical analysis of samples from these two patient groups yielded 65 metabolites that differentiated the eventual disease outcomes ( S3 Table ) . Six metabolites were identified at MSI level 1 ( Table 6 ) , and all were previously identified ( Table 5 ) . The identified metabolites were proline , alpha-linolenic acid , arachidonic acid , docosahexaenoic acid , and two lysoPCs . The metabolites identified at MSI levels 3 and 4 , which include 17 MFs at MSI level 3 , predominantly lipids , and 44 MFs at MSI level 4 , are listed in S3 Table . The potential metabolites ( MSI level 3 ) include phosphatidylcholines , diacylglycerol , phosphatidic acid , phosphatidylserine , triglycerides , and diacylglycerophosphoglycerol . Relative abundances of the six identified prognostic metabolites are presented in Fig 5 . For this analysis , the data were further processed using Agilent Mass Hunter Quantitative Analysis software B . 05 . 0 , and the results were imported into PAST ( Paleontological Statistics software package version 3 . 09 ) . The abundances of the respective metabolite in DF and DF patients who later progressed to DHF/DSS patients were statistically compared using two sample t-test for unequal variances . Each of these metabolites was elevated in abundance in the DF patients that progressed to DHF/DSS compared to those who experienced DF disease . The other metabolites listed in Table 6 and S3 Table are also candidate SMBs for progression to severe dengue disease and will be evaluated for their potential utility in predicting dengue disease outcomes .
Our studies confirm that DENV infection perturbs the human metabolome [32] . Statistical analyses indicated that many metabolites and MFs identified by HILIC-LC-MS had statistically significant differences in abundance in pairwise comparisons of the DHF/DSS , DF , and ND diagnosis groups ( Table 5 and S1 and S2 Tables ) . Cui et al . [32] demonstrated perturbation of many of the same metabolites in DF patients during the time-course of primary DENV infection . Metabolites that were perturbed in DHF/DSS and DF patients in both Nicaraguan and Mexican patients included lysoPCs ( 14:0 , 16:0 ) and long-chain polyunsaturated fatty acids such as DHA , AA , and ALA . To determine if differentiating metabolites identified by HILIC-LC-MS could be identified using a different LC-MS platform and to more thoroughly explore the metabolome , a subset of serum samples were analyzed in the Purdue Metabolite Profiling Facility ( PMPF ) using reverse phase ( RP ) -LC-MS [37] . In confirmation , 54% ( 117/288 ) of differentiating metabolites detected by HILIC-LC-MS were also detected using a T3 column ( Waters , Milford , MA ) in RP-LC-MS . All of the 13 differentiating metabolites whose identities were confirmed by LC-MS/MS ( Table 5 ) were detected by RP-LC-MS and differentiated the dengue diagnosis groups . These results provide proof of concept that differential perturbation of the serum metabolome is associated with different dengue infections and disease outcomes and that changes in relative concentrations of certain metabolites are associated with dengue diagnosis groups . Unfortunately , in this retrospective proof of concept study , a number of samples were obtained after the presumed time of defervescence and possible progression to severe disease ( Tables 1 and 2 ) . Thus , the differentiating metabolites identified in this retrospective study could represent metabolic perturbations reflecting the disease state instead of being predictive of progression to severe dengue disease . To address this issue , we compared the metabolic profiles of a subset of the Nicaraguan DF samples with those of DHF/DSS samples that were initially diagnosed as DF but then progressed to DHF/DSS ( Fig 5 ) ; all of these samples were collected by day 4 of illness ( presumably before the day of defervescence ) . Despite the small sample size , 65 metabolites differentiated the DF patients from those who progressed to DHF/DSS ( S3 Table ) , including six of the structurally confirmed metabolites reported in Table 5 ( proline , ALA , AA , DHA , and lysoPCs ( 16:0 and 18:1 ) ) ( Table 6 ) . These current candidate SMBs are not specific for dengue disease but , when combined with DENV-positive laboratory test results ( eg . , NS1 antigen or viral RNA detection ) , may provide diagnosis and prognosis of DENV infection outcomes using acute-phase serum specimens . Although these proof-of-concept prognostic metabolites are encouraging , they are based upon a small sample size and additional studies with increased numbers of patients will be needed to confirm the results . It must also be noted that these results are restricted to pediatric Nicaraguan patients . It cannot be assumed that the same metabolites will be predictive of progression to DHF/DSS in adult Nicaraguan patients or in patients from other geographic , genetic , and environmental backgrounds . Studies will be necessary to determine if these and/or alternate metabolites are predictive of progression to DHF/DSS in other patient populations . It was surprising that PCA revealed little clustering of Mexican samples by dengue diagnosis group in contrast to the Nicaraguan samples ( Fig 2A and 2B ) . Many factors have been demonstrated to condition dengue disease severity , including infecting DENV serotype and genotype and primary or secondary infections [24 , 60 , 61] , which could have confounded the analyses . The Mexican patients were infected with either DENV-1 or DENV-2; only DENV-2 was detected in Nicaraguan patients . To explore reasons for the lack of segregation of dengue diagnosis groups in PCA , we analyzed Mexican samples stratified by infecting virus serotype ( DENV-1 versus DENV-2 ) . PCA and statistical analyses revealed significant differences in the perturbation of the serum metabolome of Mexican patients attributable to the different serotypes ( Fig 3A , S4 Table ) . In this regard , the different numbers of DENV2 infections in Mexico and Nicaragua ( 22 and 59 , respectively , could have confounded the results ( Tables 1 and 2 ) . We also explored the potential role of primary versus secondary infections in perturbation of the serum metabolome of dengue patients . Immune status was only available for the Nicaraguan samples , which were stratified by primary versus secondary infection and analyzed . PCA plots revealed clustering of patients by primary versus secondary infection , and analyses revealed multiple metabolites that differentiated infection by immune status ( Fig 3B and S5 Table ) . This analysis of the Nicaraguan samples clearly demonstrates differential perturbations according to immune status , which are likely occurring in Mexican patients as well . Unfortunately immune status was only known for a few of the Mexican patients . Clearly the differences demonstrated for the Nicaraguan samples could have confounded the analyses of the Mexican samples . In addition the Mexican samples differed from the Nicaraguan samples in age distribution ( the effect of age on dengue disease severity and metabolome perturbations is addressed below ) . All of these and other factors could have contributed to the lack of clustering in the Mexican patients by diagnosis group , and additional metabolomics studies will be necessary to identify the actual mechanisms involved . Thus , although the available sample sizes were relatively small in this proof-of-concept study , PCA plots revealed clustering of patients by both infecting virus serotypes ( Fig 3A ) and by primary versus secondary infection , ( Fig 3B ) and differentiating metabolites were identified for each comparison ( S4 and S5 Tables ) . Interestingly , none of the differentiating metabolites in these two analyses overlapped , suggesting the involvement of different metabolic pathways or mechanisms . It is also interesting that these metabolites differ from those reported in Table 5 that differentiated the dengue disease diagnosis groups ( DHF/DSS versus DF , DHF/DSS versus ND and DF versus ND ) . Differentiating metabolites identified in this study provide insights into fundamental metabolic mechanisms and pathways that condition DENV infection , replication , and pathogenesis in humans , and several are potentially biologically and physiologically relevant in terms of severe disease outcomes ( DHF/DSS ) . Some are associated with lipid metabolism and regulation of inflammatory processes controlled by signaling fatty acids and phospholipids . Others are associated with immune regulation , endothelial function , and vascular barrier function , which is provocative in the context of the central role of vascular leakage in the pathogenesis of DENV infection and the possible progression to shock in DSS [3 , 4 , 7 , 30] . DENV replication is dependent on host cell lipid biosynthesis and metabolism . Viral replication complexes are enclosed in endoplasmic reticulum-derived double-membrane vesicles that organize and localize the complexes to facilitate the exchange of components with the cytosol for genome replication and virus assembly [37 , 62 , 63] . Long-chain polyunsaturated fatty acids such as DHA ( C22:6 ) and ALA ( C18:3 ) were increased in abundance in DHF/DSS versus DF and DHF/DSS versus ND groups in Nicaraguan serum samples and in early DF that progressed to DHF/DSS ( Tables 5 and 6 ) . Long-chain omega-3 polyunsaturated fatty acids such as DHA and its precursor ALA are potent anti-inflammatory agents and have been previously reported to be elevated during DENV infection [32] . DHA has been shown to decrease the production of inflammatory eicosanoids , cytokines , and reactive oxygen species [64 , 65] . This molecule can act both directly by inhibiting AA metabolism and indirectly by altering the expression of inflammatory gene products [66 , 67] . DHA also is a precursor of a family of anti-inflammatory mediators called D-series resolvins [66 , 67] . The increases in DHA levels we observed in dengue patient serum might represent the host attempt to mitigate immunopathology of dengue disease . AA and its metabolites have been shown to be elevated in plasma at different stages of infection in dengue patients [32 , 68] . This was confirmed in our results; we found AA levels elevated in DHF/DSS patients compared to DF patients in both Mexican and Nicaraguan populations . AA is mobilized from phospholipids in cell membranes and is metabolized by cyclooxygenases and lipoxygenases to pro-inflammatory eicosanoids such as prostaglandins , thromboxanes and leukotrienes [69 , 70] . We detected significant changes in abundances of several of these AA metabolites when comparing the dengue disease groups ( Tables 5 and 6 ) . A number of phospholipid metabolites differentiated dengue diagnosis groups in patients from both Nicaragua and Mexico ( Table 5 and S1 and S2 Tables ) . The increases we observed in phospholipid biosynthesis make biological sense given that host cell phospholipid metabolism is known to be influenced by DENV replication in both mosquito and human cells through DENV NS3 protein-mediated redistribution and activation of fatty acid synthase [37 , 61] . The prevalent phospholipids found to be increased primarily contain C16 and C18 unsaturated acyl chains . Palmitic acid ( C16 ) and oleic acid ( C18 ) have been found to be increased in DENV-infected mosquito cells and to facilitate production of lysoPCs . Phospholipids are precursors of lipid mediators , such as platelet activating factors ( PAFs ) and eicosanoids , which are involved in inflammatory responses [70 , 71] . LysoPCs ( 18:1 , 16:0 ) were elevated in acute-phase serum specimens of DHF/DSS patients ( Table 5 ) . These single fatty acid chain lipids are involved in alteration of membrane structures and can mediate acute inflammation and regulate pathophysiological events throughout the vasculature and at local tissue sites [72–75] . Interestingly , lysoPCs may alter homeostasis of vascular endothelium , causing endothelial cell instability , barrier dysfunction , and vascular leakage , a major component of the pathophysiology of DSS [18 , 76–78] . Previous reports demonstrated perturbation of lysoPC concentrations in DENV-infected human serum [32 , 37] . Up-regulation of the phosphatidylcholine biosynthesis pathway in acute DENV infections ( days 1–3 ) was identified as one of the predictors for progression to DHF [19] . Other metabolites from different biochemical classes differentiated dengue disease outcomes . For example , we observed lower levels of 1 , 25-dihydroxyvitamin D3 ( 1 , 25-vitD3 ) in DHF/DSS versus DF and ND in Nicaraguan patients ( Table 5 ) . Reduced levels of 1 , 25-vitD3 , with its roles in immunoregulation and vascular barrier function , could be involved in the immunopathophysiology associated with DHF/DSS [79] . A decrease in serum 1 , 25-vitD3 levels is associated with increased mortality in sepsis patients [80 , 81] . The active form of vitamin D3 ( 1 , 25-vitD3 ) can be synthesized in vascular endothelium following stimulation of vitamin D3 1α-hydroxylase activity by inflammatory cytokines . Interactions of this metabolite with endothelial cells and the reduction of 1 , 25-vitD3 observed in immune-mediated diseases by others [79 , 82] prompts speculation about the potential role of decreased concentrations of this metabolite in patients progressing to DHF/DSS . Interestingly , polymorphisms in the vitamin D receptor gene are linked with severe dengue disease outcomes [14] . Several amino acids or peptides were also found to differentiate disease outcomes . For example , proline , which can act as a modulator of the intracellular redox environment , differed in DHF/DSS and DF patients who were initially diagnosed as DF ( Fig 5 ) . Perturbations of proline in endothelial cells could affect endothelium function [83–85] . Clearly , metabolomics provides new opportunities and a powerful approach to investigate potential viral , host , pathogenic , and immunologic determinants of DENV infection and pathogenesis . Identification of metabolites that differentiate dengue disease outcomes in patients from different geographic areas , environmental conditions , genetic backgrounds , sexes , and ages [34 , 35] is an important first step in selecting SMBs for dengue diagnosis and prognosis . We have already identified a large , overlapping set of metabolites that differentiated dengue outcomes in genetically and geographically distinct populations . However , the associations were not always concordant . In some instances , a metabolite differentiated the disease outcomes in one study population but not in the other . For example , lysoPCs ( 16:0 and 18:1 ) statistically differentiated DHF/DSS and DF diagnosis groups in Nicaraguan patients but not in Mexican patients ( Table 5 ) . In other instances , a candidate SMB was either increased or decreased in abundance in serum from patients from one country and the opposite trend occurred in patients from the other country . Sixty-two of the differentiating metabolites were detected in serum specimens from the Nicaraguan and Mexican patients ( see Table 5 and S1 and S2 Tables ) . Thirty-eight of these differentiating metabolites ( denoted by ** ) had similar FC trends in Nicaraguan and Mexican patients , but 24 ( denoted by *** ) had opposite FC trends . For example , ALA and DHA exhibited positive fold-changes in differentiating DHF/DSS from DF and ND outcomes in Nicaraguan patient sera but negative trends in Mexican patients . A number of factors could have contributed to the dissimilar change trends in the two populations . For example in this exploratory metabolomics study , no Mexican patient was officially diagnosed as DSS ( although some were hospitalized and diagnosed with DHF ) , while 15% of the DHF/DSS patients in Nicaragua were classified as DSS . The lack of DSS cases in Mexico is a limitation of our study that may have confounded the identification of SMBs that differentiate DSS from non-DSS disease outcomes . In addition , there were major age differences in the two study populations . The age of the patient can condition dengue disease severity [86 , 87] . Severity of symptoms ( which can be subjective ) that influence clinical diagnosis and disease classification in the two populations could also be strongly influenced by the age of the participants . Ninety percent of the Mexican patients were >15 years of age and may have been less likely to progress to DSS even though they were diagnosed as DHF patients . All of the Nicaraguan patients were children <15 years of age . Human metabolic profiles are age-dependent [88] , and DENV pathophysiology and clinical symptomology ( e . g . , DHF ) can differ in different age groups [86 , 87 , 89] and by sex [60] . DSS is negatively correlated with age [13] . Liver inflammation ( an important target organ in DENV infection ) is more prevalent in children than in adults [90 , 91] . In this regard , we conducted a very preliminary analysis to examine the potential role of age on the DENV-infected serum metabolome . The Mexican patients with DF or DHF/DSS ranged from 1 to 62 years of age . These DENV-infected samples were stratified by age; pediatric patients <15 years of age ( N = 11 ) and adult patients >15 years of age ( N = 57 ) , and the serum metabolites were characterized by PCA ( S10 Fig ) . Because of the limited number of pediatric patients , in this preliminary analysis we used a filter of 25% metabolite presence in samples of one of the diagnosis groups instead of our standard filter of 50% metabolite presence . Despite the small number of pediatric patients , clustering of patients by age was evident ( S10 Fig ) . Clearly , age differences could have contributed to the metabolomic differences between the two groups . The effect of age on the serum metabolome during DENV-infections will be a fruitful area of research . Identification of metabolites that differentiate age groups could provide important insights into differences in the pathophysiology of DENV infections in pediatric and adult patients [86 , 87] . Other factors could also contribute to the dissimilar change trends in metabolite abundance in the two populations . Dietary differences between Nicaraguan and Mexican patients could confound results with metabolites such as ALA , which is obtained principally from dietary plant sources , and DHA , which is a metabolite of ALA . All of these factors could account for some of the metabolite differences seen in the two study populations . Further studies will be necessary to determine if metabolites such as lysoPCs and DHA are candidate SMBs for progression to DSS in older patients and in patients from other geographic areas . We are currently conducting a prospective clinical study in Managua , Nicaragua , to determine the diagnostic and prognostic potential of existing and yet to be identified candidate SMBs . MRM analysis [92] will be conducted for accurate quantification of abundance of metabolites in different disease states as part of the evaluation of the potential diagnostic utility of candidate SMBs . Super learner analysis [93] is being used to identify the most parsimonious SMB “biosignature” in acute phase serum specimens that , when combined with other laboratory and clinical information , such as NS1 antigen detection and viral RNA detection by RT-PCR , will provide the most efficient algorithms for dengue diagnosis and prognosis . This would be of enormous value for appropriate patient triaging , management and clinical care . Prospective clinical studies will allow us to increase the number of early acute-phase patients and to identify additional metabolites that are predictive of progression to severe dengue disease . Additional clinical studies potentially will also allow us to increase the number of patients who progress to DSS and to identify metabolites that differentiate DHF and DSS disease outcomes [28] . The studies will also provide insights into metabolic pathways and pathogenic mechanisms that condition DHF and DSS outcomes . Such information could potentially be exploited in the development of new therapeutics for treatment of dengue patients in danger of progressing to DSS [28] . The 3- to 4-day window from dengue disease onset to defervescence provides a unique opportunity for therapeutic intervention [3–5 , 7 , 30] . Clearly , metabolomics provides new opportunities for diagnosis and prognosis of DENV infections . | Epidemics of dengue fever ( DF ) and dengue hemorrhagic fever/dengue shock syndrome ( DHF/DSS ) are overwhelming public health capacity for diagnosis and patient care . Developing a panel of biomarkers in acute-phase serum specimens for prognosis of severe dengue disease would be of enormous value for appropriate triaging of patients for management . Metabolomics offers great potential for identification of small molecule biomarkers ( SMBs ) for diagnosis and prognosis of dengue virus ( DENV ) infections . We identified metabolites that were associated with and differentiated DHF/DSS , DF and non-dengue ( ND ) febrile illness outcomes , primary and secondary virus infections , and infections with different DENV serotypes . These metabolites provide insights into metabolic pathways that play roles in DENV infection , replication , and pathogenesis . Some are associated with lipid metabolism and regulation of inflammatory processes controlled by signaling fatty acids and phospholipids , and others with endothelial cell homeostasis and vascular barrier function . Such metabolites and associated metabolic pathways are potentially biologically relevant in DENV pathogenesis . The diagnostic and prognostic efficacy of differentiating metabolites is currently being investigated . Our goal is to identify the most parsimonious SMB biosignature that , when combined with laboratory diagnostic results , eg . , DENV NS1 or RNA detection , will provide the most efficient algorithm for dengue diagnosis and prognosis . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"liquid",
"chromatography",
"tropical",
"diseases",
"organic",
"compounds",
"pediatrics",
"metabolomics",
"metabolites",
"neglected",
"tropical",
"diseases",
"pediatric",
"infections",
"research",
"and",
"a... | 2016 | Metabolomics-Based Discovery of Small Molecule Biomarkers in Serum Associated with Dengue Virus Infections and Disease Outcomes |
Spike timing dependent plasticity ( STDP ) is a learning rule that modifies synaptic strength as a function of the relative timing of pre- and postsynaptic spikes . When a neuron is repeatedly presented with similar inputs , STDP is known to have the effect of concentrating high synaptic weights on afferents that systematically fire early , while postsynaptic spike latencies decrease . Here we use this learning rule in an asynchronous feedforward spiking neural network that mimics the ventral visual pathway and shows that when the network is presented with natural images , selectivity to intermediate-complexity visual features emerges . Those features , which correspond to prototypical patterns that are both salient and consistently present in the images , are highly informative and enable robust object recognition , as demonstrated on various classification tasks . Taken together , these results show that temporal codes may be a key to understanding the phenomenal processing speed achieved by the visual system and that STDP can lead to fast and selective responses .
Temporal constraints pose a major challenge to models of object recognition in cortex . When two images are simultaneously flashed to the left and right of fixation , human subjects can make reliable saccades to the side where there is a target animal in as little as 120–130 ms [1] . If we allow 20–30 ms for motor delays in the oculomotor system , this implies that the underlying visual processing can be done in 100 ms or less . In monkeys , recent recordings from inferotemporal cortex ( IT ) showed that spike counts over time bins as small as 12 . 5 ms ( which produce essentially a binary vector with either ones or zeros ) and only about 100 ms after stimulus onset contain remarkably accurate information about the nature of a visual stimulus [2] . This sort of rapid processing presumably depends on the ability of the visual system to learn to recognize familiar visual forms in an unsupervised manner . Exactly how this learning occurs constitutes a major challenge for theoretical neuroscience . Here we explored the capacity of simple feedforward network architectures that have two key features . First , when stimulated with a flashed visual stimulus , the neurons in the various layers of the system fire asynchronously , with the most strongly activated neurons firing first—a mechanism that has been shown to efficiently encode image information [3] . Second , neurons at later stages of the system implement spike timing dependent plasticity ( STDP ) , which is known to have the effect of concentrating high synaptic weights on afferents that systematically fire early [4 , 5] . We demonstrate that when such a hierarchical system is repeatedly presented with natural images , these intermediate-level neurons will naturally become selective to patterns that are reliably present in the input , while their latencies decrease , leading to both fast and informative responses . This process occurs in an entirely unsupervised way , but we then show that these intermediate features are able to support categorization . Our network belongs to the family of feedforward hierarchical convolutional networks , as in [6–10] . To be precise , its architecture is inspired from Serre , Wolf , and Poggio's model of object recognition [6] , a model that itself extends HMAX [7] and performs remarkably well with natural images . Like them , in an attempt to model the increasing complexity and invariance observed along the ventral pathway [11 , 12] , we use a four-layer hierarchy ( S1–C1–S2–C2 ) in which simple cells ( S ) gain their selectivity from a linear sum operation , while complex cells ( C ) gain invariance from a nonlinear max pooling operation ( see Figure 1 and Methods for a complete description of our model ) . Nevertheless , our network does not only rely on static nonlinearities: it uses spiking neurons and operates in the temporal domain . At each stage , the time to first spike with respect to stimulus onset ( or , to be precise , the rank of the first spike in the spike train , as we will see later ) is supposed to be the “key variable , ” that is , the variable that contains information and that is indeed read out and processed by downstream neurons . When presented with an image , the first layer's S1 cells , emulating V1 simple cells , detect edges with four preferred orientations , and the more strongly a cell is activated , the earlier it fires . This intensity–latency conversion is in accordance with recordings in V1 showing that response latency decreases with the stimulus contrast [13 , 14] and with the proximity between the stimulus orientation and the cell's preferred orientation [15] . It has already been shown how such orientation selectivity can emerge in V1 by applying STDP on spike trains coming from retinal ON- and OFF-center cells [16] , so we started our model from V1 orientation-selective cells . We also limit the number of spikes at this stage by introducing competition between S1 cells through a one-winner-take-all mechanism: at a given location—corresponding to one cortical column—only the spike corresponding to the best matching orientation is propagated ( sparsity is thus 25% at this stage ) . Note that k-winner-take-all mechanisms are easy to implement in the temporal domain using inhibitory GABA interneurons [17] . These S1 spikes are then propagated asynchronously through the feedforward network of integrate-and-fire neurons . Note that within this time-to-first-spike framework , the maximum operation of complex cells simply consists of propagating the first spike emitted by a given group of afferents [18] . This can be done efficiently with an integrate-and-fire neuron with low threshold that has synaptic connections from all neurons in the group . Images are processed one by one , and we limit activity to at most one spike per neuron , that is , only the initial spike wave is propagated . Before presenting a new image , every neuron's potential is reset to zero . We process various scaled versions of the input image ( with the same filter size ) . There is one S1–C1–S2 pathway for each processing scale ( not represented on Figure 1 ) . This results in S2 cells with various receptive field sizes ( see Methods ) . Then C2 cells take the maximum response ( i . e . , first spike ) of S2 cells over all positions and scales , leading to position and scale invariant responses . This paper explains how STDP can set the C1–S2 synaptic connections , leading to intermediate-complexity visual features , whose equivalent in the brain may be in V4 or IT . STDP is a learning rule that modifies the strength of a neuron's synapses as a function of the precise temporal relations between pre- and postsynaptic spikes: an excitatory synapse receiving a spike before a postsynaptic one is emitted is potentiated ( long-term potentiation ) whereas its strength is weakened the other way around ( long-term depression ) [19] . The amount of modification depends on the delay between these two events: maximal when pre- and postsynaptic spikes are close together , and the effects gradually decrease and disappear with intervals in excess of a few tens of milliseconds [20–22] . Note that STDP is in agreement with Hebb's postulate because presynaptic neurons that fired slightly before the postsynaptic neuron are those that “took part in firing it . ” Here we used a simplified STDP rule where the weight modification does not depend on the delay between pre- and postsynaptic spikes , and the time window is supposed to cover the whole spike wave ( see Methods ) . We also use 0 and 1 as “soft bounds” ( see Methods ) , ensuring the synapses remain excitatory . Several authors have studied the effect of STDP with Poisson spike trains [4 , 23] . Here , we demonstrate STDP's remarkable ability to detect statistical regularities in terms of earliest firing afferent patterns within visual spike trains , despite their very high dimensionality inherent to natural images . Visual stimuli are presented sequentially , and the resulting spike waves are propagated through to the S2 layer , where STDP is used . We use restricted receptive fields ( i . e . , S2 cells only integrate spikes from an s × s square neighborhood in the C1 maps corresponding to one given processing scale ) and weight-sharing ( i . e . , each prototype S2 cell is duplicated in retinotopic maps and at all scales ) . Starting with a random weight matrix ( size = 4 × s × s ) , we present the first visual stimuli . Duplicated cells are all integrating the spike train and compete with each other . If no cell reaches its threshold , nothing happens and we process the next image . Otherwise for each prototype the first duplicate to reach its threshold is the winner . A one-winner-take-all mechanism prevents the other duplicated cells from firing . The winner thus fires and the STDP rule is triggered . Its weight matrix is updated , and the change in weights is duplicated at all positions and scales . This allows the system to learn patterns despite changes in position and size in the training examples . We also use local inhibition between different prototype cells: when a cell fires at a given position and scale , it prevents all other cells from firing later at the same scale and within an s/2 × s/2 square neighborhood of the firing position . This competition , only used in the learning phase , prevents all the cells from learning the same pattern . Instead , the cell population self-organizes , each cell trying to learn a distinct pattern so as to cover the whole variability of the inputs . If the stimuli have visual features in common ( which should be the case if , for example , they contain similar objects ) , the STDP process will extract them . That is , for some cells we will observe convergence of the synaptic weights ( by saturation ) , which end up being either close to 0 or to 1 . During the convergence process , synapses compete for control of the timing of postsynaptic spikes [4] . The winning synapses are those through which the earliest spikes arrive ( on average ) [4 , 5] , and this is true even in the presence of jitter and spontaneous activity [5] ( although the model presented in this paper is fully deterministic ) . This “preference” for the earliest spikes is a key point since the earliest spikes , which correspond in our framework to the most salient regions of an image , have been shown to be the most informative [3] . During the learning , the postsynaptic spike latency decreases [4 , 5 , 24] . After convergence , the responses become selective ( in terms of latency ) [5] to visual features of intermediate complexity similar to the features used in earlier work [8] . Features can now be defined as clusters of afferents that are consistently among the earliest to fire . STDP detects these kinds of statistical regularities among the spike trains and creates one unit for each distinct pattern .
We evaluated our STDP-based learning algorithm on two California Institute of Technology datasets , one containing faces and the other motorbikes , and a distractor set containing backgrounds , all available at http://www . vision . caltech . edu ( see Figure 2 for sample pictures ) . Note that most of the images are not segmented . Each dataset was split into a training set , used in the learning phase , and a testing set , not seen during the learning phase but used afterward to evaluate the performance on novel images . This standard cross-validation procedure allows the measurement of the system's ability to generalize , as opposed to learning the specific training examples . The splits used were the same as Fergus , Perona , and Zisserman [25] . All images were rescaled to be 300 pixels in height ( preserving the aspect ratio ) and converted to grayscale values . We first applied our unsupervised STDP-based algorithm on the face and motorbike training examples ( separately ) , presented in random order , to build two sets of ten class-specific C2 features . Each C2 cell has one preferred input , defined as a combination of edges ( represented by C1 cells ) . Note that many gray-level images may lead to this combination of edges because of the local max operation of C1 cells and because we lose the “polarity” information ( i . e . , which side of the edge is darker ) . However , we can reconstruct a representation of the set of preferred images by convolving the weight matrix with a set of kernels representing oriented bars . Since we start with random weight matrices , at the beginning of the learning process the reconstructed preferred stimuli do not make much sense . But as the cells learn , structured representations emerge , and we are usually able to identify the nature of the cells' preferred stimuli . Figures 3 and 4 show the reconstructions at various stages of learning for the face and motorbike datasets , respectively . We stopped the learning after 10 , 000 presentations . Then we turned off the STDP rule and tested these STDP-obtained features' ability to support face/nonface and motorbike/nonmotorbike classification . This paper focuses more on feature extraction than on sophisticated classification methods , so we first used a very simple decision rule based on the number of C2 cells that fired with each test image , on which a threshold is applied . Such a mechanism could be easily implemented in the brain . The threshold was set at the equilibrium point ( i . e . , when the false positive rate equals the missed rate ) . In Table 1 we report good classification results with this “simple-count” scheme in terms of area under the receiver operator characteristic ( ROC ) and the performance rate at equilibrium point . We also evaluated a more complicated classification scheme . C2 cells' thresholds were supposed to be infinite , and we measured the final potentials they reached after having integrated the whole spike train generated by the image . This final potential can be seen as the number of early spikes in common between a current input and a stored prototype ( this contrasts with HMAX and extensions [6 , 7 , 26] , where a Euclidian distance or a normalized dot product is used to measure the difference between a stored prototype and a current input ) . Note that this potential is contrast invariant: a change in contrast will shift all the latencies but will preserve the spike order . The final potentials reached with the training examples were used to train a radial basis function ( RBF ) classifier ( see Methods ) . We chose this classifier because linear combination of Gaussian-tuned units is hypothesized to be a key mechanism for generalization in the visual system [27] . We then evaluated the RBF on the testing sets . As can be seen in Table 1 , performance with this “potential + RBF” scheme was better . Using only ten STDP-learnt features , we reached on those two classes a performance that is comparable to that of Serre , Wolf , and Poggio's model , which itself is close to the best state-of-the-art computer vision systems [6] . However , their system is more generic . Classes with more intraclass variability ( for example , animals ) appear to pose a problem with our approach because a lot of training examples ( say a few tens ) of a given feature type are needed for the STDP process to learn it properly . Our approach leads to the extraction of a small set ( here ten ) of highly informative class-specific features . This is in contrast with Serre et al . 's approach where many more ( usually about a thousand ) features are used . Their sets are more generic and are suitable for many different classes [6] . They rely on the final classifier to “select” diagnostic features and appropriately weight them for a given classification task . Here , STDP will naturally focus on what is common to the positive training set , that is , target object features . The background is generally not learned ( at least not in priority ) , since backgrounds are almost always too different from one image to another for the STDP process to converge . Thus , we directly extract diagnostic features , and we can obtain reasonably good classification results using only a threshold on the number of detected features . Furthermore , as STDP performs vector quantization from multiple examples as opposed to “one-shot learning , ” it will not learn the noise , nor anything too specific to a given example , with the result that it will tend to learn archetypical features . Another key point is the natural trend of the algorithm to learn salient regions , simply because they correspond to the earliest spikes , with the result that neurons whose receptive fields cover salient regions are likely to reach their threshold ( and trigger the STDP rule ) before neurons “looking” at other regions . This contrasts with more classical competitive learning approaches , where input normalization helps different input patterns to be equally effective in the learning process [28] . Note that “salient” means within our network “with well-defined contrasted edges , ” but saliency is a more generic concept of local differences , for example , in intensity , color , or orientations as in the model of Itti , Koch , and Niebur [29] . We could use other types of S1 cells to detect other types of saliency , and , provided we apply the same intensity–latency conversion , STDP would still focus on the most salient regions . Saliency is known to drive attention ( see [30] for a review ) . Our model predicts that it also drives the learning . Future experimental work will test this prediction . Of course , in real life we are unlikely to see many examples of a given category in a row . That is why we performed a second simulation , where 20 C2 cells were presented with the face , motorbike , and background training pictures in random order , and the STDP rule was applied . Figure 5 shows all the reconstructions for this mixed simulation after 20 , 000 presentations . We see that the 20 cells self-organized , some of them having developed selectivity to face features , and others to motorbike features . Interestingly , during the learning process the cells rapidly showed a preference for one category . After a certain degree of selectivity had been reached , the face-feature learning was not influenced by the presentation of motorbikes ( and vice versa ) , simply because face cells will not fire ( and trigger the STDP rule ) on motorbikes . Again we tested the quality of these features with a ( multiclass ) classification task , using an RBF network and a “one-versus-all” approach ( see Methods ) . As before , we tested two implementations: one based on “binary detections + RBF” and one based on “potential + RBF” . Note that a simple detection count cannot work here , as we need at least some supervised learning to know which feature ( or feature combination ) is diagnostic ( or antidiagnostic ) of which class . Table 2 shows the confusion matrices obtained on the testing sets for both implementations , leading , respectively , to 95 . 0% and 97 . 7% of correct classifications on average . It is worth mentioning that the “potential + RBF” system perfectly discriminated between faces and motorbikes—although both were presented in the unsupervised STDP-based learning phase . A third type of simulation was run to illustrate the STDP learning process . For these simulations , only three C2 cells and four processing scales ( 71% , 50% , 35% , and 25% ) were used . We let at most one cell fire at each processing scale . The rest of the parameters were strictly identical to the other simulations ( see Methods ) . Videos S1–S3 illustrate the STDP learning process with , respectively , faces , motorbikes , and a mix of faces , motorbikes , and background pictures . It can be seen that after convergence the STDP feature showed a good tradeoff between selectivity ( very few false alarms ) and invariance ( most of the targets were recognized ) . An interesting control is to compare the STDP learning rule with a more standard hebbian rule in this precise framework . For this purpose , we converted the spike trains coming from C1 cells into a vector of ( real-valued ) C1 activities XC1 , supposed to correspond to firing rates ( see Methods ) . Each S2 cell was no longer modeled at the integrate-and-fire level but was supposed to respond with a ( static ) firing rate YS2 given by the normalized dot product: where WS2 is the synaptic weight vector of the S2 cell ( see Methods ) . The S2 cells still competed with each other , but the k-winner-take-all mechanisms now selected the cells with the highest firing rates ( instead of the first one to fire ) . Only the cells whose firing rates reached a certain threshold were considered in the competition ( see Methods ) . The winners now triggered the following modified hebbian rule ( instead of STDP ) : where a decay term has been added to keep the weight vector bounded ( however , the rule is still local , unlike an explicit weight normalization ) . Note that this precaution was not needed in the STDP case because competition between synapse naturally bounds the weight vector [4] . The rest of the network is strictly identical to the STDP case . Figure 6 shows the reconstruction of the preferred stimuli for the ten C2 cells after 10 , 000 presentations for the face stimuli ( Figure 6 , top ) and the motorbikes stimuli ( Figure 6 , top ) . Again we can usually recognize the face and motorbike parts to which the cells became selective ( even though the reconstructions look fuzzier than in the STDP case because the final weights are more graded ) . We also tested the ability of these hebbian-obtained features to support face/nonface and motorbike/nonmotorbike classification once fed into an RBF , and the results are shown in Table 1 ( last column ) . We also evaluated the hebbian features with the multiclass setup . Twenty cells were presented with the same mix of face , motorbike , and background pictures as before . Figure 7 shows the final reconstructions after 20 , 000 presentations , and Table 2 shows the confusion matrix ( last columns ) . The main conclusion is that the modified hebbian rule is also able to extract pertinent features for classification ( although performance on these tests appears to be slightly worse ) . This is not very surprising as STDP can be seen as a hebbian rule transposed in the temporal domain , but it was worth checking . Where STDP would detect ( and create selectivity to ) sets of units that are consistently among the first one to fire , the hebbian rule detects ( and creates selectivity to ) sets of units that consistently have the highest firing rates . However , we believe the temporal framework is a better description of what really happens at the neuronal level , at least in ultrarapid categorization tasks . Furthermore , STDP also explains how the system becomes faster and faster with training , since the neurons learn to decode the first information available at their afferents' level ( see also Discussion ) .
While the ability of hierarchical feedforward networks to support classification is now reasonably well established ( e . g . , [6–8 , 10] ) , how intermediate-complexity features can be learned remains an open problem , especially with cluttered images . In the original HMAX model , S2 features were not learned but were manually hardwired [7] . Later versions used huge sets of random crops ( say 1 , 000 ) taken from natural images and used these crops to “imprint” S2 cells [6] . This approach works well but is costly since redundancy is very high between features , and many features are irrelevant for most ( if not all ) of the tasks . To select only pertinent features for a given task , Ullman proposed an interesting criterion based on mutual information [8] , leaving the question of possible neural implementation open . LeCun showed how visual features in a convolutional network could be learned in a supervised manner using back-propagation [10] , without claiming this algorithm was biologically plausible . Although we may occasionally use supervised learning to create a set of features suitable for a particular recognition task , it seems unrealistic that we need to do that each time we learn a new class . Here we took another approach: one layer with unsupervised competitive learning is used as input for a second layer with supervised learning . Note that this kind of hybrid scheme has been found to learn much faster than a two-layer backpropagation network [28] . Our approach is a bottom-up one: instead of intuiting good image-processing schemes and discussing their eventual neural correlates , we took known biological phenomena that occur at the neuronal level , namely integrate-and-fire and STDP , and observed where they could lead at a more integrated level . The role of the simulations with natural images is thus to provide a “plausibility proof” that such mechanisms could be implemented in the brain . However , we have made four main simplifications . The first one was to propagate input stimuli one by one . This may correspond to what happens when an image is flashed in an ultrarapid categorization paradigm [1] , but normal visual perception is an ongoing process . However , every 200 ms or 300 ms we typically perform a saccade . The processing of each of these discrete “chunks” seems to be optimized for rapid execution [31] , and we suggest that much can be done with the feedforward propagation of a single spike wave . Furthermore , even when fixating , our eyes are continuously making microsaccades that could again result in repetitive waves of activation . This idea is in accordance with electrophysiological recordings showing that V1 neuron activity is correlated with microsaccades [32] . Here we assumed the successive waves did not interfere , which does not seem too unreasonable given that the neuronal time constants ( integration , leak , STDP window ) are in the range of a few tens of milliseconds whereas the interval between saccades and microsaccades is substantially longer . It is also possible that extraretinal signals suppress interference by shutting down any remaining activity before propagating the next wave . Note that this simplification allows us to use nonleaky integrate-and-fire neurons and an infinite STDP time window . More generally , as proposed by Hopfield [33] , waves could be generated by population oscillations that would fire one cell at a time in advance of the maximum of the oscillation , which increases with the inputs the cell received . This idea is in accordance with recordings in area 17 of cat visual cortex showing that suboptimal cells reveal a systematic phase lag relative to optimally stimulated cells [34] . The second simplification we have made is to use only five layers ( including the classification layer ) , whereas processing in the ventral stream involves many more layers ( probably about ten ) , and complexity increases more slowly than suggested here . However , STDP as a way to combine simple features into more complex representations , based on statistical regularities among earliest spike patterns , seems to be a very efficient learning rule and could be involved at all stages . The third main simplification we have made consists of using restricted receptive fields and weight sharing , as do most of the bio-inspired hierarchical networks [6–10] ( networks using these techniques are called convolutional networks ) . We built shift and scale invariance by structure ( and not by training ) by duplicating S1 , C1 , and S2 cells at all positions and scales . This is a way to reduce the number of free parameters ( and therefore the VC dimension [35] ) of the network by incorporating prior information into the network design: responses should be scale- and shift-invariant . This greatly reduces the number of training examples needed . Note that this technique of weight sharing could be applied to other transformations than shifting and scaling , for instance , rotation and symmetry . However , it is difficult to believe that the brain could really use weight sharing since , as noted by Földiák [36] , updating the weights of all the simple units connected to the same complex unit is a nonlocal operation . Instead , he suggested that at least the low-level features could be learned locally and independently . Subsequently , cells with similar preferred stimulus may connect adaptively to the same complex cell , possibly by detecting correlation across time thanks to a trace rule [36] . Wallis , Rolls , and Milward successfully implemented this sort of mechanism in a multilayered hierarchical network called Vis-Net [37 , 38]; however , performance after learning objects from unsegmented natural images was poor [39] . Future work will evaluate the use of local learning and adaptative complex pooling in our network , instead of exact weight sharing . Learning will be much slower but should lead to similar STDP features . Note that it seems that monkeys can recognize high-level objects at scales and positions that have not been experienced previously [2 , 40] . It could be that in the brain local learning and adaptative complex pooling are used up to a certain level of complexity , but not for high-level objects . These high-level objects could be represented with a combination of simpler features that would already be shift- and scale-invariant . As a result , there would be less need for spatially specific representations for high-level objects . The last main simplification we have made is to ignore both feedback loops and top-down influences . While normal , everyday vision extensively uses feedback loops , the temporal constraints almost certainly rule them out in an ultrarapid categorization task [41] . The same cannot be said about the top-down signals , which do not depend directly on inputs . For example , there is experimental evidence that the selectivity to the “relevant” features for a given recognition task can be enhanced in IT [42] and in V4 [43] , possibly thanks to a top-down signal coming from the prefrontal cortex , thought to be involved in the categorization process . These effects , for example , modeled by Szabo et al . [44] , are not taken into account here . Despite these four simplifications , we think our model captures two key mechanisms used by the visual system for rapid object recognition . The first one is the importance of the first spikes for rapidly encoding the most important information about a visual stimulus . Given the number of stages involved in high-level recognition and the short latencies of selective responses recorded in monkeys' IT [2] , the time window available for each neuron to perform its computation is probably about 10–20 ms [45] and will rarely contain more than one or two spikes . The only thing that matters for a neuron is whether an afferent fires early enough so that the presynaptic spike falls in the critical time window , while later spikes cannot be used for ultrarapid categorization . At this point ( but only at this point ) , we have to consider two hypotheses: either presynaptic spike times are completely stochastic ( for example , drawn from a Poisson distribution ) , or they are somewhat reliable . The first hypothesis causes problems since the first presynaptic spikes ( again the only ones taken into account ) will correspond to a subset of the afferents that is essentially random , and will not contain much information about their real activities [46] . A solution to this problem is to use populations of redundant neurons ( with similar selectivity ) to ensure the first presynaptic spikes do correspond on average to the most active populations of afferents . In this work we took the second hypothesis , assuming the time to first spike of the afferents ( or , to be precise , their firing order ) was reliable and did reflect a level of activity . This second hypothesis receives experimental support . For example , recent recordings in monkeys show that IT neurons' responses in terms of spike count close to stimulus onset ( 100–150 ms time bin ) seem to be too reliable to be fit by a typical Poisson firing rate model [47] . Another recent electrophysiological study in monkeys showed that IT cell's latencies do contain information about the nature of a visual stimulus [48] . There is also experimental evidence for precise spike time responses in V1 and in many other neuronal systems ( see [49] for a review ) . Very interestingly , STDP provides an efficient way to develop selectivity to first spike patterns , as shown in this work . After convergence , the potential reached by an STDP neuron is linked to the number of early spikes in common between the current input and a stored prototype . This “early spike” versus “later spike” neural code ( while the spike order within each bin does not matter ) has not only been proven robust enough to perform object recognition in natural images but is fast to read out: an accurate response can be produced when only the earliest afferents have fired . The use of such a mechanism at each stage of the ventral stream could account for the phenomenal processing speed achieved by the visual system .
S1 cells detect edges by performing a convolution on the input images . We are using 5 × 5 convolution kernels , which roughly correspond to Gabor filters with wavelength of 5 ( i . e . , the kernel contains one period ) , effective width 2 , and four preferred orientations: π/8 , π/4 + π/8 , π/2 + π/8 , and 3π/4 + π/8 ( π/8 is there to avoid focusing on horizontal and vertical edges , which are seldom diagnostic ) . We apply those filters to five scaled versions of the original image: 100% , 71% , 50% , 35% , and 25% . There are thus 4 × 5 = 20 S1 maps . S1 cells emit spikes with a latency that is inversely proportional to the absolute value of the convolution ( the response is thus invariant to an image negative operation ) . We also limit activity at this stage: at a given processing scale and location , only the spike corresponding to the best matching orientation is propagated . C1 cells propagate the first spike emitted by S1 cells in a 7 × 7 square of a given S1 map ( which corresponds to one preferred orientation and one processing scale ) . Two adjacent C1 cells in a C1 map correspond to two 7 × 7 squares of S1 cells shifted by six S1 cells ( and thus overlap of one S1 row ) . C1 maps thus subsample S1 maps . To be precise , neglecting the side effects , there are 6 × 6 = 36 times fewer C1 cells than S1 cells . As proposed by Riesenhuber and Poggio [7] , this maximum operation is a biologically plausible way to gain local shift invariance . From an image processing point of view , it is a way to perform subsampling within retinotopic maps without flattening high spatial frequency peaks ( as would be the case with local averaging ) . We also use a local lateral inhibition mechanism at this stage: when a C1 cell emits a spike , it increases the latency of its neighbors within an 11 × 11 square in the map with the same preferred orientation and the same scale . The percentage of latency increase decreases linearly with the distance from the spike , from 15% to 5% . As a result , if a region is clearly dominated by one orientation , cells will inhibit each other and the spike train will be globally late and thus unlikely to be “selected” by STDP . S2 cells correspond to intermediate-complexity visual features . Here we used ten prototype S2 cell types , and 20 in the mixed simulation . Each prototype cell is duplicated in five maps ( weight sharing ) , each map corresponding to one processing scale . Within those maps , the S2 cells can integrate spikes only from the four C1 maps of the corresponding processing scale . The receptive field size is 16 × 16 C1 cells ( neglecting the side effects; this leads to 96 × 96 S1 cells , and the corresponding receptive field size in the original image is [96 / processing scale]2 ) . C1–S2 synaptic connections are set by STDP . Note that we did not use a leakage term . In the brain , by progressively resetting membrane potentials toward their resting levels , leakiness will decrease the interference between two successive spike waves . In our model we process spike waves one by one and reset all the potentials before each propagation , and so leaks are not needed . Finally , activity is limited at this stage: a k-winner-take-all strategy ensures at most two cells that can fire for each processing scale . This mechanism , only used in the learning phase , helps the cells to learn patterns with different real sizes . Without it , there is a natural bias toward “small” patterns ( i . e . , large scales ) , simply because corresponding maps are larger , and so likeliness of firing with random weights at the beginning of the STDP process is higher . Those cells take for each prototype the maximum response ( i . e . , first spike ) of corresponding S2 cells over all positions and processing scales , leading to ten shift- and scale-invariant cells ( 20 in the mixed case ) . We used a simplified STDP rule: where i and j refer , respectively , to the post- and presynaptic neurons , ti and tj are the corresponding spike times , Δwij is the synaptic weight modification , and a+ and a− are two parameters specifying the amount of change . Note that the weight change does not depend on the exact ti − tj value , but only on its sign . We also used an infinite time window . These simplifications are equivalent to assuming that the intensity–latency conversion of S1 cells compresses the whole spike wave in a relatively short time interval ( say , 20–30 ms ) , so that all presynaptic spikes necessarily fall close to the postsynaptic spike time , and the change decrease becomes negligible . In the brain , this change decrease and the limited time window are crucial: they prevent different spike waves coming from different stimuli from interfering in the learning process . In our model , we propagate stimuli one by one , so these mechanisms are not needed . Note that with this simplified STDP rule only the order of the spikes matters , not their precise timings . As a result , the intensity–latency conversion function of S1 cells has no impact , and any monotonously decreasing function gives the same results . The multiplicative term wij · ( 1 − wij ) ensures the weight remains in the range [0 , 1] ( excitatory synapses ) and implements a soft bound effect: when the weight approaches a bound , weight changes tend toward zero . We also applied long-term depression to synapses through which no presynaptic spike arrived , exactly as if a presynaptic spike had arrived after the postsynaptic one . This is useful to eliminate the noise due to original random weights on synapses through which presynaptic spikes never arrive . As the STDP learning progresses , we increase a+ and To be precise , we start with a+ = 2−6 and multiply the value by 2 every 400 postsynaptic spikes , until a maximum value of 2−2 . a− is adjusted so as to keep a fixed a+/a− ratio ( −4/3 ) . This allows us to accelerate convergence when the preferred stimulus is somewhat “locked , ” whereas directly using high learning rates with the random initial weights leads to erratic results . We used a threshold of 64 ( = 1/4 × 16 × 16 ) . Initial weights are randomly generated , with mean 0 . 8 and standard deviation 0 . 05 . We used an RBF network . In the brain , this classification step may be done in the PFC using the outputs of IT . Let X be the vector of C2 responses ( containing either binary detections with the first implementation or final potentials with the second one ) . This kind of classifier computes an expression of the form: and then classifies based on whether or not f ( X ) reaches a threshold . Supervised learning at this stage involves adjusting the synaptic weights c so as to minimize a ( regularized ) error on the training set [27] . The Xi correspond to C2 responses for some training examples ( 1/4 of the training set randomly selected ) . The full training set was used to learn the ci . We used σ = 2 and λ = 10−12 ( regularization parameter ) . The multiclass case was handled with a “one-versus-all approach . ” If n is the number of classes ( here , three ) , n RBF classifiers of the kind “class I” versus “all other classes” are trained . At the time of testing , each one of the n classifiers emits a ( real-valued ) prediction that is linked to the probability of the image belonging to its category . The assigned category is the one that corresponds to the highest prediction value . The spike trains coming from C1 cells were converted into real-valued activities ( supposed to correspond to firing rates ) by taking the inverse of the first spikes' latencies ( note that these activities do not correspond exactly to the convolution values because of the local lateral inhibition mechanism of layer C1 ) . The activities ( or firing rates ) of S2 units were computed as: where WS2 is the synaptic weight vector of the S2 cell . Note that the normalization causes an S2 cell to respond maximally when the input vector XC1 is collinear to its weight vector WS2 ( neural circuits for such normalization have been proposed in [27] ) . Hence WS2 ( or any vector collinear to it ) is the preferred stimulus of the S2 cell . With another stimulus XC1 the response is proportional to the cosine between WS2 and XC1 . This kind of tuning has been used in extensions of HMAX [26] . It is similar to the Gaussian tuning of the original HMAX [7] , but it is invariant to the norm of the input ( i . e . , multiplying the input activities by 2 has no effect on the response ) , which allows us to remain contrast-invariant ( see also [26] for a comparison between the two kinds of tuning ) . Only the cells whose activities were above a threshold were considered in the competition process . It was found useful to use individual adaptative thresholds: each time a cell was among the winners , its threshold was set to 0 . 91 times its activity ( this value was tuned to get approximately the same number of weight updates as with STDP ) . The competition mechanism was exactly the same as before , except that it selected the most active units and not the first one to fire . The winners' weight vectors were updated with the following modified hebbian rule: a is the learning rate . It was found useful to start with a small learning rate ( 0 . 002 ) and to geometrically increase it every ten iterations . The geometric ratio was set to reach a learning rate of 0 . 02 after 2 , 000 iterations , after which the learning rate stayed constant . Here we summarize the differences between our model and their model [6] in terms of architecture ( leaving the questions of learning and temporal code aside ) . We process various scaled versions of the input image ( with the same filter size ) , instead of using various filter sizes on the original image: S1 level , only the best matching orientation is propagated; C1 level , we use lateral inhibition ( see above ) ; S2 level , the similarity between a current input and the stored prototype is linked to the number of early spikes in common between the corresponding spike trains , while Serre et al . use the Euclidian distance between the corresponding patches of C1 activities . We used an RBF network and not a Support Vector Machine . | The paper describes a new biologically plausible mechanism for generating intermediate-level visual representations using an unsupervised learning scheme . These representations can then be used very effectively to perform categorization tasks using natural images . While the basic hierarchical architecture of the system is fairly similar to a number of other recent proposals , the key differences lie in the level of description that is used—individual neurons and spikes—and in the sort of coding scheme involved . Essentially , we have found that a combination of a temporal coding scheme where the most strongly activated neurons fire first with spike timing dependent plasticity leads to a situation where neurons in higher order visual areas will gradually become selective to frequently occurring feature combinations . At the same time , their responses become more and more rapid . We firmly believe that such mechanisms are a key to understanding the remarkable efficiency of the primate visual system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"primates",
"neuroscience",
"computational",
"biology"
] | 2007 | Unsupervised Learning of Visual Features through Spike Timing Dependent Plasticity |
Osteoporosis is a common polygenic disease and global healthcare priority but its genetic basis remains largely unknown . We report a high-throughput multi-parameter phenotype screen to identify functionally significant skeletal phenotypes in mice generated by the Wellcome Trust Sanger Institute Mouse Genetics Project and discover novel genes that may be involved in the pathogenesis of osteoporosis . The integrated use of primary phenotype data with quantitative x-ray microradiography , micro-computed tomography , statistical approaches and biomechanical testing in 100 unselected knockout mouse strains identified nine new genetic determinants of bone mass and strength . These nine new genes include five whose deletion results in low bone mass and four whose deletion results in high bone mass . None of the nine genes have been implicated previously in skeletal disorders and detailed analysis of the biomechanical consequences of their deletion revealed a novel functional classification of bone structure and strength . The organ-specific and disease-focused strategy described in this study can be applied to any biological system or tractable polygenic disease , thus providing a general basis to define gene function in a system-specific manner . Application of the approach to diseases affecting other physiological systems will help to realize the full potential of the International Mouse Phenotyping Consortium .
Studies of extreme phenotypes in humans have been instrumental in identifying molecular mechanisms underlying rare single gene disorders as well as common and chronic diseases including diabetes and obesity . Such studies have resulted in novel treatments that revolutionize the lives of affected individuals [1]–[4] . Collection of suitable cohorts , however , is expensive and takes many years to achieve , and progress has been limited to conditions in which simple and quantitative phenotypes can be defined [4]–[7] . By analogy , we hypothesized that an organ-specific extreme phenotype screen in knockout mice would more rapidly identify new genetic determinants of disease and also provide in vivo models to elucidate their molecular basis . The International Knockout Mouse Consortium ( IKMC ) has now established an ideal resource of mutant ES cells to test this hypothesis [8] , [9] . The skeleton represents a paradigm organ system and osteoporosis is an important global disease ideally suited to such an approach . Osteoporosis is the commonest skeletal disorder affecting hundreds of millions of people worldwide and costing tens of billions of pounds each year [10] . Between 50 and 85% of the variance in bone mineral density ( BMD ) is genetically determined [11] , but only 3% is accounted for by known genetic variation [12] and the vast majority of genes involved remain to be identified [13] . Current treatments reduce fracture risk by only 25–50% [14] , [15] and thus there is urgent need to define new pathways that regulate bone turnover and strength in order to identify novel therapeutic targets . Accordingly , application of an extreme phenotype approach to study skeletal disorders in humans has already led to discovery of SOST ( ENSG00000167941 ) and LRP5 ( ENSG00000162337 ) as critical regulators of Wnt signaling in bone [5] , [7] , [16] and resulted in development of new drugs to stimulate bone formation [17] . The Wellcome Trust Sanger Institute Mouse Genetics Project ( MGP ) is undertaking high-throughput production of knockout mice using targeted ES cells generated by the IKMC . Knockout mice are generated using a knockout-first conditional gene targeting strategy ( Figure S1 ) , in which expression from the targeted allele can be investigated by X-gal staining for LacZ gene expression [9] ( Figure S2 ) . Each mouse undergoes a broad-based primary screen to identify developmental , anatomical , physiological and behavioral phenotypes [18]–[20] . A critical challenge now is to enhance this initial screening by developing organ- or disease-specific approaches [21] that are essential to identify biologically significant and functionally relevant phenotypes rapidly and cost-effectively for the benefit of the scientific community [18] , [19] , [21] . We , therefore , developed high-throughput skeletal phenotyping methods and prospectively studied 100 consecutive unselected mutant strains from the MGP . Using this approach , we discovered nine new genetic determinants of bone mass and strength and identified a novel functional classification of bone structure . These conditional knockout mice [9] can now be used to investigate cell-specific gene function and identify new regulatory pathways in the skeleton . The strategy can be applied to other physiological systems and complex diseases , thus realizing the full potential of the International Mouse Phenotyping Consortium .
Mice generated by the MGP pipeline undergo a broad primary phenotype screen followed by terminal collection of blood and tissue at 16 weeks of age [20] . The screen is conducted on viable homozygote mutants , or heterozygotes in cases of embryonic lethality , and reports 233 variables relating to 28 physiological systems that include embyrogenesis; reproduction; growth; neurological; behaviour; sensory; skeleton; muscle; gastrointestinal and hepatobiliary; cardiovascular; endocrine; adipose; metabolism; haematopoietic; immune; skin and pigmentation; respiratory; and renal . Parameters relevant to the skeleton include body length , x-ray skeletal survey , dual energy x-ray absorptiometry ( DEXA ) analysis of BMD and biochemical measures of mineral metabolism . Tissues in which the targeted gene is expressed are determined by staining for lacZ reporter gene expression in heterozygous mice ( Figure S2 ) . To extend this broad initial screen , we incorporated novel imaging , statistical and biomechanical approaches for the specific and sensitive detection of functionally important skeletal abnormalities ( Figure S3 ) . In order to establish strain-specific reference ranges for these new approaches , limbs from 16 week-old female C57BL/6 ( B6Brd;B6Dnk;B6N-Tyrc-Brd ) wild-type mice ( n = 77 ) were obtained from 18 control cohorts . Normal ranges for six independent parameters of bone structure were obtained using Faxitron x-ray point projection microradiography and micro-computed tomography ( micro-CT ) ( Fig . 1 and Figure S4 ) . Bone mineral content ( BMC ) , bone length and cortical bone thickness were determined by x-ray microradiography and measures of trabecular bone volume per tissue volume ( BV/TV ) , trabecular number ( Tb . N ) and trabecular thickness ( Tb . Th ) by micro-CT ( Fig . 1 ) . Reference data were also obtained for six biomechanical parameters ( Fig . 2 and Figure S5 ) . The yield , maximum and fracture loads , stiffness and the proportions of energy dissipated prior to maximum load and fracture were determined from load displacement curves obtained in destructive 3-point bend tests ( Fig . 2 ) . Limbs from 16 week-old female knockout mice in an identical C57BL/6 genetic background were obtained prospectively from the MGP pipeline ( n = 100 unselected knockout strains , 2–6 mice per strain ) and analyzed by x-ray microradiography , micro-CT and 3-point bend testing . Serendipitously , one of the 100 unselected strains was a homozygous knockout of Sparc ( ENSMUSG00000018593 ) , which encodes the extracellular matrix glycoprotein osteonectin . Deletion of Sparc is known to cause low bone turnover osteopenia resulting in weak and brittle bone with a higher mineral-to-matrix ratio due to reduced bone matrix content [22] , [23] . Thus , Sparc knockout mice represented a well-characterized positive control for validation of our approach . Consistent with the reported phenotype [22] , [23] , we identified that Sparc knockout mice had reduced BMD and BMC with loss of trabecular bone but preservation of cortical bone thickness ( Fig . 1 and Table S1 ) , resulting in weak and brittle bone of reduced stiffness ( Fig . 2 ) . We also identified short stature in Sparc knockout mice ( Fig . 1A ) , a parameter not investigated in previous studies . These findings validate the use of complementary and multi-parameter imaging together with biomechanical methods as a rapid and specific phenotyping approach to identify biologically significant and functionally relevant skeletal abnormalities using a minimal number of animals ( n = 2 ) . To identify new genetic determinants of bone mass and strength , limbs from 100 knockout strains were analyzed for each structural and biomechanical variable . X-ray microradiography and micro-CT imaging identified 19 knockout strains in which at least one structural parameter was >2 . 0 standard deviations ( SD ) from the reference mean ( Fig . 3 , Figure S4 and Table S1 ) . To ensure that significant abnormal phenotypes resulting from simultaneous but smaller variances in any of the six parameters were not overlooked , Mahalanobis distances were calculated as detailed in the methods and principal component analysis performed to identify multivariate outliers [24]–[27] . These studies identified 40 strains with outlier Mahalanobis distances ( P<0 . 025 ) , 24 of which had not been identified by analysis of individual x-ray microradiography or micro-CT values alone ( Fig . 3 and Table S1 ) . The MGP broad primary phenotype screen independently annotated 17 of these knockout strains with skeletal abnormalities ( Table S1 ) . Nine of the strains were also identified as outliers by x-ray microradiography , micro-CT or statistical methods whereas 8 did not display any abnormalities ( Fig . 3 ) . The biomechanical significance of the 43 outlier phenotypes identified by imaging ( n = 19 , Faxitron and micro-CT ) and statistical ( n = 24 , Mahalanobis analysis but excluding Faxitron and micro-CT ) approaches , together with the 8 additional strains identified only in the MGP primary screen , was investigated ( Fig . 3 ) . Twelve of the 51 strains had at least one biomechanical parameter >2 . 0 SD from the reference mean ( Fig . 3 , Figure S5 and Table S1 ) . However , 2 of these 12 strains had only minor abnormalities of bone morphology in the primary screen and were normal when investigated by x-ray microradiography , micro-CT and principal component analysis . Destructive 3-point bend testing of bones from the remaining 49 strains identified a further 7 with a single outlier biomechanical parameter but no other abnormality ( Fig . 3 and Table S1 ) . In summary , the broad primary phenotype screen together with x-ray microradiography , micro-CT and statistical analysis identified knockout strains with at least one abnormal bone-related parameter . The addition of functional biomechanical testing demonstrated that 10 of these strains had major phenotypes affecting both the structure and strength of bone . Three of these carried heterozygous mutations ( Asxl1 ( ENSMUSG00000042548 ) , Setdb1 ( ENSMUSG00000015697 ) and Trim45 ( ENSMUSG00000033233 ) ) while the rest were homozygotes ( Bbx ( ENSMUSG00000022641 ) , Cadm1 ( ENSMUSG00000032076 ) , Fam73b ( ENSMUSG00000026858 ) , Prpsap2 ( ENSMUSG00000020528 ) , Slc38a10 ( ENSMUSG00000061306 ) , Sparc and Spns2 ( ENSMUSG00000040447 ) ) . The primary phenotype database ( http://www . sanger . ac . uk/mouseportal/ ) was interrogated for each of the strains identified with major skeletal phenotypes ( Table 1 ) . The sensitivities , specificities and predictive values of each phenotyping method at a statistical threshold of >2 . 0 SD and 95% confidence limit were calculated to determine their ability to identify the 10 strains with major phenotypes . X-ray microradiography was the most accurate , identifying 8 of the 10 strains . Additional use of micro-CT and Mahalanobis analysis was required to identify the remaining 2 strains ( Table S2 ) . The MGP primary phenotype screen identified 5 of the 10 abnormal strains . Thus , addition of organ-specific imaging , statistical and biomechanical analyses to the primary phenotype screen resulted in increased sensitivity and specificity ( Table S2 ) , demonstrating the advantage of a complementary multi-parameter approach . Correlations between imaging and biomechanical parameters were also determined to investigate relationships between bone structure and strength ( Figure S6 ) . Bone strength correlated strongly with BMC , cortical thickness and bone length but not with BV/TV , Tb . Th or Tb . N . Furthermore , there was no significant relationship between bone structural parameters determined by x-ray microradiography and micro-CT , thus demonstrating independence of the two techniques . Consistent with sensitivity , specificity and predictive value data ( Table S2 ) these findings demonstrate that mineralization and cortical bone parameters determined by x-ray microradiography are excellent predictors of bone strength determined by 3-point bend testing , whereas trabecular bone structure determined by micro-CT is not . Importantly , trabecular bone parameters determined by micro-CT may represent good predictors of bone strength at sites of predominantly cancellous bone such as the vertebra or in response to age-related bone loss , although these possibilities were not investigated . Biomechanical analysis of wild-type mice and the 10 knockout strains with major phenotypes enabled four distinct biomechanical categories to be identified ( Fig . 4 ) . Normal bone had a stiffness of 30 . 2±4 . 1 N/mm ( mean ± SD ) with the capability to resist loading up to a yield load of 8 . 9±0 . 9 N ( mean ± SD ) and maximum load of 10 . 7±0 . 9 N ( mean ± SD ) , prior to fracture at a load of 6 . 4±1 . 7 N ( mean ± SD ) . These material properties of normal bone represent an optimised compromise between strength and flexibility that allows dissipation of 86 . 2% of energy prior to failure and limits the structural damage at fracture ( Table S1 ) . Bones from strains with major phenotypes clustered into three abnormal biomechanical categories . Bones from Bbx , Cadm1 and Fam73b knockout mice were weak but flexible with reduced maximum load but the capability to bend and dissipate energy prior to fracture ( Fig . 4B ) . Sparc , Prpsap2 and Slc38a10 bones were weak and brittle with reduced maximum load and lacking the capability to bend and dissipate energy prior to fracture ( Fig . 4C ) . Asxl1 , Trim45 , Spns2 and Setdb1 bones were strong but brittle with an increased maximum load but were unable to bend and dissipate energy prior to fracture ( Fig . 4D ) . A plot of the proportion of energy dissipated prior to fracture versus maximum load clearly separates these categories of bone strength ( Fig . 4E ) . Further analysis demonstrated that Bbx , Cadm1 , Fam73b Sparc , Prpsap2 and Slc38a10 bones have low BMC , whereas Asxl1 , Trim45 , Spns2 and Setdb1 bones have high BMC ( Fig . 5 ) . To determine whether these three categories of abnormal bone strength were related to a further morphological parameter known to have an important role for the biomechanical properties of long bones , we investigated their relationship with mid-diaphyseal diameter ( Fig . 6 ) . As expected [28] , in WT mice mid-diaphyseal diameter correlated with fracture load and the proportion of energy dissipated prior to fracture . However , bones from the mutant strains identified with major skeletal phenotypes clustered into the same three abnormal categories , further demonstrating the validity of this functional classification . Investigation of the biological activities of the 10 genes and their possible roles in bone indicates a broad diversity of function that was not clearly related to phenotype ( Table 1 ) , thus reinforcing the importance of an unbiased screening approach for identification of gene function in both homozygous and heterozygous mutants . In summary , a functional classification of four categories of bone structure was defined . Normal bone is strong and flexible with a normal mineral content , whereas abnormal bone is either ( i ) weak but flexible with low BMC , ( ii ) weak and brittle with low BMC or ( iii ) strong but brittle with high BMC .
We have identified 10 genes with diverse and unrelated functions , the deletion of which resulted in major skeletal abnormalities . By adopting a multi-parameter phenotyping approach , we identified a new functional classification of bone structure based on its mineral content , strength and ductility that clarifies understanding of skeletal physiology and pathology , and which maps directly to human disease . As a result of evolutionary pressure , bone structure represents an optimal compromise between strength and flexibility that requires contributions from many diverse genes . Continuous bone remodeling enables the skeleton to adjust this compromise in response to changing physiological and environmental pressures [29] , [30] . The current studies demonstrate that loss of function of individual genes can disrupt this optimal compromise resulting in skeletal phenotypes that cluster into three functionally distinct categories . Postmenopausal osteoporosis is characterized by weak but flexible bone with low mineral content [31] and three of the identified knockout strains ( Bbx , Cadm1 , Fam73B ) had phenotypes in this category . Bbx encodes a conserved transcription factor that contains a SOX-TCF HMG-box [32]–[36] . Family members include SRY-related Sox genes that are implicated in skeletal dysplasias [37] , [38] and the TCF/LEF transcription factors that mediate Wnt/β-catenin signaling [39] , a key pathway implicated in osteoporosis and osteoarthritis [40]–[43] . Cadm1 encodes a trans-membrane glycoprotein adhesion molecule of the immunoglobulin superfamily [44] for which a number of disparate functions have been reported including; tumor suppression [45] , synapse development [46] , behavioral regulation [47] , T cell adhesion [48] , mast cell interactions [49] , and spermatogenesis [50] . However , no function in the skeleton has been reported . Fam73B encodes a conserved membrane protein of unknown function . These findings indicate that deletion of genes encoding proteins with diverse and unrelated functions can result in similar defects of bone strength and mineralization . Disorders of bone matrix as typified by osteogenesis imperfecta [51] are characterized by bone that is weak and brittle with low BMC , and three of the strains ( Prpsap2 , Slc38a10 , Sparc ) displayed this phenotype . Prpsap2 encodes the non-catalytic inhibitory subunit of phosphoribosylpyrophosphate synthetase [52] , and is required for synthesis of purine and pyrimidine nucleotides , the amino acids histidine and tryptophan , and the coenzyme nicotinamide adenine dinucleotide [53] . Its function in the skeleton is unknown , although a recent study proposed PRPSAP2 as a candidate oncogene in osteosarcoma tumorigenesis [54] . Slc38a10 encodes a proposed sodium-coupled neutral amino acid membrane transporter [55] that may act as a cell volume regulator in mesenchyme [56] . The severe growth defect in Slc38a10 knockout mice suggests a critical function in chondrocytes , which mediate linear growth by cell volume expansion during hypertrophic differentiation [57] . Furthermore , related transporters have already been implicated in human skeletal disease . SLC35D1 is critical for chondroitin sulphate synthesis and mutations cause Schneckenbecken skeletal dysplasia [58] . Mutations in SLC26A2 , cause four distinct chondrodysplasia syndromes [59] and emphasize the key role of these transporters in endochondral ossification . Sparc encodes the well-described extracellular matrix glycoprotein osteonectin and its deletion resulted in the characteristic and expected phenotype [22] , [23] of weak and brittle bone with low BMC . These findings highlight the importance of enzymes , transporters and structural proteins to the functional integrity of bone matrix . Diseases of high bone mass are rare and include sclerosteosis due to deletion of SOST [60] and autosomal dominant high bone mass due to gain-of-function mutations in LRP5 [61] . They are characterized by bone that is strong but brittle with high BMC , and four of the knockout strains ( Asxl1 , Setdb1 , Spns2 , Trim45 ) displayed such a phenotype . Asxl1 encodes a polycomb protein that interacts with heterochromatin protein-1 [62] and is required for regulation of Hox genes during axial patterning [63] , suggesting a role in skeletal development [64] , [65] . Indeed , ASXL1 heterozygous nonsense mutations were recently described to cause Bohring-Opitz syndrome [66] , a developmental disorder characterized by mental retardation , impaired intrauterine growth , trigonocephaly and wrist and metacarpophalangeal joint abnormalities . Although the disease mechanism is unknown , craniofacial defects identified in homozygous Asxl1 knockout mice suggest that mutations in Bohring-Opitz syndrome result in a mutant protein with dominant-negative activity . Setdb1 encodes a histone H3 methyltransferase that regulates gene silencing [67] , [68] . Although found to be expressed in cartilage but not bone in the primary phenotype screen , other studies demonstrated Setdb1 expression in osteoblasts and suggested a role in lineage commitment and differentiation [69] , [70] . Spns2 encodes a sphingosine 1-phosphate ( S1P ) transporter [71] that is essential for S1P secretion . S1P binds to the G-protein coupled receptors , S1PR1 and S1PR2 , and regulates osteoclast [72] , [73] and osteoblast [74] precursor cell recruitment and migration . Thus , control of S1P secretion by Spns2 represents a novel mechanism that couples bone resorption and formation [75] . Trim45 is a member of the tripartite protein family , many of which act as ubiquitin or SUMO E3 ligases [76]–[78] . Although restricted to brain and testis in the primary phenotype screen , human studies demonstrate that Trim45 is more widely expressed [79] . Little is known about its function , although one study indicates Trim45 interacts with AP-1 and inhibits activity of the MAP kinase pathway [79] . The physiological significance of these findings and the role of Trim45 in the skeleton are unknown , although AP-1 proteins are key regulators of osteoblast and osteoclast differentiation and function [80] , [81] . These findings emphasize the importance of lineage commitment , control of cell differentiation and coupling of both osteoblasts and osteoclasts in high bone mass disorders . In the context of osteoporosis , our identification of many new genes that determine bone strength , and which otherwise could not be predicted , is consistent with studies indicating that diverse genetic polymorphisms result in small effects on phenotype [11] , [12] , [82] , [83] . Accordingly , and in line with current understanding that only 3% of the heritability of BMD is accounted for by known genetic variation [12] , none of the genes identified in this study have been recognized in osteoporosis genome-wide association studies [84] . We hypothesize , therefore , that unbiased multi-parameter and functional phenotyping of knockout mice has the power to identify many of the major genes that determine bone strength . Ultimately , this approach is likely to identify several genes from a single signaling pathway with an important role in the control of bone mass and strength . This has the advantage of independently confirming critical pathways and the potential to identify several alternative therapeutic targets . Importantly , however , the approach has limitations . The study of knockout mice can only identify phenotypes that result from gene deletion but cannot identify genes that only cause abnormalities when they harbor gain-of-function or dominant-negative mutations . Furthermore , the strategy does not include challenges such as ageing that may reveal additional phenotypes . However , if such provocative challenges were to be incorporated into screening approaches they would inevitably increase costs and limit throughput . Our findings resulted from development and refinement of a rapid-throughput phenotyping algorithm to identify knockout mice with major abnormalities of bone structure and strength ( Figure S3 ) . The methods require bones from only two knockout mice , which first undergo digital point projection x-ray microradiography and micro-CT determination of six parameters of bone structure . Mahalanobis distance calculations and principal component analysis is performed and strains with at least one structural parameter >2 . 0 SD from the reference mean plus those with outlier Mahalanobis distances ( 95% confidence limit ) are selected for biomechanical studies . Bones from selected strains undergo destruction 3-point bend testing to determine six measures of bone strength . Application of this unbiased approach to 100 consecutive knockout strains from the MGP pipeline identified 10% with major phenotypes affecting bone strength . Subsequent consideration of the results of primary phenotype screening and biological plausibility ( Figure S3 ) allowed selection of mice to be refined . Inherent in this approach is the capability to alter the statistical stringency threshold of analyses such that the number of strains for subsequent functional studies can be adjusted according to phenotype severity . For example , if the threshold for structural parameters is increased from 2 . 0 to 3 . 0 SD , then 9 outlier strains ( rather than 19 ) are identified . Furthermore , if the confidence limit for Mahalanobis distance is increased from 95 to 99 . 7% then 21 multivariate outliers ( rather than 40 ) would be identified . Of note , Trim45 , which was recognized as an outlier only by Mahalanobis analysis , would still be identified if the confidence limit were to be increased to 99 . 7% , thus emphasizing the importance of a robust statistical method to ensure that all functional outliers are captured . Biomechanical analysis following application of these more stringent thresholds would detect 8 outlier strains including Trim45 and Sparc , resulting in the identification of 7 novel determinants of bone mass and strength rather than 9 . The intrinsic flexibility of such a bespoke approach facilitates its application to other biological systems or polygenic diseases .
All mouse studies were undertaken by Wellcome Trust Sanger Institute Mouse Genetics Project as part of the International Knockout Mouse Consortium and licensed by the UK Home Office in accordance with the Animals ( Scientific Procedures ) Act 1986 and the recommendations of the Weatherall report . All mice generated by the MGP undergo a broad primary phenotype screen ( http://www . sanger . ac . uk/mouseportal/ ) that includes measurement of body length , x-ray skeletal survey , DEXA analysis of bone mineral density and biochemical measures of mineral metabolism performed between 14–16 weeks of age , and determination of the normal tissue expression pattern of the targeted gene in 6–12 week old mice . Following primary phenotyping , lower limbs were fixed in 70% ethanol . The pattern of LacZ reporter gene expression was determined in whole mount tissue preparations from heterozygous knockout mice between 6 and 12 weeks of age . Under terminal anaesthesia , mice were perfused with fresh cold 4% paraformaldehyde ( PFA ) . Tissues were fixed for a further 30 min in 4% PFA , rinsed in phosphate buffered saline and stained with 0 . 1% X-gal for 48 hours at 4°C . Samples were subsequently fixed overnight in 4% PFA at 4°C , cleared with 50% glycerol and transferred to 70% glycerol . Specific and non-specific staining was determined in 41 tissues ( Figure S2 ) . A panel of 27 standardized images were recorded if expression was widespread ( www . sanger . ac . uk/mouseportal ) . Bones from 16 week-old mice were fixed in 70% ethanol . Soft tissue was removed from the fixed bones and digital X-ray images were recorded at 10 µm pixel resolution using a Faxitron MX20 variable kV point projection x-ray source and digital image system ( Qados , Cross Technologies plc , Sandhurst , Berkshire , UK ) operating at 26 kV and 5× magnification [85] . Magnifications were calibrated by imaging a digital micrometer . Bone mineral content , bone length and cortical bone thickness were determined with coefficients of variation ( CV ) of 1 . 7% , 2 . 0% and 5 . 1% , respectively . The relative mineral content of calcified tissues was determined by comparison with standards included in each image frame , which comprised: a 1 mm steel plate; a 1 mm diameter spectrographically pure aluminum wire; and a 1 mm diameter polyester fiber . 2368×2340 16 bit DICOM images were converted to 8 bit Tiff images using ImageJ and the histogram stretched from the polyester ( grey level 0 ) to steel ( grey level 255 ) standards . Increasing gradations of mineralization density were represented in 16 equal intervals by a pseudocolor scheme . Cortical bone thickness was determined in at least 10 locations at the mid-femoral diaphysis . Bone length was determined using ImageJ 1 . 41 software ( http://rsb . info . nih . gov/ij/ ) . Tibias were analyzed by micro-CT ( Skyscan 1172a , Skyscan , Belgium ) at 50 kV and 200 µA using a 0 . 5 mm aluminum filter and a detection pixel size of 4 . 3 µm2 [85] . Images were captured every 0 . 7° , with 2× averaging , through 180° rotation of each bone and reconstructed using Skyscan NRecon software . A volume of 1 mm3 of trabecular bone was selected as the region of interest , 0 . 2 mm from the growth plate . Trabecular bone volume as proportion of tissue volume ( BV/TV , % , CV 18 . 4% ) , trabecular thickness ( Tb . Th , mm , CV 11 . 1% ) and trabecular number ( Tb . N , mm−1 , CV 17 . 4% ) were analyzed [86] using Skyscan CT analysis software . Bones were stored and tested in 70% ethanol . Destructive 3-point bend tests were performed on an Instron 5543 materials testing load frame ( Instron Limited , High Wycombe , Buckinghamshire , UK ) using custom built mounts with rounded supports that minimize cutting and shear loads [85] . Bones were positioned horizontally and centered on custom supports with the anterior surface upward . Load was applied vertically to the mid-shaft with a constant rate of displacement of 0 . 03 mm/second until fracture . A span of 12 mm was used . Load-displacement curves were plotted and yield load , maximum load and fracture load determined . Stiffness , the slope of the linear ( elastic ) part of the load-displacement curve , was calculated by the “least squares” method . Work energy was calculated from the area under curve at both maximum load and fracture . Elastic stored energy at maximum load was determined by calculating the area of a right angled triangle with the vertex at the point of maximum load and hypotenuse with a slope equal to that of the linear phase of the load-displacement curve . Elastic stored energy at fracture was similarly calculated but with the vertex of the triangle at the point of fracture ( Fig . 2 ) . Energy dissipated at maximum load or fracture was calculated by subtracting the elastic stored energy from the work energy at maximum load or fracture . CVs for each parameter were as follows: yield load ( 9 . 8% ) , maximum load ( 8 . 5% ) , fracture load ( 26 . 6% ) , stiffness ( 13 . 6% ) , the ratio of energy dissipated at maximum load to elastic stored energy at maximum load ( 25 . 1% ) , and the ratio of energy dissipated prior to fracture to elastic stored energy at fracture ( 11 . 0% ) . C57BL/6 reference ranges were generated for all Faxitron and micro-CT measures . Outliers in multivariate data were identified using robust Mahalanobis distances [27] , which measure how far each observation is from the center of a data cluster , taking into account the shape of the cluster [87] . Robust Mahalanobis distances were calculated for the vector of multivariate observations as described [27] . Here is a robust ( i . e . relatively unaffected by outliers ) estimate of the mean vector and is a robust estimate of the covariance matrix of the data set . Under the assumption of multivariate normality , if mouse i is from the same population as the rest of the data then has a chi-squared distribution with p degrees of freedom ( where p is the number of variables ) . Robust estimates of the mean and covariance matrix are used so that potential outliers are not masked . The masking effect , by which outliers do not necessarily have a large Mahalanobis distance , can be caused by a small cluster of outliers that attract the mean and inflate the covariance in its direction . By replacing the sample mean and covariance with a robust estimate , the influence of these outliers is removed and the Mahalanobis distance is able to expose all outliers . Robust estimates of the mean and covariance matrix were calculated using the minimum volume ellipsoid method [87] . Given observations and variables , the minimum volume ellipsoid method seeks an ellipsoid containing points of minimum volume . All analysis was conducted in the statistical computing package R ( http://www . R-project . org ) . Principal component analysis was used as a method to visualize multivariate data and reveal outliers by describing variation in a set of correlated variables in terms of a new set of uncorrelated variables . These new variables or principal components are linear combinations of the original variables derived in decreasing order of importance so that the first component accounts for the most variation of all possible linear combinations . The second component is then selected so that it accounts for as much of the remaining variance as possible ( subject to it being uncorrelated with the first component ) , and so on [24] . Since the first few principal components often contain most of the variation in the data set they can be used as a lower-dimensional summary of the original variables . Normally distributed data were analyzed by Student's t test , or ANOVA followed by Tukey's multiple comparison post-hoc test . Relationships between bone structure and biomechanical measures , and between individual bone structure parameters , were determined by Pearson correlation . P values<0 . 05 were considered significant . Frequency distributions of bone mineral densities obtained by Faxitron were compared using the Kolmogorov-Smirnov test , in which P values for the D statistic in 1024 pixel data sets were D = >6 . 01 P<0 . 05 , D = >7 . 20 P<0 . 01 , and D = >8 . 62 P<0 . 001 . | Chronic disease represents a global healthcare burden but its genetic basis is largely unknown . To address this , a massive international investment is generating a resource of mutant mice to investigate the function of every gene . Although current characterization of mutants is broadbased , it lacks specificity . Here , we describe a new and rapid functional screening approach to identify genes involved in disease susceptibility . Using bone and osteoporosis as a paradigm , we identify nine new genes that determine bone structure and strength from a screen of 100 knockout mice . Deletion of five of the genes leads to low bone mass , whereas deletion of four results in high bone mass . We also report a novel functional classification that relates bone structure to bone strength and opens the field to collaborative research between material scientists , bioengineers and biologists . Our rapid throughput phenotyping approach can be applied to complex diseases in other physiological systems , thus realizing the full potential of the International Mouse Phenotyping Consortium . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"rheumatology",
"animal",
"genetics",
"integrative",
"physiology",
"anatomy",
"and",
"physiology",
"biomechanics",
"gene",
"function",
"animal",
"models",
"genetics",
"of",
"disease",
"physiological",
"processes",
"model",
"organisms",
"bone",
"homeostasis",
... | 2012 | Rapid-Throughput Skeletal Phenotyping of 100 Knockout Mice Identifies 9 New Genes That Determine Bone Strength |
Emerging evidence implies that STAT6 plays an important role in both the adaptive and innate immune responses to virus infection . Kaposi’s sarcoma-associated herpesvirus ( KSHV ) is an oncogenic γ-herpesvirus agent associated with several human malignancies , including Kaposi’s sarcoma ( KS ) and primary effusion lymphomas ( PELs ) . Previously , we demonstrated that KSHV blocks IL-4-induced STAT6 phosphorylation and retains a basal IL-13/STAT6 constitutive activation for cell survival and proliferation . However , the mechanism by which KSHV regulates STAT6 remains largely unknown . Here , we found that KSHV-encoded LANA interacts with STAT6 and promotes nuclear localization of STAT6 independent of the tyrosine 641-phosphorylation state . Moreover , nuclear localization of STAT6 is also dramatically increased in KS tissue . The latent antigen LANA induces serine protease-mediated cleavage of STAT6 in the nucleus , where the cleaved STAT6 lacking transactivation domain functions as a dominant-negative regulator to repress transcription of Replication and Transcription Activator ( RTA ) and in turn shut off viral lytic replication . Blockade of STAT6 by small interference RNA dramatically enhances expression of RTA , and in turn reduces KSHV-infected endothelial cell growth and colony formation . Taken together , these results suggest that nuclear localization and cleavage of STAT6 is important for modulating the viral latency and pathogenesis of KSHV .
Signal transducer and activator of transcription ( STAT ) is a family of latent cytoplasmic transcription factors activated by specific cytokine receptor-mediated signal transducers . Seven members of the STAT family , including STAT1 , 2 , 3 , 4 , 5a , 5b , and 6 , have been described so far [1] . STAT6 is activated by cytokines like IL-4 and IL-13 that interact with a receptor complex containing IL-4Rα chain [1] . Selective activation of STAT6 by IL-4 or IL-13 involves phosphorylation , dimerization and then translocation into the nucleus , where it binds to specific DNA elements TTC ( N3/4 ) GAA within the promoter region , activating gene transcription [2] . It has been demonstrated that STAT6 is required to induce the expression of CD23 and MHC class II , IgE isotype switching in B cells [3] , as well as differentiation of Th2 type T cells [4] . However , STAT6 blocks IL-4-dependent inhibition of IFN-γ-induced gene expression in macrophages or Th1 type T-cell differentiation , indicating that STAT6 plays a key role in negative regulation of gene expression [5 , 6] . Although little is known regarding the mechanisms of down-regulation of STAT6-dependent signaling , recent reports of STAT6 isoform with carboxyl-truncation in both bone marrow-derived mast cells and mast cell lines suggest that STAT6 could function as a dominant-negative regulator in gene expression , which , due to lack of the carboxyl-terminus , interferes with the normal ability of STAT6 to induce transcription of target genes [7 , 8] . For instance , a 70kDa carboxyl-truncated isoform of STAT6 was detected in IL-4-stimulated mast cells [7] , and this cleavage of STAT6 is induced by serine proteases in the nucleus . Interestingly , full length STAT6 ( 94kDa ) can also be cleaved at different sites to yield short STAT6 ( 60kDa and 55kDa ) in the cytoplasm of mast cells by neutrophil elastase and proteinase , respectively [8]–a phenomenon not observed in B cells . Kaposi’s sarcoma-associated herpesvirus ( KSHV ) , also known as human herpesvirus 8 ( HHV-8 ) , is the etiological agent for Kaposi’s sarcoma ( KS ) , and is causally associated with primary effusion lymphoma ( PEL ) and Multicentric Castleman disease ( MCD ) . Like other herpesviruses , KSHV infection also undergoes a two-stage life cycle: latency and lytic replication . During latency , only a limited number of genes including LANA , vFLIP , vCyclin , kaposin , and the viral microRNAs are expressed [9 , 10] , and play critical roles in cell proliferation , apoptosis , and escape of the host immune surveillance [11 , 12] . Among these genes , LANA ( encoded by ORF73 ) is the master regulator of KSHV latency . LANA not only functions as a linker to connect KSHV episome with host chromosome for maintenance of KSHV genome [13–15] , but also modulates viral and cellular gene expression by interacting with transcription factors and chromatin regulatory proteins . Moreover , LANA modulates the turnover activity of tumor suppressors like p53 and Rb which lead to chromosomal instability [16] . The switch from latency to lytic replication is mediated through another key regulator ― RTA ( Replication and Transcription Activator ) , which is encoded by ORF50 [17] . Upon induction , the mRNA transcript of RTA is expressed and acts as a transcription activator of downstream early and late genes during lytic replication for production of viral progeny [18] . In regard to the role of cytokine signaling in regulation of KSHV-mediated pathogenesis , Very little is known regarding the role of cytogenetic signaling , including STAT signaling , during KSHV latent and lytic replication , although some evidence has indicated that KSHV deregulates cytokine receptor-mediated STAT signal transduction [19–22] . In regards to STAT6 , our previous studies have shown that KSHV blocks IL-4-induced STAT6 phosphorylation favoring latency , while stimulation with IL-4 resulted in RTA expression and reactivation of viral lytic replication [22 , 23] . We recently also found that KSHV retains a basal IL-13/STAT6 constitutive activation for cell survival and proliferation [24] . However , whether STAT6 plays a role in maintaining KSHV latency , independent of IL-4 or IL-13 stimulation remains unclear . In the present study , we characterized the interaction of STAT6 and LANA , and their roles in KSHV-infected cells with regard to the mechanisms of cleaved STAT6 generation , phosphorylation status and nuclear localization . We identified that cleaved STAT6 lacking transcriptional activity acts as a dominant-negative regulator in lytic gene transcription of KSHV latently-infected cells . Distinct from other STATs , STAT6 cleavage in the KSHV latently-infected cells are due to serine proteases in the nucleus . These findings indicate that nuclear localization and cleavage of STAT6 is critical for KSHV to block lytic replication .
Previously , despite the absence of detectable phosphorylation of Y641 on STAT6 in KS tissue [24] , we clearly observed there were more nuclear localization of total STAT6 in KS tissues than in normal skin controls ( Fig 1A and 1B ) , which is further verified by the immunofluorescence assays against STAT6 and LANA ( Fig 1C and 1D ) . To further confirm if infection of KSHV enhances nuclear localization of STAT6 , endogenous STAT6 in iSLK cells with and without KSHV infection were monitored by immunofluorescence and cell fractionation assay . As shown in Fig 2A and 2B , there was an increase in the amount of nuclear localization of STAT6 in KSHV-infected cells when compared with control cells . Moreover , in HUVEC cells with KSHV primary infection , more STAT6 consistently translocated into nuclear compartment upon KSHV infection ( Fig 2C ) , further supporting the hypothesis that KSHV induces STAT6 nuclear translocation . Interestingly , we also found a cleaved isoform of STAT6 present in the nuclear compartment of both iSLK and HUVEC cells upon KSHV infection , although there was additional slow-migrated cleaved isoform of STAT6 presented in HUVEC cells ( Fig 2B and 2C ) . To further investigate if there was a similar effect STAT6 localization and cleavage in B cells , and address whether KSHV-encoded latent antigen LANA contributes to nuclear localization of STAT6 , we performed immunofluorescence assays of KSHV-infected B lymphoma PEL cells with or without LANA knockdown . As shown in Fig 3A , we observed that inhibition of LANA dramatically reduced nuclear localization of STAT6 ( from 78% to 35% ) in BC3 cells . Western blot analysis of nuclear and cytoplasm fractionation further confirmed the role of LANA in induction of STAT6 nuclear localization and cleavage ( Fig 3B ) , indicating that nuclear localization and cleavage of STAT6 is induced by LANA in response to KSHV latent infection . Our previous study demonstrated that LANA could inhibit IL-4-induced STAT6 signaling pathway and decrease STAT6 phosphorylation [23] . To address whether the nuclear localization and cleavage of STAT6 induced by LANA is independent of phosphorylation on tyrosine 641 , we performed immunofluorescence and cell fractionation assays by co-expressing LANA with wild type STAT6 or its mutant Y641F in 293 cells . Strikingly , the results revealed that both wild type STAT6 and its mutant Y641F translocated into the nucleus and was cleaved in the presence of LANA ( Fig 3C and 3D ) , although there was a moderately reduction of STAT6 nuclear localization after Y641 was mutated . To elucidate if nuclear localized STAT6 induced by LANA forms a dimer independent of phosphorylation , we performed dimerization assays using exogenous STAT6 along with IL-4 stimulation as positive control . Our results , revealed that no STAT6 dimers were induced by LANA ( Fig 3E , right panel ) . Similar results were observed using endogenous STAT6 in iSLK with or without KSHV infection ( Fig 3E , left panel ) To explore whether the nuclear translocation of STAT6 induced by KSHV is due to the interaction between STAT6 with LANA , we performed an immunoprecipitation assay in KSHV positive ( BC3 , BCBL1 ) and negative cell lines ( DG75 ) using a LANA monoclonal antibody , followed by immunoblotting against STAT6 . In accordance with our hypothesis , our results demonstrated that STAT6 was pulled down by LANA in KSHV positive cells , but not in KSHV negative cells ( Fig 4A ) . We also demonstrated that exogenous STAT6 did associate with LANA , which is mainly dependent on the amino domain of LANA ( Albeit the carboxyl terminus of LANA also presents to selectively associate with the cleaved isoform of STAT6 ) , when they were co-expressed in 293 cells ( Fig 4B ) . To further address which domain of STAT6 is required for LANA interactions , we generated three truncated mutants of STAT6 based on its functional domains , and performed a reverse immunoprecipitation assay by targeting STAT6 . As shown in Fig 4C , deletion of either the α-helics domain ( ΔN ) or the DNA-binding domain ( ΔDBD ) significantly abolished the interaction of STAT6 with LANA , when compared with wild type STAT6 or its mutants Y641F and ΔTAD ( deletion of transactivation domain ) . Given the presence of a nuclear localization sequence ( NLS ) located at the α-helics domain of STAT6 , in order to exclude the potential effect of deletion of the NLS on STAT6-LANA interactions , we evaluated the cellular localization of each of the truncated mutants in the presence or absence of LANA in 293 cells by immunofluorescence . We observed ( Fig 4D right panel and supplementary S1 Fig ) that deletion of the α-helics domain containing NLS abolished STAT6 nuclear localization regardless of whether LANA was co-expressed or not . By contrast , after the deletion of the DNA-binding domain , the nuclear localization of did not significantly change in the presence or absence of LANA when compared with wild type STAT6 or TAD deleted mutants in which the NLS are retained . This indicates that the DNA-binding domain of STAT6 is indeed required for STAT6 to interact with LANA , and that the NLS is necessary for LANA-induced nuclear translocation of STAT6 . As knockdown of STAT6 in iSLK cells with KSHV infection reduces the protein expression of LANA ( Fig 5A ) , we hoped to investigate whether the LANA-mediated nuclear translocation of STAT6 also results in an increase in LANA stability . HEK293 cells co-expressing LANA-myc with increasing amounts of FLAG-STAT6 were thus subjected to immunoblotting analysis . In agreement with our hypothesis , co-expression of STAT6 specifically enhanced the level of LANA protein in a dose-dependent manner ( Fig 5B , left panel ) . By contrast , the level of GFP-NLS-myc protein , which is driven by the same CMV promoter , did not significantly change in the presence of increasing amounts of STAT6 expression . Thus , it appears that STAT6 may stabilize LANA . To further determine the specific manner in which STAT6 contributes to LANA stability , we analyzed the protein stability of LANA co-expression with full-length STAT6 , its mutants with nuclear-localization deficient ( ΔN ) , mutation of tyrosine 641 ( Y641F ) , or lacking DNA-binding ability ( ΔDBD ) , or vector alone in 293T cells with cycloheximide treatment for 0 , 6 , 12 and 24 hours . The immunoblotting results revealed that the stability of LANA protein is significantly enhanced in the presence of full-length STAT6 , but not its mutant with nuclear-localization deficiency ΔN , ΔDBD or vector alone ( Fig 5C ) . By contrast , Y641F did not significantly impair the ability of STAT6 to stabilize LANA . Therefore , it would appear that nuclear-localized STAT6 induced by LANA enhances the stability of LANA and is independent of phosphorylation on Y641 . To determine which type of protease is critical to contribute to LANA-induced nuclear cleavage of STAT6 , HEK293 cells transfected by FLAG-STAT6 with LANA-myc or vector alone , and were subjected to treatment with different protease inhibitors followed by immunoblotting assay . Strikingly , our results revealed that LANA-mediated cleavage of STAT6 was dramatically inhibited by the serine-protease inhibitor PMSF ( an inhibitor previously also shown to block nuclear cleavage of STAT6 ) , but not other protease inhibitors ( Fig 6A , compare lane 1 , 2 , 3 with 7 , 8 , 9 ) . To further delineate the role of LANA in STAT6 cleavage , HEK293 cells were co-transfected by FLAG-STAT6 with LANA-myc or vector alone , were treated with different doses of proteasomal inhibitor MG132 for 2 hours prior to harvest . HA-tagged STAT3 was used as a parallel control . Our immunoblotting results revealed that LANA specifically induces cleavage of STAT6 but not STAT3 , and that this LANA-mediated cleavage of STAT6 was also sensitive to proteasomal inhibitor MG132 ( Fig 6B ) . To further define if LANA-mediated STAT6 cleavage requires specific interactions between STAT6 and LANA , we determined the LANA-induced cleavage profile of STAT6 truncated mutants along with full length and Y641F mutant . Consistent with our hypothesis , both ΔN ( which lacks the ability to localize in the nucleus ) and ΔDBD ( which loses the ability to associate with LANA ) STAT6 mutants did not undergo cleavage in the presence of LANA . By contrast , the ΔTAD mutant ( which retains the ability to associate with LANA and localizes in nucleus ) appeared as cleaved STAT6 band , similar to cleaved bands from full-length STAT6 ( Fig 6C , left panel ) . Unexpectedly , we also observed that the Y641F mutation greatly enhanced the LANA-induced cleavage of STAT6 , when compared with wild type STAT6 ( Fig 6C , left panel ) . In addition , based on the observation of three cleaved STAT6 bands induced by LANA , there are at least three potential cleaved sites located at the carboxyl terminus of STAT6 ( Fig 6C , right panel ) . Importantly , as shown in Fig 6D , in the absence of PMSF and MG132 treatment , we did observe that more cleaved isoforms of STAT6 in the KSHV-positive B lymphoma cells ( BC3 , BCBL1 and JSC1 ) and KSHV-infected cells ( K-BJAB , and K-iSLK ) , supporting the conclusion that KSHV induces STAT6 cleavage . To elucidate the physiologically functional link between nuclear localization and cleavage of STAT6 induced by LANA during KSHV latent infection , and given that LANA blocks cytokine IL-4-stimulated STAT6 signaling for maintaining latency [23] , we speculated that KSHV might also utilize nuclear-localized STAT6 and its cleaved isoform induced by LANA as a negative regulator to repress viral lytic gene transcription . In an attempt to validate our hypothesis , we first analyzed both the promoter sequence of the key latent antigen LANA and lytic activator RTA , and found a canonical pSTAT6-binding site within the RTA promoter but not in LANA promoter ( Fig 7A ) . This provided a clue that the nuclear-localized and cleaved STAT6 may act as a transcription repressor to inhibit RTA expression through binding to its promoter , but without affecting LANA expression . To verify this hypothesis , we generated a luciferase reporter driven by wild type RTA promoter or its mutant with pSTAT6-binding site mutation , and performed reporter assays with full length STAT6 or its truncated mutants in 293T cells . The LANA promoter reported was used as a parallel control . Our results revealed that the RTA promoter was indeed dramatically inhibited by both wild type STAT6 and its cleaved form of STAT6 ( ΔTAD ) in the presence of LANA , and this inhibition is independent of the status of Y641-phosphorylation and abolished by ΔDBD or ΔN mutant ( Fig 7B , compare lane 2 , 4 , 6 , 8 , 10 with lane 3 , 5 , 7 , 9 , 11 ) . The deletion of the STAT6-binding site within the RTA promoter ( mut RTAp-Luc ) specifically abrogated the inhibitory activity of LANA-induced nuclear localized and cleaved STAT6 in RTA transcription ( Fig 7B , compare lane 1 , 3 , 7 with lane 13 , 15 , 17 ) , although the expression of STAT6 in the absence of LANA could block the transcription of RTA to a limited extent ( which may due to the NLS localized at the amino terminus of STAT6 ) ( supplementary S2 Fig ) . By contrast , no significant response of the LANA promoter was observed when it was co-expressed with STAT6 and LANA ( Fig 7C , compare lane 2 with lane 3 ) . In addition , knockdown of STAT6 significantly blocks the inhibitory role of LANA in RTA transcription ( Fig 7D ) , suggesting an important role of STAT6 during KSHV latency . To verify the inhibitory effect of STAT6 on RTA transcription is due to STAT6 interaction with the STAT6-binding site within the RTA promoter , we performed a chromatin-immunoprecipitation ( CHIP ) assay using a wild type RTA promoter and a STAT6-binding site mutant in the presence or absence of STAT6 and LANA . Our results demonstrated that LANA dramatically enhanced the affinity of full length STAT6 and its Y641 and ΔTAD mutant bound to the RTA promoter ( Fig 8A , compare lane 3 , 5 , 7 with lane 2 , 4 , 6 ) , but not its DNA-binding domain ( ΔDBD ) or NLS-deleted domain ( ΔN ) mutant ( Fig 8A , compare lane 9 , 11 with lane 8 , 10 ) . By contrast , mutation of the STAT6-binding site significantly reduced the affinity of LANA-mediated full-length STAT6 or its ΔTAD mutant binding to the RTA promoter ( Fig 8A , compare lane 15 , 17 with 3 , 7 ) . Consistent with our observation that LANA induces nuclear localization and cleavage of STAT6 , the affinity of STAT6 binding to the RTA promoter was significantly reduced by knockdown of LANA ( Fig 8B ) . These results indicate that induction of both nuclear localization and cleavage of STAT6 is important for LANA to de-regulate functions of STAT6 . To explore whether LANA-induced nuclear localized and cleaved STAT6 in KSHV-latently infected cells interacts with the DNA element of the STAT6-binding site within the RTA promoter , we performed an in vitro DNA binding assay by individually incubating the wild-type or the mutated STAT6-binding DNA oligonucleotide , with biotinylated labeling and loading equal amounts of nuclear extracts from KSHV-infected PEL ( BC3 ) cells with or without PMSF and MG132 treatment . The DNA binding activity of both nuclear full length STAT6 and its cleaved form in BC3 cells was significant ( Fig 8C , middle panel ) , whereas little or no signal was seen using the mutant oligonucleotide ( Fig 8C , right panel ) . These results support our hypothesis that LANA-induced nuclear localized STAT6 and its cleaved form is a negative regulator of the RTA promoter by binding to its cognate DNA sequence during latency . To further determine whether introduction of STAT6 alone could block KSHV lytic reactivation , 293-Bac36 cells that harbor an intact KSHV genome were transfected with ectopically expressed wild type STAT6 or its ΔDBD mutant or vector alone , followed by treatment with or without TPA/NaB for 24 hours . Exogenous STAT6 remarkably reduced the transcription and expression of RTA , which blocks viral reactivation and virus progeny production ( Fig 9A , compare lane 1 , 2 with 3 , 4 ) . Consistently , similar results were observed by using K-iSLK cells as target cells ( supplementary S3 Fig ) . To clarify if disruption of STAT6 expression could turn over the inhibition of RTA expression and in turn impair KSHV latently-infected cell growth , STAT6 in iSLK cells and its derivative KSHV-infected cell line K-iSLK were individually knocked down by lentivirus-mediated small interference RNA against STAT6 ( shSTAT6 ) or luciferase control ( shCtrl ) , followed by immunoblotting assays and real-time monitoring of cell growth rate . The results revealed that knockdown of STAT6 in KSHV-infected cell line K-iSLK significantly enhances RTA expression ( Fig 9B , top panel ) , and in turn reduces cell growth when compared with luciferase control ( Fig 9B , low panel ) . By contrast , no significant difference was observed in the K-iSLK parent cell iSLK ( Fig 9B , middle panel ) . In addition , the results of a two-week culture period demonstrated that STAT6 knockdown alone also markedly blocked colony formation of K-iSLK cells but not iSLK cells , when compared with luciferase knockdown control ( Fig 9C ) . Similar results were observed in BC3 cells with STAT6 knockdown ( supplementary S4 Fig ) . Thus , it appears that LANA-mediated nuclear localization and cleavage of STAT6 plays an important role for KSHV to drive viral latency in host cells .
In the present study , we demonstrated a novel regulatory mechanism of KSHV-mediated STAT6 signaling in KSHV-infected cells . We found that STAT6 was induced by KSHV to translocate into the nucleus independent of tyrosine 641-phosphorylation . The nuclear localization of STAT6 was due to the interaction with the latent antigen LANA , and in turn led to STAT6 cleavage in the nucleus of KSHV-infected cells . The nuclear-localized and cleaved STAT6 , induced by LANA , inhibited the transcription of RTA , and blocked lytic replication and viral progeny production . Conversely , full length nuclear-localized STAT6 enhanced the protein stability of LANA for viral latency ( Fig 10 ) . Consistent with our previous studies [23 , 24] , these findings explain why LANA could block the IL-4-induced phosphorylation of STAT6 , and why IL-13-mediated constitutively phosphorylation of STAT6 was dramatically enhanced at the early stage ( < 3 days ) , but reduced later ( >5 days ) along with the increased expression of LANA during KSHV primary infection ( which could be due to the effect of nuclear localization of STAT6 induced by LANA ) . Our finding that LANA-mediated STAT6 cleavage in the nucleus was inhibited by PMSF , but resistant to a number of other protease inhibitors including pepstatin , leupeptin and aprotinin , suggests that STAT6 cleavage is caused by a serine protease . In addition , proteasome inhibitor MG132 could also indirectly block LANA-mediated STAT6 cleavage , further supporting this conclusion . Moreover , Y641F STAT6 , which is defect for IL-4-induced phosphorylation and less nuclear localization , translocates into the nucleus in the presence of LANA and is more sensitive to LANA-mediated cleavage . Although other proteases may be present in different cells lines we observed STAT6 cleavage in B cells , endothelial cells and epithelial cells , indicating that STAT6 cleavage is tightly regulated by KSHV infection . Taken together , our results revealed a previously uncharacterized pathway for KSHV pathogenesis which includes nuclear translocation and cleavage of STAT6 . A recent report suggested that STING is required for Sendai virus-induced phosphorylation of STAT6 on Y641 , and leads to STAT6 translocation in nucleus after virus infection [25] . Similar to the interaction of STAT6 with STING , the DNA-binding domain of STAT6 is required for interactions with LANA . However , further investigation is required to determine if LANA competes or cooperates with STING . Unlike other STATs , no detailed mechanisms for driving the nuclear import or export of STAT6 have been identified yet . We could not exclude the possibility that cleaved STAT6 induced by LANA is not generated in the nucleus but is present in the cytoplasmic fraction due to protein shuttling . For example , recent reports have shown that STAT6 phosphorylation on Tyr641 was induced by viral or parasitic infection independent of IL-4 [25–27] . However , the canonical STAT6 signaling induced by IL-4 is independent of STING or MAVS , which help viruses trigger STAT6 phosphorylation [28] . This could also address why KSHV selectively responds to IL-4 and IL-13 stimulation as observed in our previous studies [23 , 24] . There still remains some controversy with regard to how KSHV regulates unstimulated STAT6 and constitutively phosphorylated STAT6 by IL-13 during latency . The discoveries by our and other have shown that dephosphorylation of STAT6 is restricted to the nucleus [29] , and in addition to tyrosine phosphorylation , serine phosphorylation is also necessary for STAT6 to exert full transcriptional potency [30] , could be an explanation for this discrepancy . On the other hand , although the consequences of serine phosphorylation of STAT6 remain obscure , IL-4 has been shown to induce the phosphorylation of STAT6 not only on tyrosine 641 but also serine on 756 [31] . In contrast , in addition to stimulation by cytokine IL-4 or IL-13 , some studies present alternatives to the canonical induction of STAT6 by IL-15 in mast cells [32] , by PDGF in fibroblast cells [33] , or by IFNα in B cells [34] . Other posttranslational modifications such as acetylation and methylation could explain the complicated regulation of STAT6-mediated activation of expression [35 , 36] . It has been demonstrated that there exists four isoforms of STAT6 in humans , which potentially contribute to competition of activated STAT6 signaling . However , in different cell types , the cleavage of STAT6 by proteases is different . It has been suggested that the elastase family is responsible for cleavage of STAT6 in mast cells [8] , whereas the calpains family has been reported to be responsible in T cells . However , unlike the production of a 65kD cleaved STAT6 in mast cells , degradation in T cells seems to be complete [37] . Our findings revealed that STAT6 cleavage by serine proteases could be induced by KSHV infection in B cells , endothelial and epithelial cells . Consistent with previous studies [7] , our results also suggest that STAT6 cleavage is not observed in healthy B cells but appears in KSHV-infected B cells , indicating that KSHV induces STAT6 cleavage . In agreement with a previous report [38] , we observed that the cleavage positions of LANA-induced STAT6 with carboxyl terminal-truncated mutants are mainly located between 673 and 695 amino acids , and the STAT6 protease activity is present in the nucleus and inhibited by a serine protease inhibitor PMSF . Although the carboxyl truncated cleavage of STAT3 is also similarly regulated by proteolytic processing during terminal differentiation of neutrophils [39] , in contrast to STAT6 , we did not see any apparent cleavage of STAT3 . This suggests that the cleavage of STAT6 is specifically induced by KSHV . Interestingly , recent studies reported that LANA also undergoes caspase cleavage in response to oxidative stress [40] . In this study , we observed that STAT6 also pulled down the cleaved isoform of LANA at the amino terminus , and several protease inhibitors including PMSF greatly enhanced the level of LANA expression . However , further studies will be required to determine if the protein stability of LANA enhanced by its interaction of STAT6 is through blocking the cleavage of LANA . In accordance with a previous observation that STAT6-/- mice present higher virus titer than wild type control [25] , our studies found that inhibition of STAT6 by small hairpin RNA interference in KSHV-latently infected cells also reactivates lytic replication by enhancing RTA expression , which in turn leads to reduction of cell growth and colony formation of KSHV-infected cells . We also demonstrated that ectopic expression of STAT6 in HEK293 cells carrying KSHV genome is sufficient to block TPA/Sodium butyrate-induced RTA transcription and viral production , respectively . Strikingly , constitutive expression of C-terminally truncated STAT5 proteins in CD4 T cells from HIV patients on treatment have been shown to be associated with good response to therapy [41] . However , it is not known whether the truncated cleavage of STAT6 in KSHV-latently infected cells may influence the outcome of disease progression , and need to be further investigated .
De-identified 3μm paraffin-embedded KS patient tissues were obtained from public health clinical center of Fudan University . Usage of redundant cancer sample for research purpose was approved by the Hospital Medical Ethics Committee . The IRB approved protocol in which Declaration of Helsinki protocols were followed and each donor gave written , informed consent . Plasmids expressing STAT6 truncation mutants ΔN and ΔDBD was constructed by PCR amplicon ( FLAG-STAT6 as template , a gift from Jaharul S . Haque at Lerner Research Institute ) inserted into pcDNA3 . 1-FLAG-C1 with restriction enzymes EcoRI and XhoI , respectively . STAT6 mutants ΔTAD and Y641F were individually constructed by PCR-directed site mutation with 701 stop codon and Y641F . Luciferase reporter RTAp-luc with wild type RTA promoter was described previously [42] , RTAp-luc with STAT6-binding site mutation ( TTCCGCGGAA into TATATGTCTA ) was obtained by PCR-directed site mutation with RTAp-luc as template . Plasmids LANA-myc ( WT , N: 1–340 ) , GFP-LANA-C ( 930–1162 ) -myc , RFP-LANA , LANAp-Luc , and pCDNA3 . 1-GFP-NLS-myc was described previously [42 , 43] . Antibodies to STAT6 ( D3H4 , Cell signaling for WB; YE361 , Abcam for IHC and IF ) , α-tubulin ( 1C6 , Santa cruz ) , Histone H3 ( #8226 , Abcam ) , FLAG ( M2 , Sigma ) , and GAPDH ( G8140-01 , US Biological ) were used according to the manufacturers specifications . The monoclonal antibody anti-myc ( 9E10 ) and HA ( 12CA5 ) were prepared from hybridoma cultures . Mouse monoclonal antibodies against LANA and RTA were kindly provided by Ke Lan from Shanghai Pasteur Institute of CAS . PMSF , Leupeptin , Aprotinin , and Pepstatin A were purchased from Amresco . TPA was purchased from Sigma and sodium butyrate from J&K Corporation . Proteasome inhibitor MG132 was purchased from Biomol International , and Cyclohexamide ( C4859 , Sigma Inc . , St . Louis , MO ) . Chelex 100 Resin ( #142–1253 ) was from Bio-Rad . KSHV-negative ( BJAB and DG75 from American Type Culture Collection [ATCC] , Manassas , VA ) and positive ( BC3 , BCBL1 , and JSC1 from ATCC ) B-lymphoma cells , iSLK ( 1mg/ml hygromycin , 250ug/ml G418 , a gift from Shou-Jiang Gao at University of South California ) and iSLK-Bac16 ( K-iSLK , 1mg/ml hygromycin , 250ug/ml G418 and 1ug/ml puromycin , a gift from Shou-Jiang Gao at University of South California ) cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum ( FBS ) and 1% penicillin and streptomycin ( Gibco-BRL ) . HUVEC ( ATCC ) , HeLa ( ATCC ) , HEK293 ( ATCC ) , and 293/BAC36 ( a gift from Erle Robertson at University of Pennsylvania ) which harboring wild- type KSHV ( 1ug/ml puromycin ) cells , were maintained in DMEM supplemented with 10% FBS . All cell lines were incubated at 37°C in a humidified environmental incubator with 5% CO2 . 293 and 293/Bac36 cells were transfected with 1 mg/ml polyethyleneimine ( PEI ) at a ratio of 1μg plasmid DNA: 3μl PEI . iSLK and iSLK-Bac16 were transfected with Lipofectamine 2000 reagent ( Invitrogen ) according to the manufacturer’s recommendations . Cells were transfected at culture for 24hrs with cell confluence reaching 60–70% . B-cells transfection was performed with Lonza-4D nucleofector system in an optimized program CA137 . Immunofluorescence assays were performed as described previously [44] . Briefly , cells were harvested and plated in polylysine-treated coverslips for 4hrs in CO2 incubator to let cells attach to the plate . And then washed with ice-cold PBS twice , and fixed in 4% paraformaldehyde for 20 min at room temperature . After fixation , cells were washed three times in PBS and permeabilized in PBS containing 0 . 2% fish skin gelatin ( G-7765; Sigma ) and 0 . 2% Triton X-100 for 5 min , followed by primary and secondary antibody staining . Nucleus was counterstained with 4 , 6-diamidino-2-phenylindole ( DAPI ) , and coverslips were mounted with p-phenylenediamine . Cells were visualized with Leica SP8 confocal microscope . Cells were harvested , washed once with ice-cold PBS , and lysed in 600μl ice-cold RIPA buffer [150 mM NaCl , 50 mM Tris ( pH7 . 6 ) , 1% Nonidet P-40 , 2 mM EDTA , 1 mM phenylmethylsulfonylfluoride ( PMSF ) , 1 mM Na3VO3 , 1 g/ml aprotinin , 1 g/ml leupeptin , 1 g/ml pepstatin] for 30min with constant shaking at 4°C . Cell debris was removed by centrifugation at 14 , 500 rpm at 4°C for 10min . The supernatants were then transferred to a new eppendorf tube . Five percent of the supernatant was used as input . The rest lysates were then precleared with normal mouse IgG ( Invitrogen ) and protein A/G Sepharose beads by end-over-end rotation at 4°C for 1hr . After preclear , beads were spun out , washed with ice-cold RIPA buffer for four times and re-suspended with 30μl RIPA buffer . Supernatant was then incubated with primary antibody at 4°C with rotation overnight . Protein of interest complexes were captured the next day with 30μl protein A/G Sepharose beads with rotation on 4°C for another 4hrs . Beads were spun out , washed with ice-cold RIPA buffer for four times and re-suspended with 30μl RIPA buffer . For immunoblotting assays , after input lysates , immunoprecipitated ( IP ) complexes were boiled in 6xSDS loading buffer , proteins were fractionated by SDS-PAGE , and transferred to a 0 . 45-mm nitrocellulose membrane . The protein of interest in the membrane was probed with primary antibodies at 4°C with shake for overnight , followed by incubation with appropriate secondary antibodies for another 1hr at room temperature . The member was scanned with an Odyssey Infrared scanner ( Li-Cor Biosciences ) . Densitometric analysis was performed with the Odyssey scanning software . HEK293-Bac36 ( KSHV ) cells were individually induced with 20 ng/ml of tetradecanoyl phorbol acetate ( TPA ) and 1 . 5 mM sodium butyrate ( NaB ) for 2 days at 37°C with 5% CO2 . After induction , the supernatant of culture medium was collected and filtered through a 0 . 45μm filter , and viral particles were spun down at 25 , 000 rpm for 2h at 4°C . The concentrated virus was collected and used for virion quantitation by qPCR . iSLK and iSLK-Bac16 cells was individually transfected with RFP-tagged plasmid GV113 containing shControl or shSTAT6 for 48hrs . Equal amount cells were then seeded in 16 well with normalization of transfection efficiency by flow cytometer assay . Cells growth rate was real-time measured by xCELLIgence system ( ACEA Biosciences , Inc . ) . Experiments were performed in triplicate with 6 repeats each time point . For colony formation , iSLK and iSLK-Bac16 cells were trysinized and equal amount cells were dispersed to 10cm plate . After 14 days , cell culture supernatants were discarded and cells were fixed with 4% ( v/v ) formaldehyde and stained with 0 . 1% crystal violet . Colony formation in each dish were scanned by Li-Cor Odyseesy and counted . Experiments were performed in triplicate . STAT6 dimerization experiment was performed with freshly made solution of disuccinimidyl suberate ( DSS , Thermo Scientific ) . At 36hr-posttransfection , cells were harvested and washed three times with ice-cold PBS ( pH 8 . 0 ) to remove amine-containing culture media and proteins from the cells . Cells pellet was re-suspended in PBS to reach a concentration of 2 . 5x107 cells per ml , and cross-linked with 1mM DSS for 30 minutes at room temperature . The cross-link reaction was stopped by addition of 20mM Tris-HCl ( pH 7 . 5 ) and incubated for another 15 min at room temperature , followed by washing with PBS for once and lysis with RIPA for 30min at ice . The cell lysates were separated by SDS-PAGE and analyzed by immunoblotting with STAT6 antibody ( YE361 , Abcam ) for endogenous dimer of STAT6 . Cells were harvested and washed twice with ice-cold PBS followed by resuspending and lysis cell pellet in a hypotonic buffer A ( 10 mM HEPES-K+ pH 7 . 5 , 10 mM KCl , 1 . 5 mM MgCl2 , 0 . 5 DTT ) containing 0 . 1% NP-40 for 5 min on ice . The cytoplasm fraction was obtained by spinning at 3000 rpm ×5 min at 4°C . The supernatant ( cytoplasm protein ) harvested and frozen at −80°C for use . The nuclear pellets were washed twice with buffer A ( without NP-40 ) , resuspended in RIPA buffer and sonication with 30% of the maximum power output ( 10s on/off ) for two rounds , followed by 14500 rpm for 5 min at 4°C , supernatant ( nuclear protein ) was harvested and snap frozen for further use . The efficiency of cytoplasm and nuclear extraction were verified by immunoblotting with Histone H3 and α-Tubulin antibodies , respectively . Chromatin immunoprecipitation ( ChIP ) assay were performed with some modification as described previously [45–48] . At 48hr post-transfection , approximate thirty millions of cells were cross-linked with 1 . 42% ( vol/vol ) formaldehyde for 5min at room temperature for 10 min . Formaldehyde was quenched by adding 125mM glycine and incubate at room temperature for 5min . Fixed cells were scraped and washed with ice-cold PBS twice and lysed with 1ml IP buffer ( 150mM NaCl , 50mM Tris-Hcl [pH7 . 5] , 5mM EDTA , 0 . 5%NP40 ( v/v ) , 1 . 0% TritonX-100 ( v/v ) with protease inhibitors [1 mM PMSF , 10mM Na3VO3 , 2μg/ml aprotinin , 5μg/ml leupeptin , 1μg/ml pepstatin] ) and incubated for 15 min with constant shaking at 4°C . Nuclear component was obtained by centrifuge at 12 , 000g for 1min at 4°C and nuclear pellet was washed with IP buffer and re-suspended in 600 μl IP buffer . Chromatin was sheared with sonication to an average fragment size of 300 to 500 base pairs . Solubilized chromatin extracts were cleared by centrifugation at 12 , 000g for 10min at 4°C . 20μl supernatant was transferred to a new eppendorf tube to determine shearing efficiency and extract DNA as input . Rest of the sample was used for immunoprecipitation . Antibody against interest protein was added to the sheared chromatin and incubation overnight at 4°C with shaking . For mock , non-specific IgG was used as a control . The chromatin complex was precleared the next day with centrifuge at 12000g for 10min at 4°C . 90% supernatant was transferred to a new tube and chromatin complex was captured by adding protein A/G Sepharose beads with rotation at 4°C for 1hr . After incubation beads were washed five times with 1ml IP buffer to exclude non-specific binding . To elute DNA , 100ul chelex 100 ( 10% w/v ) ( Bio-Rad ) was added to the washed beads and boiled for 10 min , followed by treatment with 1μl proteinase K ( 20μg/μl ) and incubation at 56°C for 30min , and boiled for another 10min to inactivate proteinase K . condensate DNA was centrifuged at 12 , 000g for 1min at 4°C and washed with 120μl ddH20 , supernatants of the two elution were collected . Input DNA was precipitate with 2 volumes of ethanol , and washed with 70% ( v/v ) ethanol . The pellet was re-suspended in 100ul chelex 100 ( 10% w/v ) and boiled for 10min and continue processing as IP sample . Purified DNA was amplified by quantitative PCR using RTA specific primers ( forward: 5'-CCCGACTAATGAGGACAAT-3' , and reveres: 5'-TTCAAACCCATCATCTGTG-3' ) . Immunohistochemistry of STAT6 was performed on deparaffinized , formalin-fixed tissue sections using an indirect immunoperoxidase method with an automated immunostainer as described previously [24] . iSLK and iSLK-Bac16 cells with STAT6 knockdown were individually performed by transient transfection with STAT6 shRNA . STAT6 shRNA sequence ( GGGAGAAGATGTGTGAAACTCTGAA ) was inserted into lentivirus pGV113 vector which carrying a RFP protein . At 48hr post-transfection with PEI method , cells were visualized under florescence microscope and flow cytometer assay to check transfection efficiency and RNA interfering efficiency was assessed by immunoblotting against STAT6 . pGV113 vector with luciferase ( shLuc ) target ( TGCGTTGCTAGTACCAAC ) sequence was used as control . Transfected cells at 48hr post-transfection were lysed with 200ul Passive lysis buffer ( Promega ) , followed by dual luciferase reporter assay according to the manufacturer’s instruction as described previously . Renilla luciferase was used as a control to normalize the transfection efficiency . Relative luciferase activity [RLU] was expressed as fold changes relative to the reporter construct alone . Assays were performed in triplicate . Total RNA from cells was extracted using TRIzol regent ( invitrogen ) according to the manufacturer’s Instructions . 1μg RNA was reverse transcripted with a Superscript II reverse transcription kit ( Invitrogen , Inc . , Carlsbad , CA ) . After reverse transcription , 1μl cDNA was used as template for quantitative PCR . The RTA primers ( 5’-CAGACGGTGTCAGTCAAGGC-3’ and 5’-ACATGACGTCAGGAAAGAGC-3’ ) and GAPDH ( 5’-ACGACCACTTTGTCAAGCTC-3’ and 5’-GGTCTACATGGCAACTGTGA-3’ ) was used as an internal control . The cDNA was amplified in a total volume of 20ul containing 10 μl of SYBR premix Ex Taq ( Takara , Inc . ) , 0 . 5 μl each primer ( 10μM ) , 1μl cDNA and rest of RNAase free water . PCR program was running on thermocycler ( Bio-Rad Inc . ) in a procedure of 40 cycles of 1 min at 94°C , 30 s at 55°C , and 30 s at 72°C , followed by 10 min at 72°C . A melting-curve analysis was performed to verify the specificities of the amplified products . Each sample was tested in triplicate and date was summarized from three independent experiments . The relative mRNA fold changes relative to GAPDH were calculated by the threshold cycle ( CT ) method . Statistical significance of differences between means of at least n = 3 experiments was determined using Student’s t-test . | STAT6 , a member of the signal transducer and activator of transcription ( STAT ) family , has been shown to play an important role in viral infection . STAT6 activation is linked to reactivation of oncogenic herpesvirus and their associated cancers . However , the precise mechanism by which KSHV modulates STAT6 regulation remains unclear . In the present study , we demonstrate that KSHV induces nuclear localization and cleavage of STAT6 in both KSHV-infected B lymphoma and endothelial cells . Importantly , this effect is dependent on LANA ( a key latent antigen ) expression and leads to inhibition of viral lytic replication . Herein , we provide a previously uncharacterized description of how STAT6 plays an inhibitory role in the pathogenesis of oncogenic viruses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"pathology",
"and",
"laboratory",
"medicine",
"molecular",
"probe",
"techniques",
"enzymes",
"pathogens",
"immunoblotting",
"microbiology",
"enzymology",
"viruses",
"immunoprecipitation",
"dna",
"viru... | 2017 | Nuclear Localization and Cleavage of STAT6 Is Induced by Kaposi’s Sarcoma-Associated Herpesvirus for Viral Latency |
DNA sequence and local chromatin landscape act jointly to determine transcription factor ( TF ) binding intensity profiles . To disentangle these influences , we developed an experimental approach , called protein/DNA binding followed by high-throughput sequencing ( PB–seq ) , that allows the binding energy landscape to be characterized genome-wide in the absence of chromatin . We applied our methods to the Drosophila Heat Shock Factor ( HSF ) , which inducibly binds a target DNA sequence element ( HSE ) following heat shock stress . PB–seq involves incubating sheared naked genomic DNA with recombinant HSF , partitioning the HSF–bound and HSF–free DNA , and then detecting HSF–bound DNA by high-throughput sequencing . We compared PB–seq binding profiles with ones observed in vivo by ChIP–seq and developed statistical models to predict the observed departures from idealized binding patterns based on covariates describing the local chromatin environment . We found that DNase I hypersensitivity and tetra-acetylation of H4 were the most influential covariates in predicting changes in HSF binding affinity . We also investigated the extent to which DNA accessibility , as measured by digital DNase I footprinting data , could be predicted from MNase–seq data and the ChIP–chip profiles for many histone modifications and TFs , and found GAGA element associated factor ( GAF ) , tetra-acetylation of H4 , and H4K16 acetylation to be the most predictive covariates . Lastly , we generated an unbiased model of HSF binding sequences , which revealed distinct biophysical properties of the HSF/HSE interaction and a previously unrecognized substructure within the HSE . These findings provide new insights into the interplay between the genomic sequence and the chromatin landscape in determining transcription factor binding intensity .
Binding of transcription factors ( TFs ) to DNA elements is necessary to establish and maintain functional changes in gene expression levels . The mechanism by which these factors seek out and bind to their cognate motif elements remains an area of active investigation ( reviewed in [1] ) . TFs are present at cellular concentrations that allow binding to sites that are degenerate from the consensus sequences , and genomes of eukaryotes are littered with potential degenerate binding sites; however , only a small fraction of potential binding sites are recognized in vivo . Moreover , TF binding sites vary dependent upon cell type and cellular conditions . In vivo , TF binding is potentially dependent upon motif accessibility and the surrounding chromatin landscape . Therefore , determining a comprehensive set of potential genomic binding sites and quantifying the joint effects of DNA sequence and chromatin landscape upon binding intensity remains a challenge . Experimental approaches to characterize TF binding sites include assays such as ChIP-seq , protein binding microarrays ( PBM ) [2] , iterative rounds of protein-DNA binding and selection with a complex oligonucleotide library [3] , or extrapolation from DNase I hypersensitivity regions [4] . However , perhaps the most direct way to determine all potential TF binding sites within a genome is to incubate purified TF and naked sheared genomic DNA in vitro , and then specifically quantify the TF-bound DNA [5] . This in vitro method allows binding sites to be interrogated in their native sequence context without the confounding effects of chromatin and cooperation between chromatin-bound factors . It is challenging to predict in vivo TF binding accurately even when all potential in vitro binding sites have been characterized , because the chromatin landscape dramatically influences binding and it changes dynamically with development and with alterations in cellular nutrition and environment [6] , [7] . Recent TF binding site modeling efforts have considered genomic nucleosome occupancy or DNase I hypersensitivity data to account for the effect chromatin has upon in vivo TF occupancy [8]–[11] . However , these models are limited in that they rely upon genomic accessibility data and TF binding data produced under the same conditions . To date there are no data sets that describe the full set of potential TF binding sites , the chromatin state data prior to binding , and occupied binding sites in vivo , in a single inducible system . Integration of these three data sets would allow one to decouple the effect TF binding has upon chromatin state from the effect pre-existing chromatin state has upon induced TF binding . The heat shock response of Drosophila is a model system extensively used to study the general functions of sequence specific activators and how they function to regulate transcription ( reviewed in [12] ) . The master regulator of the heat shock genes , Heat Shock Factor ( HSF ) , has a modest affinity for DNA under non-stress conditions [6] , [13] , [14] , and upon stress , HSF homotrimerizes and inducibly binds to a conserved consensus motif at over 400 sites in the Drosophila genome [6] , [14] . While over 95% of the HSF binding sites contain an underlying HSF sequence motif element ( HSE ) , the vast majority of predicted genomic HSEs remain HSF–free following heat shock . Therefore , the chromatin landscape most likely plays a prominent role in determining binding of HSF . Here , we describe an experimental technique , protein/DNA binding followed by high-throughput sequencing ( PB–seq ) , to quantify the binding potential of all binding sites within a genome . We then develop a quantitative model that incorporates HSF PB–seq data , together with HSF ChIP-seq in Drosophila S2 cells [6] and S2 cell chromatin data , that accurately predicts observed in vivo HSF binding profiles . Moreover , our model allows us to quantify the relative importance of the chromatin features influencing HSF binding intensity . Finally , we develop a sequence model that uses HSF PB–seq data to characterizes the relationship between positions within the HSE and provide biophysical insight into the mechanisms by which HSF interacts with its cognate element .
We performed an in vitro binding experiment with purified HSF ( Figure S1 ) and naked , sheared genomic Drosophila DNA , to derive an accurate set of potential HSF binding sites in the Drosophila genome . HSF–bound DNA was specifically eluted and detected by high throughput sequencing ( Methods ) . The HSF PB–seq experiment yielded 68% of the sequence tags within peaks . In contrast , typical ChIP-seq protocols are more inefficient and the majority of DNA ( 60% to >99% ) sequenced is uninformative background DNA [15] . Peak calling revealed 3952 HSF–binding peaks ( p<0 . 01; 2848 peaks were common to both experimental replicates ) , which include 60% of the previously identified high-confidence HSF binding peaks in vivo [6] . The naïve expectation is that every in vivo HSF peak should have a corresponding in vitro peak , but it is not surprising to observe an incomplete overlap of in vivo by in vitro peaks , for various reasons . As will be discussed , binding sites detected in vivo but not in vitro tend to be more degenerate and have higher DNase I accessibility . Additionally , in vivo binding sites that are dependent upon cooperative interactions with pre-bound chromatin factors , long range DNA interactions , post-translational modifications of HSF [16] , higher-order chromatin structure , or bridging protein interactions [17] will not be detected in the current form of PB–seq . Underlying the in vitro binding peaks , we detected 3735 clusters of HSF binding site HSE sequences ( 2896 in peaks common to both replicates ) at 20% HSE False Discovery Rate ( FDR ) . We used clusters of co-occurring sites due to the uncertainty in HSE detection ( see Methods ) . Furthermore , the majority , 3389 clusters ( 2586 in peaks common to both replicates ) are not detectably bound in S2 cells in vivo . Figure 1 shows two examples of in vitro binding sites flanking the Cpr67B gene that are not bound in vivo . Moreover , the in vitro binding data quantifies differences in the in vitro and in vivo HSF binding intensity , such as the peaks within each of the promoters for Hsp23 and Hsp26 ( Figure 1 ) . The PB–seq experiment allows for an estimate of the relative binding intensity of each HSE , based on the number of sequence tags associated with it . To compute the dissociation constant ( Kd ) values it is necessary to have estimates for both the fraction of bound and free HSE in the PB–seq experiment . Since the PB–seq data only provides information on the bound fraction , we needed to determine the absolute Kds for two HSEs that are found within the PB–seq data in order to provide enough information to estimate the free fraction ( see Methods ) . To generate the HSF/HSE Kd measurements , we performed electrophoretic mobility shift assays ( EMSA ) . The EMSAs were performed with purified HSF and HSEs that are only modestly degenerate from the consensus . We found that HSF binds to the first HSE with ∼42 . 6 pM interval: 36 . 8–49 . 4 pM; Figure 2A and 2C ) and the second HSE with ∼224 pM affinity ( 95% confidence interval: 181–276 pM; Figure 2B and 2D ) . The resulting two absolute Kd values enabled us to transform PB–seq read depths into absolute Kd values ( Figure 2E and Methods ) . We confirmed the transformation of the relative Kd values to absolute Kds by performing band shifts with genomic HSEs of different predicted Kd values ( Figure S2 ) . The experimental verifications of the measurements are within the estimated error of the EMSA confidence interval and the variability between PB–seq replicates ( Figure S3 ) . Taken together , these measurements allow us to characterize the binding energy landscape for HSF across the entire Drosophila genome , in the absence of chromatin . Our estimated Kd values for isolated HSEs in the Drosophila genome ranged from 40–400 pM ( Figure 2E ) . These in vitro binding results demonstrate the feasibility and efficiency of combining high-throughput detection methods with classic EMSA and competition experiments to quantify the binding energy for the comprehensive set of potential genomic binding sites for a sequence-specific TF . Our data reveals substantial differences between in vivo and in vitro binding intensities ( Figure 3A ) , underscoring the role of chromatin in determining in vivo binding site selection and affinity . We found DNase I hypersensitivity was the most important predictor of HSF binding; therefore , we scaled the in vivo and the in vitro read counts so that they were approximately equal at in vivo sites with high DNA accessibility ( Methods , Figure S4 ) . After this normalization , we partitioned the binding sites that were detectable in vitro into classes: “unaffected” sites , bound at comparable affinities in vivo and in vitro ( 55 red points in Figure 3A; 2% of all sites ) ; “suppressed” sites , with reduced , but detectable , in vivo intensity ( 365 green points; 13% ) ; and “abolished” sites , below the in vivo threshold for detection ( 2223 blue points; 76% ) . In addition , sites not detectable in vivo or in vitro were labeled “background” ( 249 gray points; 9% ) , and sites with stronger relative in vivo intensity compared to in vitro were labeled “enhanced” ( 4 black points; 0 . 1% ) . PB–seq data reveals potential HSF binding sites , providing the opportunity to model the effect that non-stressed chromatin landscape has upon induced HSF binding intensity . There is a wealth of chromatin data available for S2 cells during unstressed conditions [18] , [19] , and heat-shock induced binding sites of HSF in S2 cells are also known [6] . We used DNase I hypersensitivity data [18] , MNase data [19] and ChIP-chip data for 9 factors and 21 histone modifications for unstressed Drosophila S2 cells ( Table S1 ) [18] to predict the intensity of inducibly bound in vivo HSF–bound sites ( Figure 4A , Figure S5 and Figure S6 ) . For our statistical model , we selected a rules ensemble [20] , a linear regression model in which some terms are combinations of covariates known as “rules” . This approach allowed us to capture fairly complex interactions between covariates . For example , a rule might apply when H3K27ac and DNase I hypersensitivity both exceeded designated thresholds ( value ranges can also be expressed ) . Each rule's coefficient is added to the predicted value if , and only if , the rule applies . When there is only one covariate , the model reduces to a linear regression . The Pearson's correlation coefficient between HSF ChIP-seq data for the model incorporating all these data sets was r = 0 . 62 ( Figure S6 and Figure S7 ) . As the large number of covariates brings with it some danger of overfitting , we tested combinations of the four classes of covariates: DNase I hypersensitivity , MNase , histone modifications/variants , and non-histone factors ( Figure 4B , Figure S6 , Figure S7 ) . Of notice , the correlation of the linear regression model that incorporates DNase I data was r = 0 . 64 on the test data ( Figure 4B and Figure S7B ) . Our study is consistent with a previous study that obtained r = 0 . 65 for actual and inferred TF binding intensities using a DNase I dependent model [8] . Other covariate classes produce similar , but lower , correlations . The rules model using histone modifications and histone variants yielded r = 0 . 57 ( Figure 4B and Figure S7 ) , while a rules model incorporating non-histone bound chromatin factors yielded r = 0 . 58 ( Figure 4B and Figure S7 ) . Combining covariate classes further improves the correlation to as much as r = 0 . 70 ( Figure S6 and Figure S7 ) . We also examined the Receiver Operator Curves ( ROC ) for the different covariate combinations ( Figure S8 ) and found concordant results . If we assume that the PB–seq , genomic ChIP , DNase I-seq , and MNase-seq experiments are maximally resolved and sensitive , with no experimental noise , an approximate upper bound is given by r = 0 . 90 , as observed for two HSF–ChIP-seq replicates [6] . Notably , the higher resolution of the DNase I-seq data , compared to the ChIP-chip data , may be why DNase I-seq alone is strongly predictive in the linear regression model and most influential in the rules ensemble models . Notably , we used the chromatin landscape prior to induced TF binding to predict binding intensity , whereas previous models have used the chromatin landscape present when the TF is bound in order to infer binding intensity [8] or infer binary binding events [10] , [11] ( see Discussion ) . Our data and modeling indicated that the presence of active chromatin features , such as histone acetylation and DNase I hypersensitivity , had a significant influence on the predictive power of the model , while repressive features had minimal influence ( Figure S9 ) . DNase I hypersensitivity was a strongly predictive covariate in the model when used in a simple linear regression model ( Figure 4 ) , or in combination with histone modification and non-histone factor covariates in the rules ( Figure S9E–S9G , S9J , S9K , and S9M ) . Tetra acetylation of H4 and H3K9ac were the most informative histone marks in the model that used histone variants and histone modifications as covariates ( Figure 5A ) . GAGA associated factor ( GAF ) , which has a proposed role in permitting HSF binding [21] , was the most influential factor in the HSF binding prediction model that considered all chromatin-binding factors ( Figure 5B ) . The analysis above indicates that DNA accessibility , as measured by DNase I hypersensitivity , is a primary determinant of binding intensity . Previous studies have similarly shown that TF binding sites correlate strongly with DNase I hypersensitive sites [8] , [10] , [11] , [22] . For instance , histone acetylation causes local chromatin decondensation by reducing the ionic interactions between lysine residues and DNA and promotes accessibility , but the extent to which combinations of histone marks and TFs act together to dictate chromatin accessibility is not known . Therefore , it is of interest to see whether DNA accessibility can be predicted from specific features of the chromatin landscape , such as histone modifications and non-histone chromatin bound factors . In addition , accurate predictions of DNA accessibility would be of practical use , because direct measurements are often not available . To address this question , we applied our rules ensemble framework to predict DNase I hypersensitivity ( the best available proxy for DNA accessibility ) from ChIP-chip data for histone features , non-histone chromatin bound factors , MNase data and combinations of these covariate pools ( Figure 6 ) . Tetra-acetylation of H4 and H3K9 acetylation were most influential in the model that uses histone modifications , bulk histone and histone variant intensities ( Figure S10E ) ; the correlation coefficient for this model using the test data is 0 . 51 ( Figure S11B ) . The model that uses non-histone factor ChIP-chip data obtains a correlation of 0 . 52 ( Figure S11B ) , which is consistent with TFs having characteristic DNase I hypersensitivity footprints [10] , [11] . The model that combines both histone data and non-histone data into a rules model performs the best on the test set , with a correlation of 0 . 60 ( Figure S11B ) . Repressive histone marks appear to contribute little to generating the DNase I hypersensitivity pattern ( Figure S10 ) and the lack of active chromatin marks appears to be sufficient to package DNA into inaccessible units . These models reinforces the notion that the biochemical composition of chromatin permits DNase I hypersensitivity and quantifies the contributions individual modifications , and combinations thereof , make to DNase I hypersensitivity ( Figure S11 ) . As more and higher-resolution genome-wide data becomes available , these models will be refined . PB–seq provides the opportunity to model the sequence-dependent binding preferences of a purified TF genome-wide and independent of chromatin or other factors . In the case of HSF , the consensus binding site is well characterized and consists of three pentamers , ÒAGAAN NTTCT AGAANÓ , ( here denoted pA , pB , and pC ) , each bound by a monomer of the HSF homotrimer . Note that the consensus sequences for pA and pC are identical , while the one for pB is their reverse complement . Of course , the consensus HSE is a crude summary that ignores subtleties in the base preferences at each position . A position-specific scoring matrix ( PSSM ) provides a somewhat improved description but still ignores dependencies between positions within the binding site . We sought to use genome-wide binding sites from PB–seq to produce an improved model for the sequence preferences at HSEs . We began by computing the mutual information for all pairs of HSE positions based on the identified in vitro binding sites . We found negligible evidence of correlated base preferences between positions , but we did observe that some pentamers within PB–seq peaks adhered closely to the consensus motif while others did not . This led us to formulate a probabilistic model that allows each pentamer in an HSE to closely match the consensus ( “strict” ) or diverge from it more substantially ( “relaxed” ) , and considers all possible combinations of pentamer composition ( Figure S12 ) . More specifically , we described each of the three pentamers using a two-component mixture model , with a latent variable indicating “strict” or “relaxed” binding preferences , and estimated the joint distribution of these three latent variables from the data . The model parameters—the position-specific nucleotide probabilities and prior distribution for the combinations of strict/relaxed pentamers—were estimated from the data by maximum likelihood using an expectation maximization algorithm ( see Methods ) . In fitting the model , we considered only the 1309 isolated HSEs , sequence elements that were at least 200 base pairs away from any other degenerate HSE motif , to avoid complications arising from overlapping HSEs . The model fit the data substantially better than did a simple PSSM ( lnL = −15442 vs . lnL = −15673 for the PSSM; Akaike information criterion [AIC] = 15636 vs . AIC = 15763 for the PSSM ) , suggesting that it effectively captures important dependencies between positions . Based on the estimated model parameters , we computed a posterior probability distribution over all combinations of pentamer stringency and order for each HSE ( Methods; Figure 7B ) . These values were averaged across HSEs to obtain expected genome-wide fractions of HSEs having each of the strict/relaxed pentamer combinations . We found that binding sites with strict pB and pC , and relaxed pA , were most frequent ( an expected 38% of sites ) , indicating that this configuration is preferred ( Figure 7B ) . The next most frequent configurations were a relaxed pB flanked by a strict pA and pC ( 33% ) , and a strict pA and pB combined with a weak pC ( 29% ) . Interestingly , combinations of three strict pentamers occur at negligible frequency . Indeed , only 5 out of 1309 isolated genomic HSEs matched the consensus sequence exactly , while 148 differed from it by a single mismatch . Configurations with at most one strict pentamer were also rare . Together , these results indicate that the biophysical interactions of the pentamers within the binding sites are critically dependent upon their composition and position relative to the other pentamers in an HSE . While the three estimated strict pentamer matrices were similar ( Figure 7A top ) , the relaxed matrices showed substantial differences with respect to each other ( Figure 7A bottom ) . For example , the relaxed pA matrix indicates that 70–80% of HSEs containing a weak pA have the consensus base at positions two , three and four . In contrast , position 12 in pC ( the analog of position 2 in pA ) almost invariably contains a G in all HSEs , while positions 7 and 8 in pB ( analogous to positions 3 and 4 in pA ) have only modest base preferences in HSEs containing a weak pB . This analysis indicates that each monomeric HSF/pentamer interaction has distinct biophysical properties within the context of the broader HSF/HSE interaction . We also devised a simplified model , with a single strict matrix shared by all three pentamers , and a single relaxed matrix obtained by applying a “dampening” factor to the strict matrix ( Figure S13 , Methods ) . This model further supports the strict/relaxed pentamer split ( lnL = −15908 vs . lnL = −16048 for a single-monomer PSSM; and AIC = 15952 vs . AIC = 16078 ) , although both the full model and the full PSSM fit the data better ( lower AIC ) . Moreover , not only was the simplified model still able to reproduce the posterior distributions over pentamer configurations of the full model , but it was also able to replicate synthetic patterns from simulated data ( Figure S14 ) . Finally , the preference for single pentamer degeneracy was also observed independently by comparing the pentamer-specific KL-divergence in PSSMs obtained from subsamples of HSF bound peaks ( Figure S15; Methods ) .
The PB–seq technique combined with EMSA and competition assays provides a straightforward , yet versatile and powerful framework for characterizing all potential binding sites in a genome , regardless of tissue specificity , developmental stage , or environmental conditions . Comparing in vitro and in vivo binding profiles , in the context of pre-induction genomic chromatin landscape , revealed DNase I hypersensitivity , H4 tetra-acetylation , and GAF as critical features that modulate cognate element binding intensity in vivo . Furthermore , DNase I sensitivity was found to be strongly influenced by high GAF occupancy and histone acetylation , while repressive factors were minimally influential in the statistical models . Finally , the full set of potential genomic binding sites provided a rich data set that was used to build more detailed sequence models , which tease apart substructure and features that are lost with traditional PSSM models . One initially surprising observation from our study was that 40% of the in vivo HSF peaks were not found in vitro . We believe that the limited dynamic range for quantifying in vitro binding affinity may be responsible for the lack of detectable in vitro peaks . Although we quantify in vitro binding over an order of magnitude ( 40–400 pM ) , the experimental concentrations of HSF and genomic DNA and our depth of sequencing do not permit the detection of lower affinity HSF binding sites . For instance , only eleven sequence tags would be predicted to underlie a hypothetical 5 nM HSF binding site , and these would not be distinguishable from background . Upon further examination , we find that the composite HSE representing those in vivo binding sites that were not found in vitro is more degenerate than those found using both assays ( Figure S16A ) . Moreover , the in vivo sites that were not found using PB–seq were also more accessible in vivo ( Figure S16B ) , in support of our hypothesis . Performing PB–seq at a range of protein and DNA concentrations , or increasing sequence coverage would expand the dynamic range of quantification by PB–seq . Other possible explanations for this observation include cooperative interactions with pre-bound chromatin factors , long-range DNA interactions , post-translational modifications of HSF , higher-order chromatin structure , or bridging protein interactions . The influence of DNA modifications and immediate flanking sequence do not contribute to this disparity , since we use large fragments of purified genomic DNA . Bridging protein interactions [17] , which do not involve HSF directly binding to DNA , appear not to be responsible for our results because 95% of in vivo peaks encompass at least one HSE near the peak center [6] . However , if other proteins were cooperating with HSF in vivo to enhance HSF binding intensity at low affinity binding sites , then some of these peaks may not be observed in vitro . Since our PB–seq experiment used recombinant HSF in the binding experiments , we would also not capture differences in binding site affinities that are due to post-translational modifications of HSF [16] . To overcome these potential limitations , PB–seq could be adapted to include known bridging/cooperative factors and proteins could be purified from in vivo sources to capture indirect or modification-dependent interactions . The notion that motif accessibility is driving inducible TF binding in vivo is supported by independent studies of distinct TFs: STAT1 , HSF , glucocoticoid receptor ( GR ) , and GATA1 [6] , [22]–[24] . These studies show that the chromatin landscape prior to TF binding influences inducible TF binding . In the first study , it was found that a large fraction of STAT1 induced binding sites contained H3K4me1/me3 marks prior to interferon-gamma ( IFN-γ ) induced STAT1 binding [23] . Our group previously found that inducible HSF binding sites are marked by active chromatin compared to sites that remain HSF–free [6] . A more recent study has shown that inducibly bound GR sites are marked by DNase I hypersensitive chromatin prior to GR binding [22] . Likewise , the permissive chromatin state at GATA1 binding sites is established even in GATA1 knock out cells [24] . While these correlations are instructive , no previous attempt has been made to model inducible TF binding using biological measurements of chromatin landscape present prior to TF binding . Recent models have successfully inferred TF binding profiles using DNA sequence and chromatin landscape data , generated at the same time the TF is bound [8]–[11] . However , these models do not distinguish between the influence TFs have upon local chromatin and the chromatin features that permit TF binding . In contrast , we modeled the changes between HSF in vitro binding ( PB–seq ) and in vivo binding ( ChIP-seq ) landscapes as a function of the non-heat shock chromatin state . This produced a quantitative model describing the important features that modulate the in vivo HSF binding intensity . Moreover , the use of our rules ensemble model enabled the capture of potential interactions between these chromatin features . Our study reveals that DNase I hypersensitivity and acetylation of H4 and H3K9 are strong predictors of inducible HSF binding intensities , however the molecular events and factors that precede TF occupancy to maintain accessible chromatin remain poorly characterized . For instance , the degree to which pioneering factors or flanking DNA sequence , individually or in combination , maintain or restrict accessibility remains unclear . A recent study highlights the biological consequences of maintaining the inaccessibility of TF binding sites , in order to repress expression of tissue-specific transcription factors in the wrong tissues . The authors found that ectopic expression of CHE-1 , a zinc-finger TF that directs ASE neuron differentiation , in non-native C . elegans tissue is not sufficient to induce neuron formation [25] . However , combining ectopic CHE-1 expression with knockdown of lin-53 did modify the expression patterns of CHE-1 target genes in non-native tissue , effectively converting germ line cells to neuronal cells [25] . LIN-53 has been implicated in recruitment of deacetylases , and deacetylase inhibitor treatment mimics lin-53 depletion , suggesting that LIN-53 is actively maintaining CHE-1 target sites inaccessible in germ cells . Alternatively , functional TF binding sites could be actively maintained in the accessible state . HSF binding within ecdysone genes has a functional role in shutting down their transcription [14] , and activating ecdysone-inducible genes containing inaccessible HSEs causes chromatin changes that are sufficient to allow HSF binding [6] . In this special case of HSF–bound ecdysone genes , active transcription and the corresponding histone marks are mediating access to HSEs , in order for HSF to bind and repress transcription upon heat shock . A more recent study has shown that activator protein 1 ( AP1 ) actively maintains chromatin in the accessible state , so that GR can bind to cognate elements [26] . Although TF accessibility to critical genomic sites appears to be actively maintained , many binding sites may be a non-functional result of fortuitous TFBS recognition . It has long been hypothesized that the binding affinities for TF/DNA interactions are sufficiently strong to allow promiscuous binding at the cellular concentrations of TFs and DNA [27] , [28] . There are roughly 32 , 000 HSF molecules per tetraploid S2 cell [29] and the dissociation constants for trimeric-HSF/HSE interactions are in the picomolar range ( Figure 2E ) ; therefore much of the in vivo HSF binding may be non-functional promiscuous binding . Additional investigation will further illuminate the role of chromatin context in TF binding and the mechanisms by which programmed developmental or environmental chromatin changes permit or deny TF binding . Elucidating the rules that govern accessibility is essential for predicting in vivo occupancy of TFs . Diverse transcription factors [7] , from a broad spectrum of organisms [22] , bind their sequences based on site accessibility . We found that chromatin accessibility as measured by DNase I hypersensitivity could be inferred using ChIP-chip data for various histone modifications and transcription factors . Although our model can infer accessibility based on chromatin composition , the mechanism by which accessibility originates is not addressed . Previous studies have shown that activators , such as HSF , glucocorticoid receptor , and androgen receptor bind to their cognate sites and direct a concomitant increase in local acetylation , DNase I hypersensitivity , and nucleosome depletion [6] , [22] , [30] , [31] . Androgen receptor also acts to position flanking nucleosomes marked by H3K4me2 [31] . These post-TF binding chromatin changes that occur are the result of acetyltransferase and nucleosome remodeler recruitment , both of which functionally interact with activators . For instance , both GR and GATA1 interact with the nucleosome remodeling complex Swi/Snf [32] , [33] . Concomitant increases in locus accessibility likely allow large molecular complexes such as RNA Pol II and coactivators to access the region that in turn can reinforce and maintain active and accessible chromatin . Thorough biophysical characterization of TF binding site properties is critical for accurate predictions of TF binding sites , underscoring the need for more complete models of TF binding . While the commonly used PSSM model makes the assumption of base independence , recent work has revealed that richer models providing for interactions between positions are necessary [34] , [35] . Our model captures critical features of the HSF/HSE interaction that are lost with simpler computational models , namely the interdependencies between the sub-binding sites of each HSF monomer . Consistent with our model , a series of in vitro experiments with S . cerevisiae , D . melanogaster , A . thaliana , H . sapien and D . rerio HSFs indicate that HSF from each of these species can bind to discontinuous HSEs containing canonical pentamers that contain intervening five base pair gaps [36] , [37]; interestingly , however , C . elegans HSF strictly binds to continuous HSEs that do not contain gaps [36] . The complex interactions between positions within a binding site are a critical aspect of inferring whether a polymorphism or mutation affects TF binding . These features should prove useful in providing degenerate HSE sequences for optimal co-crystallization of trimeric HSF and DNA and inferring changes in DNA sequence that affect HSF binding within and between species . In conclusion , the data and models presented here reinforce both the importance of chromatin landscape in modulating in vivo TF binding intensity and how genome wide , chromatin free , binding assays contribute to the understanding of TF sequence binding specificity .
Drosophila HSF was N-terminally tagged with glutathione s-transferase and a tobacco etch virus ( TEV ) protease cleavage site . The C-terminus of the recombinant HSF was fused to the 3xFLAG epitope . Recombinant HSF was purified from E . coli with glutathione resin as previously described [38] , with the following modifications: HSF–3xFLAG elution was achieved by addition of 6xHistidine tagged TEV protease and TEV protease was cleared from the HSF preparation using a Nickel-NTA column . Densitometry was used to show that the HSF protein preparation was 40% full length HSF–3xFLAG , and known amounts of bovine serum albumin ( BSA ) were used to quantify the HSF ( Figure S1 ) . Serial two-fold dilutions of recombinant HSF , from 3 nM ( 1 . 5 nM for the 221 pM HSE ) to 23 . 3 pM , was incubated with 200 attomoles of radiolabeled dsDNA containing modestly degenerate HSEs ( chrX:3380775–3380824 ( 224 pM ) , chr2L:5009892–500994 ( 42 . 7 pM ) , chr2R:3529792–3529841 ( 308 pM ) , chr3L 13470978–13471009 ( 221 pM ) , and chr3L:4073542–4073591 ( 97 . 5 pM ) ) and allowed to come to equilibrium for 30 minutes in a total of 10 µl of 1xHSF binding buffer ( 20 mM HEPES pH 7 . 9 , 10% glycerol , 1 mM EDTA , 4 mM DTT , 3 mM MgCl2 , 100 mM NaCl , 0 . 1% NP-40 , and 300 µg/ml BSA ) at room temperature . Binding reactions were loaded in a 3% agarose TBE ( 10 mM Tris-HCl pH 8 . 0 , 25 mM boric acid , and 1 mM EDTA ) gel and electrophoresed at 50 Volts for 2 hours . The HSF–bound probe and free probe were quantified by densitometry and the dissociation constant , Kd = ( [A][B] ) /[AB] , was estimated using a non-linear least squares method on the function [AB]/[A]total = [B]/ ( [B]+Kd ) where [AB]/[A]total is the measured shifted fraction and [B] is the free HSF trimer concentration . We incubated 600 pM HSF and 2500 ng genomic DNA ( sonicated to 100–600 bp fragment size as previously described [6] ) in 1500 µl final volume of 1xHSF binding buffer and let it come to equilibrium for an hour at room temperature . We added 20 µl ANTI-FLAG M2 affinity gel for 10 minutes and washed 8 times with 1xHSF binding buffer to remove unbound DNA , 3xFLAG peptide was added to a final concentration of 200 ng/µl to specifically elute HSF and HSF–bound DNA . The mock IP was done in the absence of recombinant HSF . We attribute the in vitro binding assay's low background to the design of the experiment . Since recombinant C-terminally 3xFLAG tagged HSF was used , the HSF–associated DNA could be specifically eluted by the addition of excess 3xFLAG peptide . In contrast , standard ChIP protocols rely on non-specific elution of all protein and DNA that binds the resin . The sample preparation was as previously described [6] , except that 15 rounds of amplification were performed in this case . The PB–seq reads were aligned to the Drosophila Genome ( BDGP R5/dm3 ) using BWA ( v 0 . 5 . 8c ) [39] . We obtained 5 , 052 , 425 uniquely aligned reads for replicate one , 4 , 694 , 846 for replicate two and 5 , 410 , 049 for the mock . Files that contain raw sequence data and uniquely aligned reads were deposited into NCBI's Gene Expression Omnibus ( GEO ) [40] , accession number GSE32570 . We called peaks using MACS ( v 1 . 3 . 7 . 1 ) [41] , both for each individual replicate and for the merged set , using a tag size of 55 bp , a starting bandwidth of 100 bp and an appropriate genome size . After experimenting with several p-value thresholds , we selected a value of p = 0 . 01 , which achieved a good tradeoff between maximizing the number of called peaks and ensuring consistency between replicates . Our results were largely unaffected by the ‘mfold’ parameter ( the threshold for fold enrichment relative to background for inclusion in the peak model ) , so we left this parameter at its default value . To improve our sensitivity in binding site detection , we made use of an ensemble of position weight matrices ( PSSMs ) , rather than a single matrix . We sampled 10 , 000 sets of 100 peaks and used the program MEME [42] for motif discovery in each set . As input , MEME was given the 100 bp sequence centered at each peak summit . We used a fixed motif width of 14 bp , a second order background Markov model estimated from the entire peak set , and the ‘zoops’ model ( zero or one site per sequence ) with the restriction that at least 75% of the sequences must contain a site . The resulting PSSMs were compared by KL-divergence against the canonical monomer PSSM ( four base pair unit with consensus AGAA ) estimated from the previously published in vivo high-confidence HSF binding sites detected by ChIP-seq [6] . In each PSSM , one of the three monomers had on average about twice the KL-divergence as the other two . Figure S15 shows a scatter plot of the KL-divergence of the PSSMs in the ensemble Each peak was scanned for matches to all PSSMs in the ensemble , allowing for overlapping sites . The score at each position was taken to be the maximum score across the ensemble . Peaks were split into three groups by GC% quantile , and for each group a 10 kbp sequence was simulated from a second order Markov model , which was then used to estimate the FDR associated with the score . In our context , an appropriate FDR threshold should strike a balance between recapitulation of in vivo results and limiting the number of spurious binding sites . In vivo results are defined by high-confidence peaks , which are ChIP-seq peaks that were called by two peak calling programs and have a corresponding binding site sequence underlying the peak [6] . Whereas , spurious sites are accounted for by limiting the average number of HSE clusters per peak ( set of potentially overlapping HSE no more than 10 bp apart from each other ) . Due to the repetitive nature of the HSE , a cluster is a better representative than a single site of a functional binding locus . We chose a 20% FDR threshold , which maximizes the fraction of peaks having a single HSE cluster while ensuring that a large fraction ( 97% ) of the high-confidence in vivo peaks contain HSEs . This threshold resulted in 3735 clusters ( 71% with a single HSE , 20% with two HSEs overlapping by 10 bp , ∼5% with two HSEs overlapping by 5 bp; see Figure S17 ) . The final set of HSE clusters was obtained by combining data from the two experimental replicates . First , a set of genomic regions was identified by intersecting the peaks from the two experimental replicates , and retaining only those peaks for which the two replicates were in close agreement ( >80% of reads fall in the overlapping region ) . We then identified the 2896 HSE clusters that fell in these regions ( ∼77% of all clusters ) . The problem of measuring the intensity of each peak is complicated by the fact that some peaks contain multiple , closely spaced clusters , whose contributions are difficult to disentangle . Furthermore , peaks often include trailing edges that are dominated by the background signal . To address these concerns we experimented with various measures of intensity based on the output produced by MACS ( wig files giving shifted read counts in 10 bp windows ) as well as the reported ‘bandwidth’ B . We considered three measures , applied to a window of radius B centered at each cluster: maximum read count , read count sum , and an “integrated” read count based on a biweight kernel ( which produces a curve at each peak that is similar to the one implied by the peak model used by MACS ) . We selected the biweight kernel measure , which does the best job of handling closely spaced clusters ( see Figure S18 ) . We assume that each HSE site i is at approximately the same initial concentration in the experiment ( [HSEi]initial = C ) . Furthermore , all sites compete to bind a shared amount of free HSF , with the remaining unbound concentration denoted by [HSF] . At the end of the experiment , a fraction of site i is bound , with concentration [HSEi : HSF] , and the remainder is unbound , with concentration [HSEi] . The dissociation constant for a particular HSE site is therefore given by: The bound HSE concentration is measured by the PB–seq experiment in terms of the number of reads at element i ( Ri ) . This leaves two unknown quantities , [HSF] and [HSEi] , in units of read counts . The first of these unknowns , [HSF] , can be eliminated by considering instead the relative Kd with respect to a known reference value ( for an HSE present in the experiment ) . To solve for [HSEi] , we express this quantity as the difference between the initial concentration C and the measured bound concentration:By substituting the expression for Kdi ( above ) and dividing by the Kd value for the reference HSE , Kdref , we obtain an expression with a single unknown , C:With the use of a reference dissociation value for a second HSE , we can solve for C and obtain estimates of the dissociation constants for all other HSE sites for which read counts are available . Replacing Kdi and Ri by the corresponding values for the second reference HSE and solving for C: Our probabilistic model for HSEs was designed to capture interactions among the binding preferences of the three monomers that form the HSF homotrimer . The model consists of three PSSM-based submodels corresponding to the three 5 bp sequences ( pentamers ) that are bound by the HSF monomers . Each of these submodels is defined by two PSSMs , one ‘strict’ and one ‘relaxed’ . These three submodels allow for eight possible combinations of strict and relaxed pentamer binding . Within each pentamer the positions are considered independent , as in standard PSSM models . Formally , let a candidate 15 bp HSE sequence Xk be composed of random variables Xi , jk where i is the pentamer index and j is the base position within that pentamer . Additionally , let each sequence have an associated unobserved random variable Yk which indicates which of the eight combinations of strict/relaxed distributions are applied the corresponding Xi , jk ( Figure S12 ) . For simplicity , our model definition assumes that the middle monomer sequence has been reverse complemented and is therefore in the same orientation as the outer monomer binding sequences . We considered two versions of the model: a sparsely parameterized ‘constrained’ version and a more parameter-rich ‘expanded’ version , as described below . The chromatin effect and DNase models are rule ensemble models , estimated using the RuleFit R package . This package was also used to estimate the relative importance of the model covariates . The covariates were obtained from modENCODE tracks , taking the mean value over a 200 bp window centered on the target point . Furthermore , these data were filtered to contain only points that had a value for every covariate used . | Transcription factors ( TFs ) bind DNA to modulate levels of gene expression . TF binding sites change throughout development , in response to environmental stimuli , and different tissues have distinct TF binding profiles . The mechanism by which TFs discriminate between binding sites in a context dependent manner is an area of active research , but it is clear that the chromatin environment in which potential binding sites reside strongly influences binding . This study used the Heat Shock TF ( HSF ) to study the effect chromatin has upon induced HSF binding . We implemented an experimental technique to quantify all potential HSF binding sites in the genome . These data were incorporated into a modeling framework along with chromatin landscape information prior to HSF binding to accurately predict the intensities of inducible HSF binding sites . DNase I hypersensitivity and tetra-acetylation of H4 were the most influential covariates in the model . The binding data enabled the development of a more complete HSF/DNA interaction model , providing insight into the biophysical interaction of HSF trimer subunits and target DNA pentamers . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computer",
"science",
"biology"
] | 2012 | Accurate Prediction of Inducible Transcription Factor Binding Intensities In Vivo |
The Mediator is a highly conserved , large multiprotein complex that is involved essentially in the regulation of eukaryotic mRNA transcription . It acts as a general transcription factor by integrating regulatory signals from gene-specific activators or repressors to the RNA Polymerase II . The internal network of interactions between Mediator subunits that conveys these signals is largely unknown . Here , we introduce MC EMiNEM , a novel method for the retrieval of functional dependencies between proteins that have pleiotropic effects on mRNA transcription . MC EMiNEM is based on Nested Effects Models ( NEMs ) , a class of probabilistic graphical models that extends the idea of hierarchical clustering . It combines mode-hopping Monte Carlo ( MC ) sampling with an Expectation-Maximization ( EM ) algorithm for NEMs to increase sensitivity compared to existing methods . A meta-analysis of four Mediator perturbation studies in Saccharomyces cerevisiae , three of which are unpublished , provides new insight into the Mediator signaling network . In addition to the known modular organization of the Mediator subunits , MC EMiNEM reveals a hierarchical ordering of its internal information flow , which is putatively transmitted through structural changes within the complex . We identify the N-terminus of Med7 as a peripheral entity , entailing only local structural changes upon perturbation , while the C-terminus of Med7 and Med19 appear to play a central role . MC EMiNEM associates Mediator subunits to most directly affected genes , which , in conjunction with gene set enrichment analysis , allows us to construct an interaction map of Mediator subunits and transcription factors .
The Mediator , first discovered by Kim et al . ( 1994 ) and Koleske et al . ( 1994 ) [1] , [2] , is a large multiprotein complex which is highly conserved in eukaryotes [3] . Yeast Mediator consists of 25 subunits , organized in 4 different modules: head , middle , tail , and kinase module . It is a general transcription factor ( TF ) that acts as an interface between gene-specific transcription factors and the core transcription machinery ( e . g . , Polymerase II ) . Mediator is required for basal transcription as well as for activated transcription or repression [4]–[6] . In the last years , many successful efforts have been made to gain insight into both structural and functional aspects [7]–[10] . However , though being a well-studied complex , the Mediator still raises a number of unanswered questions: How do the individual subunits contribute to the Mediator's functions ? How is the regulatory information transferred within the Mediator complex , and how does it convey these signals to the core transcription machinery ? Recently , “structure-function” analyses have been suggested and conducted by van de Peppel et al . ( 2005 ) and Koschubs et al . ( 2009 ) [7] , [11] . In a clustering approach , they use expression profile similarity as a proxy for physical interaction , respectively for common module membership . Their method was strikingly successful in identifying physical interactions between Mediator subunits . However , it did not exploit the fact that their data originated from active interventions into the cellular system . Such interventions followed by phenotypic measurements of a cell , as opposed to purely observational data , provide additional insight into the functions and interactions of the respective gene products . Along this line , perturbation experiments have been carried out with low-dimensional readouts ( such as cell viability or growth [12] , [13] ) as well as with high-dimensional phenotypes ( such as genome-wide expression or DNA binding measurements [14] , [15] ) . While the reconstruction of regulatory networks from observational high-dimensional gene expression data has been investigated thoroughly , e . g . , by Basso et al . ( 2005 ) , Segal et al . ( 2003 ) and Segal et al . ( 2005 ) [16]–[18] , the statistical analysis and interpretation of perturbation data is an active field of research [19] , [20] . Nested Effects Models ( NEMs ) are a class of probabilistic graphical models which are tailored for the analysis of gene expression perturbation screens [21]–[28] ( see [29] for a summary ) . They have been applied successfully to the pathway of human MCF-7 breast cancer cells [29] and to a signaling pathway in Drosophila melanogaster [21] . Here , we introduce MC EMiNEM , an efficient and robust learning algorithm for NEMs . MC EMiNEM combines a Markov Chain Monte Carlo ( MC ) sampling procedure with an Expectation-Maximization ( EM ) algorithm in NEMs . The MC EMiNEM method is freely available as a part of the R/Bioconductor package nem . When applied to gene expression data from various Mediator mutant strains , it reveals parts of the functional architecture of the yeast Mediator complex . Moreover , it predicts new interactions between its subunits and gene-specific transcription factors .
Nested Effects Models ( NEMs ) are probabilistic graphical models designed for the analysis of gene expression data from perturbation experiments . They are designed to reconstruct the dependency structure of the perturbation signals , and they perform particularly well if this structure is hierarchical [24] . The graph underlying a NEM contains two types of nodes: the perturbed entities ( the signals ) and the genes for which expression has been measured ( the effects ) . The edges of that graph describe the flow of regulatory information between the nodes . NEMs split this flow into two parts: the signals graph containing the edges between the perturbed entities , and the effects graph describing the assignment of the effect nodes to the signal nodes . We identify the graphs and with their respective adjacency matrices , . The experimental data is summarized in an matrix , where corresponds to the expression data obtained from measurements of effect upon perturbation of signal . NEMs aim at reconstructing the signals graph , assuming a particularly simple regulatory structure: The perturbation of a signal implies the perturbation of other signals that are children of . This in turn perturbs the effect nodes that are the children of the perturbed signals in the effects graph ( see Fig . 1 ) . In other words , the NEM predicts an effect of gene upon perturbation in signal exactly if there is a two-step path from to , i . e . , if . These binary predictions of our model are then linked to the actual measurements by specifying a probability model for the individual effects gene measurements , There is extensive literature on the estimation of these two distributions , see [30] , [31] . Instead of modeling the two distributions separately , it is convenient to estimate their log ratio . For each effect gene , we perform a moderated t-test comparing its expression after perturbation of signal vs . its wild type expression . A false discovery rate estimation procedure is then used to convert the p-values of the moderated t-test into a log odds matrix . This matrix can for instance be obtained using the R/Bioconductor package limma ( see Section S4 . 2 in Text S1 for details ) [32] . Consequently , a NEM is parametrized by the tuple , where is the space of binary matrices with unit diagonal , and is the space of effects graphs . We assume that the effects graph is sparse , such that each effect is linked to at most one signal ( i . e . , each column of equals either a unit base vector of dimension , or the null vector ) . According to Tresch et al . ( 2008 ) [25] , the log posterior of the signals graph is given by ( 1 ) For a derivation of Equation ( 1 ) , see also Section S1 in Text S1 . We assume edge-wise independent priors , , and , . The problem of structure learning in probabilistic graphical models is generally computationally hard ( see [33] ) . A range of methods has been proposed for the maximization of Equation ( 1 ) . It has been observed that it is very difficult to estimate the effects graph reliably . This is not surprising , since the adjacency matrix has the same dimensions as the data matrix . It is therefore desirable to reduce the number of effects a priori . Attaching a gene that never has a positive entry to a signal never increases the posterior . These genes are filtered out prior to the estimation . This step can reduce the number of effects considerably ( from about 6000 effects to roughly 3000 in the case of the Mediator experiments ) . Moreover , we extend the set of signal nodes by a so-called null node , which formally corresponds to extending by a null column . Genes that attach to the null node hence are always predicted inactive . This implements an automated feature selection mechanism within the model ( see also Section S4 . 2 in Text S1 ) . The main objective is the reconstruction of the signals graph . Several approaches try to maximize the ( marginal ) structure posterior by integrating out the hidden parameters ( for a methods review , see [29] ) . This marginalization however is a time consuming step that increases the complexity of the respective algorithms by at least a factor of , making the analysis of larger effects sets ( such as in microarray studies ) slow or even impossible . We avoid this drawback and develop an efficient Expectation-Maximization ( EM ) algorithm for the optimization of the NEM structure posterior ( EMiNEM ) , which , even for large expression data sets , is able to detect a local maximum within seconds . Since the landscape of the structure posterior is rugged ( Fig . S2 . 1 in Text S1 ) , we combine EMiNEM with mode-hopping Markov Chain Monte Carlo ( MC EMiNEM ) for an efficient optimization of the structure posterior . The MC EMiNEM method is freely available as a part of the R/Bioconductor package nem [34]–[36] . It is easy to use , and it does not require external parameters to be set manually . The only parameter that might be tuned is the weight of the sparsity prior , however moderate changes did not change the outcome qualitatively ( see also Sections S2 . 2 and S5 in Text S1 ) . A short introduction to MC EMiNEM is provided in the Supplements ( Section S5 in Text S1 , see also the nem package vignette ) . Throughout this section , the data resp . the matrix is considered given and fixed . We want to find the maximum a posteriori estimate for the signals graph , ( 2 ) This is the classical situation in which Expectation-Maximization is applicable [37] . For excellent introductions to the EM-algorithm , we recommend the tutorials of Minka ( 1998 ) , Neal et al . ( 1998 ) and Dellaert ( 2002 ) [38]–[40] . Briefly , given some guess for , the EM algorithm describes how to find an improved guess such that the sequence is monotonically increasing , and converges ( under mild additional assumptions that are met in our case ) to a local maximum of . The expectation ( E- ) step of the EM algorithm involves calculating the expected log-posterior with respect to the distribution of , given the current guess : ( 3 ) The maximization ( M- ) step of the EM algorithm then consists of finding the maximizer . This is usually a much easier task than solving Equation ( 2 ) directly . We derive an analytical solution , which leads to an efficient closed-form update step for : ( 4 ) with and . A precise definition of the variables contained in Equation ( 4 ) , together with a detailed derivation of this formula is deferred to the Supplements , Text S1 , as it involves elementary but tedious calculations . The EM algorithm is guaranteed to find a local maximum which , for unimodal distributions , equals the global optimum . In practice , the posterior landscape can be very rugged ( see also Fig . S2 . 1 in Text S1 ) . The outcome of the EM algorithm may therefore strongly depend on its initialization , and it may be far from the global optimum ( see also Fig . S2 . 2 in Text S1 ) . This raises the need to explore the set of local maxima provided by EMiNEM . To that end , we introduce MC EMiNEM . In the classical Metropolis-Hastings MCMC approach , consecutive parameter samples are drawn from the distribution . Given , a random process generates a new proposal . The Hastings ratio , a quantity that involves and , then determines the probability of acceptance ( ) or rejection ( ) of the new proposal . The MC EMiNEM algorithm instead applies an EM step to each new proposal , which maps it to the “nearest” local maximum . The acceptance/rejection step is then modified by plugging and into the Hastings ratio , instead of and . We can show that the series of local maxima associated to the underlying Markov chain is approximately drawn from , where ranges exclusively over the space of local maxima . MC EMiNEM's sampling scheme is illustrated in Fig . S2 . 3 in Text S1 . The details of the implementation as well as a theoretical justification of this method are given in Section S2 . 2 and S2 . 3 in Text S1 , respectively . Similar so-called mode hopping approaches have been established by Li et al . ( 1987 ) , Neal et al . ( 1996 ) , Wales et al . ( 1997 ) and Sminchisescu et al . ( 2003 ) [41]–[44] , with applications in areas such as protein folding [45] , nanocluster structure analysis [46] and reconstruction of signaling pathways [47] . Here , we provide a theoretical justification of their use . It is not obvious how the effects graph prior should be defined . Being most conservative , can be chosen uniform , i . e . , for all effects graphs . The posterior is then proportional to the marginal likelihood On the other side , upon availability of precise prior knowledge , can be chosen deterministic , i . e . , , for some fixed adjacency matrix . In this case , the posterior is proportional to the full likelihood . As a trade-off between these two extremes , we initialize in a data-driven fashion ( based on ) , namely ( 5 ) In an Empirical Bayes approach , we iteratively estimate and , and use these distributions as priors for the estimation of the other quantity , respectively . Our Empirical Bayes procedure is:
Extensive simulations were performed to ensure the convergence of the MCMC chain , and to verify the independence of the outcome from the initial parameter choice ( see Section S2 . 2 in Text S1 ) . The prediction quality was assessed in seven parameter settings for different noise levels and different numbers of signal nodes , with observed effect genes and a total number of edges in the signals graph . For each of these scenarios , 50 NEMs were randomly sampled ( for details see Section S3 . 1 in Text S1 ) . In each case , data was generated and afterwards analyzed with various methods: a simple EMiNEM approach without Markov Chain Monte Carlo sampling , the original NEM score [21] , the Nessy method [25] and a random sampling approach ( for details on the competing methods see Section S3 . 3 in Text S1 ) . For all methods , the sensitivity strongly depends on the noise level and the number of signal nodes ( Fig . 2A ) . MC EMiNEM performs best throughout all tested parameter settings , except for low noise where Nessy achieves a similar sensitivity . The specificity of all methods is very high , with a value above 98% in all scenarios ( see also Fig . S3 . 7 in Text S1 ) . A comparison of the method-specific run times is provided in Table S1 in Text S1 . It should be mentioned that EMiNEM itself is extremely efficient , even for large numbers of effect nodes ( one run for the Mediator data took 0 . 1 s on a standard desktop computer ) . This efficiency is a prerequisite that allows us to perform ten thousands of MCMC steps in the MC EMiNEM algorithm in an acceptable time . For a comparison of run times and scalability of the different methods , see Table S1 in Text S1 . Our approach attempts to maximize the marginal posterior . This quantity implicitly depends on the effects graph prior . Therefore , we seek a prior for which the true signals graph scores on the top end of the distribution . It has been shown that NEM models are asymptotically consistent and identifiable [25] , i . e . , given the true effects graph as a deterministic prior , the true signals graph will score best . Thus , a well-chosen effects gene prior might greatly improve the prediction outcome . We tested the following priors: a deterministic prior according to the true effects graph , our Empirical Bayes prior , the data-driven prior used for the initialization of the MCMC sampling ( see S2 . 4 ) , and a uniform effects graph prior . The quality of an effects graph prior is assessed in two ways: First , we calculate the average -distance between the prior to the true prior , where , and normalize it by dividing through the maximum gene-wise -distance , which is . Secondly , we calculate the position of within the marginal posterior distribution . Each posterior distribution was approximated by the empirical distribution of for a random sample of 5000 signals graphs . This was done for the 50 NEM samples that were generated in the most realistic simulation scenario ( 11 nodes , , see Fig . 2 A ) . The results show that the Empirical Bayes prior approaches the true prior better than the other methods , according to the -distances . Furthermore , the resulting posterior is better able to distinguish between signals graphs and to identify the true one ( the true graph is located at the , , and quantile for the uniform , data driven and Empirical Bayes prior , respectively , and at the maximum for the true effects graph; see Fig . 2 B ) . The 25 protein subunits of the Mediator are subdivided into 4 distinct modules ( head , middle , tail , kinase , see Fig . 3 ) . The tail module is believed to establish the contact to the gene-specific transcription factors , based on various TF binding domains , while the head and middle module apparently contact Polymerase II [48] . The kinase module is described as having mostly inhibitory effects on gene expression [49] . The perturbation of a central Mediator subunit can have severe consequences on the structure of the whole Mediator complex . It may cause the loss of whole modules or specific submodules [50]–[52] . The perturbation of a peripheral component might have only local effects on the Mediator structure and , consequently , have fewer effects on transcription . From the structural organization of the Mediator , we therefore expect a hierarchy of transcriptional effects upon subunit perturbations , which makes NEMs a suitable tool for their analysis . As a result of a NEM analysis , we expect the central Mediator subunits that have widespread effects upstream in the signals graph , whereas the more peripheral components should lie downstream . Due to its role as a general transcription factor involved in the formation of the transcription initiation complex , a perturbation of the Mediator can entail global changes in gene expression [53] . Such effects are completely removed by our normalization procedure and can therefore not be detected . Note that systematic variation in RNA extraction , RNA amplification , labeling and scanner calibration make it generally impossible to reliably detect global shifts in transcriptional activity by conventional methods; the absolute quantification of transcription levels requires new experimental techniques , e . g . , as proposed in Sun et al . [54] . Our focus in the present study , however , is on effects that are due to the interaction of the Mediator with gene-specific transcription factors . These effects are restricted to the target genes of the interacting transcription factors . They superimpose to the possible global effects of a Mediator perturbation , and hence become visible only after removal of the global effects . We generated expression profiles of S . cerevisiae Mediator subunit deletion mutants dMed2 , dMed15 , dMed20 , dMed31 , which were complemented by data from published intervention studies on the Mediator . Those comprise mutations of Med7 ( N- and C-terminal deletion ) , and point mutants of Med10 , Med19 , Med20 , Med21 ( see S4 . 1 ) . The raw data is available at ArrayExpress ( accession number E-MTAB-1037 ) . Although there exist even more high-quality gene expression data of Mediator mutants ( e . g . , [52] , [55] ) , we restricted our analysis to experiments that were obtained on the Affymetrix yeast 2 . 0 array under similar environmental conditions . Luckily , some data were redundant in different experiments , which enabled us to correct for batch-specific effects , and to remove outlier genes ( for data pre-processing , see Section S4 . 2 in Text S1 ) . After normalization and batch effect removal , a straightforward application of the MC EMiNEM algorithm led to identical results in 9 out of 10 independent MCMC runs; the tenth run differed only by one edge ( Fig . S4 . 1 , Fig . S4 . 2 in Text S1 ) . The runs revealed a bi-directional edge assigned to the Med10 and Med21 nodes , which means that these two subunits are indistinguishable in terms of their intervention effects . Their attached effect genes are interchangeable without affecting the model's likelihood . Therefore , according to Tresch et al . ( 2008 ) [25] , we combine the two subunits and treat them as one node ( see Section S4 . 2 in Text S1 ) . When Med10 and Med21 were combined , 10 independent MC EMiNEM runs gave identical signals graph predictions ( Fig . 3 ) . The corresponding attachment of effects to signal nodes is provided in Dataset S1 . The predicted Mediator network ( the signals graph in Fig . 3 ) agrees well with current knowledge about the Mediator structure [8] , [10]: When removing the downstream Med7N node , the signals graph is separated into three connected components that reflect the modular organization of the Mediator ( middle module: Med7C , Med19 , Med10Med21 , Med31; head module: Med20; tail module: Med2 , Med15 ) . While the overall module organization of the Mediator can also be recovered from a simple clustering analysis ( see Section S4 . 4 in Text S1 ) , MC EMiNEM reveals a much finer structure by assigning a directionality to each edge . Med7N is downstream of all other nodes , indicating that among all perturbations that were applied , it has the fewest effects on transcription . It shows that there is a set of effects ( attached to Med7N in the NEM ) whose transcription depends on an entirely intact Mediator complex . The middle module component consists of a Med7C , Med10Med21 and Med19 upstream part , and a Med31 , Med7N downstream part . Again , this conforms to its physical architecture: Med7C/Med10Med21 and Med7N/Med31 form stable complexes [8] . We conclude that the former are central architectural components , whereas the latter are peripheral . Indeed , Med7N/Med31 are only weakly attached to the middle module , and easily dissociate from it , whereas Med7C/Med10Med21 are essential for its architecture [8] . The position of Med19 yet is still unclear [56] , [57] . In our model , however , Med19 is clearly placed in the center of the middle module . The tail module interacts with gene-specific transcription factors and is structurally less analyzed [6] . The NEM includes an edge from Med15 to Med2 and thus suggests a more central role for Med15 than for Med2 , because the effects upon perturbation of Med2 are a subset of the respective Med15 effects ( see Fig . 4 and Fig . S4 . 3 in Text S1 ) . Apart from an estimate of the internal flow of regulatory information in the signals graph , MC EMiNEM returns a posterior probability of the attachment of effect genes to specific Mediator subunits ( Fig . 4 ) . The attachment of effects to signal nodes in the NEM framework does not necessarily represent a physical/direct interaction of the Mediator with the DNA . In the case of the Mediator it is sensible to assume that the coupling is mediated by transcription factors ( TFs ) . We extend the analysis of our Mediator network and infer the transcription factors by which this coupling has been achieved ( cf . [28] ) . We group the effect genes according to their attachment to signal nodes and according to the direction of expression change upon perturbation . A gene set enrichment analysis for these 16 groups then reveals interactions of gene-specific TFs with specific Mediator subunits/submodules . We used the MGSA algorithm for the enrichment analysis [58] , based on the gene-TF assignment by Mac Isaac et al . ( 2006 ) [59] ( see also Section S4 . 3 in Text S1 ) . Although the attachment of individual effects to Mediator subunits is notoriously variable ( see Fig . S3 . 5 and S4 . 6 in Text S1 ) , the gene set enrichment approach lends its robustness from combining evidence from many attached genes . The result is a map of TF-Mediator interactions , summarized in Fig . 3 and listed in Table S2 in Text S1 . The 21 TF-Mediator subunit interactions mapped by MC EMiNEM were validated using the BioGRID database [60] . Two interaction pairs were known from the literature ( YAP1-Med2 , SWI4-Med2 ) . Another eight TFs were known to interact with a Mediator subunit from the same module as the predicted interacting subunit ( [GLN3/SWI5]-Med7N , RPN4-Med7C , [SKN7/STB5/INO4/HAP3]-Med10Med21 , ASH1-Med2 ) . An interaction with the Mediator has been described for three more TFs ( [UME6/HAP4]-Med10Med21 , SUM1-Med2 ) , and eight predicted interactions were new ( MBP1-Med7C , [HSF1/SKO1]-Med10Med21 , [TEC1/YAP6/GTS1]-Med2 , [FKH2/YOX1]-Med7N ) . All target genes of TFs associated with the tail module show downregulation after perturbation , consistent with the tail's function to contact gene specific transcription factors [5] . The same holds for the target genes of TFs associated with Med7N . This is expected , as the genes attached to Med7N are those that show an effect in all perturbations ( Fig . 4 ) and therefore presumably require a completely intact Mediator . The target genes of TFs associated to the rest of the middle module show expression changes in both directions , in accordance with the middle module described as an ambiguous regulator [11] . Fig . 5 A ) offers a TF-centric excerpt on the MC EMiNEM map from Fig . 4 . It drills in to the target genes of SKO1 , which are enriched in the set of upregulated genes attached to Med10Med21 . SKO1 is both a transcriptional activator and repressor and forms a complex with the general repressor TUP1 ( Saccharomyces Genome Database [61] ) . TUP1 in turn targets Med21p [62] . A Mediator complex lacking this subunit might thus not be able to forward repressive signals , resulting in upregulated target genes of SKO1 . The transcriptional activator SWI5 has a large number of physical interactions with subunits from various Mediator modules ( Med15 , Med17 , Med18 , Med22 , [61] ) . This suggests that any change in the Mediator structure affects its interaction with SWI5 . Consequently , target genes of SWI5 should change their expression upon any Mediator subunit perturbation . Fig . 5 B ) confirms this behavior of the SWI5 targets: MC EMiNEM associates SWI5 to Med7N , because SWI5 targets are enriched in the set of downregulated genes attached to Med7N , and these are consistently downregulated in all perturbations . Similar analyses were carried out for all TFs in the MC EMiNEM map ( Figure S1; lists of genes that contribute to the respective TF enrichments are provided in Dataset S2 ) . The most striking observation is that the sign of a gene's expression change is consistent in virtually all perturbations for which MC EMiNEM predicts an effect . Since our model is completely blind with respect to the sign of regulation , the consistency in the direction of the expression changes provides compelling evidence that the signals graph reflects regulatory dependencies between Mediator subunits which are likely to be caused by structural changes . The reconstruction of interaction networks from high dimensional perturbation effects is still a challenge . We have developed MC EMiNEM , a method for the learning of a Nested Effects Model . We introduced two major improvements , namely an Expectation-Maximization algorithm for the very fast detection of local maxima of the posterior probability function . Mode hopping Markov Chain Monte Carlo sampling was then used for the efficient exploration of the space of local maxima . We applied MC EMiNEM to a combination of proper and public gene expression data obtained from Mediator subunit perturbations . It turned out that MC EMiNEM does not only shed light on structural dependencies of Mediator subunits , it also identifies interactions of gene-specific transcription factors with Mediator subunits . Our findings are consistent with the state-of-the-art knowledge about the Mediator architecture and function . By grouping of components with similar profiles , hierarchical clustering has proved tremendously useful for the analysis of expression data obtained from observational experiments . MC EMiNEM reaches beyond the identification of undirected relationships; it resolves directed regulatory structures , and it identifies gene groups with a consistent and specific response pattern . For interventional data , MC EMiNEM is thus the appropriate counterpart to clustering . | Phenotypic diversity and environmental adaptation in genetically identical cells is achieved by an exact tuning of their transcriptional program . It is a challenging task to unravel parts of the complex network of involved gene regulatory components and their interactions . Here , we shed light on the role of the Mediator complex in transcription regulation in yeast . The Mediator is highly conserved in all eukaryotes and acts as an interface between gene-specific transcription factors and the general mRNA transcription machinery . Even though most of the involved proteins and numerous structural features are already known , details on its functional contribution on basal as well as on activated transcription remain obscure . We use gene expression data , measured upon perturbations of various Mediator subunits , to relate the Mediator structure to the way it processes regulatory information . Moreover , we relate specific subunits to interacting transcription factors . | [
"Abstract",
"Introduction",
"Methods",
"Results/Discussion"
] | [
"statistics",
"microbiology",
"dna",
"transcription",
"mathematical",
"computing",
"model",
"organisms",
"mathematics",
"signaling",
"in",
"cellular",
"processes",
"transcriptional",
"signaling",
"gene",
"expression",
"computing",
"methods",
"biology",
"molecular",
"biology... | 2012 | MC EMiNEM Maps the Interaction Landscape of the Mediator |
In the last two decades , mitochondrial DNA ( mtDNA ) and the non-recombining portion of the Y chromosome ( NRY ) have been extensively used in order to measure the maternally and paternally inherited genetic structure of human populations , and to infer sex-specific demography and history . Most studies converge towards the notion that among populations , women are genetically less structured than men . This has been mainly explained by a higher migration rate of women , due to patrilocality , a tendency for men to stay in their birthplace while women move to their husband's house . Yet , since population differentiation depends upon the product of the effective number of individuals within each deme and the migration rate among demes , differences in male and female effective numbers and sex-biased dispersal have confounding effects on the comparison of genetic structure as measured by uniparentally inherited markers . In this study , we develop a new multi-locus approach to analyze jointly autosomal and X-linked markers in order to aid the understanding of sex-specific contributions to population differentiation . We show that in patrilineal herder groups of Central Asia , in contrast to bilineal agriculturalists , the effective number of women is higher than that of men . We interpret this result , which could not be obtained by the analysis of mtDNA and NRY alone , as the consequence of the social organization of patrilineal populations , in which genetically related men ( but not women ) tend to cluster together . This study suggests that differences in sex-specific migration rates may not be the only cause of contrasting male and female differentiation in humans , and that differences in effective numbers do matter .
Understanding the extent to which sex-specific processes shape human genetic diversity has long been a matter of great interest for human population geneticists [1] , [2] . To date , as detailed in Table 1 , the focus has mainly been on the analysis of uniparentally inherited markers: mitochondrial DNA ( mtDNA ) and the non-recombining portion of the Y chromosome ( NRY ) . A large number of studies have found that the level of differentiation was greater for the Y chromosome than for mtDNA , both at a global [3] and a local scale [4]–[11] , for a review see [12] . This result has mainly been explained by patrilocality , a widespread tendency for men to stay in their birthplace while women move to their husband's house [13] ( see Table 1 for more detailed interpretations ) . This hypothesis of a higher migration rate of women has been especially strengthened by the comparison of patrilocal and matrilocal populations at a local scale [14]–[17] . These studies have shown that in patrilocal populations , genetic differentiation is stronger among men than among women , while the reverse is observed in matrilocal populations . It is also noteworthy that the absolute difference between male and female genetic structure is more pronounced in patrilocal than in matrilocal populations [16] . Interestingly , while social practices seem to consistently influence the sex-specific demography at a local scale , the robustness of a sex-specific genetic structure at a global scale is still a challenging issue ( see Table 1 ) . A recent analysis of mtDNA and NRY variation at a global scale , which used the same panel of populations for both categories of markers ( an omission that was criticized in Seielstad et al . 's [3] study [18] ) showed no difference between the male and female genetic structure [19] . Consistent with this result , an analysis of the autosomal and X-linked microsatellite markers in the HGDP-CEPH Human Genome Diversity Cell Line Panel showed no major differences between the demographic history of men and women [20] . The apparent paradox between local and global trends can be resolved though , since the geographical clustering of populations with potentially different lifestyles may minimize the differences in sex-specific demography at a global scale [21] , [22] . It may also be that the global structure reflects more ancient , pre-agricultural , social patterns , as patrilocality may only have increased in human societies only with the recent transition to agriculture [12] . The higher differentiation level found on the NRY as compared to mtDNA at a local scale could also be the consequence of a higher effective number of women , for example through the practice of polygyny , a tendency for men ( but not for women ) to have multiple mates [4] , [7] , [15] , [23]–[25] , and/or through the paternal transmission of reproductive success [11] . However , the influence of such processes on genetic structure has often been considered as negligible , since realistic rates of polygyny cannot create large differences in male and female genetic structure [3] , [5] , [14] . Hence , until now , the effect of local social processes on male and female effective numbers has not been investigated directly , possibly because current methods fail to unravel the relative contribution of effective number and migration rate on the differentiation level [26] . The consequence is that the vast majority of studies fail to show whether the observed differentiation arises from sex-specific differences in migration rate , effective numbers , or both ( see Table 1 ) . New methods need therefore to be developed in order to appreciate the relative influence of sex-biased dispersal and differences in effective numbers on genetic structure . Another limitation to the use of uniparentally inherited markers stems from the fact that each of them is , in effect , a single genetic locus . For that reason , we cannot test for the robustness of the sex-specific genetic structure on these markers . We cannot either rule out the possibility that mtDNA and NRY , which contain multiple linked genes , may be shaped by selection [27] , [28] . This raises the question of whether results based on uniparentally inherited markers simply reflect stochastic variation , or real differences in sex-specific demography . To answer this question , we propose a novel approach based on the joint analysis of autosomal and X-linked markers . This multi-locus analysis has the potential of providing more robust information , as these markers give an independent picture of sex-specific demography . This approach also aims to disentangle the effects of sex-biased dispersal and effective numbers on genetic structure . In order to recognize the impact of social organization on these differences , we investigate sex-specific genetic structure in human populations of Central Asia ( Figure 1 ) , where various ethnic groups , characterized by different languages , lifestyles and social organizations , co-exist . Although all groups share a patrilocal organization , Tajiks ( sedentary agriculturalists ) are bilineal , i . e . they are organized into nuclear or extended families where blood links and rights of inheritance through both male and female ancestors are of equal importance , and they preferentially establish endogamous marriages with cousins . By contrast , Kazaks , Karakalpaks , Kyrgyz and Turkmen ( traditionally nomadic herders ) are patrilineal , i . e . they are organized into paternal descent groups ( tribes , clans , lineages ) , and they practice exogamous marriages , in which a man chooses a bride from a different clan .
We sampled 780 healthy adult men from 10 populations of bilineal agriculturalists and 11 populations of patrilineal herders from West Uzbekistan to East Kyrgyzstan , representing 5 ethnic groups ( Tajiks , Kyrgyz , Karakalpaks , Kazaks , and Turkmen ) ( see Figure 1 and Table 2 ) . We genotyped all bilineal populations , and 8 out of 11 patrilineal populations at the HVS-I locus of mtDNA , and at 11 microsatellite markers on the NRY ( for more details on the markers used , see Table 3 ) . The overall genetic differentiation was higher for NRY , as compared to mtDNA , both among the 10 bilineal agriculturalist populations , and among the subset of 8 patrilineal herder populations . Assuming an island model of population structure , this implies that female migration rate ( mf ) , and/or the effective number of females ( Nf ) , is higher than of the corresponding parameters for males ( mm and Nm ) . These results also suggest that the differences in sex-specific genetic structure are much more pronounced in the patrilineal herders than in the bilineal agriculturalists . From the above FST estimates , we obtained the female-to-male ratio of the effective number of migrants per generation ( see the Methods section for details ) : Nfmf/Nmmm≈2 . 1 for bilineal populations and Nfmf/Nmmm≈21 . 6 for patrilineal populations . The ratio in patrilineal populations is thus one order of magnitude higher than in bilineal populations . However , since each of these markers is a single genetic locus , we cannot test for the robustness of the sex-specific genetic structure on these markers . We therefore examined the amount of information contained in multi-locus data on autosomal and X-linked markers , both of which average over male and female histories . In the infinite island model of population structure with two classes of individuals ( males and females ) , we obtained the following expressions of FST ( see the Methods section for details ) : ( 1 ) for autosomal genes , and ( 2 ) for X-linked genes . A special case of interest occurs when , i . e . when the differentiation of X-linked genes exactly equals that of autosomal genes . Combining eqs ( 1 ) and ( 2 ) , we find that this occurs for , with N = Nf+Nm and m = mf+mm . Furthermore , as shown in Figure 2 , if we observe a lower genetic differentiation of autosomal markers , as compared to X-linked markers ( blue zone in Figure 2 ) , this suggests that . This may happen , e . g . , for Nf = Nm and mf = mm , i . e . for equal effective numbers of males and females and unbiased dispersal . But if autosomal markers are more differentiated than X-linked markers ( , see the red upper-right triangle in Figure 2 ) , this implies that . In this case , since mf/m and Nf/N are ratios varying between 0 and 1 , the effective number of females must be higher than that of males ( Nf>Nm ) , and the female migration rate must be higher than half the male migration rate ( mf>mm/2 ) . Hence , a prediction from this model is that when , the effective number of females is higher than that of males , whatever the pattern of sex-specific dispersal . This suggests that it is indeed possible to test for differences in effective numbers between males and females from the joint analysis of autosomal and X-linked data . We note however that when , we cannot conclude on the relative male and female effective numbers and migration rates . We tested the above prediction in the 10 bilineal agriculturalist populations and 11 patrilineal herder populations sampled in Central Asia by comparing the genetic structure estimated from 27 unlinked polymorphic autosomal microsatellite markers ( AR = 16 . 2 , He = 0 . 803 on average ) to that from 9 unlinked polymorphic X-linked microsatellite markers ( AR = 12 . 6 , He = 0 . 752 on average ) ( for more details on the markers used , see Table 4 ) . Overall heterozygosity was not significantly different between X-linked and autosomal markers , neither in the pooled sample ( two-tailed Wilcoxon sum rank test; p = 0 . 09 ) , nor in the bilineal agriculturalists ( p = 0 . 13 ) or the patrilineal herders ( p = 0 . 12 ) . The overall population structure was significantly higher for autosomal as compared to X-linked markers among patrilineal herders: ( one-tailed Wilcoxon sum rank test; ; p = 0 . 02 ) . Among bilineal agriculturalists , the result was not significant: ( p = 0 . 36 ) . From these results , and following our model predictions , we conclude that in patrilineal herders ( where ) , the effective number of females is higher than that of males . This conclusion does not hold for the bilineal agriculturalists . From our model , it is possible to get more precise indications on the sets of ( Nf/N , mf/m ) values that are compatible with our data . Rearranging eqs ( 1–2 ) , we get: ( 3 ) i . e . : ( 4 ) For any given set of ( Nf/N , mf/m ) values , we can therefore calculate from eq . ( 4 ) the expected value of for each estimate in the dataset . We can then test the null hypothesis by comparing the distribution of observed and expected values . If the hypothesis can be rejected at the α = 0 . 05 level , then the corresponding set of ( Nf/N , mf/m ) values can also be rejected . Following Ramachandran et al . [20] , we varied the values of the ratios Nf/N and mf/m ( respectively , the female fraction of effective number , and the female fraction of the total migration rate ) from 0 to 1 , with an interval of 0 . 01 between consecutive values . For each set of ( Nf/N , mf/m ) values , we applied the transformation in eq . ( 4 ) to each of the 27 locus-specific values observed . Thus , for each set of ( Nf/N , mf/m ) values , we obtained 27 expected values of , given our data . These expected values of were then compared to the 9 observed locus-specific in our dataset , and we calculated the p-value for a two-sided Wilcoxon sum rank test between the list of 27 expected values and the 9 observed in the dataset . The results are depicted in Figure 3 . Significant p-values ( p≤0 . 05 ) correspond to a significant difference between the observed and expected values , thus to sets of ( Nf/N , mf/m ) values that are rejected , given our data ( see the blue region in Figure 3 ) . Conversely , non-significant p-values ( p>0 . 05 ) correspond to sets of ( Nf/N , mf/m ) values that cannot be rejected ( see the red region in Figure 3 ) . For the patrilineal herder populations ( Figures 3A–3B ) , most sets of ( Nf/N , mf/m ) values are rejected , except those corresponding to larger effective numbers for females ( from Figures 3A–3B: Nf/N>0 . 55 , i . e . Nf>1 . 27Nm ) and mf>0 . 67mm . Because the multi-locus estimate of is significantly higher than the estimate of , we expected to find such patterns of non-significant values ( see Figure 2 ) . For the bilineal agriculturalist populations , we could not reject the hypothesis that the effective numbers and migration rates are equal across males and females or even lower in females ( see Figures 3C–3D ) . This is also reflected by the fact that the estimates of were not significantly higher than the estimates of in those populations . Finally , we have shown that the effective number of women is higher than that of men among patrilineal herders , but not necessarily among bilineal agriculturalists . Furthermore , a close inspection of the results depicted in Figures 3A and 3B reveals that , among herders , we reject all the sets of ( Nf/N , mf/m ) values for which mf<mm at the α = 0 . 10 level . This is not true for agriculturalists . This suggests that the migration rates are also likely to be higher for women than for men in patrilineal populations , as compared to bilineal populations ( compare Figures 3B and 3D ) . Although both groups are patrilocal , such a difference in sex-specific migration patterns might be expected , since patrilineal herders are exogamous ( among clans ) and bilineal agriculturalists are preferentially endogamous . For example , it was observed that in patrilocal and matrilocal Indian populations , where migrations are strictly confined within endogamous groups , sex-specific patterns were not influenced by post-marital residence [21] . While an influence of post-marital residence on the migration rate of women and men has already been widely proposed [14]–[17] ( see also Table 1 ) , the factors that may locally affect the effective number of women , relatively to that of men , are not well recognized . As seen in Table 1 , although a number of studies have compared matrilocal and patrilocal populations , few have compared contrasting groups of populations with respect to other factors as , e . g . , the tendency for polygyny [15] . Furthermore , a number of these studies lack ethnological information a priori , concerning social organization , marriage rules , etc . , which makes interpretation somewhat difficult ( see Table 1 ) . Here , we compared two groups of patrilocal populations with contrasting social organizations , and at least five non-mutually exclusive interpretations for a larger effective number of females can be invoked: There might also be non-biological explanations of our results , however , as they are based on the simplifying assumptions of Wright's infinite island model of population structure [39] . This model assumes ( i ) that there is no selection and that mutation is negligible , ( ii ) that each population has the same size , and sends and receives a constant fraction of its individuals to or from a common migrant pool each generation ( so that geographical structure is absent ) , and ( iii ) that equilibrium is reached between migration , mutation and drift . On the first point , we did not find any evidence of selection , for any marker , based on Beaumont and Nichols' method [40] for detecting selected markers from the analysis of the null distribution generated by a coalescent-based simulation model ( data not shown ) . As for the second point , we tested for the significance of the correlation between the pairwise FST/ ( 1−FST ) estimates and the natural logarithm of their geographical distances [41] . We found no evidence for isolation by distance , either for X-linked markers ( p = 0 . 47 for agriculturalists , p = 0 . 24 for herders ) , or for autosomal markers ( p = 0 . 92 for agriculturalists , p = 0 . 45 for herders ) . As for the third point , the X-to-autosomes ( X/A ) effective size ratio can significantly deviate from the expected three-quarters ( assuming equal effective numbers of men and women ) following a bottleneck or an expansion [42] . This is because X-linked genes have a smaller effective size , and hence reach equilibrium more rapidly . After a reduction of population size , the X/A diversity ratio is lower than expected , while after an expansion , the diversity of X-linked genes recovers faster than on the autosomes , and the X/A diversity ratio is then closer to unity . In the latter case , would be reduced and could then tend towards . However , neither reduction nor expansion should lead to , as we found in herder populations of Central Asia . Therefore , we do not expect the limits of Wright's island model to undermine our approach . We aimed to investigate to what extent the approach proposed here is able to detect differences in male and female effective numbers . To do this , we performed coalescent simulations in a finite island model , for a wide range of ( Nf/N , mf/m ) values . The simulation parameters were set to match those of our dataset: 11 sampled demes , 30 males genotyped at 27 autosomal and 9 X-linked markers per deme ( for further details concerning the simulations , see the Methods section ) . We used 1421 sets of ( Nf/N , mf/m ) values , covering the whole parameter space ( represented as white dots in Figure 4B ) . For each set of ( Nf/N , mf/m ) parameter values , we simulated 100 independent datasets . For each dataset , we calculated the estimates of at all loci , and we calculated the p-value for a one-sided Wilcoxon sum rank test for the list of estimates . Hence , for each set of ( Nf/N , mf/m ) parameter values , we could calculate the proportion of significant tests at the α = 0 . 05 level , among the 100 independent datasets . Figure 4 shows the distribution of the percentage of significant tests in the ( Nf/N , mf/m ) parameter space . Theory predicts that in the upper-right triangle where , we should have . One can see from Figure 4 that , given the simulation parameters used , the method is conservative: the proportion of significant tests at the α = 0 . 05 level is null outside of the upper-right triangle . However , we find a fairly large proportion of significant tests for large Nf/N and mf/m ratios which indicates ( i ) that the method presented here has the potential to detect differences in male and female effective numbers , but ( ii ) that only strong differences might be detected , for similarly sized datasets as the one considered here . We also aimed to investigate whether the results obtained here were robust to our sampling scheme , and that our results were not biased by the inclusion of particular populations . To do this , we re-analyzed both the bilineal agriculturalists and the patrilineal herders datasets , removing one population at a time in each group . For each of these jackknifed datasets , we calculated the p-value of a one-sided Wilcoxon sum rank test , as done on the full datasets . The results are given in Table 5 . We found no significant test for any of the bilineal agriculturalist groupings ( p>0 . 109 ) , which supports the idea that , in those populations , both the migration rate and the number of reproductive individuals can be equal for both sexes . In patrilineal herders , the tests were significant at the α = 0 . 05 level for 8 out of 11 population groupings . For the 3 other groupings , the p-values were 0 . 068 , 0 . 078 and 0 . 073 ( see Table 5 ) . Overall , the ratio of multi-locus estimates ranged from 1 . 7 to 3 . 5 in patrilineal herders ( and from 0 . 9 to 1 . 2 in bilineal agriculturalists ) . Although in some particular groupings of patrilineal herder populations , the difference in the distributions of may not be strong enough to be significant , we can clearly distinguish the pattern of differentiation for autosomal and X-linked markers in patrilineal and bilineal groups . Results from coalescent simulations ( see above ) suggest that this lack of statistical power might be expected for ratios close to unity . Indeed , we found that the tests were more likely to be significant for fairly large Nf/N and mf/m ratios ( the upper-right red region in Figure 4 ) which would correspond to ratios much greater than one . Importantly , our results on X-linked and autosomal markers are consistent with those obtained from NRY and mtDNA ( see Figures 3B–3D ) : in these figures , the dashed line gives all the sets of ( Nf/N , mf/m ) values that are compatible with the observed estimates . These are the sets of values that satisfy for the bilineal populations , and for the patrilineal populations , since we inferred Nfmf/Nmmm≈2 . 1 and Nfmf/Nmmm≈21 . 6 , respectively , for the two groups . For the bilineal agriculturalists ( Figure 3D ) , the set of ( Nf/N , mf/m ) values inferred from the estimates fall within the range that was not rejected , given our data on X-linked and autosomal markers . For the patrilineal herders ( Figure 3B ) , the overlap is only partial: from the NRY and mtDNA data only , low Nf/N ratios associated with high mf/m ratios are as likely as high Nf/N ratios associated with low mf/m ratios . Yet , it is clear from this figure that a large set of ( Nf/N , mf/m ) values inferred from the single-locus estimates can be rejected , given the observed differentiation on X-linked and autosomal markers . All genetic systems ( mtDNA , NRY , X-linked and autosomal markers ) converge toward the notion that patrilineal herders , in contrast to bilineal agriculturalists , have a strong sex-specific genetic structure . Yet , the information brought by X-linked and autosomal markers is substantial , since we show that this is likely due to both higher migration rates and larger effective numbers for women than for men . Our results , based on the X chromosome and the autosomes , also confirm previous analyses based on the mtDNA and the NRY , showing that men are genetically more structured than women in other patrilocal populations [3]–[10] , [14]–[17] ( see also Table 1 ) . A handful of studies have also shown a reduced effective number of men compared to that of women , based on coalescent methods [23] , [24] , but none have considered the influence of social organization on this dissimilarity ( see Table 1 ) . In some respects , our results contrast with those of Wilder and Hammer [25] , who studied sex-specific population genetic structure among the Baining of New Britain , using mtDNA , NRY , and X-linked markers . Interestingly , they found that Nf>Nm , but mf<mm , and claimed that a similar result , although left unexplored by the authors , was to be found in a recent study by Hamilton et al . [16] . This raises the interesting point that sex-specific proportions of migrants ( m ) are likely to be shaped by factors that may only partially overlap with those that affect the sex-specific effective numbers ( N ) . Further studies of human populations with contrasted social organizations , as well as further theoretical developments , are needed to appreciate this point . In order to ask to what extent our results generalize to other human populations , we investigated sex-specific patterns in the 51 worldwide populations represented in the HGDP-CEPH Human Genome Diversity Cell Line Panel dataset [43] , for which the data on the differentiation of 784 autosomal microsatellites and 36 X-linked microsatellites are available ( data not shown ) . By doing this , we found a larger differentiation for X-linked than for autosomal markers . Therefore , we confirmed Ramachandran et al . 's [20] result that no major differences in demographic parameters between males and females are required to explain the X-chromosomal and autosomal results in this worldwide sample . Ramachandran et al . 's approach [20] is based upon a pure divergence model from a single ancestral population , which is very different from the migration-drift equilibrium model considered here . In real populations , however , genetic differentiation almost certainly arises both through divergence and limited dispersal , which places these two models at two ends of a continuum . Yet , importantly , if we apply Ramachandran et al . 's [20] model to the Central Asian data , our conclusions are left unchanged . In their model , the differentiation among populations is , where t is the time since divergence from an ancestral population and Ne the effective size of the populations ( see , e . g . , [44] ) . Hence , we get for autosomal and X-linked markers , respectively . Therefore , our observation that implies that , which requires that Nf>7Nm since ( see , e . g . , [45] ) . In this case , the female fraction of effective number is larger than that of males , which is consistent with our findings in a model with migration . The HGDP-CEPH dataset does not provide any detailed ethnic information for the sampled groups , and we can therefore not distinguish populations with different lifestyles . However , at a more local scale in Pakistan , we were able to analyze a subset of 5 populations ( Brahui , Balochi , Makrani , Sindhi and Pathan ) , which are presumed to be patrilineal [46] . For this subset , we found a higher differentiation for autosomal than for X-linked markers , although non-significantly ( p = 0 . 12 ) . This result seems to suggest that other patrilineal populations may behave like the Central Asian sample presented here . Therefore , because the geographical clustering of populations with potentially different lifestyles may minimize the differences in sex-specific demography at a global scale [21] , [22] , and/or because the global structure may reflect ancient ( pre-agricultural ) marital residence patterns with less pronounced patrilocality [12] , we emphasize the point that large-scale studies may not be relevant to detect sex-specific patterns , which supports a claim made by many authors . In conclusion , we have shown here that the joint analysis of autosomal and X-linked polymorphic markers provides an efficient tool to infer sex-specific demography and history in human populations , as suggested recently [12] , [47] . This new multilocus approach is , to our knowledge , the first attempt to combine the information contained in mtDNA , NRY , X-linked and autosomal markers ( see Table 1 ) , which allowed us to test for the robustness of a sex-specific genetic structure at a local scale . Unraveling the respective influence of migration and drift upon neutral genetic structure is a long-standing quest in population genetics [48] , [49] . Here , our analysis allowed us to show that differences in sex-specific migration rates may not be the only cause of contrasted male and female differentiation in humans and that , contrary to the conclusion of a number of studies ( see Table 1 ) , differences in effective numbers may also play an important role . Indeed , we have demonstrated that sex-specific differences in population structure in patrilineal herders may be the consequence of both higher female effective numbers and female effective dispersal . Our results also illustrate the importance of analyzing human populations at a local scale , rather than global or even continental scale [2] , [19] , [21] . The originality of our approach lies in the comparison of identified ethnic groups that differ in well-known social structures and lifestyles . In that respect , our study is among the very few which compare patrilineal vs . bilineal or matrilineal groups ( see Table 1 ) , and we believe that it contributes to the growing body of evidence showing that social organization and lifestyle have a strong impact on the distribution of genetic variation in human populations . Moreover , our approach could also be applied on a wide range of animal species with contrasted social organizations . Therefore , we expect our results to stimulate research on the comparison of X-linked and autosomal data to disentangle sex-specific demography .
We sampled 10 populations of bilineal agriculturalists and 11 populations of patrilineal herders from West Uzbekistan to East Kyrgyzstan , representing 780 healthy adult men from 5 ethnic groups ( Tajiks , Kyrgyz , Karakalpaks , Kazaks , and Turkmen ) ( see Table 2 ) . The geographic distribution of the samples and information about lifestyle is provided in Figure 1 . Also living in Central Asia , Uzbeks are traditionally patrilineal herders too , but they have recently lost their traditional social organization [11] , and we therefore chose not to include any sample from this ethnic group for the purpose of this study . We collected ethnologic data prior to sampling , including the recent genealogy of the participants . Using this information , we retained only those individuals that were unrelated for at least two generations back in time . All individuals gave their informed consent for participation in this study . Total genomic DNA was isolated from blood samples by a standard phenol-chloroform extraction [50] . The mtDNA first hypervariable segment of the mtDNA control region ( HVS-I ) was amplified using primers L15987 ( 5′TCAAATGGGCCTGTCCTTGTA ) and H580 ( 5′TTGAGGAGGTAAGCTACATA ) in 18 populations out of 21 ( 674 individuals , see Table 2 ) . The amplification products were subsequently purified with the EXOSAP standard procedure . The sequence reaction was performed using primers L15925 ( 5′TAATACACCAGTCTTGTAAAC ) and HH23 ( 5′AATAGGGTGATAGACCTGTG ) . Sequences from positions 16 024–16 391 were obtained . Eleven Y-linked microsatellite markers ( see Table 3 ) were genotyped in the same individuals , following the protocol described by Parkin et al . [51] . 27 autosomal and 9 X-linked microsatellite markers ( see Table 4 ) were genotyped in the same individuals . We used the informativeness for assignment index In [52] to select subsets of microsatellite markers on the X chromosome and the autosomes from the set of markers used in Rosenberg et al . 's worldwide study [43] . This statistic measures the amount of information that multiallelic markers provide about individual ancestry [52] . This index was calculated among a subset of 14 populations , chosen from the Rosenberg et al . 's dataset [43] to be genetically the closest to the Central Asian populations ( Balochi , Brahui , Burusho , Hazara , Pathan , Shindi , Uygur , Han , Mongola , Yakut , Adygei , Russian , Druze and Palestinian ) . The rationale was to infer the information provided by individual loci about ancestry from this subset of populations , and to extrapolate the results to the populations studied here . For the X chromosome data , we pooled the ‘Screening Set10’ and ‘Screening Set52’ from the HGDP-CEPH Human Genome Diversity Cell Line Panel [53] analyzed by Rosenberg et al . [43] which represented a total of 36 microsatellites . We chose 9 markers among the 11 with the highest In . For autosomal data , we used the ‘Screening Set10’ , which represented a total of 377 microsatellites , and chose 27 markers among the 30 with the highest In . All markers were chosen at a minimum of 2 cM apart from each others [54] . PCR amplifications were performed in a 20 µl final volume composed of 1× Eppendorf buffer , 125 µM each dNTP , 0 . 5U Eppendorf Taq polymerase , 125 nM of each primer , and 10 ng DNA . The reactions were performed in a Eppendorf Mastercycler with an initial denaturation step at 94°C for 5 min; followed by 36 cycles at 94°C for 30 s , 55°C for 30 s , 72°C for 20 s , and 72°C for 10 min as final extension . Forward primers were fluorescently labeled and reactions were further analyzed by capillary electrophoresis ( ABI 310 , Applied Biosystems ) . We used the software package Genemarker ( SoftGenetics LLC ) to obtain allele sizes from the analysis of PCR products ( allele calling ) . We calculated the total allelic richness ( AR ) ( over all populations ) , the unbiased estimate of expected heterozygosity He [55] , the total number of polymorphic sites and FST for mtDNA using Arlequin version 3 . 1 . [56] . Genetic differentiation among populations for the autosomes , the X and the Y chromosome was measured both per locus and overall loci using Weir and Cockerham's FST estimator [57] , as calculated in Genepop 4 . 0 . [58] . The 95% confidence intervals were obtained by bootstrapping over loci [58] , using the approximate bootstrap confidence intervals ( ABC ) method described by DiCiccio and Efron [59] . Isolation by distance ( i . e . the correlation between the genetic and the geographic distances ) was analyzed by computing the regression of pairwise FST/ ( 1−FST ) estimates between pairs of populations to the natural logarithm of their geographical distances , and rank correlations were tested using the Mantel permutation procedure [60] , as implemented in Genepop 4 . 0 . [58] . All other statistical tests were performed using the software package R v . 2 . 2 . 1 [61] . Let us consider an infinite island model of population structure [62] , with two classes of individuals ( males and females ) , which describes a infinite set of populations with constant and equal sizes that are connected by gene flow . Then the expected values of FST for uniparentally inherited markers depend on the effective number Nm ( resp . Nf ) of adult males ( resp . females ) per population and the migration rate mm ( resp . mf ) of males ( resp . females ) per generation , as: ( see , e . g . , [63] ) . We can therefore calculate the female-to-male ratio of the effective number of migrants per generation as: . In this model , we can also compute for the autosomes and the X chromosome the reproductive values for each class ( sex ) , which are interpreted here as the probability that an ancestral gene lineage was in a given class in a distant past [64] . From these , we can obtain the well-known expressions of effective size Ne for autosomal and X-linked genes: , respectively [45] . Note that Ne is expressed here as a number of gene copies ( i . e . , twice the effective number of diploid individuals for autosomes ) . Likewise , the effective migration rate , i . e . the average dispersal rate of an ancestral gene lineage , is given by for autosomal genes , and for X-linked genes , respectively . Substituting these expressions into the well-known equation: FST≈1/ ( 1+2Neme ) [64] , we get: ( 5 ) for autosomal genes , and ( 6 ) for X-linked genes . We performed coalescent simulations , using an algorithm in which coalescence and migration events are considered generation-by-generation until the common ancestor of the whole sample has been reached ( see [65] ) . We simulated a finite island model with 50 demes , each made of N = Nf+Nm = 500 diploid individuals , with a migration parameter m = mf+mm = 0 . 2 . Using these total values for N and m , we then varied the sex-specific parameters to cover the ( Nf/N , mf/m ) parameter space evenly . Note that the parameter m is the total migration rate , which corresponds to twice the effective migration rate for autosomal markers . Hence , for each set of ( Nf/N , mf/m ) values , the total number of individuals is 500 ( although the number of females may vary from 1 to 499 ) and the effective migration rate for autosomal markers is . We chose these total values for N and m such that , for a ratio Nfmf/Nmmm = 21 . 6 ( as observed for the herder populations ) , the distribution of FST estimates on uniparentally-inherited markers in the simulations were close to the observations: for mtDNA , the 95% highest posterior density interval ( see [66] , pp . 38–39 ) for the distribution of FST estimates in the simulations was [0 . 007; 0 . 033] with a mode at 0 . 014 ( estimated value from the real dataset: among the herders ) while for the NRY , the 95% highest posterior density interval was [0 . 088; 0 . 374] with a mode at 0 . 187 ( estimated value from the real dataset: ) . Each simulated sample consisted in 330 sampled males from 11 populations ( 30 males per population ) , genotyped at 27 autosomal , 9 X-linked markers as well as 10 Y-linked markers and a single mtDNA locus . Each locus was assumed to follow a Generalized Stepwise Model ( GSM ) [67] with a possible range of 40 contiguous allelic states , except the mtDNA , which was assumed to follow an infinite allele model of mutation . The average mutation rate was 5 . 10−3 , and the mean parameter of the geometric distribution of the mutation step lengths for microsatellites was set to 0 . 2 [67] , [68] . | Human evolutionary history has been investigated mainly through the prism of genetic variation of the Y chromosome and mitochondrial DNA . These two uniparentally inherited markers reflect the demographic history of males and females , respectively . Their contrasting patterns of genetic differentiation reveal that women are more mobile than men among populations , which might be due to specific marriage rules . However , these two markers provide only a limited understanding of the underlying demographic processes . To obtain an independent picture of sex-specific demography , we developed a new multi-locus approach based on the analysis of markers from the autosomal and X-chromosomal compartments . We applied our method to 21 human populations sampled in Central Asia , with contrasting social organizations and lifestyles . We found that , in patrilineal populations , not only the migration rate but also the number of reproductive individuals is likely to be higher for women . This result does not hold for bilineal populations , for which both the migration rate and the number of reproductive individuals can be equal for both sexes . The social organization of patrilineal populations is the likely cause of this pattern . This study suggests that differences in sex-specific migration rates may not be the only cause of contrasting male and female differentiation in humans , and that differences in effective numbers do matter . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/population",
"genetics",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics"
] | 2008 | Sex-Specific Genetic Structure and Social Organization in Central Asia: Insights from a Multi-Locus Study |
Fungal pathogens have evolved diverse strategies to sense host-relevant cues and coordinate cellular responses , which enable virulence and drug resistance . Defining circuitry controlling these traits opens new opportunities for chemical diversity in therapeutics , as the cognate inhibitors are rarely explored by conventional screening approaches . This has great potential to address the pressing need for new therapeutic strategies for invasive fungal infections , which have a staggering impact on human health . To explore this approach , we focused on a leading human fungal pathogen , Candida albicans , and screened 1 , 280 pharmacologically active compounds to identify those that potentiate the activity of echinocandins , which are front-line therapeutics that target fungal cell wall synthesis . We identified 19 compounds that enhance activity of the echinocandin caspofungin against an echinocandin-resistant clinical isolate , with the broad-spectrum chelator DTPA demonstrating the greatest synergistic activity . We found that DTPA increases susceptibility to echinocandins via chelation of magnesium . Whole genome sequencing of mutants resistant to the combination of DTPA and caspofungin identified mutations in the histidine kinase gene NIK1 that confer resistance to the combination . Functional analyses demonstrated that DTPA activates the mitogen-activated protein kinase Hog1 , and that NIK1 mutations block Hog1 activation in response to both caspofungin and DTPA . The combination has therapeutic relevance as DTPA enhanced the efficacy of caspofungin in a mouse model of echinocandin-resistant candidiasis . We found that DTPA not only reduces drug resistance but also modulates morphogenesis , a key virulence trait that is normally regulated by environmental cues . DTPA induced filamentation via depletion of zinc , in a manner that is contingent upon Ras1-PKA signaling , as well as the transcription factors Brg1 and Rob1 . Thus , we establish a new mechanism by which metal chelation modulates morphogenetic circuitry and echinocandin resistance , and illuminate a novel facet to metal homeostasis at the host-pathogen interface , with broad therapeutic potential .
Invasive fungal infections have a devastating impact on human health worldwide . The most vulnerable individuals are those suffering from immune deficiencies due to chemotherapy for cancer , immunosuppression for transplants of solid organs or stem cells , or infection with HIV [1] . The incidence of deadly invasive fungal infections is on the rise , in concert with the increasing use of immunosuppressive measures and invasive medical procedures [2 , 3] . Immunocompetent individuals are also at risk , especially those in the expanding adult-onset diabetic population . Approximately 1 . 5 million people die every year from invasive fungal infections , which exceeds the death toll of malaria or tuberculosis [1] . Candida species are a leading cause of mycotic death worldwide , and account for over 85% of all hospital acquired fungal infections [2] . Candida albicans is the primary cause of systemic candidiasis with mortality rates of ~40% [4 , 5] , even with current treatment options . There is a limited repertoire of antifungal drugs available to treat human fungal infections , with the utility of current drugs restricted by problems of host toxicity , fungistatic activity , or drug resistance . There are only three major antifungal drug classes for treatment of invasive infections , with the development of novel classes of antifungals having largely stalled since the 1990s [6] . The polyenes were discovered more than 50 years ago , and have fungicidal activity due to binding and extracting ergosterol from fungal cell membranes , with host toxicity resulting from collateral effects on cholesterol in human cell membranes [7] . The first azoles were developed in the 1970s [8] , and exert fungistatic activity by inhibiting the ergosterol biosynthetic enzyme lanosterol 14α-demethylase; they are the most widely deployed class of antifungal , but are vulnerable to drug resistance given their fungistatic activity against many fungal pathogens [9] . While newer generation azoles have been introduced into the clinic more recently , they remain vulnerable to cross-resistance across the azole class [10] . The echinocandins were first introduced into the clinic in the early 2000s , and impair fungal cell wall integrity by inhibiting biosynthesis of a structural polysaccharide , ( 1 , 3 ) -β-D-glucan [11] . Echinocandins remain a front-line therapy for invasive candidiasis , and thus the emergence of echinocandin resistance in Candida poses grave concern [12 , 13] . Echinocandin resistance is increasing in prevalence in the clinic . In C . albicans , resistance is often attributable to mutations in FKS1 , which encodes the drug target ( 1 , 3 ) -β-D-glucan synthase [14–16] . Resistance phenotypes can also be modulated by compensatory changes in the fungal cell wall , as with the reduced echinocandin susceptibility that accompanies elevated chitin production [17] . Both the basal tolerance of wild-type strains to echinocandins , as well as the echinocandin resistance of FKS1 mutants is contingent upon the capacity to sense and respond to drug-induced cell wall stress . Crucial cell wall stress responses include the protein kinase C ( PKC ) cell wall integrity cascade , the calcineurin signaling pathway , and the high osmolarity glycerol response ( HOG ) pathway , which converge on regulation of chitin synthesis in response to cell wall stress [18 , 19] . Signaling through these pathways is controlled by the molecular chaperone Hsp90 , which is required for stability and activation of Hog1 , calcineurin , and the terminal mitogen-activated protein kinase ( MAPK ) in the PKC signaling pathway , Mkc1 [20–22] . Perturbation of these stress response pathways causes hypersensitivity to the cell wall stress induced by echinocandins [12 , 20 , 23 , 24] , suggesting that small molecules that inhibit key stress response regulators hold great potential as combination therapy agents to enhance antifungal efficacy and reverse drug resistance [25 , 26] . The therapeutic utility of current inhibitors of these stress response regulators as antifungals is compromised by lack of fungal selectivity , with host toxicity or immunosuppressive effects resulting from inhibition of the human counterparts . Thus , there is an urgent need for the discovery of new molecules that can increase the efficacy of echinocandins and overcome resistance . The discovery of molecules that impair drug resistance or virulence traits generates new opportunities for chemical diversity in therapeutic agents , as they are not typically explored by conventional screening approaches for molecules that target essential processes [27 , 28] . Beyond expanding the repertoire of antifungal agents , targeting drug resistance regulators offers additional benefits such as minimizing effects on the host mycobiome and reducing selection pressure for the evolution of drug resistance . These same benefits apply to targeting virulence factors [29] . For C . albicans , a central virulence factor is the capacity to transition between yeast and filamentous growth [30 , 31] , with most mutants that are locked in either morphology being attenuated in virulence [32–35] . Filamentation is induced by environmental cues such as serum , alkaline pH , and elevated CO2 in a manner that is contingent upon elevated temperature [9] . Hsp90 is a key regulator of temperature-dependent morphogenesis , with profound effects on core morphogenetic circuitry including the Ras1-protein kinase A ( PKA ) signaling cascade [36 , 37] . As has recently come to light through systematic screens of chemical combinations [28] , the discovery of molecules that enhance the activity of existing antifungals and also modulate virulence traits offers great potential to bolster the antifungal pipeline . In this study , we screened a library of 1 , 280 pharmacologically active compounds to identify those that increase the efficacy of caspofungin against an echinocandin-resistant C . albicans clinical isolate . We identified 19 compounds that potentiate the activity of caspofungin with negligible antifungal activity on their own . We focused our analysis on the broad-spectrum chelator DTPA , which had the greatest synergistic activity . Reducing levels of magnesium alone was sufficient to impair growth in the presence of caspofungin , suggesting that DTPA potentiates echinocandin activity via magnesium chelation . To identify effectors through which DTPA modulates echinocandin activity , we selected for spontaneous mutants resistant to the combination of DTPA and caspofungin . Whole genome sequencing coupled with functional analyses revealed that mutations in the histidine kinase gene NIK1 confer resistance to the drug combination . While both caspofungin and DTPA activate the MAPK Hog1 , NIK1 mutations block activation in response to either compound , suggesting that DTPA enhances the efficacy of caspofungin by modulating Hog1 signaling through Nik1 . We found that DTPA not only synergizes with caspofungin , but also induces filamentation in the absence of elevated temperature or other inducing cues . DTPA induced filamentation via depletion of zinc . To identify the circuitry through which DTPA induces filamention we performed a genetic screen of 143 homozygous deletion mutants , which revealed that DTPA-induced filamentation is contingent upon Ras1-PKA signaling , as well as the transcription factors Brg1 and Rob1 . Together , our findings establish a new mechanism by which chelation of trace metals modulates fungal morphogenesis and cellular responses to drug-induced stress , and reveals a new way in which metal homeostasis impacts host-pathogen interactions .
We first aimed to identify compounds that enhance echinocandin activity against a C . albicans caspofungin-resistant clinical isolate . We used an isolate with an FKS1T1922C homozygous mutation , resulting in the amino acid substitution of Fks1F641S , thus requiring a high concentration of caspofungin to achieve Fks1 inhibition [38] . We screened the LOPAC1280 Navigator Library . This was similar in approach to our previous screen to identify molecules that potentiate the activity of azoles , which revealed a new role for the PKC signaling pathway in enabling responses to cell membrane stress [21] . Similarly , our expectation was that a screen to identify compounds that potentiate caspofungin activity would reveal those that cripple stress response pathways crucial for responding to cell wall stress . We initially screened the library at 25 μM in RPMI medium at 30°C , in the presence or absence of 8 μg/mL caspofungin . This concentration of caspofungin inhibited the growth of the resistant clinical isolate by approximately 60% . This initial screen identified 13 compounds that were classified as strong hits based on two criteria: 1 ) they reduced growth in the presence of caspofungin by at least 90% compared to growth in the LOPAC compound alone; and 2 ) inhibited growth by less than 70% in the absence of caspofungin ( Fig 1A ) . Strong hits included compounds that target key regulators of echinocandin resistance phenotypes , such as inhibitors of calcineurin and PKC signaling [19 , 20 , 23 , 24] , validating our screen . The initial screen also identified nine compounds whose targets have not been previously implicated in echinocandin resistance phenotypes . We performed additional analyses to increase our capacity to identify molecules that modulate echinocandin resistance phenotypes , and to validate lead compounds from the initial screen . For compounds from our initial screen that were classified as weak hits based on inhibition of growth by 80–90% in combination with caspofungin compared to growth with the LOPAC compound alone , we re-screened them at twice the concentration ( 50 μM ) ; this identified three additional molecules that potentiate echinocandin antifungal activity ( Fig 1B ) . For compounds from our initial screen that inhibited growth in the absence of echinocandin by greater than 70% , we re-screened them at half the concentration ( 12 . 5 μM ) ; this identified another three compounds that enhance echinocandin antifungal activity ( Fig 1C ) . Finally , to validate all of the lead compounds from the screen , we performed minimum inhibitory concentration ( MIC ) assays in the absence or presence of 8 μg/mL caspofungin to quantify the magnitude of impact on echinocandin resistance of the clinical isolate ( S1 Fig ) . The four compounds that best potentiate caspofungin activity against the clinical isolate and whose roles in echinocandin susceptibility have not been previously described , are shown in Fig 1D . The greatest activity was observed with diethylenetriamine pentaacetic acid ( DTPA ) , a chelator of di- and trivalent cations [39] . Our discovery of a broad-spectrum chelator as the strongest potentiator of echinocandin activity suggests that targeting metal homeostasis may provide a powerful therapeutic strategy . There is precedent for the use of chelators in the treatment of fungal infections , as with ciclopirox ethanolamine for treatment of superficial mycoses [40] . The broad-spectrum antifungal activity of ciclopirox ethanolamine is thought to be due to chelation of iron [41] . To determine if the capacity to enhance echinocandin activity is a general property of chelators we performed dose response matrices ( or checkerboard assays ) with DTPA or ciclopirox ethanolamine and caspofungin against the echinocandin-resistant C . albicans clinical isolate . A fractional inhibitory concentration index ( FICI ) is an expression of the combinatorial effect of two compounds . A value ≤0 . 5 indicates synergy , while a value >0 . 5–4 . 0 indicates indifference [42] . DTPA and caspofungin were synergistic ( FICI = 0 . 375 ) , while ciclopirox ethanolamine had no impact on caspofungin resistance ( FICI = 2 . 0 ) ( Fig 2A ) . This suggests that DTPA has a distinct mode of action from other chelators in clinical use as antifungals , and that DTPA modulates echinocandin resistance phenotypes via chelation of metal cations distinct from iron . To determine the identity of the metal cation ( s ) through which DTPA modulates echinocandin resistance , we used an ion exchange resin ( Chelex 100 resin ) to deplete synthetic defined medium of its metal components . We then selectively restored metals based on the concentrations at which they are normally present in the medium , and assessed the growth of the echinocandin-resistant clinical isolate in the presence or absence of 2 μg/mL caspofungin ( Fig 2B ) . In medium containing all metal components ( Complete ) , this concentration of caspofungin did not inhibit growth . Depletion of all metals completely blocked growth in the presence of caspofungin; importantly , robust growth was observed in the absence of caspofungin , indicating that sufficient trace metals remain in the metal-depleted medium to support growth . Addition of all metals to the metal-depleted medium restored growth in the presence of caspofungin . Depleting each metal individually ( creating ‘drop out’ media by addition of metals to the metal-depleted medium ) revealed that depletion of magnesium is sufficient to impair growth of the echinocandin resistant clinical isolate in the presence of caspofungin ( P <0 . 001 , two-way ANOVA , Bonferroni correction ) ( Fig 2B ) . Conversely , addition of each metal individually to the metal-depleted medium revealed that magnesium best restores echinocandin resistance of the clinical isolate ( P <0 . 001 , two-way ANOVA , Bonferroni correction ) ( Fig 2C ) . We performed a comparable analysis with a wild-type strain of C . albicans , and confirmed that addition of magnesium to metal-depleted medium best restores growth in caspofungin ( P <0 . 001 , two-way ANOVA , Bonferroni correction ) ( S2 Fig ) . These results suggest that DTPA potentiates echinocandin antifungal activity primarily through the depletion of magnesium , and demonstrate that modulation of metal availability provides a powerful strategy to enhance antifungal activity . Because the efficacy of chelators such as DTPA does not require intracellular accumulation , this approach will evade efflux-mediated resistance . As an unbiased approach to further investigate the mechanism by which DTPA exerts synergistic activity with caspofungin , we selected for mutants resistant to the combination . We plated 2x108 cells of a standard laboratory strain and the caspofungin-resistant clinical isolate on medium containing a high concentration of DTPA ( 100 μM ) and caspofungin ( 0 . 25 μg/mL for the laboratory strain and 8 μg/mL for the resistant isolate ) . We quantified resistance to DTPA , caspofungin , and the combination using dose response matrices ( checkerboards ) . We recovered two mutants in the laboratory strain background , which had increased resistance to both DTPA and caspofungin individually , as well as in combination ( Fig 3A ) . We also recovered two mutants in the echinocandin-resistant clinical isolate background , which both had increased resistance to DTPA alone and in combination with caspofungin ( Fig 3B ) . To identify mutations that confer resistance to the combination of DTPA and caspofungin , we performed whole genome sequencing on the two mutants from each background as well as on the two parental strains . Sequence reads were aligned to the published C . albicans genome ( SC5314 , assembly 21 , with the mean depth of coverage being 106X for all assembled sequences ) . We identified only one or two non-synonymous single nucleotide variants within coding regions in the mutants relative to their parents ( Fig 3A and 3B ) . All of these mutations were heterozygous . Strikingly , three of the four sequenced mutants harbored mutations in the NIK1 gene , which encodes one of three C . albicans histidine kinases involved in the two-component phosphorelay system that regulates the Hog1 MAPK stress response pathway [43–46] . Two of the identified NIK1 mutations result in substitutions in the histidine kinase HAMP linker domain ( G73D and R440T ) , and the third Nik1 substitution is found within the ATP-binding domain ( G683V ) . The HAMP domains of CaNik1 have been found previously to mediate sensitivity to various fungicides [47] . The fourth mutant harbored a mutation in YKE2 , which encodes a component of the prefoldin complex in Saccharomyces cerevisiae [48 , 49] . Next , we performed allele replacements to functionally validate the mutations identified by whole genome sequencing of the mutants in the laboratory strain background . We replaced one native allele of NIK1 or YKE2 in the parental strain with the corresponding allele from the resistant mutants . Both the YKE2T95A and NIK1G2048T alleles conferred resistance to DTPA and caspofungin in the laboratory strain background ( S3A Fig ) . Because the phenotypic effects were observed in heterozygous mutants , the mutations likely do not cause loss of function . Consistent with this hypothesis , deletion of NIK1 did not confer resistance to either DTPA or caspofungin ( S3B Fig ) . Finally , we assessed resistance of one of the NIK1 point mutants to the combination of metal depletion and caspofungin . While the echinocandin-resistant clinical isolate ( FKS1T1922C/FKS1T1922C ) is unable to grow with the combination of metal depletion and caspofungin ( Figs 2B and 3C ) , a heterozygous NIK1G1319C mutation ( resulting in the amino acid substitution Nik1R440T ) restores this ability ( P <0 . 001 , two-way ANOVA , Bonferroni correction ) ( Fig 3C ) . Addition of magnesium also restores the ability of either strain to grow in the presence of caspofungin . Thus , specific mutations in NIK1 block the effects of metal depletion on echinocandin resistance . Nik1 is one of three C . albicans histidine kinases , along with Chk1 and Sln1 [45 , 46] . In S . cerevisiae , the sole histidine kinase Sln1 is the osmosensor that regulates the Hog1 MAP kinase cascade [50 , 51] . In C . albicans , the histidine kinases are also thought to signal upstream of Hog1 , although the exact role for Nik1 remains unknown as does the identity of the osmosensor ( reviewed in [46] and [52] ) . Deletion of HOG1 in C . albicans confers resistance to cell wall stressors such as Calcofluor White and Congo Red [53–55] , suggesting that the NIK1 point mutations may modulate resistance via Hog1 . To test this , we assayed resistance of a hog1Δ/hog1Δ homozygous deletion mutant to both caspofungin and DTPA . We found that the hog1Δ/hog1Δ mutant had increased resistance to the combination of DTPA and caspofungin compared to its parental strain , as does the NIK1/NIK1G2048T mutant ( Fig 4A ) . Furthermore , deletion of other components of the Hog1 signaling pathway such as the genes encoding Ssk1 , Ssk2 , or Pbs2 had comparable effects on resistance to DTPA and caspofungin as did deletion of Hog1 ( Fig 4B ) . This suggests that the NIK1G2048T allele may confer resistance via effects on Hog1 . Consistent with this model , introducing the NIK1G2048T allele into the hog1Δ/hog1Δ mutant did not further increase resistance to either DTPA or caspofungin ( Fig 4A ) . The finding that point mutations in NIK1 phenocopy impaired signaling through the Hog1 pathway suggests that these mutations may confer resistance to DTPA and caspofungin by blocking Hog1 activation . To determine if the NIK1G2048T mutation impairs Hog1 function , we first monitored phosphorylation of Hog1 in response to both caspofungin and the oxidative stress-inducing agent hydrogen peroxide in the laboratory strain . Hog1 is known to be activated in response to both hydrogen peroxide and caspofungin [56 , 57] . While both stressors cause an increase in Hog1 phosphorylation in the wild-type strain , only hydrogen peroxide induced Hog1 phosphorylation in the NIK1/NIK1G2048T mutant ( Fig 5A ) . This suggests that the NIK1 mutation blocks signaling of caspofungin-induced stress through the Hog1 pathway . Next , given that deletion of HOG1 confers resistance to DTPA , we tested whether DTPA activates Hog1 . All concentrations of DTPA tested induced Hog1 phosphorylation in the wild-type strain , but this activation was impaired in the NIK1 G2048T mutant , such that a higher concentration of DTPA was required to achieve activation ( Fig 5B ) . Notably , the ability of DTPA to activate Hog1 was phenocopied by depletion of magnesium ( Fig 5C ) , providing additional evidence that DTPA affects Hog1 signaling via chelation of magnesium . Further , phosphorylation of Hog1 in response to DTPA was maintained in a nik1Δ/nik1Δ homozygous deletion mutant indicating that G2048T is not a NIK1 loss-of-function mutation ( Fig 5D ) . Thus , the G2048T NIK1 mutation impairs activation of Hog1 in response to specific cellular stressors . We also monitored Hog1 activation in the echinocandin-resistant isolate ( FKS1T1922C/ FKS1T1922C ) . Although 2 μg/mL caspofungin activated Hog1 in this parental background , activation of Hog1 was blocked in the NIK1G1319C mutant that is resistant to the combination of DTPA and caspofungin ( Fig 5E ) . As with the laboratory strain , the NIK1 mutation did not alter activation of Hog1 in response to hydrogen peroxide , but did impair Hog1 activation in response to DTPA in the echinocandin-resistant isolate . Concentrations of DTPA up to 20 μM resulted in Hog1 activation in the parent , but not in the NIK1G1319C mutant ( Fig 5F ) ; Hog1 remains capable of activation in response to higher concentrations of DTPA in the N1K1 mutant ( Fig 5G ) . The NIK1 alleles in the wild-type and FKS1T1922C/ FKS1T1922C backgrounds are different , and thus may have distinct magnitudes of effect on Hog1 activation in response to DTPA . Together , these results demonstrate that both DTPA and caspofungin activate Hog1 , and that mutations in NIK1 impair this activation in response to specific cues . Given that metal cations such as magnesium are known to bind to cell wall components [58] , we the explored the possibility that DTPA might non-specifically modulate other stress response pathways or act extracellularly on the cell wall , thereby potentiating caspofungin activity . We first assessed whether DTPA impairs signaling through the PKC cell wall integrity pathway . In S . cerevisiae , exposure to caspofungin induces phosphorylation of the MAPK Slt2 [59] . We monitored phosphorylation of the C . albicans ortholog , Mkc1 , by Western blot . As expected , Mkc1 was phosphorylated in response to caspofungin ( S4A Fig ) ; DTPA did not induce activation of Mkc1 nor did it block activation in response to caspofungin . Additionally , we examined the effect of DTPA on cell wall architecture . Under conditions of cell wall stress , such as exposure to cell surface-perturbing agents , the levels of chitin and glucan in the cell wall increase dramatically [60] . We monitored cell wall chitin and glucan levels in the caspofungin-resistant isolate that was grown in RPMI and treated overnight with 50 μM DTPA or 0 . 32 μg/mL of caspofungin . Cells were stained with Aniline Blue to measure levels of exposed glucans , with Calcofluor White to measure chitin levels , or with Concanavalin A to measure levels of mannans . While treatment with the cell wall-targeting drug caspofungin drastically increased the levels of exposed glucan and chitin , treatment with DTPA had minimal effect ( S4B Fig ) . Thus , DTPA activates Hog1 signaling but does not promiscuously activate the cell wall integrity pathway or cell wall remodelling . As DTPA has profound effects on C . albicans drug resistance , we next assessed whether it also impacts another key C . albicans virulence trait: the capacity to transition between yeast and filamentous growth . We found that DTPA induced robust filamentation in wild-type cells ( Fig 6A ) . Strikingly , this yeast-to-filament transition occurs even at 30°C and does not require a concurrent shift to 37°C , as do most other filament-inducing cues . In order to identify the metal through which DTPA induces filamentation , we used the Chelex 100 ion exchange resin to deplete synthetic defined medium of its metal components . As expected , the cells grew in the yeast form in complete medium; depletion of all metals induced filamentous growth , and addition of all metal components back to the metal-depleted medium restored growth in the yeast form ( Fig 6B and S5 Fig ) . By omitting each metal individually from the metal supplement , we found that zinc-depletion induced robust filamentous growth . Conversely , addition of only zinc to metal-depleted medium repressed filamentation ( Fig 6B and S5 Fig ) . Similarly , addition of excess zinc to DTPA-treated cells inhibited filamentation in response to DTPA ( Fig 6B and S5 Fig ) . We used a lower concentration of DTPA ( 50 μM ) that still induced filamentous growth , while minimizing the amount of metal and chelator present in the culture , making the add-back feasible . Depleting the medium of the other metals individually had minimal to no effect on cellular morphology , and only depletion of all metals together or depletion of zinc alone significantly increased filamentation compared to complete medium ( S6 Fig ) ( P<0 . 0001 , one-way ANOVA , Dunnett’s test ) . We further validated that depletion of zinc induces filamentous growth in two ways . First , we found that the zinc chelator TPEN induced filamentation ( Fig 6A ) . Second , we demonstrated that doxycycline-mediated transcriptional repression of either of two zinc transporters , ZRT1 and ZRT2 [61–63] , conferred hypersensitivity to the filament-inducing effects of DTPA ( Fig 6C ) . We used a low concentration of DTPA ( 50 μM ) to provide a mild filament-inducing cue in YPD in order to enable the identification of a hyperfilamentatous phenotype . We quantified this effect by measuring the expression of HWP1 , which encodes hyphal cell wall protein 1 , by quantitative RT-PCR . Upon treatment with DOX and DTPA , HWP1 expression was increased in the ZRT1 and ZRT2 depletion strains compared to the wild type ( P<0 . 0001 , two-way ANOVA , Bonferroni correction ) . Treatment with DOX and DTPA also causes induction of ZRT1 in the ZRT2 depletion strain , and induction of ZRT2 in the ZRT1 depletion strain ( P<0 . 0001 , two-way ANOVA , Bonferroni correction ) , consistent with a transcriptional response to metal deficiency . Thus , depletion of zinc in the environment or impairment of zinc import induces C . albicans morphogenesis . Morphogenesis in response to different cues is controlled by distinct cellular circuitry . To elucidate the circuitry required for DTPA-induced filamentation , we screened a library of 143 C . albicans homozygous deletion mutants of transcription factor genes [64] . Cells were grown in rich medium in static conditions at 30°C in the absence or presence of 50 μM DTPA , and morphology was assessed by microscopy . We identified three mutants that were defective in filamentation in response to DTPA: brg1Δ/brg1Δ , rob1Δ/rob1Δ , and efg1Δ/efg1Δ ( Fig 7A ) . We also confirmed that these mutants are unable to filament in response to zinc depletion ( Fig 7B ) . These three transcription factors are key components of a network that regulates the formation of biofilms [65] , which are complex communities composed of multiple cellular morphologies that form upon adherence to surfaces [66] . Although Efg1 is a master regulator of morphogenesis and required for filamentation in response to most filament-inducing cues [67] , Brg1 and Rob1 have more specialized roles in morphogenesis . For example , although brg1Δ/brg1Δ and rob1Δ/rob1Δ mutants are defective in biofilm formation , they are capable of filamentous growth in other contexts [65] . We confirmed that the brg1Δ/brg1Δ and rob1Δ/rob1Δ mutants filament in response to diverse cues ( S7 Fig ) , suggesting that there is specificity to the role of Brg1 and Rob1 in morphogenesis induced by DTPA and zinc depletion . To further delineate functional relationships between Brg1 and Rob1 , we monitored BRG1 and ROB1 transcript levels in response to DTPA using qRT-PCR . We found that BRG1 expression is induced in response to DTPA ( P <0 . 0001 , unpaired t-test ) , while ROB1 expression is not ( Fig 7C ) . Under biofilm conditions , Rob1 is required for expression of BRG1 [65] . To determine if Rob1 also regulates BRG1 expression in response to DTPA , we examined BRG1 expression in the rob1Δ/rob1Δ mutants . Indeed , deletion of ROB1 blocked induction of BRG1 in response to DTPA ( P <0 . 001 , one-way ANOVA , Bonferroni correction ) ( Fig 7D ) . This reinforces the functional relationship between Rob1 and Brg1 , and illuminates a new role for these regulators in morphogenesis in response to perturbation of metal homeostasis . A molecular feature of filaments induced by diverse cues is reduction in the levels of the transcriptional repressor Nrg1 [34 , 68 , 69] . Removal of Nrg1 from the promoters of filament-specific genes is thought to be required for initiation and maintenance of filamentous growth [70] . BRG1 is one of the many genes repressed by Nrg1 [71] . To determine if Nrg1 degradation is also associated with DTPA-induced filaments , we used Western blot analysis to monitor levels of epitope tagged Nrg1 in cells exposed to diverse filament-inducing cues . We found that Nrg1 levels were reduced in cells exposed to each of the filament-inducing cues tested , including DTPA ( Fig 8A ) . Next we assessed whether filamentation in response to DTPA was contingent upon circuitry controlling Nrg1 levels . Nrg1 is regulated at both the transcriptional and protein level in filament-inducing conditions [69] . Upon initiation of filamentation at 37°C , NRG1 transcription is downregulated by the Ras1-PKA pathway [70 , 72] . Components of this pathway include the adenylyl cyclase Cyr1 , which synthesizes cAMP and is activated by Ras1 , and the PKA complex , which consists of two catalytic subunits , Tpk1 and Tpk2 ( reviewed in [9] ) . We assessed morphology of mutants lacking components of the Ras1-PKA pathway in response to DTPA . We found that homozygous deletion of RAS1 or CYR1 , or depletion of TPK1 in a strain lacking TPK2 blocked filamentation in response to DTPA ( Fig 8B ) . This assay utilizes 200 μM DTPA instead of 100 μM , in order to induce filamentation in the CAI4 background . Thus , Ras1-PKA signaling is required for filamentation in response to DTPA , consistent with the central role of this cascade in morphogenesis in response to diverse cues [67] . There is additional regulatory complexity to achieve rapid loss of Nrg1 upon the initiation of filamentation . The induction of filamentation at 37°C is accompanied by degradation of Nrg1 [70 , 72] , which is controlled by a pathway that consists of the E3 ubiquitin ligase Ubr1 , the transcriptional repressor Cup9 , and the kinase Sok1 [72] . Ubr1 mediates degradation of Cup9 , thereby alleviating transcriptional repression of SOK1 and enabling Nrg1 degradation upon the initiation of filamentation [72] . To determine whether this pathway is required for filamentation induced by DTPA , as is the case with filamentation at 37°C , we monitored morphology of mutants of the positive regulators Ubr1 and Sok1 . We found that Tn-sok1/Tn-sok1 transposon insertion mutants were unable to filament in response to DTPA , however , the ubr1Δ/ubr1Δ homozygous deletion mutants filamented robustly ( Fig 8C and S8 Fig ) . Nrg1 protein levels decreased in response to filamentation induced by DTPA , even in the absence of Ubr1 ( Fig 9 ) . This indicates that Ubr1 is dispensable for Nrg1 degradation in response to DTPA , and suggests that there is functional divergence in the regulation of Nrg1 degradation in response to distinct cues . Given that DTPA potentiates the activity of caspofungin in vitro and also modulates a key virulence trait , we assessed whether there might be therapeutic benefits in a murine model of disseminated candidiasis . Despite the pleiotropic effects of chelators , there is precedent for therapeutic benefits with systemic administration of the broad-spectrum chelator EDTA . EDTA has been shown to enhance the efficacy of the polyene antifungal amphotericin B lipid complex in the treatment of invasive pulmonary aspergillosis in a rat model , with no observed toxicity [73] . Therefore , we tested if a similar dose of DTPA would enhance the therapeutic efficacy of caspofungin in a murine model of systemic infection with a caspofungin-resistant clinical isolate of C . albicans ( Fig 10 ) . Most mice that were infected with 5 x 105 C . albicans cells succumbed to the infection within 5 days . Treatment with 0 . 05 mg/kg caspofungin caused a modest improvement in survival ( P = 0 . 1939 , Log-Rank Mantel-Cox Test ) , while treatment with 60 mg/kg DTPA alone did not improve survival . Strikingly , treatment with the combination of caspofungin and DTPA dramatically improved survival compared to treatment with caspofungin alone ( P = 0 . 0296 , Log-Rank Mantel-Cox Test ) . This demonstrates the feasibility of using chelators to enhance echinocandin efficacy in vivo in the treatment of echinocandin-resistant infections .
Our results reveal a new mechanism by which metal chelation potentiates echinocandin antifungal activity and modulates a key fungal virulence trait in a leading human fungal pathogen . Based on a screen of 1 , 280 pharmacologically active compounds , we identified the broad-spectrum chelator DTPA as strongly synergistic with caspofungin against a resistant C . albicans clinical isolate ( Fig 1 ) , and that DTPA enhances echinocandin activity against echinocandin-susceptible and echinocandin-resistant isolates predominantly by chelating magnesium ( Fig 2 and S2 Fig ) . By leveraging a powerful genomic approach to select for mutants resistant to the combination of DTPA and caspofungin coupled with genome sequencing we revealed that mutations in NIK1 confer resistance to the drug combination ( Fig 3 ) . This provides key molecular insight into the mechanism by which DTPA potentiates echinocandin activity , and reveals mutations in a cellular regulator that has not been previously implicated in echinocandin resistance . DTPA activates Hog1 , as does caspofungin , and mutations in NIK1 block Hog1 activation in response to both compounds ( Fig 5 ) . Inhibition of ( 1 , 3 ) -β-D-glucan synthesis and activation of Hog1 may provide parallel cell wall insults , such that the combined effect is synergistic . Beyond effects on drug resistance , we discovered that DTPA induces filamentation at 30°C in rich medium in the absence of any inducing cue , and that the effect on morphogenesis is mediated through chelation of zinc ( Fig 6 ) . Filamentation induced by DTPA is contingent upon core morphogenetic circuitry including the Ras1-PKA cascade , as well as more specialized morphogenetic regulators , including the transcription factors Rob1 and Brg1 ( Figs 7 and 8 ) . This work establishes metal chelation as a powerful strategy to probe circuitry governing cellular responses to drug-induced stress and morphogenesis , and illuminates a new facet to metal homeostasis in modulating host-pathogen interactions . Metals have complex and crucial roles in all biological systems , where they are incorporated into metalloproteins including enzymes , storage proteins , and transcription factors . They are crucial for survival , but also exert toxicity that is attributable to their capacity to engage in enzyme active sites and potentiate catalytic activity , thus their levels must be tightly controlled [74 , 75] . As a consequence , mammals exploit strategies to modulate metal availability in order to thwart pathogenic microbes [75] . As both a commensal and opportunistic pathogen , C . albicans has evolved diverse strategies to ensure metal homeostasis . This is best understood in the context of iron , which is the most abundant transition metal in the human body , and yet its limitation provides a ubiquitous strategy for innate immune defense [75] . In the iron-poor bloodstream , C . albicans activates iron uptake genes via the transcriptional activator Sef1 , which is required for virulence; conversely , in the iron-replete gastrointestinal tract , the transcriptional repressor Sfu1 represses Sef1 at both transcriptional and post-transcriptional levels , thereby dampening expression of iron uptake genes [76 , 77] . C . albicans also exploits elaborate systems to sequester iron from the host including the transport of xeno-siderophores via Sit1 , and a reductive system to acquire iron from host transferrin proteins or hemoglobin [78 , 79] . Beyond iron , zinc also has profound impacts on C . albicans biology and pathogenesis , as well as key roles in host immunity [62 , 80] . C . albicans has evolved a zinc acquisition strategy involving secretion of the zinc scavenger Pra1 , analogous to siderophore-mediated iron acquisition [62] . Our findings that the broad-spectrum chelator DTPA synergizes with caspofungin and induces filamentation , suggest that bioavailability of metals such as magnesium and zinc in distinct niches within the host may also have a profound impact on fungal morphogenesis and drug resistance . The pleiotropic effect of metal depletion provides a powerful strategy to probe cellular pathways governing virulence traits and drug resistance . In the context of drug resistance , we illuminate specific circuitry through which DTPA potentiates echinocandin antifungal activity . To date , the principal mechanism of resistance to echinocandins is mutations in the target gene , FKS1 [14] . We find that either DTPA or magnesium depletion is sufficient to reduce target-based echinocandin resistance . Notably , the concentration of magnesium present in synthetic defined medium is in excess of the concentration of DTPA required to potentiate caspofungin . However , metal cations such as magnesium bind to fungal cell wall components including glucans and mannoproteins [58] , limiting the available pool of magnesium that DTPA would need to chelate to potentiate echinocandin activity . Notably , the effects of DTPA on echinocandin activity are blocked by dominant point mutations in the histidine kinase gene NIK1 . Nik1 likely signals upstream of the Hog1 MAPK cascade , as does the Sln1 histidine kinase in S . cerevisiae [50 , 51] . Nik1 is not required for hydrogen peroxide-induced activation of Hog1 , nor does deletion of NIK1 result in constitutive Hog1 activation [81] , thus Nik1 is not likely to be the osmosensor or the key histidine kinase responsible for relaying this signal . We establish that mutations in NIK1 confer increased resistance to caspofungin and DTPA , and impair Hog1 activation in response to these cues ( Fig 5 ) . This is consistent with previous findings that deletion of specific NIK1 HAMP domains impairs Hog1 phosphorylation in response to the fungicide fludioxonil [47] . Nik1 may act as a sensor for specific stresses , such as magnesium depletion . Other histidine kinases , such as the S . cerevisiae Sln1 , have an essential Mg2+ ion bound at the active site [82] and the C . albicans Nik1 may similarly require Mg2+ . The mutations in NIK1 may allow for the ion to bind more tightly to the enzyme , or they may induce a conformational change that reduces the necessity of Mg2+ . Our results support a model in which NIK1 mutations reduce the dependence of Nik1 on magnesium , thereby blocking the effects of DTPA on echinocandin resistance . Interestingly , deletion of NIK1 does not confer hypersensitivity to either DTPA or caspofungin ( S3 Fig ) , suggesting that while dominant mutations in NIK1 are sufficient to confer resistance , DTPA must have additional targets through which it mediates toxicity . This is consistent with previous reports that while Nik1 may have a role in C . albicans virulence , deletion of NIK1 does not confer sensitivity to various antifungals [45 , 83] . The resistance phenotypes of the NIK1 mutants can be explained by the observation that impaired Hog1 signaling leads to resistance to cell wall stressors [53–55] , ( Fig 4 ) . This is likely due to de-repression of Cek1 kinase activation [54] , where Cek1 has roles in cell wall homeostasis [84] . Our findings suggests that DTPA exacerbates the cell wall stress induced by echinocandins by causing further cell wall damage through effects on Hog1 signaling ( Fig 11 ) . This finding resonates with previous work illustrating that a natural product that modulates HOG signaling in A . fumigatus also potentiates caspofungin activity [85] . Together , this reveals new functional relationships governing cell wall stress response that can be targeted to enhance antifungal activity . Beyond drug resistance , metal chelators also provide a powerful strategy to dissect cellular circuitry governing key virulence traits , such as the transition between yeast and filamentous growth . A striking feature of filamentation induced by DTPA is that it is not contingent upon an increased temperature of 37°C , in contrast to most other cues . The few other cues that are independent of elevated temperature include inhibition of the molecular chaperone Hsp90 [36] and perturbation of cell cycle regulation [86–88]; filamentation induced by these cues is independent of the key transcriptional regulator of morphogenesis , Efg1 [36 , 86 , 87] . In contrast , filamentation induced by DTPA is contingent upon Efg1 ( Fig 7A ) , suggesting that metal chelation does not induce morphogenesis by perturbing Hsp90 function or cell cycle progression . Filamentation induced by DTPA is dependent not only on Efg1 , but also on upstream PKA signaling components including Ras1 , Cyr1 , and the PKA catalytic subunits ( Fig 8B ) . Ras1-PKA signaling is also required to down regulate expression of the transcriptional repressor gene NRG1 upon initiation of filamentous growth [70 , 72] , and we observe a pronounced reduction of Nrg1 levels in filaments induced by diverse cues including DTPA ( Fig 8A ) . Depletion of Nrg1 may also be due to post-translation control , as the E3 ubiquitin ligase Ubr1 regulates degradation of the transcriptional repressor Cup9 , thereby enabling transcriptional activation of the kinase gene SOK1 and Nrg1 degradation upon release from the quorum sensing molecule farnesol at 37°C [72] . We find that Sok1 is required for filamentation induced by DTPA but Ubr1 is dispensable ( Fig 8C ) , as Nrg1 protein levels are reduced in DTPA-induced filaments even in the absence of Ubr1 ( Fig 9 ) . This suggests that DTPA may impair Cup9 function , thereby inducing Nrg1 degradation and bypassing the requirement for Ubr1 . Thus , our results highlight functional divergence in the circuitry governing post-translational regulation of Nrg1 in response to different filament-inducing cues . There is additional regulatory complexity in the circuitry through which metal chelation induces morphogenesis . We established that DTPA induces filamentation via depletion of zinc , and that this requires not only core morphogenetic signaling through the core morphogenetic Ras1-PKA cascade and through Nrg1 , but also requires the biofilm transcriptional regulators , Brg1 and Rob1 ( Fig 7 ) . This functional relationship resonates with additional connections between zinc homeostasis and biofilm maturation . For example , Zap1 is a zinc-responsive transcription factor that regulates expression of several genes involved in biofilm matrix production [89] . Deletion of ZAP1 causes filamentation defects [61] and an increase in the proportion of yeast cells in biofilms [89 , 90] . Further , overexpression of the zinc transporter ZRT2 in strains lacking Zap1 restores expression of a yeast-specific reporter to wild-type levels [90] . Our finding that DTPA induces filamentation via depletion of zinc in a manner that depends upon Rob1 , Brg1 , Ras1-PKA signaling , and Nrg1 degradation expands our appreciation of the mechanisms through which metals modulate morphogenesis ( Figs 7–9 ) . Consistent with our results , exogenous zinc has been observed to suppress C . albicans filamentation [91 , 92] , though the mechanisms involved have remained enigmatic . Our results suggest that zinc modulates complex cellular circuitry governing a key fungal virulence trait . Perturbing metal homeostasis has broad therapeutic potential for the treatment of infectious disease . Iron chelators such as ciclopirox olamine provide effective therapy for superficial mycoses [40 , 41] , and have evaded resistance despite decades of clinical use [41] . Broad spectrum chelators have also shown promise as therapeutic agents in animal models . EDTA is structurally related to DTPA , and has therapeutic benefit in a mouse model of Pseudomonas aeruginosa induced pneumonia [93] . EDTA also enhances the efficacy of polyenes in a rat model of invasive pulmonary aspergillosis , with no observed toxicity [73] . Similarly , we observe a drastic increase in the efficacy of caspofungin in combination with DTPA in a murine model of caspofungin-resistant disseminated candidiasis , demonstrating the potential utility of metal chelation as a therapeutic strategy . However , exploiting the therapeutic potential of metal limitation will likely require strategies to enhance selectivity for the pathogen and thereby minimize deleterious effects on the host . This could be achieved by targeting the pathogen’s metal import or homeostasis systems . With strategically designed high-throughput screens for molecules that target metal homeostasis [94] , there is considerable promise for discovering novel classes of antimicrobial agents . It is exquisitely clear that metal homeostasis has profound impacts on susceptibility to infectious disease , as patients with iron overload conditions due to frequent transfusions are vulnerable to infection , as are patients with inherited or acquired iron storage disorders such as those with haemochromatosis [75] . Although haemochromatosis is associated with increased susceptibility to infection with enteric Gram-negative pathogens including Vibrio vulnificus and Yersinia enterocolitica , they also have iron-deficient macrophages and consequently are more resistant to intracellular pathogens such as Salmonella enterica subsp . enterica serovar Typhi and Mycobacterium tuberculosis [95–98] . There is a growing appreciation that other metals also have a tremendous impact on microbial virulence and host immunity . Defining the mechanisms that govern metal homeostasis in hosts and pathogens , and the circuitry through which metals control key virulence traits holds great promise for revealing new therapeutic strategies for life-threatening infectious disease .
All procedures for animal research were approved by the Institutional Animal Care and Use Committee ( IACUC protocol A114-14-05 ) at Duke University according to the guidelines of the Animal Welfare Act , The Institute of Laboratory Animal Resources Guide for the Care and Use of Laboratory Animals , and Public Health Service Policy . Archives of C . albicans strains were maintained at −80°C in rich medium ( YPD: 1% yeast extract , 2% bactopeptone , 2% glucose , with 2% agar for solid medium ) with 25% glycerol . Strains were grown in YPD , synthetic defined ( SD ) medium or RPMI medium . SD was prepared as follows: 0 . 17% yeast nitrogen base without ammonium sulfate , 0 . 1% glutamic acid , 2% glucose , supplemented with arginine HCl ( 50 mg/L ) and histidine HCl ( 20 mg/L ) as required , pH 6 . RPMI was prepared as follows: 10 . 4 g/L RPMI-1640 , 3 . 5% MOPS , 2% glucose , supplemented with an additional 5 mg/mL histidine as required , pH 7 . Spider Medium and Lee’s Medium were prepared as previously described [99 , 100] . Chelex 100 resin ( BioRad ) was used to deplete SD medium of its metal components . For the growth assay , 2 . 5 g of resin was added per 100 mL of medium ( except for Fig 3C , where 2 . 0 g was used ) , and for filamentation assays , 2 . 0 g of resin per 100 mL . The medium with the resin stirred for 1 . 5 hours at room temperature and was filtered sterilized . Metals were added back to the following concentrations , as indicated: CuSO4 80 μg/L; FeCl3 400 μg/L; MnSO4 800 μg/L; ZnSO4 800 μg/L; MgSO4 1 g/L; CaCl2 0 . 2 g/L . All strains used in this study are listed in S1 Table . All oligonucleotide sequences used in this study are included in S2 Table . To select for nourseothricin ( NAT ) -resistant mutants , NAT ( Jena Bioscience ) stock solution was prepared in water at a concentration of 250 mg/mL and YPD plates were supplemented with 150 μg/mL NAT . To introduce the NIK1G2048T mutation into a wild-type strain , the plasmid pLC889 was digested with BssHII and transformed into CaLC239 . For NAT-resistant transformants , proper integration was tested by PCR with primers oLC3508/oLC275 and oLC3515/oLC274 . The SAP2 promoter was induced to drive expression of FLP recombinase to excise the NAT marker cassette [101 , 102] . Presence of the NIK1G2048T allele and absence of spurious mutations was verified by amplifying the region with the mutation ( oLC3508/oLC275 ) and sequencing with oLC3511 . To introduce the NIK1G1319C allele into the caspofungin-resistant clinical isolate ( DPL15 ) , the plasmid pLC891 was digested with BssHII and transformed into CaLC990 . For NAT-resistant transformants , proper integration was tested by PCR with primer pairs oLC3508/oLC275 and oLC3515/oLC274 . The SAP2 promoter was induced to drive expression of FLP recombinase to excise the NAT marker cassette . The presence of the NIK1G1319C allele was verified by amplifying the region with the mutation ( oLC3508/oLC275 ) and sequencing with oLC3703 . To introduce the YKE2T95A mutation into a wild-type strain , the plasmid pLC907 was digested with BssHII and transformed into CaLC239 using standard protocols . For NAT-resistant transformants , proper integration was tested by PCR with primers oLC3502/oLC275 . The SAP2 promoter was induced to drive expression of FLP recombinase to excise the NAT marker cassette . Presence of the YKE2T95A allele and absence of spurious mutations was verified by amplifying the region with the mutation ( oLC3502/oLC275 ) and sequencing with oLC3503 . To introduce the NIK1G2048T mutation into the hog1Δ/hog1Δ mutant , the plasmid pLC889 was digested with BssHII and transformed into CaLC3943 . For NAT-resistant transformants , proper integration was tested by PCR with primers oLC3508/oLC275 and oLC3515/oLC274 . The SAP2 promoter was induced to drive expression of FLP recombinase to excise the NAT marker cassette . Presence of the NIK1G2048T allele and absence of spurious mutations was verified by amplifying the region with the mutation ( oLC3508/oLC3512 ) and sequencing with oLC3511 . Absence of a wild-type allele of HOG1 was verified using oLC4151/oLC4152 . The SSK1 knockout construct was PCR-amplified from pLC49 using primer pair oLC4427/oLC4428 , containing sequence homologous to upstream and downstream regions of SSK1 , and transformed into the wild type CaLC2302 . NAT-resistant transformants were PCR tested for proper integration of the construct using primer pairs oLC275/oLC4429 and oLC274/oLC4431 . The SAP2 promoter was induced to drive expression of FLP recombinase [101 , 102] to excise the NAT marker cassette . The second SSK1 allele was deleted in the same manner . Absence of a wild-type allele of SSK1 was verified using oLC4429 and oLC4430 . To C-terminally 3xHA tag Nrg1 , the construct was amplified from pFA-ARG4 in which NRG1 and HA were cloned at the BamHI and SalI sites , with primers oLC4581 and oLC4582 . The construct was transformed into CaLC192 and arginine protrophs were tested for proper integration by PCR with primer pairs oLC4583/oLC4584 . To C-terminally HA tag Nrg1 in a wild type , the construct was amplified from pLC576 [103] with primers oLC4370 and oLC4371 and transformed into CaLC2302 . Arginine protrophs were tested for proper integration by PCR with primer pairs oLC4372/oLC4374 and oLC4375/oLC4373 . To C-terminally HA tag Nrg1 in a ubr1Δ/ubr1Δ mutant , the construct was amplified from pLC576 with primers oLC4370 and oLC4371 and transformed into CaLC4455 . Arginine protrophs were tested for proper integration by PCR with primer pairs oLC4372/oLC4374 and oLC4375/oLC4373 . Absence of a wild-type UBR1 allele was verified with primers oLC4240 and oLC4241 . All oligonucleotide sequences used in this study are included in S2 Table . Downstream of NIK1 was amplified by PCR from CaLC239 genomic DNA using primers oLC3513/oLC3514 . This PCR product and pLC49 [25] were digested with SacI and SacII and ligated . Integration was verified by sequencing with oLC274 . NIK1G2048T was amplified by PCR from CaLC3350 genomic DNA using primers oLC3509/oLC3512 . This PCR product and the plasmid pLC49+CaLC239 oLC3513/oLC3514 ( above ) were digested with KpnI and ApaI and ligated . Presence of the NIK1G2048T allele and lack of spurious mutations was verified by sequencing with primers oLC243 , oLC3703 , oLC3704 , oLC3511 , oLC3785 , oLC3786 , oLC3787 and oLC3788 . Downstream of NIK1 was amplified by PCR from CaLC239 genomic DNA using primers oLC3513/oLC3514 . This PCR product and pLC49 were digested with SacI and SacII and ligated . Integration was verified by sequencing with oLC274 . The NIK1G1319C allele was amplified by PCR from CaLC3161 genomic DNA using primers oLC3509/oLC3512 . This PCR product and the plasmid pLC49+CaLC239 oLC3513/oLC3514 ( above ) were digested with KpnI and ApaI and ligated . Presence of the NIK1G1319C allele and lack of spurious mutations was verified by sequencing with primers oLC243 , oLC3703 , oLC3704 , oLC3511 , oLC3785 , oLC3786 , oLC3787 and oLC3788 . Downstream of YKE2 was amplified by PCR from CaLC239 genomic DNA using primers oLC3505/oLC3506 . This PCR product and pLC49 were digested with SacI and SacII and ligated . Integration was verified by sequencing with oLC274 . YKE2T95A was amplified by PCR from CaLC3211 genomic DNA using primers oLC3503/oLC3504 . This PCR product and the plasmid pLC49+CaLC239 oLC3505/oLC3506 ( above ) were digested with KpnI and ApaI and ligated . Presence of the YKE2T95A allele and lack of spurious mutations was verified by sequencing with primers oLC243 and oLC275 . Resistance to single antifungal drugs or drug combinations was assayed in 96-well microtiter plates ( Sarstedt ) as previously described [20 , 21 , 104] . Assays were performed in a total volume of 0 . 2 mL/well with 2-fold dilutions of each drug in the indicated medium . Plates were incubated in the dark at 30°C before OD600 were determined using a spectrophotometer ( Molecular Devices ) , at the indicated time point . Data was displayed as heat maps using Java TreeView 1 . 1 . 6 . DTPA ( Sigma ) was dissolved in NaOH and ddH2O and the pH was adjusted to 7 . DPI ( Sigma ) and Ciclopirox ethanolamine ( ShelleckChem ) were prepared in DMSO . MRS2159 ( Sigma ) was prepared in ddH2O . NPA ( Sigma ) was prepared in methanol . Caspofungin was a generous gift from Merck and was dissolved in ddH2O . Cell pellets for whole genome sequencing were prepared by centrifuging 50 mL of overnight culture at 3 , 000 rpm for 10 minutes and washing with 40 mL of ddH2O . Sequencing libraries were prepared and alignment performed as previously described [105] . The sequence data is publicly available on the NCBI Sequence Read Archive with accession number SRX1799649 . To monitor expression of BRG1 and ROB1 , the strains were grown overnight in YPD at 30°C , diluted to an OD600 of 0 . 1 in the presence or absence of 100 μM DTPA , and grown for 24 hours at 30°C . Subsequently , untreated cells were diluted to an OD600 of 0 . 1 in YPD and DTPA-treated cells were diluted into YPD containing 100 μM DTPA and grown for 4 hours . Cultures were pelleted and frozen at -80°C . RNA extraction , complementary DNA synthesis and PCR were performed as previously described [106] . Reactions were performed in triplicate , for two biological replicates using the primers , oLC2635/oLC2636 ( BRG1 ) , oLC2637/oLC2638 ( ROB1 ) , oLC2285/oLC2286 ( ACT1 ) and oLC752/oLC753 ( GPD1 ) . All data were normalized to ACT1 and GPD1 . Data were analyzed using the BioRad CFX Manager 3 . 1 . To measure expression of HWP1 , ZRT1 and ZRT2 , the strains were grown overnight in YPD at 30°C , diluted to an OD600 of 0 . 1 in the presence or absence of 10 μM DTPA , and the presence or absence of 20 μg/mL doxycycline and grown for 24 hours at 30°C . Subsequently , cells were diluted to an OD600 of 0 . 05 in YPD with or without 20 μg/mL doxycycline and with or without 50 μM DTPA . Cells were pelleted for RNA extraction after 4 . 5 hours of growth at 30°C . Analysis was performed as above , using the primers oLC752/oLC753 ( GPD1 ) , oLC3796/oLC751 ( HWP1 ) , oLC5109/oLC5115 ( ZRT1 ) and oLC5111/oLC5112 ( ZRT2 ) . Data was normalized to GPD1 . Oligonucleotide sequences are included in S2 Table . Imaging from liquid media was performed using differential interference contrast microscopy using a Zeiss Axio Imager . MI ( Carl Zeiss ) . TPEN ( N , N , N , N-Tetrakis ( 2-pyridylmethyl ) ethylenediamine ) ( Sigma ) was prepared in ethanol . Geldanamycin ( LC laboratories , G-4500 ) was dissolved in DMSO . A caspofungin-resistant isolate ( CaLC990 ) was grown for 16 hours in RPMI in the presence or absence of 0 . 32 μg/mL caspofungin or 50 μM DTPA . Cells were washed with PBS and stained with Aniline Blue ( 0 . 05% ) , Calcofluor White ( 25 μg/mL ) or Concanavalin A ( 20 μg/mL ) . Images for a single stain were taken at the same exposure . To test for Mkc1 activation via phosphorylation , C . albicans was initially grown overnight at 30°C in YPD while shaking at 200 rpm . Subsequently , the stationary phase culture was split and adjusted to an OD600 of 0 . 1 in 2 mL YPD with or without 100 μM DTPA and grown as above for 24 hours . These stationary subcultures were each split into two and adjusted to an OD600 of 0 . 1 in 50 mL YPD . These were grown for 2 . 5 hours at 30°C , at which point each condition was treated with 125 ng/mL caspofungin for 1 hour . Cells were harvested by centrifugation and washed once with ice-cold 1×PBS . Pellets were resuspended in 200 μl lysis buffer ( 50 mM HEPES pH 7 . 5 , 150 mM NaCl , 5 mM EDTA , 1% Triton X100 , protease inhibitor cocktail ( Roche Diagnostics ) , 50 mM NaF , 10 mM Na3VO4 , 1 mM PMSF ) . Proteins were extracted and analyzed as previously described [21] . Proteins were separated by 8% SDS-PAGE and blocked with 5% bovine serum albumin in tris-buffered saline with 0 . 1% tween . Blots were hybridized with antibodies against α-phospho-p44/42 MAPK ( Thr202/Tyr204 ) ( 7:10 , 000 , Cell Signaling ) and α-PSTAIRE ( 1:5 , 000; Sigma ) . To test for Hog1 activation via phosphorylation , cells were initially grown as above . Stationary phase cultures were split and adjusted to an OD600 of 0 . 1 in 50 mL RPMI and grown for 5 hours 50 minutes at 30°C . For DTPA-treated cultures , DTPA was added at time 0 . For caspofungin-treated cultures , cells were grown for 4 hours 20 minutes , treated with caspofungin , and were left to grow for an additional 1 hour 30 minutes . For H2O2-treated cultures , 10 mM H2O2 was added for the last 10 minutes of growth . Cells were harvested and proteins were analyzed as above . Blots were hybridized with antibodies against α-phospho-p38 MAPK ( T180/Y182 ) ( 1:2 , 500 , Cell Signaling ) , α-Hog1 ( 1:1 , 2500 , AbDseroTec ) , and α-tubulin ( 1 , 1000 , Santa Cruz Biotechnology ) . To test for Hog1 activation in response to magnesium depletion , wild-type cells ( CaLC239 ) , were grown in 10 mL of SD medium either in the presence of 20 μM DTPA , or in Chelex-treated SD with all metals restored except for magnesium . Cells were grown for 5 hours . Proteins were extracted and analyzed as above . To monitor total levels of Nrg1 , strains were cultured as described above and diluted to an OD600 of 0 . 1 in 10 mL of the appropriate media ( Fig 8 ) or 25 mL YPD ( Fig 9 ) to grow for 4 . 5 hours . Protein was extracted as previously described [22] , and analyzed as above but blocked with 5% skim milk in phosphate-buffered saline . Blots were hybridized with antibodies against α-HA ( 1:5000 , Roche Diagnostics ) . Female BALB/c mice were inoculated with 5x105 cells of C . albicans strain CaLC990 , resuspended in 150 μl of sterile PBS , by tail vein injection . There were four treatment groups consisting of ten mice each , which were ordered from Charles River . Test groups were treated with vehicle ( water ) , DTPA ( Sigma-Aldrich ) alone at 60 mg/kg/dose , caspofungin ( Merck ) alone at 0 . 05 mg/kg/dose , and a combination of DTPA and caspofungin at these doses . Drug doses were administered by intraperitoneal injection starting four hours post-infection , then every 24 hours for a total of five doses . Mice were monitored daily for weight loss and overall health condition , and were euthanized upon reaching humane endpoints as defined in Duke IACUC protocol A114-14-05 . The survival curve was calculated using Prism software version 4 . | Invasive fungal infections pose a serious threat to human health worldwide , with Candida albicans being a leading fungal pathogen . Mortality is in part due to the limited arsenal of effective antifungals , with drug resistance on the rise . The echinocandins , which target the fungal cell wall , are the newest class of antifungal , and echinocandin resistance has already emerged . Here , we screened a library of 1 , 280 pharmacologically active compounds to identify those that potentiate echinocandin activity against an echinocandin-resistant isolate . The lead compound was a chelator , DTPA , which affects resistance by depleting magnesium . Genome sequencing of mutants resistant to the combination of DTPA and echinocandin revealed mutations in the gene encoding Nik1 , which signals upstream of the Hog1 stress response pathway . We established that DTPA acts through Nik1 to modulate Hog1 signaling and enhance echinocandin activity , and that this combination has therapeutic benefits in a murine model of candidiasis . We also discovered that DTPA modulates C . albicans morphogenesis , a key virulence trait . DTPA induced filamentation by chelating zinc , in a manner that is contingent upon core filamentation pathways and specialized circuitry . Thus , we establish novel roles for metal homeostasis in C . albicans pathogenesis , thereby illuminating new therapeutic strategies for life-threatening infectious disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"antimicrobials",
"cell",
"walls",
"medicine",
"and",
"health",
"sciences",
"cellular",
"stress",
"responses",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"cell",
"processes",
"microbiology",
"antifungals",
"fungi",
"model",
"organisms",
"pharmac... | 2016 | Metal Chelation as a Powerful Strategy to Probe Cellular Circuitry Governing Fungal Drug Resistance and Morphogenesis |
Streptococcus pneumoniae isolates typically express one of over 90 immunologically distinguishable polysaccharide capsules ( serotypes ) , which can be classified into “serogroups” based on cross-reactivity with certain antibodies . Pneumococci can alter their serotype through recombinations affecting the capsule polysaccharide synthesis ( cps ) locus . Twenty such “serotype switching” events were fully characterised using a collection of 616 whole genome sequences from systematic surveys of pneumococcal carriage . Eleven of these were within-serogroup switches , representing a highly significant ( p < 0 . 0001 ) enrichment based on the observed serotype distribution . Whereas the recombinations resulting in between-serogroup switches all spanned the entire cps locus , some of those that caused within-serogroup switches did not . However , higher rates of within-serogroup switching could not be fully explained by either more frequent , shorter recombinations , nor by genetic linkage to genes involved in β–lactam resistance . This suggested the observed pattern was a consequence of selection for preserving serogroup . Phenotyping of strains constructed to express different serotypes in common genetic backgrounds was used to test whether genotypes were physiologically adapted to particular serogroups . These data were consistent with epistatic interactions between the cps locus and the rest of the genome that were specific to serotype , but not serogroup , meaning they were unlikely to account for the observed distribution of capsule types . Exclusion of these genetic and physiological hypotheses suggested future work should focus on alternative mechanisms , such as host immunity spanning multiple serotypes within the same serogroup , which might explain the observed pattern .
Streptococcus pneumoniae is a human nasopharyngeal commensal bacterium and important respiratory pathogen . The ability of the pneumococcus to cause invasive disease appears critically dependent upon its polysaccharide capsule [1] , of which at least 95 immunologically-distinguishable capsular variants ( serotypes ) are known [2–7] . This structure inhibits the recognition of subcapsular protein antigens by the adaptive immune system and the binding of phosphorylcholine residues by C-reactive protein , thereby reducing the rate of complement deposition on the bacterial surface [8] . Children develop anticapsular antibodies after exposure to the bacterium , although evidence for their impact on disease risk is mixed [9] . Data indicating this immune response provides protection against nasopharyngeal carriage has been found for only a few serotypes [10–12] , though mathematical models suggest subtle effects on carriage may exist and help maintain the high level diversity of serotypes observed in pneumococcal populations [13] . In contrast to natural immunity , that induced by the seven-valent protein conjugate polysaccharide vaccine ( PCV7 ) reduces acquisition of vaccine serotypes in the nasopharynx by 50% [14] or more [15] . Following the introduction of PCV7 , a substantial fall in the carriage of the seven vaccine serotypes ( 4 , 6B , 9V , 14 , 18C , 19F and 23F ) was observed without any substantial overall reduction in the rates of pneumococcal colonisation [16 , 17] . This was primarily the result of an increase in the rate of carriage of non-vaccine type strains , termed ‘serotype replacement’ [18–20] . In some cases non-vaccine type strains are closely related to vaccine type strains from which they have been derived by ‘serotype switching’ [21] . In these cases , strains have altered their serotype through the acquisition of a different capsular polysaccharide synthesis ( cps ) locus via genetic transformation . Many serotypes , though distinguishable by certain antisera ( called "factor sera" ) , nonetheless may be clustered into “serogroups” based on cross-reactivity with other antisera; these groups have often been found to correspond to sets of similar polysaccharide structures [2] . It was originally hoped that PCV7 would provide cross-protection against ‘vaccine-related’ serotypes: those within the same serogroups as a serotype included in the vaccine [22] . However , the only case in which such an effect was observed was the protection against colonisation with serotype 6A resulting from the inclusion of serotype 6B in PCV7; several other vaccine-related types increased in prevalence post-PCV7 [23] . Contributing to this pattern was an apparent tendency for serotype switches to exchange one serotype for another within the same serogroup more often than expected by chance: of nine switches inferred using multilocus sequence typing ( MLST ) in a systematic collection of carried pneumococci from Massachusetts , three were within-serogroup ( p = 0 . 043 ) [24] . Hence genotypes successful prior to the introduction of PCV7 were able to persist expressing a similar capsule that was not recognised by vaccine-induced immunity . This was not a pattern expected a priori , as random acquisition of a new serotype from the full , diverse set of non-vaccine type capsules should usually result in a change of serogroup . This work assesses the relative likelihood of alternative explanations of the observed pattern of serotype switching based on whole genome sequencing data from this Massachusetts-based collection [25] . The first is that the detected propensity for within-serogroup serotype switching is a consequence of the constraints of genetic transformation . Pneumococcal cps loci are typically 10–30 kb in size , usually necessitating a similarly long recombination to cause a change of serotype [26] , whereas homologous recombinations have an exponential distribution of sizes with a mean length of between two and seven kilobases [27–29] . Aside from known exceptions such as serogroups 7 , 17 , 33 and 35 , the cps loci corresponding to a single serogroup are closely related [2] . Therefore , within-serogroup serotype switching may be accomplished through relatively short , more frequent , recombination events that do not replace the entire cps locus [30] . Another possibility is that patterns of serotype switching may be affected by the flanking pbp1a and pbp2x genes , which are crucial in determining β–lactam resistance . As resistance is associated with a limited number of serotypes [31] , and long recombinations that change serogroup could lead to the acquisition of resistance [32] or risk reversing any beneficial acquisition of resistance-associated pbp1a or pbp2x alleles [29 , 33] , it may be that maintenance of a strain’s β–lactam susceptibility ( or lack thereof ) affects the patterns of serotype switching . Alternatively , rather than representing a limitation on the rate of recombination , the pattern of switching may reflect the consequence of constraints imposed by epistatic interactions with other loci in the chromosome [26] . Such limitations may reflect physiological or metabolic specialisation to production of particular capsule types; alternatively , it may be important to co-ordinate the expression of particular serogroups with certain alleles of immunogenic surface proteins , in order to maintain discordant sets of antigens in distinct lineages [34] . The last explanation to be considered , which also involves the host immune response , relies on the presence of natural or vaccine-induced antibodies that cross-react with all serotypes within a serogroup . This would lead to elevated rates of co-colonisation between strains of the same serogroup , as bacteria would be confined to a subpopulation of hosts that do not exhibit immunity to either of their capsules . A consequence of this would be increased genetic exchange between members of the same lineage , including an elevated rate of within-serogroup serotype switching .
Of the 616 draft genomes previously analysed [25] , 491 fell into fifteen monophyletic sequence clusters ( SCs ) of related isolates within which serotype switching could be investigated ( Fig . 1 ) . These sequence clusters included representatives of 19 serotypes spread across eleven serogroups . In seven sequence clusters , the serotype was stable across the clade , leaving eight clusters in which at least one switch had occurred . There was no positive correlation between the number of isolates sampled within a sequence cluster and the number of serotypes it contained ( Fig . 2A ) . However , a significant correlation was observed between the number of polymorphic sites , as ascertained through a lineage-specific analysis of whole genome alignments , and the number of serotypes in a cluster ( Fig . 2B ) . When these sites were split into the number of point mutations and homologous recombination events per sequence cluster ( see Methods ) , there was no significant relationship with the former , whereas the latter had the strongest correlation of any measure of genetic diversity ( Fig . 2C and D ) . Therefore sequence clusters that exhibit serotype diversity are those that have experienced more recombination events throughout their genome over their observable evolutionary history , either as a consequence of a high rate of recombination over their recent diversification , or lower recombination rates over a longer period of time . Across the species , the 95 currently known S . pneumoniae serotypes are divided into 48 serogroups such that 2 . 2% of comparisons between different serotypes are serogroup concordant . However , many serotypes are rarely observed; the 32 serotypes observed across all 616 isolates from Massachusetts were distributed such that 3 . 0% of serotype comparisons were serogroup concordant . Seventeen serotypes were observed within the eight sequence clusters that appear to contain examples of serotype switching; 7 . 2% of all pairwise comparisons between these serotypes were serogroup-concordant , suggesting these isolates expressed a comparatively limited number of serogroups . Yet when only considering the subset of these comparisons where the different serotypes were found within the same sequence cluster , 29% were serogroup concordant ( S1 Table ) , as some individual sequence clusters were associated with multiple serotypes from the same serogroup . The significance of this enrichment was assessed by a permutation test that randomly assigned serotypes to these eight sequence clusters according to the number of serotypes originally observed in the sequence cluster . This found the high level of serogroup-concordant comparisons within sequence clusters to be statistically significant ( p = 0 . 0007 from 10 , 000 permutations ) . As sequence clusters were identified through a clustering algorithm [25] , this observation is independent of any phylogenetic analyses . Therefore this test indicates isolates with similar core genomes are more likely to express capsules of the same serogroup than expected from the overall capsular diversity of the isolate collection , even when accounting for isolates sharing the same serotype through common descent . Greater resolution can be achieved by using the phylogenetic analysis of the whole genome alignments for each of the sequence clusters in which serotype switching occurred ( S1 Text & Fig . 3 ) [25] . These attempt to reconstruct the history of the lineages more accurately by identifying recombination events and excluding the horizontally acquired polymorphisms they introduce from the point mutations used to generate the tree [29] . These reconstructions found that some within-serogroup switching occurred multiple times in parallel in the same sequence cluster: for instance , in SC9 the ancestral type of 23A was replaced by 23B on three occasions , and 23F on three occasions . Similarly within SC13 , the ancestral 6A type is exchanged for 6C on two occasions in parallel . In both cases , these examples of convergent evolution are independently supported by the divergence between the cps loci , encoding the same serotype , imported by separate capsule switching events ( S1 Fig ) . A further switch to 6C , this time from 6B , was observed within SC6 . However , the other changes within this sequence cluster cannot be reliably inferred as they occur on long branches that are difficult to reconstruct . These ‘missing data’ include at least two between-serogroup switches , but also likely within-serogroup switches given that the cluster includes 23A , 23B and 23F isolates . By contrast , each of the robustly inferred between-serogroup switches occurred only once . Hence this more detailed analysis revealed an even greater tendency to swap a serotype for one within the same serogroup than the previous results . Overall , eleven of the twenty serotype switches ( 55% ) fully defined in this population did not result in a change in serogroup . A further permutation test , in which the derived serotypes were randomly assigned to each ancestral serotype through resampling without replacement 10 , 000 times ( see Methods ) , found this result to be highly significant ( p < 0 . 0001 ) . Therefore this collection provides robust evidence that within-serogroup switching is more common than expected by chance , relative to the rate of between-serogroup switching . The first hypothesis that might explain this observation concerns the mechanics of recombination: the smaller recombinations necessary to alter a cps locus to a related sequence generating a similar serotype may be more frequent than the replacement of the entire locus . The within-serogroup switches observed in this collection all require only minor changes in the cps locus: the serogroup 23 cps loci are thought only to differ in the sequence of their wzy polymerase , as is also the case for serotypes 19A and 19F [35] . Similarly , switching between serotypes 6A and 6B can occur through a single polymorphism in the wciP gene [30] , while 6A and 6C only differ in polymorphisms in their wciN gene [35] . Including only the switch from serotype 6B to 6C from SC6 ( see S1 Text ) , the analysis of whole genome alignments used to generate Fig . 3 identified 50 putative homologous recombination events that overlapped the cps loci of the relevant reference genome sequences ( Fig . 4 ) . The nine recombinations that resulted in a change in serogroup had a median length of 42 . 7 kb , and all spanned the entire cps locus . The eleven recombinations that led to a within-serogroup switch had a shorter median length , at 30 . 4 kb ( Fig . 5A ) . These almost all spanned the 5’ region of the cps locus , but in some cases did not extend so far as the 3’ end . In two cases , this could be ascribed to the presence of the rml rhamnose synthesis operon at the 3’ end of the cps locus . Both the acquisition of the 6B capsule in SC13 , and the 6C capsule in SC6 , terminated within this gene cluster; however , the rml operon is found in several serogroups [2] , and recombinations causing between-serogroup switches have previously been observed to end within it [29] . Only three recombination events were observed where the 3’ recombination breakpoint occurred in a region that might be considered serogroup-specific . Two of these were switches from 23A to 23F within SC9 ( Fig . 4 ) . In both cases , the wzy polymerase gene that distinguished these cps loci was replaced; however , the recombinations were far more extensive than this minimal alteration , as they extended to or beyond the 5’ boundary of the cps locus . The switch to 6C within SC13 that ended before the 3’ boundary of the locus was also far more extensive than simply encompassing the wciN gene . The overall difference in the distribution of lengths between recombinations causing within- and between-serogroup switches was not significant ( Fig . 5; Wilcoxon rank sum test , W = 35 , p value = 0 . 29 ) . By contrast , the 30 recombinations inferred to overlap with the cps loci that did not affect serotype had a median length of 11 . 7 kb and were significantly shorter than both the recombinations that alter serogroup ( Wilcoxon rank sum test , W = 22 , p value = 3 . 4x10-5 ) and those that caused within-serogroup switching ( Wilcoxon rank sum test , W = 61 , p value = 0 . 0016 ) . As the majority of within-serogroup switches were caused by recombinations long enough to cause changes in serogroup , restrictions on transformation event length cannot fully explain the observed pattern of switching . There was extensive variation in the boundaries of the three classes of recombination event . For both 5’ and 3’ breakpoints outside the cps locus , the distance between the breakpoint and the cps locus edge ( Fig . 4B ) followed an approximately exponential distribution , as previously observed for experimental transformants at the cps locus [28] . However , these observed recombinations were longer than the length of transformation events observed experimentally , with a rate of decay approximately an order of magnitude lower in this work: the rate constant for the decline on the left side was 8 . 07x10-5 bp-1 ( 95% confidence interval: 8 . 06x10-5–8 . 09x10-5 bp-1 ) , while on the right side it was 6 . 75x10-5 bp-1 ( 95% confidence interval: 6 . 74x10-5–6 . 77x10-5 bp-1 ) , as opposed to the previous estimates of ~3 . 4x10-4 bp-1 . This is likely partly a consequence of the events described in Fig . 4 being composed long “mosaics” of recombinant DNA segments [28] , and also potentially representing larger “macrorecombination” events [36] . The pbp2x and pbp1a genes , crucial determinants of β–lactam resistance , are found 9–10 kb upstream and downstream of the cps locus respectively . Seven of the recombinations shown in Fig . 4 affected both , all of which caused a change of serotype . However , they were split between within- and between-serogroup switches , and none significantly affected the strain’s β–lactam resistance ( S2 Table & S2 Text ) . Hence serotype switches did not appear to be driven by selection for the acquisition of β–lactam resistance . Nevertheless , it is possible that maintenance of a genotype’s β–lactam resistance level may counterselect against recombination events that substantially alter pbp1a or pbp2x [33] . If the distribution of capsular variation does not reflect the properties of genetic transformation , then selection may be important in driving the observed pattern . Possible selective pressures include adaptive immune responses in the host population [34] , or physiological constraints relating to the bacterium . To test the latter possibility , capsule-switched variants were constructed in common genetic backgrounds and characterised through recording growth curves . The association of SC9 with serogroup 23 was investigated by knocking out the native cps locus of isolate R34-3029 , and using this genetic background to construct one mutant in which the native 23F capsule type was restored , a second through a within-serogroup switch to capsule type 23B , and two further mutants through between-serogroup switches to 18C and 6B . Capsule type 18C was selected as it was the only non-serogroup 23 capsule type observed within SC9 , whereas 6B was not found within SC9 but has similar properties to 23F [37] . Both serogroup 23 mutants were found to grow similarly quickly ( Fig . 6A ) and substantially faster than the two mutants bearing capsules of other serogroups . This suggested that these bacteria may be adapted to serogroup 23 capsules . However , comparisons of the original isolates used as the sources of the cps loci ( S2 Fig ) indicated that the apparent reduction in fitness of SC9 mutants expressing non-serogroup 23 capsule types might reflect an intrinsically slower growth phenotype associated with these capsule types . Furthermore , when the reciprocal exchange of cps loci between isolates BR1014 and R34-3029 was performed , BR1014 grew faster following restoration of its native 23B capsule relative to the variant into which the 23F capsule had been introduced ( Fig . 6A ) . Therefore changes in capsule type were generally associated with a greater reduction in growth rate than reinstatement of the original capsule type; however , this inhibition could be observed regardless of whether the new serotype was of a different serogroup or not . To test this hypothesis further , six additional mutants were constructed using two SC13 isolates in which the cps locus had been knocked out: MD5037 , originally of 6B , and MD5030 , originally of 6C ( Fig . 6B ) . The 6B , 6C and 15F capsule types were introduced into both . This comparison allowed a more precise test for interactions between the cps locus and the rest of the genome than the experiments with SC9 isolates , as SC13 isolates exhibited more consistent growth patterns , and the selected donor of the 15F capsule had a similar growth curve to the recipient isolates ( S2 Fig ) . Consistent with the observations of SC9 , the mutants with the restored native capsule grew fastest . However , both the within-serogroup and between-serogroup switched variants exhibited similarly inhibited growth rates . Hence all three capsule types replicated optimally in their native backgrounds , and more slowly when introduced into a non-native background . Therefore the only epistasis between the cps locus and the rest of the pneumococcal genome that could be inferred from these data was specific to serotype , not serogroup , and no evidence was found that the observed predominance of within-serogroup switches is explained by serogroup-specific adaptations . It remains possible that epistasis affects other phenotypes that are not easily assayed in the laboratory; it may be possible to infer these from the genomic data . One factor likely to impact on the range of capsules a strain may successfully express is its complement of carbohydrate transporters [38 , 39] , many of which have been functionally characterised [38] and are only present in a subset of the population ( S3 Fig & S3 Table ) . While capsules in different serogroups usually have distinct chemical compositions , capsules within the same serogroup often consist of the same moieties connected by different bonds . Hence strains may only be adapted to importing and processing certain carbohydrates at high enough rates to sustain capsule production , thereby inhibiting the acquisition of a capsule type with a divergent chemical composition . However , the distribution of transporters across the population does not provide an obvious explanation for the stable association between sequence clusters and serogroups ( S3 Fig ) . SCs 9 , 13 and 15 did not appear to have a smaller number of carbohydrate transporters than those in which a change of serogroup had occurred , nor relative to those not having changed serotype . The associated serogroups can be distinguished by the presence of ribitol ( serogroup 6 ) , N-acetylmannosamine ( serogroup 19 ) and glycerol ( serogroup 23 ) . Ribitol is synthesised as part of teichoic acid common to all pneumococci; glycerol is imported by GlpF , encoded by a gene ubiquitous across the sample; and N-acetylmannosamine is known to be a substrate for four transporters that are common to all isolates , and one that is absent from SC15 . These data do not support the hypothesis that the observed pattern of switching is the result of a constricted ability to acquire the requisite carbohydrate molecules , although epistatic interactions with other aspects of carbohydrate metabolism cannot be excluded . Yet this analysis only uses experimentally characterised loci , which cannot encompass novel systems that may be rare in the population . It is possible in principle to identify epistatic interactions between any loci from sequence data [40] . In bacteria , interacting accessory genome loci should be present within the same genomes more often than expected by chance , after accounting for linkage and clonal structure . Three sequence clusters , each essentially an independent genotype [41] , were universally of serogroup 6 ( SC10 , SC13 and SC14 ) . A search for clusters of orthologous genes ( COGs ) that were ubiquitous in these sequence clusters but absent from sequence clusters with no serogroup 6 representatives identified three COGs , all of which corresponded to genes within the serogroup 6 cps loci . This failure to identify serogroup 6 associated genes outside the cps locus is in keeping with the experimental data that suggests adaptation is specific to a serotype , not a serogroup . Testing for association with serotype requires that the same capsule be acquired multiple times in parallel , as in SC9 . Each triallelic polymorphic site in SC9 was tested for association of an allele with one of the three serogroup 23 serotypes . Twelve such sites were identified , all of which were within the cps locus . Additionally , no COGs in the accessory genome associated perfectly with any of these individual serotypes . Therefore these simple analyses provided no evidence of sites outside the cps locus that might epistatically interact with the expressed capsule type . A more generic strategy for identifying potential epistatic interactions was not to look for associations of specific sequences with particular cps loci , but instead quantify the overall extent of changes accompanying serotype switches . If one or more non-cps loci improved the fitness of a novel serotype variant , then it might be expected that serotype switches would be associated with elevated rates of change across the rest of the genome . In order to control for the general heterogeneity in the rates of pneumococcal recombination [36] , the total lengths of the recombinations occurring on the same phylogenetic branch as serotype switches ( both within- and between-serogroups ) were compared with those occurring on branches on which recombinations affected the cps locus but did not cause a change in serotype ( Fig . 5D ) . This found that similar levels of contemporaneous ( or near-contemporaneous ) recombination occurred whether or not there was a change in serotype or serogroup . Additionally , all three types of recombination affecting the cps locus were typically associated with sufficient numbers of base substitutions being imported by recombination ( S4 Fig ) to saturate the mismatch repair system [42] . Hence , there was no significant evidence of non-cps loci being exchanged to facilitate the acquisition of a new serotype or serogroup , nor of mismatch repair having an impact on the patterns of serotype switching .
Previous study of this collection of carried pneumococci , based on serotyping and MLST data , identified a borderline-significant tendency for serotype switches to occur within serogroups [21] . The availability of whole genome-based phylogenies increases the number of switches that can be investigated within the same set of isolates through identification of two classes of switch invisible to the MLST-based method: changes in serotype along a lineage in which the multilocus sequence type has also changed ( e . g . SC9 ) , and also multiple independent switches within the same sequence type ( e . g . in SC13 , two switches to 6C within MLST 473 ) . This increases the sample size of identified serotype switching events from nine to twenty . Also , these data provide more reliable phylogenetic information on the ancestral serotype: previously , this was based on the temporal order in which combinations of serotype and genotype were observed , which can be misleading in certain cases , such as SC9 [25] . Future population genomic datasets will determine how general the preponderance of within-serogroup switching is across the pneumococcal species; a recently published dataset of over 3 , 000 genomes systematically sampled from a Thai refugee camp , in which hosts were unvaccinated , provides an opportunity to test the reproducibility of the pattern [43] . Applying the simple permutation test to the distribution of serotypes between sequence clusters in which switching had occurred , as for the dataset described in this study , revealed the ratio of serogroup concordant to serotype-discordant comparisons to be 0 . 14 ( p = 0 . 0006 from 10 , 000 permutations ) . The serotypes in this sample are markedly different from those in the Massachusetts dataset; hence this enrichment of within-serogroup switching provides independent support for the original observation . Additionally , although only one within-serogroup switch was observed in SC15 in this dataset , the acquisition of the 19A capsule in place of the 19F type has occurred at least three further times in related isolates [33] . Multiple hypotheses were proposed as potential explanations for the observed frequency of within-serogroup switches . The first was that the pattern was caused by the constraints of genetic transformation , with short recombination events resulting in partial alteration of the cps locus being more frequent . The suggestion that most recombinations are too short to be likely to cause a change in serotype was supported by the large number of ‘silent’ recombinations affecting the cps locus , which did not change the antigenic profile of the bacterium . While some cases are likely to reflect exchanges of sequence between distinct isolates of the same serotype , many of these events affect the 5’ or 3’ boundaries of the cps locus that are the most strongly conserved across different cps loci , and therefore may originate in donors with a very different cps locus to that of the recipient . As these exchanges do not result in antigenic changes , it may be that they represent neutral diversification , or serve a role in repairing deleterious mutations . This potential for recombination between serotypes that does not cause switching should be borne in mind when designing sequence-based serotyping methodologies that do not directly target the polymorphisms within cps loci that cause differences in capsule type . Nevertheless , recombinations causing within-serogroup switching sometimes did not span the entire cps locus , as all the between-serogroup switches did . In three cases , the recombinations were sufficiently restricted in their extent so as to end within a serogroup-specific part of the cps locus; however , these were still substantially larger than the putative minimal genetic changes that could cause the same alteration of serotype . Furthermore , the enrichment for within-serogroup switching remained highly significant even when the events that did not span the entire cps locus were excluded; although this does not rule out this mechanism contributing to the observed pattern of switching , this result demonstrates it is not the sole reason underlying it . Finally , the within-serogroup switches did not appear to represent a ‘neutral’ set of typical recombination lengths , as they were still significantly larger than those which caused no alteration of serotype overall , although there may be a limit on the length of such recombinations before they are likely to stop being ‘silent’ . One possibility not considered here , subtly different to a dependence on the absolute length of a recombination , is that restriction modification systems could be limiting the transfer of cps loci . Although restriction endonucleases typically do not affect transformation , as DNA is imported in a single stranded form , in cases where genomic islands are imported , the synthesis of a complementary strand once the DNA is integrated into the genome can render the acquired locus sensitive to endonucleolysis [44] . Loci imported from a different serogroup would have greater sequence divergence , and therefore a greater set of potential target motifs that could be subject to endonucleolysis . However , this will remain difficult to assess until our knowledge of the diverse set of pneumococcal restriction-modification systems has improved [41] . Conversely , there was no evidence that serotype switching recombinations were selected for the lengths to which they affected the regions flanking the cps locus and the pbp genes . Co-incidence of serotype switches with changes in β–lactam resistance [32] was rare and there was little evidence of recombinations being constrained in the extent to which they extended into the regions flanking the cps locus . However , there are uncertainties associated with ascertaining these recombination breakpoints . One is that the evolutionary time represented by the long branches on which the changes of serogroup occur within SC1 may be sufficient to permit several rounds of transformation . This could result in overlapping recombinations being erroneously inferred to represent a single event , thereby increasing the apparent length of the recombined sequence . However , as both events affected by this bias were between-serogroup switches , this cannot account for within-serogroup switches being longer than the minimum length required for the phenotypic change . Conversely , multiple switches to the same serotype in parallel produce extensive homoplasy . Inaccurate reconstruction of such homoplasies could result in individual actual recombination events being reconstructed as multiple fragmentary events split between different branches of the phylogeny . This risks artificially shortening serotype-switching recombinations . However as this bias only affects sequence clusters in which similar cps loci are acquired multiple times , which also happen to be those in which within-serogroup switching is most common ( SC9 and SC13 ) , any such errors are conservative with regard to the finding that shorter recombinations alone cannot account for the observed pattern of within-serogroup switching . One potential example is switch 23A→23F** , as the two cps loci apparently imported by this recombination are not particularly closely related ( S1 Fig ) ; hence this inferred recombination may in fact represent only a partial fragment of two separate serotype switching recombinations . If this were correct , it would increase the frequency of within-serogroup switches further , and suggest that the short recombination currently annotated as the single switch 23A→23F** actually corresponds to two separate , longer events that cannot be resolved with the current collection . This ambiguity highlights how unlikely it is that the numbers of parallel switching events have been overestimated as a consequence of incorrect evolutionary reconstructions . Given the substantial numbers of polymorphisms caused by capsule switching recombinations in the region of the cps locus , independent acquisitions in parallel are unlikely to be inferred without strong evidence of distinct ancestry from the rest of the chromosome . Furthermore , the diversity of the cps loci themselves , not directly used in the phylogenetic inference displayed in Fig . 3 , independently support the same reconstructed pattern of switches occurring in parallel within SC9 and SC13 . As limitations to recombination do not appear to explain the observed pattern of switching , selection for certain combinations of cps loci and genetic backgrounds may explain the distribution of capsule across the population . Metabolic or physiological adaptation between the cps locus and other aspects of the pneumococcal genotype may result in a limited range of capsule types that can be successfully expressed by a given genomic background . Although serotype itself affects growth rate [45] , this study found that the rate of growth in vitro depended on both the serotype and backbone genotype in a non-additive way . Isolates grew fastest when expressing their native capsule , although R34-3029 grew similarly fast when expressing either of the tested serogroup 23 capsule loci . However , in most cases there was no difference in the level of growth inhibition when capsules of the same , or a different , serogroup were introduced , particularly when accounting for the growth curves of the donors and recipients involved in each exchange . This is consistent with the existence of epistatic interactions between the cps locus and the rest of the genome , but not with the hypothesis that such interactions were serogroup-specific . It could still be that the changes required to adapt to the acquisition of a more similar serotype are relatively small , and more likely to occur through recombination or mutation before the capsule variant is selected out of the population . Yet over the timescales easily testable in the laboratory , no evidence was found for a mechanism that would explain the frequency of within-serogroup switching . One potential confounding problem with these experiments is the chance that non-cps loci could be co-transferred , and therefore influence the fitness of the resulting capsule-switched recombinants . However , previous experiments have suggested that random co-transfer of other sequence should not systematically affect the growth rates of transformants [28] . Furthermore , it might be expected that any co-transfer from the cps locus donor might facilitate adaptation to the acquired serotype , if there were epistatic interactions between the cps locus and the rest of the genome . However , the genomic data provided no evidence of natural serotype switches being associated with unusually high levels of sequence exchanges around the rest of the genome , suggesting the co-transfer of other loci is not likely to facilitate adaptation to an altered capsule type . Furthermore , there were no signs of non-cps loci associated with pneumococci of specific serogroups or serotypes in a manner suggestive of interactions between distinct loci in the same chromosome . Therefore , the genomic and experimental data do not support the model of simple , strong epistatic interactions determining the range of serogroups that an isolate can successfully express . One longstanding hypothesis for the maintenance of “strain structure” in recombining pathogens is that host immunity produces strains that maintain discordant sets of antigens [34] . Here , individual capsular antigens found across serogroups might be associated with a particular set of subcapsular antigens , creating epistasis in the presence of host immunity that would not be observable in vitro . Alternatively , opportunities for recombination may be disproportionately available for strains in the same serogroup as a result of serogroup-wide immunity making hosts either susceptible or resistant to colonization with particular serogroups . Both these mechanisms rely on the assumption that anticapsular antibodies are protective against carriage , and that at least part of this protection is by responses that target epitopes shared across serotypes within a serogroup . Whether this could significantly impact the rate of genetic exchange between serotypes remains an open question . The existence of cross-protection by naturally-acquired anticapsular antibodies within a serogroup is made plausible by the observation of protection against serotype 6A from the 6B component of PCV7 , and by the existence of cross-reactive antibodies used in serotyping . However , we are aware of no direct evidence for protection across a serogroup by naturally-acquired anticapsular antibodies . Furthermore , the protective immunity induced by conjugate vaccines is not serogroup-wide , as demonstrated by the post-PCV7 success of some vaccine-related serotypes , such as 19A , 23A , 23B and 6C [16] . Expanded sampling in the future may allow for these and other hypotheses to be tested further , as well as permitting more detailed searches for loci that may interact epistatically with the cps locus . Such information regarding which serotypes can be readily ‘interchanged’ may well prove important in understanding the responses to past conjugate vaccine introductions , and predicting the response to future interventions .
When calculating the significance of the association of serogroups with sequence clusters , the first test resampled each of the serotypes listed in S1 Table without replacement until each sequence cluster had been assigned the same number of serotypes as observed in the actual dataset . This test was therefore independent of the phylogenies and reconstructed patterns of serotype switching . Ten thousand such permutations were carried out , and in each case the calculated test statistic was the ratio of serogroup-concordant to serotype-discordant pairwise comparisons within sequence clusters . Any pairwise comparison between identical serotypes was ignored , as it would be serologically undetectable . While recombinations between isolates of the same serotype at the cps locus may be detected in this dataset as a subset of those events that do not alter the expressed capsule type , it is not possible to easily distinguish these from recombinations donated from isolates only of the same serogroup , or of a completely unrelated serotype , which may also not affect the recombinant’s expressed capsule type , depending on how conserved different parts of the cps locus are across pneumococci of different serotypes . The p value was calculated as the proportion of permutations for which the calculated test statistic was greater than , or equal to , the value observed in the actual population ( 0 . 41 ) . When applied to the Maela dataset , the ‘secondary BAPS’ groupings were used as the equivalent of sequence clusters . The second permutation test used information from the phylogenetic reconstructions of serotype switching . The input dataset was the twenty switches for which the ancestral and derived serotypes could both be defined ( Fig . 3 ) . Ten thousand permutations were performed in which the ancestral and derived serotypes were sampled , without replacement , from those in the input dataset . In each case , the test statistic was the proportion of serotype switches that were serogroup concordant , and the p value calculated as the proportion of the permutations in which this statistic was equal to , or greater than , the observed value ( 0 . 55 ) . The phylogenies used to identify serotype switches have been reported previously [25] . Briefly , a reference genome was assembled de novo for each sequence cluster , against which Illumina reads were aligned , and polymorphisms identified [28] . Sequences imported by putative recombination events were identified and removed , and a maximum likelihood phylogeny generated from the remaining clonal frame , as described [29] and validated [46] previously . The annotation of putative mobile genetic elements allowed the number of base substitutions introduced by point mutation , homologous recombination and non-homologous recombination to be estimated . All lengths and positions of recombination events are therefore relative to the reference sequences against which the Illumina reads were originally mapped . The identification of putative homologous recombination events associated with serotype switching is detailed in the supplement S1 Text; this also describes the assessment of within-serogroup sequence diversity , and the reconstruction of changes in serotype , achieved using the maximum likelihood approach implemented in the APE R package [47] . This approach , as well as a maximum parsimony method , was also used to reconstruct the emergence of β–lactam resistance using the trees based on the clonal frame , once each isolate had been classified as either ‘resistant’ ( benzylpenicillin minimum inhibitory concentration >0 . 06 μg mL-1 ) or ‘sensitive’ . The details of this analysis are also described in S1 Text and S2 Table . The cps locus of isolates was first knocked out using a Janus cassette [48] as described previously [49] . Transformation was conducted using genomic DNA from the serotype donor and the competence stimulating peptide appropriate to the genotype . After 2 h growth , selection was performed using plates containing 500 μg ml-1 streptomycin . Colonies were replica plated to ensure loss of kanamycin resistance and expression of the altered serotype confirmed through latex agglutination ( Statens Serum Institut , Copenhagen ) . Isolates were grown to mid-log phase in THY ( Todd Hewitt broth containing 0 . 5% yeast extract ) , titered and frozen at -70°C in 10% glycerol . Growth was compared in parallel by inoculating 6 mL of THY with a starting concentration of 106 mL-1 bacteria from freshly thawed frozen stock . Cultures were incubated at 37°C with 5% CO2 . At successive time points 200 μL was removed to a 96 well plate . Optical density at a wavelength of 610 nm was measured and dilutions were plated on blood agar plates to calculate live cell densities . | Streptococcus pneumoniae is a major respiratory pathogen responsible for a high burden of morbidity and mortality worldwide . Current anti-pneumococcal vaccines target the bacterium’s polysaccharide capsule , of which at least 95 different variants ( ‘serotypes’ ) are known , which are classified into ‘serogroups’ . Bacteria can change their serotype through genetic recombination , termed ‘switching’ , which can allow strains to evade vaccine-induced immunity . By combining epidemiological data with whole genome sequencing , this work finds a robust and unexpected pattern of serotype switching in a sample of bacteria collected following the introduction of routine anti-pneumococcal vaccination: switching was much more likely to exchange one serotype for another within the same serogroup than expected by chance . Several hypotheses are presented and tested to explain this pattern , including limitations of genetic recombination , interactions between the genes that determine serotype and the rest of the genome , and the constraints imposed by bacterial metabolism . This provides novel information on the evolution of S . pneumoniae , particularly regarding how the bacterium might diversify as newer vaccines are introduced . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Selective and Genetic Constraints on Pneumococcal Serotype Switching |
Incidence of Leishmania donovani infection and Visceral Leishmaniasis ( VL ) was assessed in a prospective study in Indian and Nepalese high-endemic villages . DAT-seroconversion was used as marker of incident infection in 3 yearly surveys . The study population was followed up to month 30 to identify incident clinical cases . In a cohort of 9034 DAT-negative individuals with neither active signs nor history of VL at baseline , 42 VL cases and 375 asymptomatic seroconversions were recorded in the first year , giving an infection∶disease ratio of 8 . 9 to 1 . In the 18 months' follow-up , 7 extra cases of VL were observed in the seroconverters group ( N = 375 ) , against 14 VL cases among the individuals who had not seroconverted in the first year ( N = 8570 ) ( RR = 11 . 5 ( 4 . 5<RR<28 . 3 ) ) . Incident asymptomatic L . donovani infection in VL high-endemic foci in India and Nepal is nine times more frequent than incident VL disease . About 1 in 50 of these new but latent infections led to VL within the next 18 months .
In the Indian subcontinent 200 million people are estimated to be at risk of developing Visceral Leishmaniasis ( VL ) . VL , also known as kala-azar is a fatal parasitic disease caused by Leishmania donovani , an intracellular parasite transmitted in the Indian continent by Phlebotomus argentipes . The reported number of cases in the region was 100 , 000 per year in 2005 , but the actual figure may be 5 to 8 times higher [1]–[3] . However , most of L . donovani infections remain asymptomatic [4] . Cross-sectional surveys based on serological testing by Direct Agglutination Test ( DAT ) [5]–[9] or ELISA [9] and/or positive delayed-type hypersensitivity ( DTH ) reaction to a leishmanin skin test ( LST ) [10]–[12] show high proportions of positive persons who never reported clinical disease . It is unclear whether these asymptomatic infected persons are infectious to the sandfly vector , whether they acquire persistent immunity or develop VL later on . The proportion of L . donovani infections that result in VL disease is poorly documented as this requires large prospective epidemiological studies . In Brazil the ratio of incident infection to incident disease ranged between 6 . 5∶1 and 18 . 5∶1 [13]–[15] , whereas in Africa ratios ranging from 1∶2 . 4 to 11∶1 have been reported [16]–[19] . In the only population-based longitudinal study in South-East Asia published so far Bern et al . found a 4∶1 ratio of incident infection versus disease in Bangladesh [20] . Similar estimates are not yet available for India and Nepal , two other countries affected by VL in the Indian subcontinent . The objective of this study was to examine the relationship between L . donovani infection and clinical disease , and to estimate the probability of progressing to clinical VL in recently infected persons in India and Nepal .
A large community intervention study to measure effectiveness of long-lasting insecticide treated bednets ( LN ) for prevention of VL in India and Nepal ( KALANET , ClinicalTrials . gov NCT00318721 ) provided the opportunity to document incident infection and disease in Nepal and in India . Consent for the Kalanet study was sought at three levels: community , household and individual . Communities were duly informed about the purpose of the trial and consent was sought from village leaders for inclusion of the cluster in the trial . Written informed consent was obtained from all participants or their guardians for those under 18 years old . A literate witness signed on behalf of illiterate participants who added their thumbprint to the informed consent form . Ethical clearance was obtained from the ethical committees of BHU ( India ) , the BPKIHS ( Nepal ) , the London School of Hygiene and Tropical Medicine ( UK ) , and the University of Antwerp ( Belgium ) . The study ran from November 2006 to May 2009 in 26 highly endemic villages . Selection of those villages has been described elsewhere [21] , [22] . The cluster randomized controlled trial did not show any significant reduction in incidence of L . donovani infection or VL disease in the intervention villages compared to controls [23] . For this reason , the study subjects from both intervention and control villages were included in the current analysis . Cross-sectional serosurveys were conducted in November–December 2006 ( M0 ) , 2007 ( M12 ) and 2008 ( M24 ) in the 26 study villages . A blood sample was collected from all consenting individuals over 2 years old . Clinical follow-up was done through passive and active case detection up to May 2009 , i . e . 6 months after the last serosurvey . An incident VL case was defined as a subject with a clinical episode of VL for whom the first clinical symptoms had started after the baseline serosurvey ( November 2006 ) . From November 2006 to May 2009 , all subjects with fever lasting for two weeks or more were examined by a physician and tested with a rapid diagnostic test ( RDT ) for VL ( Kalazar Detect™ Rapid Test; InBios International , Seattle , WA ) . Most VL cases were managed free of charge at the reference treatment centers i . e . Kala-azar Medical Research Center , Muzaffarpur , India ( KAMRC ) and B . P . Koirala Institute of Health Sciences , Dharan , Nepal ( BPKIHS ) where diagnosis was further confirmed by direct microscopic examination and/or culture of bone marrow or spleen tissue aspirate . For those who were diagnosed and treated outside the two reference centers details on diagnosis and treatment were collected from the subject or his family , and double-checked with clinical records . Direct agglutination test ( DAT ) was used as a marker of L . donovani infection . We used a cut-off titre of 1∶1600 to define seropositivity . This cut-off is lower than the one used for VL diagnosis in clinical suspects ( 1∶3200 ) because we wanted to increase the sensitivity to detect L . donovani infection [24] , [25] . Seroconversion was defined as a titer increase of 2 titres or more above his/her baseline value , since a difference of one titer is considered a very common ( and accepted ) inter-observer discrepancy in routine DAT serology reading [26] . In each serosurvey , capillary blood samples were collected from all participants on Whatman 3 filter paper . DAT was performed in the laboratories of Banaras Hindu University ( BHU -Varanasi ) for India , and in BPKIHS for Nepal , using a freeze-dried version of DAT antigen composed of fixed , trypsin-treated and stained promastigotes of L . donovani prepared in ITM-A [27] . Ten per cent of the samples were repeated in the partner laboratory for quality control . DAT was conducted as described by Harith et al . [28] . Briefly , disks of 5 mm diameter containing 50 µl of blood were punched out of the filter paper and eluted overnight in a 1000 µl tube at 4°C in Phosphate Buffered Saline ( PBS – 7 . 2 ) supplemented with protein , to obtain a starting dilution of 1∶400 . Serial dilution of the eluate in 0 . 9% saline , 1% foetal bovine serum and 0 . 24 ml 2-mercaptoethanol were made in V-shaped well microtiter plates , giving a range from 1∶400 up to 1∶25600 . Results were read visually against a white background after 18 h . In order to allow comparison of our results with publications from other VL endemic areas we calculated the relation between infection and disease in two ways: 1 ) as a ratio of number of incident infections to incident VL cases in a given time period , and 2 ) as a rate ratio . The ratio of infection to clinical disease was calculated comparing the number of asymptomatic seroconversions with the number of incident VL cases in a given period . The ratio of the infection rates was calculated using the incidence rates of infection ( seroconversion ) and clinical VL over the available time period of observation ( 2 years for incident infection , two and a half years for incident VL ) . Risk of progressing to clinical VL in recently infected asymptomatic persons was calculated by comparing the cumulative incidence of VL in the one and a half year of follow-up in the cohort of seroconverters of year 1 , with the cumulative incidence of VL in those who had not seroconverted in the same year 1 . We also calculated two other incidence rates of interest: i ) the incidence rate of VL over the 2 . 5 years in the full population ( instead of the cohort with full serological data only ) , and ii ) the incidence rate of VL in year 1 in those who were DAT-positive at baseline . Confidence intervals were calculated using 95% CI formula for rates , and Pearson's χ2 for CI on rate ratios . The Kalanet study was funded by the European Commission ( contract No INCO-CT 2005-01537 , Kalanet project ) . The funders had no role in study design , data collection and analysis , decision to publish , or preparation of the manuscript .
The total study population was 21 , 267 of which 49 . 1% were women; 17 . 0% were under five years of age , 42 . 6% under fifteen . Details on the study populations in India and Nepal are provided elsewhere [21]–[23] . In brief , individual ( i . e . age and gender distribution ) and socio-economic characteristics were similar in both countries . The main difference was that the prevalence of DAT-positive results in those with no VL history at baseline was almost twice as high in India as it was in Nepal ( 18 vs 9% ) . In the study population with no VL-history ( N = 20071 ) and regardless of serostatus , 120 new VL cases were diagnosed during the study ( VL relapses excluded ) , or an overall incidence rate of VL of 2 . 4/1000 PY ( 95% CI 1 . 9–2 . 9 ) . 1 , 196 had had VL in the past and 90 . 6% of them were DAT positive . From those without prior VL and who provided a sample at baseline , 1 , 159 ( 9 . 25% ) were DAT-positive and 11 , 374 ( 90 . 75% ) were DAT-negative . A cohort of 9 , 034 baseline DAT-negative individuals with full clinical and serological dataset was retained for the infection∶disease ratio analysis , and for the calculation of the incidence rate of seroconversion . Participants' flow ( exclusion/inclusion ) is summarized in Figure 1 . Clinical and serological outcomes of the 9034 DAT-negative persons at baseline are detailed in Figure 2 . Sixty three persons developed VL in the course of the study , 42 in the first year ( Nov 2006 to Oct 2007 ) , 19 in the second ( Nov 2007 to Oct 2008 ) and 2 in the six months of follow up after the last serosurvey ( Nov 2008 to May 2009 ) giving an average incidence rate of 2 . 8/1000 person years ( PY ) ( 95% CI: 2 . 1–3 . 5 ) . Of the 61 VL cases occurring before the final serosurvey ( Nov 2008 ) , 59 had a documented seroconversion while 2 remained DAT-negative . These 2 cases were young children ( 5 and 6 years old ) suffering from malnutrition . Both cases had a RDT positive test at the time of diagnosis but DAT-negative results at the serosurveys that preceded and followed their VL episode . All calculations of incidence rates , per year , for disease and infection are shown in Table 1 . Results per country are provided in Table S1 . In the first year , 42 individuals in the cohort developed VL , 375 showed an asymptomatic seroconversion and 8617 kept their seronegative status . The infection∶disease ratio was thus 8 . 9 to 1 ( Table 1 ) . In the follow-up period ( 18 months ) , 7 of the 375 latently infected individuals developed VL , giving an incidence rate of 12 . 44/1000 PY ( 95% CI: 3 . 25–21 . 80 ) in this group . In the 8617 with no seroconversion or VL recorded in the first year , 14 new VL cases occurred in the follow-up period , giving an incidence rate of 1 . 1/1000 person years ( 95% CI: 0 . 23–1 . 94 ) . By the time of the third serosurvey ( November–December 2008 ) , 85% ( 318 ) of the asymptomatic seroconverters had turned seronegative again . In year 2 ( 12 months' period ) the infection∶disease ratio was 130 asymptomatic seroconverters against 12 new VL cases , or 10 . 8∶1 . The subsequent follow-up period ( 6 months ) was too short to calculate VL incidences in this group of seroconverters . Incidence rate of DAT seroconversion was 46 . 0 per 1000 person-years ( PY ) in the first year ( 95% CI: 41 . 6–50 . 5 ) and 16 . 4 per 1000 PY in the second ( 95% CI 13 . 7–19 . 1 ) . Over the two years , the average infection incidence rate was 31 . 6/1000 PY ( 95% CI 28 . 9–34 . 2 ) . Incidence rate of VL was 4 . 6/1000 PY in the first year ( 95% CI 3 . 2–6 . 1 ) and 2 . 1/1000 PY in the second ( 95% CI 1 . 2–3 . 1 ) . The average incidence rate of VL over 30 months in this cohort of 9034 was 2 . 8/1000 PY ( 95% CI , 2 . 1–3 . 5 ) . The ratio between the incidence rates of infection and disease in the cohort was thus 11 . 3 to 1 ( 31 . 6 divided by 2 . 8 ) . Recent seroconversion was a strong risk factor for the development of VL in these high endemic villages . Incidence of VL in year 2 in persons with recent seroconversion was 11 . 5 times higher than in the rest of the village population ( 12 . 4 vs 1 . 1/1000 PY = 4 . 66<RR<28 . 30 ) . Incidence of VL in year 1 in those who were already DAT positive at baseline ( data not shown here ) was 15 . 9/1000 PY ( 95% CI 8 . 5–23 . 2 ) , more than 3 times the 4 . 6/1000 PY incidence rate found in year 1 in individuals with negative DAT at baseline ( Relative risk 3 . 42 ( 1 . 97<RR<5 . 92; p<0 . 0001 ) ) .
Using DAT seroconversion as a marker of infection , we found incident asymptomatic infection to be eight times more frequent than incident VL disease in India and Nepal , and about 1 in 50 of these latent infections lead to VL in the next 18 months . Asymptomatic DAT-positivity detected through screening in a person with no history of VL , and especially in case of documented recent seroconversion , is a risk factor for ultimately developing VL . Further studies on latent infection are needed to better understand the serokinetics , and possibly identify markers for progression to VL . Such studies should ideally combine DTH tests , PCR and different serological tests to measure levels and timing of humoral and cellular response after exposure . | Visceral Leishmaniasis is well known as a public health problem in North-Indian Bihar state and adjacent districts in Nepal , with about 300 . 000 new cases per year . As not all infections with L . donovani lead to disease , the impact of control programs should not only be measured in numbers of VL cases , but also in number of new infections . So far there have been little or no data on the infection∶disease ratio for this region , and the evolution of these latent infections . Using DAT seroconversion as a marker of infection , we found incident asymptomatic infection to be nine times more frequent than incident VL disease in high-endemic villages in India and Nepal , and about 1 in 50 of these latent infections lead to VL in the next 18 months , while over 80% turn seronegative again within a year . Asymptomatic DAT-positivity detected through screening in a person with no history of VL , and especially in case of documented recent seroconversion , is a risk factor for ultimately developing VL . Further studies on transient and persistent asymptomatic L . donovani infection are needed to better understand their immunological patterns and serokinetics , their level of infectivity , and their potential for later progression to VL . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"leishmaniasis",
"neglected",
"tropical",
"diseases"
] | 2011 | Incidence of Symptomatic and Asymptomatic Leishmania donovani Infections in High-Endemic Foci in India and Nepal: A Prospective Study |
Pluripotent embryonic stem cells are of paramount importance for biomedical sciences because of their innate ability for self-renewal and differentiation into all major cell lines . The fateful decision to exit or remain in the pluripotent state is regulated by complex genetic regulatory networks . The rapid growth of single-cell sequencing data has greatly stimulated applications of statistical and machine learning methods for inferring topologies of pluripotency regulating genetic networks . The inferred network topologies , however , often only encode Boolean information while remaining silent about the roles of dynamics and molecular stochasticity inherent in gene expression . Herein we develop a framework for systematically extending Boolean-level network topologies into higher resolution models of networks which explicitly account for the promoter architectures and gene state switching dynamics . We show the framework to be useful for disentangling the various contributions that gene switching , external signaling , and network topology make to the global heterogeneity and dynamics of transcription factor populations . We find the pluripotent state of the network to be a steady state which is robust to global variations of gene switching rates which we argue are a good proxy for epigenetic states of individual promoters . The temporal dynamics of exiting the pluripotent state , on the other hand , is significantly influenced by the rates of genetic switching which makes cells more responsive to changes in extracellular signals .
Embryonic stem cells derived from mammalian blastocysts are pluripotent: they show an indefinite capacity for self-renewal and the ability to differentiate into every cell type constituting an adult organism [1–3] . The development of healthy tissues hinges on the ability of these pluripotent stem cells to make critical decisions determining when and into which kind of cells to differentiate in response to the environment . It is therefore unsurprising that fates of embryonic cells are decided through sophisticated biological computations by a tightly-integrated regulatory network consisting of genetic , epigenetic and signaling layers [3] . The expression of genes is subject to intrinsic noise due to finite molecular copy numbers [4] and due to randomness in extracellular environment . Thus , while at the level of the organism development is often predictable , with a well-defined order of events , at the level of single cells fate determination is fundamentally stochastic [5 , 6] . Indeed , many studies probing the transcription of pluripotency-regulating genes in single cells have found high variability in distributions of transcription factors and mRNA molecules [7–9] . Several hypotheses about the functional roles for the highly heterogeneous expression of pluripotency transcription factors ( TFs ) have been put forward . One idea is that stochastic excursions in the population levels of transcription factors help by steering cells towards distinct states [10 , 11] . A different hypothesis claims transcriptional noise to be advantageous by facilitating the exploration of the state space of a gene network such that , at any instant , a sub-population of cells is optimally primed to be responsive to differentiation signals [12] . The heterogeneity of populations of pluripotent cells has also raised some concerns that pluripotency is ill-defined at a single-cell level [12] and instead should be viewed as a macroscopic state emerginging at the level of an ensemble of cells . A comprehensive physical picture of pluripotency at the single-cell level therefore remains unclear . In this regard , the roles of modeling and computational approaches are seen as especially important for bridging the gap between our understanding of molecular dynamics of regulatory networks and phenotypic outcomes . A rapid growth in single-cell sequencing data has opened many avenues for carrying out statistical inferences of pluripotency regulating genetic networks [13–15] . In vitro studies of mouse embryonic stem cells ( mESC ) in different culture conditions , in particular , have become an ideal model system for computationally inferring gene networks and exploring mechanistic issues surrounding pluripotency and lineage commitment [15] . In a tour de force study of mESC by Dunn et al . [16] , regulatory relationships between transcription factors were uncovered through analysis of pairwise correlations in gene expression . Using mean values of RNA-seq counts as constraints , a minimal network topology was derived showing a high degree of predictive power with respect to perturbations of the network , such as gene knockouts . Network topologies in general , however , remain silent about the roles of molecular noise and dynamics in stem cell differentiation governed by stochastic biochemical reactions . Furthermore , in order to validate that the inferred network reflects true microscopic reality of cell and is not a result of overfitting , one has to ultimately test the results using mass-action-based kinetics which integrate relevant molecular factors . The key challenge lies in finding the adequate resolution for the network which is able to be predictive and does not pose insurmountable computational burden . In the present work , we outline a framework for extending Boolean resolution networks—a commonly derived product from high-throughput experimental results—into stochastic and dynamical models with microscopic resolution of promoter architecture and individual gene-switching events . The computational scheme utilizes static Boolean information about the network topology and uses novel analytical model reduction to increase the computational efficiency , allowing for extensive searches in the space of microscopic reaction rates . This framework is successfully applied to the network topology inferred by Dunn et al . [16] in order to build a mass action based stochastic dynamic model which is capable of describing both the discrete states of all the genes and the populations of transcription factors . Starting from minimal assumptions about the rates of various reactions , we find a parameter regimes where a remarkable agreement with the experimental gene expression profiles [16] is achieved for all combinations of external signals . We show that average gene-expression levels in complex regulatory networks are not a unique function of gene-switching rates , which cautions against over-interpreting Boolean-level networks and suggests strategies of inference which utilize higher moments in distribution of transcription factors . Using single cell experiments which have probed expression of pluripotency factors [7 , 8] , we are able to argue that gene switching in pluripotent states happens primarily on the intermediate scale relative to the reactions of production and degradation ( dilution ) . This regime better agrees with the diverse set of experiments available [7–9 , 16] and provides explanation for the multimodality in distributions of transcription factors and burst like expression dynamics for some genes . In the second half of the paper , armed with a predictive and physically motivated model of pluripotency network , we explore the dynamics of lineage commitment driven by withdrawal of various well documented signals ( LIF , 2i ) for maintaining the naïve state of pluripotency . We find a number of non-trivial consequences of molecular noise and gene-switching dynamics . Taking gene switching rates as a proxy for epigenetic states of individual promoters we show that global variations of gene switching rates ( mimicking global remodeling of epigenetic marks ) yields a significant leverage over stability , sensitivity and exiting dynamics of pluripotent states In particular we show that intermediate gene-switching regime generates higher sensitivity for the network when responding to external signals .
In this Section we outline the steps for constructing higher-resolution regulatory networks with explicit promoter architecture , gene states , and transcription factor copy numbers , starting from experimentally-inferred Boolean network topologies . To illustrate the potential of our modeling framework we have chosen the most comprehensive Boolean network to date which describes the regulation of pluripotency factors of mouse embryonic stem cells [16] . In the next subsection we describe how Boolean logic is converted into the molecular logic of promoter states via a set of simplifying assumptions . Once promoter logic is defined , in the second subsections we go on to define the resolution of protein to promoter interactions . These assumptions lead to a set of chemical reactions describing transitions between gene states . In the third subsection we introduce the rules defining the production and degradation of transcription factors . In the fourth subsection we give an overview of the multi-scale simulation techniques appropriate for the simulation of stochastic switching of genetic states with a single-molecule-level resolution . In particular we contrast the rigorous yet computationally inefficient individual-based model ( IB ) with a more efficient piecewise-deterministic Markov Process ( PDMP ) which we adopt and extensively validate in the present work . The Boolean model of Dunn et al . specifies twelve genes in the network . We use Gi to denote a gene and Pi to denote the corresponding functional TFs ( i ∈ {1 , 2 , … , 12} ) in the following framework . The network topology ( Fig 1 ) contains static information about the types of interactions between pairs of genes . The interactions are classified as being either repressing or activating . To study the dynamics of complex genetic networks , one has to extend the Boolean-level description to account for the molecular logic of gene regulation . This molecular logic specifies the precise relation between the binding of transcription factors to the promoters and the regulatory outcome in terms of gene activation or repression . In the case where the same sites can be bound to different transcription factors , the combinatorial nature of regulation can give rise to ambiguity in molecular logic . Indeed , even on the level of Boolean networks , the logic is ambiguous and many possible truth tables have to be enumerated in order to select an appropriate picture for the dependence of a gene on its regulating signals . Ideally , such ambiguities should be resolved by directly inferring regulatory logic . Alternatively , one could simulate different combinatorial possibilities until sufficient agreement with experiments is reached . In our case , we assume a set of rules which applies to every gene in the network . This is a simplifying assumption , the justification for which is obtained a posteriori; networks with optimized parameters yield gene expression patterns consistent with the experiments . The fact that these simplifying assumptions work surprisingly well implies a dominance of the global network topology over the local details of molecular regulatory logic of gene promoters . For each gene , we assume a fixed number N of promoter sites . The genes which are only regulated by activators are always “OFF” unless N promoter sites are bound to activators . Similarly , the genes which are only regulated by repressors are always “ON” unless N promoter sites are bound to repressors . For those genes which are regulated by both activators and repressors , their activity is up-regulated to “ON” ( resp . down-regulated to “OFF” ) only when N promoter sites are bound to activators ( resp . repressors ) , otherwise they have a “MEDIUM” activity . These rules are illustrated quantitatively in the right panel of Fig 1 . From the topology , we know that each gene is regulated ( activated or repressed ) by a subset of TFs; we denote the subset of the TFs which regulates gene i by Si ∈ {P1 , P2 , … , P12} . For example , the subset of the TFs which regulates the core gene Nanog is the set of transcription factors Klf2 , Sox2 , MEKERK ( see Fig 1 ) . The binding and unbinding of the transcription factors to the promoter sites of gene i can be summarized by the following binding and unbinding reactions: G i + P j ⇌ k off N k on Ω − 1 G i P j for P j ∈ S i , ( first TF ) ⋮ G i P j … P k + P l ⇌ N k off k on Ω − 1 G i P j … P k P l for P l ∈ S i , ( first TF ) ( 1 ) where the rates are written per the reactants , i . e . , a given Pj binds to a given unbound promoter Gi with rate NkonΩ−1 etc . The “system size” parameter Ω has been introduced which sets the scale for a typical number of TFs in the system [4 , 17] without loss of generality . The scaling of these switching reactions is chosen such that the time scale of gene switching is independent of Ω [18] . Notice that we have adopted the parallel binding mechanism [19] which assumes the transcription factors bind to any of the unbound promoter sites independently with a rate constant kon . Each gene Gi “produces” its own transcription factor Pi with a rate which depends on the promoter state of Gi . We use a one-step model of transcription [20] whereby transcription factors are produced via a unimolecular step . This approximation coarse grains several sequential intermediate reactions ( polymerase recruitment , mRNA production , RNA splicing , etc . ) into one effective step [21 , 22] . This approximation has become popular in models of protein production [18 , 23 , 24] . All of the transcription factors are assumed to have finite lifetimes set by the rate of degradation , thus ensuring the existence of stable steady states with a finite number of molecules . The absolute time is set to be ∼ 1 hr which is consistent with experimental measurements reported on the lifetimes of pluripotency transcription factors [25–28] . The reactions are G i → Ω α G i + P i , ( TF production ) P i → γ ⌀ . ( TF degradation ) ( 2 ) Here , the rate constant of the production of TFs ( Ωα ) depends on the promoter’s configuration and is determined by the molecular logic defined previously . We further assume the production rate when the gene is “ON” , “MEDIUM” , and “OFF” to be Ωαmax , Ωαm , and 0 ( see Fig 1 ) . These rates are taken to be identical among all the genes . We note that one can adopt a view with a higher resolution of the network depending on the available experimental data and the nature of the question posed . The computational framework can readily incorporate in an explicit manner more steps , for instance to model the effects of cell cycle regulation , different epigenetic states , and binding of non-coding RNAs . Herein we consider the most simplified dynamical model which describes only the promoter configurations and the population dynamics of the transcription factors . This model aims to use the optimal resolution for capturing trends in gene expression while remaining feasible for efficient stochastic simulations . After converting the Boolean topology into a higher resolution network of biochemical reactions , our next goal is to exhaustively sample a vast space of parameter space in order to identify the optimal parameter regimes with which the model best reproduces the experimental results . Biochemical reactions in gene networks are of a fundamentally probabilistic nature; this demands a stochastic description of the kinetics . The conventional mean field mass-action-based kinetics only describe the dynamical behavior in the thermodynamic limit , and ignore the stochastic effects which arise from the discreteness of the system . In our model , the system is highly discrete , as there is only a single copy of each gene , and their discrete promoter states dictate the dynamics of the transcription factors . The most rigorous way to simulate such a reaction system is by numerically solving the chemical master equation which accounts for all possible states of the network down to the level of single molecules [4 , 20] . In high dimensions , i . e . , when the number of species is large , this approach is not computationally efficient . Instead , kinetic Monte Carlo algorithms are the most straightforward way to generate sample paths of the stochastic process . We refer to the stochastic process modeling the reactions down to the single-molecule level with the well-mixed assumption as the individual-based model ( IB ) . The state of our individual-based model is characterized by the state of each gene’s promoter sites—how many sites are bound to specific kinds of transcription factors—and the integer populations of the transcription factors . The rates at which the process stochastically transitions from one state to another are specified by the reactions Eqs ( 1 ) and ( 2 ) . Fully individual-based models , however , still suffer from a steep scaling of computational time with the number of discrete system states . This fact renders them inefficient for simulating large gene networks , especially when it comes to scanning or exploring the parameter space . A wide range of approximate schemes have been employed to simulate complex gene regulatory networks [29–31] . Most conventional approximations so far have been the size-expansion methods [32 , 33] which are known for being problematic when the molecular noise induced by discrete genetic switching becomes non-negligible [18 , 34 , 35] . On the other hand , for embryonic stem cell networks it is essential to account for the stochastic nature of genetic switching events which give rise to multimodality in the probability distributions of transcription factors populations . These local maxima in the multidimensional probability distributions of a stochastic genetic network correspond to metastable states and are also known as local attractors which corresponding to distinct promoter configurations and hence likely also to distinct phenotypic states [36–38] . Theoretically a network with M independently functioning genes can generate up to ∼ 2M distinct phenotypic states [39] . Thus , even if populations of all the other species are present in large quantities , the stochastic fluctuations caused by the genetic switches ( due to stochastic binding-unbinding events of the transcriptional factors to a discrete number of promoter sites ) between ON , MEDIUM , and OFF states cannot be ignored , unless the switching is operating in the extremely fast limit compared to any other reactions , known as the “adiabatic regime” [35 , 39 , 40] . In the other cases—the non-adiabatic regime—gene switching can completely dominate the dynamics in the network [35 , 41] . Eukaryotic gene regulatory networks are often found in the intermediate regime where gene-switching events are dynamically interwoven with the rest of the reactions in the network and cannot be ignored [5 , 42 , 43] . Single-cell studies of mESC in particular have shown bursty behavior in gene expression with sudden jumps in the levels of proteins resulting in multi-modal distributions of core transcription factors [7] . The sheer volume and complexity of information that is emerging from experiments on embryonic stem cells have motivated the application of a wide variety of modeling strategies for confronting the observed patterns of gene expression in ESCs [44 , 45] . Some of the computational techniques used so far include Boolean networks [16 , 46–48] , Hopfield neural networks [49–51] , systems of coupled ordinary [52 , 53] and stochastic differential equations [10 , 54 , 55] with Hill coefficients , agent-based models [56] , individual-based models [43 , 57 , 58] and small noise approximations to individual-based models [59–61] . Many of the early computational models of embryonic stem cells focused on small fragments of pluripotency networks , typically involving bistable switches [10 , 52] . These early pioneering studies yielded many insights on stochastic decision making in regards to fate determination and self-renewal [44] . Relatively few studies have also looked at larger portions of regulatory networks while including a stochastic treatment of genetic switching dynamics by either carrying out individual-based simulations [43 , 57] or small noise approximations [59–62] . Because of the computationally demanding nature of individual-based models and the restricted validity of small noise approximations to the near-adiabatic gene switching , the full range of stochastic and dynamical regimes displayed by different gene switching time-scales has remained unexplored . Additionally with very few exceptions [43] , the kinetic parameters in many of the previous models have been not thoroughly explored or informed by data and have had to be selected from physical intuition alone . These limitations have now been overcome in the present work . Thanks to a series of recent developments in modeling of gene expression dynamics [18 , 23 , 24 , 35 , 63–65] , a novel computational framework utilizing piecewise-deterministic Markov processes ( PDMPs ) [66] has emerged as a rigorous approximation to the fully individual-based model . The idea behind PDMP is simple: reactions with large number of molecules are evolved deterministically , while reactions with smaller numbers of molecules are propagated as discrete , random switching events . This approach treats discrete genetic switching events exactly , while assuming noise due to the finite nature of populations of transcription factors to be small in comparison . As shown in the subsequent sections , the assumptions underlying the PDMP approach turn out to be sound as we go on to obtain a nearly perfect quantitative agreement with full blown kinetic Monte Carlo schemes even for the case of the complex networks of mESC operating in the intermediate gene-switching regime . The mathematical details of the PDMP are given in the next subsection . The PDMP simulations carried out in the present study show nearly O ( 10 3 ) fold faster generating stochastic trajectories compared to conventional individual-based kinetic Monte Carlo techniques ( the algorithm for generating PDMP sample paths is provided in the Method section ) . This rigorous and rapid sampling of gene-switching events has not only allowed us to investigate the stochastic dynamics of the regulatory network at a longer time scale compared to conventional kinetic Monte Carlo methods , but also enabled us to explore a vast parameter space efficiently . We have used the obtained information to derive microscopic resolution models of ESCs . The mathematical details of the PDMP described in greater detail can be found in the next few subsections . After converting the Boolean topology to a higher resolution genetic network we use a computationally efficient stochastic dynamics method ( PDMP in the present case ) to explore the kinetic parameter regimes . This is done in order to: i ) Identify all the unique steady states ii ) Obtain a set of rate coefficients that best reflects the experimental constrains iii ) Study the response of the network to gene switching dynamics the rate of which is taken as a proxy for epigenetic states . The experimental data used for constraining rate coefficients are the binarized gene expression levels of pluripotency transcription factors under well-defined culture conditions consisting of different combinations of leukemia inhibitory factor ( LIF ) , glycogen synthase kinase 3 ( CH ) , and mitogen-activated protein kinase ( PD ) . Identical culture conditions were used by Dunn et al . [16] to infer the original Boolean topology . The optimization of rate coefficients is done by defining a uniform threshold η among all the TFs and with different external conditions; TFs whose population densities are higher than η are classified as expressed , while those below the threshold are not expressed . Then , we employ the Hamming distance ( the number of discrepancies between the simulated and experimental profiles ) as a cost-function for the optimization . The Hamming distance is minimized through multiple rounds of simulations where we vary the uniform threshold η and four free and non-dimensional model parameters ( see Method section ) : the number of promoter sites per gene N , the rate constant of TFs binding to the promoter kon , rate constant for a bound TF to dissociate from the promoter koff , and the production rate αm when the gene is in the “MEDIUM” state . Through this procedure , we find that the binary expression patterns can be closely reproduced ( minimal Hamming distance = 3 ) by the model when the number of the promoter sties N ≥ 2 . We remark that , given the small number of free parameters , this is a very small amount of deviation from the experimental results; the remaining deviations are likely to have been caused by the simplifications to the logic and dynamics in the construction of our individual-based model . The qualitative features for N = 2 , 3 , … 5 cases are similar ( see S2–S7 Figs in the Supporting Information ) . We chose to present the data for the simplest case N = 2 in the manuscript . We note that various other forms of cooperativity , including cases when bound TF at the promoter recruits other TFs ( either of its own kind or other type ) , can be readily incorporated into the current computational framework by changing the association rates kon as promoter-state dependent . For simplicity , we only illustrate the most basic form of cooperativity in this pilot paper . We find for any given N there exists a “valley” in the remaining model parameter space kon , koff , αm , as illustrated in Fig 2 ( A ) . The inferred parameter αm is consistently low ( ≲ 0 . 02 ) , meaning that the “MEDIUM” production rate is almost zero . This implies that the negative regulation of those genes which are regulated both by activators and repressors ( Tfcp2l1 , Esrrb , Nanog , and Oct4 ) in the model are technically fulfilled by inhibition ( i . e . , regulating TFs by blocking the promoter sites ) , instead of actually down-regulating the production activity . In addition , the valley suggests a relationship between kon and koff , where kon ≈ 10koff . This implies a certain asymmetric time scale between the binding and unbinding processes . We chose three parameter sets in this valley of cost function , corresponding to three distinct dynamical regimes of binding and unbinding reactions between the TFs and the promoter sites: slow ( kon = 3 . 2 , koff = 0 . 2 , αm = 0 . 02 ) , intermediate ( kon = 16 , koff = 1 . 5 , αm = 0 . 01 ) , and fast ( kon = 102 , koff = 10 , αm = 0 . 005 ) compared to the time scale of the TF dynamics . We remark that as we non-dimensionalize the model using the protein degradation rate γ ( see Method section ) , the time scale of the binding and unbinding are fast , intermediate , or slow relative relative to the time scale of the protein dynamics , also set by γ . See Eq 3 . All of the parameter sets successfully reproduce the experimental gene expression patterns ( Fig 2 ) corresponding to pluripotent and lineage committed cells . The fact that the rate parameters of the network occupy finite regions and cover different regimes suggests that distinct gene expression profiles can tolerate fluctuations in reaction rates . Such rate fluctuations , reflecting the effect of extrinsic noise , are inevitable in dynamic cellular environments of embryonic cells which experience frequent epigenetic and extracellular perturbations [15 , 26 , 67] . In a way , changes in gene-switching rates can be seen as a proxy for how global epigenetic changes govern the rates of transcription factor binding to target genomic regions . From the methodology point of view , the absence of a unique regime of rates implies the following: inferring networks using only mean levels of gene expression ( as is done for Boolean networks ) may lead to the loss of valuable information contained in higher moments of distribution . Thus new approaches of inference need to be developed in order to account for broad distributions of transcription factors . For this reason , we look beyond the comparisons of mean expression levels and turn to comparing stationary distributions of transcription factors observed in experiments with the computationally generated distributions in three chosen parameter regimes identified in Fig 2 . We generate the stationary distributions of the TF concentrations in the model , with eight different environmental conditions , in the specified three parameter sets in Fig 3 . We observe qualitatively that in the “fast” parameter set the distribution of each TF is unimodal , and in the “slow” parameter set , the distribution of each TF is peaked near “0” and “1” in non-dimensional units . In the intermediate regime , on the other hand , with some environmental conditions , broader distributions of the gene expression are observed . The pluripotent state of mESCs has been the subject of intense investigations by nucleic acid-based single-cell techniques such as RNA-seq , sm-FISH , qPCR and there is now extensive data on the steady-state distributions of RNA and transcription factors maintained under pluripotency-favoring culture conditions [7 , 8 , 68] . These experiments have revealed a heterogeneous nature of gene expression with many core pluripotency factors , such as Nanog and Esrrb , having long-tailed or bimodal distributions ( see for example Fig . 5a in Ref . [7] ) . Additionally , the single-cell stochastic trajectories of TFs have shown sharp , bursty transitions implying infrequent genetic switching events [7] . Qualitatively , these experimental observations are more consistent with the intermediate regime of genetic switching as seen from Fig 3 , as in the case of slow switching nearly all of the transcription factors express bimodality and in the case of fast switching the expressions of all the factors are narrow and unimodal . In Fig 3 , consistent with experiments , we also find that under LIF gene expression is more heterogeneous than under 2i , but also that the overall levels of expression are higher [7 , 8] . In all three regimes of genetic switching supporting pluripotency ( LIF+2i , LIF+PD , LIF+CH , LIF and 2i ) , core transcriptions factors such as Nanog , Oct4 , Sox2 are highly expressed . The same factors are also repressed in conditions favoring differentiation ( CH , PD and none ) irrespective of gene-switching regime . This shows that the pluripotent and differentiated states , as determined by the pattern of gene expression , are hardwired in the architecture of the genetic network independent of genetic switching rates . Nevertheless , as we will show in the next section , the routes and dynamics of lineage commitment from pluripotent states strongly depend on the level of molecular noise generated in different gene-switching regimes . Although the PDMP approach accurately captures the effects of genetic switching , it assumes demographic noise arising from finite populations to be negligible . To test the validity of this assumption and assess the contribution of different sources of noise in establishing the steady-state distribution of the pluripotent state , we carry out individual-based simulations for the intermediate switching regime of the network , Fig 4 . In the individual-based model all reactions are treated stochastically thereby accounting for all of the sources of noise in the system . The resulting gene expression profiles follow closely those obtained by PDMP simulations , showing that the noise arising from stochastic switching events of promoter configuration accounts for the significant part of overall variability in the network . Trajectories of individual transcription factors show that indeed most of the variance in the molecular distributions are generated by genetic switching events which appear as abrupt stochastic jumps . We next ask how the steady-state gene expression patterns displayed by the gene network respond to extracellular perturbations in the form of the initiation or termination of pluripotency signals . Both dual inhibitor 2i ( PD+CH ) and Leukemia factor LIF based signaling have been shown to provide a stable environment for maintaining pluripotency of stem cells in vitro [3 , 16] . Conversely , withdrawal of either LIF or 2i leads to irreversible lineage commitment after a 24-hour period . Despite a similar ability to guard pluripotent cells against lineage commitment , these factors deploy different regulatory mechanisms reflected in distinct distributions of pluripotency factors . As a result , stem cell differentiation by withdrawal of different signals proceeds via different routes . To gain a mechanistic understanding of how the interplay of signaling , molecular noise and network architecture gives rise to the steady-state expression profiles , we study the dynamics of transitioning between the pluripotent and lineage committed states induced by rapid initiation and withdrawal of signaling conditions ( LIF , CH , PD ) . The temporal evolution of distributions of the TFs exiting ( LIF/2i withdrawal ) and entering ( LIF/2i immersion ) pluripotent states reveals rich , dynamical signatures of these transitions ( Fig 5 ) . To reveal the role of stochasticity in these transitions , we compare the intermediate regime—which is dominated by genetic switching—to the fast regime—in which transitions are largely governed by the network topology and the fluctuation of promoter configuration is almost-completely ignored . The irreversible nature of transitions manifests clearly in different routes exiting and entering the pluripotent state ( transitions to and from the None state in Fig 5 ) . In the intermediate regime of genetic switching rates , expression noise greatly facilitates transitions out of a pluripotent state by making the network more responsive to changes in environmental signaling . In contrast , in the fast-switching regime , upon withdrawal of pluripotency signals the downstream regulation happens on a much slower time scale with some factors remaining virtually unresponsive to changes of signaling . This signaling enhancement in the intermediate regime reveals the importance of molecular noise in making pluripotent states more sensitive to environmental conditions . There are qualitatively different patterns of re-entrance into pluripotent states upon LIF vs 2i addition , with LIF being much more efficient at reversing pluripotency compared to the 2i . The different potentials of signaling cultures for pluripotency reversal has been established in experiments [3] which have shown that in the later stages of commitment only the LIF is able to reverse lineage-primed cells back to their naive pluripotent states . The exit and re-entrance from pluripotency upon withdrawal and addition of LIF shows complex signatures of hysteresis and bifurcations . This suggests that there can be multiple pathways of entering or exiting pluripotency . The transition times for all signaling-induced changes of the steady states of the network are visualized on a kinetic diagram , Fig 6 . This figure shows an underlying structure to these transitions where conversion among pluripotent states takes place with lower “activation barriers” compared to transitions accompanying loss of pluripotency . In the intermediate switching regime there is a clear time scale separation between transitions which keep cells in pluripotent states and transitions out of pluripotent states . One may argue that such a time scale separation between signals inducing differentiation and pluripotency allows embryonic cells to execute developmental decisions more faithfully . In the fast-switching regime this clear time scale separation is partially lost where only the loss of all three signals is separated from the rest of the transitions . Detailed analysis of individual distributions and trajectories of transcription factors can be very informative due to their high information content . It is , however , not immediately clear how changes in the expression of individual genes contribute to global changes corresponding to different phenotypic transitions . To reveal such global changes , we project stochastic trajectories of all transcription factors onto the first two eigenvectors obtained by principal component analysis ( PCA ) of the reference pluripotent steady state ( LIF+2i ) . Most of the variance of transcription factors is well captured by the few principal components . The high-dimensional steady state of the cellular network can thus be conveniently projected onto a 2-dimensional subspace , allowing us to visualize the attractor states of the network as probability landscapes π ( PC1 , PC2 ) which are often masked by heterogeneous distributions . Comparing these probability landscapes with different gene-switching regimes reveals the distinct roles played by gene switching-induced molecular noise and the deterministic network topology in guiding the transition out of the pluripotent states ( Fig 7 ) . The intermediate gene-switching regime , once again , appears to be the more viable regime underlying pluripotent states since the probability landscape shows up as a broad attractor with interconnected states . Going towards the limit of slow switching results in the fragmentation of the landscape into states separated by high barriers . This gene switching-induced remodeling of attractors shows the potential for regulation via global epigenetic changes which are purported to act via silencing or activating entire sets of genes at once [15] . Thus one may view gene-switching rates as a proxy for genome wide acetylation/methylation patterns which can dramatically alter the access of transcription factors to key target genomic sites . The sequential removal of pluripotency-inducing signals reveals a consistent change in the size of the attractor towards occupying smaller regions on the landscape . This argues for the physical state of the network corresponding to pluripotent states to be the one with maximal variance of regulatory transcription factors where lineage commitment is accompanied by their gradual constraining and repression . A similar idea which views pluripotency as a macrostate emerging from an ensemble of cells which try to maximizes the information entropy with respect to regulatory transcription factors has been postulated before [12 , 71] . The analysis of the steady-state stochastic dynamics of the pluripotency network in this work appears in agreement with this view . Furthermore , we are able to suggest a microscopic origin of this entropic paradigm . By analyzing pairwise correlations among different transcription factors we find that signals like LIF/2i create greater independence between the expression of core transcription factors , leading them to explore a larger range of values . Hence , removal of these signals leads to more constrained and interdependent patterns of gene expression for the same transcription factors which greatly diminishes overall variance ( see S1 Fig ) .
Experimental [16 , 51 , 72] and computational studies [54 , 56 , 73 , 74] of embryonic stem cell networks have increasingly emphasized a systems-level perspective where pluripotency is a result of sophisticated biological computations done by tightly integrated set of genes , epigenetic and transcription factors [16 , 75] . The rapid growth of gene-expression data on mESCs under many different in vitro conditions has enabled reliable inferences of the underlying topology of pluriptoency regulating networks [16] . At the same time single-cell probes of gene expression in mESCs reveal significant heterogeneity and dynamism in bio-molecular populations [7 , 8] , suggesting that molecular stochasticity and non-equilibrium processes could be playing a crucial role in regulating pluripotnecy and stem cell differentiation . Thus the schematic network topology models inferred from real data , while immensely useful , are not always sufficient for rationalizing the single cell data , which is rich with stochastic and dynamical information . On the other hand , full chemical master equation based stochastic simulations of complex ESC networks with complete bimolecular inventory quickly become impractical , especially when making data-driven parameter searches and explorative simulations with a large number of external conditions . In the present work we have developed a multi-scale computational scheme for converting experimentally-inferred Boolean topologies into quantitative and predictive models of networks with a microscopic resolution of gene expression dynamics . The employed computational model is based on previously proposed hybrid stochastic dynamics approaches [18 , 23 , 24 , 35] in which the switching dynamics of individual genes are considered exactly while the rest of the biochemical reactions are approximated as deterministic processes . This hybrid-stochastic approach is approximately thousand fold faster than conventional kinetic Monte Carlo methods . This gain in computational speed allows us to simulate large scale gene regulatory networks of ESC under different culture conditions and gene-switching regimes . Thanks to rapid hybrid simulations we are able to use a standard optimization approach to exhaustively sample the space of rates and find the closest match with the experimental gene expression data collected different culturing environments [16] . The approximation introduced due to using the hybrid scheme is validated by showing excellent quantitative agreement of both steady states distributions and dynamical transition times with the fully stochastic simulations for the identified parameter sets . This agreement also shows that the switching events of genes—due to stochastic TFs binding to the promoter sites—is likely a dominant source of variance in transcription factor populations of ESC networks . Furthermore we argue that changes in gene-switching rates are a proxy for global epigenetic modifications which can alter the rates of access of transcription factors to sites buried under chromatin structures [76] . Thus , the significant remodeling of steady-state landscapes that we see by varying the global gene-switching rates suggests a powerful role for the global epigenetic changes in maintaining the stability of pluripotent states . We find that the intermediate regime , in which the gene-switching rate is comparable to the other reaction rates in the network , is most consistent with single-cell measurements [7–9] . This result has also been pointed out by Sasai et al . [43] when exploring time-scale hierarchy in stem cell network with individual-based models and concluding that experimentally observed phenotypic heterogeneity likely originates from promoter reorganization and genetic switching taking place on the comparable time-scale with the rest of the biochemical processes . In this intermediate switching regime the transcription factors show bursty dynamics which lead to heterogeneous distributions with some showing long-tailed and bimodal features . Consistent with many experiments [7–9] , our simulations show that the presence of LIF and 2i signals is crucial for maintaining the stability of pluripotent states , which in our model are defined as states with an up-regulated triad of pluripotency factors Nanog/Oct4/Sox2 . Withdrawal of either LIF/2i initiates lineage commitment via a robust pattern of reduced expression of the Nanog/Oct4/Sox2 triad in the simulations . To characterize the dynamics of lineage commitment , we have computed the transition times of going from pluripotent to differentiated steady states . We find that slower gene switching generates more heterogeneous distributions of transcription factors , which in turn makes the network more responsive to changes in signaling conditions . For instance , the response time to LIF/2i withdrawal is much faster in a more stochastic regime than in the more deterministic regime corresponding to faster gene switching time scales . Next , by carrying out principal component analysis on ensembles of gene expression profiles , we find a much simpler description of pluripotency and lineage commitment in terms of effective probability landscapes . As the signals safeguarding pluripotency are removed , these landscapes reveal a gradual narrowing of the steady-state attractor explored by the network . Thus , we see a hierarchical organization of differentiation landscapes where pluripotent states pose the largest attractor which is maintained through the extracellular signals and the molecular noise of gene switching . Given the rapid rise of information from high throughput single-cell nucleic acid based techniques ( RNA-seq , RNA-FISH , qPCR , etc . ) , we expect the microscopic resolution models , reported in the present work , to play important roles in bridging the systems-level behavior of genetic networks with the underlying molecular-level processes of binding , reaction and diffusion .
In the individual-based description of complex genetic networks studied in the present work , one models each individual reactive events as a Markov jump processes . The underlying master equation governing the Markvoian evolution of the entire network is analytically intractable and in general even numerical simulations quickly become computational inefficient once dimensionality of the system becomes too high [77] . Specifically what contributed to this inefficiency is the population scale of transcription factors for which it is common to have values on the order of Ω = 104 as is characteristic for biological cells . Thus the use of standard continuous-time Monte Carlo [17 , 78] sampling techniques becomes unfeasible especially if one wants to sample the kinetic parameter regimes for finding optimal set of rate coefficients . Fortunately the latest efforts of modeling gene expression dynamics [18 , 23 , 24 , 35 , 63 , 64 , 79] have lead to the emergence of a new class of techniques which are broadly based on using a piecewise-deterministic Markov process ( PDMP ) to approximate the individual-based model with a switching property . In this section , we briefly recapitulate the construction of the PDMP . A more thorough analysis can be found in the literature cited [18 , 23 , 24 , 35 , 63 , 64 , 79] . A PDMP is a process such that , in between discrete random switching events , the evolution of the process is deterministic . To construct the deterministic evolution of the TF populations , starting from the chemical master equations , we performed Kramers–Moyal expansion [77 , 80] in the population of TFs while maintaining the discreteness of the genetic state; we keep only the first order of the expansion . The result is a standard Liouville equation governing the deterministic flow of the distribution . The joint probability distribution of our model converges to the deterministic flow in a given genetic state and in the thermodynamic limit Ω → ∞ [80] . With the PDMP approach , the demographic noise originating from random birth-death events are neglected , so that the population density xi ( t ) of each TF evolves according to d d t x i ( t ) = α i - γ x i ( t ) , ( 3 ) where αi ∈ {0 , αm , αmax} is the production rate of the ith TF dependent on the ith gene’s configuration of promoter sites . While the evolution of the TF population density is deterministic , the binding and unbinding events of the regulating TFs to their target genes are still stochastic and formulated according to Eq 1 in the main text . We finally emphasize that the PDMP only retains the contribution of switching noise which arise from the discrete and stochastic binding and unbinding events between the TFs and the promoter sites , and ignores demographic stochasticity from the discrete production and degradation processes of the TFs . The PDMP is the limiting process when the population scale Ω → ∞ [18] , and the error bound of the description can be rigorously derived to be O ( Ω - 1 ) [81] . To simulate the stochastic binding and unbinding statistics of the promoter sites , accurate waiting times must be generated . A waiting time exists for each possible stochastic transition; the smallest of these times tells us how long the system stays in the current configuration of promoter sites , and to which promoter configuration it transitions . In general , waiting times can be generated by mapping a uniform random variable to a random time using the survival function . Since in our case the transition rates are functions of dynamical state variables , this involves the numerical integration of survival functions describing each potential transition [63] . In our case , the simple form of Eq 3 ( and thus of the transition rates ) allows us to improve upon this by generating waiting times without numerical integration , detailed in the Supplementary Information . Under the assumptions we proposed , there are initially six free model parameters: Ωαmax and Ωαm as the production rates when each of the genes has an “ON” or “MEDIUM” activity , γ as the protein degradation rate , N as the number of promoter sites , and lastly konΩ−1 and koff as the binding and unbinding rates between the TFs and the promoter sties . We remark that the population scale Ω is fixed at 104 . Through suitable non-dimensionalization of the physical time and concentrations of the TFs , we reduce the number of parameters . As can be seen from the above formulation ( Eq 3 ) , the time scale of the TF dynamics is set by the degradation rate γ . For stable proteins , the time scale of degradation is of the order of the times of the cell cycle . We therefore choose the unit of physical time such that γ is 1 . Similarly , the maximum concentration in the TF can achieve in Eq 3 is αmax/γ . We can choose a unit for the concentrations of the chemical species such that αmax = 1 , so the concentration of the TFs are always bounded in ( 0 , 1 ) . After non-dimensionalization , the model ends up with four free parameters: αm ∈ {0 , 1} as the intermediate production rate of those genes which are regulated by both activators and repressors , kon , koff as the binding and unbinding rate of the TF to the promoter sites , and N as the number of promoter sites per gene . We use these non-dimensionalized parameters to report our results in the manuscript . To narrow down the parameter regime , we match our model predictions to the experimental findings of Dunn et al . [16] in which the authors measured the TF expression under various combinations of external signals , i . e . , LIF , CH , and PD . We aim to match the model prediction to a twelve-by-five “checkerboard diagram” which records the experimentally measured expression pattern presented in Fig 2 in the main text . To achieve this goal , we performed a sweep in a vast parameter space: αm ∈ [0 , 1] , kon , koff ∈ [0 , 110] , and N = 1 , 2 … 5 . For each parameter set , we simulated 103 PDMP sample paths for a time to sufficiently reflect the stationary state , and the average TF expression levels were recorded . Because of the non-dimensionalization , the expression level ( the population density ) of each TF is a real number in between 0 and 1 . This results in a twelve-by-five real-valued matrix , which is binarized by a threshold . To find the optimal threshold , we use the number of discrepancies between the model prediction and the target matrix—the Hamming distance—as a quantitative measure . For each parameter set , an optimal threshold which minimizes the Hamming distance was then found computationally , and the minimal Hamming distance was recorded and plotted in Fig 2 in the main text as a “landscape” of how good the model captures the experimental results . We found that for N = 1 and N ≥ 2 , the global minimal Hamming distance is 5 and 3 respectively . We chose N = 2 to present our follow-up analysis , as it incorporates the capacity of modeling cooperative binding which is often modeled phenomenologically . We find the Hamming distance can be constantly as small as 3 in a vast region in the space of binding/unbinding rates when αm is small ( ≲ 0 . 02 ) . Therefore , in the manuscript we present the landscape of a fast switching regime kon ≈ 100 , an intermediate regime kon ≈ 15 and a slow switching regime kon ≈ 3 . For the three selected parameter sets , 104 sample paths of a fully individual-based model were generated by standard kinetic Monte Carlo simulations—namely Gillespie’s stochastic simulation algorithm ( SSA ) [17 , 78] . The population scale Ω for each TF is set to be 104 . A parallel analysis is carried out and the results are consistent with the predictions from using the PDMP . We report the results of the intermediate switching regime in Fig 4 of the main text . While the joint probability distributions are measured by kinetic Monte Carlo sampling , the dimensionality of the dynamical system is very high: each TF has a real-valued density , so that even if we marginalize over the genetic states the probability density is a 12-dimensional object . Although Fig 4 in the main text summarizes the marginal distributions of the real-valued TF density and contains rich information , it is desirable to visualize the results in a lower dimensional space to draw qualitative conclusions . To achieve this goal , we perform the standard principal component analysis [82] . We chose a baseline external condition to be LIF+2i; the first two principal components were computed . For the rest of the external conditions , the joint probability distributions are projected onto the plane spanned by these principal components; the results are presented in Fig 7 in the main text . To investigate dynamical transitions when the external driving conditions ( whether LIF , CH , and PD are present ) change , we prepare 105 independent sample paths with an initial external condition until the joint probability distribution converges to the stationary distribution . Then , the external condition is switched instantaneously to the second condition . We further evolve the dynamical system until stationarity for the second conditions is reached . The results are summarized in Fig 5 of the main text . To estimate the transition times between the stationary distributions with different external conditions , we measure the Jensen–Shannon distance of the marginal distribution of each TF density , at any given time during the transition to the final marginal distribution . We measure and report the first time when all 12 distances are below a threshold value of 0 . 3 , presented in Fig 6 of the main text . Both the PDMP and the IB models have been implemented in c++ , using the algorithm presented in the Supporting Information and the standard kinetic Monte Carlo algorithm [17 , 78] respectively . The network topology was hard-coded in the simulators to improve the simulation efficiency . The simulators are provided in the Supporting information . To facilitate the use of our computational work and ensure full reproducibility , the IB model is translated into standardized BioNetGen language [83 , 84] and provided in Supporting Information , along with the standard SBML format of the model exported by BNGL . | In the embryonic stage mammalian cells are pluripotent: they have not yet committed to any specific cell type . The commitment to a cell type is controlled by pluripotency networks , the bio-molecular inventory which is unsurprisingly complex , spanning a myriad of transcription factors , genes and epigenetic factors . Thanks to advances in high-throughput sequencing and related computational tools for data analysis , we are beginning to unravel basic topological features of pluripotency networks . Networks inferred from sequencing data are often cast in Boolean representations which specify the existence and nature of regulatory connections between pairs of biomolecules . While immensely useful , the Boolean networks remain silent about the stochasticity and dynamics of molecules that are indelible features of bio-molecular life inside cells . Understanding stochastic and dynamic features of pluripotency networks is crucial if one is to have a mechanistic understanding of cell fate determination rationalized in terms of fundamental physico-chemical processes . The computational framework proposed in this work offers a way of bridging the divide between Boolean networks and higher resolution views of networks in a predictive and quantitative manner . The usefulness of the framework is demonstrated by recapitulating a number of experimental trends and creating new insights about the stochastic and dynamical nature of pluripotency . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetic",
"networks",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"signaling",
"networks",
"dna",
"transcription",
"stem",
"cells",
"network",
"analysis",
"transcription",
"factors",
"computer",
"and",
"information",
"sciences",
"cell",
"... | 2018 | A stochastic and dynamical view of pluripotency in mouse embryonic stem cells |
Eukaryotic cells extend a variety of surface protrusions to direct cell motility . Formation of protrusions is mediated by coordinated actions between the plasma membrane and the underlying actin cytoskeleton . Here , we found that the single calponin homology ( CH ) domain-containing protein CHDP-1 induces the formation of cell protrusions in C . elegans . CHDP-1 is anchored to the cortex through its amphipathic helix . CHDP-1 associates through its CH domain with the small GTPase Rac1/CED-10 , which is a key regulator of the actin cytoskeleton . CHDP-1 preferentially binds to the GTP-bound active form of the CED-10 protein and preserves the membrane localization of GTP-CED-10 . Hence , by coupling membrane expansion to Rac1-mediated actin dynamics , CHDP-1 promotes the formation of cellular protrusions in vivo .
In eukaryotic cells , the actin cytoskeleton forms a cortical shell around the cell periphery , which allows the plasma membrane to undergo frequent structural changes . The Rho family of small GTPases generally act on membranes and affect the movement of these membranes by changing the membrane-associated actin cytoskeleton . Rac1 and Cdc42 induce plasma membrane protrusions , including lamellipodia and filopodia , while RhoA activation specifically stimulates the formation of actin stress fibers [1] . Through the multi-protein WAVE complex , Rac1 indirectly stimulates the formation of new actin filaments [2 , 3] . Alternatively , Rac1 and CDC42 target PAK family members , which in turn phosphorylate LIMKs ( LIM domain kinases ) , and ultimately act on the actin-severing factor cofilin [4–6] . RhoA often acts as an antagonist of Rac1 and Cdc42 , and inhibition of either RhoA or its effector ROCK can result in Rac1 activation [7–9] . The activity of Rho GTPases is strictly controlled by various regulatory proteins , including guanine nucleotide exchange factors ( GEFs ) , which convert the inactive GDP-bound enzymes into active GTP-bound forms , and GTPase-activating proteins ( GAPs ) , which stimulate GTP hydrolysis , thereby converting the active GTPases into their inactive forms . Cell motility , which involves a continuous reorganization of the cytoskeleton , must be accompanied by appropriate restructuring of the plasma membrane . Indeed , extensive research has identified a diverse array of actin regulatory proteins that are directly associated with the lipid bilayer of the plasma membrane , including vinculin [10 , 11] , talin [12 , 13] , α-actinin [10 , 14] and cofilin [15] . The plasma membrane is rich in anionic phospholipids , which form electrostatic interactions with cytosolic proteins . Among the structural domains that bind the membrane without specificity for particular phospholipids , the amphipathic helix and Bin/amphiphysin/Rvs ( BAR ) domains have membrane deformation activity [16 , 17] . The amphipathic helix domain deforms the membrane by inserting an amphipathic helix or a hydrophobic loop into the inner layer of the plasma membrane and forming additional electrostatic interactions with anionic phospholipids in the cytosolic layer [18] . Proteins in the BAR domain superfamily use their positively charged , curved surface to induce membrane deformation with different degrees of curvature [16] . In particular , the I-BARs ( Inverse BAR domains ) can deform membranes in vitro into protrusion-like shapes [19 , 20] . However , it is still uncertain whether such membrane deformations actually take place in vivo . Calponin homology ( CH ) domains have been identified in proteins with a variety of functions ranging from actin binding to signaling [21 , 22] . The type I and type II CH domains together form the actin-binding region of a large number of F-actin-interacting proteins [23 , 24] . In contrast , calponin contains a single type III CH domain , which differs markedly from the type I and type II modules . The CH domain in calponin has been implicated in the association of multiple cytoskeleton-related components , but the actin-binding ability of type III CH domains in isolation is questionable [25] . Here , we discovered that the protein CHDP-1 , which contains a single type III CH domain and shares homology to calponin in mammals , plays a crucial role in protrusion formation in C . elegans . With its additional amphipathic helix motif , CHDP-1 localizes to the cell cortex and induces the formation of membrane protrusions . Furthermore , CHDP-1 binds to the GTP-bound active form of the Rac1 homolog CED-10 and facilitates the membrane association of the CED-10 protein . Our work reveals a novel regulatory component which couples membrane expansion to actin dynamics during protrusion formation in vivo .
BDU neurons are a pair of interneurons with cell bodies situated laterally in the anterior body of C . elegans . From its cell body , each BDU neuron projects an anterior process and a posterior process . The PLM neuron has its cell body located in the tail region and sends out an anterior process to the mid-body region of the animal . Our previous work demonstrated that the posterior process of BDU connects with the anterior process of PLM through gap junctions [26] ( Fig 1A and 1B ) . During the formation of the BDU-PLM connection , both the BDU and PLM neurons produce extensively elaborated membrane protrusions to reach each other and finally form the neuronal connection . From a genetic screen to search for mutants which display a BDU-PLM disconnection phenotype , we identified xd27 ( Fig 1C ) . In xd27 animals , the cell protrusions on both BDU and PLM cells are greatly reduced ( Fig 1D and 1E ) . Although a single neurite can project out from the posterior BDU or anterior PLM cell bodies in xd27 , the typical growth-cone-like membrane expansion is missing ( Fig 1E ) . Actin filament dynamics are essential for cell surface protrusion but are not necessary for axonal elongation , which is thought to be mainly mediated by microtubule extension [27] . Intriguingly , the neurites from both BDU and PLM in xd27 can continuously extend ( Fig 1F and 1G ) . In xd27 adults , the BDU neurite length is almost indistinguishable from wild type , while the PLM neurite is slightly shorter than wild type ( Fig 1H ) . However , when we followed neurite growth during larval stages , we found that the PLM neurite elongates at a similar speed in both xd27 and wild-type animals ( Fig 1I ) , suggesting that the formation of actin-mediated cell protrusions may be specifically affected by the xd27 mutation . In addition to BDU and PLM cells , the head neurons RMED and RMEV , and the D type motor neurons DDs and/or VDs , also display weak neurite extension defects ( S1 Fig ) . However , the locomotion of xd27 is generally normal . Besides the neural deficits , xd27 animals display partial embryonic lethality and weak egg laying and distal tip cell migration defects ( S1 Fig ) . Through genetic mapping and genomic DNA sequencing , we identified C10G11 . 7 ( Fig 2A ) . C10G11 . 7 encodes a single type III CH domain-containing protein , which we named CHDP-1 ( Fig 2A ) . It shares sequence similarity with calponin in mammals ( S1F Fig ) . In addition to the CH domain , CHDP-1 contains two proline-rich motifs ( P1 and P2 ) in the N-terminal region and one amphipathic helix motif ( Helix ) close to the C-terminus ( Fig 2A ) . Proline-rich motifs widely participate in protein-protein interactions [28] , while amphipathic helix motifs directly interact with membrane phospholipids [29] . A phenylalanine to leucine change at position 117 was identified in xd27 worms ( Fig 2A ) . Another chdp-1 allele , tm4947 , which removes the amphipathic helix domain and the majority of the CH domain ( Fig 2A ) and is likely a molecular null , displays a similar membrane protrusion defect to xd27 ( Fig 2B and 2C ) , suggesting that xd27 may also act as a strong loss-of-function or null mutation . Introducing a wild-type copy of the chdp-1 gene into xd27 or tm4947 animals restores the proper morphology of developing BDU and PLM cells ( Fig 2B and 2C ) . Thus , mutations of the chdp-1 gene are indeed responsible for the BDU-PLM connection defect . The chdp-1 gene is widely expressed in almost every cell during the early embryonic stage ( Fig 2D ) . From late embryonic to adult stages , however , chdp-1 expression was gradually restricted to certain sets of neurons including BDU , ALM , AVM , PLM , PVM and PVQ ( Fig 2E ) . When chdp-1 cDNA was expressed in BDU or PLM alone , or BDU and PLM together , the BDU-PLM connection defect in chdp-1 mutants was efficiently rescued ( Fig 2F ) . In contrast , expression of chdp-1 in hypodermis or muscle failed to rescue the BDU-PLM connection defect ( Fig 2F ) , suggesting that chdp-1 functions cell-autonomously to regulate protrusion formation . To understand how CHDP-1 regulates protrusion formation , we examined the subcellular localization of CHDP-1 in vivo . A functional translational CHDP-1 fusion green fluorescent protein ( GFP ) construct was introduced into worms . During the embryonic stage or after hatching , CHDP-1 is distributed specifically at the cell periphery ( Fig 3A ) . After hatching , although some of the GFP::CHDP-1 becomes diffuse in the cytoplasm , most of the signal retains this membrane-like distribution ( Fig 3B ) . Within the PLM neuron , CHDP-1 highlights the cell periphery ( Fig 3C ) , including surface protrusions ( Fig 3D ) . Further co-labeling experiments indicated that CHDP-1 co-localizes well with the plasma membrane marker Myr-mCherry ( Fig 3C and 3D ) , suggesting that the CHDP-1 protein associates with the cell cortex . The amphipathic helix motif has been implicated in membrane bilayer association [29] . We wondered whether the amphipathic helix motif hooks the CHDP-1 protein onto the plasma membrane . To test this hypothesis , we removed the amphipathic helix motif from CHDP-1 and examined the subcellular localization of the residual CHDP-1 fragment . We found that in the absence of the amphipathic helix motif , the cortex localization of CHDP-1 is completely lost ( Fig 3E ) . In contrast , whenever the amphipathic helix motif is present , the peripheral distribution of CHDP-1 protein is largely preserved ( Fig 3E ) . Thus , the cortex localization property of CHDP-1 is probably mediated by the amphipathic helix motif . Next , we asked whether the cortex localization is important for CHDP-1 function . To address this question , we removed the amphipathic helix domain and performed rescue experiments . We found that deletion of the amphipathic helix domain largely abolished the rescue effect of CHDP-1 ( Fig 3F ) . Therefore , localization to the cell cortex is crucial for CHDP-1 function . Without CHDP-1 , robust cell protrusion is severely reduced . Furthermore , the amphipathic helix domain has been implicated in membrane deformation [16] . We therefore asked whether CHDP-1 can promote membrane protrusion . To test this possibility , we increased chdp-1 gene expression in PLM neurons using the Pmec-7 promoter . In wild-type animals , PLM cell bodies generally display smooth spindle-like shapes ( Fig 4A ) . When chdp-1 gene expression was boosted , additional membrane protrusions of various sizes appeared on the surface of PLM cell bodies ( Fig 4B–4D ) . The Pmec-7 promoter also drives chdp-1 expression in ALM , AVM , and PVM neurons , and we found that these cells also display ectopic membrane protrusions around the cell body region ( S2 Fig ) . CHDP-1 localizes to the cell cortex . Double staining of CHDP-1::GFP and Myr::mCherry indicated that when it is overexpressed , the CHDP-1 protein still tightly associates with the cortex region ( Fig 4E ) . This suggests that CHDP-1 may function at the cell margin to induce cell protrusion . The formation of cell surface protrusions is a cooperative result of both membrane expansion and actin cytoskeleton rearrangement . The expanded plasma membrane must be stabilized by the underlying cortical actin network . To address whether the ectopic membrane protrusions are genuine cell surface protrusions , we co-labeled PLM cells with Myr-GFP and mCherry-actin . We found that most of the ectopically expanded plasma membrane is indeed supported by cortical actin ( 77% , n = 236 ) ( Fig 4F and 4G ) . However , some protrusions do not contain any detectable underlying actin ( 23% , n = 236 ) , and these tend to be smaller than those with actin ( Fig 4F and 4G ) . We suspected that these actin-free protrusions may represent newly formed protrusions . Furthermore , in wild-type PLM cell bodies , the actin cytoskeleton is usually organized into stress fiber-like structures ( Fig 4H ) [30] . However , when ectopic protrusions were induced by chdp-1 over-expression , most of the actin signal was shifted to the cortex region underlying those ectopic protrusions and no stress fiber-like structures could be found in the center of the PLM cell bodies , where the nucleus is located ( Fig 4I ) . Together , these results suggest that CHDP-1 promotes the formation of cell protrusions in vivo . Membrane expansion must be coupled with the actin network to efficiently promote protrusion . Can the cortex-localized CHDP-1 protein associate with actin , thus coupling membrane expansion to actin rearrangement ? Interestingly , a tandem array of type I or II CH domains has been shown to possess actin-binding activity [23 , 24] . Therefore , we tested whether the CHDP-1 protein can self-associate by performing co-precipitation assays on 293T cells transfected with Flag- and Myc-tagged CHDP-1 . We found that CHDP-1 molecules can indeed bind to each other ( Fig 5A ) . Truncation analysis indicated that the self-association of CHDP-1 is mediated by the CH domain ( Fig 5B ) . We next tested whether CHDP-1 can bind to actin . We firstly examined whether CHDP-1 binds to monomeric ( G ) actin . Flag-tagged CHDP-1 did not co-precipitate with worm actin ACT-1 ( Fig 5C ) , suggesting that CHDP-1 does not associate with G-actin . Next , we purified the mCherry::CHDP-1 protein ( Fig 5D ) and performed co-sedimentation experiments with polymerized filamentous ( F ) actin . As shown in S2D Fig , the actin motor myosin co-sedimented with F-actin , and this myosin-actin association could be disrupted by ATP . In contrast , CHDP-1 was not sedimented with F-actin at all ( Fig 5E ) . Together , these results indicate that CHDP-1 probably does not promote the formation of cell protrusions through direct actin binding . To understand how CHDP-1 interacts with the actin cytoskeleton to regulate protrusion formation , we searched for CHDP-1-interacting molecules by performing immunoprecipitation ( IP ) followed by mass spectrometry ( MS ) . With this approach , we identified the Rho small GTPase family member Rac1/CED-10 [31] . The worm Rac1 homolog , CED-10 , is localized to plasma membranes [32] ( Fig 6A ) . In developing PLM neurons , CED-10 co-localizes with CHDP-1 at the periphery of the PLM cell body as well as on the surface of the protrusion region ( Fig 6B and 6C ) . This implies that CHDP-1 and CED-10 may interact with each other at the cell cortex in vivo . To test whether CHDP-1 directly binds to CED-10 , we performed the following experiments . Flag-tagged full-length CHDP-1 and Myc-tagged CED-10 were co-expressed in HEK293T cells . After affinity purification , the purified CHDP-1 and CED-10 were mixed together and co-immunoprecipitations were performed with Flag or Myc antibodies . In contrast to the mock-transfected samples , CED-10 was effectively co-precipitated with CHDP-1 ( Fig 6D ) , and the purified CHDP-1 protein was co-precipitated by CED-10 ( Fig 6E ) . Thus , CHDP-1 can directly bind to CED-10 . To further identify the domain of CHDP-1 that interacts with CED-10 , we performed co-IP experiments using truncated CHDP-1 fragments . We found that the CH domain of CHDP-1 is crucial for the interaction with CED-10 ( Fig 6F ) . Taken together , these results indicate that CHDP-1 may bind to CED-10 through its CH domain . Coincidently , the BDU-PLM disconnection mutant screen identified a new ced-10 allele , xd33 ( Fig 6G and 6H ) . Imaging analysis further indicated that formation of cell protrusions is considerably repressed in xd33 mutant animals ( Fig 6H ) . xd33 is a C-to-T missense mutation resulting in a serine 22-to-phenylalanine change ( S22F ) in CED-10 . To understand the role of CED-10 in BDU-PLM connection , we examined two other ced-10 mutants , n1993 and n3246 and found that they both display similar BDU-PLM disconnection defects to xd33 ( Fig 6I ) . Previous report showed that the n3246 mutation is recessive to wild type , but homozygote ced-10 ( n3246 ) animals display defects not seen in ced-10 ( n1993 ) [33] . In xd33/n3246 trans-heterozygote animals , we found that BDU-PLM disconnection defect is similar to that of xd33 homozygote . Furthermore , when the wild-type ced-10 gene was introduced into xd33 animals , the BDU-PLM connection could be largely restored ( S3 Fig ) , suggesting that the loss-of-function of ced-10 is responsible for the membrane protrusion defect in the corresponding ced-10 mutant worms . We further constructed chdp-1 ( xd27 ) ;ced-10 ( xd33 ) and chdp-1 ( tm4947 ) ;ced-10 ( xd33 ) double mutants and found that those double mutant animals displayed similar degree of BDU-PLM disconnection defect as chdp-1 ( xd27 ) animals ( S3 Fig ) , implying that chdp-1 may function together with ced-10 to regulate cell protrusion . As a member of the Rac family of GTPases , Rac1 cycles between the inactive GDP-bound form and the active GTP-bound form . We then asked how CHDP-1 interacts with different forms of CED-10 . The glycine to valine mutation at position 12 ( G12V ) , which is canonical for constitutive activation of Ras superfamily GTPases , mimics the GTP-bound active state . In contrast , the threonine to asparagine mutation at position 17 ( T17N ) mimics the GDP-bound inactive state . Co-immunoprecipitation experiments with these two mutant forms of CED-10 showed that the CHDP-1 protein precipitates much higher levels of the G12V mutant than the T17N mutant ( Fig 7A and 7B ) . This suggests that CHDP-1 preferentially associates with GTP-bound active CED-10 . Both CHDP-1 and CED-10 localize to the cell margin . How does the direct association of CHDP-1 contribute to regulation of CED-10 function ? In chdp-1 mutant animals , we found that the cortex localization of CED-10 is partially lost ( Fig 7C and 7D ) . In contrast , localization of CHDP-1 to the cell margin remains intact in various ced-10 mutants , including ced-10 ( xd33 ) , ced-10 ( n1993 ) and ced-10 ( n3246 ) ( Fig 7E and S3 Fig ) . Cell-fractionation assays further indicated that the proportion of membrane-bound CED-10 is reduced in chdp-1 mutants compared with wild-type samples ( Fig 7F ) . The reduction of membrane-bound CED-10 suggested that the level of GTP-bound , active CED-10 protein maybe decreased in chdp-1 mutants . Therefore , we further tested the GTP-CED-10 level in vivo . Active GTP-Rac1 binds to the CRIB ( Cdc and Rac interactive binding ) domain of its effector PAK , while inactive GDP-Rac1 does not [34] . Thus , we performed pull-down assays on the PAK CRIB domain . In contrast to wild type , we found that the amount of GFP-tagged CED-10 pulled down by the PAK CRIB domain was considerably decreased in chdp-1 lysates ( Fig 7G and 7H ) . Together , these data suggested that CHDP-1 facilitates the cell margin localization of CED-10 in vivo by stabilizing the active GTP-bound CED-10 protein . As the critical regulator of actin dynamics , is CED-10 required for CHDP-1-induced cell protrusion ? We showed that over-expression of CHDP-1 leads to extensive cell protrusion in PLM cell bodies . In three different ced-10 mutants , however , the ectopic protrusion on the surface of PLM cell bodies is significantly reduced . We measured the size of the PLM cell body ( Fig 7I ) . In wild-type animals , the maximum area of a PLM cell body is about 26 μm2 . In chdp-1 over-expression animals , the area of the PLM cell body is expanded to around 138 μm2 . In ced-10 mutant animals , the area of the PLM cell body is 57 μm2 ( xd33 ) , 72 μm2 ( n1933 ) and 64 μm2 ( n3246 ) respectively ( Fig 7I ) . We then examined how CED-10 coordinates with CHDP-1 on the ectopic protrusions . After co-injecting mec-7 promoter-driven rfp::chdp-1 and gfp::ced-10 constructs into wild-type animals , we found that most , if not all , CHDP-1 and CED-10 are co-distributed on the ectopic cell protrusions ( Fig 7J ) . Very few protrusions contain either CHDP-10 ( 6 . 3% ) or CED-10 ( 1 . 9% ) , suggesting a tight association of CHDP-1 and CED-10 on those ectopic protrusions . Next , we tested whether the previously described “actin-free” protrusions contain CED-10 or not . Co-labeling of PLM cells with mCherry::ACT-1 and GFP::CED-10 showed that while the majority of protrusions ( 69 . 2% ) contain both CED-10 and actin , a significant proportion ( 23 . 2% ) contain CED-10 without actin ( Fig 7K ) . Thus , CED-10 is likely more closely associated with CHDP-1 than with actin to mediate protrusion formation ( S3 Fig ) .
Coupling membrane expansion to actin filaments dynamics is essential for formation of membrane protrusions . In this report , we identified that the calponin-like protein CHDP-1 promotes membrane protrusion and associates with the Rac1/CED-10 GTPase in C . elegans ( S3 Fig ) . Two mechanisms of membrane deformation have been linked to the direct binding of proteins to the membrane . One utilizes the electrostatic interactions between the lipid-binding domains on the protein surface and the negatively charged lipids in the membrane . An outstanding example of this class of lipid-binding domain is the BAR domain . BAR domains usually form dimers and the lipid-binding surface of BAR has intrinsic curvature [35] . In particular , the I-BAR family , with its cigar-shaped dimers , can promote negative curvatures , which are frequently found inside plasma membrane protrusions [36] . In addition to interacting electrostatically with lipids , BAR proteins can also recruit other proteins , which may further modify the membrane deformation . IRSp53 ( insulin receptor phosphotyrosine 53 kDa ) consists of an N-terminal I-BAR domain , followed by a partial CRIB domain and an SH3 domain . Through the CRIB domain , IRSp53 binds to Cdc42 [37] , which regulates the formation of filopodia in mammalian cells [38 , 39] . Through its SH3 domain , IRSp53 also associates with a set of actin regulatory factors , including WAVE1 , WAVE2 , Mena , mDia1 , dynamin , and N-WASp [40] . Thus , IRSp53 allows tight coupling of membrane protrusions to actin dynamics . Direct protein insertion into the membrane is another mechanism used to alter membrane curvature . This mechanism of membrane deformation was first characterized for proteins containing an amphipathic helix; the cylindrical helix lies parallel to the membrane , exposing its hydrophilic surface to the cytosol [41] . However , it is unlikely that a single molecule or a single dimer could act alone to deform the membrane . The assembly of many molecules at the same place on the membrane may be required for deformation . Indeed , in bacteria , the peripheral membrane protein MinD , which contains a highly similar amphipathic helix to CHDP-1 , forms polymers and the polymerization of MinD proteins enhances the membrane affinity of the single amphipathic helix domain [29] . Given the fact that CHDP-1 molecules can associate with each other , it is a reasonable proposal that the assembly of chains of CHDP-1 molecules may create sufficient membrane curvature to deform the plasma membrane . What is the biological significance of the enlarged plasma membrane surface ? During the formation of neuronal connections , the distal tips of axons extend highly dynamic membrane expansions , the growth cones , to actively search for environmental cues that will guide them toward their proper target cells [27] . We previously showed that Wnt signaling is involved in BDU-PLM connection [26] . Therefore , it is possible that CHDP-1 promotes the formation of cell protrusions in order to increase the sensitivity of BDU and PLM neurons to detect Wnt signals . Interestingly , the expression of chdp-1 in either BDU or PLM alone significantly rescued the BDU-PLM disconnection defect in chdp-1 mutant animals , suggesting that the expanded growth cone of a given neuron may provide sufficient signal detection capacity to locate its target cell during the connection process . Given the fact that BDU and PLM neurites can elongate towards each other in both wild-type and chdp-1 mutant animals , we suspect that the growth cone protrusion may be specifically required for the final short-range extension of the neurite tip , but is not necessary for long-distance travel of a neurite . It is currently unclear whether similar principles apply to other neuronal connections in vivo . How does CHDP-1 couple membrane protrusion to the actin cytoskeleton ? CHDP-1 contains a single CH domain , which shares similarity to that in the mammalian calponin family . In the phylogenetic tree of CH domain-containing proteins , calponins form a separate branch together with other proteins that possess a single CH domain ( Vav , IQGAP , ARH-GEF6 , and SM22 ) [23] . Another calponin homolog , CPN-1 ( CaPoNin-1 ) , can partially fulfill the functional requirement for CHDP-1 in BDU-PLM connection ( reducing the frequency of disconnection from around 95% in chdp-1 ( xd27 ) mutant animals to 60% ) , suggesting that CHDP-1 is likely also derived from an ancestral protein in C . elegans . The isolated CH domains of calponin family proteins fail to associate with F-actin [25] . Instead , they bind to a variety of cytoskeleton and signaling components , including tubulin , intermediate filament , ROCK ( Rho-associated kinase ) , ERK1 , and ERK2 [25] . It has therefore been suggested that calponin functions as a scaffolding protein for cytoskeletal structures and/or adaptor molecules in signaling pathways . Here , we found that the CH domain of CHDP-1 does not bind to either G- or F-actin , but instead binds to the Rac1 homolog CED-10 . This suggests that CHDP-1 may influence actin cytoskeleton dynamics by directly regulating Rac1/CED-10 . Intriguingly , CHDP-1 preferentially associates with GTP-bound active CED-10 . In addition , the plasma membrane localization of CED-10 is impaired in the absence of CHDP-1 . Coincidently , expression of constitutively active CED-10 in C . elegans promotes extensive branching [42] , suggesting that Rac1 plays an important role in actin polymerization . Another interesting observation is that CHDP-1 molecules may associate with each other also through the CH domain . We suspect that the assembly of chains of CHDP-1 molecules may create sufficiently large membrane protrusions , as well as providing anchor sites for numerous CED-10 molecules , which will efficiently recruit cortical actin to stabilize the protrusions . Thus , CHDP-1 promotes plasma membrane protrusion and stabilizes the expanded membrane by promoting actin cytoskeleton rearrangement through interaction with Rac1/CED-10 .
Culture and manipulation of C . elegans strains were performed using standard methods . Mutants used in this studies are: LGI , chdp-1 ( xd27 ) , chdp-1 ( tm4947 ) ; LGII , juIs76 ( Punc-25::GFP ) ; LGIV , ced-10 ( n1993 ) , ced-10 ( xd33 ) , ced-10 ( n3246 ) , kyIs262 ( Punc-86::MYR::GFP , Podr-1::dsRed ) . The xd27 and xd33 mutations were isolated from kyIs262 ( Punc-86::Myr::GFP ) animals treated with EMS . A total of 2 , 500 mutagenized haploid genomes were screened . The isolated strains were outcrossed with the N2 strain at least four times . Corresponding chdp-1 DNA fragments were amplified from N2 genomic DNA to perform the chdp-1 cloning experiment . chdp-1 cDNA was amplified by reverse transcription . ced-10 cDNA was obtained from Dr . Yuji Kohara . GFP sequence was inserted into Pchdp-1::CHDP-1 constructs . For protein co-localization , the RFP fragment was inserted into the Pmec-7::CHDP-1 construct . The mec-7 promoter was inserted into Pced-10::GFP::CED-10 construct to replace the ced-10 promoter . In general , plasmid DNAs of interest were injected at 10-50ng/μl and the co-injection markers Podr-1::RFP , rol-6 or Podr-1::GFP were injected at 10-50ng/μl . The corresponding transgenes and constructs are listed in ( S1 Table ) . Animals were mounted on 2% agar pads in M9 buffer containing 1% 1-phenoxy-2-propanol and examined by fluorescence microscopy unless indicated otherwise . Fluorescence photographs were taken using a Zeiss Axioimager A1 with an AxioCam digital camera and Axiovision rel . 4 . 6 software ( Carl Zeiss ) or an IX81 Olympus inverted confocal microscope . Contact was considered defective if the BDU neurite failed to touch the PLM neurite . Neurite and worm body lengths were measured with NIH ImageJ software . The length of each neurite was traced from the center of the cell body to the tip of neurite . The length of each animal was measured from the center of the PLM cell body to the tip of the nose . The relative BDU or PLM neurite length is defined as neurite length/animal length . The PLM cell body was measured using ImageJ software . To measure the co-distribution of actin , CED-10 and CHDP-1 on ectopic protrusions , fluorescence photographs were taken using confocal microscopy . The size of individual protrusions was measured with NIH Image J software using the line-tracing tool . All data are shown as mean ± SD . Statistical analyses were performed with Student’s t-test . For each genotype , more than 20 animals were analyzed . At different time points , eggs were mounted on thin 1% agar pads in M9 buffer; gentle pressure was applied to crack the egg shells , and embryos were examined immediately by fluorescence microscopy . To measure the membrane protrusion area , newly hatched L1 animals were collected and photographs were taken within 10 min of hatching . The expansion area was defined by linking the first branching or elaboration site of the neurite with the tips of all the branches or filopodia . The area was measured with NIH ImageJ software . Some chdp-1 ( xd27 ) embryos die at the later elongation stage . Therefore we dissected gravid adults and obtained embryos at early developmental stages ( from the 16-cell stage to gastrulation stage ) . Then the collected embryos were mounted on thin 1% agar pads in M9 buffer and examined immediately by fluorescence microscopy . Embryos in which at least half of the cells contained diffuse GFP::CED-10 or GFP::CHDP-1 signal were considered as showing defective localization of the GFP::CED-10 or GFP::CHDP-1 marker . All images were collected using Zeiss Axioimager A1 with an AxioCam digital camera and Axiovision rel . 4 . 6 software ( Carl Zeiss ) . For each sample , at least 50 embryos were scored . Newly hatched L1 animals were anesthetized with 0 . 1mM levamisole in M9 buffer and mounted on 2% agar pads at 22°C . Time-lapse images were captured every 1 . 5 min using a Delta-vision Core imaging system ( Applied Precision ) with an UPLSApo 100 6/1 . 40NAoil-immersion objective and a Photometrics CoolSnap HQ camera . Deconvolution and analysis of images were performed with Softworx ( Applied Precision ) . Worm lysates were prepared from the N2 strain carrying either xdIs23 ( Pchdp-1::GFP ) or xdIs48 ( Pchdp-1::GFP::CHDP-1 ) . Specifically , worms were collected and lysed with a Dounce homogenizer ( Cheng-He Company , Zhuhai , China ) [43] in pre-chilled homogenizing buffer ( 50mM Tris-Cl pH8 . 0 , NaCl 150mM , 0 . 5% sodium deoxycholate , 1% Triton-X 100 , protease inhibitor [Roche] ) , then incubated for 15 min on ice , and centrifuged at 12 , 000 rpm for 15 min at 4°C . The supernatants were incubated with GFP polyclonal antibody ( Rabbit , Abcam , ab290 , 1:1 , 000 dilution ) overnight , and then incubated with protein A agarose beads ( Cat#17-0780-01 , GE ) for 4 hr at 4°C . The pellet was washed three times in washing buffer ( 50mM Tris-Cl pH8 . 0 , 150mM NaCl , 1% NP-40 , 1 mM PMSF ) , and then boiled in sample buffer for 5 min . The boiled samples were resolved by SDS-PAGE gel and protein bands of interest were excised from SDS-PAGE gel and in-gel digestion was performed using a well-established protocol with slight modifications [44] . Briefly , the proteins embedded in gel slices were reduced with 10mM DTT and alkylated with 55mM iodoacetamide , and then digested overnight with sequencing grade trypsin ( Sigma , USA ) at 37°C . The tryptic peptides were analyzed by mass spectrometry using a TripleTOF 5600 mass spectrometer ( AB SCIEX , Canada ) coupled to an EksigentNanoLC . Peptide identification and quantification were performed with the ProteinPilot 4 . 2 software ( AB SCIEX ) . The C . elegans proteome sequences ( Uniprot ) was used as the database and the mass tolerance was set to 0 . 05 Da for the database search . The false discovery rate ( FDR ) analysis was performed using the software PSPEP integrated with ProteinPilot . About 10 cytoskeleton-related proteins were identified in the mass spectrometry analysis . GST pull-down and co-immunoprecipitation assays were then performed to verify the putative protein-protein interactions . Among the tested proteins , CED-10 displayed consistent CHDP-1-binding activity in both GST pull-down and co-immunoprecipitation assays . cDNA fragments were amplified and cloned into modified pcDNATM3 . 1/myc-HIS ( - ) or pFLAG-CMV-2 vectors through standard procedures . HEK293T cells were cultured in DMEM medium supplemented with 12% FBS . Plasmid transfections were carried out using Lipofectamine 2000 ( Invitrogen ) . 24 hours after transfection , cells were harvested and lysed for 30 min at 4°C . After centrifugation , the supernatants were incubated with anti-FLAG M2 affinity gel beads at 4°C overnight and then washed three times with washing buffer and incubated with SDS-sample buffer . Samples were resolved by standard immunoblotting techniques . To quantify the binding affinity between CHDP-1 and different CED-10 mutant proteins ( G12V and T17N mutant forms ) , the absolute signal intensity of individual protein bands was analyzed with ImageJ . The relative protein level was achieved by subtracting the respective FLAG control . For binding experiments with truncated CHDP-1 , the following fragments were used: FN ( AA 1–146 ) , CH ( AA 121–242 ) and FC ( AA 243–338 ) . All the co-immunoprecipitation experiments were repeated at least three times . The Actin Binding Protein Spin-Down Biochem Kit ( Cat . # BK001 ) from Cytoskeleton Inc was used to detect the F-actin binding activity of CHDP-1 . The recombinant proteins Flag::mCherry::CHDP-1 and Flag::mCherry were expressed in HEK293FT cells . 24 hours after transfection , cells were harvested and lysed for 30 min at 4°C . After centrifugation , the corresponding supernatants were incubated with anti-FLAG M2 affinity gel beads at 4°C for purification of tagged proteins . F-actin was prepared from muscle actin and suspended in general actin buffer at room temperature for 1 hour . Proteins at a molar ratio of 1:3 ( recombinant protein: F-actin ) were incubated at 25°C for 30 minutes and pelleted at 15 0000g for 1 . 5 hours . Subsequently , protein samples were resolved by SDS-PAGE gel . The muscle myosin was used as a positive control for F-actin binding assay . Wild-type and chdp-1 ( xd27 ) worms expressing Pced-10::GFP::CED-10 were collected and washed in M9 buffer . Then 500μl lysis buffer ( 250 mM sucrose , 50 mM Tris–HCl pH6 . 8 , 1 mM EDTA ) were added and the samples were homogenized with a Dounce homogenizer ( Cheng-He Company , Zhuhai , China ) ( Chen et al . , 2010 ) on ice for 15 min . The nuclear pellet was removed by centrifuging at 3 , 000 rpm for 10 min at 4°C . The supernatant was further centrifuged at 40 , 000 rpm for 1 hr . The new supernatant was collected as the cytosolic fraction . The pellet was further washed and centrifuged at 40 , 000 rpm for 45 min to obtain the membrane fraction . All samples were boiled with 2xSDS loading buffer then subjected to 10% SDS-PAGE gel analysis . PAK-GST protein beads ( human P21 activated kinase PBD ) were used to detect the GTP-bound form of CED-10 . Worm lysates were prepared from wild type and chdp-1 ( xd27 ) mutant animals carrying Pced-10::GFP::CED-10 . Worms were lysed in pre-chilled homogenizing buffer with a Dounce homogenizer ( Cheng-He Company , Zhuhai , China ) and incubated for 15 min on ice , then centrifuged at 12 , 000 rpm for 15 min at 4°C . The supernatants were incubated with PAK-GST protein beads ( Cat#PAK02 , Cytoskeleton Inc ) for 4 hr at 4°C . The pellet was washed three times and then boiled in sample buffer for 5 min . The boiled samples were resolved by SDS-PAGE and detected with anti-GFP antibody . The pull-down experiments were repeated three times . The absolute intensity of each protein band was quantified with ImageJ . The relative protein level was determined by normalizing each sample with the corresponding input , and then normalizing each chdp-1 ( xd27 ) sample with the paired WT sample . | In response to intra- and extracellular cues , remodeling of the sub-membranous cortical actin cytoskeleton constantly reorganizes the plasma membrane . Thus , distinct types of actin-rich invaginations or protrusions , such as filopodia and lamellipodia , enable cells to explore territory and pull themselves around . Extensive research has shown that the plasma membrane is tightly coupled to the motility machinery . However , how the continuous reorganization of the actin cytoskeleton is coupled with appropriate restructuring of the plasma membrane at the molecular level in vivo is unclear . Here , we identified that the single calponin homology ( CH ) domain-containing protein CHDP-1 promotes the formation of cell protrusions in C . elegans . CHDP-1 localizes to the cell cortex and through its calponin homology ( CH ) domain , CHDP-1 directly binds to the master actin regulator Rac1/CED-10 and enhances the membrane localization of the active CED-10 protein . Thus , we discovered a novel CHDP-1/Rac1 module which effectively couples membrane expansion to cortex actin dynamics in vivo . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"invertebrates",
"caenorhabditis",
"neurites",
"neuroscience",
"animals",
"membrane",
"proteins",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"cellular",
"structures",
"and",
"organelles",
"neuronal",
"dendrites",
"cytoskeleton",
"research",
"and"... | 2016 | The Calponin Family Member CHDP-1 Interacts with Rac/CED-10 to Promote Cell Protrusions |
Collective behaviour is a widespread phenomenon in biology , cutting through a huge span of scales , from cell colonies up to bird flocks and fish schools . The most prominent trait of collective behaviour is the emergence of global order: individuals synchronize their states , giving the stunning impression that the group behaves as one . In many biological systems , though , it is unclear whether global order is present . A paradigmatic case is that of insect swarms , whose erratic movements seem to suggest that group formation is a mere epiphenomenon of the independent interaction of each individual with an external landmark . In these cases , whether or not the group behaves truly collectively is debated . Here , we experimentally study swarms of midges in the field and measure how much the change of direction of one midge affects that of other individuals . We discover that , despite the lack of collective order , swarms display very strong correlations , totally incompatible with models of non-interacting particles . We find that correlation increases sharply with the swarm's density , indicating that the interaction between midges is based on a metric perception mechanism . By means of numerical simulations we demonstrate that such growing correlation is typical of a system close to an ordering transition . Our findings suggest that correlation , rather than order , is the true hallmark of collective behaviour in biological systems .
Intuition tells us that a system displays collective behaviour when all individuals spontaneously do the same thing , whatever this thing may be . We surely detect collective behaviour when all birds in a flock fly in the same direction and turn at the same time [1] , as well as when all spins in a magnet align , giving rise to a macroscopic magnetization [2] , [3] . On the other hand , we would not say that there is any collective behaviour going on in a gas , despite the large number of molecules . The concept of collective behaviour seems therefore closely linked to that of emergent collective order , or synchronization . Indeed , explaining how order spontaneously arises from local inter-individual interactions has been one of the major issues in the field [4]–[6] . The case of insect swarms is tricky in this respect . Several species in the vast taxonomic order Diptera ( flies , mosquitoes , midges ) form big swarms consisting largely of males , whose purpose is to attract females [7] , [8] . Swarming therefore has a key reproductive function and , in some cases , relevant health implications , the obvious , but not unique , example being that of the malaria mosquito , Anopheles gambiae [9]–[11] . It is well-known that swarms form in proximity of some visual marker , like a water puddle , or a street lamp [7] . Swarming insects seem to fly independently around the marker , without paying much attention to each other ( see Video S1 ) . For this reason , the question of whether swarms behave as truly collective systems is debated [4] , [12] . In fact , it has even been suggested that in Diptera there is no interaction between individuals within the swarm and therefore no collective behaviour at all [13] , [14] . Although other studies observed local coordination between nearest neighbours [15] , [16] , it remains controversial whether and to what extent collective patterns emerge over the scale of the whole group . Clarifying this issue is a central goal in swarms containment [17] , [18] . In absence of quantitative evidence telling the contrary , the hypothesis that external factors , as the marker , are the sole cause of swarming and that no genuine collective behaviour is present , is by far the simplest explanation . We must , however , be careful in identifying collective behaviour with collective order . There are systems displaying important collective effects both in their ordered and in their disordered phase . An example is that of a ferromagnet near the critical temperature i . e . the temperature below which a spontaneous magnetization emerges: the collective response of the system to an external perturbation is as strong in the disordered phase slightly above as it is in the ordered phase slightly below In fact , once below the critical temperature , increasing the amount of order lowers the collective response [2] , [3] . Similarly , in animal behaviour it is possible to conceive cases in which individuals coordinate their behavioural reactions to environmental stimuli , rather than the behaviours themselves; conversely we may expect that a group too heavily ordered , i . e . with a very large behavioural polarization , may respond poorly to perturbations , because of an excessive behavioural inertia . Hence , although certainly one of its most visible manifestations , emergent order is not necessarily the best probe of collective behaviour . The crucial task for living groups is not simply to achieve an ordered state , but to respond collectively to the environmental stimuli . For this to happen , correlation must be large , namely individuals must be able to influence each other's behavioural changes on a group scale . The question then arises of whether correlation in biological systems is a consequence of collective order or whether it can be sustained even in absence of order . The question is relevant because the way individuals in a group synchronize their behavioural fluctuations ( correlation ) is possibly a more general mechanism than the synchronization of behaviour itself ( order ) . All experimental studies performed up to now , however , concerned highly synchronized groups ( as bird flocks , fish shoals and marching locusts [19]–[21] ) , which displayed both order and correlation . Hence , the question of whether or not order and correlations are two sides of the same phenomenon remained open until now . Here , we attempt to give an answer to this question by experimentally studying large swarms of insects in the field . As we will show , despite the lack of collective order , we do find strong correlations , indicating that in biological systems collective behaviour and group-level coordination do not require order to be sustained .
We perform an experimental study of swarms of wild midges in the field . Midges are small non-biting flies belonging to the order Diptera , suborder Nematocera ( Diptera:Chironomidae and Diptera:Ceratopogonidae - see Methods ) . The body length of the species we study is in the range – Swarms are found at sunset , in the urban parks of Rome , typically near stagnant water . As noted before [7] , we find that swarms form above natural or artificial landmarks . Moving the landmark leads to an overall displacement of the swarm . The swarms we studied range in size between and individuals ( see Table S1 in Text S1 ) . To reconstruct the 3d trajectories of individual insects we use three synchronized cameras shooting at frames-per-seconds ( trifocal technique – Fig . 1e and Methods ) . Our apparatus does not perturb the swarms in any way . The technique is similar to the one we used for starling flocks [22] , with one notable difference . To reach the desired experimental accuracy we need to know the mutual geometric relations between the three cameras very accurately . In the case of flocks , this could be achieved only by an a priori alignment of the cameras . In the case of swarms , though , we proceed differently . After each swarm acquisition , we pin down the geometry of the camera system by taking multiple images of a calibrated target ( Fig . 1f ) . This procedure is so accurate that the error in the 3d reconstruction is dominated by the image segmentation error due to the pixel resolution . If we assume this to be equal to pixel ( typically it is smaller than that because midges occupy many pixels ) , we make an error of in the determination of the distance between two points apart from each other ( a reference value for nearest neighbour distance ) . The absolute error is the same for more distant points , making the relative precision of our apparatus even higher . This accuracy makes the determination of the correlation functions we study here very reliable . The -tracking of each midge is performed by using the recursive global optimization method described in [23] . This recursive algorithm dramatically reduces the complexity of the tracking problem , effectively overcoming the limit of other tracking methods [24] , [25] , and allowing the reconstruction of large swarms , up to midges , for long time , up to frames . Sample reconstructions are shown in Fig . 1b and in Video S2 . Compared to previous field [11] , [15] , [26] and lab [27]–[29] studies , data collected and analysed in the present work have the advantage to span among swarms of different sizes and densities . Swarms are in a disordered phase . The standard order parameter normally used in collective behaviour is the polarization , where is the number of midges in the swarm and is the velocity of insect The polarization measures the degree of alignment of the directions of motion; it is a positive quantity and its maximum value is The average polarization over all swarms is quite small , ( see Fig . 2 and Table S1 in Text S1 ) . As a reference , in starling flocks we find on average [19] . The probability distributions of the polarization fully confirms the swarms' lack of translational order and the stark difference with flocks ( Fig . 2 ) . Clearly , swarms are not in a polarized state . Translation is not the only possible collective mode , though . For example , it is well-known that fish schools can produce rotating ( milling ) configurations . Moreover , a group can expand/contract , giving rise to dilatational ( or pulsive ) collective modes . For this reason we have defined and measured also a rotational and a dilatational order parameter ( see Methods ) . We find , however , that these quantities too have very small values ( Fig . 2 ) . The time series , on the other hand , show that the order parameters can have rare , but strong fluctuations , during which their value may become significantly larger than that of an uncorrelated system ( Fig . 2 ) . These large fluctuations are a first hint that non-trivial correlations are present . The connected correlation function measures to what extent the change in behaviour of individual is correlated to that of individual at distance Correlation is the most accessible sign of the presence of interaction between the members of a group . The absence of interaction implies the absence of correlation . Conversely , the presence of correlation implies the presence of an effective interaction ( see Text S1 , Section I ) . Correlation can be measured for different quantities , but in the case of midges , as with birds and other moving animals , the principal quantity of interest is the direction of motion . To compute the connected correlation we first need to introduce the velocity fluctuations , namely the individual velocity subtracted of the overall motion of the group , ( see Methods for the detailed definition of and ) . This fluctuation is a dimensional quantity , hence it is unsuitable to compare the correlation in natural vs numerical systems , as we shall do later on . We therefore introduce the dimensionless velocity fluctuation , ( 1 ) The connected correlation function is then given by , ( 2 ) where if and zero otherwise , and is the space binning factor . The form of in natural swarms is reported in Fig . 3: at short distances there is strong positive correlation , indicating that midges tend to align their velocity fluctuations to that of their neighbours . After some negative correlation at intermediate distances , relaxes to no correlation for large distances . This qualitative form is quite typical of all species analysed ( see Fig . 3 ) . The smallest value of the distance where crosses zero is the correlation length , that is an estimate of the length scale over which the velocity fluctuations are correlated [19] . The average value of this correlation length over all analysed swarms is , This value is about times larger than the nearest neighbours distance , whose average over all swarms is ( see Fig . 3 and Table S1 in Text S1 ) . Previous works noticed the existence of pairing manoeuvres and flight-path coordination between nearest neighbours insects [4] , [15] , [16] . Our results , however , indicate that midges within a natural swarm influence each other's motion far beyond their nearest neighbours . The collective response of the swarm depends crucially on two factors: how distant in space the behavioural change of one insect affects that of another insect ( spatial span of the correlation ) and how strong this effect is ( intensity of the correlation ) . To combine these two factors in one single observable we calculate the cumulative correlation up to scale , ( 3 ) where is the Heaviside function , It can be shown ( see Text S1 , Section II ) that this dimensionless function is related to the space integral of the correlation function Hence , reaches a maximum where vanishes , i . e . for ( see Fig . 3 ) . This maximum , is a measure of the total amount of correlation present in the system . In statistical physics is exactly equal to the susceptibility , namely the response of the system to an external perturbation [30] , [31] . In collective animal behaviour , we do not have a quantitative link between integrated correlation and response , so that calling susceptibility is not strictly correct . However , if the probability distribution of the velocities is stationary , we can follow a maximum entropy approach [32] and still find that the total amount of correlation in the system , is related to the way the group responds collectively to a perturbation ( see Text S1 , Section II ) . The value of for midge swarms is reported in Fig . 3 . In order to judge how significant is the correlation function and how large is the susceptibility in natural swarms , we need an effective zero for these quantities , i . e . some null hypothesis baseline . As we have seen in the Introduction , the minimal assumption is that all individuals in the swarm interact with an external landmark independently from each other . Following Okubo [4] ( but see also [27] and [16] ) , we therefore simulate a ‘swarm’ of non-interacting particles performing a random walk in a three-dimensional harmonic potential ( see Methods ) . Visually , the group behaviour of this Non-interacting Harmonic Swarm ( NHS ) looks remarkably similar to that of a real swarm ( see Video S2 and S3 ) : all ‘midges’ fly around the marker and the group lacks collective order . This similarity , however , is deceptive . In the NHS , the correlation function simply fluctuates around zero , with no spatial span , nor structure ( Fig . 3 ) . Moreover , the susceptibility in the NHS is extremely small , whereas the susceptibility in natural swarms is up to times larger than this non-interacting benchmark ( Fig . 3 ) . We conclude that swarming behaviour is not the mere epiphenomenon of the independent response of each insect with the marker . Despite the lack of collective order , natural swarms are strongly correlated on large length scales . There are big clusters of midges that move coherently , contributing to the ‘dancing’ visual effect of the swarm . The only way this can happen is that midges interact . What kind of interaction is that ? To understand the nature of the interaction , we study the susceptibility across swarms of different densities . Interestingly , we find that increases when the average nearest neighbour distance , decreases ( Fig . 4 ) . Denser swarms are more correlated than sparser ones . This result indicates that midges interact through a metric perceptive apparatus: the strength of the perception decreases with the distance , so that when midges are further apart from each other ( larger ) the interaction is weaker and the susceptibility is lower . This is at variance with what happens in starling flocks: starlings interact with a fixed number of neighbours , irrespective of their nearest neighbour distance [33]; such kind of topological interaction does not depend on the group density , hence the susceptibility does not depend on the nearest neighbour distance . Fig . 4 , on the other hand , shows that midges interact metrically , namely with all neighbours within a fixed metric range , Hence , in swarms the number of interacting neighbours increases with decreasing ( increasing density ) , and as a consequence of this increased amount interaction , the system becomes also more correlated . In a system ruled by metric interaction we expect all lengths to be measured in units of the perception range , This implies that the natural variable for the susceptibility is the rescaled nearest neighbour distance , The problem is that we are considering different species of midges , likely to have different metric perception ranges . The simplest hypothesis we can make is that is proportional to the insect body length ( which we can measure ) , so that This hypothesis is confirmed by the data: the susceptibility is significantly more correlated to the variable ( P-value ) than to ( P-value - see Methods for the definition of P-value ) . The fact that the natural variable is is a further indication that the interaction in swarms is based on a metric mechanism . The difference in the nature of the interaction between flocking birds and swarming midges ( topological vs . metric ) is possibly due to the significant differences between vertebrates and arthropods . Topological interaction , namely tracking a fixed number of neighbours irrespective of their distance , requires a level of cognitive elaboration of the information [33] more sophisticated than a metric interaction , where the decay of the effective force is merely ruled by the physical attenuation of the signal with increasing metric distance . In other words , within a metric mechanism the range of the interaction is fixed by a perceptive cut-off , rather than a cognitive one . Metric interaction is known to be more fragile than topological one against external perturbations [33] , and indeed it is far more likely to observe the dispersion of a swarm in the field than that of a flock . This may be the reason why the presence of an external marker is crucial for the swarming behaviour of midges [13] . The experimental observations of a non-trivial connected correlation and of a large susceptibility indicate that midges are effectively interacting with each other by acting on their directions of motion . This does not exclude , of course , that other types of interaction are present . First of all , the empirical observation that the swarm uses a visual marker as a reference for maintaining its mean spatial position , strongly suggests that each individual interacts with the marker . Besides , it is certainly possible that effective positional attraction-repulsion forces between midges , as those described in [34] , exist . However , the directional correlations indicate that insects are also effectively interacting by adjusting their velocities . Moreover , the fact that these correlations are positive for short distances means that midges tend to align their direction of motion . This fact may seem surprising , because alignment interactions typically lead to the formation of ordered ( polarized ) groups , which is clearly not the case for midges . Swarms are disordered , and yet interacting and highly correlated systems . Is this a paradox ? In fact , it is not . An alignment interaction does not per se lead to global order in the group . In all models where imitation of the neighbours is present , the onset of long-range order depends on the value of some key tuning parameter . In a ferromagnet , this parameter is the temperature namely the amount of noise affecting the interaction between the neighbouring spins . At high temperature the system is in a disordered state , whereas by lowering one reaches a critical temperature below which an ordering transition occurs . In models of active matter there is another parameter tuning the transition between disorder and order , that is density or , equivalently , nearest neighbour distance: the system gets ordered once the nearest neighbour distance falls below some transition value . The crucial point is that , in general , the correlation of the system tends to be very large around the transition point , irrespective of whether the system is in the ordered or in the disordered phase . Hence , even a disordered system can display large correlations , provided that it is not too far from an ordering transition . In what follows , we want to show that this is indeed what happens with midge swarms . The simplest model based on alignment interaction that predicts an order-disorder transition on changing the density is the Vicsek model of collective motion [35] . In this model each individual tends to align its direction of motion to that of the neighbours within a metric perception range , The rescaled nearest neighbour distance , is the control parameter: for low noise , the model predicts a transition from a disordered phase ( low polarization ) at high values of ( low density ) , to an ordered phase ( large polarization ) at low values of ( high density ) [35]–[37] . We numerically study the Vicsek model in three dimensions . As we have seen , real swarms hold their average position with respect to a marker; to reproduce this behavioural trait we introduce an harmonic attraction force that each individual experiences towards the origin ( see Methods ) . Also in central potential the model displays an ordering transition: at large density , for the system is ordered and it has large polarization ( Video S4 ) . On the other hand , the polarization is low in the disordered phase , ( Fig . 5 ) . However , the correlation function is non-trivial when is sufficiently close to ( Fig . 5 ) , indicating the existence of large clusters of correlated individuals , which can be clearly detected in Video S5 . We calculate the susceptibility in the same manner as we did for natural swarms , in the disordered phase , and find a clear increase of on lowering ( Fig . 5 ) . This increase of the susceptibility is coherent with the existence of an ordering transition at It has been shown that , unless is much larger than the values analysed here , the transition in the Vicsek model is characterized by a clear second order phenomenology ( the nature of the transition for is still debated - see [37]–[39] ) . As a consequence , the susceptibility is expected to become very large approaching and to follow the usual scaling relation of critical phenomena [38] , ( 4 ) A fit to equation ( 4 ) of the 3d-Vicsek data is reported in Fig . 5 , giving and a transition point , The reason for the growth of approaching in the Vicsek model is quite intuitive . The model is metric , so that at large namely when the nearest neighbour distance is much larger than the interaction range very few individuals interact with each other , and coordination is small . The smaller becomes , the larger the number of particles within the mutual interaction range , thus promoting the correlation of larger and larger clusters of particles . For this reason the correlation length and the susceptibility grow when the nearest neighbour distance decreases . When approaches its critical value , the coordinated clusters become as large as the whole system , so that the groups orders below The low order parameter , the non-trivial correlation function , and especially the increase of on decreasing the nearest neighbour distance , are phenomenological traits that the metric Vicsek model shares with natural swarms . We conclude that a system based solely on alignment can be in its disordered phase and yet display large correlations , as midge swarms do . It is interesting to note that by approaching the ordering transition a compound amplification of the correlation occurs: when the nearest neighbour distance , decreases , the spatial span of the correlation , increases , so that the effective perception range in units of nearest neighbour distance , is boosted . We emphasize that we are not quantitatively fitting Vicsek model to our data . Our only aim is to demonstrate a general concept: large correlation and lack of global order can coexist even in the simplest model of nearest neighbours alignment , provided that the system is sufficiently close to an ordering transition . The consistency between our experimental data and the Vicsek model suggests that an underlying ordering transition could be present in swarms as well . An ordering transition as a function of the density has been indeed observed in laboratory experiments on locusts [21] , fish [40] and in observations of oceanic fish shoals [41] . In these cases , both sides ( low and high density ) of the ordering transition were explored . However , midge swarms in the field are always disordered , living in the low-density/high- side of the transition . Locating a transition point having data on just one side of it , is a risky business . The reason why we want to do this here is because it will allow us to give a rough estimate of the metric range of interaction in midges , which can be compared with other experiments . If a Vicsek-like ordering transition exist , we can use equation ( 4 ) to fit the swarms data for ( Fig . 4 ) . As we already mentioned , we do not know the value of the metric perception range , in swarms . Therefore , we use as scaling variable where is the body length . Although the fit works reasonably well ( Fig . 4 ) , the scatter in the data is quite large; hence , given the non-linear nature of the fit , it would be unwise to pin down just one value for the parameters , and we rather report confidence intervals . The fit gives a transition point in the range , with an exponent in the range , ( larger exponents correspond to lower transition points ) . Interestingly , there is an alternative way to locate the ordering transition that does not rely on the fit of Let us establish a link between pairs of insects closer than the perception range and calculate the size of the biggest connected cluster in the network . Given a swarm with nearest neighbour distance the larger the larger this cluster . When exceeds the percolation threshold , a giant cluster of the same order as the group size appears [42] . We calculate the percolation threshold in swarms ( Fig . 6 and Methods ) and find The crucial point is that varying the perception range at fixed nearest neighbour distance is equivalent to varying at fixed Hence , at fixed there is an equivalent percolation threshold of the nearest neighbour distance , such that for a giant cluster appears . Clearly , It is reasonable to hypothesise that the critical nearest neighbour distance is close to the maximal distance compatible with a connected network , given A sparser network would cause the swarm to lose bulk connectivity . Therefore , given a certain perception range the ordering transition occurs at values of the nearest neighbour distance close to its percolation threshold , At this point we have two independent ( and possibly equally unreliable ) estimates of the transition point in natural swarms of midges: the first one in units of body-lengths , the second one in units of interaction range , Putting the two together we finally obtain an estimate of the metric interaction range in units of body-lengths , The body length of the species under consideration is in the range , This implies a perception range of a few centimetres , depending on the species . This crude estimate of the midge interaction range is compatible with the hypothesis that midges interact acoustically . In [43] the male-to-male auditory response in Chironomus annularius ( Diptera:Chironomidae ) was studied and it was found that the range of the response was about not too far from our estimate . Similar measurements in mosquitoes ( Diptera:Culicidae ) show that the auditory perception range is about [44] , which is again compatible with our determination of the interaction range in midge swarms .
We have shown that natural swarms of midges lack collective order and yet display strong correlations . Such correlations extends spatially much beyond the inter-individual distance , indicating the presence of significant cluster of coordinated individuals . This phenomenology is incompatible with a system of non-interacting particles whose swarming behaviour is solely due to the attraction to an external landmark . We conclude that genuine collective behaviour is present in swarms . We stress that the existence of correlation , and therefore of inter-individual interaction , is not in contradiction with the fact that a swarm almost invariably forms in proximity of a marker . The effect of the marker ( external force ) is merely to keep the swarm at a stationary position with respect to the environment . However , as we have shown in the case of the non-interacting swarm , this stationarity ( which superficially would seems the only visible trait of swarming ) , cannot by itself produce the observed strong correlations . By using Vicsek model as a simple conceptual framework , we have shown that this coexistence of disorder and correlation is a general feature of systems with alignment interaction close to their ordering transition . We should be careful in interpreting our data as proof that explicit alignment is the main interaction at work in swarms . What we can say is that non-trivial alignment correlation implies effective alignment interaction . However , how this effective alignment interaction is achieved in terms of sensorimotor processes is hard to tell . In fact , as we have already remarked , it is possible that models purely based on repulsion/attraction positional forces , lead to correlations similar to the ones we reported here . Hence , as always when dealing with animal behaviour , it is important to keep in mind the intrinsically effective nature of any interaction . The Vicsek model provides the simplest and most compelling description of collective behaviour when effective alignment is present and this fact is not hindered by the real , non-effective nature of the interaction giving rise to the observed correlations . Our results suggest that correlation , rather than order , is the most significant experimental signature of collective behaviour . Correlation is a measure of how much and how far the behavioural change of one individual affects that of other individuals not directly interacting with it . Our data show that in swarms correlations are so strong that the effective perception range of each midge is much larger than the actual interaction range . If the change of behaviour is due to some environmental perturbations , such large correlation guarantees that the stimulus is perceived at a collective level . A notion of collective behaviour based on correlation is more general and unifying than one based on order . For example , bird flocks and insect swarms look like completely different systems as long as we stick to collective order . However , once we switch to correlation , we understand that this big difference may be deceptive: both flocks and swarms are strongly correlated systems , in which the effective perception range , or correlation length , is far larger than the interaction range [19] . In this perspective , the striking difference in emergent order between the two systems , namely the fact that flocks move around the sky , whereas swarms do not , may be related to different ecological factors , rather than to any fundamental qualitative difference in the way these systems interact . Strong correlations similar to those found in bird flocks and midge swarms have also been experimentally measured in neural assemblies [45] . This huge diversity - birds , insects , neurons - is bewildering but fascinating , and it suggests that correlation may be a universal condition for collective behaviour , bridging the gap between vastly different biological systems .
Data were collected in the field ( urban parks of Rome ) , between May and October , in and in We acquired video sequences using a multi-camera system of three synchronized cameras ( IDT-M5 ) shooting at fps . Two cameras ( the stereometric pair ) were at a distance between and depending on the swarm and on the environmental constraints . A third camera , placed at a distance of from the first camera was used to solve tracking ambiguities . We used Schneider Xenoplan lenses . Typical exposure parameters: aperture , exposure time Recorded events have a time duration between and seconds . No artificial light was used . To reconstruct the 3d positions and velocities of individual midges we used the techniques developed in [23] . Wind speed was recorded . After each acquisition we captured several midges in the recorded swarm for lab analysis . A summary of all swarms data can be found in Table S1 in Text S1 . We recorded swarms of midges belonging to the family Diptera:Ceratopogonidae ( Dasyhelea flavifrons ) and Diptera:Chironomidae ( Corynoneura scutellata and Cladotanytarsus atridorsum ) . Midges belonging to the family Chironomidae were identified to species according to [46] , the ones belonging to the family Ceratopogonidae were identified according to [47] and [48] . Specimens used for identification were captured with a hand net and fixed in alcohol , cleared and prepared according to [49] . Permanent slides were mounted in Canada Balsam and dissected according to [50] . Species identification was based on morphology of the adult male , considering different characters , as wing venation , antennal ratio ( length of apical flagellomere divided by the combined length of the more basal flagellomeres ) and genitalia , which in Diptera are named hypopygium ( a modified ninth abdominal segment together with the copulatory apparatus - see Fig . 1 ) . Let be the set of coordinates at time and at the next time step . To simplify the notation we set The velocity vector of insect is defined as , To compute the connected correlation function we need to subtract the contribution of all collective modes from the individual velocity . We identify three collective modes: translation , rotation and dilatation ( expansion/contraction ) . Translation: Let be the position of the centre of mass , and the position of the -th object in the centre of mass reference frame . By subtracting the centre of mass velocity , from the individual velocity , we obtain the translation-subtracted fluctuation , ( 5 ) Rotation: The optimal rotation about the origin is defined [51] as the orthogonal matrix which minimizes the quantity By subtracting the overall translation and rotation , the velocity fluctuation is , ( 6 ) Dilatation: The optimal dilatation is defined [51] as the scalar that minimizes the quantity After subtracting the optimal translation , rotation and dilatation , the velocity fluctuation is finally given by , ( 7 ) where with we have indicated the contribution to the velocity of of all three collective modes . The rotational order parameter is defined as , ( 8 ) where is the projection of on the plane orthogonal to the axis of rotation , the operator indicates the cross product , and is a unit vector in the direction of the axis of rotation . In ( 8 ) , is the angular momentum of midge with respect to the axis In a perfectly coherent rotation , all individuals would have angular momenta parallel to the axis , so that In a non-coherent system , some of the projections of the angular momentum on would be positive and some negative , so Note that is the axis of rotation defined in the previous section , computed using Kabsch algorithm [51] . The dilatational order parameter is defined as , ( 9 ) and it measures the degree of coherent expansion ( positive ) and contraction ( negative ) of the swarm . In a perfectly coherent expansion/contraction would be parallel to and so the scalar product in equation ( 9 ) will be for an expansion and for a contraction . In the study of flocks [19] , we normalized by its limiting value for which is equivalent dividing it by the value in the first bin . In that way the normalized correlation function tends to for so that its value is amplified . In the study of flocks we were only looking at the correlation length , which is not altered by such a normalization . However , here we will be interested in both the range and the intensity of the correlation , so we must not amplify artificially the correlation signal . Normalising the fluctuations as in ( 1 ) is equivalent normalising the correlation function by its value at exactly i . e . for which is different from its limit for The NHS is an elementary model of non-interacting particles performing a random walk in a three-dimensional harmonic potential . The dynamics of each particle is defined by the Langevin equation , ( 10 ) where is the position of the -th particle at time is the mass , the friction coefficient , the harmonic constant and is a random vector with zero mean and unit variance , with Clearly , in this model there is no interaction between particles . The parameter tunes the strength of the noise . The equation of motion is integrated with the Euler method [52] . We simulated the NHS in the critically damped regime ( ) , which gives the best similarity to natural swarms . The number of particles is set equal to that of the natural swarm we want to compare it with . Parameters have been tuned to have a ratio between the distance travelled by a particle in one time step ( frame ) and the nearest neighbour distance comparable to natural swarms , Let us define a data set as a collection of pairs of variables , with ( for example , the susceptibility as a function of the rescaled nearest neighbour distance - Fig . 4 ) . The null hypothesis is that are independent variables . Let us call the Spearman's rank correlation coefficient for a set of data and the probability distribution of in the case of pairs of independent variables . Given the empirical data , we calculate the Spearman's rank correlation coefficient and get a certain value , The P-value is defined as the probability that the statistical test we are using ( Spearman ) gives a result at least as extreme as the one actually observed , provided that the null hypothesis is true . Hence , the P-value is given by , ( 11 ) Basically , the P-value is telling us how likely it is that the degree of correlation that we observe is just the result of chance . In absence of an a priori model of the noise , we estimate by a permutation test [53] , [54]: using the original paired data , we randomly redefine the pairs to create a new data set where the are a permutation of the set we calculate the Spearman's rank correlation coefficient of this new randomized data set; we iterate this permutation times; we compute the fraction of permutations that give This fraction is equal to the P-value of the data set under consideration [53] . We performed numerical simulations of the Vicsek model in 3d [35]–[38] , [55] . The direction of particle at time is the average direction of all particles within a sphere of radius around ( including itself ) . The parameter is the metric radius of interaction . The resulting direction of motion is then perturbed with a random rotation ( noise ) . Natural swarms are known to form close to a marker and to keep a stationary position with respect to it [13] . To mimic this behaviour we modified the Vicsek model by adding an external harmonic force equal for all particles . This potential also grants cohesion , without the need to introduce an inter-individual attraction force [4] , [16] , [27] . The update equation for velocities is therefore given by , ( 12 ) where is the spherical neighbourhood of radius centred around is the normalization operator , and performs a random rotation uniformly distributed around the argument vector with maximum amplitude of The term is the harmonic force directed towards the origin . For we recover the standard Vicsek model . The update equation for the positions is , Thanks to the central force we can use open boundary conditions . All particles have fixed velocity modulus Each simulation has a duration of time steps , with initial conditions consisting in uniformly distributed positions in a sphere and uniformly distributed directions in the solid angle . After a transient of time steps , we saved 500 configurations at intervals of 1000 time steps in order to have configurations with velocity fluctuations uncorrelated in time . The control parameter of interest is where is the nearest neighbour distance , which is tuned by The model displays a transition to an ordered phase when We studied the susceptibility for different values of To observe the power-law behaviour of predicted by the model we performed standard finite-size scaling [38]: at each fixed value of the system' size we calculated and worked out the maximum of the susceptibility and its position we finally plotted vs . parametrically in to obtain the function in Fig . 5 . The noise , affects the position of the transition point [35]–[37] , but this is irrelevant for us , because we do not use any quantitative result from the model to infer the biological parameters of real swarms . The data reported in Fig . 5 have For each frame we run a clustering algorithm with scale [56]: two points are connected when their distance is lower than For each value of we compute the ratio between the number of objects in the largest cluster and the total number of objects in the swarm ( Fig . 6 ) . The percolation threshold , is defined as the point where a giant cluster , i . e . a cluster with size of the same order as the entire system , forms [42] . We define as the point where The percolation threshold scales with the nearest neighbour distance , ( Fig . 6 ) . Strictly speaking , the percolation argument only holds at equilibrium , because in a system where particles are self-propelled there may be order even at low density [36] . However , at low values of the noise , we still expect the percolation argument to give a reasonable , albeit crude , estimate of the perception range . | Our perception of collective behaviour in biological systems is closely associated to the emergence of order on a group scale . For example , birds within a flock align their directions of motion , giving the stunning impression that the group is just one organism . Large swarms of midges , mosquitoes and flies , however , look quite chaotic and do not exhibit any group ordering . It is therefore unclear whether these systems are true instances of collective behaviour . Here we perform the three dimensional tracking of large swarms of midges in the field and find that swarms display strong collective behaviour despite the absence of collective order . In fact , we discover that the capability of swarms to collectively respond to perturbations is surprisingly large , comparable to that of highly ordered groups of vertebrates . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"behavioral",
"ecology",
"ecology",
"and",
"environmental",
"sciences",
"ecology",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"biophysics",
"biophysical",
"simulations"
] | 2014 | Collective Behaviour without Collective Order in Wild Swarms of Midges |
There is great interindividual variability in HIV-1 viral setpoint after seroconversion , some of which is known to be due to genetic differences among infected individuals . Here , our focus is on determining , genome-wide , the contribution of variable gene expression to viral control , and to relate it to genomic DNA polymorphism . RNA was extracted from purified CD4+ T-cells from 137 HIV-1 seroconverters , 16 elite controllers , and 3 healthy blood donors . Expression levels of more than 48 , 000 mRNA transcripts were assessed by the Human-6 v3 Expression BeadChips ( Illumina ) . Genome-wide SNP data was generated from genomic DNA using the HumanHap550 Genotyping BeadChip ( Illumina ) . We observed two distinct profiles with 260 genes differentially expressed depending on HIV-1 viral load . There was significant upregulation of expression of interferon stimulated genes with increasing viral load , including genes of the intrinsic antiretroviral defense . Upon successful antiretroviral treatment , the transcriptome profile of previously viremic individuals reverted to a pattern comparable to that of elite controllers and of uninfected individuals . Genome-wide evaluation of cis-acting SNPs identified genetic variants modulating expression of 190 genes . Those were compared to the genes whose expression was found associated with viral load: expression of one interferon stimulated gene , OAS1 , was found to be regulated by a SNP ( rs3177979 , p = 4 . 9E-12 ) ; however , we could not detect an independent association of the SNP with viral setpoint . Thus , this study represents an attempt to integrate genome-wide SNP signals with genome-wide expression profiles in the search for biological correlates of HIV-1 control . It underscores the paradox of the association between increasing levels of viral load and greater expression of antiviral defense pathways . It also shows that elite controllers do not have a fully distinctive mRNA expression pattern in CD4+ T cells . Overall , changes in global RNA expression reflect responses to viral replication rather than a mechanism that might explain viral control .
There has been a recent effort to identify the genomic determinants of susceptibility to HIV-1 infection , control of viral replication , and disease progression [1] . Genetic analyses have identified over the years a number of validated variants in candidate genes , while a recent genome-wide association study [2] highlighted the dominant role of variants in the MHC region in the control of viral setpoint ( the steady state of viral replication after infection ) and disease progression . Other genome-wide studies [3]–[5] confirmed the variants identified in the first genome-wide analysis . These variants collectively explain up to about 13% of the variation in viral setpoint , indicating that other biological determinants of control have yet to be identified . Here our focus is on determining the contribution of variable gene expression to viral control , and to relate it to genomic DNA polymorphism . There have been a number of transcriptome studies in HIV-1 target cells ( CD4+ T cells , monocytes/macrophages ) , non-targets such as NK cells and B cells , and of dendritic cells and total peripheral blood mononuclear cells ( PBMCs ) ( reviewed in [6] , and recent publications [7]–[11] ) . These studies provide insight into gene expression changes associated with virus replication and persistence . Studies are limited by the number of genes interrogated , or by the number of individuals investigated . These limits notwithstanding , microarray data have yielded novel mechanisms of HIV-mediated pathogenesis . Transcriptome analyses of cell lines transfected with individual viral proteins or mutant viruses have also been reported ( reviewed in [6] ) . This study aims at coupling a large scale assessment of gene expression in purified CD4+ T cells from HIV-1 infected individuals , with genome-wide genotype data tested for association with viral setpoint . Integrating gene expression data with results from genome-wide association studies may help prioritize fine-mapping efforts and provide shortcuts to disease biology [12] . Therefore , the goals of the study are the description of the expression program associated with HIV-1 in vivo , the identification of mRNAs that are differentially expressed in individuals that present effective control of viral replication , and the search for cis-acting variation in differentially expressed genes . Expression polymorphism due to single nucleotide polymorphisms ( SNPs ) that influence mRNA levels has received increasing attention for the understanding of phenotypes in health and disease ( reviewed in [12] ) . Genome-wide screens , most generally done in cell lines , have established the relevance of cis-acting SNPs in expression polymorphism [13] . However , little is known regarding expression polymorphism and HIV-1 disease . In order to have power to detect correlations , we have considered a large sample set of purified CD4+ T cells from individuals with known date of infection and carefully determined viral load results . Transcription analysis was done at the time of viral setpoint , so that samples are representative of the steady-state replication for a given individual , and across the full range of viral setpoint in an infected population . For a large subset of participants , we also established the transcription profile after initiation of antiretroviral therapy ( ART ) to assess the modulation of expression upon effective control of viremia . Thus , this study represents a first attempt at assessing , genome-wide , the genotype-to-transcriptome-to-clinical phenotype associations in HIV-1 disease .
We identified 298 hybridization probes that were significantly correlated with viral load ( FDR-adjusted p-value<0 . 01 ) . This resulted in a list of 260 genes , since multiple probes are used for some of the genes . The majority of these ( n = 209 ) were positively associated with viral setpoint , while a smaller group ( n = 51 ) was negatively associated ( Supplementary Table S1 ) . We used ( unsupervised ) clustering to group the expression profiles of the samples for these 260 genes , and found that they showed distinct behavior in individuals with effective virus control ( reflected in low viral setpoint ) as compared with individuals showing poor control of viral replication ( Figure 1 ) . In an analysis that considered viral load at the precise date of transcriptome analysis instead of setpoint , the results were comparable , implying that the expression profile is representative for the period of analysis ( three months to three years after seroconversion ) , Figure 1 . The analysis included various parameters as covariates ( clinical center , gender , age , CD4 T cell viability and laboratory date , and microarray chip batch - sentrix ID ) . The CD4 T cell value at the time of sampling was found to be closely correlated with viral setpoint ( Pearson's correlation of −0 . 5 and a p-value = 1 . 195e-10 ) , which made difficult to separate their effects on the data . The 149 genes that are shared between analysis using CD4 T cell count , and the analysis using viral setpoint are indicated in Supplementary Table S1 . The main gene clusters exhibiting a positive correlation with viral setpoint ( i . e . , increasing gene expression with increasing viral load ) , as defined by STRING , DAVID and IPA , were the interferon pathway , the proteasome , and cell cycle genes ( Figure 2 and Supplementary Table S2 , S3 ) . Conversely , among genes that exhibited a negative correlation with viral setpoint ( Supplementary Table S1B ) , no pathway enrichment was identified . A separate analysis that used a gene-by-gene modeling approach resulted in a list of significant genes that was shorter ( 44 genes ) but highly concordant with the output of the empirical Bayes analysis described above ( Supplementary Table S4 ) : we therefore used the empirical Bayes results for subsequent analyses . Because the CD4+T cell composition may vary depending on the degree of viral replication [7] , we re-analyzed the data controlling for CD25 expression ( encoding IL2RA as marker of activation ) , or CD62L , CD40L , CD11a , and CD27 ( markers that distinguish naive from memory CD4+ T cells ) . Although several additional significant genes were found using each of the above markers as covariates , the overall expression profile did not vary significantly ( see for example data from analysis adjusted by CD25 in Supplementary Table S5 ) . These analyses indicate the existence of a clear expression program associated with high viral load , but fail to identify definite gene networks associated with viral control . We observed a linear association between increasing expression of interferon signaling and interferon-stimulated genes ( ISGs ) and increasing viral setpoint . We compiled a list of 40 genes implicated in the interferon response [14] ( Supplementary Table S6 ) . Seventeen genes were significantly associated with viral setpoint after FDR adjustment at the 0 . 01 level , and 12 were associated at a p-value of 0 . 05 . These 29 genes comprise most of the signaling and ISGs , but notably exclude the interferon genes themselves and the interferon receptors ( Figure 3 ) . This analysis points to a de-regulated interferon response that associates with an ineffective antiviral response . We similarly examined in detail a list of selected genes reported to be involved in HIV-1 life cycle or pathogenesis ( see Methods for explanation of candidate selection ) [15] . Of this list , 138 genes were matched to probes , with four having a FDR-adjusted significant association with viral setpoint , p-value <0 . 01: TRIM22 , IRF7 , RANBP1 , and APOBEC3G . An additional 12 genes had FDR-adjusted p-values <0 . 05 , and a further 26 had nominal p-values <0 . 05 ( Supplementary Table S7 ) . Genes of the intrinsic cellular defense against retroviruses ( TRIM5α , TRIM22 , TRIM19/PML , APOBEC3G , APOBEC3F , APOBEC3H , PPIA/Cyclophilin A , BST2/Tetherin ) were all upregulated with increasing viral load , which is consistent with their general dependence on the interferon pathways . A number of chemokines and chemokine receptors were also positively modulated with increasing viremia . We also identified differentially expressed genes that are present in both the current analysis and studies that used siRNA or shRNA to identify HIV-1 dependency factors [16]–[19] ( Supplementary Table S8 ) . The significant association of a number of genes and pathways with viral setpoint was further assessed by observing the changes in transcriptional profile in CD4+ T cells after viral suppression . We found statistical support for differential expression of 247 probes ( FDR-adjusted p-value <0 . 01 ) between treated and untreated-noncontroller individuals . The list of genes involved had an extensive overlap with the list of genes associated with viral setpoint in the transcriptome analysis above ( Supplementary Table S9 ) . The list also shares 97 genes with the recent study by Li et al . [20] on changes in the lymph node transcriptome profile upon initiation of ART . This analysis indicates that successful treatment appears able to recapitulate the cellular state of a well-controlled individual , since we did not find support for any probes being differently expressed between successfully treated and untreated-controller individuals . To compare the treated and untreated individuals with uninfected individuals , we clustered the expression profiles from samples from a selected group of individuals , including elite controllers ( viral load <50 copies/ml ) , samples from successfully treated individuals and their paired untreated samples , and from the three uninfected individuals ( measured in triplicate , one triplicate failed analysis ) . For this , we restricted analysis to the 260 genes found to be differently expressed by viral setpoint . As shown in Figure 4 , both the successfully treated and uninfected individuals tended to cluster with the controllers individuals . Two of the individuals from healthy donors were most tightly grouped with several of the untreated individuals that have the lowest level of virus at setpoint ( i . e . the elite controllers ) , while one uninfected individual showed a profile that is less extreme , but still most similar to the viral control profile . Treated individuals also preferentially grouped with the viral control pattern , although the majority showed a mid-range expression level and a smaller fraction grouped with elite controllers and uninfected individuals . A bootstrapping analysis showed support ( p-value 0 . 06 ) for the consistency of the top-level groupings with one group containing all the uninfected individuals and the majority of the treated and elite controller individuals , while the other group contained mostly non-controller individuals . This indicates that the expression levels of individuals with the best viral control closely resemble those of uninfected individuals . Among a total of 1 . 3 million association tests comparing 399 , 626 gene-centric SNPs ( some SNPs were within 100 kb of multiple transcripts ) with 28 , 828 individual probes measuring a total of 18 , 059 unique transcripts , we detected 782 study-wide significant associations ( SNP-probe associations ) below the threshold p-value of 3 . 8×10−8 . Stepwise linear regression was used to prune out redundant associations of SNPs with a particular probe . This step resulted in evidence for cis-regulation of 208 unique probes , 157 of which were regulated by multiple SNPs in high linkage disequilibrium between SNPs included in the analysis ( 51 signals of unique SNPs with a transcript , and 731 signals arising from the regulation of 157 transcripts by multiple non-unique SNPs ) . These 208 associations included 193 SNPs that modulate 190 genes in CD4+ T cells , with the overlap occurring because of probe cross-hybridization , and also several probes detecting the same gene ( Supplementary Table S10 ) . This list of study-wide significant associations was compared to the list of genes whose expression was found associated with viral load at setpoint . Among genes under differential expression during HIV-1 infection , several showed evidence for cis-regulation ( Supplementary Table S11 ) but only one , involving the interferon stimulated OAS1 , reached study-wide significance . OAS1 was found to be regulated by an intronic SNP ( rs3177979 ) located near exon 6 ( Supplementary Figure S1 ) . Lower expression was associated with the rs3177979 GG genotype . The association was detectable in treated and untreated individuals; however the expression level was lower in samples from treated individuals . The association of this SNP with OAS1 transcript expression is also detectable in PBMCs collected from uninfected controls [21] . We did not observe an association of OAS1 rs3177979 with viral setpoint in the study ( untreated ) population . However , given the potential interest of genetic polymorphism in OAS1 , we also assessed the association between rs3177979 and HIV-1 outcomes in a large population of 2362 individuals [5] . The association p-values were 0 . 05 for an association of the OAS1 SNP and viral setpoint and 0 . 09 for HIV-1 disease progression , but differences were subtle: mean HIV-1 load was 4 . 11 log10 viral copies/ml for the AA genotype , 4 . 07 for AG , and 4 . 01 for GG . Because rs3177979 is in linkage disequilibrium with rs10774671 , a SNP associated with a splicing variant ( [22] and Text S1 ) reported to have greater activity against West Nile virus [23] , we re-genotyped the population for this putative functional SNP , without finding any stronger association: we have therefore no definitive evidence of an association of cis-acting genetic variation in OAS1 with HIV-1 viral control or disease progression . One additional gene , RANBP1 , encoding a Ran GTPase-binding protein that interferes with Rev-mediated expression of HIV-1 [24] , presented both increased expression at higher setpoint , and a cis-acting SNP ( rs2008591 ) that modulates its expression ( Supplementary Table S11 ) . We assessed the association between rs2008591 and viral setpoint and disease progression in the large population of 2362 individuals [5] . Here , rs2008591 did not associate with viral setpoint ( p = 0 . 45 ) or disease progression ( p = 0 . 35 ) . Overall , these analyses identified a significant number of cis-acting genetic variants influencing gene expression in CD4+ T cells; however , expression polymorphism , genome-wide or among genes that are modulated during HIV-1 infection , did not contribute in a significant fashion to viral control .
This study represents the largest effort to date to characterize the mRNA expression profile in CD4+ T cells in vivo in HIV-1 infected individuals . The study population , only including individuals with known date of seroconversion or elite controllers , represents the complete range of viral load control: from undetectable viral load to sustained high levels of viral replication . The study also analyzed changes in transcriptome upon successful antiretroviral therapy . In addition , we searched for cis-acting variants – SNPs that would possibly associate with the observed differences in gene expression in the course of HIV-1 infection . Overall , changes in RNA expression reflect responses to viral replication rather than a mechanism that might explain control of viral replication . As such , the reactive transcriptome profile we observed shares common responses with other viral infections , eg . to dengue virus [25]–[28] ( Supplementary Table S12 ) . In vivo HIV-1 infection results in a distinctive mRNA transcriptome profile in CD4+ T cells that involves 260 genes in an analysis that differentiates individuals with high and those with low viral setpoint . Under conditions of high viral load , there is a distinct upregulation of the interferon pathways , cell cycle and the ubiquitin-proteasome degradation machinery . The study confirms and extends previous analyses of in vitro infection of T cell lines , or of CD4+ T cells in vivo that were performed on a limited number of individuals [7]–[10] , [29] , [30] . This study underscores that the observed increase in transcription of ISGs is not associated with a better control of viremia [7] . This contrasts with the reported efficacy and possible therapeutic role of interferon ( IFN-α , IFN-α2β ) suggested by results from in vitro studies , while exogenous administration of interferon in clinical trials led to doubts about its efficacy in the clinical setting ( reviewed in [31] ) . Our observations lend support to the hypothesis that interferon activation plays a deleterious role in retroviral pathogenesis , as proposed by many recent reports ( reviewed in [31] ) . Elevated ISG expression is associated with disease progression in pathogenic SIV infection of non-human primates [32]–[35] , while the type I interferon response subsided after peak viral load during non-pathogenic infection [36] , [37] . Sedaghat et al . [7] compared the transcriptional programs of in vivo-activated CD4+ T cells from untreated HIV-positive individuals with those of activated CD4+ T cells from HIV-negative individuals . From this study , they concluded that CD4+ T cells from infected individuals are in a hyperproliferative state that is modulated by type I interferons , and that this would lead , during chronic infection , to CD4+ T-cell preferential differentiation and depletion . Imbeault et al . [10] suggested that interferon could lead to a sustained increase in p53 mRNA levels and therefore to a higher susceptibility of CD4+ T cells to pro-apoptotic signals . Herbeuval and Shearer [31] proposed that interferon , through binding to its receptor on primary CD4+ T cells would result in membrane expression of the TNF-Related Apoptosis-Inducing Ligand , TRAIL , death molecule leading to the selective death of HIV-exposed CD4+ T cells . More recently , Sato et al . [38] showed that type I interferon induce proliferation and exhaustion in hematopoietic stem cells; chronic and excessive type I interferon signaling may cause hematopoietic stem cells reduction . Overall , interferon response appears a poorly effective antiretroviral mechanism , and may actually contribute to HIV-1 disease [7] , [39] . Among genes previously associated with HIV-1 pathogenesis , the analysis identified a number of significant associations , in particular for genes of the intrinsic cellular defense against retroviruses . Many of these respond to interferon , and thus have the same profile of increased expression with increasing viral load as ISG . Thus , these genes appear ineffective both by their poor specificity against HIV-1 and by the apparent limited response of HIV-1 to increasing titration of the transcripts . We also analysed genes issued from four genome-wide siRNA/shRNA screens [16]–[19] . Fifteen genes that were associated with decreased cellular permissiveness to infection after silencing , were upregulated with increasing viremia in vivo in the current study . They deserve further inspection for a role in HIV-1 pathogenesis . Although the scope of the present work was not to complete a meta-analytical study of all available genome-wide transcriptome studies and siRNA screens [40] , we are aware of the interest to progressively integrate large scale datasets [41] . We aimed at identifying patterns of gene expression associated with effective viral control . However , the nature of the analysis could not establish whether high levels of viral replication would lead to the observed transcriptional profile , or whether genetic modifiers of transcriptional profile were determinants for the control of viral replication . This was addressed first by comparing the transcriptional profile of CD4+ T cell from elite controllers with that of successfully treated individuals and healthy donors . Here , we observed that the expression profile of genes associated with active viral replication was , after effective treatment , similar to that of individuals with spontaneous control of viral replication , and close to that of healthy donors . This suggests that infection drives gene expression rather than the contrary . Second , we tested the hypothesis that genetic variants influence expression levels of genes , thus leading to differences in viral control . The analysis identified a number of variants that would possibly act in cis to modulate gene expression – most notably a variant in OAS1 that has been associated with improved control of West Nile virus infection [23] . It may be argued that if a variant influences expression of a gene , and if expression of that gene correlates with viral load , then the two analyses will be partially redundant . However , we emphasize that this approach allows for independent information because the variation in expression of few if any genes is determined exclusively by cis-acting variants . In addition , the identification of strong cis-acting variants would contribute to disentangle causation and correlation . Thus if a gene expression correlates with viral load , it could be that the change in expression is a response to the amount of virus , or it that the gene directly controls the viral level . In the former case , a cis-acting variant will show no association with viral load , whereas in the latter it will . In the present study , none of the candidate cis-acting SNPs , or SNPs in the implicated genes was associated with differences in viral setpoint in a genome-wide association analysis . These results do not contradict current evidence of mechanisms of viral control through differences in expression levels of particular genes , most notably CCR5 [42] . Rather , the analysis indicates that polymorphisms in genes implicated in the differential expression programs do not represent a strong source of variation at the population level . There are a number of technical and conceptual limits to the study . The study failed to identify a transcriptome profile characteristic of elite controllers . This may be attributed to the large scale approach , as the current technology covers a total of 25 , 440 annotated human genes . While this allows for pathway or network analyses , it may fail in the identification of subtle expression changes , in particular at the level of the individual gene . On one hand , the analysis would require greater study power ( ie , additional elite controllers ) to compensate the penalty of correction for multiple testing . On the other hand , the precision of several of analyses described earlier in this section could be improved through the added resolution of new technology such as RNA-Seq [31] , or the targeted multiplexed measurement of gene expression in selected pathways [43] . High-throughput deep sequencing results in a superior dynamic range , and allows quantitative analysis of coding and non-coding region transcripts , such as small RNAs . It should also be pointed out that the use of cryopreserved cells may result in changes in the transcriptome and in transcript stability . However , this allowed the investigation of a large number of samples from seroconverting individuals in batch analyses . We argue that the internal consistency of the results and the general agreement across studies supports the robust nature of the transcription profiles that were generated . In conclusion , while this study suggests that the generalized upregulation of ISG , an important component of viral defense , does not lead to consistently improved viral control throughout the course of infection , it does not implicate any specific gene expression network in viral control . There are several possible explanations for these observations . First , the most important cellular populations for determining control may be effector cells such as CD8+ T cells or NK cells whose expression patterns have not been evaluated here . Second , the key expression patterns that determine eventual control may be only detectable early in infection and thus largely missed in studies focusing on cells taken during the setpoint period . These possibilities argue strongly that the next phase of expression work in the study of HIV-1 control must focus on large scale analysis of isolated populations of effector cells taken from individuals as early in the course of infection as possible and in a standardized fashion . We believe the approach taken here provides a general template for such studies .
Study participants were followed in the Swiss HIV Cohort Study ( www . shcs . ch ) . The Genetics Project of the Swiss HIV Cohort Study was approved by the ethics committees of all participating centers , and the permission for genomic work was approved by the Institutional Review Board/Ethics Committee of the University Hospital of Lausanne . Participants gave written , informed consent for genetic testing . 198 HIV-1 infected individuals from the Swiss HIV Cohort study with a known date of seroconversion ( n = 182 ) , or elite controllers ( n = 16 ) were included in the study . Seroconversion was defined on the basis of a documented positive test and date and a documented negative test less than two years before the first positive test . The viral setpoint was calculated for each participant by using a median of 4 ( range 2 to 8 ) plasma HIV-1 RNA determinations obtained in the absence of antiretroviral treatment between 3 months and 3 years after seroconversion , as previously described [2] . See Text S1 for the detail definition of viral setpoint and of elite controllers . When available , HIV-1 infected participants contributed samples during stable viral setpoint before and under effective ART ( median [IQR] from treatment initiation to sample collection was 1297 ( 434–2730 ) days ) . In addition three healthy blood donors provided three control samples used as biological replicas . Quality control steps at the level of cellular viability , RNA integrity , microarray and hybridization quality , and data analysis led to a final number of 190 samples from 153 participants and 8 samples from 3 healthy controls ( 68% of valid samples , 78% of successful recruitment ) . The demographic characteristics of the patients and the flow chart of enrollment and sample validation is presented in Text S1 . Representative examples of QC checks are presented in Supplementary Figure S2 . CD4+ T cells were positively selected from frozen PBMCs ( median time [IQR] of cryopreservation was 616 [333–1448] days ) using magnetically labeled CD4 microbeads and subsequent column purification according to the manufacturer's protocol ( Miltenyi Biotec ) . CD4+ T cell purity , verified by flow cytometry , was 95 . 6% ( 86 . 4–98 . 1% ) [median ( range ) ] . CD4+ T cell viability was assessed by the trypan blue dye exclusion method using the Vi-CELL ( Beckman Coulter ) . Total RNA was extracted from purified CD4+ T cells using mirVana miRNA isolation kit ( Ambion ) according to the manufacturer's protocol for total RNA extraction . RNA amount was estimated by spectrophotometry using the Nanodrop 1000 ( Thermo Fisher ) . RNA quality was determined by Agilent RNA 6000 pico kit on an Agilent 2100 Bioanalyzer . We used cryopreserved samples because of the interest to analyse a large population of seroconverting individuals during the precise window of stable viral setpoint . Samples were collected between 1995 and 2007 , and investigated in 2008 . The median ( range ) of CD4+ T cell viability for samples that were successfully analysed was 78 . 5% ( IQR 70 . 5–85 . 3 ) . Viability was minimally dependent on time of cryopreservation , and more dependent on collection center . These covariates were included in the analyses ( see below ) . 200 ng of total RNA was amplified and labeled using the Illumina TotalPrep RNA Amplification kit ( Ambion ) . cRNA quality was assessed by capillary electrophoresis on Agilent 2100 Bioanalyzer . Expression levels of over 48 , 000 mRNA transcripts were assessed by the Human-6 v3 Expression BeadChips ( Illumina ) . Hybridization was carried out according to the manufacturer's instructions . Genome-wide SNP data had been generated from genomic DNA using the HumanHap550 Genotyping BeadChip ( Illumina ) with 555 , 352 SNPs [2] . We screened the literature for genes associated with biology of HIV-1 ( reviewed in [44]–[47] and recent studies [15] , [48] , [49] ) , as well as HIV-1 dependency factors emerging from genome-wide siRNA screens [16]–[19] , and genes considered polymorphic and involved in HIV-1 pathogenesis ( compiled in www . hiv-pharmacogenomics . org ) . For the three large siRNA screens , that resulted in over 600 candidates , we restricted analysis to ( i ) genes identified in at least two of three screens , or to ( ii ) genes with SNPs that reached a nominal significant p value in a recent genome-wide association study of determinants of susceptibility to HIV-1 [2] . Bead summary data was output from Illumina's BeadStudio software without background correction , as this has previously been shown to have detrimental effects [50] . Data pre-processing , including a variance-stabilizing transformation [51] and robust-spline normalization were applied as implemented in the lumi package [52] of R . Four outlier samples identified based on aberrant expression of control probes and aberrant median-interquartile range values compared to other samples were removed . We applied an empirical Bayes analysis approach within a linear mixed-model framework to identify associations between variation in gene expression and in viral setpoint . The Empirical Bayes approach has been developed to model the variation profiles of all genes and use that information as prior knowledge to better estimate the variance of each gene expression [53]–[55] . In addition , we used a more conservative gene-by-gene modeling approach for result comparison with the empirical Bayes approach . We controlled for variation caused by gender , age , CD4+ T cell viability , location of sample collection , and laboratory batch effects . Effect of chip batch was modeled as a random effect; all others were fixed or continuous . All samples from untreated individuals were tested for association of expression with viral setpoint . We used a false discovery rate ( FDR ) method [56] to control for multiple testing . Probes selected for further analysis had an FDR-adjusted p-value <0 . 01 . A separate analysis compared expression in samples from treated and untreated individuals , using a similar mixed-model approach as above , but also incorporating viral load as a factor in the analysis . We tested for effect of treatment by separately comparing samples from treated individuals to each of the untreated groups , using the limma ( linear models for microarray data ) package in R with FDR adjustment as above . This analysis explicitly excluded samples from the same individuals because the statistical approach did not allow control for both the correlation between paired samples and the strong correlation ( batch ) effect of chip . To compare samples from treated and untreated individuals with samples from uninfected controls , we clustered the expression profiles for a selected group of individuals . We performed 1000 replicate clusterings on the Pearson correlation coefficient , using the “ward” clustering method as implemented in the pvclust package in R . The Search Tool for the Retrieval of Interacting Genes/Proteins ( STRING ) ( http://string . embl . de/ ) was used to identify known and predicted interactions ( derived from four sources: genomic context , high-throughput experiments , co-expression , and previous knowledge ) . DAVID Bioinformatic resources ( http://david . abcc . ncifcrf . gov/ ) using the annotation sources GOTERM-BP ( biological process ) , and GOTERM-MF ( molecular function ) identified functional categories [57] . Ingenuity Pathway Analysis 7 . 0 ( IPA ) ( http://www . ingenuity . com/ ) was used for the analysis of pathway enrichment . Analysis was limited to genes significantly associated with viral load ( FDR p-value <0 . 01 ) . Normalized expression data was exported for all untreated , HIV-1 infected individuals ( n = 125 ) . Only probes that targeted fully annotated genes were included in the analysis . A principal components analysis was run to assess batch effects . The cis-screen consisted of a scan for common SNPs , within 100 kb of the defined gene start and stop positions , for effects on transcript expression levels . The analysis was limited to SNPs with a minor allele frequency greater than 0 . 04 , requiring at least ten alleles to be present to detect associations with a low false positive rate . This analysis was performed using a standard linear regression , incorporating age , gender , and 11 eigenstrat axes to correct for population stratification . In total , there were 1 , 330 , 529 tests run , therefore using a Bonferroni correction , a p<3 . 8×10−8 was used to declare a statistically significant association . All microarray results have been deposited in the Gene Expression Omnibus database ( GSE18233 ) . | There has been recent progress in understanding the genetic factors that modulate susceptibility to HIV-1 infection . Genetic variation explains to a certain extent differences in disease progression among individuals . Less is known regarding the contribution of differences in gene expression to viral control . The present study evaluated , genome-wide , gene expression levels in CD4+ T cell , the main target of HIV-1 . Thereafter , it searched for genetic variants that would modify gene expression . Specific expression profiles associated with high levels of viremia—in particular , the upregulation of genes of the antiviral defense . In contrast , no expression profile associated with effective viral control . Multiple genetic variants modulated gene expression in CD4+ T cells; however , none had a strong influence on viral control . This integrated genome-wide assessment suggests that viral replication drives gene expression rather than expression pointing to mechanisms of viral control . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/population",
"genetics",
"infectious",
"diseases/hiv",
"infection",
"and",
"aids",
"genetics",
"and",
"genomics/gene",
"expression"
] | 2010 | Genome-Wide mRNA Expression Correlates of Viral Control in CD4+ T-Cells from HIV-1-Infected Individuals |
Numerous plant viruses that cause significant agricultural problems are persistently transmitted by insect vectors . We wanted to see if apoptosis was involved in viral infection process in the vector . We found that a plant reovirus ( rice gall dwarf virus , RGDV ) induced typical apoptotic response during viral replication in the leafhopper vector and cultured vector cells , as demonstrated by mitochondrial degeneration and membrane potential decrease . Fibrillar structures formed by nonstructural protein Pns11 of RGDV targeted the outer membrane of mitochondria , likely by interaction with an apoptosis-related mitochondrial protein in virus-infected leafhopper cells or nonvector insect cells . Such association of virus-induced fibrillar structures with mitochondria clearly led to mitochondrial degeneration and membrane potential decrease , suggesting that RGDV Pns11 was the inducer of apoptotic response in insect vectors . A caspase inhibitor treatment and knockdown of caspase gene expression using RNA interference each reduced apoptosis and viral accumulation , while the knockdown of gene expression for the inhibitor of apoptosis protein improved apoptosis and viral accumulation . Thus , RGDV exploited caspase-dependent apoptotic response to promote viral infection in insect vectors . For the first time , we directly confirmed that a nonstructural protein encoded by a persistent plant virus can induce the typical apoptotic response to benefit viral transmission by insect vectors .
In mammals , viral infection can induce or activate apoptosis , a process of programmed cell death , which generally is important in the regulation of viral pathogenesis [1] . Apoptosis is a normal process during development and aging to regulate cell populations in multicellular organisms [2–3] . Caspases , a family of cysteine proteases , are crucial proteases responsible for the execution of the apoptotic cascade , while the inhibitor of apoptosis protein ( IAP ) serves as a pivotal regulator of apoptosis [4] . Apoptosis is triggered either via an extrinsic death receptor or an intrinsic mitochondria-dependent pathway [5–6] . The initial event of mitochondria-dependent apoptosis is the loss of mitochondrial membrane potential , leading to the release of apoptosis-related factors associated with the mitochondrial membranes [7–10] . Later , the chromatin is cleaved into nucleosomal fragments , and apoptotic bodies are generated [11] . These fundamental stages are first elucidated for mammalian systems , due to the important function of apoptosis in development and diseases [2] . Although apoptosis is commonly involved in viral pathogenesis , some viruses appear to have evolved to exploit this mechanism to promote their survival and replication in different ways [12–14] . Thus , the role of apoptosis in host–virus interactions is diverse among different viruses . Many plant viruses that cause significant agricultural problems are transmitted via insect vectors such as thrips , aphids , leafhoppers and planthoppers in a persistent manner [15] . Growing evidence has shown that the persistent transmission of viruses causes only a limited adverse effect , rather than pathogenesis in their insect vectors [15–20] . We now know that a conserved small interfering RNA ( siRNA ) antiviral response is triggered by the replication of viruses in the insect vectors to modulate a metastable balance between viral accumulation and adverse effects , allowing for viral persistence and highly efficient spread in nature [15 , 21–24] . Generally , persistent infection by arthropod-borne viruses ( arboviruses ) can induce apoptosis in mosquito and Drosophila vectors , but it is usually restricted to a low level to avoid serious damage to the insects [14 , 25–27] . The apoptosis induced by arboviruses may serve as an innate antiviral mechanism to protect against or benefit viral transmission by insect vectors [12 , 14] . The cytopathologic changes caused by virus-induced apoptosis may also damage functionally relevant tissues and organs , decreasing insect fitness . Using the terminal deoxynucleotidyl transferase dUTP nick-end labeling ( TUNEL ) assay , Huang et al . demonstrated that apoptotic signs appeared in limited regions of the principal salivary glands in brown planthopper infected with rice ragged stunt virus , a plant reovirus [28] . However , the functional roles for apoptosis induced by persistent plant viruses in insect vectors are still poorly understood . This gap in knowledge is , in part , attributable to the lack of reliable tools such as insect vector cell cultures for real-time analysis of virus-induced apoptosis in insect vector cells . In the present study , we used continuous cell cultures derived from insect vectors to trace the apoptosis process induced by rice gall dwarf virus ( RGDV ) , a plant reovirus that was first described in 1979 in Thailand and caused a severe disease of rice in southern China and Southeast Asia [29] . RGDV is mainly transmitted with high efficiency in a persistent-propagative manner by the leafhopper vector Recilia dorsalis ( Hemiptera: Cicadellidae ) [17] . RGDV has icosahedral and double-shelled particles approximately 65 to 70 nm in diameter [24] . Its genome consists of 12 double-stranded RNA ( dsRNA ) segments , encoding six structural proteins and six nonstructural proteins [29] . The outer capsid shell is composed of the major outer capsid protein P8 and the minor outer capsid protein P2 [24] . Nonstructural protein Pns11 assembles into fibrillar or tubular structures to facilitate viral spread in insect vectors [17 , 30–31] . The leafhopper R . dorsalis and its cultured cells support the efficient propagation of RGDV to a high titer in a persistent and nonlethal infection that causes limited damage [17 , 19] . We have demonstrated that the siRNA antiviral pathway modulates the persistent infection of RGDV in R . dorsalis , thus preventing serious harm [24] . However , we still do not know how RGDV regulates physiological processes in R . dorsalis to permit effective viral propagation . Previously , we found that RGDV infection could directly remodel and utilize a variety of cellular structures and pathways for efficient propagation in its insect vectors [30–32] . For example , RGDV particles were often observed to associate with the bundles of fibrillar structures to facilitate viral infection [30–31] . Furthermore , RGDV particles are distributed close to the periphery of degenerated mitochondria during viral replication of vector cells , suggesting that mitochondria might support the energy demands of viral propagation [33] . The degeneration of virus-associated mitochondria surprised us , suggesting that RGDV infection may induce apoptotic response in insect vector cells and adversely affect the insects . Here , by applying a system that combines insect vector cell cultures , immunofluorescence and electron microscopy , we revealed that the fibrillar structures that are composed of nonstructural protein Pns11 of RGDV , targeted mitochondria and activated typical apoptotic response to promote viral infection and transmission by insect vectors .
RGDV exerts an adverse effect on its vector R . dorsalis , including reduced survival , emergence , fecundity and longevity of adults [19] . To explore how the virus causes these adverse effects , we first investigated whether RGDV infection caused apoptotic changes in continuous cultured cells of R . dorsalis , which were originally established from embryonic fragments dissected from eggs [17] . At 72 h post-inoculation ( hpi ) with RGDV at a multiplicity of infection ( MOI ) of 1 , bright field microscopy showed that the infected cultured insect cells had slight cytopathological changes , such as cell clumping and loss of confluent monolayers ( S1 Fig ) . Electron microscopy showed that RGDV-infected cells had apoptotic characteristics , including crescent-shaped nuclei and condensed and marginalized chromatin , compared with the round nuclei and finely dispersed chromatin in mock-infected cells ( Fig 1A–1C ) . At this time , virus-containing apoptotic bodies , the typical characteristic of the end stage of apoptosis were present ( Fig 1D ) . The mitochondria in RGDV-infected cells appeared to be degenerating , and cristae were diffuse and indistinct ( Fig 1E and 1F ) . Bundles of fibrillar structures , absent in virus-free cells , were in contact with the periphery of these degenerated mitochondria ( Fig 1F ) . Some RGDV particles were closely associated with the free ends of the virus-induced fibrillar structures and along their edges ( Fig 1F ) . Thus , RGDV infection caused cytopathological changes that included the hallmarks of apoptosis . To better understand the prevalence of RGDV-induced apoptotic response in cultured cells , we examined 400 cells in mock- or RGDV-infected treatment to count the number of apoptotic cells using electron microscopy and found that 34 . 5% of the cells were apoptotic , significantly higher than in the mock treatment ( Fig 1G ) . Therefore , RGDV specifically induced specific morphology changes of apoptosis in cultured cells of R . dorsalis . One of the key characteristics of apoptosis at an early stage is the disruption of the mitochondrial membrane potential [2 , 34] . The degeneration of mitochondria in RGDV-infected cells is probably caused by a change of the mitochondrial membrane potential , which initiates a mitochondria-dependent apoptotic cascade . We thus used the JC-1 assay , a flow cytometry-based method widely used to detect any changes in mitochondrial membrane potential . Continuous cultured cells of R . dorsalis were inoculated with RGDV at a MOI of 1 . At 48 hpi , RGDV infection caused the mitochondrial membrane potential to decrease in 42 . 7% of the cultured leafhopper cells , which was significantly higher than in the mock-infected treatment ( Fig 2A ) . Thus , RGDV infection induced early-stage apoptotic response . At 72 hpi , the infected cells were then examined for nucleosomal fragmentation , a hallmark event of later-stage apoptosis [11] . At this later time , a clear ladder of DNA fragments was detected in RGDV-infected cells , but not in the mock , which had a single intact chromosomal DNA band ( Fig 2B ) . Furthermore , the TUNEL assay , widely used to detect apoptotic bodies [35] , showed positive apoptotic signals in virus-infected regions , but none in the uninfected cells ( Fig 2C ) . We calculated that about 43% of infected cells were TUNEL-positive ( Fig 2D ) . These results strongly indicated that RGDV infection specifically induced the early and later events of apoptotic response in insect vector cells . Because RGDV infection of insect vector cells has been shown to induce the formation of various inclusions composed of nonstructural proteins for viral replication or spread [29–31 , 36] , we used immunoelectron microscopy to investigate which viral nonstructural proteins ( Pns4 , Pns7 , Pns9 , Pns10 , Pns11 or Pns12 ) were involved in the formation of virus-associated fibrillar structures along the degenerated mitochondria . Immunoelectron microscopy showed that Pns11-specific IgG specifically recognized the fibrillar structures surrounding the degenerated mitochondria ( Fig 3A ) . Confocal microscopy further confirmed that some Pns11-specific fibrillar structures colocalized with the mitochondria , which were stained by MitoTracker Red in virus-infected R . dorsalis cells ( Fig 3B ) . Thus , the fibrillar structures composed of Pns11 of RGDV apparently targeted the mitochondria and may induce mitochondrial degeneration during viral infection of insect vector cells . Previously , we showed that expression of RGDV Pns11 alone can induce the formation of fibrillar or tubule-like structures in cells of the nonhost Spodoptera frugiperda ( Sf9 ) [17] . To determine whether Pns11 of RGDV had an inherent ability to target and induce mitochondrial degeneration , Sf9 cells were inoculated with recombinant baculovirus that expressed Pns11 . Confocal microscopy demonstrated that some Pns11-specific fibrillar structures colocalized with the mitochondria stained by MitoTracker Red in the cytoplasm of Sf9 cells at 48 hpi ( Fig 3C ) . Immunoelectron microscopy further showed that the fibrillar structures-associated mitochondria were degenerating , with diffuse and indistinct cristae in the cytoplasm of Sf9 cells at 48 hpi ( Fig 3D ) . These results illustrated that RGDV Pns11 had an inherent ability to target and induce mitochondrial degeneration in the absence of viral infection . To characterize the cytopathological effect of Pns11 of RGDV in Sf9 cells , we stained cells with trypan blue to observe the cellular phenotypes and measure cell viability using a cell counter . Bright field microscopy showed that the viability of Sf9 cells had greatly decreased after Pns11 accumulation at 72 hpi , indicating the relative cytotoxicity of Pns11 ( S2A and S2B Fig ) . We then used rhodamine 123 , a specific fluorescent dye used to assess mitochondrial membrane potential [37] , to determine whether Pns11 alone induces the collapse of mitochondrial membrane potential . We found that Pns11 caused an 87 . 1% decrease in total rhodamine 123 fluorescence intensity of Sf9 cells at 72 hpi , compared with that of the mock ( Fig 3E ) , suggesting that Pns11 alone potentially reduced mitochondrial membrane potential . Because the disruption of the mitochondrial membrane potential can lead to cytochrome c release [7–10] , we then determined whether the cytochrome c could translocate from the mitochondrial to cytosol . The immunofluorescence assay demonstrated that at 72 hpi , the cytochrome c in Pns11-expression Sf9 cells was largely localized within the cytosol , but not together with the mitochondria ( S2C Fig ) . Thus , Pns11 alone potentially induced the release of apoptosis-related factors and the subsequent apoptotic cascade . A yeast two-hybrid ( Y2H ) assay was then used to screen a cDNA library of R . dorsalis to identify putative mitochondrial factors interacting with RGDV Pns11 . From this library screen , 116 colonies of 207 positive ones were randomly sequenced . Finally , 36 sequences were annotated using the BLASTX program in GenBank . Among these candidates , an apoptosis-related protein named voltage-dependent anion channel ( VDAC ) ( also called mitochondrial porin ) captured our attention . The VDAC , a class of porin channel located on the outer mitochondrial membrane , serves as a major diffusion pathway for ions and metabolites [38] . This protein plays a crucial role in apoptosis [39] . At the early stage of apoptosis , the VDAC increases the permeability of mitochondrial membrane to allow the release of apoptotic factors , such as cytochrome c and apoptosis-inducing factor , then initiates the apoptotic cascade [39] . Therefore , the VDAC of R . dorsalis was analyzed further . Based on the transcriptome data from R . dorsalis in our lab , the BLASTX searching method in the GenBank demonstrated the 90% similarity of putative full-length open reading frame ( ORF ) of VDAC ( 846 bp long ) with counterpart of Homalodisca vitripennis ( S3A Fig ) . Then this putative ORF of VDAC was amplified , and the gene sequence was deposited in GenBank with accession number MG241500 . The predicted protein product ( 282 amino acid residues ) possessed characteristic porin3 domains ( S3B Fig ) , which can form a β-barrel to span the mitochondrial outer membrane [40] , but did not show the significant predicted transmembrane domains analyzed with TMHMM . The phylogenic analysis revealed that the amino acid sequence of VDAC clustered with those of other insect species in the order Hemiptera ( S3A Fig ) . Both Gal4 transcriptional activator-based and membrane-based yeast two-hybrid ( MbY2H ) systems were applied and revealed the strong interaction between VDAC and Pns11 ( Fig 4A ) . We then used a glutathione S-transferase ( GST ) pull-down assay to confirm such interaction . The GST-tag and His-tag were fused to the N-terminal of Pns11 and VDAC , respectively , to express fusion proteins GST-Pns11 and His-VDAC . The result showed that the purified GST-Pns11 pulled down His-VDAC from cell lysates ( Fig 4B ) . By contrast , no such interaction was obtained with the purified GST ( Fig 4B ) . Thus , RGDV Pns11 specifically interacted with the VDAC , suggesting that the VDAC was likely the target protein mediating the binding of Pns11-specific fibrillar structures with the mitochondria . To confirm this possibility , we then performed RNA interference ( RNAi ) experiments to test the effect of the reduced expression of VDAC gene on the interaction between Pns11 and VDAC . Cultured cells of R . dorsalis were transfected with synthesized dsRNA targeting the gene of VDAC ( dsVDAC ) , then infected with purified RGDV ( MOI of 1 ) . RT-qPCR assay demonstrated that dsVDAC treatment caused approximately 70% or 20% reduction in the relative expression of VDAC or Pns11 at 48 hpi , respectively ( Fig 4C ) , suggesting that the formation of Pns11-specific fibrillar structures was independent of VDAC . Immunofluorescence and immunoelectron microscopy further demonstrated that at 48 hpi , dsVDAC treatment significantly inhibited the colocalization of Pns11-specific fibrillar structures with the mitochondria in cultured cells of R . dorsalis ( Fig 4D and 4E ) . Taken together , our results indicated that the target of Pns11-specific fibrillar structures with the mitochondria may depend on the specific interaction of RGDV Pns11 with an apoptosis-related mitochondrial outer membrane protein . For clarifying that the apoptotic pathway is induced by RGDV infection , IAP and two caspase orthologs , caspase-2-like ( CASP2L ) and caspase-8-like ( CASP8L ) , were first identified in transcriptome data from R . dorsalis . The full-length ORFs of CASP2L , CASP8L and IAP genes of R . dorsalis were amplified , and each gene sequence was deposited in GenBank ( accessions MG241499 , MG241497 and MG241498 , respectively ) . Phylogenic analysis showed that the amino acid sequences of CASP2L and CASP8L clustered with those of other insect species ( S4 Fig ) . To determine the expression profiles of apoptosis-related genes during viral infection , cultured cells of R . dorsalis were infected with purified RGDV ( MOI of 1 ) . At 48 hpi , an RT-qPCR assay showed that the expression of three apoptosis-related genes was increased significantly ( Fig 5A ) . We further determined that the treatment with the broad-spectrum caspase inhibitor Z-VAD-FMK also significantly inhibited relative gene expression of RGDV P8 at 48 hpi , but did not significantly affect viability of R . dorsalis cells ( Fig 5B ) . Thus , the typical apoptotic response induced by RGDV infection was caspase-dependent . To confirm this result , we also silenced gene expression for CASP2L , CASP8L or IAP by RNAi to inhibit or induce apoptotic responses , respectively . Continuous cultured cells of R . dorsalis were treated with dsRNAs targeting the CASP2L , CASP8L or IAP genes or the gene for green fluorescence protein ( GFP ) ( dsCASP2L , dsCASP8L , dsIAP or dsGFP ) . At 8 h post transfection , cells were inoculated with RGDV at a MOI of 0 . 1 . At this low MOI , the early viral infection rate was low ( about 20–30% ) , and the spread of viruses among R . dorsalis cells could be easily monitored . We also tested the efficiency of the knockdown of the targeted genes 48 h post-transfection with dsRNAs in continuous cultured cells ( S5A Fig ) . By 48 hpi , after examining at least 1000 cells , immunofluorescence microscopy indicated that treatment with dsCASP2L or dsCASP8L decreased the percentage of infected cells from an average of 65% to 20% , compared with the dsGFP-treated cells control , respectively ( Fig 5C and 5D ) . In contrast , treatment with dsIAP increased the percentage of infected cells from an average of 65% to 90% , compared with dsGFP-treated cells ( Fig 5C and 5D ) . As expected , the number of TUNEL-positive cells was positively correlated with viral infection ( Fig 5C–5E ) . Meanwhile , in dsCASP2L- , dsCASP8L- or dsGFP-treated uninfected cells , no specific TUNEL signals appeared ( Fig 5C ) . However , about 10% of cells in the dsIAP-treated uninfected cells were TUNEL-positive , significantly lower than in the dsIAP-treated infected cells ( Fig 5C ) . RT-qPCR assay demonstrated an approximately 70% reduction in the relative expression of CASP2L , CASP8L or IAP after treatments with dsCASP2L , dsCASP8L or dsIAP , respectively , at 48 hpi ( Fig 5F ) , indicating that transfection with the dsRNA specific for these genes indeed triggered RNAi in R . dorsalis cells . RT-qPCR assay showed that the treatment with dsIAP increased gene expression of P8 by more than 2-fold ( Fig 5F ) . By contrast , gene expression of P8 was reduced by the treatment with dsCASP2L or dsCASP8L by about 3- or 2-fold , respectively ( Fig 5F ) . Northern blot analysis showed that the treatment of dsCASP2L and dsCASP8L resulted in a marked reduction of the synthesis of viral mRNAs , but the treatment of dsIAP increased the synthesis of viral mRNAs at 72 hpi ( Fig 6A ) . Expectedly , the synthesis of viral genome dsRNAs and the accumulation of viral proteins were also decreased by treatment with dsCASP2L or dsCASP8L , but were increased by treatment with dsIAP ( Fig 6B and 6C ) . The pattern of genomic dsRNA segments separated from purified RGDV virons is considered as the reference [41] . Thus , the apoptotic response induced by RGDV infection was beneficial for viral infection . In addition , dsIAP treatment of virus-free cells induced a low level of apoptotic response , while dsIAP treatment under viral infection significantly induced distinct apoptotic response . Totally , our results confirmed that RGDV infection inherently activated an apoptotic response in its insect vector cells . To determine whether RGDV infection caused apoptotic response in intact insect vector , the intestines of nonviruliferous or viruliferous R . dorsalis adults were tested using the TUNEL assay . It is known that RGDV initially infects the filter chamber epithelium of the intestines by 2 days post-first access to diseased plants ( padp ) , then directly crosses the basal lamina into the visceral muscles at 4 days padp , from where it spreads throughout the entire intestines at 6 days padp [42] . By 10 days padp , RGDV is extensively present in the salivary glands [42] . Usually , our test confirms that about 70% of insects can transmit RGDV to healthy rice seedlings after a latent period of 10 days . At 6 days padp , TUNEL-positive signs could be detected in limited areas of virus-infected intestines in about 70% of viruliferous insects , while few TUNEL-positive cells were found in the nonviruliferous insects ( Fig 7A and 7B ) . However , we could not detect specific TUNEL-positive signs in the virus-infected salivary glands of viruliferous insects at 10 days padp . Thus , the RGDV-induced apoptotic response appears to be restricted to a low level to avoid serious damage to R . dorsalis , similar to results reported previously for arboviruses in mosquito vectors [14 , 25–27] . Electron microscopy showed that the epithelial cells of virus-free intestines had normal histology and ultrastructure , including intact and orderly microvilli , evenly distributed chromatin , and abundant mitochondria with tightly involuted cristae ( Fig 7C ) . In contrast , at 10 days padp , morphological changes in the intestinal epithelial cells of viruliferous R . dorsali were evident , including cytoplasmic reduction and vacuolization , damaged or decreased number of microvilli , and shrunken or crescent-shaped nuclei ( Fig 7D and 7E ) . Furthermore , the degenerated mitochondria with indiscernible cristae were surrounded by bundles of Pns11 fibrillar structures in virus-infected intestines , compared with the intact mitochondria in nonviruliferous control ( Fig 7F and 7G ) . This abnormal cytopathology of virus-infected intestinal epithelium further suggested that RGDV infection also induces typical mitochondrial-dependent apoptotic response in intact insect vectors . To further investigate the effects of apoptotic response on viral infection in insect vectors , from 6 to 14 days padp , we sampled 30 live viruliferous leafhoppers daily and then examined the gene expression profiles for apoptosis-related factors ( CASP2L and IAP ) and major outer capsid protein P8 of RGDV by RT-qPCR assay . During the latent period for RGDV in its insect vectors , before 10 days padp [42] , the transcript levels of apoptosis-related genes increased , and then decreased ( Fig 8A ) . Similarly , the transcript level of viral major outer capsid protein P8 gene also increased quickly to peak at 10 days padp , and then remained relatively stable ( Fig 8A ) . Although changes in the relative expression of apoptosis-related genes are not informative regarding the biology of apoptosis in an insect , our results suggest a positive association between viral infection and the expression of apoptosis-related genes in the insect vector . We then manipulated apoptotic response using RNAi to explore the role of apoptosis during viral infection in insect bodies . Viral titers in 30 live viruliferous leafhoppers after dsRNAs microinjection were examined daily . We tested the efficiency of the knockdown of the targeted genes after 2 days post microinjection of dsRNAs in intact insects ( S5B Fig ) . Time-course experiments showed that the profiles of mean number of RGDV P8 gene copies in all dsRNAs treatments were similar ( Fig 8B ) . From 6 to 14 days pdap , the mean number of RGDV P8 gene copies in the dsIAP , dsCASP2L and dsGFP treatment was 5 . 45 × 109 , 1 . 01 × 108 and 8 . 53 × 108 copies/μg insect RNA , respectively ( Fig 8B ) . These results demonstrated that blocking the apoptotic pathway inhibited viral infection , while promoting the apoptotic pathway facilitated viral infection . Thus , we concluded that RGDV induced and utilized apoptosis for viral infection in R . dorsalis vectors . We then calculated the mortality of 100 viruliferous or nonviruliferous insects daily after dsRNA microinjection . Our preliminary test indicated that the microinjection itself had little effect on the mortality rates of nonvirulifeorus R . dorsalis adults ( S6A Fig ) . For both nonviruliferous and viruliferous insects , the mortality of dsIAP-treated R . dorsalis was higher than those treated with dsGFP or dsCASP2L . For example , at 6 days post microinjection , approximately 43 . 0% of the dsIAP-treated viruliferous insects were dead , compared with mortality rates of about 33 . 0% for the dsGFP-treated and about 24 . 0% for dsCASP2L-treated viruliferous insects ( S6B Fig ) . However , at the same number of days post microinjection , the mortality rates for the dsCASP2L- , dsGFP- and dsIAP-treated nonviruliferous insects were approximately 11 . 3% , 16 . 3% and 26 . 0% , respectively ( S6B Fig ) , supporting the fact that IAP is necessary to maintain cell viability in insects [43] . Furthermore , our attempt to silence IAP resulted in a simultaneous increase in apoptotic response and death in viruliferous R . dorsalis , illustrating the possible positive association between virus-induced apoptotic response and insect mortality .
Many important plant viruses are persistently transmitted via insect vectors with limited harm to the insects . Here , we used the plant reovirus RGDV and leafhopper vector R . dorsalis to determine how the apoptotic response was activated during persistent viral transmission by insect vectors . We first demonstrated that RGDV infection induced typical apoptotic characteristics in cultured leafhopper cells , including degeneration of mitochondria , decrease in mitochondrial membrane potential , and appearance of condensed chromatin , chromosomal DNA fragments and virus-containing apoptotic bodies ( Figs 1 and 2 ) , verifying that RGDV triggered the typical apoptotic response in insect vector cells . The expression of caspases ( CASP2L and CASP8L ) and IAP genes were up-regulated during viral infection in cultured leafhopper cells ( Fig 5 ) . The knockdown of caspase gene expression using RNAi blocked apoptotic response and led to the significant inhibition of synthesis of viral mRNAs and genome RNAs , and the accumulation of viral proteins in insect vector cells ( Figs 5 and 6 ) . However , the knockdown of IAP gene expression using RNAi promoted apoptotic response , causing a significant increase of these viral replication processes in insect vector cells ( Figs 5 and 6 ) . Thus , RGDV exploit the apoptotic response in a caspase-dependent pathway to promote viral replication in insect vector cells . In insect bodies , RGDV infection also triggered a similar apoptotic response restricted to a low level , which appeared to benefit viral replication , but may have damaged functionally relevant tissues and organs , decreasing the fitness of the vectors ( Figs 7 and 8 ) . Thus , the typical apoptotic response can be induced and facilitated viral accumulation during viral replication and transmission by the insect vectors . As we observed for the RGDV–R . dorsalis combination , the similar apoptotic response is also restricted to a low level in mosquito vectors [27] . In general , the induction of apoptosis promoted viral infection but also harmed the insects . Thus , viruses have evolved some mechanisms to avoid stimulating extensive apoptotic responses in the bodies of insect vectors . During the latent period for RGDV in insect vectors , before 10 days padp [42] , the expression levels of apoptosis-related genes ( CASP2L , CASP8L and IAP ) increased , then decreased ( Fig 8A ) , indicating that the apoptotic response was activated during replication and then was suppressed . Such synchronous gene expression for CASP2L , CASP8 and IAP suggested that virus-induced apoptotic response was critically modulated during viral infection of insect vectors . For controlling the excessive viral accumulation and avoiding obvious pathology , IAP , the inhibitor of apoptosis was induced by viruses to restrict the apoptotic response to a low level . Other as-yet unknown anti-apoptotic mechanisms might also be activated to block or restrict apoptotic response . Furthermore , a conserved siRNA antiviral response was triggered by RGDV infection to control viral propagation , avoiding excessive viral accumulation past the pathogenic threshold in insect vectors [24] . Previously , we also found that cellular structures and pathways such as microtubules , intermediate filaments or autophagy were induced by RGDV infection , promoting viral infection but also causing some insect cytopathology [44 , 45] . It appeared that all these mechanisms were involved in modulating a metastable balance between viral accumulation and adverse effects , allowing the virus to be persistently transmitted by insect vectors . The multifunctional mitochondrion not only plays an essential role in host immune responses but also serves as an important control point in the regulation of apoptosis [7] . After apoptotic signaling , the mitochondrial membrane potential is lost , and apoptosis-related factors are released [8–10] . Numerous viral proteins , including Vpr protein of human immunodeficiency virus ( HIV ) , X protein of hepatitis B virus ( HBV ) , PB1-F2 of influenza A virus ( IAV ) , and NS4A of hepatitis C virus ( HCV ) , have been reported to directly target mitochondria to activate mitochondrial apoptosis in mammalian host cells [46–50] . VDAC , an outer membrane protein of mitochondria , is often activated during apoptosis and is targeted by viral proteins to initiate apoptosis [46 , 51–52] . In fact , VDAC is a porin channel that serves as a major diffusion pathway for ions and metabolites to control mitochondrial membrane permeabilization [39] . Our data revealed that the fibrillar structures composed of nonstructural protein Pns11 of RGDV could target the mitochondria , induce mitochondrial degeneration and decrease the mitochondrial membrane potential in the absence of viral replication ( Figs 3 and 4 ) , suggesting that RGDV Pns11 was responsible for initiating virus-induced apoptotic response . Such attachment of virus-induced fibrillar structures with mitochondria possibly was mediated by the specific interaction of RGDV Pns11 with the VDAC ( Figs 3 and 4 ) . How the association of VDAC with Pns11 of RGDV is involved in the induction of apoptotic response during viral replication in insect vectors is still unknown . Our study is the first to directly confirm that a nonstructural protein encoded by a persistent-propagative plant virus induce the apoptotic response in an insect vector to promote viral infection and transmission . Based on the results described , we propose that RGDV exploits the apoptotic mechanism for efficient infection in insect vector cells . However , there are still many unknowns in the apoptotic process . For example , how does RGDV trigger and exploit the apoptotic response for efficient infection and how does it evolve such a strategy to enable persistent infection ? In some cases , viruses may directly exploit apoptotic bodies for their dissemination and subsequent infection of a mammalian host or an insect vector [9 , 12 , 53 , 54] . In our study , whether packaging of the RGDV virions within apoptotic bodies protects them from insect immune mechanisms was not determined . In insect vectors , one possible consequence of apoptosis is that physical barriers within the insect are weakened . We thus deduced that the apoptotic response is exploited by RGDV to overcome multiple tissue and membrane barriers to enable efficient infection of its leafhopper vectors . In addition , we still do not know how virus-induced apoptotic response is restricted to a low level in insect vectors . New approaches based on the reverse genetics systems for plant reoviruses and on CRISPR/Cas9 technologies for the leafhopper vector , combined with continuous insect vector cell lines will provide new opportunities to unravel the molecular mechanisms for virus-induced apoptotic response in the RGDV–R . dorsalis system .
Nonviruliferous individuals of the leafhopper R . dorsalis were collected from Guangdong Province in southern China . The continuous cultured cell line derived from R . dorsalis was originally established from embryonic fragments dissected from insect eggs and maintained on growth medium as described previously [17] . The R . dorsalis cell line supported a uniform and synchronous viral infection , enabling the early viral replication process to be traced [42] . RGDV was purified from infected cultured insect vector cells as described previously , and resuspended in His-Mg buffer ( 0 . 1 M histidine , 0 . 01 M MgCl2 , pH 6 . 2 ) [17] . Synchronous infection of continuous cultured cells by RGDV was initiated as described by Wei et al . [55] . When the cultured monolayer of leafhopper cells on a coverslip ( 15 mm diameter ) reached 80% confluency , cells were inoculated with purified RGDV at a MOI of 0 . 1 or 1 for 2 h , washed twice , and covered with growth medium at 25°C . His-Mg buffer-treated cells served as controls . Rice samples infected with RGDV were initially collected from Guangdong Province . Rabbit polyclonal antisera against intact viral particles , major outer capsid protein P8 , and nonstructural proteins Pns11 and Pns12 were prepared as described previously [29 , 32 , 36] . IgGs were purified from specific polyclonal antisera , then conjugated to rhodamine or fluorescein isothiocyanate ( FITC ) , according to the manufacturer’s instructions ( Thermo Fisher ) . The antibody against cytochrome c was obtained from BD Biosciences . Sf9 cells infected with recombinant baculovirus vector containing Pns11 of RGDV have been previously described [17] . In brief , the coding region of the ORF for RGDV Pns11 was amplified by PCR . The purified product was cloned into Gateway vector pDEST8 ( Thermo Fisher ) to construct plasmid pDEST8-Pns11 . Then the recombinant baculovirus vector was introduced into E . coli DH10Bac ( Thermo Fisher ) to generate a recombinant bacmid . The isolated recombinant bacmid was used to transfect Sf9 cells in the presence of Cellfectin II ( Thermo Fisher ) according to the manufacturer’s instructions . After a high-titer baculoviral stock was generated , amplification of the viral stock was scaled up to an appropriate volume for cellular infection on coverslips or in flasks . At 48 hpi , Sf9 cells , growing on coverslips and infected with recombinant bacmids , were treated for immunoelectron microscopy . Sf9 cells inoculated with empty baculovirus vector served as negative controls . They were also fixed , permeabilized , and immunolabeled with Pns11-specific IgGs conjugated to FITC ( Pns11-FITC ) for immunofluorescence microscopy of Pns11 overexpression and mitochondria . Sf9 cells , growing in flasks and infected with recombinant bacmids , were harvested for cell viability tests using 0 . 4% trypan blue solution at 72 hpi . Cell images and counts were made in an automated cell counter ( Counter Star ) . Virus-infected cultured R . dorsalis cells growing in a monolayer on coverslips and intestines dissected from viruliferous R . dorsalis or Sf9 cells were fixed , dehydrated , and embedded , and ultrathin sections were cut as previously described [31] . For immunoelectron microscopy , sections were immunolabeled with the Pns11-specific IgGs as the primary antibody , followed by treatment with goat anti-rabbit IgG conjugated with 15-nm-diameter gold particles as the secondary antibody ( Abcam ) , as previously described [31] . The MitoProbe JC-1 Assay Kit for Flow Cytometry ( Thermo Fisher ) was used to measure the change in mitochondrial membrane potential in cultured cells of R . dorsalis at 48 hpi . Briefly , approximately 1 × 106 cells were collected and suspended in 1 mL PBS . JC-1 was added to cells at a final concentration of 2 μM , and after incubation at 37°C for 30 min , cells were washed once with warm PBS , then resuspended in PBS and immediately examined with a flow cytometer ( BD FACS Calibur ) . Data from three independent biological experiments were analyzed using Cellquest software and displayed as a dot plot of JC-1 green fluorescence ( x-axis ) against red fluorescence ( y-axis ) . Changes in mitochondrial membrane potential in Sf9 cells infected with the recombinant baculovirus expressing Pns11 were measured using rhodamine 123 fluorescence ( Thermo Fisher ) at 72 hpi . Sf9 cells inoculated with empty baculovirus vector served as a negative control . In brief , approximately 1 × 106 Sf9 cells were harvested and suspended in PBS , then incubated with rhodamine 123 at a concentration of 1 μM in PBS at 37°C for 1 h . Cells were washed with PBS , then resuspended in PBS and analyzed immediately using the flow cytometer . Data from three independent biological experiments were analyzed and displayed as a plot of fluorescence intensity of rhodamine ( x-axis ) against cell number ( y-axis ) . At 72 hpi , cultured cells of R . dorsalis were harvested , and DNA was extracted using the Cell Apoptosis DNA Ladder Detection Kit ( KeyGEN BioTECH ) , according to the manufacturer’s instructions . Chromosomal DNA fragments were separated using 2% agarose gel electrophoresis , and DNA ladders were visualized by ethidium bromide staining . R . dorsalis cells cultured in a monolayer on coverslips were fixed in 4% v/v paraformaldehyde and treated with 0 . 2% v/v Triton-X , as previously described [56] . Then the DeadEnd Fluorometric TUNEL System ( Promega ) was used for TUNEL staining . According to the manufacturer’s instructions , samples were treated with the equilibration buffer in the kit at room temperature for 10 min , then incubated with rTdT incubation buffer at 37°C for 60 min . Thereafter , the reaction was terminated by adding 2× sodium citrate in the kit , and then incubated with viral particle-specific IgG conjugated to rhodamine ( virus-rhodamine ) . Nick-end-labeling of nucleosome fragments with fluorescein-dUTP and viral infection were visualized using a confocal microscope . For observing TUNEL signals during viral infection of insect vectors , second instars of R . dorsalis were fed on diseased rice plants for 2 days and then transferred to healthy rice seedlings . At different days padp , the intestines or salivary glands were dissected and processed for TUNEL assay , as described above . Meanwhile , samples were immunolabeled with virus-specific IgGs conjugated to rhodamine ( virus-rhodamine ) and actin dye phalloidin-Alexa Fluor 647 carboxylic acid ( Invitrogen ) , then processed for immunofluorescence microscopy as described previously [17] . Three independent biological replicates were conducted and analyzed . Virus-infected cultured R . dorsalis cells or Pns11-expressing Sf9 cells growing on coverslips were incubated with MitoTracker Red CMXRos ( Thermo Fisher ) for 45 min using standard growth conditions . After the staining solution was carefully removed , cells were fixed in 4% v/v paraformaldehyde and permeabilized in 0 . 2% v/v Triton-X , immunolabeled with Pns11-FITC , then examined with immunofluorescence microscopy . To set up the immunofluorescence microscopy parameters , digital images ( 1024×1024 pixels ) were captured with either 488 nm excitation ( emission filters ) or 543 nm excitation . They were acquired with a 63 oil-immersion objective . Samples in the same group possessed the same parameters of immunofluorescence microscopy to unitize the background . A Matchmaker Gold Yeast-two-hybrid system ( Clontech , USA ) was used for Y2H screening . The cDNA library derived from R . dorsalis or the VDAC gene was constructed in the pGADT7 vector for prey plasmids . Full-length ORF of RGDV Pns11 was cloned in the pGBKT7 vector as a bait plasmid , which was then used to transform yeast strain AH109 to confirm the absence of self-activation and toxicity . Thereafter , the prey and bait were used to cotransform AH109 , and transformants were screened on the SD double-dropout ( DDO ) medium ( SD/-Leu/-Trp ) , SD triple-dropout ( TDO ) medium ( SD/-His/-Leu/-Trp ) and SD QDO medium ( SD/-Ade/-His/-Leu/-Trp ) . Positive clones were selected on QDO/X plates containing X-α-Gal ( 20 μg/mL ) to detect β-galactosidase activity . The interaction of pGBKT7-53 with pGADT7-T served as a positive control and that of pGBKT7-Lam with pGADT7-T served as a negative control . We then used the DUALmembrane starter kit ( Dualsystems Biotech ) to detect the interaction between membrane-associated VDAC and RGDV Pns11 according to the manufacturer’s instructions . Full-length ORF of RGDV Pns11 or VDAC was inserted into bait vector pBT3-STE or prey vector pPR3-N . Thereafter the bait and prey were used to transform the yeast strain NMY51 , and transformants were screened on the TDO and QDO medium . The clones were streaked on QDO/X plates containing X-Gal for color formation in a β-galactosidase assay . The pTSU2-APP/ pNubG-Fe65 interaction served as a positive control , and the pTSU2-APP/ pRR3N served as a negative control . A GST pull-down assay was performed as previously described [57] . The Pns11 gene of RGDV was cloned in the pGEX-3x vector to construct a plasmid expressing the GST fusion protein as a bait ( GST-Pns11 ) . The full-length ORF of the VDAC from R . dorsalis was cloned into the pHM4 vector to construct a plasmid expressing the His fusion protein as a prey ( His-VDAC ) . Recombinant proteins GST-Pns11 and GST were respectively expressed in the Escherichia coli stain BL21 . Lysates were then incubated with glutathione-Sepharose beads ( Amersham ) and subsequently , with the recombinant protein His-VDAC . Finally , eluates were analyzed using GST-tag and His-tag antibodies ( Sigma ) , respectively , in a Western blot assay . Cultured R . dorsalis cells in a monolayer with 80% confluency were inoculated with RGDV at a MOI of 1 . 0 for 2 h . His-Mg buffer-treated cells served as controls . At 48 hpi , the cells were collected at the same time . For viral acquisition by insects , about 500 nonviruliferous second instar nymphs of R . dorsalis were fed on RGDV-infected rice plants for 2 days , then transferred to healthy rice seedling . From 6 to 14 days padp , 30 leafhoppers were daily collected at the same time . Total RNA was extracted from cells or insects using TRIzol Reagent ( Thermo ) according to the manufacturer’s instructions . For synthesizing first-strand cDNA , total RNA was primed with oligo-dT primer and reverse transcribed with M-MLV Reverse Transcriptase ( Promega ) . The qPCR assays were performed in a Mastercycler Realplex4 real-time PCR system ( Eppendorf ) using GoTaq qPCR Master Mix kit ( Promega ) with efficient and specific primers ( S1 Table ) . For relative quantitation , the transcriptional level of the actin gene from the leafhopper was used as the control for each qPCR assay . Relative levels of genes were qualitatively analyzed using the 2−ΔΔCt method . For absolute quantification , the number of RGDV P8 gene copies and CASP2L and IAP gene copies were calculated as the log of the copy number/μg insect RNA based on a standard curve of the RGDV P8 gene , CASP2L gene and IAP gene , respectively . The T7 RNA polymerase promoter was added to the forward primer and reverse primer at the 5′ and 3′ terminal to amplify a region of about 500–900 bp in each gene ( S1 Table ) . PCR products were transcribed into dsRNAs in vitro using the T7 RiboMAX ( TM ) Express RNAi System , according to the manufacturer’s protocol ( Promega ) . Purified dsRNAs were examined using agarose gel electrophoresis to determine their integrity and quantified by spectroscopy . Three microliters of Cellfectin II Reagent ( Thermo ) and 4 μg dsGFP , dsVDAC , dsCASP2L , dsCASP8L or dsIAP were diluted individually in 25 μL LBM without antibiotics and fetal bovine serum , and mixed gently together at room temperature for 20 min . Then the dsRNA–lipid complex was incubated with the cultured cells of R . dorsalis in a monolayer ( at 80% confluency ) for 8 h . Thereafter , cells were inoculated with purified RGDV ( MOI of 0 . 1 ) for 2 h . Infected cells were processed for immunofluorescence or were harvested for RT-qPCR detection at 48 or 72 hpi . Alternatively , at 72 hpi , total proteins were extracted from infected cells and further analyzed by immunoblotting with antibodies against P8 and Pns12 , respectively . Insect actin was detected with actin-specific antibodies ( Sigma ) as a control . Furthermore , viral genome dsRNAs were isolated from cultured cells , as described previously [58] . In brief , viral genome dsRNAs from cell lysates were isolated at 72 hpi using the modified CF11 cellulose chromatography procedure . Cell lysates of each dsRNA treatment were mixed with 1× STE ( 0 . 1 M NaCl , 0 . 05 M Tris , 0 . 001 M EDTA , pH 6 . 8 ) , 10% SDS , and 2× STE-saturated phenol . Following centrifugation , the aqueous phase was recovered and added the ethanol to a final volume of 16 . 5% . Then CF11 cellulose ( Whatman ) was loaded and vortexed to mix thoroughly . The pellets after the centrifugation were washed with 1×STE in 16 . 5% ethanol . Finally , the dsRNAs were eluted from CF11 with 1×STE . Separation of genomic dsRNA segments were loaded on 1 . 0% agarose gel . The pattern of genomic dsRNA segments separated from purified RGDV virons was considered as the reference [41] . For the northern blots , at 72 hpi , the total RNAs from cultured cells were extracted with TRIzol Reagent ( Thermo Fisher ) . The DIG High Prime DNA labelling and Detection Starter KitⅠ ( Roche ) was used to examine the transcript level of RGDV P8 and Pns12 . In brief , about 500 bp DIG-labeled DNA probe of P8 or Pns12 about were generated after the incubation of denatured PCR products ( S1 Table ) and DIG-High Prime for 20 h at 37°C . About 5 μg total RNA of each dsRNA treatment were loaded and detected for the transcript level of P8 or Pns12 . The 5 . 8S rRNA stained with Methylene blue were served as a control to confirm loading of equal amounts of RNA in each lane . In addition , the relative abundance of CASP2L , CASP8L or IAP genes in virus-free cultured cells of R . dorsalis after different times of transfection with dsRNAs was quantified by RT-qPCR as described already . Three independent biological replicates were conducted and analyzed . Cultured R . dorsalis cells were treated for 4 h with 25 μM pancaspase inhibitor Z-VAD-FMK ( Promega ) dissolved in DMSO . Cells were treated with DMSO as the control . Cells were then inoculated with purified virus ( MOI of 0 . 1 ) , then assayed by RT-qPCR at 48 hpi , as described above . Three independent biological replicates were conducted and analyzed . Generally , first or second instar nymphs of R . dorsalis are the most efficient stage for acquiring RGDV from infected rice plants [19] . Furthermore , RGDV takes at least 2 days to infect the intestinal epithelium of R . dorsalis , so the synthesized dsRNAs are microinjected directly into the insect abdomen for efficient dsRNA delivery instead of oral feeding [59] . We first allowed 700 nonviruliferous second instar nymphs to feed on RGDV-infected rice plants for 2 days to acquire viruses , then microinjected them with synthesized dsGFP , dsCASP2L or dsIAP ( about 0 . 05 μg/insect ) using a Nanoject II Auto-Nanoliter Injector ( Spring ) . The treated insects were transferred to healthy rice seedlings until they were assayed . Viral titers in 30 live leafhoppers treated with dsRNAs were assayed daily by RT-qPCR for viral gene copies of the major outer capsid protein P8 . The equation of y = -3 . 349x +49 . 258 ( y = the logarithm of plasmid copy number to base 10 , x = Ct value , R2 = 0 . 9995 ) was used to calculate the viral genome copy as the log of the copy number per microgram of insect RNA [60] . For calculating mortality of insects , 100 insects of each dsRNA treatment were individually fed on a healthy rice seedling in one glass tube after microinjection . The dead insects were counted at the same time each day . The mortality rate was calculated as the number of dead dsRNAs-treated insects in the total number of dsRNAs-treated insects . In addition , the relative abundance of CASP2L and IAP genes in 10 nonviruliferous leafhoppers was also estimated by RT-qPCR assay . Three independent biological replicates were conducted and analyzed . All data for cultured cells and some data from insects , including percentage of TUNEL-positive intestines and relative expression of genes , were analyzed with a two-tailed t-test in GraphPad Prism 7 . The data for insect mortality were analyzed using SPSS , version 17 . 0 . Percentage data were arcsine square-root-transformed before analysis . Multiple comparisons of the means were performed using Tukey’s honestly significant difference ( HSD ) test and a one-way analysis of variance ( ANOVA ) . Data were back-transformed after analysis for presentation in the text and figures . | Of the approximately 700 known plant viruses , more than 75% are transmitted by insects . Numerous plant viruses can replicate inside the cells of the insects . Unlike in the plant hosts , the viruses do not seem to cause disease in the insect vectors that carry them . Here , we report that the replication of a plant reovirus , rice gall dwarf virus ( RGDV ) , activated the apoptotic response in limited areas of leafhopper vectors during viral replication . Interestingly , fibrillar structures constituted by nonstructural protein Pns11 , which is encoded by RGDV , targeted the mitochondria and induced apoptotic response in the absence of viral replication , possibly via the specific interaction of RGDV Pns11 with an apoptosis-related mitochondrial outer membrane-associated protein . Our findings further suggest that the activation of apoptotic response facilitates efficient viral infection , whereas inhibition of apoptotic response blocks viral infection in insect vectors . This work presents a novel discovery that a plant reovirus induces typical apoptotic response and thus promotes its transmission by insect vectors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"viral",
"transmission",
"and",
"infection",
"cell",
"processes",
"biological",
"cultures",
"microbiology",
"vector-borne",
"diseases",
"viral",
"structure",
"cell",
"cultures",
"mitochondria",
"bioenergetics",
"... | 2019 | Fibrillar structures induced by a plant reovirus target mitochondria to activate typical apoptotic response and promote viral infection in insect vectors |
CD8 T cell responses have three phases: expansion , contraction , and memory . Dynamic alterations in proliferation and apoptotic rates control CD8 T cell numbers at each phase , which in turn dictate the magnitude of CD8 T cell memory . Identification of signaling pathways that control CD8 T cell memory is incomplete . The PI3K/Akt signaling pathway controls cell growth in many cell types by modulating the activity of FOXO transcription factors . But the role of FOXOs in regulating CD8 T cell memory remains unknown . We show that phosphorylation of Akt , FOXO and mTOR in CD8 T cells occurs in a dynamic fashion in vivo during an acute viral infection . To elucidate the potentially dynamic role for FOXO3 in regulating homeostasis of activated CD8 T cells in lymphoid and non-lymphoid organs , we infected global and T cell-specific FOXO3-deficient mice with Lymphocytic Choriomeningitis Virus ( LCMV ) . We found that FOXO3 deficiency induced a marked increase in the expansion of effector CD8 T cells , preferentially in the spleen , by T cell-intrinsic mechanisms . Mechanistically , the enhanced accumulation of proliferating CD8 T cells in FOXO3-deficient mice was not attributed to an augmented rate of cell division , but instead was linked to a reduction in cellular apoptosis . These data suggested that FOXO3 might inhibit accumulation of growth factor-deprived proliferating CD8 T cells by reducing their viability . By virtue of greater accumulation of memory precursor effector cells during expansion , the numbers of memory CD8 T cells were strikingly increased in the spleens of both global and T cell-specific FOXO3-deficient mice . The augmented CD8 T cell memory was durable , and FOXO3 deficiency did not perturb any of the qualitative attributes of memory T cells . In summary , we have identified FOXO3 as a critical regulator of CD8 T cell memory , and therapeutic modulation of FOXO3 might enhance vaccine-induced protective immunity against intracellular pathogens .
The ability of the immune system to respond rapidly and vigorously to antigen re-exposure is termed immunological memory , which is one of the tenets of adaptive immunity [1] , [2] . Induction of memory B and T cells is the basis of immunological memory induced by infections or vaccinations [2]–[4] . As compared to naïve T cells , memory T cells are hyper-reactive to antigenic stimulation and swiftly proliferate and/or differentiate into effector cells to confer protective immunity expeditiously [5]–[8] . The ability of memory T cells to confer protective immunity depends upon the number and quality of memory T cells [5] , [9]–[13] . Understanding the mechanisms that regulate the quantity and quality of T cell memory is fundamentally important for the development of effective vaccines . During a CD8 T cell response , engagement of the TCR , along with appropriate co-stimulatory and inflammatory signals , activate naïve T cells to proliferate and differentiate into effector cells [1] , [4] , [8] , [13] , [14] . In the case of LCMV infection , the peak of T cell expansion is reached at 8–10 days after infection , and the majority of the newly generated effector cells present at the peak of the response are short-lived and fated for deletion [15]–[17] . But , a small subset of the effectors , termed memory precursor effector cells ( MPECs ) , possesses the potential to survive and differentiate into long-lived memory cells [16] , [17] . The number of memory CD8 T cells generated depends largely upon the magnitude of the expansion of MPECs during the T cell response . Substantial progress has been made in deciphering the extracellular signals and transcription factors that regulate the differentiation of MPECs [1] , but the signaling pathways that govern the number of MPECs , their differentiation into memory CD8 T cells , and the maintenance of CD8 T cell memory are not fully understood . The FOXO family of transcription factors plays a crucial role in governing cellular proliferation , apoptosis , energy metabolism , and stress resistance in response to dynamic alterations in stress and abundance of nutrients and growth factors in many cell types [18]–[26] . In mammals , the FOXO family consists of at least four members: FOXO1 , FOXO3 , FOXO4 , and FOXO6 [22] , [27] , [28] . The activity of FOXOs is regulated by post-translational modifications , most notably , phosphorylation [25] , [26] , [29]–[31] . In resting , quiescent cells , FOXOs exist in a hypo-phosphorylated state and localize to the nucleus to control the transcription of their target genes such as p27Kip1 , p21Cip1 , p300 , BIM , Fas ligand that are involved in regulating cellular proliferation or apoptosis [25] , [32]–[43] . However , in response to stimulation with growth factors or cytokines that stimulate the PI3K/AKT signaling pathway , activated Akt phosphorylates FOXO leading to its exclusion from the nucleus and degradation by proteolysis in the cytoplasm [18] , [25] , [29] , [44] , [45] . As a result , the transcription of FOXO target genes is diminished , which in turn facilitates cell cycle entry and/or survival . T cells express FOXO1 and FOXO3 , and there has been a recent surge in interest to elucidate the importance of FOXOs in regulating T cell biology [46]–[53] . Phosphorylation-mediated inactivation of FOXOs and downregulation of p27Kip1 appear to be obligatory steps for T cells to enter the cell cycle in response to TCR engagement [24] , [41] , [53] . FOXO1 has been shown to control several aspects of T cells including the expression of adhesion molecules like L-selectin and CCR7 , cytokine receptors like the IL-7 receptor , development of regulatory T cells , and protection against autoimmunity [51] , [54] . Unlike FOXO1 , the role of FOXO3 in T cell homeostasis is less well understood . It was first reported that global deletion of FOXO3 results in lymphadenopathy and spontaneous activation of T cells , but other independent studies have failed to confirm these results , possibly due to differences in the genetic background of mutant mice [19] , [40] , [49] . Nonetheless , elegant studies from the Hedrick group have showed that FOXO3 inhibits the primary expansion of CD8 T cells in the spleen by regulating IL-6 production by dendritic cells [49] . However , the T cell intrinsic role of FOXO3 in regulating various phases of the polyclonal multi-epitope-specific CD8 T cell response to an acute viral infection in lymphoid and non-lymphoid organs remains to be determined . In this study , using global and conditional T cell-specific FOXO3 knockout mice , we have systematically examined the role of FOXO3 in regulating the: ( 1 ) expansion and function of polyclonal antigen-specific CD8 T cells in lymphoid and non-lymphoid organs; ( 2 ) antigen-driven in vivo proliferation of virus-specific CD8 T cells; ( 3 ) differentiation of CD8 T cells into short-lived effector cells ( SLECs ) and MPECs; ( 4 ) contraction of antigen-specific CD8 T cells in lymphoid and non-lymphoid organs; ( 5 ) the numbers and function of memory CD8 T cells; ( 6 ) proliferative renewal of memory CD8 T cells; ( 7 ) secondary CD8 T cell responses and protective immunity . These studies show that FOXO3 regulates the clonal expansion of polyclonal CD8 T cells in a tissue-specific fashion by T cell intrinsic mechanisms . The enhanced expansion of CD8 T cells was clearly not due to an increased proliferation rate , but was instead associated with reduced cellular apoptosis . Furthermore , we show that FOXO3 deficiency markedly enhances the size of the memory CD8 T cell compartment without affecting the phenotype or quality of memory CD8 T cells . These findings have implications in the design of effective vaccines that engender potent and effective protective cellular immunity against intracellular pathogens and tumors .
FOXO3 has emerged as a key regulator in a number of physiological outcomes , including metabolism , ageing , and vascular reactivity [21]–[23] , [27] , [46] , [55] . More recently , in the immune system , there is increasing evidence that FOXO3 is a critical regulator of T cell homeostasis [46] , [47] , [52] , [53] , [55] . While a number of signaling pathways and post translational modifications may be involved in controlling the activity of FOXO3 , phosphorylation mediated through a PI3K-Akt centric signaling module primarily regulates FOXO3 function in T cells [29] , [45] , [56] , [57] . In order to fully understand the role of the Akt/FOXO3 axis in the control of CD8 T cell homeostasis , we developed a phospho-specific , flow cytometric method to quantify in vivo phosphorylation levels of key proteins implicated in FOXO3 activity: the upstream kinase Akt , FOXO3 itself , as well as a potential downstream substrate of Akt , the kinase , mammalian target of Rapamycin ( mTOR ) ( Figure 1A ) . Figure 1A illustrates the kinetics of phosphorylation of Akt ( Thr308 ) , FOXO1/3 ( T24/T32 ) , and mTOR ( S2448 ) in antigen-specific CD8 T cells during an acute LCMV infection . The phosphorylation of Akt ( Thr308 ) and mTOR ( S2448 ) was highest at day 5 post-infection ( PI ) , and subsided to steady-state levels by days 10 and 8 PI respectively . Interestingly , the phosphorylation dynamics for FOXO1/O3 were different from that of Akt and mTOR; the phosphorylation levels for FOXO1/O3 dropped between days 5 and 8 PI , but gradually increased back to steady-state levels by days 10–15 PI . Note that the phosphorylation kinetics of NP396-specific CD8 T cells after day 8 PI was slightly delayed , as compared to those in GP33-specific CD8 T cells . In summary , these findings demonstrate that the in vivo phosphorylation levels of Akt , FOXO1/O3 , and mTOR are highly dynamic , and of note , the phosphorylation kinetics of Akt correlated with mTOR phosphorylation but not with phosphorylation of FOXO1/O3 during a T cell response to LCMV . However , the specific role of FOXO3 in regulating different phases of the CD8 T cell response has not been carefully examined . Infection of mice with LCMV elicits a potent , multiple epitope-specific , CD8 T cell response wherein virus-specific CD8 T cells are distributed to both lymphoid and non-lymphoid organs [58] , [59] . To examine the role of FOXO3 in regulating CD8 T cell responses to an acute viral infection , groups of wild type ( +/+ ) and FOXO3-deficient ( FOXO3−/− ) mice were infected with LCMV . At day 8 PI , we quantified CD8 T cells that are specific to three immuno-dominant LCMV epitopes in lymphoid ( spleen and lymph nodes ) and non-lymphoid ( liver ) tissues . Figure 1B shows that the numbers of LCMV-specific CD8 T cells in spleens of FOXO3−/− mice were significantly ( P<0 . 05 ) higher than in +/+ mice . Surprisingly , the numbers of virus-specific CD8 T cells in the lymph nodes and liver of FOXO3−/− mice were comparable to those in +/+ mice ( Figure 1B ) . These data suggested that FOXO3 downregulates the accumulation of CD8 T cells in a tissue-specific fashion during an acute LCMV infection . Next , we assessed whether FOXO3 deficiency affected the cell surface phenotype of LCMV-specific effector CD8 T cells at day 8 PI . The expression levels of CD44 , CD62L , CD27 , and CD122 on FOXO3−/− LCMV-specific CD8 T cells were comparable to those on +/+ CD8 T cells ( Figure 2A ) . The population of LCMV-specific CD8 T cells in the spleen is comprised of at least two subsets of effector cells based on the cell surface expression of IL-7 receptor α ( CD127 ) and KLRG-1: the SLECs ( KLRG-1HI/ CD127LO ) a majority of which are destined for apoptosis , and the MPECs ( CD127HI/KLRG-1LO ) , which are poised to differentiate into memory CD8 T cells [16] , [17] , [60] . Figure 2B shows that FOXO3 deficiency significantly enhanced the absolute numbers of both SLECs and MPECs in the spleen at day 8 PI . Of note is the marked increase in the total number of MPECs in FOXO3−/− mice . In summary , these data suggested that FOXO3 deficiency increased the clonal expansion of CD8 T cells without disrupting the differentiation of SLECs and MPECs . To detect possible differences in the functionality of antigen-specific CD8 T cells from +/+ and FOXO3−/− mice , we assessed their ability to produce the cytokines IFNγ , TNFα , and IL-2 in response to antigenic stimulation directly ex vivo ( Figure 2C ) . The MFIs of IFNγ staining for FOXO3−/− and +/+ effector CD8 T cells were similar ( Figure 2C ) . Additionally , the percentages of epitope-specific , CD8 T cells that produced two cytokines ( IFNγ and TNFα ) or three cytokines ( IFNγ , TNFα , and IL-2 ) in FOXO3−/− mice were similar to those in +/+ mice ( Figure 2C ) . The MFIs for TNFα and IL-2 were also comparable between +/+ and FOXO3−/− CD8 T cells ( data not shown ) . As a surrogate marker of the lytic function of effector CD8 T cells , we compared granzyme B expression between +/+ and FOXO3−/− LCMV-specific effector CD8 T cells directly ex vivo . The levels of granzyme B in FOXO3−/− CD8 T cells were similar to those in +/+ CD8 T cells ( Figure 2D ) . Taken together , data in Figures 1 and 2 suggested that FOXO3 deficiency increased the expansion of virus-specific CD8 T cells without affecting their phenotype or function . Consistent with normal CD8 T cell effector function , LCMV control in FOXO3−/− mice was similar to that in +/+ mice ( not shown ) . To rigorously address whether the greater number of LCMV-specific CD8 T cells in the spleen of FOXO3−/− mice was due to increased proliferation , we utilized two different approaches , Ki67 staining and BrdU incorporation . First , we quantified proliferation of LCMV-specific CD8 T cells by staining for the nuclear antigen Ki67 directly ex vivo ( Figure 3A ) during the peak clonal expansion phase ( days 6 and 8 PI ) of the CD8 T cell response . For the second approach , we measured BrdU incorporation by LCMV-specific CD8 T cells in vivo from days 6–8 PI ( Figure 3B ) . As shown in Figures 3A and 3B , the percentages of of Ki67+ve LCMV-specific CD8 T cells and BrdU+ve LCMV epitope-specific CD8 T cells in spleens , liver , and lymph nodes of FOXO3−/− mice were similar to those in +/+ mice . Measurement of Ki67 expression at day 5 PI also showed that FOXO3 deficiency did not alter the proliferation of LCMV-specific CD8 T cells ( Figure S1 ) Taken together , these data indicated that enhanced proliferation was not sufficient to explain the increase in antigen-specific CD8 T cells in FOXO3−/− mice during LCMV infection . CD8 T cell homeostasis is not simply determined by alterations in proliferation , it is also regulated by alterations in cell death . During the clonal expansion phase of the CD8 T cell response , there is concomitant proliferation and apoptosis [61] , and therefore the magnitude of clonal expansion is dependent upon the relative rates of proliferation and apoptosis . To assess cellular apoptosis , we determined the percentages of Annexin VHI LCMV-specific CD8 T cells in spleens of +/+ and FOXO3−/− mice at day 6 PI , directly ex vivo . Data in Figure 4A shows that the percentages of Annexin VHI LCMV-specific CD8 T cells in spleens of FOXO3−/− mice were significantly lower than in spleens of +/+ mice . These data suggested that FOXO3 controls the accumulation of effector CD8 T cells by promoting cellular apoptosis during an acute LCMV infection . During antigen-driven proliferation , the competing pro-apoptotic effects of TGF-β and the anti-apoptotic effects of IL-15 regulate the apoptotic rate of CD8 T cells [61] . Because apoptosis of LCMV-specific CD8 T cells was dampened by FOXO3 deficiency ( Figure 4A ) , and IL-15 has been reported to induce phosphorylation of FOXO3 [62] , we hypothesized that FOXO3 might downregulate clonal expansion by reducing the viability of CD8 T cells that are deprived of IL-15 . The balance in the levels of the anti-apoptotic molecule Bcl-2 and the pro-apoptotic molecule BIM controls the susceptibility of a T cell to apoptotic stimuli [36] , [38] , [39] , [63]–[65] . Since BIM is a target gene for FOXO3 , we tested whether deficiency of FOXO3 might lead to lower expression of BIM in proliferating CD8 T cells ( day 6 PI ) , by comparing the levels of BIM in +/+ and FOXO3−/− LCMV-specific CD8 T cells after culture with or without IL-15 . As shown in Figure 4B and Figure S2 , BIM expression in +/+ CD8 T cells was higher than in FOXO3−/− CD8 T cells , when cultured in media without IL-15 . However , IL-15 reduced BIM expression in +/+ CD8 T cells to levels seen in FOXO3−/− CD8 T cells; IL-15 did not affect BIM expression in FOXO3−/− CD8 T cells . These data suggested that FOXO3 might control BIM expression in IL-15-deprived CD8 T cells . When analyzing for BIM in direct relation to Bcl-2 , we observed that after 6 days of infection , LCMV-specific CD8 T cells from +/+ mice exhibit an increased BIM to Bcl-2 ratio , as compared to FOXO3−/− CD8 T cells ( Figure 4B ) . Thus , we propose that FOXO3 might downregulate the accumulation of proliferating CD8 T cells by inducing BIM expression . Next , we examined whether FOXO3 deficiency regulated the contraction of CD8 T cells in lymphoid and non-lymphoid tissues during an acute LCMV infection . Virus-specific CD8 T cells were quantified in spleen , liver , and lymph nodes at days 8 , 11 , 15 , and 30 PI ( Figure 5A ) . Overall , the slopes of the contraction curves for LCMV-specific CD8 T cells in FOXO3−/− mice were comparable to those of +/+ mice . Thus , the contraction of LCMV-specific CD8 T cells was minimally affected by FOXO3 deficiency . Next , we assessed whether FOXO3 regulated proliferation of LCMV-specific CD8 T cells during the contraction phase . In vivo BrdU incorporation studies showed that the percentages of LCMV-specific CD8 T cells that incorporated BrdU between days 8–11 or 12–15 PI in FOXO3−/− mice were similar to those in +/+ mice ( Figure 5B ) . The number of memory CD8 T cells is a function of the magnitude of expansion and contraction during the T cell response [5] . Here , we determined whether increased clonal expansion of MPECs in FOXO3−/− mice ( Figure 2B ) translated to inflation of LCMV-specific memory CD8 T cells in lymphoid and non-lymphoid organs . We observed that FOXO3−/− mice exhibit a substantial increase in the numbers of NP396- ( P<0 . 001 ) , GP33- ( P<0 . 01 ) , and GP276- ( P<0 . 04 ) specific CD8 T cells in spleen at 180 days PI ( Figure 6A ) . Interestingly , there was no detectable increase in the numbers of LCMV-specific memory CD8 T cells in either the liver or lymph nodes at day 180 PI ( Figure 6A ) . It should be noted that the magnitude of increase in the number of memory CD8 T cells in FOXO3−/− mice reflected the increased accumulation of MPECs during the primary response ( Figure 2B ) . High numbers of memory CD8 T cells were maintained in FOXO3−/− mice stably until at least day 300 PI ( data not shown ) . These data strongly imply that FOXO3 plays an important role in downregulating the magnitude of CD8 T cell memory in the spleen following an acute viral infection . Phenotypic analysis of LCMV-specific memory CD8 T cells in FOXO3−/− and +/+ mice suggested that FOXO3 deficiency did not affect the expression of molecules that control T cell trafficking ( CD44 ) or cytokine receptors ( CD122 and CD127 ) ( Figure 6B ) . In addition , assessment of CD62L levels on antigen specific CD8 T cells indicated that both central ( CD62Lhigh ) and effector ( CD62Llow ) memory frequencies were unaffected by FOXO3 deficiency ( Figure 6C ) . Furthermore , functional analysis of antigen-triggered cytokine production did not reveal alterations in cytokine producing ability of LCMV-specific memory CD8 T cells from FOXO3−/− mice when compared to their +/+ counterparts ( Figure 6D ) . In summary , data presented in Figure 6 showed that FOXO3 deficiency increased the quantity of CD8 T cell memory without affecting the quality . It is well established that memory CD8 T cells are maintained for extended periods of time by proliferative renewal , driven by homeostatic cytokines IL-7 and IL-15 [11] , [56] , [66]–[68] . Although it is known that IL-7 and IL-15 signaling triggers phosphorylation of FOXO3 [56] , [62] , the effect of FOXO3 deficiency on proliferative renewal of memory CD8 T cells has not been examined . Using three approaches , we compared the cytokine-driven proliferative renewal of memory CD8 T cells in +/+ and FOXO3−/− mice . First , in vivo BrdU incorporation studies showed that the percentages of BrdU+ve LCMV-specific memory CD8 T cells in FOXO3−/− mice were comparable to those in +/+ mice ( Figure 7A ) . Likewise , the percentages of Ki67+ve LCMV-specific CD8 T cells were unaffected by FOXO3 deficiency ( Figure 7A ) . To further examine the effect of FOXO3 deficiency on the proliferative renewal of memory CD8 T cells , CD8 T cells from the spleens of LCMV-immune +/+ and FOXO3−/− mice were labeled with CFSE and adoptively transferred into congenic uninfected mice . Thirty days after cell transfer , flow cytometric analysis of CFSE staining revealed that donor LCMV-specific CD8 T cells from +/+ and FOXO3−/− mice proliferated equally in the recipient mice ( Figure 7B ) . Taken together , data in Figure 7 provided strong evidence that FOXO3 deficiency did not alter the homeostatic turnover of LCMV-specific memory CD8 T cells . To observe the effect of FOXO3 deletion on recall responses of memory CD8 T cells and protective immunity , LCMV-immune +/+ and FOXO3−/− mice were challenged with LCMV clone 13 , a strain of LCMV that establishes a chronic infection in naïve immunocompetent mice . At day 5 after challenge , the numbers of NP396- ( P<0 . 01 ) , GP33- ( P<0 . 01 ) and GP276- ( P<0 . 03 ) specific CD8 T cells in FOXO3−/− mice ( Figure 8A ) were significantly higher than in +/+ mice . The increased number of LCMV-specific CD8 T cells in spleens of LCMV clone 13-Challenged FOXO3−/− mice correlated with the increased numbers of memory CD8 T cells ( Figure 6A ) . As in a primary infection ( Figure 1B ) , we did not see a statistically significant increase in the numbers of LCMV-specific CD8 T cells in either the liver or lymph nodes ( Figure 8A ) . Staining for Ki67 illustrated that FOXO3 deficiency did not affect the proliferation of LCMV-specific CD8 T cells during a secondary response ( Figure 8B ) . Secondary effector CD8 T cells in spleens of FOXO3−/− mice produced comparable levels of IFNγ , TNFα , and IL-2 ( Figure 8C ) . LCMV titers in tissues were comparable in +/+ and FOXO3−/− mice , which indicated that protective immunity was not compromised in the absence of FOXO3 ( Figure 8D ) . To address whether FOXO3 has a T cell intrinsic role in regulating polyclonal CD8 T cell responses to LCMV , we used a cre-loxP knockout strategy to generate the FOXO3L mice that lacked FOXO3 only in T cells [69] , [70] . T cell-specific loss of FOXO3 in FOXO3L mice was confirmed by western blot and flow cytometry ( Figure S3 ) . FOXO3L and littermate +/+ mice were infected with LCMV and virus-specific CD8 T cell responses were analyzed at day 8 PI . In the spleen , FOXO3L mice exhibited a statistically significant increase in the numbers of LCMV-specific CD8 T cells ( NP396 P<0 . 02; GP33 P<0 . 01; GP276 P<0 . 04 ) over their +/+ littermate controls ( Figure 9A ) . It should be noted that the observed increase in the expansion of CD8 T cells in global FOXO3−/− mice ( Figure 1B ) was fully recapitulated in FOXO3L mice ( Figure 9A ) . To assess whether greater accumulation of effector CD8 T cells in spleens of FOXO3L mice was driven by increased proliferation , we measured Ki67 expression and BrdU incorporation during the clonal expansion phase of the CD8 T cell response to LCMV . At day 6 PI , the percentages of both Ki67+ve and BrdU+ve LCMV-specific CD8 T cells in FOXO3L mice were comparable to those in +/+ mice , which suggested that enhanced accumulation of effector CD8 T cells in FOXO3L mice are not linked to an altered proliferation rate ( Figure 9B ) . Likewise , percentages of Ki67+ve CD8 T cells at day 5 PI were comparable in +/+ and FOXO3L mice ( Figure S4 ) . However , the percentages of Annexin VHI LCMV-specific CD8 T cells in spleens of FOXO3L mice were significantly lower than in +/+ mice ( Figure 9B ) . These data suggested that FOXO3 controls the accumulation of CD8 T cells during a primary response by regulating apoptosis of proliferating cells . Conditional deficiency of FOXO3 in T cells did not affect the cell surface phenotype ( data not shown ) or antigen-triggered cytokine production by LCMV-specific effector CD8 T cells at day 8 PI ( Figure 9C ) . Since both CD8 and CD4 T cells lack FOXO3 activity in FOXO3L mice , it could be argued that increased clonal expansion of CD8 T cells might result from enhanced CD4 T cell help in FOXO3L−/− mice . To address this issue , we depleted CD4 T cells in +/+ and FOXO3L mice and quantified CD8 T cell responses to LCMV in the absence of CD4 T cells . At day 8 PI , >95% of the CD4 T cells were depleted in spleen of both +/+ and FOXO3L mice ( data not shown ) . The number of LCMV-specific CD8 T cells in spleen of CD4-depleted FOXO3L mice was substantially higher than in CD4 T cell-depleted +/+ mice ( Figure 9D ) . Thus , in the apparent absence of CD4 T cells , FOXO3 deficiency in CD8 T cells was sufficient to increase the accumulation of virus-specific CD8 T cells during an acute LCMV infection . To determine whether deletion of FOXO3 , exclusively from the T cell compartment , would affect CD8 T cell memory generation , FOXO3L and +/+ mice were infected with LCMV and virus-specific memory CD8 T cells were quantified in lymphoid and non-lymphoid tissues at 180 days PI . We observed a significant increase in the number of memory CD8 T cells that are specific to the three immuno-dominant LCMV epitopes ( NP396 P<0 . 02; GP33 P<0 . 01; GP276 P<0 . 05 ) in spleens of FOXO3L mice over the +/+ mice ( Figure 10A ) . These data indicated that FOXO3 regulates the magnitude of CD8 T cell memory by T cell intrinsic mechanisms . Next , we examined whether conditional deficiency of FOXO3 in T cells affected the phenotypic and functional attributes of memory CD8 T cells . The expressions of CD44 , CD62L , CCR7 and LFA-1 on +/+ and FOXO3L memory CD8 T cells from spleen , liver and lymph nodes were similar ( Figure 10B ) . Additionally , the relative proportions of effector ( CD62LLo ) and central ( CD62LHI ) memory CD8 T cells were unaffected by FOXO3 deficiency ( Figure 10B ) . In response to antigenic stimulation , virus-specific memory CD8 T cells in LCMV-immune FOXO3L−/− mice produced IFNγ , TNFα , and IL-2 at levels comparable to those in +/+ mice ( Figure 10C ) . As was the case in our studies of global FOXO3-deficient mice ( Figure 7 ) , memory CD8 T cells in FOXO3L mice exhibited no difference in homeostatic turnover , as evidenced by BrdU incorporation from day 120 PI mice , pulsed for 8 days and by parallel Ki67 staining ( Figure 10D ) . Taken together , data presented in Figure 10 strongly suggested that FOXO3 regulates the quantity , with no apparent loss in quality , of CD8 T cell memory by T cell intrinsic mechanisms .
The FOXO transcription factors are important regulators of cell cycle progression , apoptosis , and energy metabolism [21] , [23] , [26] , [28] , [56] . In the T cell compartment , FOXOs have been implicated in regulating homing of T cells , cytokine receptor expression , and development of regulatory T cells [48] , [51] , [53] , [54] . While it has been reported that FOXO3 may control CD8 T cell expansion , albeit through non T cell intrinsic mechanisms , a role for FOXO3 in memory T cell survival has also been posited [49] . What has not been thoroughly addressed , however , is the T cell intrinsic role of FOXO3 in governing different facets of the physiological polyclonal T cell response to foreign antigens , including the in vivo generation and maintenance of CD8 T cell memory . In the present study , we have systematically examined the T cell-intrinsic role of FOXO3 in controlling the expansion , contraction , and memory phases of the polyclonal CD8 T cell response to an acute viral infection . These studies have provided strong evidence supporting a T cell intrinsic role for FOXO3 in limiting the magnitude of expansion and the number of memory CD8 T cells in a tissue-specific fashion during a physiological response to an acute LCMV infection . These findings have advanced our mechanistic understanding of CD8 T cell homeostasis , and are expected to have implications in the development of effective vaccines . FOXOs are known to maintain cellular quiescence by mechanisms including the induction of cell cycle inhibitors like p27KIP1 and p21Cip1 [30] , [34] , [37] , [38] . Downregulation of FOXO activity is believed to be an obligatory step for cell cycle entry in response to mitogenic stimuli [30] . The observed phosphorylation of FOXO1/O3 in LCMV-specific CD8 T cells was readily detectable at day 5 PI but exhibited a sharp decline by day 8 PI . The drop in the phosphorylation in FOXO1/O3 between days 5 and 8 PI coincides with declining viral load and decreased antigenic stimulation . However , we observed a rebound in the phosphorylation of FOXO1/O3 between days 8 and 11 PI . What controls the dynamics of FOXO1/O3 phosphorylation during a CD8 T cell response ? In addition to TCR signaling , FOXO1/O3 phosphorylation is regulated by signaling via cytokine receptors such as IL-2 , IL-7 , and IL-15 [38] , [56] , [66] . It is possible that FOXO1/O3 is phosphorylated by different extracellular signals at different phases of the T cell response . For example , during the phase of antigen-driven proliferation , IL-7R expression is known to be very low [68] , and TCR signaling along with IL-2/IL-15 might drive the phosphorylation of FOXO1/O3 . However , after antigen clearance , IL-7 signaling might drive phosphorylation on the surviving IL-7 receptor-expressing MPECs , and eventually their resultant memory cells . FOXO3 has been reported to suppress expansion of CD8 T cells indirectly by inhibiting IL-6 production by dendritic cells [49] . In this report , however , the T cell intrinsic role of FOXO3 was not assessed in polyclonal CD8 T cells , and monoclonal TCR transgenic CD8 T cells may not always mimic the responses of polyclonal CD8 T cells . In the present study , global FOXO3-deficient mice exhibit increased expansion of polyclonal CD8 T cells specific to multiple epitopes during an acute LCMV infection . Furthermore , infection of T cell-specific conditional FOXO3-deficient mice with LCMV , fully recapitulated the enhanced CD8 T cell expansion seen in global FOXO3−/− mice; even in the absence of CD4 T cells . Therefore , our data implies that FOXO3 suppresses CD8 T cell expansion in vivo by T cell intrinsic and extrinsic mechanisms . This inference is also supported by the reported T cell intrinsic regulation of regulatory T cell development by FOXO1 and FOXO3 [71] , [72] . One of the most interesting findings presented in this manuscript is that the effect of FOXO3 deficiency on CD8 T cell expansion and memory is observed in the spleen , but not in the liver or lymph nodes . The enhanced accumulation of effector CD8 T cells preferentially in spleens of FOXO3−/− mice could not be explained by tissue-specific differences in BIM expression in CD8 T cells directly ex vivo; regardless of the tissue ( spleen , liver , or lymph nodes ) , BIM levels in FOXO3−/− CD8 T cells were slightly lower than in +/+ CD8 T cells ( Figure S5 ) . Additionally , the selective increase in the number of memory CD8 T cells in spleens of FOXO3−/− mice could not be linked to alterations in the expression of molecules such as CD62L , LFA-1 , CCR7 , and CD44 that regulate T cell trafficking ( Figure 10B ) . Tissue-specific effects were observed in both global and T cell-specific conditional FOXO3 knockout mice , which suggests that the local immunological milieu influences the effects of FOXO3 in T cells . It is possible that FOXO3-deficient T cells are hyper-responsive to cellular and environmental cues unique to the spleen during or after cessation of antigen-driven proliferation . Tissue-specific alterations in CD8 T cell homeostasis are not unique to FOXO3 deficiency because lymph node-specific effects on CD8 T cell numbers have been reported in mice deficient for Fas and BIM [64] . Future experiments will address the mechanisms underlying the tissue-specific effects of FOXO3 in regulating CD8 T cell homeostasis . FOXO3 is known to regulate both proliferation and apoptosis by controlling the transcription of genes like p27Kip1 , p130 , BIM , and Fas ligand [26] , [34] , [35] , [63] , [64] . Therefore , the observed increase in the accumulation of LCMV-specific CD8 T cells in FOXO3−/− mice could be attributed to altered proliferation and/or apoptosis . Analysis of in vivo proliferation by multiple strategies indicated that FOXO3 deficiency did not alter proliferation rates of LCMV-specific CD8 T cells in vivo . Clonal expansion of CD8 T cells is associated with concomitant proliferation and apoptosis , therefore , cellular accumulation is the result of the proliferation rate exceeding the apoptotic rate . The competing effects of TGF-β and IL-15 are known to dictate the apoptotic rate of proliferating CD8 T cells , but the signaling mechanisms involved are not well defined [61] . We theorized that IL-7/IL-15 deprivation during antigen-driven proliferation might diminish FOXO3 phosphorylation , and augment the expression of BIM . We find that at day 6 PI , apoptosis of LCMV-specific CD8 T cells was significantly reduced in spleens of FOXO3−/− mice . Additionally , IL-15 deprivation was indeed associated with higher BIM levels in +/+ CD8 T cells than in FOXO3−/− CD8 T cells , which suggested that proliferating FOXO3−/− CD8 T cells might be less susceptible to cytokine withdrawal-induced apoptosis during the expansion phase of the CD8 T cell response . During the early contraction phase ( day 8–11 PI ) , a substantial number of LCMV-specific CD8 T cells are still in cycle ( Figure 5A ) , but during this interval the apoptotic rate presumably exceeds the proliferation rate resulting in a net loss of CD8 T cells . Interestingly , FOXO3 deficiency minimally altered the contraction of LCMV-specific CD8 T cells . These data suggest that mechanisms controlling apoptosis of CD8 T cells during expansion and contraction are likely distinct . Remarkably , the numbers of memory CD8 T cells in the spleen of both FOXO3−/− and FOXO3L mice were substantially higher than in +/+ mice . The number of memory CD8 T cells is dictated by the magnitude of expansion ( clonal burst size ) and contraction of effector CD8 T cells [5] . The magnitude of increase in the number of memory CD8 T cells in FOXO3−/− or FOXO3L mice reflects enhanced expansion of MPECs ( Figure 2B ) during the primary CD8 T cell response . Importantly , enhancement in the number of memory CD8 T cells induced by FOXO3 deficiency was not associated with detectable alterations in phenotype or effector function . Memory CD8 T cells in LCMV-immune FOXO3−/− mice exhibit strong recall responses and provide effective immunity against a persistent LCMV infection . Thus , FOXO3 deficiency increased the quantity of CD8 T cell memory without affecting their phenotype or effector functions . Memory CD8 T cells are maintained by IL-7 and IL-15-driven proliferative renewal and phosphorylation of FOXO3 is an integral component of the signaling circuitry triggered by IL-7/IL-15 signaling [11] , [54] , [66] , [67] . Additionally , we have previously shown that deficiency of the cell cycle inhibitor p27Kip1 , a target gene for FOXO3 , enhances the homeostatic turnover of memory CD8 T cells [43] . Surprisingly , despite the suggested importance of FOXO3 in regulating the homeostasis of memory T cells , FOXO3 deficiency exerted minimal effects on the proliferative renewal of antigen-specific memory CD8 T cells in vivo [48] . Studies of human memory T cells have ascribed a negative regulatory role for FOXO3 in the persistence of memory CD4 T cells and FOXO3 deficiency would be expected to increase the number of effector memory cells [48] , [73] . However , FOXO3 deficiency did not affect the relative proportions of central and effector memory CD8 T cells . It is plausible that FOXO3 might regulate the persistence of central/effector memory CD4 T cells , and not CD8 T cells . Alternatively , FOXO3 function may be redundant in maintaining fully differentiated memory CD4 and CD8 T cells . In conclusion , this manuscript documents that FOXO3 plays a critical role in controlling the clonal burst size and the magnitude of CD8 T cell memory by T cell intrinsic mechanisms . Furthermore , the enhanced number of memory CD8 T cells induced by FOXO3 deficiency , is maintained for extended periods without compromising its quality . These findings have important implications in vaccine development , and suggest that modulation of FOXO3 activity during the expansion phase might be a fruitful strategy to bolster vaccine-induced CD8 T cell memory and protective immunity .
The generation and characterization of the global FOXO3-deficient ( FOXO3−/− ) mice on the C57BL/6 ( B6 ) background have been described previously [40] . The control wild type B6 ( +/+ ) mice were either littermates or purchased from the National Cancer Institute ( Bethesda , MD ) . Derivation of mice carrying the floxed FOXO3 alleles has been described elsewhere [69] , [70] . Mice carrying the floxed FOXO3 alleles were bred with the CD4-Cre mice at UW-Madison to generate the T cell-specific FOXO3−/− ( FOXO3L ) mice . Littermate +/+ mice were used as controls with the FOXO3L mice . Mice used in experiments were between the ages of 6–8 weeks and all experiments were performed in accordance with the protocols approved by the University of Wisconsin School of Veterinary Medicine Institutional Animal Care and Use Committee ( IACUC ) . The animal committee mandates that institutions and individuals using animals for research , teaching , and/or testing much acknowledge and accept both legal and ethical responsibility for the animals under their care , as specified in the Animal Welfare Act ( AWA ) and associated Animal Welfare Regulations ( AWRs ) and Public Health Service ( PHS ) Policy . Mice were infected with 2×105 PFU of lymphocytic choriomeningitis virus ( LCMV ) Armstrong strain by intraperitoneal ( IP ) injection . Mice that have recovered from an infection with LCMV Armstrong were challenged with LCMV-Clone 13 ( 2×106 PFU by intravenous injection ) . Tissue viral titers were quantified by plaque assay with Vero cell monolayers [74] . Single cell suspensions of splenocytes were stained with antibodies for surface markers including CD8 , CD44 , CD122 , CD127 , CD62L , CCR7 , LFA-1 and KLRG-1 ( BD Biosciences , Franklin Lakes NJ , eBIOSCEINCE , San Diego CA or Southern Biotech , Birmingham AL ) in conjunction with MHC I tetramers ( Db ) specific for the class I-restricted LCMV epitopes , NP396 , GP33 , and GP276 as previously described [15] . Cells were fixed in 2% paraformaldehyde ( PFA ) and acquired in a FACSCalibur or LSR II flow cytometer ( BD Biosciences , Franklin Lakes NJ ) . To quantify intracellular cytokine production , splenocytes were incubated for 5 hours at 37°C with LCMV epitope viral peptides in the presence of Brefeldin A . After stimulation , cells were first incubated with antibodies for surface markers . Next , cells were permeabilized and stained for intracellular cytokines ( IFNγ , IL-2 and TNFα ) using the Cytofix/Cytoperm kit ( BD Biosciences , Franklin Lakes NJ ) . The percentages of cytokine-producing cells were quantified by flow cytometry . Splenocytes were stained for cell surface markers as above . After cell surface staining , cells were fixed and permeabilized using Phosflow lysis and Phosflow PermWash I reagents ( BD Biosciences , Franklin Lakes NJ ) according to the manufacturer's recommendations . Next , cells were blocked for 30 minutes on ice in blocking buffer ( 10% normal goat serum in 2%BSA/PBS ) and subsequently stained with either phospho-specific antibodies ( Cell Signaling Technology , Danvers MA; P-Akt [T308] , P-FOXO1/3 [T24 , T32] , P-mTOR [S2448] ) or non-phospho state-specific antibodies ( Akt , mTOR , FOXO3 , BIM ) . As negative controls for staining , antibodies were pre-incubated/blocked with their specific antigenic peptide for 1 hr at room temperature before adding on to the cells . Following incubation with antibody or peptide blocked antibody , cells were washed twice and incubated with secondary antibody ( Goat anti-Rabbit ALEXA488; Sigma-Aldrich , St . Louis MO ) for 40 minutes . Cells were washed and fixed with 2% PFA . The levels of phospho-specific staining were quantified by flow cytometry . Specific levels of staining ( Corrected Mean Fluorescence Intensity [MFI] ) were calculated using the formula: difference of observed MFI for the phospho-specific protein and peptide-blocked control , divided by the peptide-blocked control . To assess in vivo proliferation of antigen specific cells , mice were administered an IP injection of 1 . 5 mg of 5-Bromo-2′-deoxyuridine ( [BrdU] MP Biomedicals , Solon OH ) followed by exposure to 0 . 8 mg/ml of BrdU in drinking water for the rest of the pulse period . Splenocytes were stained for surface markers and determination of BrdU positive cells was performed using a BrdU staining kit ( BD Biosciences ) . Splenocytes were stained for surface markers and MHC I tetramers as described above . After surface staining , cells were fixed and permeabilized using FACS Lysing Solution and FACS Permeabilization Solution 2 reagents ( BD Biosciences , Franklin Lakes NJ ) and subsequently incubated with antibodies against Ki67 , Bcl-2 or Granzyme B ( BD Biosciences , Franklin Lakes NJ ) for 45 minutes at room temperature . Virus-specific CD8 T cells staining positive for Ki67 , Bcl-2 or Granzyme B were visualized using a FACSCalibur flow cytometer . Data is expressed as either a percentage of antigen-specific CD8 T cells positive for the respective protein or MFI for the indicated protein . After 6 days PI , splenocytes from +/+ and FOXO3−/− or FOXO3L mice were isolated and stained with anti-CD8 and MHC-I tetramers as described above except no red blood cell lysis was performed . Annexin V staining ( BD Biosciences , Franklin Lakes NJ ) was then carried out according to the manufacturer's protocol , with the exception that all staining was performed on ice . The percentage of Annexin V high cells amongst antigen-specific CD8 T cells was determined by flow cytometry and expressed accordingly . Mice were depleted of CD4 T cells through IP administration of 100 µg of the monoclonal antibody GK1 . 5 ( eBioscience , San Diego CA ) at days 0 and 4 relative to LCMV infection . T cells and non-T cells were purified from spleens of +/+ and FOXO3L mice using the anti-CD90 . 2 based MACS cell separation system ( Miltenyi Biotec , Auburn CA ) . Purity of cells was >93% . Cells were subsequently lysed in buffer ( 50 mM HEPES , 100 Mm NaCl , 10 mM EDTA , 10 Mm NaF , 4 Mm Na ( PO4 ) 2 , 1% Triton X-100 , 5 µg/ml Aprotinin , 1 Mm Phenylmethylsulfonylflouride ) , sonicated , and total protein levels in each lysate were determined by the Bicinchoninic Acid protein assay . ( Sigma-Aldrich , St . Louis , MO ) . 20 µg samples were loaded and resolved on a 10% SDS-PAGE . Total levels of FOXO3 protein in each sample were detected using a Rabbit primary antibody specific for FOXO3 ( Cell Signaling Technology , Danvers MA ) followed by a Donkey anti Rabbit F ( ab ) 2 fragment HRP-conjugated secondary antibody ( Thermo Fisher , Rockford IL ) Bands were visualized using chemiluminescence reagents ( Thermo Fisher , Rockford IL ) and presented by use of an HP Deskscan system ( Hewlett-Packard , Palo Alto CA ) . Blots initially probed for FOXO3 were subsequently stripped and re-probed with β-Actin ( Sigma-Aldrich , St . Louis MO ) to serve as a loading control . | CD8 T cells are vital for controlling infections with viruses , intracellular bacteria and protozoa . Induction of T and B cell memory is the basis of vaccinations and cellular immunity to intracellular pathogens depends upon the number and quality of memory CD8 T cells . Understanding the mechanisms that control various facets of CD8 T cell memory is of fundamental importance for development of effective vaccines . In this study , we have identified the transcription factor FOXO3 as a crucial regulator of the magnitude of CD8 T cell memory . During a T cell response , FOXO3 limits the number of memory CD8 T cells by inhibiting the accumulation of memory precursor effector cells that give raise to long-lived CD8 T cells . Loss of FOXO3 activity in T cells led to a durable increase in the number of memory CD8 T cells , and the functional quality of FOXO3-deficient memory CD8 T cells was unaffected by FOXO3 deficiency . Thus , our studies suggest that targeting FOXO3 activity may be a fruitful strategy to augment vaccine-induced CD8 T cell memory and protective immunity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"immunology",
"biology"
] | 2012 | FOXO3 Regulates CD8 T Cell Memory by T Cell-Intrinsic Mechanisms |
Soil-transmitted helminths , such as Trichuris trichiura , are of major concern in public health . Current efforts to control these helminth infections involve periodic mass treatment in endemic areas . Since these large-scale interventions are likely to intensify , monitoring the drug efficacy will become indispensible . However , studies comparing detection techniques based on sensitivity , fecal egg counts ( FEC ) , feasibility for mass diagnosis and drug efficacy estimates are scarce . In the present study , the ether-based concentration , the Parasep Solvent Free ( SF ) , the McMaster and the FLOTAC techniques were compared based on both validity and feasibility for the detection of Trichuris eggs in 100 fecal samples of nonhuman primates . In addition , the drug efficacy estimates of quantitative techniques was examined using a statistical simulation . Trichuris eggs were found in 47% of the samples . FLOTAC was the most sensitive technique ( 100% ) , followed by the Parasep SF ( 83 . 0% [95% confidence interval ( CI ) : 82 . 4–83 . 6%] ) and the ether-based concentration technique ( 76 . 6% [95% CI: 75 . 8–77 . 3%] ) . McMaster was the least sensitive ( 61 . 7% [95% CI: 60 . 7–62 . 6%] ) and failed to detect low FEC . The quantitative comparison revealed a positive correlation between the four techniques ( Rs = 0 . 85–0 . 93; p<0 . 0001 ) . However , the ether-based concentration technique and the Parasep SF detected significantly fewer eggs than both the McMaster and the FLOTAC ( p<0 . 0083 ) . Overall , the McMaster was the most feasible technique ( 3 . 9 min/sample for preparing , reading and cleaning of the apparatus ) , followed by the ether-based concentration technique ( 7 . 7 min/sample ) and the FLOTAC ( 9 . 8 min/sample ) . Parasep SF was the least feasible ( 17 . 7 min/sample ) . The simulation revealed that the sensitivity is less important for monitoring drug efficacy and that both FLOTAC and McMaster were reliable estimators . The results of this study demonstrated that McMaster is a promising technique when making use of FEC to monitor drug efficacy in Trichuris .
Worldwide , infections with soil-transmitted heminths ( STHs ) such as Trichuris trichiura , Ascaris lumbricoides and hookworms ( Ancylostoma duodenale and Necator americanus ) are of major importance for public health , particularly in tropical and subtropical countries where climate factors combined with poor environmental , domestic and personal hygiene ease transmission [1] , [2] . Current efforts to control STH infections involve periodic mass drug treatment of people , particularly of school-aged children , in all endemic areas [3] . Since these large-scale interventions are likely to intensify as more attention is addressed to the importance of these neglected diseases [4] , monitoring drug efficacy will become indispensable in order to detect the emergence of resistance [5] , [6] and/or identify confounding factors affecting the drug efficacy [7] . Thus far , the efficacy of anthelmintics has mostly been monitored qualitatively based on the cure rate . However , the fecal egg count reduction test ( FECRT ) is presently under examination for monitoring the drug efficacy quantitatively [8] , implying the need for a sensitive detection technique which will allow the accurate estimation of the infection intensity based on fecal egg counts ( FEC ) . Various techniques have been used for the detection of STH eggs , yet studies comparing detection techniques based on FEC are scarce . Moreover , little attention has been addressed to their feasibility for mass diagnosis under field conditions ( poorly equipped laboratories and short of professionally trained personnel ) and their ability to estimate the efficacy of administered drugs , in particular in different settings of pre-drug administration infection intensities . The ether-based concentration method [9] and the Kato-Katz thick-smear technique [10] are the most commonly used techniques . The latter was initially designed for the diagnosis of Schistosoma eggs and not for STH such as Trichuris , Ascaris and hookworms . Due to the lack of other quantitative techniques and the importance of Schistosoma in public health , this method became also commonly used for the detection of STH [11] . However , this has important drawbacks when the objective of a study is to examine STH simultaneously . The most important one is the diverse clearing time of the different eggs of STH , eggs of hookworms in particular which impedes further standardization of this technique in large-scaled studies at different study sites [12] , [13] . Utzinger et al . ( 2008 ) [14] recently evaluated the FLOTAC technique for the diagnosis of hookworms in stools of African schoolchildren . This quantitative technique , which has been recently described both for human and veterinary medicine [15] , proved to be more sensitive than the traditionally used techniques . Other candidate techniques for monitoring drug efficacy are the McMaster technique , the precursor of FLOTAC , and the Parasep Solvent Free ( SF ) . McMaster is a quantitative flotation technique which is commonly used in veterinary parasitology and can be easily performed without a centrifuge apparatus [16]–[18] . Parasep SF is a single use , disposable enclosed concentration technique recently developed by DiaSys Europe . In contrast to the traditional ether-based concentration technique , a fat dispersion chamber is used to separate the fat content , therefore reducing the need for chemical reagents in the field . A study was conducted to evaluate the performance of different techniques for monitoring drug efficacy in Trichuris . To this end , the ether-based concentration technique , the Parasep SF , the FLOTAC and the McMaster were compared for the detection of Trichuris eggs in stool of nonhuman primates , focusing on validity , feasibility under field conditions and ability to estimate the ‘true’ drug efficacy using a statistical simulation . Non-human primates are an appropriate model , since these animals share the same STH and have a comparable fecal composition [19] , [20] .
The study was conducted at a Dutch sanctuary for exotic animals . All nonhuman primates involved belonged to the family of Old World monkeys and were representatives of barabary macaques ( Macaca sylvanus ) , vervet monkeys ( Chlorocebus pygerythrus ) , rhesus monkeys ( Macaca mulatta ) , crab-eating macaques ( Macaca fascicularis ) , Sunda pig-tailed macaques ( Macaca nemestrina ) , grivet monkeys ( Chlorocebus aethiops ) and Hamadryas baboons ( Papio hamadryas ) . The animals were housed in 22 groups of one to 15 animals . A total of 100 fecal samples were randomly collected by the animal caretakers on a single occasion in January 2008 . Three grams of feces were suspended in distilled water and strained through a layer of surgical gauze to withhold large debris . After sedimentation for 1 h and centrifugation at 800 g for 5 min , the sediment was suspended in 3 ml of distilled water . An aliquot of 500 µl was randomly assigned to each of the four detection techniques . The feasibility of the 4 the techniques was evaluated on a total number of 90 samples randomly assigned to 3 experienced laboratory technicians . Each of the laboratory technicians processed a set of 1 , 2 , 4 and 8 samples . The time needed to prepare the samples and to clean the devices was measured six times for each set of samples . The preparation period started when the aliquots were distributed and ended when the sample was ready for examination . During the cleaning period all used materials were either disposed of , for single use components , or cleaned in the case of recyclable devices . Examination of the slides or chambers was timed individually for all samples . Each of the samples were examined with all techniques by the same laboratory technician . The estimates of the ‘true’ drug efficacies ( TDE ) by the quantitative techniques was studied in a statistical simulation in R ( version 2 . 4 . 0 , The R Foundation for Statistical Computing ) . To this end , difference between the TDE and the estimated drug efficacy ( EDE ) was examined in two strata with different pre-drug administration infection intensities . In this simulation , the TDE for Trichuris varied from 10 to 90% [22] . Because these values are likely to vary between individuals ( ‘individual’ TDE ) an additional variation was postulated for each value of TDE , ranging from 1 to 5% ( standard deviation ) [7] . Only the positive samples of the quantitative techniques in the present study ( a positive result at any technique ) were included as pre-drug administration samples . These samples were stratified into a stratum with low pre-drug administration intensity ( 0<FEC≤50 EPG ) and one with a high pre-drug administration intensity ( FEC>50 EPG ) . The cut off of 50 EPG was used , since this was the highest detection limit of all the techniques . The samples in each strata were combined into 4 different sample sets of 100 , 250 , 500 and 1000 samples to examine the effect of sample size on the efficacy estimates . The individual post-drug administration infection intensities for each technique were obtained by multiplying the individual pre-drug administration FEC of each technique with ( 100%-individual TDE ) . The obtained individual post-drug administration FEC of each technique were corrected according the results of the validity . The estimated efficacy of each individual for each technique was calculated as the difference of individual pre- and post-drug administration FEC divided by the individual pre-drug administration FEC obtained by each technique . The difference between the TDE and the individual EDE ( bias ) for each detection technique , standard deviation , sample set was examined within each stratum of pre-drug administration infection intensities . To assess the validity of the techniques both qualitative and quantitative aspects were compared . The agreement in qualitative test results between the 4 techniques was measured using the kappa statistic ( κ ) ( SAS 9 . 1 . 3 , SAS Institute Inc . ; Cary , NC , USA ) . Both sensitivity and negative predictive value ( NPV ) were estimated for each method . To this end , a positive result at any technique was considered as a ‘true’ positive test result ( specificity = 1 , positive predictive value = 1 ) . The agreement in quantitative test results was estimated by the Spearman rank correlation coefficient ( SAS 9 . 1 . 3 , SAS Institute Inc . ; Cary , NC , USA ) . In addition , the Wilcoxon signed rank test was assessed to test for differences in FEC between the techniques . For this end , a Bonferroni pair-wise comparison procedure was performed and the level of significance was set at 0 . 0083 . Furthermore , samples were subdivided according infection intensity based on guidelines of the WHO [23] . A linear mixed model was built using the Tukey pair-wise comparison test to evaluate the feasibility of the 4 techniques ( SAS 9 . 1 . 3 , SAS Institute Inc . ; Cary , NC , USA ) . This model examined the differences in time taken for preparing and cleaning a sample between the different sample sets and the 4 techniques . The differences in time taken for the microscopic examination of the samples prepared by the 4 techniques and the effect of FEC on time taken were studied separately . This study was embedded in a annual parasitological survey of the animals housed at the sanctuary for exotic animals AAP . No approval of the Ethics Committee was needed since all samples were randomly collected during cleansing of the enclosures and none of the animals involved were immobilized neither physically or chemically ( European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes , Strasbourg , 18 . III . 1986 ) .
Based on FLOTAC , a total of 47 positive samples of which 23 had a FEC of no more than 50 EPG were included in the simulation . Additional samples were randomly chosen to add to the strata ( 2 for the low pre-drug administration infection intensity stratum , 1 for the high pre-drug administration infection intensity stratum ) , resulting in two subsets of each 25 samples . Each subset was combined 4 , 10 , 20 and 40 times to obtain the different sample sets of 100 , 250 , 500 and 1000 , respectively . Both the ether technique and Parasep SF were withdrawn for further analysis , since these detection techniques resulted in significant lower FEC compared to FLOTAC and McMaster . Based on the results of the validity for McMaster , the individual post-treatment FEC were corrected using the following conditions . At first , all positive post-drug administration samples which had no more than 5 EPG were negative for McMaster . Samples with a FEC of more than 5 EPG , but no more than 50 , were positive with a probability of 0 . 40 and the FEC was set on 50 EPG if positive . The FEC was rounded off to the nearest multiple of 50 , in all other cases . Although FLOTAC revealed to be 100% sensitive , positive samples with no more than 2 EPG were positive with a probability of 0 . 40 and FEC was set on 2 EPG if positive . In all other cases , the FEC was rounded off to the nearest multiple of 2 . Figure 4 describes the bias ( difference between TDE and the individual EDE ) in 2 strata with different pre-drug administration infection intensities . Overall , there is a large bias in estimating the drug efficacy when low pre-drug administration infection intensities were included . The mean bias in the stratum including FEC not higher than 50 EPG ranged from −20 . 8 to +16 . 1% for FLOTAC and from −45 . 0 to +20 . 0% for McMaster . Moreover , this bias is likely to change over TDE . Low TDE ( ≤60 . 0% ) were overestimated and high TDE ( >60 . 0% ) were underestimated . This is in contrast to the stratum where only FEC higher than 50 EPG were considered , in particular for FLOTAC which resulted in an accurate estimation of the TDE ( median = 0 . 0% , range = −0 . 6; +0 . 7% ) . The differences for McMaster varied from −6 . 4 to +2 . 3% ( median = −1 . 1% ) , and decreased with the TDE-values of at least 50% ( median = −0 . 3% , range = −3 . 6; +2 . 1% ) . In average , McMaster resulted in higher efficacies than FLOTAC , since the bias was more negative than the efficacy estimated by FLOTAC in the majority of the ‘true’ drug efficacies in both strata . Both the different standard deviations and sample sets did not affect the estimates of both techniques ( Figure 5 ) , since the bias remained unchanged over the different standard deviations and sample sets .
In the present study , four techniques were compared for the qualitative and quantitative detection of Trichuris in stools of nonhuman primates , as well as their feasibility for mass diagnosis under field conditions . In addition , their ability to give accurate estimates of the ‘true’ drug efficacy was studied based on a statistical simulation . Overall , the observed prevalence of Trichuris in these animals was 47% ( 95% CI: 37–57% ) and remained unchanged when the test results of the 4 techniques were combined . Although the test properties might be overestimated due to the absence of a diagnostic ‘gold’ standard , it is clear that FLOTAC is the most sensitive technique , followed by the Parasep SF and the ether-based concentration technique . McMaster is the least sensitive , because it often fails to detect low FECs due to its relative high detection limit ( 50 EPG ) . Multiple comparisons of the 4 techniques revealed a linear correlation in FEC . Nevertheless , both the Parasep SF and ether techniques are likely to be less appropriate for an accurate estimation of FEC , since the McMaster and the FLOTAC detected significantly more eggs . The time taken for preparing samples and cleaning between samples was the lowest for the McMaster . FLOTAC was the most time- and labour-consuming technique over the different sample sets and therefore seems to be less feasible in large-scale studies . Samples obtained by the Parasep SF protocol were the hardest to examine , because of the large amount of fecal debris recovered with the eggs . This resulted in decreased clarity and a much greater length of time required to examine the samples . McMaster slides were examined at the highest speed . The most important reasons explaining the difference in time required for reading between McMaster and the other techniques are the surface of the slides ( McMaster: 1 cm2 versus other techniques: 3 . 24 cm2 ) and the detection limit ( McMaster: 50 versus other techniques: 2 ) , as FEC had no significant effect on the time needed to examine the samples using McMaster . Moreover , McMaster slides were likely to be less contaminated by fecal debris , since only a small proportion of already diluted samples was examined . Overall , McMaster was the most feasible and does not need any centrifuge apparatus , in contrast with the other techniques , which clearly emphasizes its usefulness in poorly equipped and often short-staffed laboratories . Estimating drug efficacies should be done on samples with high FEC ( >50 EPG ) , as including samples with lower FEC may result in a significant bias . As a consequence , sensitivity as a criterion for detection techniques for monitoring drug efficacy is less important . FLOTAC resulted in the most accurate estimates of TDE in the stratum of high FEC . However , the bias when using McMaster is minimal ( median = −1 . 1% ) and is comparable to FLOTAC if the TDE ( median = −0 . 3% ) is at least 50% . Although the results presented in this study were obtained from stool of nonhuman primates , these will also be applicable in human parasitology . These animals not only share the same STH species with humans , they also have a similar fecal composition [19] , [20] . Moreover , the prevalence of Trichuris is comparable to those of previous epidemiological studies in pre-school children , where in average 39% of the subjects were infected [3] . Furthermore , the distribution of the FEC found in these animals was similar . Based on WHO guidelines , 23 . 4% ( FLOTAC ) to 30 . 0% ( McMaster ) of the Trichuris infections fell into the moderate FEC range , where this was roughly 25% in Zanzibari infants [13] . Not including the Kato-Katz method in the present study is a major shortcoming . However , the present study suggests that McMaster is likely to be more feasible . The microscopic view is clear and all parasites can be examined simultaneously , which is in contrast to the Kato-Katz technique due to a different clearing time of the different STH . Based on previous study where both FLOTAC and Kato-Katz were compared for the detection of hookworm [14] , we expect that both the sensitivity and the FEC of the Kato-Katz will be comparable to McMaster . McMaster is commonly used for both diagnosis and drug efficacy monitoring programs of gastrointestinal parasites in livestock , including Ascaris and hookworm [16]–[18] , [24] , but its usefulness in detecting these STH in public health still needs to be confirmed by further studies in endemic areas where also A . lumbricoides and hookworms are present . In conclusion , this study indicates that McMaster holds promise as the method of choice for monitoring drug efficacy , since sensitivity appeared to be a less important criterion . It is a quantitative technique which can be easily performed under field conditions and gives reliable estimates of TDE which are at least 50% . | Worldwide , millions of people are infected with soil-transmitted helminths , particularly in developing countries . Efforts to control these infections involve periodic mass drug treatment in endemic areas . Since these large-scale interventions are likely to intensify , monitoring of drug efficacy has become a key issue in order to detect the emergence of resistance . At present , the drop in infection intensity is under examination for monitoring the drug efficacy . However , studies comparing detection techniques based on infection intensities are scarce . Moreover , little attention has been addressed to their feasibility and their ability to estimate drug efficacies . We have compared different techniques for the detection of whipworm ( Trichuris ) in simian stool samples based on prevalence , infection intensities , feasibility and ability to estimate the ‘true’ drug efficacy . We have found that techniques often fail to detect low infection intensities and that not all techniques are appropriate for estimating infection intensities . The time needed to obtain a test result varied from 3 . 9 to 17 . 7 min/sample . Finally , accurate estimates of drug efficacy were only obtained in high pre-drug administration infection intensities . To conclude , along with accurate estimates of infection intensities , feasibility is a considerable criterion for the detection techniques used in drug efficacy monitoring programs . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"public",
"health",
"and",
"epidemiology/preventive",
"medicine",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"public",
"hea... | 2009 | Field Validity and Feasibility of Four Techniques for the Detection of Trichuris in Simians: A Model for Monitoring Drug Efficacy in Public Health? |
Haiti's cholera epidemic has been devastating partly due to underlying weak infrastructure and limited clean water and sanitation . A comprehensive approach to cholera control is crucial , yet some have argued that oral cholera vaccination ( OCV ) might result in reduced hygiene practice among recipients . We evaluated the impact of an OCV campaign on knowledge and health practice in rural Haiti . We administered baseline surveys on knowledge and practice relevant to cholera and waterborne disease to every 10th household during a census in rural Haiti in February 2012 ( N = 811 ) . An OCV campaign occurred from May–June 2012 after which we administered identical surveys to 518 households randomly chosen from the same region in September 2012 . We compared responses pre- and post-OCV campaign . Post-vaccination , there was improved knowledge with significant increase in percentage of respondents with ≥3 correct responses on cholera transmission mechanisms ( odds ratio[OR] 1 . 91; 95% confidence interval[CI] 1 . 52–2 . 40 ) , preventive methods ( OR 1 . 83; 95% CI 1 . 46–2 . 30 ) , and water treatment modalities ( OR 2 . 75; 95% CI 2 . 16–3 . 50 ) . Relative to pre-vaccination , participants were more likely post-OCV to report always treating water ( OR 1 . 62; 95% CI 1 . 28–2 . 05 ) . Respondents were also more likely to report hand washing with soap and water >4 times daily post-vaccine ( OR 1 . 30; 95% CI 1 . 03–1 . 64 ) . Knowledge of treating water as a cholera prevention measure was associated with practice of always treating water ( OR 1 . 47; 95% CI 1 . 14–1 . 89 ) . Post-vaccination , knowledge was associated with frequent hand washing ( OR 2 . 47; 95% CI 1 . 35–4 . 51 ) . An OCV campaign in rural Haiti was associated with significant improvement in cholera knowledge and practices related to waterborne disease . OCV can be part of comprehensive cholera control and reinforce , not detract from , other control efforts in Haiti .
In October 2010 , a cholera outbreak began in the Artibonite and Centre Departments of Haiti [1] . By December , cholera had been identified in all 10 departments of Haiti and has since reached neighboring countries [2] , [3] . Cholera is an acute , watery diarrheal infection caused by the bacterium Vibrio cholerae of the O1 or O139 serogroup; and it can rapidly lead to severe dehydration and death if untreated . However , effective therapy can decrease mortality rate from more than 50% to less than 0 . 2% [4] . Efforts to control the cholera outbreak have been hampered by weak health systems and lack of clean water and adequate sanitation in Haiti . In 2008 , only 17% of Haiti's population used improved sanitation facilities while 12% had access to piped , treated water [5] . In addition , conditions in Haiti further deteriorated on January 12 , 2010 when the country suffered a devastating 7 . 0-magnitude earthquake that killed thousands and rendered approximately 2 million individuals homeless [6] . Pockets of densely populated areas resulting from internal migration after the earthquake likely contributed to an explosive outbreak in Haiti . Rural areas and urban slums were particularly vulnerable to the rapid spread of a waterborne disease such as cholera . Furthermore , Haiti's population had no prior exposure or immunity to V . cholerae [7] . Moreover , analysis of the V . cholerae strain in Haiti revealed a variant strain ( serotype Ogawa , biotype El Tor ) known to be associated with more severe illness [8] , [9] . Between October 2010 and May 2013 , there were over 600 , 000 cases of infection and more than 8 , 000 cholera deaths reported [10] . In 2011 , the cholera epidemic in Haiti accounted for 58% of all cholera cases and 37% of all cholera deaths reported to the World Health Organization ( WHO ) [11] . A comprehensive approach is necessary to fight the cholera epidemic in Haiti and proven cholera control measures include: active case finding , improving water and sanitation , and widespread hygiene education [12]–[14] . In addition , there are two safe oral cholera vaccines ( OCV ) , approved by the WHO for use in cholera endemic areas [15] . Some have argued that cholera vaccination might detract from other prevention efforts and result in diminished hygiene practices among vaccine recipients [16]–[18] . Yet , there is no evidence indicating that cholera vaccination reduces hygiene practice . Knowledge , Attitude , and Practice ( KAP ) surveys have been used in various settings to assess existing knowledge and hygiene practices relevant to prevention and transmission of diarrheal diseases , including cholera [19]–[22] . KAP surveys have also been employed in areas of cholera outbreak to measure uptake of knowledge and behavioral changes in response to educational activities aimed at cholera control [23] , [24] . In December 2010 , a KAP survey was conducted in resource-limited communities of Port-au-Prince , Haiti to assess the effectiveness of public health campaigns on cholera education [24] . The study showed high knowledge of cholera signs and transmission mechanisms as well as improvement in water treatment practices after the outbreak . However , there have been no studies evaluating the effect of vaccination campaigns for waterborne , diarrheal diseases on knowledge and practices related to these diseases . We aimed to assess the impact of an OCV campaign on knowledge of cholera and health practice related to waterborne illness in rural Haiti . We hypothesized that the campaign , which had been implemented with an educational component , would lead to improved knowledge and behavior critical for cholera control and therefore had served to bolster efforts in the fight against cholera in Haiti .
Ethics Statement: Data were collected as part of a public health campaign; therefore informed consent was not required from survey respondents . Institutional Review Board approval was received from Partners Healthcare for post-hoc analysis of the de-identified dataset . We analyzed data from the rural 5th section of St . Marc , also known as Bocozel ( Figure S1 ) , in the Artibonite Department of Haiti , where between May and June 2012 , the non-profit organization , Partners In Health , carried out a pilot OCV campaign in support of the Haitian Ministry of Health [25] . In February 2012 , prior to vaccine implementation , a census was undertaken in Bocozel , resulting in enumeration of 9 , 517 households . Empty households were visited twice , and if neighboring households could not provide information to confirm that a third visit was warranted , the household was not counted in the census . During the census , every 10th household was invited to participate in a baseline survey on knowledge and practices regarding cholera and waterborne disease . The survey gathered information on sociodemographic characteristics; knowledge about means of cholera transmission , preventive measures , and water treatment modalities; practices related to frequency of water treatment and hand washing; type of toilet access; and source of drinking water . Knowledge questions prompted respondents to provide as many answers as they could to the following questions: “How can a person get cholera ? ” “What can you do to avoid getting cholera ? ” and “What are the methods of treating water that you drink ? ” Examples of appropriate responses for cholera transmission mechanisms included: “drinking untreated water , ” “eating uncooked food , ” and “dirty hands . ” For cholera prevention methods , suitable answers included: “treat water , ” “eat cooked food , and “wash hands . ” For hygiene practices , respondents were asked to choose the option that described their frequency of water treatment among: “always , ” “almost always , ” “often , ” “sometimes , ” and “almost never . ” Respondents were also asked to report the number of times they washed their hands with soap and water daily . Knowledge questions were directed to the individual responding , and practice questions were related to the household . Trained enumerators ( locally recruited Haitians who had completed high school ) administered surveys to one adult individual ( male or female , ≥18 years ) identified by members of the household as the head or , in the absence of head of household , a representative of the household . Enumerators received a 2-day training on the use of hardware and software used for data collection as well as the survey modules . Refresher trainings were conducted prior to the administration of each vaccine dose . The OCV campaign was executed in 2 phases with individuals aged 10 years and above targeted in the first phase , and children between the ages of 1 and 10 years targeted in the second phase . The campaign is described in detail elsewhere [25] . Prior to the campaign , meetings with key stakeholders , community focus groups , and Ministry of Health representatives led to the generation of key messages about cholera prevention and cholera vaccine that were used as part of the vaccination campaign ( Table S1 ) . Before and throughout the period of vaccination , educational information was disseminated verbally via radio shows , sound trucks , town criers , local television and was printed on T-shirts and posters . Members of the vaccination team were encouraged to share education messages at every contact with the public . These messages were also communicated by enumerators to household members in the census , after all data collection was complete . Education information was thus provided directly to at least one representative of all enumerated households . All vaccine recipients received the same information during vaccination days , and the entire community received information during the period of the campaign . Printed educational information was not a major focus of the campaign because of low literacy rates in the region . In September 2012 , after the vaccine campaign , a follow-up survey was conducted to estimate vaccination coverage , and as a secondary objective , to evaluate knowledge and practice about cholera . De-identification of pre-vaccine survey data precluded resurveying the same participants; therefore , a list of 600 households was randomly generated from the 9 , 517 households enumerated during the census using a random number generator in Microsoft Excel . The same survey tool used in the pre-vaccination phase was administered to these households in addition to questions about receipt of cholera vaccine . The same enumerators collected census data and conducted both surveys with the exception of a few staff who were not available at the second time point . We analyzed results from both surveys using Statistical Analysis System ( SAS 9 . 3 ) . Chi-square and Wilcoxon rank-sum tests were used to compare knowledge and practice variables from the pre- and post-vaccination surveys . We used multivariable logistic regression analysis to ( 1 ) evaluate changes in knowledge of cholera prevention and transmission and hygiene practices after the vaccine campaign; ( 2 ) examine whether proxies for socioeconomic status ( i . e . ever having attended school and access to electricity at home ) were associated with these outcomes; and ( 3 ) assess whether cholera knowledge was associated with hygiene practices . Multivariable models included a variable for survey ( 1 versus 2 ) , ever having attended school , and electricity access in the home . To assess for confounding , we first identified baseline variables that were differentially distributed between the two surveys and were associated with any outcome at a p-value≤0 . 05 . These variables ( farming occupation , latrine , open defecation ) were then included in the multivariable models and those that altered the effect estimate for the survey variable by >10% were retained in the final model .
Vaccine coverage is described in detail elsewhere and was estimated between 76 . 7–92 . 7% of the population of the region , with the lower limit of the range estimated by census and registration data and the upper limit estimated from Survey2 [25] . A total of 41 , 242 individuals received 2-dose series of the OCV . Of the 518 Survey2 respondents , 480 ( 92 . 7% ) [95% CI 90 . 1%–94 . 6%] reported receipt of at least one dose of the cholera vaccine , and 419 ( 80 . 8% ) [95% CI 77 . 3%–84 . 0%] provided their vaccination cards for verification . Baseline demographic characteristics for pre-and post-vaccine survey respondents were generally similar ( Table 1 ) ; however statistically significant differences between the two time points were observed for household size , number of people sharing a toilet , toilet type , and having a farming occupation . 65 . 2% of Survey1 respondents reported use of latrine compared to 46 . 9% in Survey2 . Farming was the most common occupation representing 69 . 5% of Survey1 respondents and 76 . 1% in Survey2 . Nearly all respondents pre-vaccine ( 99 . 1% ) and post-vaccine ( 99 . 6% ) had heard of cholera . A high level of knowledge was defined as greater than the median number of correct answers in Survey1 ( Table 2 ) . A significantly higher proportion of Survey2 respondents ( 63 . 8% ) knew ≥3 correct modes of cholera transmission compared to 48 . 1% in Survey1 ( p<0 . 0001 ) . A similar pattern was observed with cholera prevention questions . Pre-vaccination , 50 . 0% of respondents provided ≥3 correct answers on how to avoid cholera compared to 64 . 5% post-vaccine ( p<0 . 0001 ) . Finally , a higher percentage of individuals in Survey2 ( 44 . 1% ) knew ≥3 means of water treatment compared to Survey1 ( 22 . 6% ) with p<0 . 0001 ( Figure 1 ) . None of the differentially distributed baseline variables significantly changed the effect estimates for any outcome; therefore , only the socioeconomic proxy variables ( ever having attended school and access to electricity at home ) , and no additional variables , were included as covariates in the final multivariable models . For cholera knowledge , post-vaccination surveys were associated with a statistically significant increase in the odds of providing at least 3 correct responses on means of cholera transmission ( odds ratio [OR] 1 . 91; 95% CI 1 . 52–2 . 40; p<0 . 0001 ) . For cholera prevention measures , the odds ratio of knowing 3 or more correct answers in Survey2 compared to Survey1 was 1 . 83 ( 95% CI , 1 . 46–2 . 30; p<0 . 0001 ) . Similarly , there was also greater odds of knowing ≥3 ways to treat water in Survey2 relative to Survey1 ( OR 2 . 75; 95% CI , 2 . 16–3 . 50; p<0 . 0001 ) . Ever having attended school and electricity access in the home , were not generally associated with increased knowledge ( Table 3 ) ; however , we did observe a positive relationship between access to electricity in the home and knowing 3 or more means of avoiding cholera of borderline statistical significance ( OR: 1 . 37; 95% CI: 1 . 00–1 . 89 ) . The percentage of respondents who reported “always” treating their water increased from 49 . 4% in Survey1 to 62 . 0% in Survey2 ( p<0 . 0001 ) . The most common reasons provided for not always treating water were related to access to products . 35 . 9% had “no products” in Survey1 and 49 . 2% reported the same reason in Survey2 . Products were “hard to get” for 28 . 2% and 35 . 0% of respondents in Survey1 and Survey2 respectively . Regarding hand washing practices , 46 . 7% of Survey2 respondents reported hand washing with soap and water >4 times a day compared to 41 . 1% in Survey1 ( p 0 . 05 ) . We observed decreased use of river water in Survey2 ( 42 . 7% ) versus Survey1 ( 48 . 0% ) , although this was not statistically significant ( p 0 . 06 ) . Multivariable regression analysis of hygiene practice revealed that relative to the pre-vaccination period , post-vaccination participants were more likely to report always treating water ( OR1 . 62; 95% CI , 1 . 28–2 . 05; p<0 . 0001 ) . Similarly , odds of washing hands with soap and water >4 times a day was increased in Survey2 relative to Survey1 ( OR1 . 30; 95% CI , 1 . 03–1 . 64; p 0 . 03 ) . Higher socioeconomic status , as measured by ever having attended school and access to electricity , was associated with increased odds of always treating water and hand washing with soap and water >4 times a day ( Table 3 ) . There were no confounding variables associated with practice questions . Knowledge of water treatment as a means of preventing cholera was associated with the practice of always treating water ( OR 1 . 47; 95% CI , 1 . 14–1 . 89; p 0 . 003 ) . Overall , there was no statistically significant association between knowledge of hand washing as a cholera preventive measure and practice of frequent hand washing ( OR 1 . 10; 95% CI , 0 . 82–1 . 46; p 0 . 53 ) . However , in stratified analyses , knowledge of hand washing as a preventive measure was significantly associated with the practice of washing hands >4 times a day post-vaccine ( OR 2 . 47; 95% CI 1 . 35–4 . 51; p 0 . 003 ) but not pre-vaccine ( OR 0 . 85; CI 0 . 61–1 . 19; p 0 . 35 ) .
After an integrated cholera vaccination campaign in rural Haiti , surveys demonstrate a significant increase in knowledge of cholera transmission and prevention mechanism as well as improvement in practices of water treatment and frequent hand washing , which are critical for curbing the spread of diarrheal diseases such as cholera . This provides evidence that oral cholera vaccination can be part of comprehensive cholera control and can reinforce , rather than detract from , other prevention activities in Haiti . | In October 2010 , Haiti experienced a cholera outbreak that is now considered one of the largest cholera epidemics in recent history . A comprehensive approach is necessary to successfully fight the epidemic and proven methods for controlling cholera include improving access to clean water and sanitation as well as widespread hygiene education . In addition , there are two safe cholera vaccines approved for use . The authors conducted surveys before and after a cholera vaccination campaign , that included a public health educational component , in rural Haiti; surveys addressed knowledge of cholera and hygiene practices such as hand washing and water treatment , which are crucial for preventing waterborne diseases such as cholera . The authors found that after the vaccination campaign , knowledge of cholera improved significantly . There was also significant increase in reported hand washing and water treatment post vaccination . Furthermore , there was an association between knowledge and hygiene practices . Therefore , this study demonstrates that cholera vaccination can be a complementary tool in the fight against cholera in Haiti and will not detract from other control efforts . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | Cholera Vaccination Campaign Contributes to Improved Knowledge Regarding Cholera and Improved Practice Relevant to Waterborne Disease in Rural Haiti |
Rhodopsin mistrafficking can cause photoreceptor ( PR ) degeneration . Upon light exposure , activated rhodopsin 1 ( Rh1 ) in Drosophila PRs is internalized via endocytosis and degraded in lysosomes . Whether internalized Rh1 can be recycled is unknown . Here , we show that the retromer complex is expressed in PRs where it is required for recycling endocytosed Rh1 upon light stimulation . In the absence of subunits of the retromer , Rh1 is processed in the endolysosomal pathway , leading to a dramatic increase in late endosomes , lysosomes , and light-dependent PR degeneration . Reducing Rh1 endocytosis or Rh1 levels in retromer mutants alleviates PR degeneration . In addition , increasing retromer abundance suppresses degenerative phenotypes of mutations that affect the endolysosomal system . Finally , expressing human Vps26 suppresses PR degeneration in Vps26 mutant PRs . We propose that the retromer plays a conserved role in recycling rhodopsins to maintain PR function and integrity .
Rhodopsins are G protein-coupled receptors that function as light sensors in photoreceptors ( PRs ) , and defective trafficking of rhodopsins often leads to PR degeneration in humans and flies [1]–[5] . Because vision is not required for animal survival , previous studies in Drosophila mostly focused on viable mutations that specifically impair PR function [1] . However , it is likely that numerous additional players encoded by essential genes have remained unidentified . We performed an eye-specific mosaic genetic screen [6] and found that loss of subunits of the retromer causes light-induced PR degeneration . The retromer , a hetero-multimeric protein complex , retrieves specific proteins from endosomes , thereby preventing the degradation of these proteins in lysosomes [7]–[9] . The retromer is composed of Vps26 , Vps29 , Vps35 , and certain sorting nexins ( Snx ) ( Figure 1A , [7]–[9] ) . Most subunits are evolutionarily conserved ( Figure 1A , [7]–[9] ) . Mutations in some subunits ( Vps35 or Snx3 ) of the retromer have been shown to decrease the abundance of Wntless ( Wls ) and impair the secretion of Wingless ( Wg ) , a ligand of the Wnt signaling pathway [10]–[14] . Wls is a transmembrane protein that binds to Wg and is required for Wg secretion [15] , [16] . Impaired retromer function leads to excessive degradation of Wls in lysosomes , severely reducing Wg secretion and signaling [10]–[14] . The retromer has also been shown to maintain the levels of Crumbs , a transmembrane protein required for maintaining the apicobasal polarity in some tissues [17] , [18] . More recently , mutations in human VPS35 have been shown to cause a dominant inherited form of Parkinson's disease ( PD ) [19] , [20] . However , the retromer has not been implicated in neurons of the visual system in flies or vertebrates . The Drosophila compound eye comprises ∼800 hexagonal units named ommatidia [1] , [2] , [21] , [22] . Each ommatidium contains eight PRs ( R1–R8 ) that express rhodopsin proteins [1] , [2] , [21]–[23] . Rhodopsin 1 ( Rh1 ) is the major rhodopsin that is primarily expressed in R1–R6 [1] , [2] , [21] , [22] . It is synthesized and folded in the endoplasmic reticulum ( ER ) and transported to rhabdomeres , the stacked membranous structures in PRs , via the secretory pathway [1] , [2] , [21] . The proper transport of Rh1 from ER to rhabdomeres requires molecular chaperones [24]–[30] and Rab GTPases [24]–[33] . Binding of opsins to chromophores [34]–[40] as well as protein glycosylation and deglycosylation [41]–[44] are essential for Rh1 folding and maturation . Mutations in genes involved in Rh1 synthesis , folding , or transport can result in defective PR development or PR degeneration [24] , [25] , [32] , [41]–[43] , [45]–[51] . Phototransduction in the PRs relies on the activation of Rh1 by photons ( Figure S1A , [52] ) . Active Rh1 ( metarhodopsin , M* ) activates phospholipase C ( PLC ) [53] , which hydrolyzes phosphatidylinositol 4 , 5-bisphosphate ( PIP2 ) to produce diacylglycerol ( DAG ) [54] . DAG or its metabolites can activate Transient Receptor Potential ( TRP ) and TRP-like cation channels that lead to depolarization of the PRs [55]–[59] . Similar to fly PRs , the vertebrate intrinsically photosensitive retinal ganglion cells ( ipRGCs ) use melanopsin ( a homolog of fly Rh1 ) as the light sensor and requires PLC and TRPC channels for activation [60]–[63] . ipRGCs project their axons to specific brain areas to control circadian rhythms [64]–[67] or pupillary light reflex [64] , [68] , [69] . Tight regulation of Rh1 activity upon light exposure is required to maintain the integrity of PR cells [1] , [2] . M* can be converted into its inactive form upon exposure to orange light [52] . In addition , a significant portion of active Rh1 can be endocytosed and degraded in lysosomes [70]–[72] . Mutations that abolish Rh1 deactivation [72]–[74] or impair the endolysosomal system [71] , [75] , [76] can cause PR degeneration due to Rh1 accumulation . However , it is unknown whether Rh1 can be retrieved from the endolysosomal compartments and whether impaired Rh1 recycling leads to PR degeneration . Here , we show that loss of the fly Vps26 or Vps35 causes early-onset PR degeneration . We show that retromer subunits are expressed in PRs in flies and melanopsin-expressing ipRGCs in the mouse retina . In fly mutant PRs , the numbers of late endosomes and lysosomes are significantly elevated . The PR degenerative phenotypes are dependent on exposure to light and the presence of Rh1 . Our data indicate that the fly retromer recycles Rh1 , preventing Rh1 retention in the PR cell bodies and shunting Rh1 from being degraded in lysosomes , thereby promoting Rh1 redelivery to rhabdomeres . In summary , the retromer recycles Rh1 , prevents an overload of the endolysosomal pathway , and salvages a substantial fraction of Rh1 from degradation in flies . It may also play a similar role in ipRGCs in the retina of vertebrates .
To identify mutations that cause PR degeneration , we generated large mutant clones of essential genes in the eyes ( Figure S1B ) and screened for mutants that exhibit an age-dependent decrease in amplitude of electroretinogram ( ERG ) recordings [6] , [77] , [78] . Mutations in a complementation group XE52 , which we later mapped to Vps26 ( Figure S2A ) , result in a progressive loss of ERG amplitudes compared to the control ( Figure 1B–C and Figure S1C ) . Reduced amplitudes in ERG recordings may result from aberrant PR development , a defective phototransduction pathway , or PR degeneration [1] . To address if the ERG defects are due to aberrant PR development , we examined the light response of dark-reared control ( y w FRT19Aiso , the isogenized y w FRT19A flies , or iso; see Materials and Methods ) and two Vps26 mutant alleles . Upon rearing the flies in the dark , the Vps261 and Vps262 mutants exhibit nearly normal ERG responses and the ERG amplitude and on- and off-transients in Vps261 mutant PRs are slightly reduced when compared to control ( Figure 1B–C and Figure S1C ) . In addition , imaging of newly eclosed flies ( see definition in Materials and Methods ) revealed that the overall morphology of ommatidia is similar to control , although loss of a rhabdomere is observed at low frequency in the Vps261 mosaic clones ( Figure 1D–E and Figure S1D–E ) . These data indicate that mutations in Vps26 do not significantly affect PR development , allowing nearly normal ERGs in young flies . To assess the impact of the Vps26 mutations in aged flies , we raised the animals either in a 12-h light/dark cycle ( LD , L = 1 , 800 lux ) or in constant darkness ( DD ) and recorded the ERGs at 0 , 2 , 4 , 10 , and 21 d . In LD , both alleles show a loss of on- and off-transients , indicating a loss of synaptic transmission . They also exhibit a reduced depolarization by day 2 , and a progressive loss of the ERG amplitude with age ( Figure 1B–C and Figure S1C ) . At day 21 , the light-induced PR depolarization in Vps261 and Vps262 mutant PRs is almost completely absent ( Figure 1B–C and Figure S1C ) . Strikingly , 3-wk-old DD-reared Vps261 and Vps262 mutants exhibit nearly normal ERG responses , similar to newly eclosed flies ( Figure 1B–C and Figure S1C ) . The PR morphology of Vps261 and Vps262 mutants is severely affected in flies kept in LD for 21 d ( Figure 1D–E and Figure S1D–E ) . In addition , we constructed an EGFP-tagged genomic rescue construct ( Vps26-gEGFP; see below and Figure S2A ) that fully rescues the lethality ( Figure S2B ) , ERG , and morphological defects of both Vps261 and Vps262 alleles ( Vps261 GR in Figure 1B–E; Vps262 GR in Figure 1C and S1C–E ) . However , PR morphology is nearly wild-type when the flies are raised in DD ( Figure 1D–E and Figure S1D–E ) . In summary , PR degeneration in the Vps26 mutant clones is light-dependent . To identify the mutations in XE52 complementation group consisting of two alleles , we performed duplication , deficiency , and P[acman] rescue mapping ( Yamamoto et al . , in preparation ) , narrowing the locus to a 21 kb genomic region ( Figure S2A ) . The XE52 alleles fail to complement the lethality of P{EP}Vps26G2008 ( Figures S2A–B , [79] ) , a transposable element inserted in the 5′-UTR of the Vps26 gene . Sanger sequencing revealed mutations in conserved amino acid residues in both alleles: a nonsense mutation ( K264* ) and a missense mutation ( V267D ) in Vps261 and Vps262 , respectively ( Figure 1F ) . To test if Vps26 is responsible for both the lethality and PR degenerative phenotypes in XE52 alleles , we generated the Vps26-gEGFP genomic rescue construct mentioned above ( Figure S2A ) as well as a UAS-Vps26 full-length cDNA construct [80] . Both the genomic rescue and ubiquitously expressed cDNA ( tub-Gal4>UAS-Vps26 ) rescue all phenotypes ( Figure 1B–E , Figure S1C–E , Figure S2B , and unpublished data ) , confirming that the phenotypes are caused by loss of Vps26 . Vps26 is a subunit of the retromer ( Figure 1A , [81] , [82] ) . Because loss of retromer function in Vps35 and Snx3 mutants has been previously documented to lead to defects in Wg signaling [10]–[14] , we tested whether the Vps26 mutants isolated in our screen also exhibit this phenotype . We generated mutant clones of Vps261 in wing discs of third instar larvae and observed a failure to secrete Wg in the wing margin cells . Furthermore , we observed concomitant loss of Senseless ( Sens ) expression , a downstream effector of the Wingless pathway ( Figure S2C , [83] ) . These data confirm the loss of retromer function in the Vps26 mutants . The subunits of the retromer complex have been shown to primarily associate with endosomes in vertebrates [7]–[9] . However , the in vivo subcellular localization of the retromer in Drosophila has not been documented . We therefore raised a polyclonal antibody against the full-length fly Vps26 protein ( Figure S2D–E ) and costained with antibodies against Avalanche ( Avl , fly homolog of Syntaxin 7 , an early endosomal marker , [84] ) , Rab5 ( an early endosomal marker , [85] ) , Rab7 ( a late endosomal marker , [76] ) , Protein Disulfide Isomerase ( PDI , an ER marker , [86] ) , GM130 ( a cis-Golgi marker , [87] ) , Peanut Agglutinin ( PNA , a marker for the trans-Golgi network ( TGN ) , [88] ) , or Rab11 ( a TGN or post-Golgi secretory vesicle marker , [31] ) . Vps26 shows substantially greater colocalization with Avl , Rab5 , and Rab7 than the ER , Golgi , and secretory vesicle markers ( Figure S2F–G ) . This distribution is seen in both wing discs and adult PRs ( Figure S2F–G and unpublished data ) , indicating that Vps26 associates preferentially with early and late endosomes in different tissues in vivo . Vps35 acts as a scaffolding protein and physically interacts with Vps26 in the retromer ( Figure 1A , [81] , [82] , [89] ) . RNAi knockdown of vps35 has been shown to reduce Vps26 protein levels in HEK293 cells [13] . Consistent with work in cultured cells , we find that mutant clones of the Vps35 null allele ( Vps35MH20 , [13] ) in eye discs cause a nearly complete loss of Vps26 ( Figure 2A ) , showing that the presence of Vps35 is required for Vps26 in vivo . Next , we performed ERG recordings and bright field imaging on large mutant clones of Vps35MH20 in flies aged for 0 , 4 , 10 , and 21 d and examined function and morphology of Vps35MH20 mutant PRs . As shown in Figure 2B–C , the ERG phenotypes of Vps35MH20 are very similar to those of Vps26 shown in Figure 1 . In addition , mutations in the two genes exhibit very similar morphological defects in aged PRs ( compare Figure 2D–E with Figure 1D–E and Figure S1D–E ) . These results indicate that the retromer is required to maintain PR function and morphology . Loss of the retromer therefore leads to light-dependent PR degeneration . Sorting nexins can be involved in retromer-mediated protein trafficking [7]–[9] , [90] , [91] . To test whether these proteins are involved in the maintenance of PRs , we performed ERGs on Snx1d1 and Snx3d1 null alleles ( Figure S3 , [10] ) . Upon keeping the flies for 10 d in LD , we did not observe a difference in ERG amplitudes of Snx1d1 or Snx3d1 mutants ( Figure S3A–B ) , although we observed a partial loss of the on-transient in Snx3d1 ( Figure S3A , C ) . The data indicate that these sorting nexins are not part of the complex or play a minor role . In addition , we tested whether the WASH ( Wiskott–Aldrich Syndrome protein and SCAR Homolog ) complex , a protein complex that has been shown to be required for the transport of several retromer cargos in vertebrate cells and the fly tracheal tubes [8] , [90] , [91] , is involved in PR degeneration . As shown in Figure S4 , RNAi-mediated loss of a subunit of the WASH complex , Washout ( the single WASH ortholog in Drosophila , [92] ) , does not affect PR maintenance and function . Hence , the retromer can have different compositions in different cells . The retromer retrieves membrane protein cargos from the endolysosomal pathway [7]–[9] . To assess whether endosomes and lysosomes are affected in the absence of the retromer , we performed transmission electron microscopy ( TEM ) and examined the ultrastructure of Vps26 and Vps35 mutant PRs . Vps261 and Vps35MH20 flies raised in LD for 4 d show obvious ultrastructural defects ( Figure 3B , D ) in PRs when compared to control ( Figure 3A ) . These defects include a dramatic increase in late endosomes ( inset 1 in Figure 3B , D , Figure 3F , and yellow arrowheads in Figure S5B , D; [93] ) and lysosomes ( inset 2 in Figure 3B , D , Figure 3F , and arrows in Figure S5B–C , E; [94] ) . However , PRs of the dark-reared Vps261 or Vps35MH20 flies ( Figure 3C , E–F and Figure S6B–E ) do not show an increased number of late endosomes or lysosomes when compared to control ( Figure 3A , F , Figure S5A , and Figure S6A ) . These data show that the endolysosomal pathway is severely affected in retromer mutants upon light exposure . To address whether the trafficking of endocytosed Rh1 is affected in the retromer mutants , we assessed Rh1 distribution in PRs that were exposed to light . Two protocols have been used for Rh1 immunostaining in fly PRs: whole mount preparations [76] and eye sections [41] . Previously , we showed that the whole mount protocol is more sensitive than staining of sections for Rh1 present in the cell body , while sections are better than whole mount staining for assessing Rh1 in rhabdomeres [6] . Here , we slightly modified the whole mount protocol to increase Rh1 signals in the PR cell body by treating the tissues with higher concentrations of Triton X-100 . As shown in Figure S7 , Rh1 signals are obviously enhanced upon increasing Triton X-100 concentration ( top panels ) compared to low Triton X-100 treatment ( bottom panels ) . Here , we use the high Triton X-100 protocol for our immunostaining assays . As previously reported [6] , [76] , immunostaining of Rh1 in dark-reared wild-type control PRs reveals a typical crescent-shaped staining pattern at the base of the rhabdomeres , due to the densely stacked rhabdomeric membranes that do not permit antibody penetration in whole mount preparations ( Figure 4A , top panels ) . Similar to this control , we observed that the distribution ( Figure 4A , top panels ) and levels ( Figure 5A–B ) of Rh1 are not affected in Vps261 mutants prior to light exposure . These data are in agreement with the observation that dark-reared Vps261 mutants exhibit nearly normal ERGs and PR morphology ( Figure 1B–E ) . Upon a 12-h light exposure , Rh1 is internalized in both control and Vps261 mutant PRs , as shown by the dramatic increase in Rh1 levels in the PR cell bodies ( Figure 4A , middle panels ) . This increase in Rh1 is not due to the changes in levels of Rh1 ( unpublished data ) , but rather because Rh1 becomes increasingly accessible to the antibodies upon endocytosis [6] . In addition , this light exposure is not pathological as control flies exhibit normal ERGs and PR morphology ( Figure S8 ) . However , upon a 6-h recovery in the dark , the Rh1 signal in the cell body is restored to the original levels in control but not in Vps261 PRs ( Figure 4A , bottom panels ) . Similarly , although less obvious , we also observed Rh1 accumulation in Vps261 mutant PRs upon light exposure in the sectioned eye samples ( Figure S9 ) . Hence , the Rh1 subcellular localization is altered in Vps261 mutants . To assess where in the cell body Rh1 accumulates in Vps261 mutants , we transiently exposed the flies to light and monitored the subcellular localization of Rh1 using markers . We briefly exposed the flies to blue light for 20 min followed by 5-min exposure to orange light ( Figure 4B ) to acutely induce endocytosis of a small amount of Rh1 . This light condition is not pathological as the ERG responses and morphology of control PRs are normal ( Figure S8 ) . Upon recovery in the dark for 4 h , Rh1 can be detected in very bright punctae that often colocalize with Rab7 or Lamp1::GFP , a marker for lysosomes [95] , in Vps261 PRs ( Figure 4B–C ) , implicating that Rh1 is trapped in the endolysosomal compartments . Rh1 retention in the endolysosomal pathway in Vps261 mutants upon an acute light exposure may be due to reduced Rh1 recycling and/or decreased lysosomal degradation . If Rh1 recycling is decreased , one would expect a reduction in Rh1 levels in Vps261 mutants compared to controls upon light exposure as an increased amount of internalized Rh1 will be delivered to lysosomes for degradation . On the other hand , if Rh1 degradation in lysosomes is impaired , one would predict that Rh1 abundance is elevated in Vps26 mutant PRs when compared to controls exposed to light . To distinguish between these two possibilities , we exposed the dark-raised flies to white light ( 1 , 800 lux ) for 16 h . Prior to this light exposure , the Rh1 levels are similar in control and Vps261 mutants ( Figure 5A–B ) , indicating that Rh1 expression is not affected in Vps261 mutants . Upon a 16-h exposure to white light , Rh1 levels in Vps261 mutant PRs are significantly decreased when compared to control ( Figure 5A–B ) . Reduced Rh1 levels in Vps261 mutants upon light exposure do not appear to be due to impaired Rh1 synthesis as the Rh1 levels are restored after 6-h recovery in the dark ( Figure 5A–B ) . Moreover , the ERG responses of Vps261 mutants are not affected upon this light exposure ( Figure 5C–D ) . The data indicate that , in the absence of Vps26 , lysosomes are able to degrade internalized Rh1 to maintain PR function when briefly exposed to light . However , chronic light exposure ( LD for several days ) severely affects the endolysosomal system ( Figure 3 and Figure S5 ) because accumulated Rh1 in endosomes and/or lysosomes persistently challenges the endolysosomal system [2] . To examine whether the retromer is recruited to Rh1 punctae upon endocytosis in the PRs , we performed colocalization studies of Vps26 and Rh1 . In the dark , Rh1 and Vps26 signals do not overlap ( Figure 5E ) . In contrast , numerous Rh1 punctae colocalize with Vps26 in the PR cell body upon transient light exposure ( Figure 5E ) . These data indicate that endocytosed Rh1 is a cargo of the retromer . If PR degeneration in retromer mutants is caused by aberrant accumulation of Rh1 in the endolysosomal pathway , reducing the amount of endocytosed Rh1 or the abundance of Rh1 itself may suppress PR degenerative phenotypes . Dynamin ( encoded by the shi gene ) is required for Rh1 endocytosis upon light exposure [73] . shits1 , a temperature-sensitive dynamin allele , encodes a mutant protein that is functional at 18°C but loses its function at elevated temperature [96] . Although impaired dynamin function leads to death of adult animals at 26°C or higher temperature ( unpublished data ) , shits1 mutant flies are viable and do not exhibit ERG defects at 24°C ( Figure 6A–B ) . This permits us to conditionally reduce dynamin function in adult flies without impairing their light response . To address whether internalization of Rh1 is critical for the PR degeneration in Vps261 flies , we raised the Vps261 , shits1 double mutants at 18°C and shifted the newly eclosed flies to 24°C to decrease dynamin function . We then performed ERGs on Vps261 , shits1 double mutants kept in LD ( Figure 6A ) . As shown in Figure 6A–B , reducing dynamin function significantly suppresses the loss in ERG amplitude in Vps261 mutants at day 4 in LD , indicating that Rh1 endocytosis plays an important role in PR degeneration upon loss of retromer function . To address if a reduction in Rh1 suppresses the demise of the PRs , we tested flies that lack one copy of the Rh1 gene . Strong hypomorphic or null Rh1 alleles impair rhabdomere development in homozygous animals , as Rh1 plays a key signaling role in maintaining the structure of rhabdomeres [97] , [98] . However , flies that are heterozygous for Rh1 loss-of-function alleles do not show aberrant PR development even when the Rh1 protein levels are reduced by 50% ( Figure S10 and [50] ) . To assess whether a reduction of 50% in Rh1 levels is able to suppress the Vps26-associated PR degenerative phenotypes , we tested ERGs in mutants that carry a single copy of the Rh1 null allele ( ninaE17 , [99] ) . As shown in Figure 6C–D , flies that carry a single copy of Rh1 exhibit a significant suppression of the ERG defects seen in Vps261 mutant flies that have been kept in LD for 4 d . To strongly reduce Rh1 levels , we next raised Vps261 mutants on low vitamin A containing fly food . Vitamin A is a dietary precursor of 11-cis 3-hydroxyretinal , the chromophore that is covalently linked to the Rh1 opsin [100] . A decrease in vitamin A leads to loss of the chromophore . Because the chromophore is essential for Rh1 folding and stability [39] , [101] , absence of vitamin A leads to reduced Rh1 levels in flies . Although this manipulation still allows the formation of rhabdomeres , the size is dramatically reduced due to the requirement of Rh1 as a major structural component [102] . Indeed , a reduction in dietary vitamin A leads to diminished rhabdomere size in both control and Vps261 mutants . Remarkably , vitamin A deprivation leads to a very strong suppression of the PR degeneration observed in Vps261 mutants at day 10 in LD ( Figure 6E–F ) . In addition , the reduced ERG amplitudes of Vps261 are strongly suppressed upon vitamin A deprivation ( Figure 6G ) . In summary , decreasing Rh1 levels suppresses the retinal degeneration associated with Vps261 . This also suggests that elevated levels of Rh1 trapped in the endolysosomal pathway , rather than alterations in the levels of Rh1 in the rhabdomeres per se , underlie PR degeneration in Vps261 mutants . Rh1 accumulation in late endosomes causes PR degeneration in flies that lack PLC activity ( encoded by the no receptor potential A ( norpA ) gene , [76] , [103] ) ( Figure 7A ) . If the retromer retrieves Rh1 from endosomes , increasing retromer levels might alleviate PR degeneration in norpA flies . We therefore overexpressed the fly Vps35 ( actin-GAL>UAS-Vps35 , UAS-w RNAi ) or Vps26 ( actin-GAL>UAS-Vps26 , UAS-w RNAi ) in norpAP24 mutants . Upon continuous exposure to white light for 6 d , norpAP24 mutants exhibit reduced numbers of rhabdomeres compared to control flies ( Figure 7B , left and middle left panels and Figure 7C ) . In addition , the ommatidia organization is often aberrant in norpAP24 mutants ( Figure 7B , middle left panel ) , indicating that the PRs are undergoing a progressive demise . Expression of Vps35 or Vps26 significantly restores rhabdomere numbers ( Figure 7B , middle right and right panel and Figure 7C ) , suggesting that this manipulation is sufficient to suppress PR degeneration caused by Rh1 accumulation . In addition , expression of human Vps26B also suppresses the norpAP24 phenotypes ( unpublished data ) . These data indicate that rerouting some of the Rh1 slated for lysosomal degradation delays PR degeneration . Mutations that impair lysosomal function can also lead to Rh1-induced PR degeneration [76] . We therefore assessed whether increasing Vps35 or Vps26 expression suppresses PR degeneration in mutants that affect a subunit of the AP-3 complex . AP-μ3 , encoded by carmine ( cm ) , is required for the proper biogenesis of lysosomal-related organelles [104] . Upon 6 d of light exposure , cm1 mutant PRs showed degenerative features [76] , including loss of rhabdomeres and disruption of the ommatidia organization ( Figure 7D , bottom left panel and Figure 7E ) . In addition , cm1 mutant PRs exhibit decreased ERG amplitudes ( Figure 7F–G ) . Similar to norpAP24 mutants , increasing Vps35 or Vps26 in cm1 mutants restores the number of rhabdomeres ( Figure 7D–E ) and suppresses ERG defects ( Figure 7F–G ) . These data indicate that increased retromer function can not only prevent PR degeneration in mutants with impaired late endosomal trafficking but also suppress PR degeneration caused by defective lysosomal function . Because the genes that encode the retromer proteins are evolutionarily conserved ( Figure 1A , [7]–[9] ) , we investigated whether the vertebrate Vps26 and Vps35 may also function in the visual system . As shown in Figure S11A , vps26A is expressed in the mouse retina based on Northern blots . In addition , we stained the retina of wild-type control or heterozygous mice containing a lacZ inserted in the vps35 gene [105] . Most of the melanopsin-expressing ipRGCs ( 92% ) also express β-GAL ( Figure S11B ) . To test whether human Vps26 homologs can substitute for the function of fly Vps26 in PRs , we ubiquitously expressed the human vps26A or vps26B cDNA in the Vps261 mutant background . Both human Vps26A and Vps26B were able to rescue the lethality of Vps261 mutant flies ( Figure 8A–D ) . We also recorded ERGs from adult flies kept in LD for 21 d . As shown in Figure 8A–B , we observed a full rescue of the loss of ERG amplitude and on–off transients . In addition , PR degeneration in Vps261 mutants is fully suppressed by overexpressing human Vps26 proteins ( Figure 8C–D ) . These results show that the human Vps26 homologs are able to perform the same function as the fly Vps26 protein .
The retromer recycles specific proteins from endosomes to the TGN or plasma membrane [7]–[9] . Here we show that , although the loss of the retromer does not obviously affect eye development , mutations in Vps26 or Vps35 genes lead to strong light-dependent PR degeneration ( Figure 1B–E , Figure 2B–E , and Figure S1C–E ) . The demise of Vps261 and Vps35MH20 PRs is associated with a significant increase in the number of late endosomes and lysosomes upon light exposure ( Figure 3 and Figure S5 ) , showing that the endolysosomal pathway is strongly affected when Rh1 recycling by the retromer is impaired . Indeed , Rh1 accumulates in late endosomes or lysosomes in the mutant PRs ( Figure 4B–C ) . Although Rh1 can be degraded and the function of Vps261 mutant PRs is not abolished upon short light exposure ( Figure 5A–D ) , chronic exposure to light is detrimental to Vps261 mutants because persistent Rh1 accumulation in the endolysosomal pathway is toxic to PR cells [71] , [75] , [76] . Hence , reducing Rh1 endocytosis or Rh1 levels in the rhabdomeres suppresses PR degeneration upon prolonged light exposure ( Figure 6 ) . Interestingly , increasing Vps35 or Vps26 in mutants that show PR degeneration due to Rh1 accumulation in the endolysosomal compartments [76] suppresses the degenerative phenotypes ( Figure 7 ) . In summary , the retromer is required to retrieve Rh1 from endosomes to maintain PR function and integrity ( Figure 8E ) . How does Rh1 internalization affect the endolysosomal pathway in retromer mutants ? One possibility is that lysosomes in the mutants are unable to cope with the increased levels of internalized Rh1 over time as Rh1 is one of the most abundant proteins in PRs [106] , [107] . This in turn triggers an increase in the number of lysosomes ( Figures 3B , D , F and Figure S5B–C , E ) , the accumulation of aberrant lysosomes ( Figure S5C ) , and the accumulation of endolysosomal intermediates , including late endosomes ( Figure 3B , D , F and Figure S5B , D ) . Alternatively , loss of the retromer may increase the flux in the endolysosomal pathway , which overpowers the rate of endolysosomal maturation and leads to an adaptive response that eventually leads to the expansion of these compartments . Both pathways can lead to an apparent accumulation of Rh1 in the cell body when stained and analyzed by fluorescence microscopy ( Figure 4 and Figure S9 ) . Defective regulation of Rh1 can lead to the demise of PRs via apoptosis [72] , [108] . Does apoptosis play a critical role in the PR degeneration in retromer mutants ? We argue that this is not the case based on the following observations . First , the retromer mutants exhibit progressive PR degeneration over a 3-wk period ( Figure 1B–C , Figure 2B–C , and Figure S1C–E ) , whereas apoptosis typically occurs within hours [109] . Second , mutants that lead to PR loss via apoptosis lose most PRs within ommatidia by the engulfment of surrounding glial cells [110] . However , the degenerating Vps26 and Vps35 mutant PRs are not removed , although their morphology is very severely disrupted . Indeed , they can still be identified in 3-wk-old flies ( Figure 1D , Figure 2D , Figure 8C , and Figure S1D ) , indicating a lack of engulfment by surrounding cells [111] . Third , overexpressing p35 , a pan-caspase inhibitor of apoptosis [108] , fails to suppress the PR degeneration in Vps26 mutants upon light exposure ( unpublished data ) . Indeed , Rh1 accumulation in different subcellular compartments triggers different cellular responses , which leads to PR degeneration of varying severity [2] . Accumulation of Rh1 in the endolysosomal pathway seems particularly toxic to PRs but often does not cause the removal of PRs for many weeks . Loss of Wg affects eye development [112] , whereas loss of Crumbs leads to short and/or fused rhabdomeres [113] , [114] . Although the retromer can recycle Wls or Crumbs in some fly tissues [10]–[14] , [17] , [18] , Vps261 or Vps35MH20 mutants do not exhibit obvious eye developmental defects . It is therefore very likely that the composition of the eye retromer is different from the wing retromer . Indeed , loss of Snx3 does not cause obvious degenerative phenotypes when compared to loss of Vps26 or Vps35 ( compare Figure S3 with Figures 1 and 2 ) , yet these proteins are all essential for the recycling of Wls and wing development [10]–[14] . Many players required for phototransduction in Drosophila are conserved in a phototransduction cascade in vertebrate ipRGCs . These include melanopsin , PLC , and the TRP channels [60]–[63] . This phototransduction pathway plays a role in photoentrainment of circadian rhythms [64]–[67] and the control of the pupillary light reflex [64] , [68] , [69] . We find that vps35 is expressed in 92% of the melanopsin-expressing RGCs in the mouse retina ( Figure S11B ) . This may be an underestimate due to technical difficulties with β-GAL immunostaining in mouse tissues . In addition , the human Vps26 proteins are able to substitute for the function of the fly homolog in PRs ( Figure 8A–D ) . Because vertebrate melanopsin has very similar photochemical properties to fly rhodopsin [115] , [116] , the retromer may play a conserved role in vertebrate ipRGCs . As deletion of vps35 leads to early embryonic lethality [105] , a vps35 conditional KO in ipRGCs will need to be established to address its role in ipRGCs . The retromer has been implicated in human neurodegenerative disease , including Alzheimer's disease ( AD ) and PD [19] , [20] , [117] . In AD , a retromer deficiency has been proposed to affect subcellular distribution of β-secretase , which leads to increased amyloid-beta ( Aβ ) deposits and defective neuronal function [105] , [118] . In PD , a missense mutation in VPS35 ( D620N ) has been shown to cause an autosomal dominant late onset form of the disease [19] , [20] . The Vps35 D620N mutant protein appears to function as a dominant negative , and Vps35 and LRRK2 ( Leucine-Rich Repeat Kinase 2 ) have been shown to interact [119] . It will therefore be interesting to assess if these mutants affect ipRGCs as PD patients often have sleep issues [120] .
Vps261 and Vps262 alleles were isolated as described [6] , [78] . Male large duplications ( ∼1–2 Mb ) covering the X chromosome [121] were crossed with female y , w , mut* , P{neoFRT}19Aisogenized flies that were balanced with FM7c , Kr-GAL4 , UAS-GFP ( Kr>GFP ) . For the XE52 group , the lethality of alleles was rescued by Dp ( 1;Y ) y267g19 . 2 ( 1A1;2B17-18 + 20A3;20Fh ) . XE52 alleles complemented all the available deficiencies covered by Dp ( 1;Y ) y267g19 . 2 [122] , [123] . Therefore , we turned to cross the XE52 alleles with 80 kb P[acman] duplications that cover the gaps among deficiencies [124] and found three overlapping duplications ( Dp ( 1;3 ) DC446 , Dp ( 1;3 ) DC033 , and Dp ( 1;3 ) DC034 ) that rescued the lethality of the mutants . These P[acman] duplications share a ∼21 kb interval . We then performed PCR sequencing for genes localized to this region and identified mutations in Vps26 . A lethal insertion P{EP}Vps26G2008 in Vps26 failed to complement both XE52 alleles . Rh117 [84] , shits1 [81] , y , w , GMR-hid , cl , P{neoFRT}19A [125] , norpAP24 [65] , cm1 [89] , w1118 [126] , UAS-w RNAi [127] , and GMR-w-RNAi13D [128] were obtained from the Bloomington Drosophila Stock Center ( BDSC ) . Vps35MH20 is a gift from John-Paul Vincent [13]; Snx1d1 and Snxd3 are gifts from Xinhua Lin [10]; UAS-Lamp1::GFP is a gift from Helmut Krämer [95]; and the UAS-GFP RNAi is a gift from Norbert Perrimon [129] . UAS-RNAi stocks against the wash or wls genes were obtained from the Vienna Drosophila RNAi Center ( VDRC ) . shits1 was used to generate the y , w , Vps261 , shits1 , P{neoFRT}19A recombinant , whereas the GMR-w-RNAi13D was used to generate the y , w , GMR-w-RNAi13D , P{neoFRT}19A and y , w , Vps261 , GMR-w-RNAi13D , P{neoFRT}19A recombinants . For ERG , bright field , and TEM experiments , y , w , Vps261 , P{neoFRT}19A/FM7c , Kr>GFP , or y , w , Vps262 , P{neoFRT}19A/FM7c , Kr>GFP females were crossed with cl ( 1 ) , P{neoFRT}19A/Dp ( 1;Y ) , y+ , v+ [130]; ey-FLP males , whereas w ; Vps35MH20 , P{neoFRT}42D/CyO females were crossed with y , w , ey-FLP , GMR-lacZ ; w+ , cl , P{neoFRT}42D males . w; Snx1d1 , P{neoFRT}FRT40A and w; Snx3d1 , P{neoFRT}FRT82B mutants were crossed with y , w , ey-FLP , GMR-lacZ; w+ , cl , P{neoFRT}40A and y , w , ey-FLP , GMR-lacZ; w+ , cl , P{neoFRT}82B , respectively . White patches indicate mutant clones . Flies in which at least 90% of the eyes are mutant were examined . To produce clones for controls , homozygous iso were crossed with cl ( 1 ) , P{neoFRT}19A/Dp ( 1;Y ) y+ , v+; ey-FLP; y , w; P{neoFRT}42D were crossed with y , w , ey-FLP , GMR-lacZ; w+ , cl , P{neoFRT}42D; and y , w; P{neoFRT}82B were crossed with y , w , ey-FLP; w+ , cl , P{neoFRT}82B . All the flies were raised in constant darkness before experiments . For immunostaining in wing imaginal discs of third instar larvae , y , w , Vps261 , P{neoFRT}19A/Dp ( 1;Y ) 901 males were crossed with ubi-GFP , cl , FRT19A ; hh-GAL4 , UAS-FLP [13] females to generate large mutant clones in the posterior region . In eye imaginal discs of third instar larvae , w; Vps35MH20 , P{neoFRT}42D/CyO , Kr>GFP [13] females were crossed with y , w , ey-FLP , GMR-lacZ; ubi-GFP , cl ( 2 ) , P{neoFRT}42D . For immunostaining , we crossed iso or y , w , Vps261 , P{neoFRT}19A/Dp ( 1;Y ) 901 males with y , w , GMR-hid , cl , P{neoFRT}19A; ey-FLP females to remove w+ wild-type cells in the eyes . To generate Vps261 clones that express Lamp1::GFP , we crossed the y , w , Vps261 , GMR-w-RNAi13D , P{neoFRT}19A; actin-GAL4>UAS-Lamp1::GFP females with cl ( 1 ) , P{neoFRT}19A/Dp ( 1;Y ) , y+ , v+ [130]; ey-FLP males to remove eye pigments that are generated by the mini-w+ markers in the actin-GAL4 and UAS-Lamp1::GFP transgenic constructs . Crosses were kept in the dark to prevent flies from being exposed to light before experiments . The newly eclosed flies ( 1 d after eclosion ) were used for all the experiments . Following are the genotypes used in our analyses: controls — y , w , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/+ ( = iso in Figures 1B–E , 3A , 6 , 8A–D , S1B , S5A , S6A , S8 , and S10 ) ; y , w , P{neoFRT}19A/y , GMR-hid , cl , P{neoFRT}19A; ey-FLP/+ ( = iso in Figure 4A , Figure 4B , top left panel , Figure 4C , top panel , Figure 5A–B , E , Figure S7 , and Figure S9 ) ; y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}42D , Vps35MH20/P{neoFRT}42D , ubi-GFP , cl ( GFP positive tissues in Figure 2A ) ; y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}42D/P{neoFRT}42D , w+ , cl ( = FRT42D in Figure 2B–E ) ; y , w , GMR-w-RNAi13D , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/actin-GAL4>UAS-Lamp1::GFP ( = iso in Figure 4B , bottom left panel and Figure 4C , bottom panel ) ; w1118 ( Figure 7B–C ) ; w; actin-GAL4/UAS-w RNAi ( Figure 7D–G ) ; y , w , Vps261 , P{neoFRT}19A/ubi-GFP , cl , P{neoFRT}19A; hh-GAL4>UAS-FLP/+ ( GFP positive tissues in Figure S2C–D ) ; y , w , P{neoFRT}19A ( = iso in Figure S2F–G ) ; y , w , ey-FLP , GMR-lacZ/y , w; P{neoFRT}82B/P{neoFRT}82B , w+ , cl ( = FRT82B in Figure S3A–B ) ; GMR-GAL4>UAS-w RNAi/UAS-GFP RNAi ( = GMR>GFP RNAi in Figure S4A–C ) ; Rh1-GAL4/UAS-GFP RNAi ( = Rh1>GFP RNAi in Figure S4D–E ) . Generation of Vps26 mutant clones — y , w , Vps261 , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/+ ( = Vps261 in Figures 1B–E , 3B–C , 6 , 8A–D , S1B , S5B–C , S6B–C , and S10 ) ; y , w , Vps262 , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/+ ( = Vps262 in Figures 1C and S1B–E ) ; y , w , Vps261 , P{neoFRT}19A/y , GMR-hid , cl , P{neoFRT}19A; ey-FLP/+ ( = Vps261 in Figure 4A , Figure 4B , top right panel , Figure 4C , top panel , Figure 5A–D , and Figure S9 ) ; y , w , Vps261 , GMR-w-RNAi13D;P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/actin-GAL4>UAS-Lamp1::GFP ( = Vps261 in Figure 4B , bottom right panel and Figure 4D , bottom panel ) ; y , w , Vps261 , shits1 , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP/+ ( = Vps261 , shits1 in Figure 6A–B ) ; y , w , Vps261 , P{neoFRT}19A/cl ( 1 ) , P{neoFRT}19A; ey-FLP; Rh117/+ ( = Vps261; Rh117/+ in Figure 6C–D and Figure S10 ) ; y , w , Vps261 , P{neoFRT}19A/ubi-GFP , cl , P{neoFRT}19A; hh-GAL4>UAS-FLP/+ ( GFP negative tissues in Figure S2C–D ) . Generation of Vps35 mutant clones — y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}42D , Vps35MH20/P{neoFRT}42D , ubi-GFP , cl ( GFP negative tissues in Figure 2A ) ; y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}42D , Vps35MH20/P{neoFRT}42D , w+ , cl ( = Vps35MH20 in Figures 2B–E , 3D–E , S5D–E , and S6D–E ) . Generation Snx1 or Snx3 mutant clones in the eyes — y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}40A , Snx1d1/P{neoFRT}40A , w+ , cl ( = Snx1d1 in Figure S3 ) ; y , w , ey-FLP , GMR-lacZ/w; P{neoFRT}82B , Snx3d1/P{neoFRT}82B , w+ , cl ( = Snx3d1 in Figure S3 ) . Rescued Vps26 mutants — y , w , Vps261 , P{neoFRT}19A/Y; Vps26-gEGFP/actin-GAL4>UAS-w RNAi ( = Vps261 GR in Figure 1B–E ) ; y , w , Vps262 , P{neoFRT}19A/Y; Vps26-gEGFP/actin-GAL4>UAS-w RNAi ( = Vps262 GR in Figure 1C and Figure S1C–E ) ; y , w , Vps261 , P{neoFRT}19A/Y; actin-GAL4>UAS-w RNAi/UAS-hvps26A ( = Vps261 hvps26A OE in Figure 8A–D ) ; y , w , Vps261 , P{neoFRT}19A/Y; actin-GAL4>UAS-w RNAi/UAS-hvps26B ( = Vps261 hvps26B OE in Figure 8A–D ) ; y , w , Vps261 , P{neoFRT}19A/Y; Vps26-gEGFP/+ ( Figure S2B ) ; y , w , Vps262 , P{neoFRT}19A/Y; Vps26-gEGFP/+ ( Figure S2B ) ; y , w , Vps261 , P{neoFRT}19A/Y; tub-GAL4>UAS-Vps26 ( Figure S2B ) ; y , w , Vps262 , P{neoFRT}19A/Y; tub-GAL4>UAS-Vps26 ( Figure S2B ) . Overexpression of LacZ , Vps35 , or Vps26 — w , norpAP24/Y; actin-GAL4>UAS-w RNAi/UAS-Vps35 ( = norpAP24 Vps35 OE in Figure 7B–C ) ; w , norpAP24/Y; actin-GAL4>UAS-w RNAi/UAS-Vps26 ( = norpAP24 Vps26 OE in Figure 7B–C ) ; w; actin-GAL4>UAS-w RNAi/UAS-lacZ ( = control lacZ OE in Figure 7D–G ) ; w; actin-GAL4>UAS-w RNAi/UAS-Vps35 ( = control Vps35 OE in Figure 7D–G ) ; w; actin-GAL4>UAS-wRNAi/UAS-Vps26 ( = control Vps26 OE in Figure 7D–G ) ; cm1/Y; actin-GAL4>UAS-w RNAi/UAS-lacZ ( = cm1 lacZ OE in Figure 7D–G ) ; cm1/Y; actin-GAL4>UAS-w RNAi/UAS-Vps35 ( = cm1 Vps35 OE in Figure 7D–G ) ; cm1/Y; actin-GAL4>UAS-w RNAi/UAS-Vps26 ( = cm1 Vps26 OE in Figure 7D–G ) . RNAi knockdown of wash — w; GMR-GAL4>UAS-w RNAi/UAS-wash RNAi ( = GMR>wash RNAi in Figure S4A–C ) ; w; Rh1-GAL4/UAS-wash RNAi ( = Rh1>wash RNAi in Figure S4D–E ) . Miscellaneous — w , shits1 ( = shits1 in Figure 6A–B ) ; w; Rh117/+ ( = Rh117/+ in Figure 6C–D and Figure S10 ) ; w , norpAP24 ( = norpAP24 in Figure 7B–C ) ; w , P{EP}Vps26G2008/Y ( = P{EP}Vps26G2008 in Figure S2B ) . For whole mount staining , we dissected fly tissue samples in ice cold PBS and fixed them with 4% paraformaldehyde at room temperature for 30 min as described [6] . For tissue sections , we fixed fly tissues with 4% paraformaldehyde , dehydrated the tissues with acetone , and embedded them in LR white resin ( Electron Microscopy Sciences ) . Images were captured with a Zeiss LSM 710 confocal microscope . Antibodies were used at the following concentrations: rabbit anti-Rab5 ( abcam ) , 1∶500; rabbit anti-Avl [84] , 1∶500; rabbit anti-Rab7 [76] , 1∶500; mouse anti-PDI ( abcam ) , 1∶100; rabbit anti-GM130 ( abcam ) , 1∶500; rabbit anti-Rab11 ( abcam , [6] ) , 1∶500; chicken anti-GFP ( abcam ) , 1∶1 , 000; mouse anti-Rh1 [4C5 , Developmental Studies Hybridoma Bank ( DSHB ) ] , 1∶50; mouse anti-Wash ( P3H3 , [131] ) , 1∶20; guinea pig anti-Vps26 , 1∶1 , 000 ( this study ) ; biotin-conjugated PNA ( Vector Labs ) , 1∶1 , 000; Alexa 488-conjugated phalloidin ( Invitrogen ) , 1∶100; and Alexa 405- , Alexa 488- , Cy3- , or Cy5-conjugated secondary antibodies ( Jackson ImmunoResearch ) , 1∶600 . Fly heads were homogenized in 1× Laemmli sample buffer ( Bio-Rad ) containing 2 . 5% β-mercaptoethanol ( Sigma-Aldrich ) . Tissue samples were loaded into 10% gels , separated by SDS-PAGE , and transferred to nitrocellulose membranes ( Bio-Rad ) . Antibodies were as follows: mouse anti-actin ( ICN Biomedicals ) , 1∶2 , 500; mouse anti-Rh1 ( 4C5 , DSHB ) , 1∶1 , 000; and rabbit anti-TRP [58] . Full-length fly Vps26 cDNA was cloned from LD29140 by PCR with primers 5′-CTTGGATCCATGAATTTCCTGGGATTCGGCCA-3′ and 5′-CTGGCGGCCGCTCAATCGGTGGCCAACGGC-3′ . PCR products were then digested with BamHI ( NEB ) and NotI ( NEB ) and ligated with a pET28a ( + ) vector containing a Hisx6 tag at the N′-terminus ( Novagen ) . The Vps26-pET28a ( + ) construct was transformed into BL21 ( DE3 ) pLysS bacteria ( Invitrogen ) . Expression of Hisx6-Vps26 was induced with 0 . 5 mM isopropylthio-β-galactoside ( IPTG , Sigma-Aldrich ) , and the recombinant protein was purified with a Ni-NTA column ( Qiagen ) and inoculated into guinea pigs to generate polyclonal antibodies ( Cocalico Biological Inc . ) . In general , mosaic flies were grown in the dark at 25°C , and the newly eclosed flies were selected and shifted to a 12-h light/dark cycle ( LD , 1 , 800 lux in the light cycle ) or constant darkness ( DD ) at 25°C for the indicated periods . In the shits suppressor assay , animals were raised in the dark at 18°C to avoid reduced dynamin protein function . Newly eclosed flies were then shifted to 24°C in the dark for 12 h . Flies were then kept in LD or DD at 24°C . For aging experiments , flies were flipped to new vials every 3 d to maintain food conditions , and the positions of the vials were randomly shuffled to normalize light exposure . For pulse-chase assays , flies were kept in a box with a blue LED [6] for 20 min and then an orange LED for 5 min . Low vitamin A–containing food was made according to [134] . In brief , a solution containing 10% dry yeast , 10% sucrose , 0 . 02% cholesterol , and 2% agar was microwaved and poured into fly vials . After drying at room temperature overnight , fresh low vitamin A food was used for experiments . To prepare fly retina samples for TEM and bright field imaging , fly heads were dissected and fixed at 4°C in 4% paraformaldehyde , 2% glutaraldehyde , 0 . 1 M sodium cacodylate , and 0 . 005% CaCl2 ( PH 7 . 2 ) overnight; postfixed in 1% OsO4; dehydrated in ethanol and propylene oxide; and then embedded in Embed-812 resin ( Electron Microscopy Sciences ) with vacuum attachments . PRs were then sectioned and stained in 4% uranyl acetate and 2 . 5% lead nitrate . TEM images of PR sections were taken using a JEOL JEM 1010 transmission electron microscope with an AMT XR-16 mid-mount 16 mega-pixel digital camera , while bright field pictures were captured using the Zeiss Imager . Z2 light microscope with an AxioCamMRm digital camera . The morphology and quantity of rhabdomeres , endosomes , and lysosomes were determined in Image J . ERG recordings were performed as described [6] . Briefly , adult flies were glued to a glass slide , a recording probe was placed on the surface of the eye , and a reference probe was inserted in the thorax . A 1-s flash of white light was given , and the response was recorded and analyzed using the AXON™-pCLAMP®8 software . Mouse retina cDNA was prepared as described [135] . vps26A was PCR-amplified using primers 5′-GAGTTTTCTTGGAGGCTTTTTTGGTCC-3′ and 5′-TTACATCTCAGGCTGCTCCGCAGAGG-3′ and a 30 ng cDNA template . The PCR product was cloned into pBluescript by blunt-end ligation and verified by sequencing . DNA probes were generated from gel-purified insert ( excised with EcoRI and HindIII ) by random-prime labeling with [α-32P]dCTP using the DECAprimeII kit ( Ambion ) . Unincorporated nucleotides were removed with a MicroBio-Spin 30 column ( Bio-Rad ) . For Northern blot analysis , total RNA was prepared from retinas of 3–4-mo-old CD-1 mice by homogenization in TRI reagent ( Ambion ) using a motorized pestle and passage through a 20-gauge needle , extraction with 1-bromo-3-chloro-propane , and purification with the RNeasy kit ( Qiagen ) . We resolved 10 µg of total RNA on a formaldehyde-agarose gel and transferred it to a Hybond-N+ nylon membrane ( GE Healthcare ) by capillary transfer . Membranes were prehybridized in ULTRAhyb ( Ambion ) for 1 h , then incubated at 42°C overnight with a ∼2 . 5×107 cpm denatured probe diluted in ULTRAhyb . Blots were washed several times in 2× SSC+0 . 1% SDS and several times in 0 . 1× SSC+0 . 1% SDS . Washes were performed at 42–45°C . For imaging , blots were exposed to a storage phosphor screen and scanned on a Typhoon ( GE Healthcare ) . For immunohistochemistry , wild-type or vps35 lacZ KI [105] mouse eyes were fixed in 4% paraformaldehyde in PBS for 45 min , washed in PBS , then stored overnight in 30% sucrose in PBS . The cornea and lens were removed , and eyecups were embedded in OCT ( Tissue-Tek ) and flash-frozen . Cryosectioned retina were fixed in 2% paraformaldehyde in 1× PBS for 10 min and stained with the indicated antibodies: chicken anti–β-GAL ( abcam , 1∶1 , 000 ) and rabbit α-melanopsin ( Advanced Targeting Systems , 1∶5 , 000 ) . Alexa 488- or Cy3-conjugated secondary antibodies ( Jackson ImmunoResearch , 1∶600 ) were used for immunostaining . Images were captured using the Zeiss LSM 710 confocal microscope and analyzed using the Image J software . Two-tailed Student's t test was used to analyze experimental results . | Upon light exposure , rhodopsins—light-sensing proteins in the eye—trigger visual transduction signaling to activate fly photoreceptor cells . After activation , rhodopsins can be internalized from the cell surface into endosomes and then degraded in lysosomes . This mechanism prevents constant activation of the visual transduction pathway , thereby maintaining the function and integrity of photoreceptor cells . It is not known , however , whether these internalized rhodopsins can be recycled . Here , we show that the retromer , an evolutionarily conserved protein complex , is required for the recycling of rhodopsins . We find that loss of key retromer subunits ( Vps35 or Vps26 ) causes rhodopsin mislocalization in the photoreceptors and severe light-induced photoreceptor degeneration . Conversely , gain of retromer subunits can alleviate photoreceptor degeneration in some contexts . Human retromer components can stand in for depleted fruit fly retromer , suggesting that this complex plays a role in recycling light sensors in both vertebrate and invertebrate photoreceptors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"neuroscience",
"visual",
"system",
"biology",
"and",
"life",
"sciences",
"sensory",
"systems"
] | 2014 | The Retromer Complex Is Required for Rhodopsin Recycling and Its Loss Leads to Photoreceptor Degeneration |
Alternative 3′ and 5′ splice site ( ss ) events constitute a significant part of all alternative splicing events . These events were also found to be related to several aberrant splicing diseases . However , only few of the characteristics that distinguish these events from alternative cassette exons are known currently . In this study , we compared the characteristics of constitutive exons , alternative cassette exons , and alternative 3′ss and 5′ss exons . The results revealed that alternative 3′ss and 5′ss exons are an intermediate state between constitutive and alternative cassette exons , where the constitutive side resembles constitutive exons , and the alternative side resembles alternative cassette exons . The results also show that alternative 3′ss and 5′ss exons exhibit low levels of symmetry ( frame-preserving ) , similar to constitutive exons , whereas the sequence between the two alternative splice sites shows high symmetry levels , similar to alternative cassette exons . In addition , flanking intronic conservation analysis revealed that exons whose alternative splice sites are at least nine nucleotides apart show a high conservation level , indicating intronic participation in the regulation of their splicing , whereas exons whose alternative splice sites are fewer than nine nucleotides apart show a low conservation level . Further examination of these exons , spanning seven vertebrate species , suggests an evolutionary model in which the alternative state is a derivative of an ancestral constitutive exon , where a mutation inside the exon or along the flanking intron resulted in the creation of a new splice site that competes with the original one , leading to alternative splice site selection . This model was validated experimentally on four exons , showing that they indeed originated from constitutive exons that acquired a new competing splice site during evolution .
The human genome sequencing project has led to the understanding that total gene number is not indicative of a higher level of phenotypic complexity , as the number of human genes is ~25 , 000 , only slightly higher than the nematode ( ~19 , 000 genes ) and lower than rice ( ~40 , 000 genes ) [1 , 2] . The mechanism that was proposed to resolve this discrepancy is alternative splicing , in which several mRNA isoforms are generated from a single gene through the alternative selection of 3′ss and/or 5′ss , producing several functional proteins [3 , 4] . There are five major forms of alternative splicing: exon skipping ( also known as cassette exons ) , in which the exon as a whole either is included in the mature mRNA transcript or is skipped . Exon skipping is the most common alternative splicing event and accounts for 38% of conserved alternative splicing events between human and mouse . Alternative 3′ss exons ( A3Es ) and 5′ss exons ( A5Es ) account for ~18% and ~8% of the human–mouse conserved events , respectively . These exons are flanked on one side by a constitutive splice site ( fixated ) and on the other side by two ( or more ) competing alternative splice sites , resulting in an alternative region ( extension ) that either is included in the transcript or is excluded . Intron retention accounts for fewer than 3% of the human–mouse conserved alternative splicing events , whereas the remaining ~33% are different types of complex events [5 , 6] . Four splice signals are essential for accurate splicing: 5′ and 3′ splice sites ( 5′ss and 3′ss ) , the polypyrimidine tract , and the branch site sequence [7] . However , these signals solely can't support proper splice site selection and proper splicing . Cis-acting regulatory elements called exonic splicing enhancers ( ESEs ) and exonic splicing silencers ( ESSs; ESEs + ESSs are also termed ESRs ) were found to be involved in the regulation of the alternative splicing process [8] . These elements have the major effect when located in proximity to the alternative splice site [9–12] . Previous studies focused their research mainly on cassette exons in the quest of identifying the regulatory mechanism governing their splicing and the evolutionary background that led to their creation [13 , 14] . In addition , other studies have tried to examine the biological significance and evolutionary mechanism of intron-retained exons [15 , 16] . On the other hand , A3Es and A5Es are relatively more poorly characterized , even though they were found to be related to several diseases caused by aberrant splicing generated by mutations [17–19] . Thus , it is important to widen the knowledge regarding their regulation , characteristics , and evolutionary origin . Examination of A3Es and A5Es reveals that their alternative region is highly diverse in terms of length . However , ~50% of A3Es have alternative splice sites that are exactly 3 bp apart ( known as the NAGNAG motif ) . This was one of the reasons that led several studies to examine a subset of A3Es and A5Es ( containing a short alternative region , 1–4 bp in length ) , in an attempt to learn whether they are controlled by a highly complex regulatory mechanism and serve as a fine-tuning mechanism for protein functionality [20–22] or are merely a meter of noise of the splicing machinery [23] . A previous study showed that in A3Es , an additional polypyrimidine tract ( PPT ) between the two 3′ss appears only when the distance between them is more than 8 bp , whereas in A5Es , the GT dinucleotide located 4 bp downstream of the authentic GT in the 5′ss sequence ( GTNNGT ) can serve as a competing 5′ss [24] . Other studies have aimed to develop tools that will enable us to identify new A3Es and A5Es based on different criteria [25 , 26] . However , the evolutionary mechanism that led to their creation is largely unknown . Comparative genomics has been used as a powerful tool for the identification of functional biological features . The guideline directing this approach is that a high conservation level implies an essential biological function—structural or regulatory [27] . Human–mouse comparison is the most commonly used in research , because it is estimated that the two species diverged 75–130 million years ago and , hence , had enough time to accumulate mutations on the one hand while still maintaining high levels of homology on the other . In addition , nearly all human genes ( 99% ) have a mouse ortholog with a high ( 88% ) protein coding sequence resemblance [28] . Moreover , their genomes have been explored widely and their transcriptome covered by millions of EST sequences , making it easier to perform large-scale analyses . It has been previously shown that constitutive and alternative cassette exons differ from each other in several features , such as: splice site strength , exon length , conservation level between human and mouse ( exon identity ) , divisibility-by-3 ( symmetry ) , and KA/KS ratio test [6 , 14 , 29] . These discriminating features were also used to explore the regulatory mechanism that is required to ensure the proper splicing of these exons . A3Es and A5Es can produce ( at least ) two splice transcripts: one contains the extension , and the other excludes it . These transcripts can be formed in different ratios , i . e . , one can be more abundant ( major form ) compared with the other ( minor form ) . Thus , the more common form of each exon can allude to its evolutionary background . Thus , we examined each of the aforementioned characteristics in A3Es and A5Es , subdivided into two subgroups according to the common form , and compared them with constitutive and alternative cassette exons to understand the overall regulatory mechanism that directs the proper selection of one alternative splice site over the other . Our analysis was based on a unified dataset of human–mouse orthologous exons comprising constitutive exons , alternative cassette exons , A3Es , and A5Es [6 , 30 , 31] . The analysis was expanded further to multiple species alignment to acquire knowledge on the evolutionary background of A3Es and A5Es from both subgroups . Our results suggest an evolutionary model for the creation of new A3Es and A5Es . According to this model , A3Es and A5Es originated from constitutive exons that acquired a new functional splice site as a result of a mutation in the exon or along the intron . This splice site started to compete with the authentic splice site , leading to alternative splice site selection . We provide computational and experimental evidence supporting this hypothesis .
To identify unique characteristics of A3Es and A5Es , we examined several features that can differentiate these exons . We compiled four datasets of human–mouse orthologous exons , including their flanking introns: 45 , 553 constitutively spliced exons , 757 alternative cassette exons , 530 A3Es , and 232 A5Es ( see Materials and Methods ) . Each of the A3Es and A5Es was represented by both “short” form ( without the alternative extension ) and “full” form ( with the alternative extension ) . We first extracted the 3′ss and 5′ss of all the exons in the datasets and examined their strength ( Figure 1 ) . We found that while alternative cassette exons were shown to have relatively weaker splice sites in comparison with constitutive exons [32–35] , A3Es and A5Es present a strong splice site in the fixated exon's side ( i . e . , 5′ in A3Es and 3′ in A5Es ) , resembling constitutive exons and statistically differing from alternative cassette exons ( Mann-Whitney , p = 4 . 39E-05 and p = 8 . 67E-05 for A3Es and A5Es , respectively ) , and a weak splice site in the alternative exon's side ( i . e . , 3′ in A3Es and 5′ in A5Es ) . The alternative sites of A3Es and A5Es are even weaker than the splice sites flanking alternative cassette exons ( Figure 1; see Tables 1 and 2 for statistical analysis and Figure S1 for mouse results ) . Thus , A3Es and A5Es contain a strong anchor at the constitutive splice site and suboptimal splice sites at the altered sites . Next , we tried to understand why the splicing machinery selects one splice site more frequently than the other , when the two are located on the same exon . We therefore determined which is the major and which is the minor splice site , based on ESTs coverage , requiring at least one transcript representing each of the forms ( see Materials and Methods ) . The average number of cDNAs/ESTs per event is 78 . 27 and 56 . 12 cDNAs/ESTs for A3E and A5E , respectively . In addition , the percentage of events that were represented by a small number of cDNAs/ESTs is 1% and 2 . 3% for A3E and A5E , respectively . It is worth mentioning that we base this method of major/minor determination on the assumption that the cDNA/EST coverage of the splice forms properly represents the cellular situation . We subdivided each of the A3Es and A5Es into two subgroups , according to their extension inclusion level: ( i ) exons whose major splice site is within the exon—extension inclusion level is less than 40% ( group 1 ) ; and ( ii ) exons whose major splice site is at the end of the exon—extension inclusion level is more than 60% ( group 2 ) . We observed that in A3Es ( both group 1 and group 2 ) and in group 2 A5Es , the major splice site was significantly stronger than the minor splice site ( Wilcoxon , p = 0 . 01 , p = 2 . 37E-06 , p = 0 . 007 for group 1 A3Es , group 2 A3Es , and group 2 A5Es , respectively ) . However , in the case of group 1 A5Es , both alternative splice sites had a similar splice site score . To examine this discrepancy further , we examined ESEs' , ESSs' , and ESRs' density in the 15 basepairs of the exonic sequences immediately upstream of the major and minor alternative 5′ss ( Figure 2A ) , because it has been found that ESEs , ESSs , and ESRs in close proximity to the altered site have the major effect on alternative splicing [9–11] . It is noteworthy that even though the average distance between the two alternative splice sites is 63 . 71 ( mean = 22 ) , we cannot absolutely rule out the possibility that one regulatory element affects both alternative splice sites . Interestingly , we revealed that in the case of group 1 A5Es , there is a higher ESE density upstream of the major splice site than upstream of the minor splice site ( Wilcoxon , p = 0 . 02 and p = 0 . 03 for ESRs and ESEs , respectively; Figure 2B and 2C , right panel ) . Conversely , in group 2 A5Es , an opposite trend can be observed , in which there is a higher ESE density upstream of the minor splice site than upstream of the major splice site ( Wilcoxon , p = 0 . 001 and p = 0 . 02 for ESRs and ESEs , respectively ) . This phenomenon was demonstrated using both Fairbrother et al . ESEs and Goren et al . ESRs ( Figure 2B and 2C , respectively , left panel , [10 , 36] ) . As opposed to ESEs , we found a lower ESS density upstream of the major splice site in group 1 exons and a very high density upstream of the minor splice site ( Wilcoxon , p = 1 . 80E-04; Figure 2D , right panel; [37] ) . In group 2 exons , the results were vice versa , that is , there was a lower ESS density upstream of the minor splice site and a very high density upstream of the major splice site ( Wilcoxon , p = 7 . 65E-08; Figure 2D , left panel ) . These results imply that A5Es require rigorous regulation for the proper selection of alternative splice sites , which , in the case of group 1 A5Es , assures the preferential selection of the major splice site despite the similar score of the two alternative sites . On the other hand , in the case of group 2 A5Es ( in which the major splice site is stronger than the minor splice site ) , the regulatory mechanism is required to support minor splice site partial selection ( see Figure S2 for mouse data ) . Thus , when the two alternative splice sites are of similar strength , a delicate interplay between ESE and ESS presumably directs splice site selection between the major and minor sites . Further , we compared several additional features among the four datasets . Among the examined features were: exon length , conservation level between human and mouse ( exon identity ) , divisibility-by-3 ( symmetry ) , and KA/KS ratio test . We found that the A3Es ( both group 1 and group 2 ) show more resemblance to constitutive exons than to alternative cassette exons ( Table 3 ) . For instance , the average length of a constitutive exon is 132 . 46 bp ( median = 121 ) , which was found to be very similar to that of A3Es—136 . 6 ( median = 128 . 5 ) and 147 . 02 ( median = 129 . 5 ) for group 1 and group 2 exons , respectively . Statistically , the A3Es show no significant difference from constitutive exons . However , the difference from alternative cassette exons is significant ( Mann-Whitney , p = 4 . 7E-07 and p = 1 . 04E-20 for group 1 and group 2 exons , respectively ) . Examination of the human–mouse conservation level showed the same trend . In this case , the A3Es are significantly different from constitutive exons ( Mann-Whitney: p = 0 . 02 and p = 0 . 04 for group 1 and group 2 exons , respectively ) . However , the difference from alternative cassette exons is much more significant ( Mann-Whitney , p = 2 . 16E-05 and p = 4 . 56E-20 for group 1 and group 2 exons , respectively ) . The symmetry analysis also showed significant difference from alternative cassette exons ( χ2 , p = 2 . 85E-09 and p = 4 . 69E-14 for group 1 and group 2 exons , respectively ) and resemblance to constitutive exons ( group 1 exons even presented lower levels of similarity ) . The KA/KS ratio test was previously introduced as a tool for the identification of exons within genomic regions [38] . Briefly , the test examines whether the tested exon has evolved under purifying selection . A prospective research showed that this tool is more proficient in identifying constitutive exons rather than alternative cassette exons , which tend to fail the KA/KS ratio test . Thus , it can serve as a discriminative tool between constitutive and alternative cassette exons [39] . As anticipated , constitutive and alternative cassette exons show significantly different KA/KS ratio test failure percentages ( 7 . 59% and 33 . 33% , respectively; χ2 , p = 8 . 51E-83 ) . Regarding A3Es , 12% of group 1 exons and 9 . 92% of group 2 exons failed the test , showing resemblance to constitutive exons and a significant difference from alternative cassette exons ( χ2 , p = 0 . 41/0 . 33 and p = 0 . 03/4 . 67E-07 for group 1/group 2 versus constitutive and alternative cassette exons , respectively ) . This same trend of resemblance to constitutive exons can also be observed in group 1 A5Es ( group 1 exons ) and in the length and symmetry features of group 2 A5Es ( group 2 exons ) . However , in the case of group 2 A5Es , the identity level and KA/KS ratio test features show more significant difference from the constitutive exons than from the alternative cassette exons ( Table 3 ) . The analysis presented above shows that in the case of A3Es ( both group 1 and group 2 ) and group 1 A5Es , all features examined indicate a resemblance to constitutive exons , whereas group 2 A5Es present inconsistent results ( see Discussion ) . We then checked the selected features in the extension sequence separately ( the sequence between the two alternative splice sites ) . We found a high percentage of symmetry ( 71 . 19% and 63 . 81% for A3Es and A5Es , respectively ) , showing resemblance to alternative cassette exons and a significant difference from constitutive exons ( χ2 , p = 0 . 114 and p = 1 . 17E-42 for alternative cassette and constitutive exons , respectively ) . This is consistent with the notion that selective pressure is applied against the formation of an asymmetrical sequence between two alternative splice sites , because an asymmetrical sequence would shift the open reading frame and , hence , could lead to premature translation termination . Thus , the extension sequences show characteristics of cassette exons ( Table 4 ) . Intronic sequences flanking constitutive exons show low conservation levels between human and mouse , whereas intronic sequences flanking alternatively spliced exons show 88% and 80% conservation of 103 and 94 bp on average for the upstream and downstream introns , respectively [40] . Also , tissue-specific cassette exons are flanked by highly conserved intronic sequences , even more conserved than cassette exons [41] . However , intronic sequences flanking the alternative side of A3Es and A5Es are characterized by a high conservation level , similar to alternative cassette exons , while intronic sequences flanking the fixated side are characterized by a relatively low conservation level , similar to constitutive exons [6] . A similar observation was also made on the NAGNAG motif subgroup of A3Es [20] . We thus examined the conservation of the intronic regions flanking A3Es and A5Es using a different method for the sequences alignment . Briefly , we used the local alignment program Sim4 to identify the human–mouse conserved regions . It is noteworthy that the intronic region flanking the alternative side was defined as the sequence downstream of the distal 5′ss ( “most” downstream 5′ss ) and upstream of the proximal 3′ss ( “most” upstream 3′ss ) for A5Es and A3Es , respectively ( see Materials and Methods ) . Figure 3A demonstrates the percentage of upstream and downstream introns that are conserved between human and mouse ( left and right panels , respectively ) . As expected , low conservation is observed in the intronic region flanking the fixated side ( median length of 30 bp and 37 bp for A3Es and A5Es , respectively ) . However , there is a high conservation level in the intronic region flanking the alternative side ( median length of 42 bp and 57 bp for A3Es and A5Es , respectively ) . Consequently , this analysis provides additional evidence for the importance of the intronic portion located immediately adjacent to the alternative site , whereas the fixated side is almost free from that constraint . Presumably , this conserved intronic sequence is involved in the subtle regulation of alternative splice site selection . Another phenomenon that can be observed is that for A5Es , the conservation level in the alternative side is as high as in alternative cassette exons . However , A3Es show lower conservation levels in the alternative side , but higher than constitutive exons . To further examine this observation , we decided to divide the A3Es into two subgroups . It has been previously shown that an additional PPT appears only when the two alternative 3′ss are at least 8 bp apart [24] . Thus , we used this value as a cutoff for the subdivision . Surprisingly , repeating the analysis on the two subgroups showed that A3Es whose alternative splice sites are up to 8 bp apart show a conservation level similar to constitutive exons ( average conservation length of 44 . 01 bp and average conservation level of 82 . 65 ) , whereas A3Es in which the distance between the two 3′ss is longer than nine nucleotides show a conservation level that resembles alternative cassette exons ( average conservation length of 69 . 39 bp and average conservation level of 86 . 16; Figure 3B , left panel; Figure S4A ) . We conducted the same analysis for A5Es using 6 , 8 , 9 , and 20 bp as cutoffs , and the most significant change in the behavior of the two subgroups was observed using 8 or 9 bp as a cutoff ( unpublished data ) . A5Es in which the distance between the two 5′ss is longer than nine nucleotides show a conservation level that resembles alternative cassette exons ( average conservation length of 71 . 18 bp and average conservation level of 84 . 70 ) , whereas the subgroup of A5Es whose alternative splice sites are up to 8 bp apart show a conservation level that is between the constitutive and alternative cassette exons ( Figure 3B , right panel; Figure S4B ) . These results imply that there is a difference in the importance of the intronic regulatory elements according to the distance between the alternative splice sites: exons that present a long distance between the two competing splice sites rely on intronic sequences for proper alternative splice site selection , whereas exons that present a short distance between the two competing splice sites are relatively less dependent on those intronic sequences . These findings provide an additional interpretation to the regulatory implications of short distance alternative splice sites ( see also [20 , 22 , 23] ) . Tables 5 and 6 summarize the entire statistical analysis . We also examined the flanking introns conservation level for group 1 and group 2 exons separately . However , no significant difference was found between these two subgroups ( unpublished data ) , suggesting that selection of the major/minor splice site is not regulated via the intronic sequences . In addition , we observed a “local peak” in the conservation curve of the flanking introns around 75 bp and 60 bp for upstream and downstream introns , respectively , suggesting the preference of regulatory elements in these regions . Such a peak was previously reported in tissue-specific alternative cassette exons [41] . Based on the splice site score analysis , and the flanking intronic conservation , we suggest that A3Es and A5Es are a hybrid of constitutive and alternative cassette exons , where the alternative side of the exon ( i . e . , 3′ in A3Es and 5′ in A5Es ) resembles alternative cassette exons , and the fixated side ( i . e . , 5′ in A3Es and 3′ in A5Es ) resembles constitutive exons . The analysis of the characteristic features conducted on the exon as a whole showing resemblance to constitutive exons with the extension showing resemblance to alternative cassette exons further supports this hypothesis . Alu elements are primate-specific retrotransposones , ~300 bp long , that have been shown to play a significant role in gene expression regulation [42] . Alternative cassette exons are flanked by longer intronic sequences , compared with constitutively spliced exons [5 , 43] . Density analysis of Alu sequences revealed that introns flanking alternative cassette exons demonstrate a significantly higher density of Alu sequences , compared with constitutive exons [44] , which was also supported by our analysis ( 3 . 29/3 . 03 and 2 . 12/2 Alu sequences per intron for upstream/downstream introns of alternative cassette and constitutive exons , respectively . Mann-Whitney , p = 0 . 03/9 . 39E-05 ) . When the density of Alu elements was examined in the flanking introns of A3Es , we found a resemblance to the constitutive exons with 2 . 44/1 . 98 Alu sequences for upstream and downstream introns , respectively ( Mann-Whitney , p = 0 . 40 and p = 0 . 91 for upstream and downstream introns , respectively ) , which differs significantly from alternative cassette exons ( Mann-Whitney , p = 0 . 048/0 . 02 for upstream and downstream introns , respectively ) . In contrast , A5Es showed an Alu density that resembles alternative cassette exons . However , these results were not significant . We conducted a human–mouse KA/KS ratio analysis of the exonic region that is present in both the spliced variants ( the short form , or core ) of each exon , versus the extension sequence . The analysis was performed for group 1 and group 2 exons separately . Based on the KA/KS ratio results , which indicated that the extension sequence behaves differently for group 1 and group 2 exons ( Figures 4 and 5 , panels ( i ) ) , we set to examine the evolutionary constraints affecting A3Es and A5Es . Altogether , we randomly selected and examined 13 cases for which we found the homologue exon and flanking exons sequences in the HomoloGene database among seven organisms ( see below ) . Three of these cases showed putative functional alternative splice sites among all seven organisms , but no support was found for the alternative state for some of the organisms , probably because of a low EST/cDNA coverage of this region . In the remaining ten cases we found a functional alternative splice site only in some of the species . Figures 4 and 5 show four of such cases: one example representing a group 1 exon and one example representing a group 2 exon for both A3Es and A5Es . For each of the selected exons and flanking splice sites , we used the HomoloGene database to extract orthologous genes in seven different organisms [45]: human ( Homo sapiens ) , mouse ( Mus musculus ) , rat ( Rattus ) , opossum ( Didelphis virginiana ) , chicken ( Gallus gallus ) , xenopus ( Xenopus tropicalis ) , and zebrafish ( Danio rerio ) ; and generated a multiple sequence alignment among these seven species using ClustalW [46] . In group 1 A3E ( human PRPF3 gene ) , we found the KA/KS ratio of the major ( short ) form to be 0 . 001 ( KA = 0 . 0002 , KS = 0 . 1677 ) and the KA/KS ratio of the extension to be 99 ( KA = 0 . 0476 , KS = 0 . 0005 . Such a high ratio indicates that synonymous substitutions practically do not occur; Figure 4A , panel ( i ) ) . The multiple alignment of this exon revealed that the major ( short ) form is relatively conserved among the seven species ( 63 . 16% ) , while the extension showed poor conservation ( 13 . 33% ) . In addition , we found that at the minor splice site , a functional AG dinucleotide exists only among the mammals , whereas in chicken , xenopus , and zebrafish a nonfunctional 3′ss is present: either AA or GG dinucleotides in that site ( Figure 4A , panel ( ii ) ) . We also observed that in human , mouse , rat , and chicken , a relatively strong PPT exists at the minor splice site , while in xenopus and zebrafish , no such sequence is present . Based on a known evolutionary tree , the most plausible scenario is that a mutation from A to G occurred , which created a functional AG 3′ss in human , mouse , rat , and opossum and resulted in the emergence of a new A3E . We noticed that no evidence ( EST or cDNA ) for alternative splicing was found for rat or opossum . However , this is likely a consequence of relatively small amounts of deposited mRNA and EST sequences that cover that locus . Next , we validated these results experimentally in six of the seven species ( human , mouse , rat , chicken , xenopus , and zebrafish ) . We designed primers for the flanking exons in each of these organisms . Total RNA was extracted from brain tissues of human , mouse , and rat , from a whole organism of chicken and zebrafish , and from xenopus' oocytes . This was followed by RT–PCR ( reverse transcription–polymerase chain reaction ) analysis and sequencing of all PCR products . The pairs of primers used in the RT–PCR were designed specifically to amplify a similar size of PCR product for the extension–inclusion isoform among the different organisms of the same ortholog . The experimental results validated that the alternative form was unique to human , mouse , and rat , which , as indicated above , have a functional minor splice site ( Figure 4A , panel ( iii ) ; lanes H , M , R , which are abbreviations for human , mouse , and rat , respectively ) . In chicken , xenopus , and zebrafish , which were shown not to have an additional functional splice site , the exon is constitutively spliced ( Figure 4A , panel ( iii ) ; panels C , Z , X , which are abbreviations for chicken , xenopus , and zebrafish , respectively ) . We note , that whenever the alternative 3′ss or 5′ss isoforms were not detected , longer exposure of the gels did not indicate its presence . However , this did not rule out completely that this alternative isoform ( or selection of another cryptic splice site ) might be detected under different conditions . Based on the phylogenetic relationships among the analyzed organisms and these results , we conclude that the alternative splice variant is a derived form , and the constitutive spliced variant is the ancestral one . Moreover , the constitutive form observed for chicken , xenopus , and zebrafish is also the major mRNA product of the human , mouse , and rat alternative forms . This indicates that the major form , which is the ancestral conserved isoform , is also the major mRNA product after the transition from constitutive to alternative splice site selection . Presumably , this is needed to prevent selection against the alternative isoform . Unlike the group 1 A3E , in the group 2 A3E ( human UBQLN4 gene ) , the KA/KS ratio of the extension is similar to that of the minor ( short ) form ( Figure 4B , panel ( i ) ) . The multiple alignment shows that the exon is conserved among the seven species throughout its full length ( 62 . 5% ) and that only human , mouse , and rat present a functional minor splice site ( AG ) , whereas opossum , chicken , and zebrafish contain TG in the reciprocal positions , and xenopus contains CG in the reciprocal positions . The minor splice site is within the exon , which is relatively conserved among the seven species . Thus , we find a relatively strong PPT at the minor splice site in all seven species . The minor 3′ss was presumably created during evolution following a mutation in the first position of the AG 3′ss ( Figure 4B , panel ( ii ) ) . The experimental results verify that only the human , mouse , and rat present two alternatively spliced forms , which is compatible with the bioinformatic analysis , while the chicken and zebrafish ( no PCR product was observed for xenopus ) present only one constitutive splice form ( Figure 4B , panel ( iii ) ) . The different size of the chicken PCR product is due to the fact that the forward primer was designed on the A3E , because the upstream exon does not exist in the University of California Santa Cruz Genome ( UCSC ) Browser ( see Table S1 ) . Again , the major mRNA product is the major form , which is the constitutive and only mRNA product in chicken and zebrafish , suggesting that it is the ancestral form . Based on the phylogenetic relationships among the analyzed organisms , we concluded that the alternative splice variant is a derived form , and the constitutive spliced variant is the ancestral one . A corresponding trend can be observed in group 1 A5E ( human ACTR6 gene; Figure 5A ) and group 2 A5E ( human NCOR1 gene; Figure 5B ) . The chicken PCR product is different in size , because the forward primer was designed on the A5E , as the upstream sequence is not fully mapped in the UCSC Genome Browser ( see Table S1 ) . The tested exons are alternatively spliced in human , mouse , and rat . In these species , the ancestral form is also the major mRNA product , but these exons are constitutively spliced in chicken , xenopus , and zebrafish ( Figures 4 and 5 , panel ( iii ) ) . Based on the phylogenetic relationships among the analyzed organisms and the above results , we concluded that the examined cases of A3Es and A5Es originated from a constitutive ancestral form ( originally including the alternative extension or excluding it ) . Mutations upstream or downstream of the ancestral splice site created a new functional splice site that competes with the original splice site , but with a lower selection level . This competition results in an alternative selection between those sites . This hypothesis is consistent with the high level of similarity found between A3Es and A5Es and constitutive exons . It is worth mentioning that although we only demonstrated four cases , in which the origin of alternative 3′ss and 5′ss is from previously constitutively spliced exons , we believe that since these cases are representatives of the four exon groups , and their characteristics apply to all exons , we thus suspect that a major portion of the A3Es and A5Es originated by this process .
Here we used bioinformatic tools to examine A3Es and A5Es that are conserved between human and mouse and , thus , are potentially functionally meaningful events . We divided each of the A3Es and A5Es datasets into two subgroups: one in which the major splice form is spliced using the splice site within the exon ( group 1 ) and the other in which the major splice form is spliced using the splice site at the end of the exon ( group 2 ) , according to each of their alternative forms inclusion level . We used these four subgroups to examine unique characteristics that distinguish these types of splicing events . It has been previously shown that constitutive and alternative cassette exons differ in several features , such as exon length , conservation level between human and mouse ( exon identity ) , divisibility-by-3 ( symmetry ) , and KA/KS ratio test [6 , 14 , 29 , 39] . We demonstrated that A3Es and A5Es , as a whole , show a resemblance in the aforementioned characterizing features to constitutive exons and differ from alternative cassette exons . These findings are statistically significant for group 1 and group 2 A3Es , as well as for group 1 A5E exons . Further , we showed that when examining splice site strength , flanking introns conservation , and Alu density in the fixated side ( i . e . , 5′ in A3Es and 3′ in A5Es ) and the alternative side ( i . e . , 3′ in A3Es and 5′ in A5Es ) separately , we find that the fixated side shows a resemblance to constitutive exons , whereas the alternative side shows a resemblance to alternative cassette exons . Thus , we suggest that A3Es and A5Es can be regarded as an intermediate state between constitutive and alternative cassette exons . Moreover , we showed that the alternative side of the A5Es requires precise regulation , presumably to achieve proper splice site selection . Previously , it was shown that ESE and ESS elements have the major effect on splicing when located in proximity to the alternative splice site [9–11] . In the case of group 1 A5Es , this regulation is essential , because both alternative splice sites are of similar strength; high ESE/ESR and low ESS density in proximity to the major splice site are probably the major factors that govern the level of each site usage in splicing . On the other hand , in group 2 A5Es , the major splice site is stronger than the minor splice site . Thus , the ESE/ESR density is higher , and the ESS density is lower in proximity to the minor splice site . This provides relative predominance and enables the selection level of the minor splice site and , thus , presumably regulates alternative splice site selection . It is noteworthy that examination of the ESR density in proximity to the alternative splice sites in A3Es found no significant results ( unpublished data ) . This may be because ~50% of A3Es belong to the NAGNAG family [20 , 21 , 24] , and , thus , the alternative forms differ from each other by only 3 bp , which makes the region scanned for ESRs overlap between the two alternative forms . Also , it is possible that , in A3Es , the ESRs play a minor role in the decision of major/minor 3′ss selection and that screening downstream of the branch site determines alternative 3′ss usage [47 , 48] . Relying on the assumption that sequence conservation implies functional properties , we conducted an analysis on the conservation of flanking intronic sequences . In agreement with other reports [6 , 20] , we showed that the conservation level of the intronic sequence flanking the alternative side resembles the conservation level observed in alternative cassette exons . This analysis also revealed that , while the conservation level in the downstream intron flanking A5Es is relatively similar to alternative cassette exons , the conservation level of the upstream intron flanking A3Es is in between the conservation level of constitutive and alternative cassette exons . To resolve this discrepancy , we decided to divide the A3Es dataset into two subgroups that are secluded by the length of the extension sequence . The chosen cutoff was 8 bp , which is based on a previous report showing that an additional PPT appears only when the two alternative splice sites are at least 8 bp apart [24] . We were intrigued to find that the subgroup whose extension is 8 bp or less presents a low intronic conservation level that resembles constitutive exons , while the subgroup whose extension is 9 bp or larger presents a high intronic conservation level which resembles alternative cassette exons . This suggests that regulation via the intronic sequences on alternative 3′ss occurs only when the alternative splice sites are not in close proximity , whereas selection of splice sites that are located in close proximity may be “noise” of the splicing mechanism [23] . We conducted the same analysis on A5Es and found that the subgroup of exons whose extension is 8 bp or less shows a conservation level that is between constitutive and alternative cassette exons , and the other subgroup showed a high conservation level that is similar to ( and even higher than ) alternative cassette exons ( Figure 3 and Tables 5 and 6 ) . The KA/KS ratio is used for the estimation of the selective forces acting on proteins [14] . That is , KA/KS ≪ 1 indicates purifying selection and KA/KS ≫ 1 indicates positive selection . We used this estimation tool to examine whether the selective forces throughout evolution acted in a different way on the core exon ( i . e . , the exon sequence not including the region that is alternatively spliced ) compared with the alternative region ( i . e . , the region which is either included or not ) . We performed the analysis for the subgroup of exons for which the major splice site is within the exon ( group 1 ) and for the subgroup of exons for which the major splice site is at the end of the exon ( group 2 ) separately . In both group 1 and group 2 exons , the core exon presented KA/KS ≪ 1 , indicating high purifying selection . In group 1 exons , the extension ( i . e . , the alternative region ) generally presented KA/KS ≫ 1 , indicating that it is under positive selection , so that the extension form is free to evolve and eventually potentially acquire a new function . In contrast , in the case of group 2 exons , the alternative region generally presented KA/KS ≪ 1 , which was relatively similar to the core KA/KS , indicating that the same selective forces have acted on both regions . This raises the possibility that , in the case of group 1 exons , the extension region was newly introduced during evolution , while in the case of group 2 exons the extension region was part of the original exon . Based on this analysis , we suggest that dynamic evolution is the key for the emergence of a new A3E or A5E . We propose that a certain fraction of the A3Es and A5Es originated from a previously constitutively spliced exon that acquired a competing splice site upstream or downstream of the ancestral authentic splice site . As a result , both sites are selected , thus giving birth to an alternative 5′ss or 3′ss exon . We believe that selective forces acted to ensure the symmetry of the created alternative region ( the region between the authentic splice site and newly created splice site ) , to maintain the same reading frame as that of the ancestral form , presumably by the selection of events that ensure this symmetry . We also assume that , initially , the selection of the new splice site was very rare and lacked biological functionality , but during the course of evolution such functionality was acquired , in a process that is called exaptation [49] . Thus , an ancestral form including the extension ( i . e . , alternative splice site is within the original ancestral exon sequence ) will show a similar KA/KS ratio throughout the exon , because the same evolutionary purifying selection forces took effect before and after the appearance of the alternative splice site . However , an ancestral form excluding the extension ( alternative splice site is inside the flanking intron ) will show a higher KA/KS ratio in the extension ( alternative added sequence ) than in the original short form , because the evolutionary purifying selection forces began operating at a much later time . We have shown that in four tested features , the A3Es and A5Es are similar to constitutive exons ( see above ) . However , in one subset of group 2 A5Es , the exons are more similar in their conservation level and in their KA/KS ratio to alternative cassette exons than to constitutive exons . We believe that the evolutionary model described previously can also serve as a possible explanation for these results . That is , the alternative 5′ss appeared during evolution within exons that were already under purifying selection to maintain the coding sequence . Thus , the exon that includes the alternative region is the evolutionary ancient one , and the exon that excludes the alternative region is the new addition . Another possibility is that A3Es and A5Es can also originate from alternative cassette exons in which one of the splice sites ( 5′ss in A3Es and 3′ss in A5Es ) is strengthened , because of mutations , and an existing or a newly generated cryptic splice site began to be selected as an alternative splice site . We decided to examine our hypothesis of constitutive origin on a limited number of exons manually; one example represented a group 1 exon and one example represented a group 2 exon for both A3Es and A5Es . We used multiple species alignment of seven species for each of the selected exons and their flanking splice sites to pinpoint the molecular evolutionary changes . The multiple alignments show that while in chicken , xenopus , and zebrafish only one functional splice site can be observed ( AG and GT for 3′ss and 5′ss , respectively ) , in human , mouse , rat , and , sometimes , in opossum , another alternative functional splice site was created by a mutation during evolution . It suggests that , based on a known evolutionary tree ( see Figure S3 ) , the appearance of new functional alternative splice sites as a result of mutations in the discussed examples is mammal-specific ( not necessarily in all mammals ) . We further validated these cases experimentally , showing that the most plausible scenario is that A3Es and A5Es were originated from previously constitutive exons . In these exons , an alternative splice site emerged in the lineage leading to mammals , following a mutation that created a functional 3′ss or 5′ss ( AG or GT , respectively ) . After the creation of the alternative site , selection pressures show differences between sites created outside of the ancestral exon or within the ancestral exon . In the case of group 1 exons relative to the ancestral form , the extension is poorly conserved ( between the seven species ) and presents KA/KS ≫ 1 , while in the group 2 exons , the extension is relatively highly conserved and presents KA/KS < 1 . Although our analysis experimentally validates the hypothesis for these cases , a broader evolutionary examination and large-scale analysis is still required to examine the scope and magnitude of this phenomenon . In conclusion , we examined the characteristics of A3Es' and A5Es' splicing forms , showing that A3Es and A5Es contain an anchor splice site that is as strong as that of constitutive exons and alternative sites that are weaker than cassette exons . We propose an evolutionary dynamic model in which A3Es and A5Es originated from ancestral constitutive exons that following mutation/s , a new alternative splice site appeared , and started competing with the ancient one for splice site selection . This model is supported by bioinformatic analysis showing that A3Es and A5Es are similar to constitutive exons . It was validated for four exons by multiple species' comparison and experimental validation . We also present that when two alternative 5′ss are of a similar strength , a delicate ratio between ESE/ESS located immediately upstream of each splice site determines which one is the major or minor selected site . In addition , A3Es and A5Es whose alternative splice sites are at least 9 bp apart show a high intronic sequences conservation level , indicative of the participation in the splicing regulation of these exons .
We assembled a unified dataset of human–mouse orthologous exons [6 , 30 , 31] . The dataset included 45 , 553 constitutively spliced exons , 757 cassette exons , 530 A3Es , and 232 A5Es . For each of the A3Es and A5Es , obtained from Sugnet et al . ( 2004 ) , we extracted the DNA and mRNA sequences using the BLAST program and produced an exon/intron structure using the Sim4 program . The Sim4 program receives a DNA sequence and an mRNA sequence and projects the mRNA onto the DNA considering the exon/intron architecture and the flanking consensus splice sites . This led to acquisition of both forms of A3Es and A5Es , including their respective flanking intronic sequences . For the Carmel et al . ( 2004 ) and Sorek et al . ( 2004 ) datasets , this information was already extracted . The dataset of A3Es and A5Es was divided into two subgroups: ( i ) exons whose major splice site is within the exon and ( ii ) exons whose major splice site is at the end of the exon . These subgroups were defined as “group 1” and “group 2 , ” respectively ( Figure 1 ) . The determination of whether an exon is in one subgroup or another was made by locating all the mRNA and EST sequences hits received from blasting that exon and the flanking exons against all the sequences in the UCSC database ( http://genome . ucsc . edu ) , requiring at least one transcript representing each form . The number of long forms , i . e . , those that include the alternative region ( NL ) , versus short forms , i . e . , those that exclude the alternative region ( NS ) , were counted , and the extension inclusion percentage ( i . e . , the fraction of mRNAs and ESTs containing the alternative region ) was calculated as [NL / ( NL + NS ) ] * 100 . Exons whose extension inclusion percentage was less than 40% were attributed to “group 1 , ” and exons whose extension inclusion percentage was more than 60% were attributed to “group 2 . ” Exons whose extension inclusion percentage was in the range of 40% to 60% were not included in either group , so as to prevent false predictions resulting from a low EST coverage or borderline cases . 3′ss and 5′ss scores were extracted using a program based on the “Analyzer Splice Tool” server ( http://ast . bioinfo . tau . ac . il/SpliceSiteFrame . htm ) . The program was adjusted for multisequence analysis , using the algorithm of Shapiro and Senapathy ( 1987 ) . Only canonical splice sites ( i . e . , AG in positions −2 and −1 for 3′ss and GT in positions +1 and +2 or GCA in positions +1 to +3 for 5′ss ) were considered . The analysis was executed for the human and mouse splice sites separately . A region of 100 bp for each flanking intron was selected . The intronic region flanking the alternative side was defined as the sequence downstream of the distal 5′ss ( “most” downstream 5′ss ) and upstream of the proximal 3′ss ( “most” upstream 3′ss ) for A5Es and A3Es , respectively . A homology examination was performed using Sim4 [50] with its default parameters . Briefly , this program detects exact matches of length 12 and extends them in both directions with a score of 1 for a match and −5 for a mismatch , stopping when extensions no longer increase the score . For homology of 100 bp long , the following 100 bp were also examined ( and so on ) . The length of the homology was defined as the sum of lengths of the homology regions identified by Sim4 ( L1 . . n ) , and the weighted identity percentage was calculated as ( Σi=1 . . n Li * Ii ) / Σi=1 . . n Li , where Ii is the identity percentage for the I'th homology region . The relative percentile was calculated for each of the sequences . The A3Es and A5Es datasets were then both divided into two subgroups , according to the length of the extension ( number of bp between the alternative splice sites ) when the cutoff used was 8 bp . The analysis was then also conducted for each subgroup separately . The exon and flanking splice site sequences of rat , opossum , chicken , xenopus , and zebrafish were extracted , using the UCSC Genome Browser ( http://genome . ucsc . edu ) . The sequences of the seven species ( including human and mouse ) were aligned using the ClustalW program , with its default parameters ( [46] , http://www . ebi . ac . uk/clustalw ) . Major and minor ( authentic—i . e . , supported by mRNAs or ESTs; potential—i . e . , not supported by mRNAs and ESTs and noncanonical ) 3′ss and 5′ss were located and marked . Chicken ( 5-d embryo ) and zebrafish ( adult ) were disrupted in TRIzol ( Sigma , http://www . sigmaaldrich . com ) with a Polytron homogenizer ( PT-MR2100 Kinematica , http://www . kinematica . ch ) , or with a hand-held motor-pestle ( Kimble-Kontes , http://www . kimble-kontes . com ) for the xenopus' oocytes . After complete homogenization of the tissue , the total RNA was isolated . The samples were treated with 2U of RNase-free DNase ( Ambion , http://www . ambion . com ) . RT was performed on 1–2 μg total RNA using RT–AMV ( avian myeloblastosis virus , Roche , http://www . roche . com ) following the manufacturer's protocol . We used commercial brain cDNA for human , mouse , and rat ( BioChain , http://www . biochain . com ) . Endogenous PCR amplification was performed , using Taq polymerase ( BioTools , http://www . btools . com ) and species-specific primers ( see Table S2 for the sequences of the primers ) . Amplification was performed for 30 cycles , consisting of denaturation for 30 s at 94 °C , annealing for 45 s at the appropriate Tm , and elongation for 1–2 min at 72 °C . The spliced cDNA products were separated in 3%–3 . 5% agarose gel . We note that , in the case of UBQLN4 ( Figure 4B ( iii ) ) and NCOR1 ( Figure 5B ( iii ) ) exons , the RT–PCR failed to amplify the cDNA of the xenopus . All PCR products were eluted and sequenced .
Accession numbers from the National Center for Biotechnology Information ( http://www . ncbi . nlm . nih . gov/RefSeq ) are: Homo sapiens PRPF3 ( NM_004698 ) , Gallus gallus ( NM_001031390 ) , Homo sapiens UBQLN4 ( NM_020131 ) , Mus musculus ( NM_033526 ) , Gallus gallus ( NM_001031373 ) , Xenopus tropicalis ( NM_001037720 ) , Danio rerio ( NM_213356 ) , Homo sapiens ACTR6 ( NM_022496 ) , Xenopus tropicalis ( NM_001016472 ) , and Homo sapiens NCOR1 ( NM_006311 ) . Accession numbers from The National Center for Biotechnology Information GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) are Mus musculus ( BC026607 ) , Rattus ( CB586372 ) , Xenopus tropicalis ( CX437471 ) , Danio rerio ( CK028262 ) , Rattus ( CB747083 ) , Mus musculus ( CK617284 ) , Rattus ( CK365697 ) , Danio rerio ( BC045961 ) , Rattus ( CB546074 ) , Gallus gallus ( BU326580 ) , Xenopus tropicalis ( BC099620 ) , and Danio rerio ( CK015548 ) . The accession number from the EMBL Nucleotide Sequence Database ( http://www . ebi . ac . uk/embl ) is Gallus gallus ( AJ719762 ) . The accession number from the DNA Data Bank of Japan ( http://www . ddbj . nig . ac . jp ) is Mus musculus ( AB093281 ) . | Alternative splicing is the mechanism that is responsible for the creation of multiple mRNA products from a single gene . It is considered a key player in genomic complexity achievement . Alternative 3′ and 5′ splicing events in which part of the exon is alternatively included or excluded in the mRNA constitute a significant part of all alternative splicing events , and yet little is known regarding their regulation mechanism and the evolutionary background that led to their creation . We show that alternative 3′ and 5′ splice site exons resemble constitutive exons . However , their alternative sequence resembles alternative cassette exons . Comparative genomics spanning seven vertebrate species suggests an evolutionary model in which the alternative state is a derivative of an ancestral constitutive exon , where a mutation inside the exon or along the flanking intron resulted in the creation of a new splice site that competes with the original one , leading to alternative splice site selection . This model was validated experimentally , showing that during evolution mutations shifted constitutive exons to undergo alternative 3′ and 5′ splicing . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"evolutionary",
"biology",
"molecular",
"biology",
"vertebrates"
] | 2007 | The Emergence of Alternative 3′ and 5′ Splice Site Exons from Constitutive Exons |
The posttranslational modifiers SUMO and ubiquitin critically regulate the DNA damage response ( DDR ) . Important crosstalk between these modifiers at DNA lesions is mediated by the SUMO-targeted ubiquitin ligase ( STUbL ) , which ubiquitinates SUMO chains to generate SUMO-ubiquitin hybrids . These SUMO-ubiquitin hybrids attract DDR proteins able to bind both modifiers , and/or are degraded at the proteasome . Despite these insights , specific roles for SUMO chains and STUbL in the DDR remain poorly defined . Notably , fission yeast defective in SUMO chain formation exhibit near wild-type resistance to genotoxins and moreover , have a greatly reduced dependency on STUbL activity for DNA repair . Based on these and other data , we propose that a critical role of STUbL is to antagonize DDR-inhibitory SUMO chain formation at DNA lesions . In this regard , we identify a SUMO-binding Swi2/Snf2 translocase called Rrp2 ( ScUls1 ) as a mediator of the DDR defects in STUbL mutant cells . Therefore , in support of our proposal , SUMO chains attract activities that can antagonize STUbL and other DNA repair factors . Finally , we find that Taz1TRF1/TRF2-deficiency triggers extensive telomeric poly-SUMOylation . In this setting STUbL , together with its cofactor Cdc48p97 , actually promotes genomic instability caused by the aberrant processing of taz1Δ telomeres by DNA repair factors . In summary , depending on the nature of the initiating DNA lesion , STUbL activity can either be beneficial or harmful .
The Small Ubiquitin-like Modifier ( SUMO ) is a posttranslational modifier ( PTM ) that critically regulates most aspects of cell growth . For example , key roles for SUMOylation in genome stability , transcription , proteostasis , and embryonic development have been identified [1–11] . In fission yeast , SUMO is covalently attached to lysine residues within its targets via a conserved enzymatic cascade of E1 activating ( Fub2:Rad31 heterodimer ) , E2 conjugating ( Ubc9 ) , and E3 ligase factors ( Pli1 , Nse2; [1] ) . Pli1 and Nse2 drive the majority of SUMOylation , and are members of the evolutionarily conserved PIAS class of SUMO E3 ligases [12–14] . Pli1-dependent SUMOylation requires a non-covalent SUMO:Ubc9 complex [12] , which is also required to generate SUMO chains by the sequential addition of SUMO to N-terminal lysine residues of SUMO ( K14 and K30 in fission yeast; [12 , 15–17] ) . Interestingly , despite catalyzing more than 90% of cellular SUMOylation , deletion of Pli1 in fission yeast does not cause overt growth defects or sensitivity to genotoxins [12 , 18] . However , cells lacking Pli1 have defects in centromere silencing , and also have elongated telomeres due to the loss of Tpz1 SUMOylation [14 , 18–20] . Despite a relatively restricted target pool , Nse2-mediated SUMOylation is required for genotoxin resistance and genome stability , but not telomere length control [12–14 , 21–24] . Countering the activities of Pli1 and Nse2 , the SUMO proteases Ulp1 and Ulp2 ( also called SENPs ) remove SUMO from target proteins , and disassemble poly-SUMO chains [25–27] . Deletion of Ulp2 causes a striking accumulation of SUMO conjugates , especially high molecular weight ( HMW ) species that correspond to SUMO chains [28] . Abrogating SUMO chain formation by mutating SUMO's N-terminal acceptor lysine residues rescues many of the severe phenotypes caused by Ulp2 deletion , which include sensitivity to various genotoxic agents , genome instability and temperature sensitivity [12 , 28] . Moreover , deletion of Pli1 , which catalyzes SUMO chain formation , also rescues ulp2Δ phenotypes [29] . Importantly therefore , in contrast to global hypoSUMOylation e . g . in pli1Δ cells , uncontrolled SUMO chain formation has relatively dire consequences for the cell . SUMO chains are also key substrates for the SUMO-targeted ubiquitin ligase ( STUbL ) family of E3 ubiquitin ligases [30–34] . STUbLs ubiquitinate SUMO chains , both “capping” them [35] and creating a dual signal for the recruitment of proteins able to simultaneously bind both SUMO and ubiquitin e . g . the AAA ATPase p97/Cdc48-Ufd1-Npl4 that remodels chromatin-associated complexes , and the BRCA1 cofactor RAP80 [36 , 37] . STUbL and p97/Cdc48 together constitute a mechanism for the extraction and/or degradation of SUMO/ubiquitin modified proteins from chromatin or other protein complexes [36 , 38–40] . For example , the SUMOylated and chromatin bound fraction of FANCI , a Fanconi Anemia complex protein , is controlled by STUbL and p97-dependent extraction from chromatin and proteasomal degradation [41] . In addition , the poly-SUMOylated forms of the DDR factors MDC1 and RPA are also subject to STUbL-dependent regulation at DNA lesions [42–45] . The accumulation of poly-SUMOylated MDC1 and RPA at lesions in STUbL deficient cells correlates with defects in downstream events in the DDR e . g . Rad51 loading , and a blockade to homologous and non-homologous recombination repair . In these studies SUMO conjugation defective mutants of each factor were used to infer the impact of their poly-SUMOylation and STUbL-dependent processing on DNA repair . However , such mutants are refractory to mono- and poly-SUMOylation , as well as other possible lysine modifiers including ubiquitin . Therefore , questions remain about the specific contribution ( s ) of SUMO chains and STUbL in regulating the DDR . As mentioned , Pli1-dependent SUMOylation of Tpz1 regulates telomere length homeostasis [19 , 20] . SUMOylation also controls the cold sensitive phenotype of fission yeast lacking the telomere shelterin factor Taz1 ( human TRF1/2 ) [46] . Specifically , taz1Δ cells are cold sensitive due to telomere "entanglement" and subsequent chromosome breakage during mitosis; phenotypes that can be prevented by reducing SUMOylation of the RecQ DNA repair helicase Rqh1 [46 , 47] . Thus , SUMOylated Rqh1 is toxic to cells lacking Taz1 , but potential roles for SUMO chains and STUbL in this toxic DNA repair process remain untested . Overall , diverse STUbL targets have been identified , and the SUMO-dependent degradation of proteins appears to be pivotal for the output of a number of cellular pathways , including DNA repair [7 , 38 , 48] . However , if the SUMO-dependent degradation of certain DNA repair proteins was required , then preventing SUMO chain formation should yield phenotypes similar to those of cells lacking STUbL i . e . genome instability and lethality [30–34 , 42 , 44 , 45 , 49] . On the contrary , cells containing SUMO chain blocking mutations lack overt phenotypes in fission yeast e . g . SUMOK14 , 30R [12] , and analogous SUMO mutants in budding yeast have very mild phenotypes when compared to STUbL deficiency [28 , 50] . In light of the above data , we hypothesized that a critical physiological role of STUbL in the DNA damage response ( DDR ) is to prevent a local toxic buildup of SUMO chains . By manipulating SUMO pathway homeostasis and assessing the DNA damage response ( DDR ) to distinct genomic lesions we obtain key support for the above model . Furthermore , we identify the SUMO-interacting Swi2/Snf2 family DNA translocase , Rrp2 , a candidate mediator of SUMO chain toxicity in STUbL mutant cells . Finally , in addition to the known positive functions of STUbL , we show that at certain DNA lesions its activity can actually cause genome instability .
We previously showed that SUMO chains spontaneously accumulate in a single large subnuclear focus in STUbL mutant cells , suggesting that STUbL prevents SUMO pathway nucleation and rampant SUMO chain formation [36] . To test this more directly , we expressed both LacI-RFP-SUMO and LacI-GFP in fission yeast that have a cluster of LacO repeats integrated at the lys1 locus of chromosome 1 ( ~256 repeats , [51] ) . LacI-GFP was expressed from the nmt1 promoter whereas LacI-RFP-SUMO was expressed from the attenuated nmt81 promoter , both under repressive conditions ( + thiamine [52] ) . This sets up a competition wherein the more abundant LacI-GFP protein occupies the LacO repeats , forming visible foci and largely excluding LacI-RFP-SUMO ( Fig 1A ) . In wild-type cells , as anticipated the LacI-GFP protein can be readily visualized as a focus on the LacO repeats , whereas LacI-RFP-SUMO is diffuse and pan-nuclear ( Fig 1A ) . In contrast , LacI-GFP and LacI-RFP-SUMO are both frequently present in co-localizing nuclear foci in slx8-29 cells that have compromised STUbL activity ( Fig 1A ) . Quantification revealed that LacI-GFP and LacI-RFP-SUMO foci colocalize in ~18% of wild-type cells , and this colocalization was strikingly increased to ~80% in slx8-29 cells , consistent with a role for STUbL in antagonizing SUMO chain nucleation ( Fig 1B ) . Notably , expression of LacI-RFP-SUMOK14 , 30R with LacI-GFP reduced the colocalization of GFP and RFP signals to ~40% in slx8-29 cells , indicating that the observed SUMO foci contain SUMO chains . That LacI-RFP-SUMOK14 , 30R expression does not completely suppress focus formation to the levels of LacI-RFP-SUMO foci in wild-type cells is expected based on the residual limited capacity of SUMOK14 , 30R to form SUMO chains through alternative internal lysine residues [36] . Therefore , STUbL can antagonize SUMO chains when they are nucleated on chromatin , even in the absence of active DDR signaling . If countering SUMO chain nucleation and growth is a key function of STUbL , then reducing SUMO chain formation should lessen cellular dependency on STUbL . Indeed , deleting the SUMO chain catalyst Pli1 strongly suppresses STUbL mutant phenotypes , and also makes the normally essential fission yeast STUbL dispensable for growth ( Fig 2A & [30] ) . We next asked if simply reducing the levels of wild-type SUMO would make it limiting for chain formation , thereby phenocopying pli1Δ . To this end , we replaced the endogenous SUMO promoter with a weaker constitutive promoter [53] . In stark contrast to sumoΔ cells , those expressing less SUMO ( tetO7-SUMO ) appear wild-type for growth and genotoxin resistance ( Fig 2B ) . We then determined the effect of tetO7-SUMO on cells with compromised STUbL activity . Strikingly , tetO7-SUMO potently rescued the temperature and genotoxin sensitivity of slx8-29 cells ( Fig 2B ) . Under conditions where slx8-29 cells are completely growth inhibited , tetO7-SUMO slx8-29 double mutants grow almost as well as tetO7-SUMO single mutants . In addition , tetO7-SUMO similarly suppressed the genotoxin sensitivity of ulp2Δ cells , which lack the SUMO chain editing protease Ulp2 ( S1A Fig , [28] ) . The phenotypic rescues of slx8-29 and ulp2Δ cells were accompanied by a reduction in the HMW SUMO chains that normally accumulate in these mutants ( S1B Fig , [54] ) . Supporting the SUMO chain specificity of slx8-29 suppression by pli1Δ and tetO7-SUMO , the SUMOK14 , 30R chain mutant also suppresses slx8-29 ( Fig 2C , [12] ) . We previously identified a role for STUbL in the repair of topoisomerase 1 ( Top1 ) -dependent lesions [22] . In the absence of Tdp1 , an enzyme that catalyzes the reversal of Top1-DNA adducts , cells require STUbL activity for removal of Top1-DNA adducts via a parallel repair pathway ( Fig 2D , [22] ) . This dependency can be partially alleviated by introducing the SUMO chain reducing mutant SUMOK14 , 30R ( Fig 2D ) . Notably , tetO7-SUMO is more effective than SUMOK14 , 30R at rescuing the synthetic slx8-29 tdp1Δ phenotype ( Fig 2D ) . This difference may be due to the residual SUMO chain-forming capacity of SUMOK14 , 30R [12 , 36] . HMW SUMO species in slx8-29 tdp1Δ cells are reduced by tetO7-SUMO and SUMOK14 , 30R , correlating with the extent of rescue ( Fig 2E ) . Overall , these data indicate that "low dose SUMO" ( tetO7-SUMO ) limits toxic Pli1-dependent SUMO chain formation in STUbL mutant cells , similar to but more potently than SUMOK14 , 30R . It is somewhat surprising that SUMO chain formation follows an apparently passive mechanism , likely driven by the normally high local concentrations of SUMO . In this setting , local pathway control by STUbL and deSUMOylating enzymes would appear critical . Such a role for STUbL is well supported by the focal accumulation of SUMO chains in fission yeast STUbL mutants ( Fig 1 & [36] ) . Our data are consistent with SUMO chains in STUbL mutant cells exerting an inhibitory effect on DNA repair . To identify potential SUMO chain "effectors" , we purified fission yeast proteins that have affinity for SUMO chains [36] . In this way we made the initial identification of Cdc48-Ufd1-Npl4 as a STUbL co-factor that recognizes both SUMO and ubiquitin modifications on its targets [36] . Another SUMO chain interactor we identified was the Swi2/Snf2-related DNA translocase called Rrp2 in fission yeast and Uls1 in budding yeast . Both Rrp2 and Uls1 bind SUMO [32 , 55] , and Uls1 has been described as a STUbL due to its interaction with SUMO chains and the presence of an E3 ubiquitin ligase RING domain ( although ubiquitin ligase activity is yet to be demonstrated ) [32 , 56] . An elegant study in budding yeast suggested that Uls1 could be an antagonist of the Slx8-based STUbL [56] , but straightforward interpretation of this model was complicated by the synthetic sickness of slx8Δ and uls1Δ [32 , 57 , 58] . In light of the foregoing , we tested for a genetic interaction between the slx8-29 and rrp2Δ mutations . Deletion of Rrp2 did not cause overt phenotypes such as temperature or genotoxin sensitivity ( Fig 3A ) . Strikingly however , rrp2Δ instead suppressed the temperature and genotoxin sensitivities of slx8-29 cells ( Fig 3A ) . Suppression of slx8-29 phenotypes by rrp2Δ was accompanied by a reduction in the levels of HMW SUMO conjugates ( Fig 3B ) . On the other hand , ectopic expression of Rrp2 strongly induced the formation of HMW SUMO species , in wild-type but not SUMO chain deficient backgrounds ( Fig 3C ) . Moreover , deletion of Rrp2 mitigated the synthetic sickness and camptothecin sensitivity of slx8-29 and tdp1Δ mutations ( Fig 3D ) . The simplest interpretation of these data is that Rrp2 inhibits SUMO chain-directed STUbL activity , which is further supported by the analysis of crosstalk between Uls1 and STUbL in budding yeast [56] . Together with the known functions for Rrp2 and Uls1 in modulating DNA repair [55 , 59–61] , our results reveal Rrp2 as an excellent candidate mediator of the DDR-inhibitory effects of SUMO chains in slx8-29 cells . The telomeres of cells lacking Taz1 are elongated and become "entangled" at low growth temperatures causing chromosome breakage during mitosis , G2 cell cycle checkpoint activation , and cell death [46 , 47 , 62] . The precise mechanism of taz1Δ telomere entanglement is unknown . However , it is promoted by DNA processing factors such as the RecQ DNA helicase Rqh1 and the DNA exonuclease Exo1 [46 , 63] , and can be suppressed by certain Topoisomerase II mutations [64] . Interestingly , all of these factors are regulated by SUMO conjugation [46 , 65–67] , and reducing Rqh1 SUMOylation suppresses taz1Δ telomere entanglement [46] . Notably , the impact of SUMO chains on taz1Δ phenotypes has not been determined . Therefore , we assessed the effects of a panel of SUMO pathway-related mutants on the cold sensitivity of taz1Δ cells . Strikingly , pli1Δ taz1Δ and nup132Δ taz1Δ double mutant cells grow well at low temperature compared to taz1Δ cells ( Fig 4A ) . Previously , Pli1 deletion was suggested to not suppress taz1Δ cold sensitivity [46] . Therefore , we further validated our results by re-expressing Pli1 in pli1Δ taz1Δ cells , which confirmed that Pli1 activity promotes taz1Δ cold sensitivity ( Fig 4B ) . The similar effect of deleting Pli1 or Nup132 on taz1Δ cold sensitivity likely reflects the STUbL-mediated degradation of Pli1 in nup132Δ cells [29] . Finally , SUMOD81R that abolishes the SUMO:Ubc9 non-covalent complex and reduces Pli1-dependent SUMOylation also suppresses taz1Δ ( Fig 4C , [12] ) . We previously showed that Pli1 has separable roles in mono-SUMOylation and SUMO chain formation [12] . We therefore asked which function of Pli1 causes taz1Δ cold sensitivity . Notably , reducing SUMO chain formation with either the SUMOK14 , 30R or tetO7-SUMO allele also rescues taz1Δ cold sensitivity to a similar degree as pli1Δ ( Fig 4C ) . Therefore , Pli1-dependent SUMO chains , not mono-SUMOylation events , cause taz1Δ cold sensitivity . Reducing Rqh1 SUMOylation has been shown to rescue taz1Δ cold sensitivity [46] . Therefore , we considered that reduced Rqh1 SUMOylation could explain the rescue of taz1Δ phenotypes by tetO7-SUMO . Notably however , the Rqh1-SM SUMOylation defective mutant renders cells camptothecin sensitive [46] , whereas tetO7-SUMO cells are not ( Fig 4D ) . Therefore , Rqh1 SUMOylation is intact in tetO7-SUMO cells , as are other key SUMOylation events that control genotoxin resistance ( e . g . Fig 2B & 2D ) , and telomere length e . g . Tpz1 SUMOylation ( see below ) . Based on the suppression of taz1Δ phenotypes by SUMO pathway mutants that compromise SUMO chain formation , we tested if STUbL activity plays a role in this context . As can be seen , slx8-29 taz1Δ double mutant cells grow robustly at low temperature , unlike the taz1Δ single mutant ( Fig 5A ) . Moreover , ufd1-1 , a mutation that compromises the Cdc48/p97-dependent processing of STUbL targets [36] also rescues taz1Δ cold sensitivity ( Fig 5A ) . Expression of either Slx8 or the human STUbL RNF4 reverses the suppression of taz1Δ by slx8-29 , as does expression of Ufd1 in ufd1-1 taz1Δ cells ( Fig 5B ) . Therefore , SUMO chain-dependent STUbL and Cdc48 ( p97 ) activity has pathological consequences at taz1Δ telomeres . At low temperatures taz1Δ cells undergo checkpoint-dependent cell cycle arrest due to chromosome breakage in mitosis [47] . In fission yeast , cell cycle arrest is readily visualized as an increase in cell length . Therefore , we compared cell lengths of wild-type , taz1Δ and double mutants including tetO7-SUMO taz1Δ , slx8-29 taz1Δ and ufd1-1 taz1Δ at 19°C . As can be seen , taz1Δ cells are highly elongated compared to each of the double mutants , which are instead similar to wild-type in length ( Fig 5C & 5D ) . Given that these mutants are checkpoint proficient , SUMO chains , STUbL and Cdc48/p97 activity drive the chromosome damage and checkpoint activation in taz1Δ cells [22 , 36] . As the SUMO pathway controls telomere length [14 , 18–20] , we asked if suppression of taz1Δ phenotypes is due to restoration of normal telomere homeostasis in the SUMO pathway and taz1Δ double mutants e . g . slx8-29 taz1Δ . However , telomere length analysis revealed no overt effects on telomere length , with telomeres in each double mutant being highly elongated , as in taz1Δ cells ( Fig 6A ) . This finding is similar to that reported for the Rqh1 SUMOylation mutant Rqh1-SM , which also rescues taz1Δ cold sensitivity without affecting taz1Δ telomere length [46] . Recently , Pli1-dependent SUMOylation of the telomere factor Tpz1 was shown to antagonize telomerase activity by recruiting the Stn1-Ten1 complex to telomeres [19 , 20] . Therefore , deletion of Pli1 causes telomere elongation , which can be readily visualized by Southern analysis of chromosome termini ( Fig 6A , [14 , 18–20] ) . Interestingly , in addition to the expected telomere elongation in pli1Δ cells , our Southern analyses indicated telomere homeostasis defects in some other SUMO pathway mutants ( Fig 6A ) , which we analyzed further . We have shown that SUMOD81R abolishes the non-covalent SUMO:Ubc9 complex that fuels Pli1 activity and , that Pli1 is degraded by STUbL activity in cells lacking Nup132 ( human NUP133; [12 , 29] ) . In agreement with these analyses , telomeres in SUMOD81R and nup132Δ cells are longer than wild-type ( Fig 6B ) . In contrast , the telomeres of tetO7-SUMO , slx8-29 and ufd1-1 cells are essentially wild-type in length ( Fig 6B ) . Therefore , telomere length regulation through Pli1-dependent SUMOylation of Tpz1 [19 , 20] is retained in tetO7-SUMO cells , as well as those with STUbL pathway dysfunction . Interestingly however , telomeres in SUMO chain reducing SUMOK14 , 30R cells are clearly elongated when compared to those in wild-type cells ( Fig 6B ) . This unexpected finding either reveals a new role for SUMO chain-modified Tpz1 in telomere length maintenance , or indicates that SUMO lysine residues 14 and 30 are subject to other forms of regulatory posttranslational modification that impact telomere length . If SUMO chains promote STUbL activity at taz1Δ telomeres ( e . g . Fig 7A ) , then elevated SUMO conjugates should be detectable at telomeres using anti-SUMO chromatin immunoprecipitation and quantitative PCR ( ChIP-qPCR ) . Indeed , ChIP-qPCR revealed a strong enrichment of SUMO conjugates at taz1Δ versus wild-type telomeres , particularly in cells grown at 19°C ( Fig 7B ) . These telomeric SUMO conjugates are predominantly chains , as they are reduced to near wild-type levels in both taz1Δ tetO7-SUMO and taz1Δ SUMOK14 , 30R double mutant cells ( Fig 7C ) . Consistent with the known SUMO chain-directed activity of STUbL [7 , 29 , 38 , 68] , telomeric SUMO chains were further elevated in taz1Δ slx8-29 double mutant over taz1Δ cells ( Fig 7C ) . Therefore , SUMO chains generated on taz1Δ telomeres engage STUbL activity . A SUMO conjugated form of RecQ helicase Rqh1 was shown to promote telomeric entanglement in taz1Δ cells [46] . Therefore , we considered the possibility that SUMOylated Rqh1 contributes to the elevated SUMO signal detected at taz1Δ telomeres ( Fig 7B and 7C ) . However , neither rqh1Δ nor the Rqh1 SUMOylation deficient mutant Rqh1-SM [46] had a detectable affect on telomeric hyper-SUMOylation in taz1Δ cells ( Fig 7D ) . Moreover , because Rqh1-SM rescues taz1Δ cold sensitivity [46] , hyper-SUMOylation is not a consequence of Rqh1-dependent taz1Δ telomere entanglement ( Fig 7D ) . Together , our data indicate that the dysfunctional telomeres in taz1Δ cells are subjected to increased SUMO chain conjugation and subsequent STUbL activity . In turn , this likely promotes telomeric entanglements through downstream DNA repair factors such as Rqh1 [46] .
The highly conserved STUbL family has critical roles in the DNA damage response ( DDR ) and genome stability [38 , 48 , 69] . Importantly , STUbL could support the DDR in two quite distinct ways ( i ) STUbL may act in a sequential DNA repair cascade to remove specific poly-SUMO conjugates in a "programmed manner" [48] , or ( ii ) As the DDR promotes localized "group SUMOylation" of proteins at DNA lesions [70] , we hypothesized that STUbL may also ( or instead ) antagonize local SUMO chain formation . If left unchecked by STUbL , these SUMO chains could inhibit downstream DDR events . Together , our analysis supports inhibition of localized SUMO chain formation as a major role for STUbL in fission yeast . That is , if "programmed" SUMO chain and STUbL-dependent removal of DDR factors were critical for normal DNA repair , then blocking SUMO chain formation would cause DNA repair defects analogous to those of STUbL mutants . This is clearly not the case ( see Fig 2 & [12 , 30 , 36 , 71] ) . Moreover , inhibiting SUMO chain formation strongly reduces the need for STUbL in DNA repair , genome stability and cell growth ( Fig 4 & [12 , 30 , 36 , 71] ) . Therefore , although mono-SUMOylation plays critical roles in genome stability and the DDR , SUMO chain formation in STUbL mutant cells has a predominantly negative impact . STUbL could antagonize SUMO chains in a number of ways including ( i ) its ability to "cap" SUMO chains by ubiquitinating the amino-terminal lysine residues that are acceptors for SUMO chain growth [35] ( ii ) its ability together with Cdc48/p97 to promote the local degradation/extraction of the SUMO ligases Pli1 and SIZ1 [29 , 72] or ( iii ) its ability to degrade/extract other SUMOylated proteins that are "licensed" for SUMOylation e . g . MDC1 at DNA lesions , and therefore constantly nucleate the SUMOylation machinery [48] . SUMO chains in STUbL mutant cells clearly have a negative effect on genotoxin resistance , but how they exert this effect was undefined . Intriguingly , we identified the Swi2/Snf2 translocase Rrp2 in a SUMO chain-binding proteomic screen [36] , and further analysis indicates that it contributes to SUMO chain toxicity in STUbL mutant cells . Based on the known functions of Rrp2 ( ScUls1 ) [55 , 56 , 60 , 61] and our analysis of SUMOylation levels upon its deletion or overexpression , Rrp2 likely antagonizes STUbL activity at SUMO chains . In the future , it will be interesting to determine if Rrp2 also contributes to DNA repair inhibition by SUMO chains in slx8-29 cells . In this regard , both Rrp2 and Uls1 have been shown to regulate DNA repair [55 , 60] . For example , Uls1 inhibits the non-homologous end joining of telomeres in the presence of SUMO chains [60] . Furthermore , Rrp2 and Uls1 can inhibit Rad51-dependent homologous recombination [55 , 61] . Therefore , Rrp2 and its orthologs in other species could contribute to the DDR inhibition observed upon STUbL inactivation ( Fig 8 ) . So far , the analysis of SUMO chain function has depended on the use of SUMO mutants that reduce chain extension e . g . SUMOK14 , 30R [12 , 28 , 50] . However , these mutants are refractory not only to modification by SUMO itself , but also by other posttranslational modifications such as ubiquitination . Indeed , STUbL ubiquitinates the lysine residues used for SUMO chain formation in budding yeast , thereby "capping" growth of SUMO chains [35] . Here , we show that limited expression of wild-type SUMO ( tetO7-SUMO ) provides robust suppression of the phenotypes caused by STUbL and Ulp2 mutants ( Fig 2 & S1 Fig ) . Importantly , growth , genotoxin resistance and telomere length appear normal in tetO7-SUMO cells , demonstrating that key SUMOylation events required for genome stability are intact ( Figs 2 & 6 ) . For example , Tpz1 SUMOylation is lost in pli1Δ cells causing telomere elongation [19 , 20] , whereas tetO7-SUMO telomeres are wild-type in length . These striking results reinforce the model in which STUbL antagonizes local SUMO pathway activity at DNA lesions to prevent toxic SUMO chain buildup ( Fig 8 ) . Surprisingly , unlike tetO7-SUMO , the SUMO chain blocking mutant SUMOK14 , 30R causes telomere elongation . Therefore , although both tetO7-SUMO and SUMOK14 , 30R largely bypass the need for STUbL or Ulp2 activity , the mutations in SUMOK14 , 30R additionally impact Pli1 and Tpz1-dependent telomere length regulation [19 , 20] . How SUMOK14 , 30R impacts telomere length is unknown , but may be through defects in its ability to form short SUMO chains or to be ubiquitinated . This supports the utility of tetO7-SUMO in the analysis of SUMO pathway homeostasis , and in corroborating phenotypes ascribed to SUMO chains using SUMOK14 , 30R and analogous mutants . Analysis of taz1Δ telomere phenotypes provides new insight into dysfunctional telomere management by the SUMO pathway , and unexpected deleterious effects of STUbL activity . Taz1TRF1/TRF2-depleted telomeres become hyper-SUMOylated at low growth temperature ( Fig 7 ) . These telomeric SUMO conjugates contain SUMO chains , as they are strongly reduced in both tetO7-SUMO and SUMOK14 , 30R backgrounds . Strikingly , the genome instability and cell death caused by unscheduled repair of taz1Δ telomeres is suppressed by mutations in the SUMO pathway that compromise SUMO chain-dependent STUbL activity . Our data indicate that Pli1-dependent SUMO chain formation engages STUbL and Cdc48 ( p97 ) at taz1Δ telomeres ( Fig 8 ) . In turn , this makes taz1Δ telomeres permissive for pathological processing by Rqh1 and other DNA repair factors [46 , 47] . STUbL-dependent processing of structurally distinct DNA lesions in budding yeast has recently been characterized [73–75] . A common theme is emerging: SUMO accumulates at persistent lesions such as irreparable DNA double strand breaks , eroded telomeres in cells lacking telomerase , or triplet repeat tracts that are difficult to replicate [73–75] . These heavily SUMOylated lesions are then processed by STUbL activity at the nuclear pore , promoting alternative repair pathways . Perhaps most relevant to our study is the STUbL-mediated processing of eroded telomeres in telomerase negative budding yeast , which promotes the emergence of so called Type II survivors [73] . Type II survivors maintain telomere terminal tracts in manner dependent on a number of DNA repair factors , including the RecQ helicase Sgs1 . In this regard , it is noteworthy that the fission yeast Sgs1 homolog , Rqh1 , promotes taz1Δ telomere entanglement [46] . Therefore , it is tempting to speculate that STUbL processing of short eroded telomeres ( telomerase delete ) , or those that are highly elongated and difficult to replicate ( taz1Δ ) , allows access to similar DNA repair machinery , including RecQ helicases . In the case of short eroded telomeres , this RecQ-dependent processing is beneficial to the cell and improves fitness [73] , whereas it entangles the elongated telomeres of taz1Δ cells , leading to genome instability . Analysis of cytotoxic Top1-DNA adduct ( Top1cc ) processing provides additional evidence for SUMO chain-dependent inhibition of DNA repair in STUbL mutant cells ( Fig 2D; [22 , 36] ) . We showed that STUbL and Cdc48 ( p97 ) cooperate in Top1cc repair in cells that lack the Top1cc reversing enzyme Tdp1 [36] . Combining STUbL and Tdp1 mutations causes Top1-dependent genome instability and hypersensitivity to camptothecin [36 , 76] . Importantly , the sickness of slx8-29 tdp1Δ cells is strongly suppressed by tetO7-SUMO , and albeit less efficiently , by SUMOK14 , 30R ( Fig 2D; [22 , 36] ) . Therefore , the accumulation of SUMO chains in slx8-29 tdp1Δ cells inhibits Top1cc repair by the alternative mechanism i . e . Rad16-Swi10XPF-ERCC1 cleavage [22 , 36] . Deleting the SUMO chain-binding DNA translocase Rrp2 also partially suppresses the slx8-29 tdp1Δ phenotype ( Fig 3 ) , revealing a candidate mediator of repair inhibition in this context . In the absence of exogenous stress , STUbL may also maintain the functionality of chromatin loci that are normally heavily SUMOylated e . g . heterochromatic centromeres [14 , 18 , 77] . Indeed , in a recent proteomic analysis of STUbL and Cdc48 substrates in fission yeast , proteins located at centromeres and telomeres were over represented [40] . Moreover , deleting the major SUMO E3 ligase for these loci , Pli1 , bypasses the essential role ( s ) of STUbL in fission yeast ( this study & [30] ) . Overall , integrating our current data with the known DDR-inhibitory effects of STUbL and Cdc48/p97 dysfunction ( e . g . [42–45 , 48 , 49] ) , unchecked SUMO chains at DNA lesions appear to antagonize downstream repair processes ( Fig 8 ) . In this setting , the Rrp2 DNA translocase emerges as an excellent candidate mediator of SUMO chain toxicity . We also reveal that STUbL activity can destabilize the genome , depending on the nature of the initial DNA lesion . Given the highly conserved nature of STUbL and its cofactors , antagonizing localized SUMO chain accumulation is likely to be a conserved role of STUbL across species .
Standard media and growth conditions for S . pombe were used as described previously [78] . All strains ( Table 1 ) are of genotype ura4-D18 leu1-32 unless otherwise stated . For Spot Assays , cells were grown at 25°C to logarithmic phase ( optical density at 600 nm [OD600] of 0 . 6 to 0 . 8 ) , spotted in 5-fold dilutions from a starting OD600 of 0 . 5 on plates supplemented with the relevant drug . The plates were incubated at 19 to 35°C for 3 to 6 days . The chimera LacI-mCherry-SUMO was cloned by serial insertions of LacI and mCherry-SUMO into a pREP81 vector at the NdeI , then SalI and BamHI sites , respectively . The primers used are listed in Table 2 . The E . coli LacI with the last five amino acid truncated to disable tetramerization [79] was amplified using Omn406 and Omn411 primers from the E . coli genome . LacI was cloned in frame with mCherry-SUMOGG amplified using Omn083 and Omn408 from pREP41-mCherry-SUMO ( pMN087 ) . The sequence containing the Pnmt8 promoter and LacI-mCherry-SUMOGG was amplified using Omn423 and Omn424 primers , and transferred , by Gibson Assembly [80] , to pJK148 that has been linearized with PstI and SacI , to generate pNB075 . To mutagenize SUMO from wild-type to SUMOK14 , 30R , site-directed mutagenesis was performed using the JP243/ JP244 ( K14R ) , oNB112/oNB113 ( K30R ) primers to make pJK148-Pnmt8-LacI-mCherry-SUMOK14 , 30R ( pNB077 ) . The pNB075 or pNB077 plasmid was linearized by digesting with NruI and integrated at the leu1 locus by homologous recombination . For live imaging , cells were grown in filter-sterilized EMM medium supplemented with leucine , uracil , arginine , and histidine ( LUAH ) to logarithmic phase . Bright-field and fluorescence images of live cells were acquired using a Nikon Eclipse microscope with a 100x Plan Apochromat DIC H oil immersion objective and a Photometrics Quantix charge-coupled device camera . Images were analyzed with NIH ImageJ software . TCA precipitation of total proteins and Western blotting were carried out as previously described [29] . The membrane was blocked in 1% w/v non-fat milk in phosphate buffer saline solution with 0 . 1% v/v Tween-20 , probed with antibodies against α-tubulin ( Sigma ) and S . pombe SUMO [12] , followed by HRP- or IRDye-conjugated secondary antibodies , and detected either using an ECL Dura system ( Pierce ) on film; or scanning on an ODYSSEY scanner ( Li-Cor ) . Fission yeast genomic DNA was digested with EcoRI , resolved on a 1% agarose gel , and then transferred onto a Hybond-XL membrane ( GE Healthcare ) using a TurboBlotter ( Whatman ) . The membrane was UV autocrosslinked at 120 mJ/cm2 ( Stratalinker 1800 ) . The 0 . 3-kb ApaI-EcoRI fragment of pTELO [81] was used as a template to generate a probe using the Takara Random Primer DNA Labeling Kit ( Takara ) . Logarithmic phase cells ( 25 OD600 units ) were fixed in 1% ( w/v ) formaldehyde for 25 min at room temperature . The reaction was stopped by adding 2 . 5 M glycine to a final concentration of 125 mM and incubate for 5 min at room temperature while shaking . Cells were washed in ice-cold Tris-buffered saline ( TBS ) solution . Cell pellet was resuspended in 0 . 4 mL of FA lysis buffer ( 50 mM HEPES , pH 7 . 6 , 150 mM NaCl , 1 mM EDTA , 0 . 1% sodium deoxycholate , 1% Triton X-100 , 0 . 1% SDS ) , supplemented with 1 mM PMSF , 20 mM N-ethylmaleimide , and Complete protease inhibitor tablet EDTA-free ( Roche ) , and lysed by beating with silica–zirconia beads four times at 5 . 0 m/s for 20 s in a FastPrep-24 ( MP Biochemicals ) . After clarification by centrifugation for 10 min at 16 , 000 x g in a microfuge at 4°C , pellet was re-suspended in 0 . 3 mL of the supplemented FA lysis buffer . After sonication for 30 s ( 30 s on , 30 s off ) in 16 cycles using a Bioruptor Pico ( Diagenode ) , the average size of sheared chromatin was less than ~300 bp , and subsequently cleared of insoluble cell debris by centrifugation at 16 , 000 x g for 10 min . After adding FA lysis buffer to bring the total volume to 1 mL , 10 μL of the clarified extract was saved as an input control . Immunoprecipitation was performed at 4°C for 2 h with protein G magnetic Dynabeads ( Thermofisher ) that had been coupled with α-SUMO antibody ( 1 μg per sample ) . Reverse crosslinking was performed as described [81] . Recovered DNA from input or ChIP was used as template for SYBR Green-based real-time PCR ( Bio-Rad ) . Fold enrichment values were calculated based on ΔCt between ChIP and input using the jk380/jk381 primer pairs of subtelomere [81] and an act1 ( actin ) gene fragment as background control ( Omn228 , Omn229 primers , Table 2 ) . The values were expressed as ChIP/input ( subtelomere ) normalized with ChIP/input ( act1+ ) as described [82] . | Since its discovery in 2007 , SUMO-targeted ubiquitin ligase ( STUbL ) activity has been identified as a key regulator of diverse cellular processes such as DNA repair , mitosis and DNA replication . In each of these processes , STUbL has been shown to promote the chromatin extraction and/or degradation of SUMO chain modified proteins . However , it remains unclear whether STUbL acts as part of a "programmed" cascade to remove specific proteins , or antagonizes localized SUMO chain formation that otherwise impedes each process . Here we determine that SUMO chains , the major recruitment signal for STUbL , are largely dispensable for genotoxin resistance in fission yeast . Moreover , when SUMO chain formation is compromised , the need for STUbL activity in DNA repair is strongly reduced . These results indicate a primary role for STUbL in antagonizing localized SUMO chain formation . Interestingly , we also find that STUbL activity can be toxic at certain genomic lesions that induce extensive local SUMOylation . For example , STUbL promotes the chromosome instability and cell death caused by deprotected telomeres following Taz1TRF1/2 deletion . Together , our data suggest that STUbL limits DNA repair-inhibitory SUMO chain formation , and depending on the nature of the genomic lesion , can either suppress or cause genome instability . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"chemical",
"compounds",
"chromosome",
"structure",
"and",
"function",
"organic",
"compounds",
"telomeres",
"fungi",
"model",
"organisms",
"sumoylation",
"experimental",
"organism",
"systems",
"basic",
"amino",
"acids",
"amino",
"acids",
"protein",
"structure",
"dna",
... | 2017 | SUMO-targeted ubiquitin ligase activity can either suppress or promote genome instability, depending on the nature of the DNA lesion |
Monocyte dysfunction by filarial antigens has been a major mechanism underlying immune evasion following hyporesponsiveness during patent lymphatic filariasis . Recent studies have initiated a paradigm shift to comprehend the immunological interactions of Wolbachia and its antigens in inflammation , apoptosis , lymphocyte anergy , etc . Here we showed that recombinant Wolbachia heat shock protein 60 ( rWmhsp60 ) interacts with TLR-4 and induces apoptosis in monocytes of endemic normal but not in chronic patients . Higher levels of reactive oxygen species ( ROS ) induced after TLR-4 stimulation resulted in loss of mitochondrial membrane potential and caspase cascade activation , which are the plausible reason for apoptosis . Furthermore , release in ROS owing to TLR-4 signaling resulted in the activation of NF-κB p65 nuclear translocation which leads to inflammation and apoptosis via TNF receptor pathway following the increase in IL-6 and TNF-α level . Here for the first time , we report that in addition to apoptosis , rWmhsp60 antigen in filarial pathogenesis also induces molecular senescence in monocytes . Targeting TLR-4 , therefore , presents a promising candidate for treating rWmhsp60-induced apoptosis and senescence . Strikingly , induction of autophagy by rapamycin detains TLR-4 in late endosomes and subverts TLR-4-rWmhsp60 interaction , thus protecting TLR-4–mediated apoptosis and senescence . Furthermore , rapamycin-induced monocytes were unresponsive to rWmhsp60 , and activated lymphocytes following PHA stimulation . This study demonstrates that autophagy mediates the degradation of TLR-4 signaling and protects monocytes from rWmhsp60 induced apoptosis and senescence .
Lymphatic filariasis is a debilitating parasitic infection with nematodes Wuchereria bancrofti and Brugia malayi , associated with varied clinical outcomes , such as lymphedema , hydrocele , or elephantiasis [1] . Chronicity by such metazoan parasites often foists crippling morbidity and incapacitating disability with profound economic , social and political consequences [2] . Abundant data from the literatures suggest that immunopathological reactions that play a major role in the development of filarial infection result from the cumulative effects of inflammation evoked mainly by the microfilarial stage of worms that invade tissues [3] . The most enthralling aspect in disease pathogenesis is individuals with subclinical conditions have compromised antigen-specific T-cell responsiveness with diminished proliferation and IFN-γ expression in response to antigen stimuli [4] and altered functions of antigen-presenting cells ( APCs ) [5–7] . Antigen-mediated T regulatory mechanisms are supposed to be potentiated majorly by the APCs . Impaired production of regulatory cytokines ( IL-12 and IL-10 ) by APCs [6] , monocyte dysfunction [8] and apoptosis [9] has been reported earlier during filarial infection . Furthermore , monocyte function of the patients is diminished based on adherence , spreading and phagocytic efficiencies in response to a bacterial stimulus [10] . Hence , failure of APCs function has been implicated to underlie this T-cell unresponsiveness . Filarial parasites rely on endogenous Wolbachia for embryogenesis , growth , survival and also contribute to pathogenesis of filarial disease [11] . Outbreak of disease pathogenesis after host inflammatory response provoked by death of the parasite and severe systemic inflammation following chemotherapy are attributed to the release of Wolbachia into the circulation [12] . Few reports suggest that Brugia malayi and Onchocerca volvulus extracts induce inflammation while the parasites that cause rodent filariasis devoid of Wolbachia failed to induce inflammation [13] . Also extracts from Wolbachia found to replicate these inflammatory effects [13] . These reports imply the crucial role of Wolbachia exposure in immune responses origination and the development of filarial pathology . Similarly , adverse reactions during microfilaricidal treatment has been associated with increase in inflammatory IL-6 and TNF-α [12] production by APCs where toll like receptor-4 ( TLR-4 ) signaling appeared to be operative [14] . Hence , the possibility that the Wolbachia antigens such as LPS , surface protein , heat shock protein 60 ( hsp60 ) , CpG motifs and peptidoglycan may play a role in immune regulation that necessitates investigation . Apart from its ascribed primary function as intracellular molecular chaperone , heat shock proteins also elicit a potent pro-inflammatory response and , therefore , has been proposed as a danger signal of stressed or damaged cells [15–17] . Both human and bacterial hsps are found to stimulate and regulate innate and acquired immune responses during pathogenesis that leads to severe autoimmune disorders [18] and chronic inflammation [19 , 20] . In this perspective , Wolbachia hsp60 has been shown to evoke IgG1 antibody response in chronic patients [21] . In addition , we earlier have shown that Wolbachia hsp60 induces pro-inflammatory cytokine production and apoptosis in monocytes [9] and T-cell unresponsiveness , [22] a hallmark status during filarial infection . Until now , endotoxin-like molecules , present in the nematode extracts , were thought to be the major inducers of these responses [23] . A recent study has found the striking similarity of innate immune responses to human hsp60 and LPS [24 , 25] . Few reports even suggest that like LPS , hsp60 can interact with TLR-2 and TLR-4 [26] and provoke inflammation and apoptosis . Similar observations were also documented in filarial conditions , where filarial and Wolbachial extract exhibit interaction with TLR-2 and TLR-4 [23] . Despite considerable evidence that Wmhsp60 can evoke pronounced innate immune response , the mechanisms and the pathways responsible remains unknown . In the present communication , we demonstrate the molecular mechanism that underlies rWmhsp60 induced monocyte dysfunction and T-cell unresponsiveness , more importantly , the involvement of TLR-4 signaling . In addition , the present study has highlighted rapamycin , a clinically approved drug-induced autophagy as one of the inherent methodology to subvert inflammation provoked by monocyte apoptosis and senescence by limiting rWmhsp60–TLR-4 interaction . This strategy might even strengthen the current Mass Drug Administration ( MDA ) program by using rapamycin synergistically with diethylcarbamazine citrate ( DEC ) .
Ten asymptomatic amicrofilaremic endemic normal ( EN ) and five individuals with active infection and/or chronic lymphatic pathology ( CP ) were included in this study . Standardized histories were obtained and physical examinations were done on all the participants in and around Chennai , India , an area endemic for W . bancrofti infection . Patients were recruited through Apollo Hospitals , Chennai , India , after obtaining informed consent with protocols approved by the Institutional Review Board of the Anna University . All the individuals were screened for the presence of circulating filarial antigens by Og4C3 mAb ELISA , a marker of W . bancrofti infection and adult worm burden [27] ( TropBio , Townsville , Australia ) ( Table 1 ) . Peripheral blood mononuclear cells ( PBMCs ) were isolated from the study population by density gradient centrifugation with Pancoll from Pan-Biotech ( Aidenbach , Germany ) . Cells were washed in RPMI 1640 medium , containing HEPES 25 mM and 80 mg gentamicin for 10 min at 1200 rpm and resuspended in the medium supplemented with 10% fetal calf serum ( Pan-Biotech ) . Monocytes were then purified from the upper interface of a hypotonic Percoll density gradient ( 1 . 129g/mL ) . Purified monocytes were resuspended in RPMI 1640 medium and the purity was found to be at least 90% as assessed by fluorescent microscopy using FITC-conjugated antihuman CD14 antibody . Finally , PBMCs or purified monocytes were stimulated with 100 ng/ml of LPS ( Sigma ) or 10 mM of cyclohexamide or 5μg of rWmhsp60 for 24 h and cultured in a humidified 5% CO2 incubator at 37°C . Recombinant Wolbachia heat shock protein 60 ( rWmhsp60 ) was expressed and purified as described previously [21] . In brief , Wolbachia hsp60 gene was PCR amplified from B . malayi genomic DNA and cloned in pRSET-A vector for expression of the recombinant protein . Recombinant plasmid was then transformed into BL21 ( DE3 ) Escherichia coli host , and the expression was induced using 1 mM IPTG , followed by purification using immobilized metal affinity chromatography . Assessment of endotoxin contamination was done using a Limulus amoebocyte lysate assay , which showed <1 pg of LPS/10 mg of protein . Dose–response study was carried out with normal healthy volunteers to determine the optimum cytotoxic dose of rwmhsp60 using 3-[4 , 5-dimethylthiazol-2-yl]-2 , 5-diphenyl tetrazolium bromide ( MTT ) from Life Technologies ( Waltham , Massachusetts , USA ) . PBMCs and monocytes were stimulated with various concentrations of rWmhsp60 ( 1 , 2 , 5 , 10 and 20 μg/ml ) for 48 h . After treatment , cells were incubated with MTT ( 10 μl , 5 mg/mL ) at 37°C for 4 h and then with DMSO at room temperature for 1 h . The plates were read at 490 nm on a scanning multiwell spectrophotometer . As the optimum concentration was found to be 5 μg/ml of rWmhsp60 , MTT assay was again performed in the presence and absence of polymyxin B sulphate , to check the endotoxin contamination . PBMCs ( 106/ml ) from EN and CP were seeded in 24-well culture plates , stimulated with rWmhsp60 and 100 mM Cyclohexamide ( CHX ) from Sigma ( St . Louis , USA ) and incubated for 24 h . CHX-treated cells served as positive control . After 24 h , the cells were then pelleted at 4°C ( 500×g ) for 5 min , washed twice with PBS , and resuspended in 100 ml of 1× annexin-V–binding buffer ( 0 . 1 M Hepes/NaOH at pH 7 . 4 , 1 . 4 M NaCl , 25 mM CaCl2 ) with 2 mg/ml of annexin-V-FITC ( BD Biosciences , San Jose , CA , USA ) and 2 mg/ml of propidium iodide ( PI ) ( Merck , NJ , USA ) . Finally , cells were washed and resuspended in 2× PBS , acquired on a Becton Dickinson FACS Calibur ( BD Biosciences ) and analyzed using Cell Quest software ( BD Biosciences ) . For acquisition , forward and side scatter gates were adjusted to acquire monocytes and lymphocytes separately , and the data analysis was performed using FlowJo software ( Tree star , San Carlo , CA ) . Caspase-3 activity was measured in isolated monocytes of EN following rWmhsp60 and CHX treatment . The activity of caspase-3 was calculated from cleavage of the fluorogenic substrate AC-DEVD-AMC from EMD Millipore ( Billerica , Massachusetts , USA ) . After 24 h incubation , rWmhsp60- and CHX-treated cell lysates were incubated with substrate solution ( caspase-3 substrate AC-DEVD-AMC 20 mg/mL , HEPES 20 mM , glycerol 10% , dithiothreitol 2 mM , pH 7 . 5 ) for 1 h at 37°C , and cleavage of the caspase-3 substrate was measured at an excitation wavelength of 390 nm and an emission wavelength of 460 nm . Activity was expressed as percentage fold increase in relative fluorescence unit ( RFU ) . For analysis of caspase cascade and autophagic events , total cell lysates were prepared as reported previously [28] . Briefly , cells were treated with rWmhsp60 , CHX for apoptosis experiments , and with rWmhsp60 in the presence and absence of rapamycin for autophagy experiments . After 24 h incubation , cells were lysed with 50 mM Tris–HCl ( pH 8 . 0 ) , 150 mM NaCl , 1% Triton X-100 , and 100 μg/mL phenylmethylsulfonyl fluoride and 1 μg/mL aprotinin . Proteins were separated by SDS/PAGE ( 15% polyacrylamide ) , transferred to a nitrocellulose membrane , blocked with 5% milk powder in TBST ( 50 mM Tris , 150 mM NaCl and 0 . 05% Tween 20 , pH 7 . 4 ) and probed with the appropriate antibody . Blots were developed by the NBT/BCIP substrate ( nitrobluetetrazolium salt/5-bromo 4-chloro 3-indolyl phosphate; USB Cleveland , OH , USA ) . For NF-κB experiments , monocytes were pre-incubated in the presence and absence of N-acetyl cysteine ( NAC ) ( Sigma ) for 30 min before the addition of rWmhsp60 or LPS . Cells were harvested after 24 h , and the remaining steps are carried out as described earlier . To visualize total intracellular levels of ROS and mROS superoxide , immunofluorescence assay was performed in monocytes of EN . Log-phase cells were grown on 24-well plates and treated with rWmhsp60 or CHX as indicated . The culture medium was removed , and the cells were washed with PBS and incubated with CM-H2DCFDA from Life Technologies ( to measure the total cellular H2O2 levels ) Life Technologies at a final concentration of 2 . 5 mM and/or MitoSOX ( to measure the mitochondrial superoxide levels ) ( Life Technologies ) at a final concentration of 5 mM in serum-free RPMI1640 for 30 min at 37°C . For immunofluorescence assay , the cells were mounted with ProLong Gold Antifade Reagent with DAPI ( Life Technologies ) , and images were acquired using a LSM-710 confocal microscope ( Carl Zeiss , Jena , Germany ) . For fluorescent spectrometer analysis , cells were treated with DCF2-DA , digested , and analyzed using 495 nm excitation and 527 nm emission filters . Mitochondrial membrane potential changes were determined by the uptake of tetramethylrhodamine , ethyl ester ( TMRE ) ( Life Technologies ) fluorescence . After treatment , monocytes were harvested and incubated with 1 μM TMRE at 37°C for 15 min in the dark . Cells were washed and resuspended in PBS and analyzed using 550-nm emission filter in UV fluorescent spectrometer . After 24 h treatment with rWmhsp60 and CHX , monocytes were lysed , and the total RNA was extracted according to the RNeasy Mini kit manufacturer’s protocol ( Qiagen , New Delhi , India ) . RNA was dissolved in 20 ml of RNase-free water . Reverse transcription of RNA was performed in a final volume of 40 ml containing 0 . 25 mM mix of the 4 deoxynucleotide triphosphates ( dATP , dGTP , dTTP and dCTP ) ( New England Biolabs , MA , USA ) ; 1 reverse transcriptase buffer ( 50 mM Tris–HCl , pH 8 . 3 , 75 mM KCl , 3 mM MgCl2 ) ; 8 mM DTT; 20 U RNase inhibitor ( Life Technologies ) ; and 200 U of MMLV–reverse transcriptase ( New England Biolabs ) followed by incubation of the tubes at 37°C for 60 min . The reverse transcription reaction was stopped by heating the tubes at 90°C for 5 min . The cDNAs were snap chilled in ice for 5–10 min and stored at 20°C until use . Real-time quantitative RT-PCR was performed in an ABI 7500 sequence detection system ( Life Technologies ) using TaqMan Assays on Demand reagents for Bcl-2 , Bax , Bid , Bad , GAPDH , and an endogenous 18S ribosomal RNA control . The end point used in real-time PCR quantification is CT that is the threshold cycle during the exponential phase of amplification , according to the manufacturer’s protocol . Quantification of gene expression was performed using the comparative CT method ( Sequence Detector User Bulletin 2; Life Technologies ) and reported as the fold change relative to the housekeeping gene . To calculate the fold change , the CT of the housekeeping gene ( 18S rRNA ) was subtracted from the CT of the target gene to yield the ΔCT . Change in the expression of the target gene as a result of antigenic exposure was expressed as 2-ΔΔCT , where ΔΔCT = ΔCT of stimulated − ΔCT of unstimulated . Along with fold change , basal level expression of the same genes was also assessed . Monocytes from EN were plated on the cover slips and cultured at up to 50–60% confluence . After treatment , cells were washed with PBS , and fixed with fresh 4% paraformaldehyde solution for 15 min at room temperature . Cells were then washed twice with PBS , followed by incubation in 10% normal rabbit serum–blocking solution for 20 min at room temperature in a humidified chamber . Cells were incubated with specific primary antibodies against NF-κB p65 ( Cell Signaling Technologies , Danvers , MA , USA ) for 2 h at room temperature in a humidified chamber . Cells were washed 3 times in PBS and incubated with Alexa Fluor 488–conjugated goat anti-rabbit IgG ( Life Technologies ) for 45 min at room temperature in a humidified chamber . The cells were then washed in PBS , mounted with ProLong Gold Antifade Reagent with DAPI ( Life Technologies ) . Images were acquired using Carl Zeiss LSM-710 confocal microscope with 10 fields of view . Above-mentioned protocol is also followed for TLR-4 surface expression studies with TLR-4 ( Life Technologies ) and LC-3 ( Cell signaling Technologies ) specific antibodies . Monocyte culture supernatants for cytokine profiling was recovered after 24 h stimulation and kept frozen until batch analysis . The levels of cytokines IL-6 and TNF-α in the culture supernatants were measured conventional by ELISA following manufacturer’s protocol ( Life Technologies ) . The results were expressed as fold change over control . The concentration of interferon-γ was determined using an IFN-γ Quantikine enzyme-linked immunosorbent assay kit ( R&D Systems , Minneapolis , MN , USA ) according to the manufacturer’s instructions . The crystal structures of human TLR-2-TLR-1 ( 2Z7X ) complex , TLR-4 ( 3FXI ) and TLR-9 ( 4QDH ) were acquired from the Protein Data Bank . Wolbachia hsp60 structure was obtained from The Protein Model Portal database designed with E . coli GroEL ( 1FYA ) as template . TLR and Wmhsp60 structures were submitted as receptor and ligand , respectively , to HEX and ClusPro , protein–protein docking servers for complex prediction . The analysis was conducted using a default parameter . Interprosurf was further used for refinement and global energy values scoring , and the models with lowest energy were selected for further analysis . UCF Chimera software was used to plot predicted intermolecular interactions . After 24 h treatment of monocytes with rWmhsp60 , cells were incubated with saturating amounts of anti-TLR2-FITC or anti-TLR4-APC or anti-TLR9-PE and isotype-matched nonbinding control mAb as per manufacturer’s protocol . All the antibodies were purchased from eBiosciences ( San Deigo , CA , USA ) . Cells were then acquired on a FACSCalibur ( BD Biosciences ) and analyzed using Kaluza software . For blocking experiments , monocytes were cultured for 2 h , in the presence of anti-TLR-2 or TLR-4 or TLR-9 antibody or rapamycin ( 50 nM ) . After this time , cells were treated with rWmhsp60 for additional 24 h , and monocyte apoptosis was analyzed in FACS Calibur using annexin-V/PI staining technique , described previously . Monocytes ( 5×105 /well ) were cultured in 24-well culture plates . After 24 h of incubation with rWmhsp60 or rapamycin or both , the cells were washed with 2× PBS and incubated with 0 . 05 mmol/L monodansylcadverine ( MDC ) ( Sigma ) at 37°C for 1 h , and the change in fluorescence was observed by Eclipse TS100 fluorescence microscope from Nikon ( Tokyo , Japan ) at an excitation wave length 380 nm with an emission filter of 525 nm . Sa-β-gal activity was detected as previously described [29] . Cells were washed once with PBS ( pH 7 . 2 ) , fixed with 0 . 5% glutaraldehyde ( PBS , pH 7 . 2 ) , and washed in PBS supplemented with 1 mM MgCl2 . Cells were stained in X-gal solution ( 1 mg/ml ) , 0 . 12 mM K3Fe [CN] 6 , 0 . 12 mM K4Fe [CN] 6 , 1 mM MgCl2 in PBS at pH 6 . 0 ) overnight at 37°C and observed using Nikon eclipse TS100 phase contrast microscope . For , co-culture experiments , the monocytes were isolated as described earlier and plated out at 1 × 106 cells/well in 24-welled plates as triplicates , treated with rapamycin for 2 h , and then washed with PBS . Furthermore , these monocytes were subjected to rWmhsp60 stimulation for 24 h . The monocytes ( 104 cells/ml ) were then co-cultured with monocyte-depleted PMBC ( 106 cells/ml ) cells in a final volume of 500 μl medium in 24-well plates and stimulated with PHA , ( 5 μg/ml ) for 48 h . [3H] Thymidine from DuPont NEN ( Boston , MA ) was added to the cells , at a final concentration of 1 μCi/ml , 3 h prior to the termination of the experiment . The cells were harvested , and the cell lysates were put onto the filter ( Glass microfibre filters; Whatman International Ltd , Maidstone , England ) and were washed three times with 5% trichloracetic acid and one time with acetone on a filter . The cell pellet on the filter was dissolved in 10 ml of Biofluor ( Fisher Scientific , Pittsburgh , PA ) , and the radioactivity was determined by liquid scintillation counting . The cell growth was expressed as stimulation index . Data analyses were performed using GraphPad PRISM ( GraphPad Software , Inc . , San Diego , CA ) . Mean ± SD were used for measurements of central tendency . Comparisons were made using either the Kruskal–Wallis ( non-parametric ANOVA ) test with Dunn’s multiple comparisons ( unpaired comparisons ) or the Wilcoxon signed-rank test ( paired comparisons , “*” denotes p < 0 . 05 .
The B . malayi Wolbachia hsp60 gene was amplified from the genomic DNA of B . malayi , cloned in pRSET-A ( Life technologies ) , and the rWmhsp60 was purified by IMAC as reported previously [21] ( Fig 1A ) . Sera from patients with active microfilaria and chronic pathology reacted with rWmhsp60 in Western blot , whereas sera from asymptomatic endemic and non-endemic normal individuals did not exhibit any reactivity [21] . To investigate the effect of rWmhsp60 on PBMCs and monocytes , cells were cultured in 96-well plates at a concentration of 2 × 105 cells/ml and treated with 1 , 2 , 5 , 10 and 20 μg/ml of rWmhsp60 for 48 h . MTT assay revealed a dose-dependent increase in inhibition until 5 μg/ml of rWmhsp60 ( Fig 1B ) . Further increase in the concentration did not have any significant influence in inhibition . Cytotoxicity observed with rWmhsp60 ( 5μg/ml ) in the presence and absence of polymyxin-B sulfate indicated that the contaminating LPS did not contribute to the proliferative responses and the observed immune regulatory effect of rWmhsp60 was not due to LPS contamination . Exposure of phosphatidyl serine on the outer membrane , a key event during apoptosis , was evaluated by annexin-V-FITC staining using flow cytometric analysis . Lymphocytes and monocytes were gated separately from PBMCs based on scattering of light . In order to exclude the inner membrane staining , cells were also stained with PI . Early apoptotic cells ( E ) were annexin-V-FITC+/PI− , late apoptotic cells ( L ) were annexin-V-FITC+/PI+ and dead cells ( D ) were annexin-V-FITC−/PI+ . Although , there was marginal increase ( 7% ) in early apoptotic monocytes from CP patients , rWmhsp60 treatment , however , did not induce any significant change in early , late apoptotic and dead monocytes . In contrast , monocytes of EN exhibited an increase in late ( 18%; p = 0 . 0027 ) apoptotic and dead cells ( 23%; p = 0 . 0031 ) on rWmhsp60 stimulation compared with the untreated monocytes ( Fig 2A , 2B , 2C and 2D ) . As expected , CHX treatment resulted in elevated levels of late apoptotic and dead cells in monocytes of both EN ( 28%; p = 0 . 00317 and 30%; p = 0 . 00293 ) and CP ( 15% , p = 0 . 0154 and 22% , p = 0 . 00197 ) . Interestingly , lymphocytes did not show any signs of apoptosis in both EN and CP [9] , thus establishing the effect of rWmhsp60 as restricted to monocytes of EN . Hence , the mechanism of rWmhsp60-induced apoptosis was assessed using monocytes of EN in further experiments . Most apoptotic cell death process is associated with the activation of complete caspase cascade and p53 levels . Hence , to explore the caspase cascade activation events in monocytes of EN , rWmhsp60-treated monocyte cell lysates were subjected to immunoblot analysis with antibodies to cleaved and total Caspase-3 , Caspase-9 , PARP and p53 . Prominent increase in the cleaved forms of casapase-3 , caspase-9 and PARP compared with the total forms provide strong evidence that rWmhsp60 induces apoptosis via caspase-dependent mechanism . Further , western blot analysis exhibited increased expression of p53 on rWmhsp60 treatment ( Fig 2E and 2F ) . A similar trend was also observed with CHX treatment on monocytes of EN . As caspase-3 is a key effector molecule in apoptosis activation , we investigated its enzymatic activity using AC-DEVD-AMC caspase-3 fluorogenic substrate . rwmhsp60 treatment of monocytes from EN for 24 h markedly increased ( 2 . 3 fold; p = 0 . 0028 ) the proteolytic activity of caspase-3 as assessed using AC-DEVD-AMC fluorogenic substrate ( Fig 2G ) . Also , CHX stimulation significantly increases the caspase-3 activity in EN ( 2 . 5 fold; p = 0 . 0046 ) as analyzed using AC-DEVD-AMC substrate . Collectively , these results augment that rWmhsp60 induces apoptosis in monocytes of EN via caspase-dependent mechanism . ROS signals were known to control variety of responses including apoptosis . To determine its relationship in rwmhsp60-mediated apoptosis , we assessed the spontaneous ROS and mROS generation in monocytes of EN followed by the change in ROS production after rWmhsp60 treatment using DCF2-DA and MitoSOX staining . The results revealed that the EN monocytes had spontaneous basal levels of ROS and mROS , and rWmhsp60 and CHX stimulation increased ROS and mROS production ( Fig 3A ) as evident from the increased number DCF2-DA and MitoSOX-stained monocytes . As mitochondria play an important role in the intrinsic pathway of apoptosis , mitochondrial membrane potential was quantified using tetramethylrhodamine ethyl ester ( TMRE ) . rWmhsp60 treatment resulted in increased number of EN monocytes with depolarized mitochondria compared with the control monocytes which had the intact mitochondrial integrity ( Fig 3B ) . The fluorimetric analysis with DCF2-DA confirmed that rWmhsp60 ( 3-fold; p = 0 . 0029 ) and CHX ( 2-fold; p = 0 . 0015 ) stimulations resulted in increased ROS levels in monocytes of EN ( Fig 3C ) . Also , quantitative analysis with TMRE ( Fig 3D ) resulted in loss of mitochondrial membrane potential on rWmhsp60 stimulation ( 3-fold; p = 0 . 0085 ) . As expected , CHX stimulation increased the depolarization of mitochondria in monocytes of EN ( 2-fold; p = 0 . 0043 ) . In addition , to determine whether the increased or decreased uptake of TMRE were due to the differences in the mitochondrial density or originates from the changes in the membrane integrity , the gene expression of the intrinsic pathway markers were studied . Our analysis revealed a marked increase in pro-apoptotic genes ( Fig 3E , 3F and 3G ) bax ( 8-fold; p = 0 . 0057 ) , bad ( 2 . 5-fold; p = 0 . 0039 ) and bid ( 5-fold; p = 0 . 0028 ) , respectively . Also , relatively decreased expression of anti-apoptotic gene ( Fig 3H ) , bcl-2 ( 3 fold; p = 0 . 0037 ) , was observed on rWmhsp60 stimulation . A similar trend was observed with CHX treatment , bax 10-fold ( p = 0 . 0017 ) , bid 3-fold ( p = 0 . 0034 ) , bad 13-fold ( p = 0 . 0019 ) and bcl-2 0 . 8-fold ( p = 0 . 0043 ) . In conclusion , these results suggest that increase in ROS , mROS and loss in mitochondrial membrane integrity may contribute to rWmhsp60-mediated apoptosis . To examine the effect of rWmhsp60-mediated ROS on nuclear translocation of NF-κB , the immunofluorescence staining of NF-κB-p65 was performed using confocal microscopy to provide the exact location of NF-κB in rWmhsp60 and LPS-treated and-untreated monocytes . Nuclear translocation of NF-κB was observed in EN monocytes treated with rWmhsp60 and LPS ( Fig 4A ) . In contrast , NF-κB resides predominantly in the cytoplasm of the untreated monocytes and , thus , confirms the activation and translocation of NF-κB on rWmhsp60 stimulation . Further to examine the effect of ROS generation on NF-κB , Western blot analysis was performed following concurrent incubation of monocytes with NAC for 1 h . Increase in the NF-κB levels were observed on rWmsp60 treatment , whereas NAC greatly inhibited NF-κB production ( Fig 4B ) . These data suggest that ROS influence NF-κB transcriptional activation following rWmhsp60 stimulation . Expression of TNF-α and IL-6 was regulated by transcription factor , NF-κB . To determine whether rwmhsp60 can modulate TNF-α and IL-6 production , culture supernatants of rWmhsp60-treated and untreated monocytes were assayed for TNF-α and IL-6 protein levels by ELISA . As expected , rWmhsp60 stimulation increased the production of IL-6 ( 4 . 2-fold; p = 0 . 005 ) ( Fig 4C ) and TNF-α ( 3 . 7-fold; p = 0 . 029 ) ( Fig 4D ) comparable with LPS-induced IL-6 ( 3 fold; p = 0 . 0061 ) and TNF-α ( 3 . 5 fold; p = 0 . 0089 ) production . Taken together , these results suggest that ROS and mROS mediate NF-κB translocation into nucleus and initiate the production of pro-inflammatory cytokines TNF-α and IL-6 . To determine the mechanism by which rWmhsp60 activates the apoptotic cascade via ROS , we examined the possible role of TLRs by docking Wmhsp60 with TLR-2 , TLR-4 and TLR-9 using HEX and Cluster pro-docking server . We predicted the structure of Wmhsp60 using homology-modeling approach with SWISS-PDB . TLR-2 , TLR-4 , and TLR-9 3D structures were obtained from RCSB . As a result , Wmhsp60 fits precisely with the binding moiety of TLR-4 with the lowest binding energy of about −440kcal ( HEX ) and −882 . 6 kcal ( Cluspro ) and not with TLR-2 and TLR-9 ( Fig 5A ) . Also , non-availability of the complete structure of TLR-2 and TLR-9 was a great hindrance to make further assessment . In addition , the surface expressions of TLR-2 , TLR-4 and TLR-9 on rWmhsp60 stimulation were assessed with fluorescent-tagged antibodies using flow cytometer . As shown in the Fig 5B , exposure of human monocytes for 2 h to rWmhsp60 dramatically increased the surface expression of TLR-4 and not the TLR-2 and TLR-9 . Furthermore , increased TLR-4 surface expression in the rWmhsp60-treated monocytes was supported and confirmed by confocal microscopy studies ( Fig 5C ) . To confirm that rWmhsp60 was , indeed , signaling through TLR-4 , we pre-treated monocytes with anti-TLR-4 monoclonal antibodies , 2 h before stimulation with rWmhsp60 . Anti-TLR-4 pretreatment significantly reduced the late apoptotic ( 6%; p = 0 . 0075 ) and dead cells ( 6 . 5%; p = 0 . 0084 ) compared with rWmhsp60 treatment that exhibited increased late ( 18%; p = 0 . 0058 ) apoptotic and dead cells ( 23%; p = 0 . 0029 ) ( Fig 5D and 5E ) . Collectively , these experiments synergistically augment that rWmhsp60-induced monocyte apoptosis is triggered via TLR-4 signaling pathway . In attempt to characterize the mechanism by which rWmhsp60 induces cell death other than apoptosis , we have examined for autophagy and senescence in rWmhsp60-treated monocytes . To detect autophagy , rWmhsp60 stimulated monocytes were subjected to MDC staining . Rapamycin-treated monocytes that served as a positive control exhibited high intensities ( 75%; p = 0 . 0015 ) of MDC staining ( Fig 6A and 6B ) , whereas rWmhsp60 stimulated monocytes that failed to uptake MDC and remained as similar to control monocytes , suggesting that rWmhsp60 does not induce autophagy in monocytes . Senescence was characterized as SA-β-Gal–positive cells . Increase in SA-β-Gal–positive cells ( 85%; p = 0 . 0019 ) were detected in response to rWmhsp60 treatment ( Fig 6C and 6D ) versus 2% in control cells . As expected , rapamycin does not show any significant increase in SA-β-Gal–positive cells ( 8% ) . This confirms that rWmhsp60 induces monocyte senescence in addition to apoptosis . In order to reduce the p53-mediated apoptosis and senescence-induced inflammation in rWmhsp60-stimulated monocytes , we hypothesized to induce p53-mediated autophagy using rapamycin prior to rWmhsp60 treatment . Also , rapamycin is known to sequester and degrade the pro-inflammatory TLR signaling complex in autophagy-dependent manner . Rapamycin-treated monocytes served as a positive control that exhibited high intensities ( 75%; p = 0 . 0075 ) of MDC staining , whereas rWmhsp60-stimulated monocytes failed to uptake MDC and remained as similar to control monocytes ( Fig 7A and 7B ) . In addition , rWmhsp60 stimulation of monocytes post-rapamycin treatment ( 82% ) does not influence any change in MDC staining compared with rapamycin-treated monocytes . This clearly suggests that rapamycin-induced autophagosome formation in monocytes was not altered on rWmhsp60 stimulation . In addition , we investigate the localization of TLR-4 on rapamycin and rWmhsp60 treatment using Confocal microscopy . Indeed , rWmhsp60 stimulation does not induce vesicular localization of TLR-4 and LC3 , whereas rWmhsp60 stimulation following rapamycin pretreatment exhibited co-localization of LC3 with TLR-4 . Also , it is evident from the Fig 7C that rapamycin pretreatment sequesters TLR-4 into endosomal vesicle and degrades the TLR signaling cascade . As a result , rWmhsp60 fails to activate TLR signaling cascade following rapamycin treatment . Similarly , Western blot analysis revealed that the expression of mammalian target of rapamycin ( m-TOR ) by rWmhsp60 was abrogated by rapamycin pretreatment ( Fig 7D ) . Furthermore , expression of LC-3 an autophagy marker following rapamycin treatment was not altered by rWmhsp60 . Thus , this study provides the evidence of involvement of autophagy in rapamycin-induced hyporesponsiveness to rWmhsp60 and is mediated by degradation of TLR-4 . Furthermore , to analyze , rapamycin pretreatment protects rWmhsp60-induced apoptosis in EN monocytes , annexin-V/PI analysis was performed ( Fig 7E ) . As expected , EN monocytes exhibited increased late ( 18%; p = 0 . 0027 ) apoptotic and dead cells ( 23%; p = 0 . 0031 ) on rWmhsp60 stimulation . In contrast , rapamycin treatment prior to rWmhsp60 stimulation significantly decreased the late ( 6 . 5%; p = 0 . 0017 ) apoptotic and dead ( 7 . 5%; p = 0 . 0025 ) cells . Thus , rapamycin significantly reduced the rWmhsp60-induced apoptosis by 27% ( annexin V , annexin V + PI and PI-positive cells; Fig 7F ) . Similarly , increase in SA-β-Gal–positive cells ( 78%; p = 0 . 0026 ) were detected in response to rWmhsp60 , whereas rapamycin pretreatment decreased the SA-β-Gal–positive cells to 10% ( p = 0 . 086; Fig 8A and 8B ) . Further , to determine whether rapamycin can modulate rWmhsp60-induced pro-inflammatory TNF-α and IL-6 production , culture supernatants of EN monocytes treated with rapamycin and/or rWmhsp60 were assayed for TNF-α and IL-6 protein levels by ELISA . As expected , rWmhsp60 stimulation resulted in perceptible increase in the production of IL-6 ( 4 . 2 fold; p = 0 . 005 ) and TNF-α ( 3 . 7 fold; p = 0 . 029 ) ( Fig 8C and 8D ) . In contrast , rWmhsp60-stimulated monocytes post-rapamycin treatment significantly decreased TNF-α and IL-6 protein levels comparable with the control . Collectively , these results demonstrate that autophagy protects monocytes from rWmhsp60-induced apoptosis and senescence by the induction of autophagy . Finally , MDC and β-Gal staining revealed that rapamycin also induces differentiation of monocytes . Finally , we investigated whether rapamycin treatment on monocytes can activate lymphocytic cells . As expected , co-culturing of monocytes with monocyte-depleted PBMCs and rapamycin-treated monocytes with monocyte-depleted PBMCs resulted in the similar levels of proliferation to PHA stimulation . However , co-culturing of rWmhsp60-treated monocytes abrogated the PHA-induced proliferation of monocyte-depleted PBMCs . Nevertheless , rapamycin pretreatment to the monocytes reversed this effect and showed a 9-fold ( p = 0 . 019 ) increase in the proliferation following PHA stimulation ( Fig 9A ) . Furthermore , increase in proliferation was substantiated with increase in IFN-γ levels ( Fig 9B ) in the culture supernatants of the above-mentioned stimulated cultures . These results suggest that rWmhsp60 downregulates T-cell activation by monocytes and rapamycin pretreatment abrogated this effect .
Wolbachia in filariae is gaining importance as it is implicated with inflammatory responses and pathogenesis of filarial infections [30] . More precisely , adverse inflammatory reactions and diminished APC function , attributed during disease conditions , are also observed with Wolbachial antigens [11–13] . This could be one among many existing explanation for T-cell hyporesponsiveness , a hallmark status in filarial pathogenesis . However , till date , there is dearth of proper evidence in its support to unravel the probable potentials of Wolbachial antigens in filarial pathogenesis . Consequently , sensing the gravity , we earlier demonstrated that Wolbachial antigens ( hsp60 and WSP ) induce T-cell suppression and monocyte apoptosis in normal healthy population [9 , 22] . Similar T-cell anergy [4] and impaired monocyte functions [6–8] were reported earlier in filarial patients; however , the molecular mechanisms underlying these functions remain elusive . Hence , the present study was particularly designed to account for this lacuna by investigating Wolbachial hsp60-modulated apoptosis in endemic normals and CPs . Flow cytometric analysis with annexin-V and PI has revealed that monocytes of EN were more susceptible to rWmhsp60-induced apoptosis , whereas lymphocytes of EN and CP and monocytes of CP exerted resistance to rWmhsp60-induced apoptosis , thus establishing its specificity to monocytes of EN . Possible reason for this differential susceptibility between monocytes of EN and CP could be due to the binding ability of hsp60 to TLR [31] and suppressed expression of TLRs in CP [32] . Resistance offered by lymphocytes of EN and CP may be due to lesser expression of TLRs in lymphocytes in comparison with monocytes [33 , 34] . Similar monocyte apoptosis mediated by filarial antigens was previously reported [35 , 9] . This prompted us to elucidate the mechanism of rWmhsp60-mediated apoptosis in monocytes of EN . Recent studies with Wolbachial surface antigen by Brattig and studies with other hsps limited our search for hsp’s surface receptors to TLR-2 , -4 and -9 [23] . In this report , we showed that rWmhsp60 can interact with TLR-4 , by the upregulation of TLR-4 but not TLR-2 and TLR-9 surface expression , confirmed by flow cytometric analysis and confocal microscopy . In addition , blocking experiments with anti-TLR4 antibodies inhibited apoptosis , proving a sharp note that rWmhsp60 induces apoptosis via TLR-4 . Furthermore , docking studies with TLR-4 also provided evidence that Wmhsp60 can directly bind to the TLR-4 . These results are in broad agreement with the demonstration of TLR-4 as a receptor for human hsp60 [31] . Apart from identifying receptor for Wmhsp60 , the present study to acquire insights into molecular events in activation pathway of monocyte will be of crucial interest since the nature of the ligand determines the phenotype of activation . There are enormous previous evidences suggesting that ROS can act as a secondary messenger [36] and control various signaling molecules downstream of TLR [37] . Increased ROS and mROS levels and altered mitochondrial bioenergetics were observed in apoptotic monocytes as a result of Wmhsp60–TLR-4 interaction . Apoptosis mediated by mitochondrial membrane potential loss was regulated and controlled by Bcl-2 family of pro-apoptotic Bax , Bid and Bad and anti-apoptotic Bcl-2 markers . An elevation in the levels of pro-apoptotic markers relative to concomitant Bcl-2 expression in rWmhsp60-treated monocytes confirms that apoptosis is through mitochondria-mediated pathway . Similarly , activation of caspases , the major downstream events following mitochondrial membrane potential loss , is responsible for many of the molecular changes in the cell undergoing apoptosis [38] . Activation of caspase cascade and PARP degradation by rWmhsp60 provided the strong evidence that the cell death is initiated via mitochondrial disruption pathway and caspase cascade activation . There are numerous reports supporting this mechanism . Furthermore , there are few evidences on ROS contribution to translocation and transcriptional activation of NF-κB [39] . In our study , rWmhsp60 stimulation led to the translocation of NF-κB into the nucleus of monocytes . In addition , we found that rWmhsp60-induced consistent expression pattern of NF-κB was abrogated on NAC treatment , which suggests the role of ROS in NF-κB production and translocation . Furthermore , NF-κB translocation resulted in transcriptional activation of pro-inflammatory cytokines TNF-α and IL-6 . Excessive release of TNF-α may further lead to apoptosis via TNF-R [40] . Hence , both the death receptor and the mitochondrial pathways are likely to be involved in rWmhsp60-induced apoptosis . Altogether , these results suggest that rWmhsp60 interacts with TLR-4 and results in apoptosis mediated by ROS generation and subsequent TNF-α release . Thus , Wmhsp60-exposed monocytes undergo immune dysfunction at an early time point , reducing the window of time during which they might interact with T-cells . p53 , besides a key apoptosis regulator , is known to promote and impede autophagy and senescence by the ability to control metabolic stress through regulation of ROS and mTOR activity [41] . Elevated p53 expression confirms apoptosis and suggests that rWmhsp60 may also regulate other death mechanisms , apart from apoptosis . rWmhsp60 promotes mTOR phosphorylation , suggesting the absence of autophagy . X-gal staining and p16 expression synergistically augment that rWmhsp60 induces senescence in monocytes . This study , for the first time , reported that rWmhsp60 , antigen involved in filarial pathogenesis , induces senescence in monocytes along with apoptosis . These senescent monocytes were the additional source of inflammatory factors other than apoptotic bodies that may ameliorate the degree of chronicity [41–43] . Constitutive stress by rWmhsp60 leads to persistent p53 activity that induces apoptosis and senescence , whereas pretreatment with rapamycin may tip the balance toward autophagy , favoring anti-inflammatory effects . Rapamycin is an inhibitor of m-TOR [44] , downstream target of PI3K , and an important factor in TLR signaling [45] . Also , rapamycin was shown to block TLR-2—and TLR-4–mediated TNF-α and IL-6 production in neutrophils [46] . These studies , therefore , suggest that autophagy may act to limit TLR-mediated inflammatory conditions . Rapamycin pre-treatment sequesters TLR-4 into late endosomes which prevented TLR-4–rWmhsp60 interaction , thereby inhibiting TLR-4–mediated apoptosis and senescence . Furthermore , the lymphocyte activation study reveals that monocytes pretreated with rapamycin were unresponsive to rWmhsp60 and induces the proliferation of lymphocytes along with increase in IFN-γ levels . This study demonstrated that rapamycin pretreatment of monocytes addresses the key events during filarial pathogenesis and help monocytes and lymphocytes to uphold its function . This further corroborates with the earlier findings by others to limit excessive inflammation triggered by TLRs during sepsis and other inflammatory disorders [47] . Collectively , rWmhsp60 acts via TLR-4–dependent pathway to disrupt the maturation process of monocytes by inducing apoptosis and senescence . The resulting monocytes have functional defects that may have an impact on their ability to stimulate specific T-cell responses . This suggests that parasite may take advantage of this process to evade the immune response and establish its niche . As a counteractive measure , we propose that rapamycin pretreatment could help monocytes to retain its function even after rWmhsp60 exposure . The proactive interference of rapamycin instigates autophagy and subverts TLR-4 molecular signaling crosstalk with rWmhsp60 . This strategy undermines inflammation-mediated via apoptosis and senescence and contributes to novel therapeutic applications . As rapamycin is specific to TLR-mediated mechanism and not to rWmhsp60 , administration of rapamycin in combination with other anti-filarial drugs might help to control the adverse TLR-mediated inflammatory reactions following microfilaricidal treatment . | Despite knowing the significance of Wolbachia in helminth infections , our understanding of immunity and pathogenesis remains incomplete . Therefore , considering the gravity of the problem , the present study provides evidence that Wolbachia heat shock protein 60 induces apoptosis and senescence through TLR-4 . Also , binding of rWmhsp60 to TLR-4 triggered caspase cascade activation following , ROS-mediated mitochondrial potential loss . Moreover , we found that nuclear translocation of NF-κB p65 was predominantly related to TLR-4 expression and resulted in apoptosis- and senescence-mediated inflammation via TNF-α and IL-6 . Hence , we hypothesized that modifying TLR-4 expression may provide a plausible target for designing antiparasitic drugs . Here we have shown that induction of autophagy by rapamycin destabilizes TLR-4 expression and protects monocytes from rWmhsp60-induced apoptosis and senescence . In addition , rapamycin-induced monocytes were unresponsive to rWmhsp60 and triggered lymphocyte activation after PHA stimulation . Thus , synergistic usage of rapamycin with existing anti-filarial drugs might reduce the TLR-mediated inflammatory reactions following microfilaricidal treatment . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Autophagy Protects Monocytes from Wolbachia Heat Shock Protein 60–Induced Apoptosis and Senescence |
Leprosy reactions , reversal reactions/RR and erythema nodosum leprosum/ENL , can cause irreversible nerve damage , handicaps and deformities . The study of Mycobacterium leprae-specific serologic responses at diagnosis in the cohort of patients enrolled at the Clinical Trial for Uniform Multidrug Therapy Regimen for Leprosy Patients in Brazil/U-MDT/CT-BR is suitable to evaluate its prognostic value for the development of reactions . IgM and IgG antibody responses to PGL-I , LID-1 , ND-O-LID were evaluated by ELISA in 452 reaction-free leprosy patients at diagnosis , enrolled and monitored for the development of leprosy reactions during a total person-time of 780 , 930 person-days , i . e . 2139 . 5 person-years , with a maximum of 6 . 66 years follow-up time . Among these patients , 36% ( 160/452 ) developed reactions during follow-up: 26% ( 119/452 ) RR and 10% ( 41/452 ) had ENL . At baseline higher anti-PGL-I , anti-LID-1 and anti-ND-O-LID seropositivity rates were seen in patients who developed ENL and RR compared to reaction-free patients ( p<0 . 0001 ) . Seroreactivity in reactional and reaction-free patients was stratified by bacilloscopic index/BI categories . Among BI negative patients , higher anti-PGL-I levels were seen in RR compared to reaction-free patients ( p = 0 . 014 ) . In patients with 0<BI<3 , ( 36 RR , 36 reaction-free ) , higher antibody levels to PGL-I ( p = 0 . 014 ) and to LID-1 ( p = 0 . 035 ) were seen in RR while difference in anti-ND-O-LID positivity was borderline ( p = 0 . 052 ) . Patients with BI≥3 that developed ENL had higher levels of anti-LID-1 antibodies ( p = 0 . 028 ) compared to reaction-free patients . Anti-PGL-I serology had a limited predictive value for RR according to receiver operating curve/ROC analyses ( area-under-the-curve/AUC = 0 . 7 ) . Anti LID-1 serology at baseline showed the best performance to predict ENL ( AUC 0 . 85 ) . Overall , detection of anti-PGL-I , anti-LID-1 and anti-ND-O-LID antibodies at diagnosis , showed low sensitivity and specificity for RR prediction , indicating low applicability of serological tests for RR prognosis . On the other hand , anti-LID-1 serology at diagnosis has shown prognostic value for ENL development in BI positive patients . ClinicalTrials . gov NCT00669643
Leprosy is a complex dermato-neurologic disease caused by Mycobacterium leprae that presents a wide spectrum of clinical manifestations characterized by distinct bacteriologic , immunologic and histopathologic features [1] . On one pole , tuberculoid leprosy ( TT ) is characterized by few skin lesions , low or absent bacilloscopic index ( BI ) , strong M . leprae-specific Th1-type cell-mediated immunity ( CMI ) and low or absent specific antibodies . On the other extreme of the spectrum , lepromatous leprosy ( LL ) is characterized by multiple disseminated skin lesions , high BI , Th2-type immunity with vigorous antibody production and low or absent M . leprae-specific CMI . Additionally , immunologically unstable borderline forms ( borderline tuberculoid/BT , borderline borderline/BB and borderline lepromatous/BL ) lie in the middle of the spectrum combining features of both poles [2] . Leprosy treatment , known as multidrug therapy ( MDT ) is based on different combinations of antibiotics for paucibacillary ( PB ) and multibacillary ( MB ) leprosy , classified according to the number of skin lesions ( < = 5: PB leprosy; >5: MB leprosy ) . MB leprosy patients are prescribed with 12 monthly-supervised doses of rifampicin , dapsone and clofazimine plus self-administered daily doses of dapsone and clofazimine . PB leprosy patients are treated with six monthly-supervised doses of rifampicin and dapsone plus self-administered daily doses of dapsone [3] . In 2007 , an open-label , randomized clinical trial was designed and conducted to compare the regular MDT ( R-MDT ) proposed by WHO and a uniform MDT regimen ( U-MDT ) consisting of six doses of rifampicin , dapsone and clofazimine for PB and MB leprosy patients . Clinical monitoring is still under way in Brazil regarding mainly the development of reactions and relapses ( Clinical Trial for Uniform Multidrug Therapy Regimen for Leprosy Patients in Brazil , U-MDT/CT-BR ) [4–8] . One of the main difficulties in the clinical management of leprosy patients is the development of leprosy reactions that can occur anytime during the chronic course of the disease: before diagnosis , during treatment and even years after treatment release [2 , 9 , 10] . Leprosy reactions represent immunologically mediated episodes of acute inflammation that if not diagnosed and treated promptly can cause irreversible impairment of nerve function and permanent incapacities [11] . There are two major types of leprosy reactions: type 1 reaction ( T1R ) or reversal reaction ( RR ) which is associated with Th1-type immunity and type 2 reaction ( T2R ) represented mainly by erythema nodosum leprosum ( ENL ) which is related to Th2-type immune responses [9 , 12] . Currently , there is no laboratory test able to predict the emergence of leprosy reactions among recently diagnosed patients . Leprosy serology , comprises the well known detection of IgM antibodies against the phenolic glycolipid I ( PGL-I ) , a M . leprae specific cell-wall antigen . Since the PGL-I identification , several studies have been extensively performed to understand the use of this antigen in diagnostic tests and the immune response in leprosy , but there are still many knowledge gaps to be filled [13] . More recent IgG based tests to newly M . leprae-recombinant protein antigens have been described [14 , 15] . In leprosy , the seropositivity of IgM and IgG tests reflects the patients’ bacillary load with low positivity rates in PB patients and high positivity in MB patients [16 , 17] . After the decodification of M . leprae genome , over 200 new recombinant proteins have been screened in serology and cell mediated tests aiming the development of new diagnostic tests for leprosy [15 , 18–23] . Results from serological screenings in different endemic areas in the world have highlighted the significant reactivity of ML0405 and ML2331 proteins , which were later fused and named LID-1 antigen ( Leprosy IDRI Diagnostic-1 ) . LID-1 was shown to retain the immunogenicity of the original proteins with high positivity rates reported in MB leprosy patients from different endemic sites [18 , 24] . PGL-I and LID-1 antigens represent important specific antigens for leprosy serology and their association was suggested to improve the sensitivity of serology for MB leprosy [25] . In this regard , a single fusion complex ( ND-O-LID ) resulting from the conjugation of the synthetic mimetic disaccharide of PGL-I ( ND-O ) and LID-1 was constructed and shown to detect most MB leprosy by both a lateral flow platform and by ELISA [26 , 27] . Our results from a previous semi-quantitative serological study with the entire cohort of U-MDT/CT-BR showed that results of rapid lateral flow test ( ML Flow ) to detect IgM antibodies to PGL-I antigen at diagnosis had low sensitivity and specificity to predict the development of leprosy reactions during follow-up [28] . This previous results obtained using the ML Flow , prompted us to investigate the predictive value for leprosy reactions of the quantitative serology to new M . leprae protein antigens as LID-1 and ND-O-LID compared to the well known anti PGL-I serology . This investigation used a robust sera bank of the U-MDT/CT-BR , a well characterized cohort of leprosy patients monitored for more than 6 years regarding the development of leprosy reactions and relapses . [4 , 6–8 , 29] .
Our study group included 452 out of 753 patients who were enrolled at the U-MDT/CT-BR ( March 2007-September 2013 ) in two highly endemic areas for leprosy in Brazil ( Fortaleza , Ceará , northeast; Manaus , Amazonas , north region ) . So far , patients have been followed for a total person-time of 780 , 930 person-days , i . e . 2139 . 5 person-years , with a maximum of 6 . 66 years follow-up time [8] . At enrollment , all leprosy patients had complete dermato-neurological evaluation by clinicians with vast expertise in leprosy diagnosis . The following laboratory tests were performed at diagnosis: ML Flow rapid test , slit skin smear , histopathology of biopsies from leprosy skin lesions . For research purposes , patients were categorized according to a modified Ridley-Jopling ( R&J ) classification system considering clinical features , histopathology of skin lesions and the slit skin smear bacterial index ( BI ) . Mitsuda tests and BI of the skin lesions were not performed . In this case-control study , we have compared serology results at diagnosis from patients that developed reactions during follow-up ( RR and ENL , without other complications ) and patients that remained reaction-free during entire monitoring . From the original group of 753 patients , exclusions ( n = 301 ) were due to: unavailability of serum sample at diagnosis ( n = 46 ) ; reaction at diagnosis ( RR = 16 ) , reaction associated or not with neuritis ( n = 184 ) or other clinical manifestations such as orchitis , arthritis and lymphadenopathy ( n = 55 ) . The socio-demographic , clinical and laboratory characteristics of our study group are similar to the previously described characteristics of the entire group [30] ( S1 Table ) . In our study-group , 26% patients developed RR ( 119/452 ) , 10% had ENL ( 41/452 ) while 64% remained reaction-free ( 292/452 ) . The majority of patients was male , the age of patients who developed RR was similar compared to reaction-free patients ( median age of 43 and 41 years respectively , p = 0 . 44 ) , while patients who developed ENL were younger ( median age = 35 years ) compared to reaction-free and RR patients ( p = 0 . 003 and p = 0 . 001 , respectively ) . Most reactional patients was classified as BL leprosy ( 62% , 99/160 ) , whereas most reaction-free patients was classified as BT leprosy ( 60% , 176/292 ) . The baseline serological profiles to PGL-I , LID-1 and ND-O-LID antigens in different Ridley & Jopling groups is in agreement with previous reports [15 , 18–23] . Higher levels of IgM anti-PGL-I and IgG anti-LID-1 antibodies were detected in MB leprosy ( BB , BL , LL ) compared to PB leprosy ( TT , BT ) ( p<0 . 0001 ) . Higher positivity of IgM and IgG responses to ND-O-LID antigens was found in BL and LL patients compared to TT , BT groups ( p<0 . 0001 ) ( S1 Fig ) . Serum IgM antibodies to PGL-I were detected by ELISA performed as described previously [31] . PolySorp 96-well plates ( Nunc , Roskilde , Denmark ) were coated with 0 . 01 μg/mL of the semi-synthetic analogue of PGL-I ( NT-P-BSA; batch: Nara XVI-61 , Dr Fujiwara , Japan ) , or BSA and blocked with PBS-Tween containing 1% BSA . Serum samples diluted 1/300 in PBS-T containing 10% normal goat serum/NGS ( Sigma-Aldrich , St . Louis , USA ) were incubate and washing , horse radish peroxidase/HRP-conjugated anti-human IgM ( Immuno Chemicals , St . Louis , Missouri , USA ) was added . After incubation and washing , peroxidase color substrate ( TMB , Sigma Aldrich , St . Louis , USA ) , was added and the reaction was quenched by the addition of 2 . 5 N H2SO4 , when the optical density/OD at 450 nm from reference serum reached an OD value of 0 . 6 ( Bio-Rad microplate reader , Life Science , Hercules , CA , USA ) . The final OD was calculated by subtracting the OD values of BSA coated wells from OD values of NT-P-BSA wells . The cut-off was defined as OD > 0 . 25 as previously described [16] . Serum IgG antibodies to LID-1 ( batch: 14 November 2011 , Dr Duthie , USA ) and the single fusion complex ( ND-O-LID- batch: 17 August 2012 , Dr Duthie , USA ) were detected by ELISA . Polysorp 96 well plates ( Corning Costar , NY , USA ) were coated with 1μg/mL LID-1 or with 0 . 25 μg/mL ND-O-LID . Blocking was performed with PBS-T 1% BSA . Serum samples diluted 1/200 in PBS-T-10% NGS were added in duplicate and incubated for 2 hours at room temperature . Plates were washed and incubated for 1 hour with HRP-conjugated anti-human IgG ( Southerm Biotech , Birmingham , AL ) for anti LID-1 serology and anti-human IgG ( Southerm Biotech , Birmingham , AL ) plus anti-human IgM ( Immuno Chemicals , St . Louis , Missouri , USA ) for anti- ND-O-LID serology . After washing , reactions were developed with peroxidase color substrate ( KPL , Gaithersburg , MD , USA ) and quenched by the addition of 1 N H2SO4 . The optical density was determined ( Bio-Rad microplate reader , Life Science , Hercules , CA , USA ) at 450 nm . For anti LID-1 serology the cut-off was calculated to be 2 times the standard deviation of the OD of sera from healthy endemic controls ( EC ) , such that samples with OD > 0 . 3 were considered positive [18] . As previously described , the anti-ND-O–LID serology threshold for positive responses was considered OD > 0 . 923 [20] . The results of serologic tests were expressed as the mean OD of duplicates . The baseline serologic profiles of PB and MB patients enrolled in the U-MDT/CT-BR were analyzed according to the clinical outcomes reported during clinical follow-up , which included reactional ( RR and ENL ) and reaction-free patients . In this study group , the frequency of leprosy reactions was similar in the first year after diagnosis compared to subsequent monitoring ( p>0 . 05 ) , therefore , the consolidated data analysis , considered the predictive value of baseline serology for the development of leprosy reactions observed any time during the entire monitoring up to 6 . 66 years after MDT . Serological results were analyzed based on the patients´ bacteriological index ( BI ) which was stratified into 3 categories: negative BI , BI higher than zero and lower than 3 ( 0> BI < 3 ) and BI equal or higher than 3 ( BI≥ 3 ) as previously reported [8] . Statistical significance was assessed using the Kruskal-Wallis one-way analysis of variance for comparison of multiple groups and by the Mann-Whitney U test for comparison between two groups . Chi square ( χ2 ) was used to compare positivity rates . Receiver Operating Curves ( ROC ) were determined with 95% confidence interval ( CI ) ( GraphPad Prism , version 5 ) and results were considered statistically significant when p values <0 . 05 were obtained . The UMDT/CT-BR trial was performed considering international ( Helsinki ) and Brazilian research regulations involving human beings and approvals were obtained from all the regional research ethical committees involved and from the National Committee for Ethics in Research ( CONEP ) of the National Health Council/ Ministry of Health , Brazil ( protocol # 001/06 ) . Additionally , approved by the human and animal research ethics committee from Federal University of Goiás ( CEMHA/HC/UFG#166/2011 ) . Written informed consent was obtained from all patients prior to inclusion in the study . For patients under 18 years old , written parental consent was obtained . Data confidentiality was strictly guaranteed and all patients were informed they were free to leave the study and opt for the regular MDT treatment ( ClinicalTrials . gov identifier: NCT00669643 ) .
Higher IgM anti PGL-I seropositivity was observed in reactional patients compared to reaction-free ones ( 73% , versus 39%; p<0 . 0001 ) . Similarly , the median ODs were higher among reactional compared to reaction-free patients ( 0 . 498 versus 0 . 128 , respectively; p<0 . 0001 ) . Regarding anti-LID-1 and anti-ND-O-LID serology in reactional versus reaction-free patients , both the positivity rates and the antibody levels , measured by median OD , were higher in reactional patients: IgG antibodies to LID-1 ( 84% vs 46% positivity , median ODs = 1 . 43 and 0 . 261 respectively , p>0 . 0001 ) , IgM and IgG antibody levels to ND-O-LID antigens ( 67% vs 32% positivity , median ODs = 1 . 34 and 0 . 493 respectively , p<0 . 0001 ) ( Fig 1 ) . However , 2–6% reaction-free patients had outliers serological results for the three antigens tested . We have investigated if the time elapsed between blood intake and the development of reaction has had any effect on serology as ongoing “subclinical reactions” could have impacted the antibody titers of patients that manifested reactions within the first year of monitoring ( n = 68 ) . For these analyses we have separated patients into three groups according to the time between these two events ( diagnosis/intake and development of reactions ) : 90 days , 90–180 days and 180–365 days . These analyses showed that regardless of the time in which reactions occurred after the first year of leprosy diagnosis , patients that further developed reactions showed highest the O . D medians for all antigens evaluated compared to reaction-free patients ( S2 Fig ) . At diagnosis higher anti-PGL-I positivity rates were seen in patients who developed both ENL ( 78% , 32/41 ) and RR ( 76% , 90/119 ) ( p = 0 . 376 ) compared to reaction-free patients ( 39% , 113/292; p<0 . 0001 ) . Similarly , anti-LID-1 positivity was higher in reactional ( ENL: 95% , 39/41; RR: 80% , 95/119 ) compared to reaction-free patients ( 46% , 134/292 ) ( p>0 . 0001 ) . Regarding LID-1 serology at baseline , patients that developed ENL during follow-up had higher positivity than patients that manifested RR ( p = 0 . 01 ) . Serological responses to ND-O-LID were higher in ENL patients ( 88% , 36/41 ) compared to RR patients ( 60% , 71/119; p = 0 . 0004 ) and reaction-free ones ( 32% , 93/292; p<0 . 0001 ) ( Fig 2 ) . For all three M . leprae antigens tested , the ELISA results showed a gradual increase in the median ODs . The lowest median OD was seen in reaction-free patients , increasing in RR patients and the highest values were seen in ENL patients . For PGL-I serology , the median OD among reaction-free patients was 0 . 128 ( range 0–2 . 38 ) while among patients that developed RR and ENL the median ODs were 0 . 45 ( range 0–1 . 45 ) and 0 . 58 ( range 0–1 . 41 ) , respectively ( reaction-free versus RR and reaction-free versus ENL , p<0 . 0001 ) . For LID-1 serology , reaction-free patients presented a median OD of 0 . 26 ( range 0 . 07–3 . 13 ) while higher antibody levels were seen in patients that developed RR ( median OD = 1 . 29; range 0 . 08–3 . 36 ) and ENL ( median OD = 1 . 9; range 0 . 18–3 . 27 ) . Differences in antibody levels ( measured by the median OD ) were statistically significant when comparing reactional patients ( RR or ENL ) with reaction-free group ( p<0 . 0001 ) and when comparing RR and ENL groups ( RR versus ENL , p<0 . 0001 ) ( Fig 2 ) . The IgG and IgM antibody levels to ND-O-LID were also higher among reactional patients ( ENL , median OD = 1 . 77 , range 1 . 47–3 . 32 and RR , median OD = 1 . 11 , range 0–3 . 25 ) when compared to reaction-free patients ( median OD = 0 . 49; range 0–3 . 36 , p<0 . 0001 ) . Higher antibody levels were detected in ENL patients compared to RR ( p = 0 . 005 ) . As we observed a gradual increase in antibody levels from reaction–free to RR and ENL patients and since in leprosy , antibody responses are positively correlated with patients’ BI , we have analyzed the impact of BI in serology from reactional versus reaction-free patients . The analysis of BI and serological levels to M . leprae-specific antigens for both reactional and reaction-free groups clearly indicated that for RR and reaction-free patients , serological levels were associated with BI ( S3 Fig ) . However , for patients that further developed ENL and who had high BIs ( >3 ) , despite inter-individual serologic variability , antibody levels had not correlation with BI . Therefore we have analysed , serologic responses to PGL-I , LID-1 and ND-O-LID in reactional and reaction free patients stratified according to the bacterial index range: BI = 0 , 0<BI<3 and BI≥3 . The BI negative patients were mostly reaction-free ( n = 196 , 89% ) while 23 developed RR during clinical monitoring and no case of ENL was reported in this group . Serology of reaction-free BI negative leprosy patients ( 33 TT , 161 BT , 2 BB ) compared to RR patients ( 2 TT , 20 BT , 1 BL ) shower higher anti-PGL-I positivity rate in RR ( 27% versus 49% , respectively , p = 0 . 02 ) . The difference in anti- PGL-I antibody levels ( median ODs ) between reaction-free and RR patients was also statistically significant ( median ODs of 0 . 04 and 0 . 213 respectively , p = 0 . 014 ) ( Fig 3 ) . Anti-PGL-I positivity of BI negative patients that developed RR was higher compared to reaction-free ones . However , among reaction-free patients , 10–15% had high outlier serologic results to PGL-I , LID-1 , ND-O-LID indicating an important overlap of serologic responses from reaction-free patients and patients who developed RR . Extended clinical monitoring of these patients reaction-free patients with high seropositivity for 2 further years ( until December 2015 ) did not reveal the development of reactions nor relapse in any of these patients . This group of patients was composed by equal numbers of reaction-free patients ( n = 36; 13 BT , 7 BB , 15 BL , 1 LL ) and patients that developed RR ( n = 36; 8 BT , 3 BB , 25 BL ) . Seropositivity to PGL-I and LID-1 antigens was higher in reactional patients compared to reaction-free patients , 72% and 78% of patients who developed RR during follow-up were positive at diagnosis while among reaction-free patients , 50% and 53% were positive ( p = 0 . 026 and p = 0 . 012 respectively ) . Anti-ND-O-LID positivity rates were similar in reaction-free and RR groups ( p>0 . 05 ) ( Fig 4 ) . For anti-PGL-I and anti-LID-1 serology , patients who developed RR had higher median ODs compared to reaction-free ones ( PGL-I , median ODs = 0 . 46 and 0 . 25; p = 0 . 014 and LID-1 , median ODs = 1 . 1 and 0 . 52; p = 0 . 035 ) . For anti-ND-O-LID serology , the difference in antibody levels between groups was marginal ( p = 0 . 052 ) ( Fig 4 ) . This group of patients was composed by reaction-free patients ( n = 60; 34 LL , 2 BT , 24 BL ) and patients that developed either RR ( n = 60; all BL ) or ENL ( n = 41; 28 LL , 13 BL ) . Although RR patients presented higher seropositivity rate for all tested antigens , this difference was only significant for anti-LID-1 serology ( p = 0 . 01 ) . However , reaction-free and RR patients had similar medians of ODs for all tested antigens ( p>0 . 05 ) . Anti-ND-O-LID baseline positivity rates were higher in ENL patients compared to reaction-free patients ( 88% versus 72% respectively , p = 0 . 03 ) . Also , for LID-1 serology , patients that developed ENL had higher median OD than reaction-free patients ( OD median = 1 . 91 and 1 . 58 , respectively; p = 0 . 028 ) ( Fig 5A , 5B and 5C ) . The accuracy of serology at diagnosis , using PGL-I , LID-1 and ND-O-LID antigens , to predict the development of leprosy reactions during follow-up was analyzed by ROC curve . According to this analysis , a test that gives an area under the curve ( AUC ) above 0 . 7 is considered satisfactory [32] . For RR , comparison of the antibody responses in RR versus reaction-free patients , only anti PGL-I serology presented AUC above 0 . 7 ( Fig 6 ) but all antigens presented similar AUC . Establishing a specificity of 80% ( 95% CI: 75–84% ) the sensitivity is 44% ( 95% CI: 35–54% ) for a cut-off OD > 0 . 521 . Sensitivity , specificity and the AUC of M . leprae-specific serology is detailed in S2 Table . Regarding ENL during follow-up , all three serologic tests showed acceptable results . For anti PGL-I and anti ND-O-LID serology , establishing a specificity of 80% ( 95% CI: 75–84% ) the sensitivity is 58% ( 95% CI: 42–73% ) for a cut-off O . D > 0 . 520 and 1 . 527 , respectively . Anti LID-1 serology at baseline showed the best performance to predict ENL ( AUC = 0 . 847 ) . Setting a specificity of 80% ( 95% CI: 75–84% ) and sensitivity of 71% ( 95% CI: 55–84% ) the cut-off point was O . D > 1 . 5 ( Fig 7 and S2 Table ) .
Leprosy reactions are immune inflammatory episodes that can cause irreversible handicaps , incapacity and deformities and no prognostic marker is currently available [9] . In the current study , the use of quantitative M . leprae specific ELISAs with PGL-I , LID-1 and NDO-LID antigens , that have high sensitivity for multibacillary patients , have shown the potential application of anti-LID-1 serology at diagnosis for ENL prediction . This finding is in accordance with previous reports indicating that ENL occurs in multibacillary patients that have abundant antibody production [33–35] . A previous serological study of leprosy patients enrolled in U-MDT/CT-BR showed that the qualitative ML Flow test at baseline had limited sensitivity and specificity to predict whether patients would develop RR or ENL during follow-up [30] . The ML flow test uses PGL-I antigen and the ELISA anti PGL-I results from the current study did not show predictive value for ENL confirming previous results . The new finding of anti LID-1 serology showing predictive value for ENL is probably related to the antigen employed in a quantitative assay . While , PGL-I is a phenoglicolipid that induces IgM antibodies , LID-1 is a di-fusion recombinant protein originated from two highly immunogenic M . leprae antigens ( ML2331 and ML0405 ) . In the current study , M . leprae-specific ELISA results confirmed the limitation observed with ML flow results to predict RR . This limitation probably reflects the fact that RR mainly occurs in a context of strong CMI with little or no impact on the antibody production . Initially our analysis at diagnosis showed higher seropositivity and antibody levels to PGL-I , LID-1 and ND-O-LID antigens in patients that developed reactions during follow-up compared to the ones that remained reaction-free . Further stratification of leprosy serology data according to the type of leprosy reaction showed higher antibody levels to all three antigens in both ENL and RR versus reaction-free patients . A gradual increase in both seropositivy rate and median ODs was seen from reaction-free to RR patients while ENL patients had the highest values . However , for patients that remained reaction-free throughout the follow-up , a variable rate ( 2–15% ) had high serologic responses which were considered outliers . Therefore , our results showed a significant overlap between serologic responses of patients that further developed RR and the ones that remained reaction-free a finding that emphasizes the limitation of serology as a predictor of RR . The impact of patients´ bacillary load in leprosy-specific serology has been well described [14 , 18 , 24 , 36] . Also , higher BI has been shown to be an important factor for the development of reactions [12 , 37 , 38] . To understand the impact of patients´ BI in this different serologic pattern observed from reaction-free to ENL , serology data of reactional ( RR , ENL ) and reaction-free patients was analyzed according to distinct BI range . These analyses showed a positive correlation between the rate of reaction and the BI: the frequency of reactions was low among BI negative patients ( 10%; 23/219 ) , increasing to 50% ( 36/72 ) in patients with intermediary BI ( 0<BI<3 ) reaching 66% in reactional patients with high BI ( ≥3 ) . These results corroborate the influence of BI in the development of leprosy reactions , as previously described [12 , 37 , 38] . Corroborating this observation , a previous U-MDT/CT-BR report showed that patients with high BI ( ≥3 ) had a higher frequency of reactions compared to patients with BI <3 throughout the follow-up and that recurrent reaction were associated with high BI ( ≥3 ) [7] . It is well known that in patients with high bacillary load the start of leprosy treatment is characterized by a massive release of mycobacterial antigens [39 , 40] which can stimulate an exacerbated immune response , including antibody production and trigger leprosy reactions . ENL is a severe , often difficult to manage immunological complication of borderline lepromatous ( BL ) and lepromatous leprosy ( LL ) that can be triggered by specific treatment [33] . Over 50% of lepromatous leprosy patients and 25% of borderline lepromatous leprosy patients experienced ENL prior to MDT [34] . ENL is characterized by exacerbated humoral immune response with increased synthesis of IgG1 [52] and transient activation of cellular immunity , demonstrated by Th1 type cytokine production [34] . In our study increased levels of anti-LID-1 antibodies at diagnosis were observed in patients who developed ENL during monitoring compared with reaction-free patients . A previous study showed that the LID-1 fusion protein is recognized by M . leprae specific antibodies and induces cellular immunity measured by IFNγ production [22] . A recent study among MB patients that presented RR or ENL at diagnosis or during MDT showed that high and persistent levels of anti-LID-1 was associated with the occurrence of ENL at diagnosis or during MDT [53] . Similarly , as indicated by the high AUC by the ROC analysis , anti-LID-1 serology at diagnosis could identify patients susceptible to develop ENL with 71% sensitivity and 80% specificity . The demonstrated capacity of original proteins that compose LID-1 fusion protein ( ML2331 and ML0405 ) to induce both M . leprae specific humoral and cellular immune responses in leprosy patients supports the highest accuracy of anti LID-1 serology for ENL prediction . As part of ENL , immune complex formation and deposition that occur in tissues may cause a decrease in circulating antibodies levels [54] . Serological analysis of sequential samples collected during monitoring of patients from U-MDT/CT-BR might clarify the dynamic of anti-LID-1 serology during ENL . The immunopathogenesis of RR is characterized by Th1 type immunity and increased pro-inflammatory cytokines , as IP-10 , IFNγ , IL-1 , IL-2 and IL-12 [41] . The association of high levels of anti-PGL-I antibodies and higher risk to develop of RR is controversial . Some studies associated high levels of anti-PGL-I antibodies at diagnosis or after treatment and higher risk to RR development [42–44] . However , other investigations showed similar anti-PGL-I levels among patients that developed RR and reaction-free patients [45–47] . Although , our results indicated higher antibody responses in RR patients compared to reaction-free patients , 10–15% of the reaction-free patients with negative BI had high outlier antibody levels , however these levels were lower compared to BI positive patients . Therefore , despite the statistically significant difference in positivity rate and medians of OD in BI negatives that developed RR and reaction-free ones , an important overlap was observed in antibody levels between these two groups . These results highlight that regardless of the further development of leprosy reactions , there is a high inter-individual variability of serological responses in reaction-free , RR and ENL patients and this variability results in extensive standard deviation . Also , in the group of patients with higher BI ( ≥ 3 ) there was no difference in antibody responses between patients that manifested RR and patients that remained reaction-free . Overall , these results and the area under the curve given by the ROC analysis for RR strongly indicate that serological markers present a questionable applicability for RR prognosis . The identification of laboratory markers to predict the occurrence of leprosy reactions remains a priority in leprosy research aiming to prevent irreversible sequelae . Several studies have been carried out aiming to find a diagnostic and prognostic biomarker for leprosy reactions . CXCL10 and IL-6 were shown to be potential plasma markers for the diagnosis of RR and IL-7 , PDGF-BB and IL-6 for ENL diagnosis [48] . The elevation of CXCL10 levels was associated with episodes of RR , however without positive predictive value [49] . A recent study evaluated TNFα , anti-ceramide , anti-S100 , IgG and IgM anti-PGL-I , IgG1 and anti-LAM IgG3 antibodies in leprosy patients before , during and after the reaction episode . This study showed that preceding the RR episode 47% patients had increased levels of markers and the association of two to four markers detected 70% of patients that developed RR . The markers that showed the highest elevation were anti-ceramide , TNFα , anti-PGL-I and anti-S100 antibodies , suggesting that the association of these markers may enhance the sensitivity to predict RR [50] . Recently , biomarker profiles associated with the onset of RR were described in cohorts of patients from Bangladesh , Brazil , Ethiopia and Nepal who had peripheral blood mononuclear cells ( PBMCs ) stimulated and anti PGL-I antibodies measured . High IFN-γ , IP-10 , IL-17- and VEGF production by M . leprae-stimulated PBMC peaked at diagnosis of RR and the ratio of these pro-inflammatory cytokines versus IL-10 could be useful for the early diagnosis of RR and for evaluating treatment efficacy . Nevertheless , anti-PGL-I serology was not useful for the diagnosis of RR , but could help treatment monitoring [51] . Overall our results showed low applicability of anti-PGL-I serology for the prognosis of leprosy reactions . Thus , together with our previous ML flow study , our results indicate low applicability of serology for both diagnosis and prognosis of RR . Therefore , other plasma biomarkers associated with anti-PGL-I serology could potentially increase the sensitivity of anti PGL-I to predict RR . Despite the existence of seropositive patients with high BI who did not develop reaction and seronegative patients with low BI that developed reactions , anti-LID-1 serology at diagnosis has shown prognostic value for ENL development in BI positive patients: 71% sensitivity and 80% specificity . These apparently paradoxal results indicate that besides the impact of the bacillary load on the immune responses and on the risk to develop reactions , other yet unidentified factors are probably implicated in the susceptibility to manifest leprosy reactions . | Leprosy is a debilitating dermato-neurologic disease caused by Mycobacterium leprae . One of the main difficulties in the clinical management of leprosy patients is the development of leprosy reactions which are immune inflammatory episodes that can cause irreversible handicaps , incapacities and deformities . There are two major types of leprosy reactions: reversal reaction ( RR ) and erythema nodosum leprosum ( ENL ) . Currently , there is no laboratory test able to predict the emergence of leprosy reactions among recently diagnosed patients . In order to investigate laboratory markers for the occurrence of leprosy reactions , we investigated the prognostic value of serologic responses to M . leprae antigens ( PGL-I , LID-1 , ND-O-LID ) in 452 leprosy patients enrolled at the Clinical Trial for Uniform Multidrug Therapy Regimen for Leprosy Patients in Brazil/U-MDT/CT-BR . At diagnosis higher anti-PGL-I , anti-LID-1 and anti-ND-O-LID seropositivity rates were seen in patients who developed ENL and RR compared to reaction-free patients . The anti-PGL-I serology at diagnosis show low sensitivity to predict RR and anti-LID-1 serology at diagnosis has shown prognostic value for ENL development . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"tropical",
"diseases",
"bacterial",
"diseases",
"neglected",
"tropical",
"diseases",
"... | 2017 | Leprosy reactions: The predictive value of Mycobacterium leprae-specific serology evaluated in a Brazilian cohort of leprosy patients (U-MDT/CT-BR) |
Cells of the innate immune system act in synergy to provide a first line of defense against pathogens . Here we describe that dendritic cells ( DCs ) , matured with viral products or mimics thereof , including Epstein-Barr virus ( EBV ) , activated natural killer ( NK ) cells more efficiently than other mature DC preparations . CD56brightCD16− NK cells , which are enriched in human secondary lymphoid tissues , responded primarily to this DC activation . DCs elicited 50-fold stronger interferon-γ ( IFN-γ ) secretion from tonsilar NK cells than from peripheral blood NK cells , reaching levels that inhibited B cell transformation by EBV . In fact , 100- to 1 , 000-fold less tonsilar than peripheral blood NK cells were required to achieve the same protection in vitro , indicating that innate immune control of EBV by NK cells is most efficient at this primary site of EBV infection . The high IFN-γ concentrations , produced by tonsilar NK cells , delayed latent EBV antigen expression , resulting in decreased B cell proliferation during the first week after EBV infection in vitro . These results suggest that NK cell activation by DCs can limit primary EBV infection in tonsils until adaptive immunity establishes immune control of this persistent and oncogenic human pathogen .
Epstein-Barr ( EBV ) is a lymphotropic γ-herpes virus infecting over 90% of the human adult population [1 , 2] . A striking feature that the virus shares with other γ-herpes viruses is its oncogenic potential . This transforming property can be observed in vitro and in vivo both in immunocompetent and more frequently in immunosuppressed individuals . In the latter group , EBV causes tumors such as post-transplant lymphoproliferative disease and immunoblastic lymphoma , whereas , nasopharyngeal carcinoma , Hodgkin's disease and endemic Burkitt's lymphoma are the most prominent EBV-associated malignancies in immunocompetent individuals [3] . However , in most individuals , the lifelong chronic infection with EBV is fortunately free of complications due to effective immune control , primarily mediated by CD4+ and CD8+ T cells [4] . In addition to protective T cell immunity in healthy virus carriers , several lines of evidence suggest a role for innate lymphocytes in the resistance against EBV-associated malignancies . Firstly , in male patients with X-linked lymphoproliferative disease ( XLP ) , who frequently succumb after primary EBV infection to EBV-induced lymphomas , a mutation in the SAP gene leads to defective recognition of EBV-transformed B cells by NK cells [5–7] . While SAP mutations not only affect NK cell function , this defective recognition most likely contributes to loss of EBV-specific immune control . Secondly , IL-2-activated peripheral blood NK cells have been shown to restrict EBV-induced B cell transformation in vitro [8–10] . Thirdly , NK cell depletion from PBMCs prior to adoptive transfer into SCID mice , rendered the animals more susceptible to tumor development after transfer of EBV-transformed B cells [11] . Fourthly , activated NK cells have been shown to lyse lytically EBV replicating B cells [12] . Fifthly , a novel primary immunodeficiency with a specific NK cell defect was recently reported to be associated with EBV-driven lymphoproliferative disease [13] . Therefore , NK cells may be involved in the early phase of the EBV-specific immune response . NK cells are innate lymphocytes that play an important role in the control of infections and the immune surveillance of tumors [14] . In particular , early after primary viral infections they are thought to limit the viral burden until virus-specific T cells are able to eliminate the infection or control viral titers at low levels [15] . During infection with the β-herpes virus murine cytomegalovirus ( MCMV ) , NK cells have been shown to be crucial for limiting viral replication [16] . Indeed , lack of the activating NK cell receptor Ly49H , involved in the recognition of the MCMV m157 protein , confers susceptibility to fatal MCMV infection in BALB/c mice , whereas Ly49H-expressing C57BL/6 mice survive infection with high viral titers [17] . Nonetheless , NK cells need to receive additional signals to successfully protect against MCMV infection and these are provided by DCs [18–22] . Notably , NK cells produce cytokines such as IFN-γ , proliferate and increase their cytotoxicity upon activation by both myeloid and plasmacytoid DCs [23] . Therefore , DCs seem to activate NK cells early after infection in order to restrict pathogen replication until the adaptive immune system establishes long-lasting immune control . Although there are significant similarities between murine and human NK cell activation by DCs , human and mouse NK cell phenotypes and functions differ substantially and these differences also influence NK cell/DC interactions . Two main functional NK cell subsets have been distinguished in humans: CD56dimCD16+ NK cells readily lyse susceptible target cells , but secrete only low levels of cytokines after activation . In contrast , CD56brightCD16− NK cells produce large amounts of cytokines upon stimulation and acquire cytotoxicity only after prolonged activation [24 , 25] . Around 90% of human peripheral blood NK cells are CD56dimCD16+ , whereas CD56brightCD16− NK cells constitute less than 10% of the peripheral blood NK cell pool . However , in secondary lymphoid organs such as tonsils and lymph nodes , the CD56brightCD16− NK cells are the dominant subset [26–28] . Interestingly , human DCs primarily activate this NK cell subset [29 , 30] . Because the tonsils are the primary site of EBV infection , we investigated whether the DC/NK cell crosstalk could trigger NK cells to limit B cell transformation during EBV infection . We show here that human NK cells from peripheral blood , spleen , and tonsils limit the outgrowth of EBV-infected B cells after activation by DCs in vitro . Monocyte-derived DCs , which were matured with viral double-stranded RNA or its analog polyinosine-polycytidylic acid ( polyI:C ) to produce high levels of IL-12 , stimulated strong IFN-γ secretion and proliferation of mainly the CD56brightCD16− NK cell subset . These NK cells then significant restricted B cell transformation by EBV . Tonsilar NK cells were more efficient in inhibiting EBV-induced B cell transformation in vitro than peripheral blood NK cells and secreted high IFN-γ levels , which proved sufficient to limit B cell-transformation despite their low frequency at the site of primary infection . Restriction of B cell transformation by EBV was primarily due to IFN-γ secretion by DC-activated NK cells , and due to regulation of the EBV latency program by this cytokine . Interestingly , myeloid DCs , matured by exposure to EBV , were also able to elicit IFN-γ secretion by NK cells to levels protective against EBV-induced B cell transformation . Contrary to the original hypothesis that NK cells control pathogens via spontaneous cytotoxicity , which gave rise to the name of this innate lymphocyte subset , we demonstrate that these NK cell responses require activation by DCs and are mediated by cytokines . These data provide the first evidence for a direct antiviral effector function of NK cells in secondary lymphoid tissues , which might limit EBV infection until the adaptive immune system efficiently controls it .
The outcome of the crosstalk between NK cells and DCs depends strongly on the activation status of both cell types [27] . We investigated how differently matured human monocyte-derived DCs vary in their capacity to elicit proliferation and IFN-γ secretion by NK cells in order to define the optimal activation conditions for anti-viral NK cell responses . DCs were matured using a standard mixture of proinflammatory cytokines ( IL-1β , IL-6 , TNF-α and PGE2: cyt DC ) , the TLR3 and mda-5 ligand polyI:C ( polyI:C DC ) , and the TLR4 ligand LPS ( lps DC ) . In addition , DCs were matured with poly I:C supplemented with proinflammatory cytokines ( IL-1β and TNF-α ) and type I and II interferons ( DC1 ) to generate DC1 cells according to Mailliard and colleagues [31] . As expected , all of these maturation stimuli were able to significantly up-regulate MHC class II and costimulatory molecules such as CD80 , CD83 and CD86 ( Table S1 ) . However , when we compared different DC preparations for their capacity to activate NK cells from peripheral blood , we found that DCs matured with polyI:C were far superior in NK cell activation compared to the other DC preparations ( Figure 1A ) . polyI:C DCs and DC1s induced strong proliferation with 40–75% of the NK cells cycling after 6 days . Further characterization showed that the number of NK cells producing IFN-γ increased significantly ( 4-fold to 10-fold ) , when they were activated by polyI:C DCs and DC1s compared to immature DCs ( iDCs ) or cyt DCs , respectively ( Figure 1B ) . In line with previous findings , the CD56brightCD16− NK cells were preferentially stimulated by these DC preparations to proliferate and secrete IFN-γ ( Figure 1 ) . These data demonstrate that DCs matured by polyI:C are efficient stimulators of CD56bright CD16− NK cells , the NK cell subset enriched in human secondary lymphoid tissues . To characterize the mechanism of DC-mediated NK activation , we compared the production of NK cell stimulatory cytokines by DCs after maturation . Previous work indicated important roles for IL-12 , IL-15 , and IL-18 in the activation of NK cell proliferation and IFN-γ secretion [28 , 32] . We detected only little ( <50 pg/ml ) or no secretion of the bioactive form of IL-12 ( IL-12p70 ) by cyt DCs or iDCs , respectively , whereas polyI:C DCs and DC1s produced high amounts of this cytokine ( up to 5000 pg/ml ) ( Table 1 ) . LPS-matured DCs produced intermediate amounts of IL-12p70 . Monocyte-derived DCs matured with EBV-derived dsRNA produced slightly higher IL-12 levels than LPS matured DCs ( up to 200pg/ml ) . Addition of recombinant IL-12p70 at comparable levels ( ≥100pg/ml ) induced both proliferation and IFN-γ production by NK cells ( data not shown ) . IL-15 secretion and surface expression of IL-15 and IL-15Rα were also primarily induced by incubation of DCs with polyI:C ( Tables 1 and S1 ) . IL-18 secretion was not detectable for all DC preparations tested ( Table 1 ) . In line with these observations , IFN-γ secretion of NK cells stimulated by DC1s was mainly dependent on IL-12 ( 80% , p < 0 . 01 ) and to a lesser degree on IL-18 ( 15% , p < 0 . 05 ) ( Figure S1A ) . Similar results were obtained with polyI:C DCs ( data not shown ) . Consistent with the hypothesis that polyI:C-matured DCs mainly elicited IFN-γ secretion of NK cells via their high levels of secreted IL-12 , this NK cell/DC interaction was not sensitive to transwell separation ( Figure S1B ) . Furthermore , we found that NK cell proliferation upon coculture with DC1s ( Figure S1C ) and polyI:C DCs ( data not shown ) could be blocked with an IL-12 specific antibody by 90% or 80% , respectively ( p < 0 . 01 for both ) , and again transwell experiments showed that direct cell contact was not required ( data not shown ) . Blocking of IL-15 also significantly decreased numbers of surviving NK cells when combined with anti-IL-12 antibodies ( p < 0 . 01 ) , however blocking of IL-15 or IL-18 alone did not significantly decrease survival ( Figure S1D ) . In addition , antibody blocking of IL-2 did not influence DC induced proliferation and IFN-γ production by NK cells ( Figure S2 ) . These data suggest that polyI:C and EBV-derived dsRNA elicit IL-12 production , which in turn stimulates NK cell proliferation and IFN-γ production by NK cells . In contrast to previous studies focusing on IL-2 activation of NK cells , we addressed the question whether NK cells restrict EBV-induced B cell transformation after activation by cells of the innate immune system . Because interactions between NK cells and DCs have been shown to be essential for virus control in murine models of herpes virus infections [19 , 33 , 34] , we specifically investigated whether DCs can activate NK cells to limit EBV-mediated B cell transformation . For these experiments we used monocyte-derived DCs , since they can be generated at sufficient numbers to allow functional experiments . When we infected purified B cells with EBV and cocultured them with resting purified peripheral blood NK cells at ratios of 5:1 ( NK to B cells ) we could not observe any restriction of B cell transformation . Similarly , addition of iDCs or cyt DCs induced only limited NK cell-mediated inhibition of EBV-transformed B cell outgrowth ( 6 or 13% , respectively ) ( Figure 2A–2C ) . However , in cultures with NK cells and polyI:C DCs or DC1s , we observed a 49 or 55% ( both p < 0 . 01 ) reduction of the number of transformed B cells , respectively ( Figure 2B and 2C ) . The DC preparations used in this study had no significant direct effect on B cell transformation by EBV ( data not shown ) . This demonstrated that DC-activated NK cells can inhibit B cell transformation by EBV . Tonsils are the primary infection sites for EBV and harbor enriched populations of CD56brightCD16− NK cells , which can be efficiently activated by DCs . Therefore , we investigated whether tonsilar NK cells can restrict EBV-induced B cell transformation . For this purpose , we depleted tonsilar mononuclear cells of CD3+ T cells by cell sorting and compared the numbers of transformed B cells after EBV infection to cultures that were depleted of both CD3+ T and CD56+ NK cells ( Figure 2D and 2E ) . Without addition of DCs and with the addition of allogeneic iDCs or cyt DCs , we did not observe any significant difference in the number of transformed B cells after 12 days ( Figure 2D and 2E ) . However , we observed a 35% or 42% ( p < 0 . 03 and p < 0 . 01 ) reduction of the number of transformed B cells after addition of allogeneic polyI:C DCs or DC1s , respectively ( Figure 2D and 2E ) . These data suggest that tonsilar NK cells are able to restrict B cell transformation by EBV after stimulation by polyI:C DCs and DC1s . NK cells are present in tonsils at lower frequencies than in peripheral blood ( 0 . 3% compared to 10% ) , corresponding to only 3000 NK cells in a tonsilar B cell transformation assay with 6 × 105 B cells ( NK to B cell ratio of 1:200 ) . However , the ratio between CD56dimCD16+ and CD56brightCD16− NK cells is almost reversed between these organs with 75% of all NK cells being CD56brightCD16− in tonsils and only 5% being CD56brightCD16− in peripheral blood [27] . To compare directly the abilities of NK cells from blood and tonsils to restrict B cell transformation , we tested different ratios of blood NK cells to B cells . In addition to an intermediate ratio ( NK:B = 1:1 ) , we mimicked NK cell to B cell ratios found in tonsils ( NK:B = 1:200; 60% B and 0 . 3% NK cells ) . We also added sufficient peripheral blood NK cells to mimic the CD56brightCD16− NK cell to B cell ratio found in tonsil ( NK:B = 1:13 . 3; CD56brightCD16− NK cells are 15fold enriched in bulk tonsilar NK cells compared to blood NK cells: 75% in tonsil and 5% in blood ) . Already at a ratio of 1:1 we could only detect a limited 6% reduction of B cell transformation compared to controls ( Figure 3A and 3B ) . These findings indicate that tonsilar NK cells are 1000-fold more efficient in controlling B cell transformation than peripheral blood NK cells . To study which subset of NK cells limits B cell transformation , we sorted NK cell subsets from blood and activated them with DC1s . While sorted CD56dimCD16+ NK cells did not mediate restriction of B cell transformation ( NK:B = 4 . 5:1 ) , low numbers of CD56brightCD16− NK cells ( NK:B = 1:2 ) were found to inhibit B cell transformation ( 34% ) similar to 10-fold higher numbers of bulk NK cells ( Figure 3A and 3B ) . Next , we sorted NK cell subsets from tonsils and observed that again CD56brightCD16− NK cells but not CD56dimCD16+ NK cells were efficiently limiting B cell transformation after activation by DC1 ( Figure 3C and 3D ) . At 10-fold lower numbers , tonsilar CD56brightCD16− cells inhibited B cell transformation more ( 48 . 3% vs . 34% ) than their counterparts in blood . Since CD56brightCD16− NK cells are 15-fold enriched in tonsils compared to blood , NK cell mediated restriction is achieved with at least 150-fold lower bulk NK cell numbers in this organ . Further increasing the number of tonsilar NK cells 5-fold ( NK:B = 1:40 ) led to restriction of B cell transformation by 69 . 7% ( Figure 3C and 3D ) . Moreover , high numbers of splenic CD56brightCD16− NK cells ( NK:B = 5:1 ) , a mixture of blood and secondary lymphoid organ NK cells , when activated with matured autologous myeloid CD11chigh DCs isolated from spleen , restricted B cell transformation by 67% ( Figure 3E and 3F ) . Since we observed that at lower ratios of blood NK cells to B cells , even purified CD56brightCD16− peripheral blood NK cells were unable to limit B cell transformation after activation by polyI:C DCs and DC1s , whereas tonsilar NK cells were still able to restrict EBV-induced B cell transformation , we concluded that tonsilar CD56brightCD16− NK cells are functionally different from their counterparts in peripheral blood , and inhibit EBV-induced B cell transformation more efficiently . Human NK cells from secondary lymphoid organs such as tonsils produce IFN-γ rapidly upon activation and this antiviral cytokine contributes directly to control early infection in murine models of herpes virus infection [33] . When we compared the production of IFN-γ upon NK cell/DC coculture , we observed that NK cells from tonsil and lymph node produced significantly more IFN-γ than their equivalents from blood or spleen ( Figure 4A ) . Comparing CD56brightCD16− NK cells , tonsilar and lymph node cells produced 5-fold more IFN-γ than peripheral blood cells , which amounted to a 50-fold difference when bulk NK cell cultures were analyzed due to the enrichment of CD56brightCD16− NK cells in these organs . It had recently been reported that IL-18 exposed blood NK cells develop into a CD56brightCD83+CCR7+ NK cell subset with superior IFN-γ production [35] . In order to test if an enrichment of this NK cell subset could account for the superior ability of tonsilar NK cells to produce IFN-γ , we analyzed CD83 and CCR7 expression on tonsilar NK cells ( Figure S4 ) . Confirming our previously published data [28] , we found no CCR7 expression on tonsilar NK cells , and only a minor population expressed CD83 . Therefore , an enrichment of CD83+CCR7+ NK cells with superior IFN-γ production does not explain why NK cells from secondary lymphoid organs produce more IFN-γ than their peripheral blood counterparts . Higher IFN-γ secretion by tonsilar NK cells was also apparent when we then quantified the levels of IFN-γ in the culture supernatants of the B cell transformation assay after 12 days . We detected high levels in cocultures of NK cells from blood with polyI:C DCs ( 960 pg/ml; data not shown ) or DC1s ( 1560 pg/ml; Figure 4C ) , but only with the highest numbers of NK cells ( NK:B = 5:1 ) . However , IFN-γ levels were even higher in B cell transformation cultures with bulk tonsilar NK cells and polyI:C DCs ( 1140 pg/ml ) or DC1s ( 2500 pg/ml ) , and this IFN-γ secretion was NK cell dependent ( Figure 4B ) . Furthermore , we detected similar levels of IFN-γ also in cultures with sorted CD56brightCD16− NK cells from blood , tonsil and spleen ( Figure 4C and 4D and data not shown ) , reaching up to 4000pg/ml IFN-γ concentrations with purified tonsilar CD56brightCD16− NK cells . Therefore , only DC/NK cell co-cultures with polyI:C matured DCs and either high peripheral blood or low tonsilar NK cell numbers produce IFN-γ concentrations above 1000 pg/ml , and only these high IFN-γ levels correlate with control of EBV-transformed B cells . In order to estimate the contribution of NK cell-produced IFN-γ on control of EBV-infected B cells , we added recombinant IFN-γ to purified and EBV-infected B cells from blood , spleen and tonsil . We detected fewer transformed B cells in cultures with high levels of IFN-γ ( from 1000 to 10000 pg/ml ) compared to controls with low levels ( from 10 pg/ml to 200 pg/ml ) of IFN-γ or without IFN-γ ( Figure 4E ) . Indeed , 42% restriction of B cell transformation ( p < 0 . 02 ) was observed in tonsilar B cell cultures with 5000 pg/ml IFN-γ , an IFN-γ concentration that was produced by tonsilar NK cells upon culture with polyI:C DCs and DC1s , and similar to measured IFN-γ levels in tonsil cell cultures ( Figure 4A and 4D ) . However , IFN-γ mediated restriction of B cell transformation by EBV only limits , but does not eradicate EBV infection , since even with IFN-γ concentrations exceeding 10 , 000 pg/ml , we never observed more than 80% inhibition of B cell transformation by EBV ( data not shown ) . Finally , DC activated NK cell were still able to mediate restriction when separated by transwell from EBV-infected B cells , and we could block inhibition of B cell transformation by over 60% using blocking antibodies against IFN-γ ( Figure 4F ) . Therefore , IFN-γ contributes to NK cell-mediated restriction of EBV-induced B cell transformation . In order to extend our findings from monocyte-derived DCs to human blood DCs and from polyI:C to maturation by EBV , we exposed sorted human CD11c+ myeloid DCs to EBV particles directly ( DC:EBV MOI = 1:1 ) . We observed 189 ± 20 pg/ml IL-12 secretion and upregulation of the maturation marker CD83 upon coculture of myeloid DCs with EBV ( Figure 5A and 5B ) . Both infectious and heat-inactivated EBV elicited this DC maturation ( Figure 5A ) . DC maturation by EBV was not due to endotoxin contamination of the EBV virus preparations , since we detected less than in 0 . 1 ng endotoxin in 1 × 105 EBV RIU , a concentration insufficient for human DC maturation ( Figure 5A ) . Furthermore polymyxin B , which inhibits TLR4 stimulation by LPS [36] , had no effect on EBV-mediated DC maturation , but significantly inhibited DC maturation by LPS ( Figure 5A ) . While EBV induced IL-12 levels were lower than IL-12 concentrations in response to polyI:C and to high levels of LPS ( 1286 ± 188 pg/ml and 763 ± 87 pg/ml , respectively; Figure 5A ) , EBV-matured DCs stimulated purified autologous NK cells to secrete IFN-γ in excess of 4000 pg/ml via IL-12 ( Figure 5C ) . These IFN-γ concentrations are high enough to inhibit B cell transformation by EBV in vitro ( Figure 4E ) . These data suggest that human myeloid DCs can be matured by EBV and then activate NK cells to produce protective amounts of IFN-γ . In order to investigate how IFN-γ restricts B cell transformation , we infected B cells with EBV and compared total cell numbers , proliferation and expression of EBV-encoded genes between untreated and IFN-γ-treated cells . We started observing significant differences in cell numbers from day 4 ( Figure 6A ) . This coincided with beginning EBV-infected B cell proliferation , which was delayed when cells were treated with IFN-γ ( Figure 6B ) . Comparing expression of different EBV-encoded genes showed that EBER1 , EBNA2 , and EBNA1 were similarly up-regulated in untreated and IFN-γ-treated EBV-infected B cells . In contrast , we observed a delayed up-regulation of LMP1 in IFN-γ-treated cultures ( Figure 6C and data not shown ) . Quantitative RT-PCR demonstrated that expression of LMP1 was reduced by 28% and 49% at days 3 and 5 post infection , while at later stages similar LMP1 levels were observed with and without IFN-γ addition , when normalized to GAPDH ( Figure 6D ) . Notably , recombinant IFN-γ did not mediate restriction of B cell transformation when added later than 96h after EBV infection ( Figure 6E ) , and also did not inhibit growth of established EBV-transformed B cell lines ( data not shown ) . We did not observe any effect of IFN-γ on B cell viability as measured by counting live cells up to day 6 post-infection , and levels of the EBV receptor CD21 were not affected by IFN-γ treatment ( data not shown ) . Therefore , we suggest that DC activated NK cells limit B cell transformation by EBV via regulation of EBV latent infection , at least in part via delaying LMP1 expression via IFN-γ .
NK cells and DCs are central figures in the innate immune response , and have been shown to interact in early phases of murine herpes virus infections [18 , 19 , 34 , 37] . In contrast to the mouse , humans possess the CD56brightCD16− subset of NK cells , which rapidly secretes high IFN-γ levels and strongly proliferates upon activation by DCs [29] . These NK cells are enriched in secondary lymphoid organs like tonsils and lymph nodes , and are therefore strategically positioned to rapidly respond to pathogens at these sites [26–28] . Such a pathogen is the human tumor virus EBV , which enters the human body through the tonsils after transmission via saliva exchange . Within tonsils of healthy virus carriers , the proliferation program , which is also observed in in vitro EBV-infected B cells , was found in naïve B cells , which travel through the perifollicular T cell zone and follicular mantle zone to encounter antigen [38 , 39] . These areas were also described to harbor or are close to DC/NK cell interactions [30 , 40 , 41] . Our data suggest that at these sites , human DCs can activate preferentially CD56brightCD16− NK cells , which then become able to limit EBV-mediated B cell transformation , mainly by secretion of IFN-γ , and regulate the proliferation program of EBV latency via this cytokine . Restriction of EBV-induced B cell transformation by NK cells probably curtails EBV infection until it can be efficiently immune controlled by the adaptive immune system . These results suggest for the first time an important effector function for tonsilar NK cells early in the primary immune response against human persistent and oncogenic EBV . iDCs patrol the periphery and act as sentinels for the immune system [42] . Upon direct infection by a pathogen or uptake of pathogen-containing material in conjunction with a maturation stimulus they migrate to secondary lymphoid organs carrying information both in the form of a particular maturation pattern and pathogen constituents . DC maturation changes drastically the properties of DCs converting them into potent activators of both the innate and adaptive immune system . One group of DC receptors that detect pathogenic determinants and trigger the activating functions of DCs are TLRs [43] . In murine herpes virus infections , it has been shown that different pathways synergize for the activation of immune responses against these pathogens . TLR9-deficient mice as well as TLR3- and TLR2-deficient mice have increased MCMV titers , suggesting that the immune system uses complementing recognition systems in herpes virus infection [18 , 20 , 44] . Similarly , EBV might activate human DCs by means of several pathways . In addition to TLR9-activating CpG-motifs , by which EBV activates human plasmacytoid DCs [45] , the EBV genome supports convergent transcription , which occurs also in other DNA viruses such as herpes simplex virus-1 ( HSV-1 ) [46 , 47] . The resulting virally encoded dsRNAs have been isolated from HSV-1-infected cells [48] . Indeed , dsRNA from the convergently transcribed LMP1 and LMP2A antigens of EBV was able to stimulate IL-12 secretion . The measured IL-12 amounts were similar to levels secreted by CD11c+ DCs after exposure to EBV particles ( Figure 5A; Table 1 ) . Consequently , we suggest that myeloid DCs can detect EBV either directly , or indirectly through EBV-derived dsRNA as TLR3 and mda-5 agonists during primary infection and subsequently initiate the immune response by activating NK cells and priming of T cells [49 , 50] . Along these lines , DC-activated NK cells might have limited B cell transformation by EBV in a previous study from our lab , but complete regression of transformed B cells was only achieved after DC dependent priming of EBV-specific T cell responses and was also observed with purified CD4+ and CD8+ T cells [51] . In line with previous studies , we find that phenotypical markers such as MHC class II and costimulatory molecules are equally upregulated with various DC maturation stimuli , while cytokine secretion profiles varied dramatically between different DC maturation conditions . In particular , IL-12p70 , a potent stimulator of NK cells and Th1 responses , was produced at higher levels by DCs exposed to EBV , or matured with EBV-derived dsRNA or with maturation cocktails containing the dsRNA analog polyI:C , compared to immature DCs or DCs matured with proinflammatory cytokines or LPS . Nonetheless , even low levels of IL-12 secreted by LPS-matured DCs through directed secretion into the synapse between NK cells and DCs have been shown to activate NK cells [52] . Hence , our data suggest that myeloid DCs stimulate NK cell responses during primary EBV infection at least in part via IL-12 . Two main functional subsets of NK cells have been described in humans , while counterparts for these NK cell populations have not been identified in the mouse so far [25] . The CD56dimCD16+ subset is mainly responsible for natural cytotoxicity and antibody-dependent cell mediated cytotoxicity ( ADCC ) , while the CD56brightCD16− subset has been characterized by its unique capacity to produce high amounts of immunoregulatory cytokines , such as TNF , IFN-γ and GM-CSF , upon activation [25] . IFN-γ production by CD56brightCD16− NK cells , as well as NK cell proliferation are rapidly induced by DCs [30] . In addition , DCs also augment cytotoxicity of this subset after prolonged activation [53] . Several studies indicate a role for NK cells in the control of EBV infection and in particular in early primary immune responses [8–10 , 54] . However , all in vitro studies up to now have used blood NK cells activated by IL-2 , which is , at least in humans , mostly secreted by activated T cells and therefore presumably not present during innate immune responses . Hence , we focused on NK activation by DCs as a physiological NK cell stimulus present during the early immune response to primary virus infections . Furthermore , previous studies on the role of NK cells during EBV infection mainly emphasized cytotoxicity , but not IFN-γ secretion of NK cells , although lymph node- and tonsil-resident NK cells primarily release cytokines after activation [26 , 28 , 30] . Suggesting a prominent role for NK cell-derived cytokines early during EBV infection , earlier studies have found that recombinant IFN-γ is protective for several days during initial B cell transformation by EBV , whereas in contrast , type I interferons inhibit transformation only during the first hours after infection [55] . Moreover , EBV-specific CD4+ and CD8+ T cells have been reported to mediate regression of EBV-transformed B cells despite low to undetectable cytolytic activity [56 , 57] . Our data support the hypothesis that tonsilar NK cells restrict efficiently EBV-induced B cell transformation via their superior ability to produce IFN-γ upon DC activation . Although IL-12 and IFN-α secreted by polyI:C DCs and DC1s upregulate cytotoxicity of NK cells ( Figure S3 ) , autologous EBV-transformed B cells are not efficiently killed by activated NK cells due to their high MHC class I expression ( data not shown and [12] ) . Instead , the results of this study support a prominent role for IFN-γ in the innate immune response to EBV by NK cells . First , addition of recombinant IFN-γ to the B cell transformation assay decreased the number of transformed B cells . Second , IFN-γ levels , sufficient to restrict EBV-induced B cell transformation , were found in our cocultures of EBV-infected B cells with DCs and NK cells . Finally , blocking of IFN-γ in B cell transformation assays with NK cells significantly decreased the protective effect of NK cells . Hence , we conclude that this cytokine significantly contributes to innate resistance against primary EBV infection . In addition to its direct antiviral activity , IFN-γ secreted by DC-activated NK cells might also shape the EBV-specific adaptive immune response favoring a Th1-polarization which is observed in EBV-positive individuals [58–61] . Therefore , contrary to the original paradigm that NK cells respond primarily with cytotoxicity without prior activation , which gave this lymphocyte subset its name , we demonstrate that their main function against a relevant human pathogen consists of cytokine secretion after activation by DCs . EBV transforms B cells by the coordinate expression of EBV latency genes that provide signals for B cell survival and proliferation . Of the eight latent EBV antigens , LMP1 has been suggested to be the main oncogene of the virus , causing epithelial cell transformation in vitro and B cell transformation in vivo [62–64] . Therefore , IFN-γ induced down-regulation of LMP1 transcription could be one mechanism by which DC-activated NK cells limit EBV-induced B cell expansion . The fairly late transcription of LMP1 , compared to other EBV latent antigens , during the establishment of EBV latency could also explain why IFN-γ can restrict B cell transformation by EBV during the first days of primary EBV infection , while others and we found that IFN-γ was not able to inhibit proliferation of fully EBV-transformed lymphoblastoid cell lines ( LCLs ) ( [55] and data not shown ) . Similarly , IFN-γ secretion by NK cells was shown to limit MCMV infection during the first week of infection , and reduced immediate early or late MCMV gene transcription , depending on the infected cell type [33 , 65] . Therefore , tonsilar NK cells might limit latent EBV infection by IFN-γ mediated down-regulation of LMP1 until adaptive T cell immune responses can eliminate fully EBV-transformed B cells . In summary , we suggest that myeloid DCs stimulate NK cells during EBV infection primarily via their ability to secrete IL-12 . Activated NK cells are then able to mediate restriction of EBV-mediated B cell transformation . Tonsilar NK cells , which , like lymph node NK cells , produce higher levels of IFN-γ than their peripheral blood counterparts after activation by DCs , are superior in inhibiting EBV-induced B cell transformation in vitro by down-regulating important components in the proliferation program of EBV latency . These results suggest a novel and important effector function for tonsilar CD56brightCD16− NK cells upon DC activation in the primary immune response against EBV . Beyond EBV infection , our data suggest that humans have a strategically well-positioned population of NK cells that directly combats pathogen entry at mucosal sites and might restrict pathogens until they can be cleared or controlled by adaptive immunity .
The following directly-labeled monoclonal antibodies were used for flow cytometry: anti-CD3 ( clone SK7 ) , anti-CD11c ( B-ly6 ) , anti-CD19 ( HIB19 ) , anti-CD21 ( B-ly4 ) , anti-CD23 ( M-L233 ) , anti-CD56 ( B-159 ) , anti-CD80 ( L307 . 4 ) , anti-CD86 ( IT2 . 2 ) , anti-HLA-DR ( TU36 ) , anti-IFN-γ ( B27 , all BD Biosciences ) , anti-CD16 ( 3G8 , Caltag ) , anti-CD25 ( PC61 5 . 3 ) , anti-CD83 ( HB15a , both Beckman Coulter ) , and anti-IL-15 ( 34599 , R&D Systems ) . Cells labeled with goat-polyclonal anti-IL-15Rα ( R&D Systems ) were stained with AlexaFluor-488 rabbit-anti-goat-IgG ( Molecular Probes ) . The following monoclonal antibodies were used for antibody-mediated blocking: anti-IL-12 ( clone 24910 ) , anti-IL15 ( 34593 , both R&D Systems ) , and anti-IL-18 ( 125–2H , MBL International ) . IgG1 ( MOPC-21 , BioLegend ) was used as control . All tonsils , lymph nodes and spleens were obtained as part of Institutional Review Board-approved protocols . Tonsils were collected immediately after surgery from patients undergoing tonsilectomy for chronic inflammation . Tonsils were not acutely inflamed at the time of removal . Spleens and lymph nodes were procured by the regional Organ Procurement Organization from brain-dead donors after obtaining informed consent from appropriate individuals . Soon after their removal , tissues were mechanically dissociated to obtain single cell suspensions and were then filtered through a 75-μm nylon cell strainer to exclude undissociated fragments . Debris and dead cells were eliminated using Ficoll/Hypaque ( Amersham Pharmacia ) discontinuous gradient centrifugation . Single cell suspensions were then extensively washed and cryopreserved . PBMCs were isolated from leukocyte concentrates ( New York Blood Center ) by density-gradient centrifugation on Ficoll/Hypaque . CD14+ cells were isolated from PBMCs by positive magnetic cell separation ( MACS , Miltenyi Biotec ) and cultured for 5 days in RPMI1640 + 1% single donor plasma + IL-4 and GM-CSF according to standard protocols [30] . The CD14− cells were frozen for later isolation of B cells and NK cells . Splenic DCs were isolated as previously described by flow cytometric sorting using a BD FACSVantage SE cell sorter [30] . To isolate CD11c+ cells from blood , PBMCs were overlayed with an Optiprep gradient ( 1 . 080 to 1 . 049 ) and centrifuged for 30 min at 700×g . Low-density fractions were collected and CD11c+ DC were further enriched by depletion of CD14+ , CD3+ , CD8+ , and CD19+ cells by MACS . CD11c+ DCs were purified by flow cytometric sorting using a BD FACS Aria cell sorter by isolating lin- ( CD3 , CD14 , CD19 , and CD56 ) , HLA-DR+ , and CD11c+ cells . Purify after sorting was regularly higher than 99 . 5% . DCs were matured for 2d in medium with IL-4 , GM-CSF , and i ) 10 ng/ml IL-1β , 1 , 000 units/ml IL-6 , 10 ng/ml TNF-α , and 1 μg/ml prostaglandin E2 ( cyt DC ) , ii ) 25 μg/ml polyinosine-polycytidylic acid ( polyI:C , Invivogen ) ( polyI:C DC ) , iii ) 25 ng/ml IL-1β , 50 ng/ml TNF-α , 3 , 000 IU/ml IFN-α , 500 pg/ml IFN-γ , and 25 μg/ml polyI:C ( DC1 ) , iv ) 250 ng/ml LPS ( Sigma ) ( lps DC ) . CD11c+ DCs were exposed to AGS-cell derived EBV at an MOI of 1 . Maturation of DCs was monitored by flow cytometry using anti-CD25 , anti-CD80 , anti-CD83 , anti-CD86 , and anti-HLA-DR . Secretion of cytokines was quantified using IL-12p70 ELISA , IL-15 ELISA ( both R&D Systems ) and IL-18 ELISA ( Bender Medsystems ) . Frozen CD14− PBMCs were thawed , washed and B cells were isolated by positive selection using CD19− Microbeads ( Miltenyi Biotec ) . NK cells were isolated from either CD19− or CD14− fractions by negative selection using the NK cell Isolation Kit II ( Miltenyi Biotec ) according to the manufacturer's instructions . The purity of the isolated B cells and NK cells was higher than 90% and contained less than 5% contaminating T cells as determined by flow cytometry . For other experiments , B cells , NK cells , and NK cell subsets were isolated by flow cytometric sorting using a BD FACSVantage SE cell sorter . The EBV+ marmorset cell line B95–8 was seeded at 2 × 105 cells / ml and cultured for 12 d in RPMI1640 + 10% FCS + gentamycin without refeeding . Virus-containing supernatant was centrifuged at 2000 rpm for 10 min and passed through a 0 . 45 μm filter . EBV+ AGS cells were used to produce EBV as previously described , and virus was further concentrated and purified by ultracentrifugation [66] . AGS-derived EBV preparations contained less than 0 . 1 EU/ml endotoxins as measured by LAL test ( Cambrex Corporation ) . A 1 . 3 kB fragment spanning the coding region of LMP1 in one direction and part of the first intron of LMP2A in the other direction was cloned into pGEM ( Promega ) between the T7 and SP6 promoter . ssRNA was generated using linearized plasmid and the Riboprobe Combination System SP6/T7 ( Promega ) . After verification of integrity of RNA by gel electrophoresis , ssRNA was purified using RNeasy Kit ( Qiagen ) and quantified by Nanodrop . Finally , equal amounts were annealed in siRNA buffer ( Dharmacon ) to generate dsRNA and successful annealing was confirmed by gel electrophoresis . Isolated NK cells were labeled with 1 μM CFSE in PBS plus 0 . 1% BSA for 10 min at 37°C . After washing twice with RPMI1640 + 5% human serum + gentamycin , NK cells ( 2 . 5 × 105 cells in 96 well plate ) were cultured with 500 IU/ml IL-2 or autologous DCs at a ratio of 5:1 for 6 days at 37°C in RPMI1640 + 5% human serum + gentamycin . In selected experiments , isotype control antibody ( 5 μg/ml ) or blocking antibodies against IL-12 , IL-15 ( 5 μg/ml each ) , and IL-18 ( 1 μg/ml ) were added to the cultures at the beginning and on day 3 of culture . Where indicated , DCs ( bottom ) were separated from NK cells ( top ) by 0 . 4 μm pore membranes ( Corning ) . CFSE fluorescence and CD16 staining was evaluated on CD3−CD56+ cells by flow cytometry . Where indicated , B cells were CFSE labeled as described above and proliferation of CD19+CD20+ cells was evaluated with and without 10 , 000 pg/ml IFN-γ addition after infection with EBV as CFSE dilution by flow cytometry . Live cell numbers were determined by trypan blue exclusion . For intracellular staining of IFN-γ , isolated NK cells from blood ( 2 . 5 × 105 cells in 96 well plate ) were incubated with 500 IU/ml IL-2 or autologous DCs at a ratio of 1:2 at 37°C in RPMI1640 + 5% human serum + gentamycin . In selected experiments , isotype control antibody ( 5 μg/ml ) or blocking antibodies against IL-12 , IL-15 ( 5 μg/ml each ) , and IL-18 ( 1 μg/ml ) were added . Brefeldin-A was added after 6 h of coculture and additional 6 h later cells were harvested and stained with anti-CD3 , anti-CD56 and anti-CD16 . After fixing the cells with 2% paraformaldehyde , they were permeabilized and stained with anti-IFN-γ . In other experiments , DCs ( bottom ) were separated from NK cells ( top ) by 0 . 4 μm pore membranes and IFN-γ production was compared to NK-DC cocultures without separation after 20h using an IFN-γ ELISA ( Mabtech ) . To directly compare IFN-γ production , sorted NK cell subsets from blood , spleen , tonsil and lymph node ( 1 × 104 cells CD56brightCD16− , 1 × 105 CD56+CD16+ ) were then cultured with allogeneic or autologous DCs at a ratio of 1:10 or 1:1 , respectively . After 20 h IFN-γ levels were determined using ELISA . Isolated B cells ( 1 × 105 cells in 48 well plate ) were cultured in RPMI1640 + 5% human serum + gentamycin , infected with EBV and isolated NK cells were added at indicated numbers . In experiments including DCs , they were added at B cell to DC ratios of 1:1 . In other experiments , B cells and DCs ( bottom ) were separated from NK cells ( top ) by 0 . 4 μm porous membranes . After 12 d , numbers of transformed B cells were quantified by counting live cells via trypan blue exclusion and determining the ratio of CD19+CD21+CD23+ cells to total live cells by flow cytometry . Restriction of B cell transformation was calculated by comparing numbers of transformed B cells between respective samples with and without NK cells; % Restriction of B cell transformation = ( 1 − total transformed B cell number of sample with NK cells/total transformed B cell number of sample without NK cells ) × 100 . Cryopreserved tonsilar mononuclear cells were thawed , washed , and then stained with anti-CD3 and anti-CD56 . Cells were then depleted ( i ) of CD3+ cells ( Tonsil -T ) or ( ii ) of CD3+ and CD56+ cells ( Tonsil -NK-T ) by flow cytometric sorting using a BD FACSVantage SE cell sorter . The number of sorted cells per condition was adjusted according to the ratio between sorted cells and input cell numbers ( 1 × 106 cells in 48 well plate ) . Then , the cells were infected with EBV and , where indicated , DCs were added at ratio of total cell to DC of 10:1 . After 12 d , numbers of transformed B cells were quantified by counting live cells via trypan blue exclusion and determining the ratio of CD19+CD21+CD23+ cells to total live cells by flow cytometry . Restriction of B cell transformation was calculated by comparing numbers of transformed B cells between respective samples with and without NK cells; % Restriction of B cell transformation = ( 1 − total transformed B cell number of sample with NK cells/total transformed B cell number of sample without NK cells ) × 100 . RNA was isolated from non-treated and IFN-γ-treated B cells infected with EBV at indicated time point using the RNeasy Kit ( Qiagen ) according to the manufacturer's instructions . 10 ng of total RNA were used for semi-quantitative RT-PCR using the OneStep RT-PCR-Kit ( Qiagen ) and gene-specific primers using the following program: 50°C ( 30 min ) , 95°C ( 15 min ) , followed by 35 cycles of 95°C ( 30 sec ) , 55°C ( 30 sec ) , and 72°C ( 30 sec ) . Primers sequences were as follows: Actin 5′-CAAGAGATGGCCACGGCTGCT , Actin 3′-TCCTTCTGCATCCTGTCGGCA , EBNA1 5′-GAGCGTTTGGGAGAGCTGAT , EBNA1 3′-CATTTCCAGGTCCTGTACCT , EBNA2 5′-CATAGAAGAAGAAGAGGATGAAGA , EBNA2 3′-GTAGGGATTCGAGGGAATTACTGA , EBER1 5′-AAAACATGCGGACCACCAGC , EBER1 3′-AGGACCTACGCTGGCCCTAGA , LMP1 5′-AGGTTGAAAACAAAGGAGGTGACCA , LMP1 3′-GGAACCAGAAGAACCCAAAAGCA , LMP2a 5′-ATGACTCATCTCAACACATA , LMP2a 3′-CATGTTAGGCAAATTGCAAAA . To generate cDNA for RealTime-PCR , 500 ng total RNA were reverse transcribed per 50 μl reaction using the TaqMan Reverse Transcription Reagent ( Applied Biosystems ) , using the following program: 25°C ( 10 min ) , followed by 50°C ( 30 min ) and 75°C ( 5 min sec ) . Amplification was performed in a final volume of 20 μl , containing 2 μl cDNA from the reversed transcribed reaction , primer mixture ( 0 . 25 μM each of sense and antisense primers ) , and 10 μl of 2× SYBR Green Master Mix ( Applied Biosystems ) . PCR was performed in ABI 7900HT Sequence Detection System ( Applied Biosystems ) using following program: 95°C ( 15 min ) , followed by 40 cycles of 95°C ( 15 sec ) , 55°C ( 30 sec ) , and 68°C ( 30 sec ) . The final mRNA levels of LMP1 were normalized to GAPDH using the comparative CT method . Primers sequences for RealTime-PCR were as follows: GAPDH 5′-AGCCACATCGCTCAGACAC , GAPDH 3′-GCCCAATACGACCAAATCC , LMP1 5′-AGGTTGAAAACAAAGGAGGTGACCA , LMP1 3′-GGAACCAGAAGAACCCAAAAGCA . To evaluate the cytolytic activity after DC activation , NK cells were cocultured for 2 d with DC1s at a ratio of 5:1 with or without blocking antibodies against IL-12 and type I interferon receptor ( anti-CD118 , clone MMHAR-2 , PBL Biomedical Laboratories ) . The NK cell-sensitive lymphoblastoid cell line LCL 721 . 221 , which does not express surface HLA class I molecules , was used as target cell . Cytotoxicity assays were performed , as previously described [28] . Briefly , target cells were labeled with PKH26 ( Sigma-Aldrich , St . Louis , MO ) , and then incubated with NK cells at different NK cell/target cell ratios . After 6 h , cells were harvested; TO-PRO-3 , a membrane-impermeable DNA stain , was added to each culture ( 1 μM final concentration ) ; and cells were finally analyzed by flow cytometry . Background and maximum TO-PRO-3 staining were obtained by incubation of target cells with medium and detergent , respectively . The percent specific lysis was calculated as ( % TO-PRO-3+PKH26+ cells in NK/target cell co-culture − % TO-PRO-3+PKH26+ cells in medium ) / ( % TO-PRO-3+PKH26+ cells in detergent − % TO-PRO-3+PKH26+ cells in medium ) × 100% . Statistical analyses were performed with the paired two-tailed Student t-test . The p-value of significant differences is reported . Plotted data represent mean plus standard deviation ( SD ) , unless otherwise stated .
Accession numbers mentioned in this study are from Swiss-Prot ( http://www . ebi . ac . uk/swissprot/ ) and from GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/ ) , respectively: hIL-2 , P60568; hIL-12p35 , P29459; hIL-12p40 , P29460; hIL-15 , P40933; hIL-15Rα , Q13261; hIL-18 , Q14116; hIFN-γ , P01579; hIFNα/β-R , P17181; hCD3 , P07766; hCD16 , P08637; hCD25 , P01589; hCD56 , P13591; hCD80 , P33681; hCD83 , Q01151; hCD86 , P42081; EBNA1 , P03211; EBNA2 , P12978; LMP1 , P03230; LMP2a , P13285; EBER1 , J02078; hβ-actin , P60709; hGAPDH , P04406 . | Epstein-Barr virus ( EBV ) establishes a persistent infection in nearly all human adults . Due to its tumor causing potential EBV infection has to be continuously controlled by the immune system in virus carriers . We demonstrate here that in the first week after infection , when other EBV-specific immune responses are still being recruited , human natural killer ( NK ) cells are able to prevent transformation of the main host cell type by EBV , the human B cell . Especially NK cells of tonsils , the primary site of EBV infection , inhibit B cell transformation by EBV after they have been activated by dendritic cells ( DCs ) . For this protective function , EBV can directly stimulate DCs to efficiently activate NK cells . Interestingly , NK cells primarily prevent B cell transformation by EBV via secretion of the anti-viral cytokine IFN-γ , and NK cells from tonsils and lymph nodes produce 5-fold more of this cytokine than their peripheral blood counterparts . These data suggest that specialized NK cells in tonsils , the mucosal entry site of EBV , can be efficiently stimulated by EBV-activated DCs , and then limit EBV-induced B cell transformation until EBV-specific immune control by other components of the immune system is established . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"viruses",
"none",
"virology",
"in",
"vitro",
"immunology",
"homo",
"(human)"
] | 2008 | Tonsilar NK Cells Restrict B Cell Transformation by the Epstein-Barr Virus via IFN-γ |
While we have good understanding of bacterial metabolism at the population level , we know little about the metabolic behavior of individual cells: do single cells in clonal populations sometimes specialize on different metabolic pathways ? Such metabolic specialization could be driven by stochastic gene expression and could provide individual cells with growth benefits of specialization . We measured the degree of phenotypic specialization in two parallel metabolic pathways , the assimilation of glucose and arabinose . We grew Escherichia coli in chemostats , and used isotope-labeled sugars in combination with nanometer-scale secondary ion mass spectrometry and mathematical modeling to quantify sugar assimilation at the single-cell level . We found large variation in metabolic activities between single cells , both in absolute assimilation and in the degree to which individual cells specialize in the assimilation of different sugars . Analysis of transcriptional reporters indicated that this variation was at least partially based on cell-to-cell variation in gene expression . Metabolic differences between cells in clonal populations could potentially reduce metabolic incompatibilities between different pathways , and increase the rate at which parallel reactions can be performed .
The aim of this study was to analyze metabolic phenotypes and quantify the assimilation of nutrients at the level of individual bacterial cells . The motivation was to contribute towards filling a major gap in our understanding of the metabolism of bacteria and other microorganisms . Many basic principles and underlying mechanisms of bacterial metabolism are well known [1] . We know , for a given culture environment , the metabolic pathways that are expressed and the fluxes of metabolites through these pathways [2 , 3] . What we know little about , however , is whether each individual cell expresses all these metabolic pathways , or whether there are substantial and functionally relevant metabolic differences between cells . The notion of phenotypic differences between genetically identical microbial cells has emerged as a concept in biology in the 1970s [4] . Research on phenotypic heterogeneity received renewed attention in the last two decades , with the rise of new and improved tools for single-cell analysis [5 , 6 , 7] . Quantitative analysis of gene expression at the single-cell level has led to the discovery of substantial levels of variation that are independent of genetic and environmental differences . One source of this variation is that many cellular processes are based on molecules that occur in small numbers per cell , and processes that occur at low rates; consequently , fluctuations in these small numbers or variations in these rates lead to substantial phenotypic variation between single cells , and within one cell over time [5 , 7 , 8 , 9] . What is the evidence for phenotypic variation in metabolism , and how might such variation matter for the dynamics in microbial populations ? Previous studies have found that intermediate concentrations of nutrients or other inducers of metabolic activity lead to variation in expression of metabolic genes and , generally , variation in metabolism [10 , 11 , 12] . Other experiments have found that cells differ in their growth rates during growth in constant environmental conditions [13] , or during controlled switches from one growth condition to another [14 , 15 , 16] . Genome-wide screening for phenotypic variation showed that genes involved in carbon utilization , energy production and metabolism have higher levels of phenotypic variation than genes of most other functional classes in E . coli [17] . Variation in the expression of metabolic genes and its influence on bacterial growth and metabolic activity has been shown in model bacterial organisms and environmental isolates [12 , 18] . In addition , metabolic diversity has also been reported between cells that belong to the same operational taxonomic unit ( but are not necessarily genetically identical ) from diverse environmental samples [19 , 20 , 21 , 22] . Such variation in metabolism can potentially have functional consequences . On one side , phenotypic variation in metabolism can result in some individuals growing much slower than population average , and thus lowering the population growth rate . On the other hand , there are possible benefits . One possible benefit is that variation allows a genotype to cope with fluctuating environments; if individuals carrying the same genotype express different metabolic phenotypes , this increases the chance that at least some individuals will fortuitously express phenotypic traits that allow them to continue to grow and divide in case the environment shifts suddenly [12 , 14 , 23 , 24 , 25] . Another potential benefit is that if individual cells in a clonal population differ in their metabolic phenotypes and specialize on different subsets of metabolic reactions , they could potentially benefit from performing these processes faster or more efficiently . This would be expected if the metabolic specialization of single cells resolved incompatibilities and biochemical conflicts [26 , 27 , 28] . Phenotypic differentiation in metabolism could thus provide benefits of specialization to single cells while providing the clonal population as a whole with the benefits of being a generalist , namely the ability to consume a wide set of nutrients [26 , 28 , 29] . Whether individual cells show substantial metabolic variation in steady state conditions , and how this impacts the dynamics in microbial populations , is largely unknown . We thus set out to measure metabolic activities of single bacterial cells . As a model system , we used Escherichia coli growing on a combination of the two carbon sources , glucose and arabinose , in chemostats . At high carbohydrate concentration , cells only utilize glucose , and the uptake and catabolism of arabinose is blocked through inducer exclusion , a mechanism of carbon catabolite repression [30] . Catabolite repression vanishes at low sugar concentrations , where carbohydrate availability limits the population sizes that can be attained [31 , 32] . Our goal was to measure uptake and assimilation of both sugars at the level of single cells , to ask whether cells differ from each other in their total sugar uptake , and whether individual cells specialize on one of the two substrates . Most established approaches in systems biology are not well suited to answer these questions . The common single-cell methods allow quantifying gene expression and phenotypic traits that can be assessed optically , such as cell growth , division , survival and differentiation , but they do not allow quantifying actual uptake and assimilation of nutrients . We used an alternative approach that has mostly been used for analyzing microorganisms in natural communities: labeling with stable isotope-labeled nutrients , and the subsequent detection of assimilated nutrients by means of nanometer-scale secondary ion mass spectrometry ( NanoSIMS ) . This technique allows precise quantification of the isotope content of individual cells [33 , 34 , 35] . In parallel , we analyzed transcriptional reporters for genes involved in the uptake of glucose and arabinose to infer underlying molecular mechanisms of variation in assimilation .
Our main focus was on determining the rate at which individual bacterial cells consume and assimilate the two sugars that were available in our experiments , glucose and arabinose . Specifically , we wanted to know whether individual cells differ in the relative amount of each of the two sugars they assimilate , and whether they assimilate different total amounts of the two sugars combined . We worked with clonal populations of E . coli strains that were derived from the commonly used laboratory strain MG1655 [36] . Clonal populations were grown in mini-chemostats [2] in media containing arabinose and glucose as the sole externally added carbon sources . We ran a set of replicated chemostats under conditions where population size was limited by the carbon source , and another set where population size was limited by the nitrogen source and hence under conditions of carbon excess , while keeping the population size unchanged ( see Materials and Methods , and S1 File , ‘Supplementary Methods’ ) . As comparison , we also analyzed clonal populations growing in batch cultures ( S1 Fig ) . To analyze metabolic activity of single cells , we used sugars that were labeled with stable isotopes: we used arabinose in which 99% of all carbon atoms were the stable 13C isotope , and glucose in which 97% of all hydrogen atoms were 2H , the deuterium isotope . Assimilation of arabinose thus leads to the accumulation of 13C into the bacterial biomass , and assimilation of glucose leads to the accumulation of deuterium . We used nanometer-scale secondary ion mass spectrometry ( NanoSIMS ) to quantify isotope enrichment in biomass of bacterial cells . Due to its high resolution , NanoSIMS provides precise intracellular measurements of various stable isotopes [33 , 34 , 35] . We first performed control experiments in chemostats where bacterial strains were grown on only one of the two sugars ( see S1 File , 'Maximal incorporation of the stable isotope-labeled sugars in chemostats' ) , and analyzed the isotopic labeling that they maximally achieved . Interestingly , these experiments showed that bacteria grown on 13C-arabinose for ten generations ( during which more than 99% of the biomass is newly formed ) reached a 13C labeling of about 70% , and thus contained about 30% of 12C ( S1 Table ) . The unlabeled carbon assimilated during this experiment is most likely derived from assimilable organic carbon ( AOC ) , which is a collective term describing the fraction of labile dissolved organic carbon molecules that is readily assimilated by microorganisms , resulting in their growth [37] . It usually consists of a broad range of low molecular weight organic molecules such as sugars , organic acids , and amino acids . In our chemostat setup , AOC could originate from culture containers , medium components , and laboratory air [37]; in control experiments where we grew E . coli in batch cultures without adding sugar , we found that AOC alone can support growth of approximately 106 cells/ml ( S2 Fig , ‘Bacterial growth on AOC’ in S1 File ) . AOC made a considerable fraction of the assimilated carbon because the sugar concentration supplemented in the medium was relatively low , i . e . 20 μM ( see S1 File for further discussion: ‘AOC contamination in the chemostat setup’ ) . Bacteria grown on 2H-glucose for ten generations reached a 2H labeling of about 14% ( S1 Table ) . The rest of their hydrogen was composed of 1H , most likely derived both from AOC molecules as well as through hydrogen exchange with water [13] . These results allowed us to estimate the uptake of sugars from sugar mixtures , as will be discussed below . We then grew bacteria in chemostats that contained both arabinose and glucose , to determine the metabolic behavior of single cells grown on mixed substrates . We supplemented the chemostats with unlabeled sugars until they reached steady state , and then switched to 13C-arabinose and 2H-glucose ( S3 Fig ) . After an incubation period corresponding to 65% of the generation time we retrieved the cells from the chemostats and analyzed them with NanoSIMS . One expects that these cells contain isotopes of carbon and hydrogen from different sources . The two sugars provide a heavy isotope of C and H , respectively , but they also provide the light isotope of the other element; in addition , light isotopes stem from biomass assimilated before isotope labeling , and from AOC molecules as well as from water ( in the case of hydrogen ) . To correct for these effects , we constructed a mathematical model that described the isotope composition of individual cells based on input from all these sources ( see S2 File , ‘Mathematical Model’ ) . The model also included that cells assimilate AOC and exchange hydrogen with water , but did not include variation between cells in these parameters , since we had no measurements that would allow estimating this for single cells . This model allowed estimating the sources of carbon and hydrogen for the cells grown in our chemostats; these estimates are depicted in Fig 1A . We will describe further below how we also used this model to estimate sugar assimilation for individual cells . We then analyzed the 13C excess atom fraction as a marker for arabinose assimilation and the 2H excess atom fraction as a marker for glucose assimilation for each cell . In chemostats that were carbon-limited we observed variation in the excess atom fractions of 13C and 2H between cells , and also variation in the ratio of these two isotope fractions ( Fig 1B ) . This indicates that bacteria consumed both sugars simultaneously , but that there was heterogeneity between cells in the total amount and the relative fraction of the two sugars assimilated . In contrast , cells from chemostats that were nitrogen-limited did not assimilate 13C from arabinose , as they grew under conditions of carbon excess . This observation is consistent with a large body of work on catabolite repression: many bacteria growing in the presence of high concentrations of glucose and a second sugar consume exclusively glucose [30] . While this analysis revealed variation in content of the two isotope markers between individual cells , it did not provide direct quantitative information about the rate at which the two sugars were assimilated; to obtain quantitative information , we needed to take into account the different sources of the two isotopes , as illustrated in Fig 1A . From the mathematical model described above we derived a proxy for the degree of assimilation specialization , which we refer to as s ( see S2 File , ‘Mathematical Model’ ) . A value of s = 0 indicates that cells are not specializing on either sugar , and consume and assimilate the sugars in the proportion they are available . s = -1 indicates that a cell only assimilates arabinose , and s = 1 indicates that a cell only assimilates glucose . We discovered large and unimodal variation in the degree of specialization between individual cells . While most cells did assimilate measurable amounts of both sugars , the ratio between glucose and arabinose that individual cells assimilated varied substantially ( Fig 2A ) . This variation was unimodal in the sense that the degree of specialization of individual cells varied around one common mean value; we found no evidence of two distinct groups of cells , with one group specializing on glucose and the other on arabinose . For comparison , we conducted the same analysis with a strain of E . coli that was recently isolated from an environmental source [38] , and found qualitatively similar results ( Table 1 , S4 and S5 Figs ) . This indicated that the large unimodal variation in the degree of specialization that we observed was independent of adaptation to laboratory conditions . The mathematical model did not only allow to estimate the relative assimilation of the two sugars , but also allowed estimating the total sugar consumption and assimilation of each cell , and thus the increase of its biomass through cellular growth during the incubation period with 13C-arabinose and 2H-glucose . This analysis again revealed large variation in growth rates between individual cells growing in carbon-limited chemostats ( Figs 2B and S6 ) . Two estimates illustrate the magnitude of this variation: the growth rate of the 20% most active cells exceeded the growth rate of the 20% least active cells by the factor of 5 . 5 . And if all cells would grow as fast as the most active 20% , the population would grow 2 . 14 times faster . These results reveal large variation between cells in their total metabolic activity and growth , similar to what has recently been reported by [13] . Of note , variation in growth rate was more than three times higher in chemostat cultures than in batch cultures ( S7 Fig; for the analysis of isotope enrichment in batch cultures see S2 File , ‘Mathematical Model’ ) . This suggests that restricted growth in carbon- or nitrogen-limited chemostats promotes heterogeneity in the utilization of carbon sources , which translates into growth rate heterogeneity . We see this result as an interesting contrast to the conventional perspective that cells growing in chemostats are in physiological steady state [1] . The next question we addressed was whether the degree of specialization on glucose versus arabinose depended on the rate at which single cells grew . Did all cells show large variation in specialization , independently of how fast they grew ? To address this question , we analyzed the proxy for assimilation specialization , s , as a function of the single-cell growth rate . This analysis revealed that preference for sugar assimilation did not depend on the single-cell growth rate ( Fig 2C ) , and that the large variation in specialization that we observed was a robust pattern across cells with different growth rates . Interestingly , we observed that most cells assimilated more carbon from arabinose than from glucose ( Fig 2A ) . This does not mean that they consumed more arabinose , since both glucose and arabinose are expected to be depleted in such carbon-limited chemostats [31 , 32] . Rather , it is possible that glucose was primarily used for energy production ( with its carbon atoms not being assimilated into biomass ) and arabinose primarily as a cellular carbon source , similar to what has been found in a recent study that investigated simultaneous utilization of methanol and succinate by Methylobacterium extorquens [39] . Overall , these direct measurements of metabolic activity based on stable isotope-labeled nutrients revealed large variation in total assimilation between single cells , as well as in the relative proportion of the two sugars assimilated . This raises the question about the cellular basis of this variation in metabolism . Specifically , we asked whether the variation in metabolic activity was mirrored in the expression of genes involved in sugar uptake and assimilation , and was potentially caused by variation in gene expression between cells . To this end we used fluorescent reporters to analyze transcription in single cells . To investigate transcription of sugar uptake genes , we worked with strains that carried fluorescent transcriptional reporters for genes involved in the uptake of glucose and arabinose . Glucose can be taken up by at least five different systems in Escherichia coli , including the three main glucose transporters PtsG/Crr , ManXYZ and MglBAC [40 , 41 , 42] . Furthermore , E . coli has two arabinose transport systems , AraE and AraFGH [43 , 44] . We tested transcriptional reporters for these genes [45] and found that their expression patterns varied depending on the availability and concentration of carbon sources ( S8 and S9 Figs ) . Moreover , bacterial growth rate , i . e . dilution rate in chemostats , also affected expression patterns of transcriptional reporters ( S10A Fig ) . From this set of reporters , we selected two genes for more detailed analysis , namely the glucose phosphotransferase system PtsG and the arabinose-proton symporter AraE; these genes showed a broad dynamic range of expression when tested in different conditions ( S10 Fig ) , and previous studies had reported that these two genes have a high degree of phenotypic variation in clonal populations [17 , 46] . To analyze the variation and covariation in the expression of these two genes in single cells , we constructed a strain in which two different fluorescent proteins served as transcriptional reporters for the two genes; this strain carried ParaE-gfp and PptsG-mCherry on its chromosome . We used fluorescence microscopy to quantify the levels of GFP and mCherry in individual cells . We then analyzed the expression of these transcriptional reporters both by comparing population means across different carbon source conditions as well as by comparing individual cells within a given condition . We found consistent changes in mean expression between chemostat conditions ( Fig 3A ) : the promoter of ptsG was generally more active when bacteria were grown in the presence of glucose , while the promoter of araE was active when arabinose was present and not when only glucose was present . When analyzing the signal of the transcriptional reporter at the level of single cells ( Figs 3B and S11 ) , we found large variation in reporter expression levels between cells . Again , and importantly , this variation was unimodal around the population mean . We found no evidence for the emergence of two discrete groups of cells that each would specialize on the expression of genes to take up one of the two sugars ( of note , the apparent formation of discrete groups with different gene expression phenotypes in cultures grown on 20 μM arabinose in Fig 3B is a consequence of differences in gene expression patterns between different replicate populations , rather than caused by the co-existence of differently specialized cells within the same replicate population ) . The analysis of transcriptional reporters of single cells growing on mixtures of arabinose and glucose revealed a positive association between araE expression and ptsG expression ( Fig 3B , Table 1 ) . This observation is in line with the notion of substantial ‘global noise’ in gene expression [8] . Global noise manifests as variation in overall gene expression levels between individual cells , and can arise from differences between cells in the concentration of cellular components that modulate rates of transcription and translation . One possible source of global phenotypic noise is variation between cells in growth rates . Many bacterial promoters show an increase in activity with increasing growth rates [47 , 48 , 49] . One would thus expect that even among genetically identical cells growing under homogeneous conditions there might be an association between single-cell growth rates and overall transcriptional activity . As a first approach to address this question , we tested for an association between the expression levels of the reporters for the sugar transporters ptsG and araE and the ribosomal gene rpsM , which encodes the ribosomal protein S13 . The rationale of this experiment was to use the expression of ribosomal proteins as a proxy for growth rate [47 , 50] . This analysis revealed a significant correlation between the expression of reporters for sugar transporters and for rpsM ( Table 1 ) , in line with the notion that the activities across different types of promoters within a given cell were positively correlated , and potentially also correlated with the cell’s growth rate . The observation of large degrees of cell-to-cell variation both in metabolic activity and in the expression of sugar uptake genes raised the question about how these two types of cellular processes were linked: was the relative specialization of a cell for glucose or arabinose mirrored in the expression levels of genes involved in uptake of glucose and arabinose measured at the same time ? Such a pattern would indicate a direct link between a cell’s expression level of ptsG and araE and its uptake and assimilation of glucose and arabinose , respectively . To address this question , we used an experimental approach to link the analyses of gene expression and sugar assimilation: we used fluorescence microscopy to analyze gene expression patterns in clonal groups of bacteria , and then used NanoSIMS on the same group of bacteria to analyze substrate assimilation at the single-cell level . For each cell , we therefore acquired four pieces of quantitative information: expression of ParaE-gfp , expression of PptsG-mCherry , assimilation of 13C-arabinose and assimilation of 2H-glucose ( Fig 4A ) . We then tested for a statistical association between each cell’s expression level of ptsG and its glucose assimilation , and of its araE expression and arabinose assimilation . We found no significant direct correlation between the transcriptional signal of the transporter gene and the assimilation of the corresponding sugar , neither in carbon-limited chemostats nor in carbon-excess batch cultures ( Fig 4B ) . However , we found evidence that each cell’s specialization in gene expression was linked to its specialization in sugar assimilation in carbon-limited chemostats ( Fig 4C ) . Specifically , we observed that cells that transcribe more ptsG than araE also tend to assimilate more glucose than arabinose . This analysis is based on two proxies of ‘specialization’ that we derived . First , we calculated the ‘transcriptional specialization’ of individual cells; we standardized reporter gene expression levels across different measurements and transformed them to obtain a proxy for the degree of specialization in gene expression . Second , we calculated the ‘sugar specialization’ of these cells in the same way , by standardizing and transforming isotope enrichment levels to obtain a proxy for the degree of specialization in sugar assimilation ( see Methods , ‘Analysis of filters with fluorescence microscopy and NanoSIMS‘ ) . We found that cells that have higher PptsG-mCherry signal than ParaE-gfp signal also have higher 2H excess atom fraction compared to 13C excess atom fraction . In other words , a cell that specializes on ptsG expression rather than araE expression tends to specialize on glucose assimilation rather than arabinose assimilation , and vice versa ( Fig 4C ) . This indicates that in environments containing different sugars , the cell’s preference in sugar uptake and assimilation is at least partially based on specialization in transcriptional activity .
Our first main finding is that clonal cells in carbon-limited chemostats growing in the presence of two sugars show variation in cellular carbon assimilation rates ( Fig 1B ) , and therefore presumably variation in the rate at which single cells grow ( Fig 2B ) . This result is in line with results from recent chemostat experiments that were also based on measuring the incorporation of stable isotopes [13] or based on measuring fluorescent sugar analogues [51] , and thus supports the growing notion that well-mixed chemostat cultivation can promote cell-to-cell heterogeneity [51] . There is one potential additional aspect of metabolic heterogeneity that our approach did not allow to quantify: it is possible that cells growing in carbon-limited chemostats could also differ in the assimilation of AOC . Cell-to-cell variation in assimilation and growth rate can also arise in situations where cells grow on only one substrate . Our experiments in which clonal populations were grown in nitrogen-limited chemostats with excess of carbon sources indicates that even though carbon catabolite repression forces cells to utilize solely glucose , cells differ in the amount of assimilated glucose , which then manifests as the variation in single-cell growth rate ( Fig 1B ) . It has been argued that limitation in other nutrients affects the degree of glucose utilization [52] . In our experiments , it is possible that limitation in nitrogen poses restrictions on glucose metabolism that do not affect every cell equally . We also found that the variation in growth rate is significantly higher in chemostat cultures than in batch cultures . It is plausible that growth restriction imposed by the chemostat dilution rate generates an additional layer of heterogeneity that could manifest as variation in the rate at which single cells grow . Similar findings have been reported in recent studies that revealed the highest cell-to-cell variation in growth rates in the slowest growing cultures [13 , 18] . The dilution rate can significantly affect gene expression patterns ( S10A Fig ) , as well as intracellular fluxes [2] and the physiology of chemostat populations [31 , 32] , and it seems conceivable that these effects could contribute to variation in growth rates . Our second main finding is that cells that grow in the presence of two sugars do not all assimilate similar amounts of these two sugars; rather , there is strong variation between single cells in their specialization on glucose versus arabinose ( Fig 2A ) . While we did not find two metabolically distinct groups ( but rather a continuous variation between cells in their sugar specialization ) , this result gives credence to the idea that there is substantial variability in the metabolic reactions performed by cells in clonal populations . As discussed in the Introduction , this notion of individual cells differing in the sets of reactions they perform raises the question whether such specialization can reduce biochemical conflicts within cells and increase the rate at which single cells grow . Different cellular mechanisms could cause such cell-to-cell variation in sugar assimilation . Metabolic differences between genetically identical cells are often attributed to cell-to-cell variation in the transcription and translation of the relevant genes [7 , 9 , 28] . Alternatively , and not mutually exclusively , metabolic variation could potentially reflect differences in distribution of nutrient transporters due to for instance molecular sieving effect [53] or inner membrane protein crowding [54] , and asymmetric partitioning [55] of the nutrient transporters as well as cytoplasmic macromolecules for the storage of energy or carbon [56] during cell division . Here , we asked whether we can explain part of the cell-to-cell variation that we observed in our NanoSIMS analysis by variation in the transcriptional activity between cells . Such an analysis is interesting because the combination of fluorescence reporters and NanoSIMS allows to directly test for a link between transcriptional activity and actual metabolism–something that is usually difficult to do . Interestingly , when focusing on the assimilation of individual sugars from the mixtures of arabinose and glucose , we found no direct correlation between transcriptional initiation of the genes encoding the sugar transporters and assimilation of the respective sugars ( Fig 4B ) . A first potential reason for this lack of a direct correlation is the redundancy of sugar transporters . E . coli has five different transporters to take up glucose [40 , 41] and two different transport systems for arabinose [43 , 44] . The expression of these systems is adjusted according to the carbon source concentrations in the environment and bacterial growth rates . It is thus plausible that individual cells employ different transporters to a different extent ( S9 Fig ) [46] . One experimental approach to investigate the role of transporter redundancy in our system would be to delete all genes encoding alternative sugar transporters from the genome . However , deleting these genes is expected to cause pleiotropic effects on several phenotypic traits including carbon utilization , as recently investigated in more details at the level of bulk populations [42] , and is thus not easily feasible . A second potential reason for the discrepancies between the transcriptional signal and a cell's actual activity could be caused by post-transcriptional regulation , including regulation of the mRNA stability , regulation of the production and the stability of the protein , and through post-translational modifications as for instance acetylation [57] . Indeed , it has been shown that ptsG expression is post-transcriptionally regulated by the small RNA SgrS [41] ( post-transcriptional regulation of araE expression has not been reported [44] ) . Furthermore , even if transcriptional variation does translate into differences in enzyme concentrations and activities between cells , this does not necessarily lead to metabolic differences [3 , 58 , 59] . Almost all enzymes that are part of the central metabolic pathway are overabundant in the cell’s cytoplasm [60] , so that enzyme variation should have little effect on the flux through a pathway ( however , see [18] for a contrary example for how variation in the expression of an enzyme can directly affect bacterial growth ) . Despite this lack of a direct correlation between transcription of a particular transporter gene and the assimilation of the sugar taken up by the corresponding transporter ( Fig 4B ) , we found evidence for a different link between transcription and assimilation ( Fig 4C ) . As described in the Results , we observed that cells that specialize on transcribing ptsG ( encoding the glucose transporter ) over transcribing araE ( encoding the arabinose transporter ) also tended to specialize on the assimilation of glucose over arabinose ( Fig 4C ) . How is this consistent with the fact that we did not find a direct correlation between transcription of a transporter-encoding gene and the assimilation of the sugar taken up by the respective transporter ? One possible explanation for this apparent inconsistency is the high level of variation in total carbon assimilation between cells that we observed ( Fig 1B ) . If this variation in total assimilation is unrelated to transcription of ptsG and araE , then it is expected to obscure any direct correlation between transcription of a transporter-encoding gene and assimilation of the corresponding sugar . Determining the degree to which individual cells specialize on one sugar versus the other , as we did , removes this effect of variation in the total assimilation–and this analysis might therefore reveal the link between transcription and assimilation that manifests in Fig 4C . In summary , metabolic variation between cells might fundamentally be determined–at least in part–by what sets of genes these cells transcribe: cells that specialize more on the transcription of a gene encoding a certain sugar uptake system also show stronger specialization in the assimilation of this sugar . From a broader perspective , we see an interesting connection between cell-to-cell variation in clonal populations and the functioning of genetically diverse microbial communities . One major question about diverse microbial communities is how different types of microorganisms complement each other to together form a distributed metabolic network . It is interesting to ask whether such a distributed metabolism can also arise–to a lesser extent–within clonal populations . If genetically identical cells in clonal populations perform different sets of metabolic processes , this could potentially alleviate metabolic incompatibilities and allow individuals to reap the benefits of metabolic specialization . One example of a distributed metabolism in microbial communities is a process called cross-feeding: one microorganism partially degrades a primary resource and generates a metabolic intermediate that is excreted and used as a resource by a second microorganism . Although this process has been mostly implicated to occur between genetically different populations [61 , 62] , there is some evidence of phenotypic cross-feeding . Transcriptional analysis suggests that under certain conditions clonal populations of E . coli [46] , Pseudomonas putida [63] , and potentially Rhodopseudomonas palustris [64 , 65] might differentiate into subpopulations that specialize in different parts of a catabolic pathway . We see our study of single-cell assimilation as a contribution towards understanding the extent and the relevance of metabolic differences between genetically identical cells . Using stable isotope-labeled substrates in combination with single-cell mass spectrometry and mathematical modeling offers a powerful method for such an analysis . This method complements the more established approach of using transcriptional or translational fluorescent reporters alone . Furthermore , the method does not require genetic manipulation , and is thus applicable to all the diverse types of microbes that can be cultured in the laboratory . We expect that such studies will reveal how widespread metabolic diversity is in clonal populations , and how important it is for the activities and metabolic potential of these populations .
All experiments were performed with Escherichia coli K-12 MG1655 [36] and its derivatives ( see S2 Table ) . We also used a natural isolate of Escherichia coli , an enteroaggregative pathogenic strain 55989 ( CRBIP14 . 5 ) [38] , obtained from CRBIP-Institute Pasteur , Paris , France . A 2473 base-pair long dual reporter system was made by total synthesis by DNA2 . 0 in Menlo Park , CA , USA ( S1 Sequence ) . More details are provided in S1 File , ‘Supplementary Methods’ . Frozen strains were streaked once on LB agar plates to obtain single colonies . A single colony was inoculated overnight at 37°C in minimal media containing 1x M9 salts ( Sigma-Aldrich ) , 1 mM MgSO4 ( Fluka ) and 0 . 1 mM CaCl2 ( Sigma-Aldrich ) , supplemented with 3 mM D-glucose ( Glc ) ( Sigma ) , 3 mM L-arabinose ( Ara ) ( Sigma-Aldrich ) and 5% ( v/v ) LB complex broth ( total 4 ml ) ( Sigma-Aldrich ) . The 1000-fold diluted overnight cultures were used to inoculate precultures ( total 4 ml ) in defined minimal media containing 47 . 76 mM Na2HPO4 ( Sigma ) , 23 . 6 mM KH2PO4 , 8 . 56 mM NaCl , 20 . 2 μM NH4Cl , 1 mM MgSO4 ( all from Fluka ) and 0 . 1 mM CaCl2 . Sugar concentration was as follows: 10 μM Glc and 10 μM Ara in carbon-limited media , 3 mM Glc and 3 mM Ara in carbon-excess media , solely 20 μM Glc or solely 20 μM Ara for single-substrate studies . Carbon-excess chemostat cultures were nitrogen-limited ( limitation in NH4Cl ) thus all chemostats had similar bacterial population size . The precultures were grown for 12 hours until mid-exponential phase ( for growth data see S5B Fig ) , and 1 ml of precultures was used to inoculate glass mini-chemostats [2] . The dilution rate for chemostats operation was increased in 2 steps , from the minimal speed of the inflow pump ( IPC-N model , Ismatec , IDEX Health & Science , Germany ) until 0 . 15 h-1 was reached after 14 h . Chemostats were harvested after 5 volume changes , which is the minimum number of volume changes suggested for reaching the steady-state in the respective mini-chemostat system [2 , 46] . The outflow pump ( IP model , Ismatec , IDEX Health & Science , Germany ) was attached to a syringe that was positioned at the liquid surface in the chemostats , and operated at 20 ml per minute . This kept a constant volume and supplied the culture with flow of filter-sterilized , water-saturated air [2] . Each mini-chemostat contained a magnetic stir bar inside; a rack with the mini-chemostats was placed in a waterbath at 37°C , and the waterbath was placed on a magnetic stirrer . Continuous supply of air-bubbles together with magnetic stirring facilitated adequate mixing of the bacterial cultures in the mini-chemostat system . The volume of chemostat cultures was in total 5 . 6 ml and the pump rate was set to 0 . 84 ml h-1 , which means that bacterial generation time was 4 . 62 hours , and that chemostats needed to operate 6 . 67 hours to get one volume change . pH during the experiments remained constant at pH = 7 . 1 . The average OD600 of cultures grown in carbon-limited and carbon-excess chemostats was 0 . 008 ( with the standard errors of the mean of 0 . 001 and 0 . 0007 , respectively ) . Colony count was on average 2 . 45 x 106 ml-1 in carbon-limited and 2 . 4 x 106 ml-1 in nitrogen-limited carbon-excess chemostats . After completing five volume changes in mini-chemostats , media-flow was switched to media bottles containing stable isotope-labeled carbon sources , which were in the same concentrations as in the unlabeled media . The defined minimal medium was supplemented with D-glucose labeled with deuterium 2H ( D-[UL-2H12]-glucose , product number 616338 , 97% labeling , Sigma-Aldrich , Switzerland ) and L-arabinose labeled with 13C ( L-[UL-13C5]-arabinose , 99% labeling , ANAWA Trading SA , Switzerland ) . The cultures were run for additional 3 hours ( 45% of volume change , corresponding to 65% of generation time ) in the mini-chemostats and then harvested . As a control , we also prepared cultures of strain NN114 grown in carbon-limited chemostats without incubation with stable isotope-labeled sugars . More details are provided in S1 File , sections 'Fixation of bacterial samples' , 'Maximal incorporation of the stable isotope-labeled sugars in chemostats' , and 'AOC contamination in the chemostat setup' . Batch cultures of strain NN114 were grown in media supplemented with 3 mM Glc and 1 . 5 mM Ara; other components of the medium were as used for chemostat cultivation . Exponential populations had starting A600 ( absorbance at 600 nm ) of 0 . 001 and total volume of 200 μl , and were incubated in the plate-reader Eon ( BioTek ) at 37°C , with continuous shaking . The populations were growing for 9 h until the cells assimilated approximately half of supplemented Glc leaving 1 . 5 mM unlabeled Glc and Ara in the cell suspension . Then we added isotopically labeled sugars 2H-Glc and 13C-Ara to a final concentration of 1 . 5 mM each , by adding 2 μl of a 150 mM stock solution to the cell suspension . This resulted in a medium composition of 3 mM Glc and 3 mM Ara , both labeled approximately with 50% 2H and 13C , respectively , at the time when the stable isotope incubation started . We used absorbance , glucose , and yield measurements to more precisely estimate the isotopic labeling of glucose in each replicate . After 36 minutes , the cultures were fixed in formaldehyde and stained as described in S1 File , section ‘Fixation of bacterial samples’ , and applied on filters for NanoSIMS measurements , as described in S1 File , section ‘Sample preparation for NanoSIMS’ . As a control , we also prepared reference filters with unlabeled cells grown in the same conditions , which we used to determine “background” or natural isotopic composition and “filter-reference” ( see the section below , ‘NanoSIMS data analysis’ ) . This filter contained cells of strain NN114 growing in batch culture with Glc and Ara , without incubation with stable isotope-labeled sugars . By using NanoSIMS , we collected count data for the mass fragments 1H , 2H , 12C and 13C , together with the detection of secondary electrons ( Esi ) that provides information about surface topography ( Fig 4 ) . Images were processed using Matlab-based script Look@NanoSIMS [66] . The images were first corrected for a possible drift of the sample during the measurement , and then the counts in each pixel were accumulated over the measured z-planes . We discarded planes that were acquired incorrectly . Cell outlines ( region of interest , ROI ) were drawn according to secondary electron images . The accumulated counts c were averaged over the area of a cell , and the uncorrected atom fractions for carbon , X ( 13C ) cell/uncorrected , and hydrogen , X ( 2H ) cell/uncorrected , were calculated as ratios: X ( 13C ) cell/uncorrected=c[13C]/ ( c[13C]+c[12C] ) X ( 2H ) cell/uncorrected=c[2H]/ ( c[2H]+c[1H] ) To infer measurement error , Poisson percentage and standard errors of the ratio were determined by Look@NanoSIMS . These statistical errors represent the theoretical precision of the reported mean ratio , and the exact calculation is described in [66] . For statistical analysis of stable isotope-labeled cells we used only those cells with Poisson percentage errors below 10% for c[2H]/c[1H] , and below 1% for c[13C]/c[12C] ( S12 Fig ) . For each image we calculated atom fractions outside ROI , X ( 13C ) filter/cells , X ( 2H ) filter/cells . We also measured the reference filter with unlabeled cells to calculate atom fractions outside ROI ( X ( 13C ) filter/reference , X ( 2H ) filter/reference ) and atom fractions of the unlabeled cells as the background isotopic composition ( X ( 13C ) cells/natural = 0 . 009928 , X ( 2H ) cells/natural = 0 . 0002642 ) . For each image , the atom fractions were normalized to correct for day-to-day variation of the instrument by subtracting from the uncorrected atom fractions the difference between the measured atom fractions outside ROI of the analyzed filter ( X ( 13C ) filter/cells , X ( 2H ) filter/cells ) , and atom fractions outside ROI of the reference filter ( X ( 13C ) filter/reference , X ( 2H ) filter/reference ) as: X ( 13C ) cell=X ( 13C ) cell/uncorrected−[X ( 13C ) filter/cells-X ( 13C ) filter/reference] X ( 2H ) cell=X ( 2H ) cell/uncorrected−[X ( 2H ) filter/cells-X ( 2H ) filter/reference] The obtained corrected atom fractions ( X ( 13C ) cell , X ( 2H ) cell ) were used to calculate the excess atom fractions ( XE ( 13C ) cell , XE ( 2H ) cell ) by subtracting the background isotopic composition: XE ( 13C ) cell=X ( 13C ) cell-X ( 13C ) cells/natural XE ( 2H ) cell=X ( 2H ) cell-X ( 2H ) cells/natural More details are provided in S1 File , sections 'Sample preparation for NanoSIMS' , and 'NanoSIMS measurements' . | This study addresses a fundamental question in bacterial metabolism: do all individuals in a clonal population express the same metabolic functions , or do individuals specialize on different metabolic functions and assimilate different substrates ? Reports about stochastic gene expression in bacterial populations raise the possibility that transcriptional differences between individuals translate into different metabolic behaviors , but the prevalence and magnitude of such effects is currently not known . Here , we quantified the assimilation of two isotope-labeled sugars by single Escherichia coli cells using nanometer-scale secondary ion mass spectrometry , an analytical approach seldom used in systems biology . By comparing sugar assimilation and gene expression dynamics , we were able to differentiate the metabolic profiles of individual cells . We observed a previously hidden level of cell-to-cell variation in metabolism: cells differed both in the total amount of sugar they assimilated , as well as with respect to which of the two sugars they preferentially assimilated . Intriguingly , a cell’s preference in sugar assimilation was only partially based on specialization in gene expression . Taken together , this study is a step towards understanding the magnitude and the relevance of metabolic differences between genetically identical cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"physiology",
"carbohydrate",
"metabolism",
"chemical",
"compounds",
"particle",
"physics",
"biological",
"cultures",
"carbohydrates",
"cell",
"metabolism",
"organic",
"compounds",
"glucose",
"glucose",
"metabolism",
"dna",
"transcription",
"atoms",
"composite",
"... | 2017 | Cell-to-cell variation and specialization in sugar metabolism in clonal bacterial populations |
Auditory neurons encode stimulus history , which is often modelled using a span of time-delays in a spectro-temporal receptive field ( STRF ) . We propose an alternative model for the encoding of stimulus history , which we apply to extracellular recordings of neurons in the primary auditory cortex of anaesthetized ferrets . For a linear-non-linear STRF model ( LN model ) to achieve a high level of performance in predicting single unit neural responses to natural sounds in the primary auditory cortex , we found that it is necessary to include time delays going back at least 200 ms in the past . This is an unrealistic time span for biological delay lines . We therefore asked how much of this dependence on stimulus history can instead be explained by dynamical aspects of neurons . We constructed a neural-network model whose output is the weighted sum of units whose responses are determined by a dynamic firing-rate equation . The dynamic aspect performs low-pass filtering on each unit’s response , providing an exponentially decaying memory whose time constant is individual to each unit . We find that this dynamic network ( DNet ) model , when fitted to the neural data using STRFs of only 25 ms duration , can achieve prediction performance on a held-out dataset comparable to the best performing LN model with STRFs of 200 ms duration . These findings suggest that integration due to the membrane time constants or other exponentially-decaying memory processes may underlie linear temporal receptive fields of neurons beyond 25 ms .
The tuning properties of auditory neurons are commonly described by a spectro-temporal receptive field ( STRF ) model , which characterizes the linear dependence of the neural response on the sound spectrum at a range of latencies [1–14] . A static non-linearity is often applied to the output from the linear STRF—this linear non-linear ( LN ) model estimates the responses of neurons significantly better than the linear estimate [15 , 16] . While STRF models are somewhat successful in explaining the dependence of neural responses on the past few hundred milliseconds of stimulus history [17] , these models do not show how this temporal aspect of receptive fields might be implemented biologically . The most direct biological interpretation of the STRF model would involve auditory cortical neurons receiving inputs at a range of simple delays spanning out to 200 ms; however , this is biologically unrealistic . The onset latencies of neurons in the ventral division of the medial geniculate body ( vMGB ) , which provides the primary ascending input to primary auditory cortex ( A1 ) , are typically less than ~14 ms in the ferret [18] , the animal species that provided the data that were modelled in this study . In the cat vMGB response latencies are typically less than ~20 ms [19] , in the mouse less than ~10 ms [20] , and in the marmoset less than ~40 ms [21] . These values are far less than the duration of stimulus history which influences the responses of A1 neurons . Although a few vMGB units do show onset latencies beyond this range ( in cat up to 70 ms [19] , in mouse up to 60 ms [20] , and in marmoset up to 300 ms [21] ) , this small fraction is unlikely to be the main determinant of the dependence of the neural response in primary auditory cortex on stimulus history beyond a few tens of milliseconds . This is because there are other plausible mechanisms , such as the dynamic temporally-integrative properties intrinsic to neurons . In this study , using recordings from anesthetised ferrets , we asked how much of the dependence on recent stimulus history—the temporal receptive fields , of primary auditory cortical neurons can instead be explained by certain of these simple dynamical aspects of neurons–an approach more consistent with the known biology . To model a neuron’s response , we constructed a neural network model whose output is the weighted sum of the responses of multiple units , each of which resembles an LN model . However , the response of each unit of the network was modified in accordance with a dynamic firing-rate equation . This low-pass filters the unit’s response ( by convolving with an exponential decay impulse response ) , providing a simple exponentially decaying memory . This integrative characteristic can be related to the capacitance and resistance of nerve cell membranes , and the individual time constant of each unit can be related to the membrane time constant of real neurons [22] . The dynamic aspect of our network can alternatively be interpreted in terms of other neurobiological dynamical phenomena , such as channel-based neural adaptation , short term synaptic plasticity , or recurrent network properties . We found that our biologically-motivated dynamical model can accurately capture the temporal receptive fields of neurons without the need for a wide span of latencies of input .
In vivo electrophysiological recordings in ferrets were performed under ketamine ( 5 mg/kg/h ) and medetomidine ( 0 . 022 mg/kg/h ) anaesthesia . All animal procedures were performed under license from the UK Home Office and were approved by the University of Oxford Committee on Animal Care and Ethical Review . Models were fitted to single-unit neural responses of anesthetized ferrets to natural sound stimuli . Altogether , 20 sound clips were presented containing human speech in different languages , other animal vocalizations ( e . g . ferret ) and environmental sounds ( e . g . wind and water ) . All clips were 5 s long with a sampling rate of 48 , 828 . 125 Hz and with root mean square intensity ranging from 75 to 82 dB SPL . The electrophysiological data used in this study were taken from a series of experiments used in a previous study by Harper et al . [23] . Briefly , recordings were made from the primary auditory cortical areas , A1 and the anterior auditory field ( AAF ) of 6 adult pigmented ferrets ( 5 females and 1 male ) under ketamine ( 5 mg/kg/h ) and medetomidine ( 0 . 022 mg/kg/h ) anaesthesia for 20 repeats of the 20 natural sound clips played in random order . All animal procedures were performed under license from the United Kingdom Home Office and were approved by the local ethical review committee . In total , 56 penetrations resulted in 549 single and multi-units , of which 284 were single units . A single unit was taken for analysis only if its activity was driven by the stimulus according to the noise ratio ( see below ) . For each unit , the number of spikes was counted in each 5 ms time bin and averaged over repeats , to provide a response profile yn ( t ) , where t is time , and n is the clip number . The total number of time bins in a clip is T . For simpler notation , we will drop the subscript n , unless we note otherwise . The noise ratio was calculated as a measure of how much the response of each unit is dependent on the stimuli [9 , 24] . The noise ratio was measured over all 20 stimuli and 20 repeats . Any unit with a noise ratio > 40 were excluded from the study . Also , only putative single units were used . This resulted in 73 neurons for the study . The models receive the input as a cochleagram , a spectrogram-like transformation of the sound waveform . The cochleagram approximates the spectral filtering performed by the auditory periphery and was calculated as follows [23 , 25] . For each sound clip , the amplitude spectrum was measured using 10-ms Hanning windows , overlapping by 5ms . The number of frequency channels was then reduced by weighted summation using overlapping triangular windows to provide 34 log-spaced channels ( 500 Hz to 22 , 627 Hz center frequencies , adapted from melbank . m ( http://www . ee . ic . ac . uk/hp/staff/dmb/voicebox/voicebox . html ) ) . Next , a log function was applied to the value in each time-frequency bin , and any values below a low threshold was set to the threshold . The cochleagram was normalized to zero mean and unit variance over the training set ( see Cross-validation and model testing ) . The input xfq ( t ) at time t was the chunk of cochleagram preceding t , given by xfq ( t ) = Kf ( t−q+1 ) , where f is the frequency channel and q is the latency and Kf ( t ) is the cochleagram . The two stage LN model consists of a linear model followed by a sigmoid output nonlinearity . Linear stage: a ( t ) =∑f , qwfqxfq ( t ) +b ( 1 ) where a ( t ) is the model neuron's activation , wfq is the input weight and b is the baseline activation . Both w and b are free parameters of the model and were estimated by regressing neural response y ( t ) against input x ( t ) using glmnet [26] . To overcome overfitting , the weights were regularized using an L1-norm ( LASSO regularization ) . Thus a ( t ) can be seen as the best linear estimate of y ( t ) from x ( t ) . The regularization hyperparameter λ controlled the strength of the regularization . To select λ , the model was subjected to k-fold cross-validation ( see Cross-validation and model testing ) . Nonlinear stage: The nonlinear stage of the model was a parameterized sigmoid nonlinearity: ν ( t ) =ρ11+exp ( −a ( t ) −ρ3ρ2 ) +ρ4 ( 2 ) The four parameters ρi of the function were fitted by minimizing the squared error between v ( t ) and y ( t ) . The NRF model is an artificial neural network with a single hidden layer of 20 hidden units ( HU ) which converge onto an output unit ( OU ) . Each unit of the model is similar to a single LN model . The activation of the j-th HU , aj ( t ) is , aj ( t ) =∑f , qwjfqxfq ( t ) +bj ( 3 ) The output of the j-th HU , vj ( t ) is , vj ( t ) =g ( aj ( t ) ) ( 4 ) where g ( aj ( t ) ) is a sigmoid non-linear activation function , 1/ ( 1+exp ( −aj ( t ) ) ) . The HUs feed these outputs to the OU . The activation function of the OU ao ( t ) is , ao ( t ) =∑jwjvj ( t ) +bo ( 5 ) The output vo ( t ) of the OU is , vo ( t ) =g ( ao ( t ) ) ( 6 ) which is the model’s estimate of the neuron’s response at time t . A neuron’s response depends on the activity of its presynaptic neurons through the total synaptic current they evoke . However , a neuron’s membrane potential , and consequently its firing rate , is not an instantaneous function of this current . Due to the capacitance and resistance of the neural cell membrane , the membrane potential and hence the firing rate is approximately a low-pass filtered version of the synaptic current . A common simple dynamic model of neurons that captures this phenomenon is the differential equation ( Eq 7 ) known as the firing-rate equation [22] . Here , τcont is the membrane time constant , v is post-synaptic firing rate , I ( tcont ) is the synaptic current at time tcont , F ( ) is a nonlinear function that relates synaptic current to the post-synaptic firing rate , and tcont is time . Here we use tcont and τcont with the superscript ‘cont’ to denote that for this equation these variables operate in continuous time rather than discrete time . τcont determines how rapidly firing rate approaches steady state for a constant synaptic current , and can be related to the membrane time constant [22] . Although the relationship is not strict , we shall call τcont the membrane time constant . This is equivalent to convolving the instantaneous output of function F ( ) with an exponential decay impulse response . Eq 7 can be discretized by the Euler method , where c = 1 time-bin = 5 ms . We redefine v and I to depend on time-bin number , t , rather than continuous time ( t = tcont / c , where t is an integer ) . Hence , the discretized form of the firing rate equation is: v ( t ) = ( 1−1τ ) v ( t−1 ) +1τF ( I ( t−1 ) ) ( 8 ) The dynamic network model incorporates units that behave like Eq 8 . In recognition that the characteristics of units in our model may not exactly correspond to biophysical properties of neurons described by the firing rate equation , and to be consistent with the NRF model , we replace F ( ) with g ( ) and the current I ( t−1 ) with activation a ( t ) . As a result , activation aj ( t ) of the j-th HU of the DNet model is the same as the NRF model , aj ( t ) =∑f , QwjfQxfQ ( t ) +bj ( 9 ) However , the output vj ( t ) is different from that of the NRF model , incorporating the discretized firing-rate equation . Similarly , the activation ao ( t ) of the OU is , ao ( t ) =∑jwjvj ( t ) +bo ( 11 ) and the output vo ( t ) of the OU is , vo ( t ) = ( 1−1τo ) vo ( t−1 ) +1τog ( ao ( t ) ) ( 12 ) Here , τj and τo are the membrane time constants of j-th hidden unit and output unit respectively . The firing-rate equation ( Eq 7 ) assumes that the time constant that governs the relationship between the current and the firing rate is substantially larger than the time constant that governs the decay of post-synaptic potentials . However , if we assume the reverse , i . e . , that the synaptic time constant is substantially larger , a different simple model of the relationship between the firing rate and the current becomes appropriate [22] . This model is given by: τscontdIdtcont=−I+z ( tcont ) ( 13 ) v=F ( I ) ( 14 ) Here , I is synaptic current and v is post-synaptic firing rate , z ( tcont ) is the immediate influence of the presynaptic spikes on the synaptic current , F ( ) is the nonlinear function of synaptic current to post-synaptic firing rate , τscont is synaptic time constant , and tcont is time . Similar discretization , like that used for the DNet model ( see above ) , can be used for Eq 13 to use this for the units of a neural network model . Again , replacing the synaptic current , I , with the activation a , the activation of hidden units of the sDNet model can be modelled as the following: aj ( t ) = ( 1−1τsj ) aj ( t−1 ) +1τsj ( ∑f , QwjfQxfQ ( t ) +bj ) ( 15 ) And the hidden unit output , replacing F ( ) with g ( ) , is given by: vj ( t ) =g ( aj ( t ) ) ( 16 ) Hence , in the sDNet model , the activation of each HU , rather than the output of the nonlinearity , is convolved with an exponential decay impulse response . Similarly , for the output units , ao ( t ) = ( 1−1τso ) ao ( t−1 ) +1τso ( ∑jwjvj ( t ) +bo ) ( 17 ) And the output , vo ( t ) =g ( ao ( t ) ) ( 18 ) Here , τsj and τso are the synaptic time constants of j-th hidden unit and output unit respectively . The free parameters wjfq , wj , bj , and bo ( also τj and τo for the DNet and τsj and τso for the sDNet ) were optimized by minimizing the squared error between vo ( t ) and y ( t ) subject to L1-regularization of the weights . Thus , the objective function is given by , E=12NT∑n , t ( vo , n ( t ) −yn ( t ) ) 2+λ ( ∑j , f , q|wjfq|+∑j|wj| ) ( 19 ) Here , n is included to indicate clip number , but was left out of other equations for simplicity . N is the number of clips used in training . All models except for the LN model are fitted by minimizing the objective function with respect to the free parameters using the sum-of-function optimizer algorithm [27] . In using this algorithm , we take one clip to be one minibatch . The optimization algorithm requires calculation of the error gradients in respect to each of the parameters . For the NRF model , error gradients are calculated using standard chain rule ( backpropagation ) . For the dynamic models , the process is similar ( see below ) . For the network models , before training , the weights were initialized by modified Glorot initialization from a uniform distribution ranging from—1fq+l to + 1fq+l , where fq is the number of input weights to a HU and l = 1 is the number of output weights from a HU . The biases were initialized similarly [28] . For the DNet model , to prevent the 1τ term running into mathematical error ( when τ = 0 ) , we substitute 1τ with h ( d ) =11+d2 . Parameters , dj and do ( for hidden and output units ) were each independently initialized with the square root of a random variable from an exponential distribution with a mean of 1 . Similar , substitution is used for the sDNet model . To estimate the parameters , the gradients of the error with respect to various parameters are calculated using a method similar to the one used for real time recurrent learning [29] . Substituting 1τ=11+d2=h ( d ) , the dynamic equations for HUs ( Eq 10 ) and OU ( Eq 12 ) of the DNet model become respectively: vj ( t ) = ( 1−h ( dj ) ) vj ( t−1 ) +h ( dj ) g ( aj ( t ) ) ( 20 ) vo ( t ) = ( 1−h ( do ) ) vo ( t−1 ) +h ( do ) g ( ao ( t ) ) ( 21 ) Using the chain rule , the gradient of the error term , E ( t ) at time t in respect to the weights of the output layer , wj , ∂E ( t ) ∂wj=∂E ( t ) ∂vo ( t ) ∂vo ( t ) ∂wj ( 22 ) ∂vo ( t ) ∂wj= ( 1−h ( do ) ) ∂vo ( t−1 ) ∂wj+h ( do ) g′ ( ao ( t ) ) νj ( t ) ( 23 ) The prime denotes the derivative of the function . Similarly , the gradient of the error term , E ( t ) in respect to the bias of the output layer , bo , ∂E ( t ) ∂bo=∂E ( t ) ∂vo ( t ) ∂vo ( t ) ∂bo ( 24 ) ∂vo ( t ) ∂bo= ( 1−h ( do ) ) ∂vo ( t−1 ) ∂bo+h ( do ) g′ ( ao ( t ) ) ( 25 ) Finally , the gradient of the error term , E ( t ) with respect to the parameter , do , ∂E ( t ) ∂do=∂E ( t ) ∂vo ( t ) ∂vo ( t ) ∂do ( 26 ) ∂vo ( t ) ∂do= ( 1−h ( do ) ) ∂vo ( t−1 ) ∂do−h′ ( do ) vo ( t−1 ) +h′ ( do ) g ( ao ( t ) ) ( 27 ) Hence , the gradient is passed forward from the current time step to the next time step to be used in calculation of the gradient at that new time step . At t = 1 , the values of ∂vo ( t−1 ) ∂wj , ∂vo ( t−1 ) ∂bo and ∂vo ( t−1 ) ∂do are undetermined . So , at time t = 1 , the values of these terms are set to zero . The gradients of the error term , E ( t ) , with respect to the rest of the parameters of the DNet model and the sDNet model are obtained similarly . From the data for the 20 sound stimuli , 4 were chosen as a test set which was not used during training and cross-validation . The cross-validation set ( the remaining 16 stimuli ) was used to fit the models using k-fold cross validation , where k = 8 . The cross-validation set was randomly divided into a training set of 14 stimuli and a validation set of 2 stimuli . The model was trained on the training set for 18 different values of the hyperparameter λ . A log spaced range of lambda values was used , but with a somewhat lower density at the extremes . For the LN model , the exact values of λ used were: 1 . 00 x 10−1 , 2 . 00 x 10−2 , 1 . 17 x 10−2 , 6 . 84 x 10−3 , 4 . 00 x 10−3 , 2 . 34 x 10−3 , 1 . 37 x 10−3 , 8 . 00 x 10−4 , 4 . 68 x 10−4 , 2 . 74 x 10−4 , 1 . 60 x 10−4 , 9 . 36 x 10−5 , 5 . 41 x 10−5 , 3 . 20 x 10−5 , 6 . 40 x 10−6 , 1 . 28 x 10−6 , 2 . 56 x 10−7 , and 5 . 12 x 10−8 . For the rest of the models , the values of λ used were: 1 . 00 x 10−3 , 2 . 00 x 10−4 , 1 . 17 x 10−4 , 6 . 84 x 10−5 , 4 . 00 x 10−5 , 2 . 34 x 10−5 , 1 . 37 x 10−5 , 8 . 00 x 10−6 , 4 . 68 x 10−6 , 2 . 74 x 10−6 , 1 . 60 x 10−6 , 9 . 36 x 10−7 , 5 . 41 x 10−7 , 3 . 20 x 10−7 , 6 . 40 x 10−8 , 1 . 28 x 10−8 , 2 . 56 x 10−9 , and 5 . 12 x 10−10 . For each of the fitted models , neural responses were then predicted for the validation set , and the correlation coefficient between the actual neural responses and the prediction was measured . This process was repeated 8 times for different non-overlapping validation sets . The model was then retrained with the whole cross-validation set using the λ value that provided the highest mean correlation coefficient over all 8 folds . Next , the retrained network was used to predict the neural responses to the test set . All the correlation coefficients and normalized correlation coefficients shown are for this held out test set , and all the model parameters shown are for the retrained network . To ensure that all models receive the same amount of training data , models with different latency spans of STRFs are provided with exactly the same durations of neural response . This was done by clipping the necessary number of time points from the beginning of the neural response data . The hidden layer of each of the networks contained 20 HUs , but because their weights were L1-regularized , they tended to develop substantive weights only if they explained aspects of the neural response that were not explained by any other units . As a result , many HUs developed weights close to zero , and thus the network models tended to have only a small number of effective HUs . The ‘effectiveness’ of each unit j was calculated [23] as the variance over time of the unit’s weighted output wovj ( t ) . Any HU with variance greater than 5% of the sum of the variances of all 20 HUs is considered effective . The IE score [23] measures whether the degree to which a HU is excitatory or inhibitory , on a scale between -1 and 1 . Most inhibitory is IE = -1 , when all of the input weights are negative or zero , and the output weight is positive , or when all of the input weights are positive or zero and the output weight is negative . Most excitatory is IE = 1 , when all of the input weights are positive or zero , and the output weight is positive , or when all of the input weights are negative or zero and the output weight is negative . Because the membrane time constant is modelled as 1+d2 and the duration of each time step is 5 ms , the value of membrane time constant in ms is 5 ( 1+d2 ) . After fitting a network model , HUs with time constant greater than the length of the STRF are selected . The output weights connecting these units to the output unit are then set to zero . As a result , HUs with large time constants become inactive during the test process . Everything else in the model remained unchanged .
For each neuron , an LN model was fitted to estimate its firing rate as a function of the preceding cochleagram . The LN model consists of an STRF , followed by a sigmoid nonlinearity ( Fig 2A ) . As mentioned in the Methods , the STRF is the weighted sum of the cochleagram over frequency components at a span of latencies . The effect of different spans of latencies ( lengths ) of the STRF was investigated . For the held-out test set , and averaged over the 73 neurons , the mean normalized correlation coefficient , mean CCnorm [30 , 31] , between the actual neural responses and the prediction of LN models with 25 , 50 , 100 , 200 , and 400 ms long STRFs are respectively 0 . 51 , 0 . 64 , 0 . 69 , 0 . 71 , and 0 . 71 ( Fig 2B ) , where 1 . 0 is the maximum achievable prediction given the estimated neuronal noise and 0 . 0 indicates that there is no correlation between actual and predicted neural responses . The means of non-normalized Pearson correlation coefficients ( CC ) are 0 . 40 , 0 . 50 , 0 . 54 , 0 . 55 , and 0 . 55 respectively . Fig 2C shows STRFs of some example neurons estimated by the LN model . A network receptive field model ( the NRF model ) was also developed and fitted to each neuron . This model consisted of the weighted sum of multiple LN-like units . This model is the same as a previously published version [23] , except that it uses sigmoid nonlinearities instead of scaled tanh non-linearities . A sigmoid non-linearity was used to be consistent with the dynamic network model that we will come to , which requires non-negative outputs and hence also uses sigmoid non-linearities . The NRF model we use here is essentially an artificial neural network with a single hidden layer , using a sigmoid non-linearity in its hidden units and its single output unit ( Fig 3A ) . The NRF model makes use of multiple component STRFs , which allows modelling of neuronal sensitivity to non-linear interactions of multiple features [23] . STRFs of at least 400 ms are needed for best performance in predicting neural responses ( Fig 3B ) . For the test set , the mean CCnorm between neural responses and the model predictions made by NRF models with 25 , 50 , 100 , 200 and 400 ms long component STRFs are 0 . 51 , 0 . 64 , 0 . 70 , 0 . 72 and 0 . 73 , respectively ( mean CC 0 . 40 , 0 . 50 , 0 . 54 , 0 . 55 and 0 . 56 , respectively ) . The NRF model predicts the neural responses equal to or better than the LN model over all durations examined , although by a more modest amount than seen in an earlier study [23] . This may be because the NRF model in this study uses a sigmoid rather than tanh nonlinearity , or because the earlier study [23] used a different way of selecting the test set . Fig 3C shows the input weights ( analogous to STRFs ) for the ‘effective’ HUs ( see Methods ) of the model-fit for one example neuron , obtained using the NRF model . Fig 3D and Fig 3E show the ‘effectiveness’ ( see Methods ) and IE score ( see Methods ) of the units shown in Fig 3C . Although both the LN and NRF models can capture dependence on stimulus history using a span of input latencies extending up to at least 200 ms , there is little biological evidence for such a wide span of latencies in the auditory thalamic neurons that feed to auditory cortex ( see Introduction ) . With the aim of better understanding the biological underpinnings of stimulus history dependence in auditory cortical responses , a model with a dynamic aspect was developed that integrates the output of its component units over time in a manner similar to those of real neurons ( see Methods ) . Each unit in the DNet model includes exponentially-decaying response integration , motivated by the integrative properties of real neuronal membranes ( Fig 4A ) . This model is in essence the NRF model but with dynamic units , each modelled by the dynamic firing-rate differential equation ( Eq 7 ) [22] . That is , each unit of the model applies an exponentially decaying impulse response to the output of their non-linear activation function . Each unit’s decay rate is governed by its individually fitted time constant , τ ( see Methods ) . We found that the mean CCnorm between the actual neural responses and the predictions of the DNet model with 25 , 50 , 100 , 200 and 400 ms STRFs are respectively 0 . 71 , 0 . 71 , 0 . 70 , 0 . 68 and 0 . 68 ( Fig 4B ) ( mean CC 0 . 55 , 0 . 55 , 0 . 54 , 0 . 52 and 0 . 52 , respectively ) . Because the 25-ms model performs the best and the biological range of latencies are typically less than 30 ms[32] , we focused on the 25-ms DNet model for further analysis . Fig 4C shows the input weights ( analogous to STRFs ) for the ‘effective’ hidden units ( see Methods ) of the model fits for 8 example neurons , using a DNet model with 25-ms long STRFs . Fig 4D and 4E show the ‘effectiveness’ ( see Methods ) and IE score ( see Methods ) of the units shown in Fig 4C . The 25-ms DNet model out performs the DNet with longer ( 200 ms and 400 ms ) STRFs , as measured by prediction of neural responses on a held-out test set averaged over all 73 neurons ( Fig 5A ) . It also achieves performance similar to the best performing LN model , but performs a little worse than the best performing NRF model ( Fig 5A ) . This can be seen qualitatively in the time course of the prediction for an example neuron and stimulus in Fig 5B . A neuron by neuron comparison shows that this result is consistent for most neurons ( Fig 5C and 5D ) . Also , the 25-ms LN and NRF models are outperformed by a large margin by the 25-ms DNet model ( Fig 5A ) . A neuron by neuron comparison shows that this result is also consistent for almost all neurons ( Fig 5E and 5F ) . Note that the DNet model is equivalent to the NRF model when the discretized time constant for every DNet model unit is at the smallest possible value ( τ = 1 , d = 0 ) . Hence , because all possible NRF models are a subset of all possible DNet models , the DNet should in theory always equal or outperform the NRF model . The fact that this is not the case suggests that , at least in some cases , the very best fits are not being found for the DNet model , perhaps as a consequence of some overfitting . That the DNet model performs worse for large latency spans , where there are more variables , is again probably due to challenges in fitting the DNet model , likely due to some overfitting to the training set . To understand how much of an effect HUs of the DNet model with long time constants have on the model’s predictions , the fitted model was modified by ‘knocking out’ ( see Methods ) HUs with time constants longer than the duration of their STRFs . This was achieved by setting the output weights of those units to zero , while keeping all other parameters intact . We found that the 25-ms DNet model is the DNet whose CCnorm is most reduced by this alteration ( Fig 6A ) . A single-unit dynamic model was also developed to examine the effect of a time constant in an LN-like single STRF model . Prediction performance of this model was found to be worse than the LN model ( Fig 6B ) . This indicates that both long time constants and a network architecture are needed for the 25-ms DNet model to achieve prediction performance similar to that of standard models . Although the DNet model had 20 hidden units , the L1 regularization eliminated unnecessary units . Hence there were fewer effective HUs , and for the 25-ms DNet model this number was found to be between 2 and 8 ( Fig 6C ) . We defined an effective HU as one whose effectiveness ( variance of weighted output , see Methods ) exceeded a certain threshold . In the 25-ms DNet model , the time constants of the effective HUs ranged between 5 and 475 ms , often far greater than the duration of the STRFs ( Fig 6D ) . The time constants of the OUs of the DNet model ranged from 5 to 130 ms ( Fig 6E ) . To investigate whether long time constants are associated with effective HUs , for each neuron HUs were ranked in order of effectiveness and the mean of the time constants of all HUs in the same rank ( over all neurons ) were calculated . This showed that high-ranked HUs tend to have long time constants ( Fig 6F ) . To test whether a relationship exists between the time constant and unit’s inhibitory/excitatory nature , we classified the effective HUs of the 25-ms DNet model into excitatory or inhibitory using an IE score ( see Methods ) . If the output weight is positive , an IE score of 1 means a unit with entirely excitatory influence and an IE score of -1 means a unit with entirely inhibitory influence . We found that almost all units having large time constants are inhibitory in nature ( Fig 7 ) . In the DNet model , the integration governed by the time constant occurs after each unit’s nonlinearity . Although the time constant is conceived and implemented in a manner consistent with a membrane constant , within the network it may capture other dynamic neural phenomena such as synaptic time constants , forms of a channel-based or synaptic adaptation , or certain forms of network recurrency . An alternative network model places the integration just before the nonlinearity , which is more consistent with the time constant being synaptic in nature [22] . To investigate how much of the neural activity can be explained by such a model , we modified the DNet model to form the sDNet model ( synaptic DNet model ) , with each unit governed by Eqs 13 and 14 [22] . The sDNet model achieves best performance in predicting neural responses when the duration of the STRFs is 200 ms ( CCnorm = 0 . 71 ) ( Fig 8A ) . Although the sDNet with 25-ms STRFs predicts neural responses better ( CCnorm = 0 . 67 ) than the 25-ms LN model ( CCnorm = 0 . 51 ) and the 25-ms NRF model ( CCnorm = 0 . 51 ) , it is worse than the 25-ms DNet model ( CCnorm = 0 . 71 ) . This suggests that a large percentage of the variance of neural responses can be explained by a time constant before non-linearity , but that having a time constant after the non-linearity further improves prediction . While the largest time constants of the effective HUs of the sDNet model can be as large as ~470 ms ( Fig 8B ) , the distribution of time constants of output units have smaller values compared to the DNet model ( Fig 8C ) . As with the DNet model the more effective HUs of the sDNet model tend to have longer time constants ( Fig 8D ) and units with long time constants tend to be inhibitory in nature ( Fig 8E ) .
We find that a model of primary auditory cortical neurons which incorporates exponentially-decaying memory processes–the DNet model–can capture some of the temporal aspects of the receptive field using very short ( 25 ms ) component STRFs in a network . Standard models such as the LN model ( and the NRF model ) represent temporal receptive fields of neurons using delay lines whose latencies extend up to a few hundred milliseconds . Our results suggest that , beyond 25ms , the temporal receptive field can be modelled more succinctly by exponentially-decaying memory processes than by delay lines . When auditory neurons are fitted using a standard LN model , the resulting STRFs often contain narrowband inhibitory fields that decay over approximately 200 ms . In contrast , we find that the STRFs of the DNet model have short inhibitory fields with long time constants . This suggests that these temporally extended inhibitory fields in STRFs may reflect neural mechanisms that can be approximated by exponentially-decaying memory processes . Although they have not been described in the ferret , a few vMGB units do show onset latencies beyond a few tens of milliseconds: in cat latencies up to 70 ms have been described [19] , in mouse up to 60 ms [20] , and in marmoset up to 300 ms [21] . Areas other than vMGB that project to cortex may also include long latency responses . Hence , we cannot rule out the possibility that long-latency inputs at least partially explain the dependence of auditory cortical neurons on stimulus history–indeed it is likely that they do to some extent . However , these long-latency responses are relatively rare and we show that it is possible to explain a lot of the dependence of the neural response in primary auditory cortex by exponentially-decaying memory properties consistent with those properties intrinsic to neurons , such as membrane dynamics . There are many biological mechanisms which may underlie exponentially-decaying memory processes . Neurons have various intrinsic dynamic integrative and adaptive characteristics originating from basic membrane dynamics , channel-based adaptation , post-synaptic potentials and short-term synaptic plasticity [22] . Recurrent network effects could also underlie such memory processes . In a previous study , a model with synaptic depression has been shown to partially capture the integration of stimulus history in A1 [33] . Here , we found that a better prediction of neural responses can be achieved by a model based on the integrative properties of neural membranes ( the DNet model ) than by the one based on the dynamics of synaptic potentials ( the sDNet model ) . We note , however , that although the DNet model was developed with membrane dynamics in mind , aspects of all the above listed dynamic processes could be captured by the same exponentially-decaying memory in the DNet . These different processes are not exclusive of each other , and combinations of them may contribute to the effects we find . The firing rate of A1 neurons has been demonstrated to depend on stimulus history up to ~4 s in the past [17] . This dependence cannot be fully captured by linear models such as STRFs ( 17 ) . This indicates that there are some nonlinear dynamic processes that provide history dependence which cannot be captured either by a linear or LN models , nor by the DNet model . Perhaps this is because the DNet model has limited capacity to capture temporal dynamics that are not exponential in nature . Much of the relationship between stimulus and neural responses results from neurons being embedded in a network structure and from the nonlinear properties of the units in this network . From the auditory periphery to the A1 , there are numerous feedforward , feedback and lateral connections . Many previous models of auditory cortical neurons have not explicitly represented this network character . However , some recent models have incorporated non-linear contextual effects [14 , 25 , 34–36] . Furthermore , a few models do explicitly incorporate multiple receptive fields [37–43] or more extensive network structure [23] , which perform better in predicting neural responses of A1/AAF neurons than STRF models [23 , 37] . When modelling the dependence of neural responses on stimulus history , one approach is to include the history of a neuron’s spikes and those of its recorded neighbours in a generalized linear model ( GLM ) model [35 , 44] . However , unlike the DNet model , this approach does not model multi-layer network structure with hidden units that are inferred but not recorded . Modelling hidden network structure has been applied to the salamander retina , where in vitro responses to simple artificial stimuli are well predicted by a network model ( LNFDSNF ) with various non-linearities and response delays mimicking the retinal pathway [45] . The LNFDSNF model is particularly relevant to our study because it is a network model whose units have some memory of their past activity , though the form of this memory differs from that of the DNet model . The LNFDSNF model units have a feedback kernel that acts before the nonlinearity—in contrast , the DNet model units integrate their activity after the nonlinearity using an infinite impulse response that decays exponentially . The succinct and biologically-motivated description of neural responses provided by the DNet model provides a basis from which to develop further models to account for the ~30% of the CCnorm that remains to be explained . One variant of the model that can be developed involves a network with multiple hidden layers where the units each have just a single latency ( i . e . each unit in the first hidden layer uses just a spectral receptive field at a single delay ) . This approach provides an opportunity to build deeper neural-network-like models with fewer parameters , potentially overcoming overfitting problems for deep network models of neural responses [46] . This network can provide a basis for new network models that incorporate a greater range of intrinsic dynamic processes more explicitly , for example modelling in each unit its basic membrane dynamics , synaptic depression and channel-based adaptation together , with different appropriate time constants in each case . Such models may show more predictive power . Neurons are intrinsically dynamic systems , each with memory of past events . Furthermore , neurons are embedded in networks . We show that models incorporating both these characteristics can be valuable for explaining moment to moment in vivo responses of neurons in the auditory cortex . This is an arguably more biologically plausible way of capturing the history dependence of neural responses than a set of delay lines . In artificial neural networks , memory is usually implemented through recurrent connections between units . However , real biological neurons also have intrinsic memory processes with particular characteristics . Including such single-unit memory has potential advantages for explaining the in vivo responses of biological neurons and may also be useful in artificial neural networks . In summary , the dynamic network model we present here provides a biologically-plausible alternative encoding model to the classic LN model , as well as a basis for further development of encoding models with hidden units with diverse intrinsic dynamic processes . | The responses of neurons in the primary auditory cortex depend on the recent history of sounds over seconds or less . Typically , this dependence on the past has been modelled by applying a wide span of time delays to the input , although this is likely to be biologically unrealistic . Real neurons integrate the history of their activity due to the dynamical properties of their cell membranes and other components . We show that a network with a realistically narrow span of delays and with units having dynamic characteristics like those found in neurons , succinctly models neural responses recorded from ferret primary auditory cortex . Because these integrative properties are widespread , our dynamic network provides a basis for modelling responses in other neural systems . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"auditory",
"cortex",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"brain",
"vertebrates",
"neuroscience",
"learning",
"and",
"memory",
"animals",
"mammals",
"bacterial",
"diseases",
"cognition",
"network",
"analysis",
"memory",
"cellular",
"structures... | 2019 | A dynamic network model of temporal receptive fields in primary auditory cortex |
The dynamics of tumor progression is driven by multiple factors , which can be exogenous to the tumor ( microenvironment ) or intrinsic ( genetic , epigenetic or due to intercellular interactions ) . While tumor heterogeneity has been extensively studied on the level of cell genetic profiles or cellular composition , tumor morphological diversity has not been given as much attention . The limited analysis of tumor morphophenotypes may be attributed to the lack of accurate models , both experimental and computational , capable of capturing changes in tumor morphology with fine levels of spatial detail . Using a three-dimensional , agent-based , lattice-free computational model , we generated a library of multicellular tumor organoids , the experimental analogues of in vivo tumors . By varying three biologically relevant parameters—cell radius , cell division age and cell sensitivity to contact inhibition , we showed that tumor organoids with similar growth dynamics can express distinct morphologies and possess diverse cellular compositions . Taking advantage of the high-resolution of computational modeling , we applied the quantitative measures of compactness and accessible surface area , concepts that originated from the structural biology of proteins . Based on these analyses , we demonstrated that tumor organoids with similar sizes may differ in features associated with drug effectiveness , such as potential exposure to the drug or the extent of drug penetration . Both these characteristics might lead to major differences in tumor organoid’s response to therapy . This indicates that therapeutic protocols should not be based solely on tumor size , but take into account additional tumor features , such as their morphology or cellular packing density .
Organoid cultures are the three-dimensional ( 3D ) in vitro experimental systems in which individual cells grow and self-organize into multicellular structures that recapitulate the morphology and , to some extent , the functionality of the organ from which they are derived [1 , 2] . The complexity of organoid structure rises from single-layered hollow spherical breast or prostate acini [3 , 4] to the branched breast or salivary ducts [5 , 6] to the multilayered colon polyps [7] or brain lobules [8] . These in vitro cultures are utilized to address a wide range of biological questions , such as testing the mechanisms of tissue development and homeostatic maintenance and the initiation of malignant transformations and cancer growth dynamics , as well as anti-cancer treatments [9–11] . The Nature Publication Group recognized the novelty and importance of the organoid culture system by naming the organoid experiments as the Method of the Year in 2017 . Tumor organoids can be derived either from tumor cell lines or a patient’s tumor cells . They are used as tumor surrogates , as they resemble the biological and biophysical features of tumors and metastases better than 2D cell cultures . Also , these in vitro cultures can be controlled more efficiently than mouse experiments . Various tumor types can be currently cultured as organoids , including pancreatic [12] , liver [13] , gastrointestinal [14] , prostate [15] , brain [16] or breast [17] tumors . Tumor organoids can acquire diverse morphologies even if they are derived from the same organ . Kenny et al . showed that tumor organoids derived from twenty-five different breast cancer cell lines developed into four different multicellular shapes ( round , mass , grape-like and stellate ) after four days in culture [18] . Harma et al . obtained similar results for organoids derived from prostate cancer cell lines [19] . This morphological diversity may arise due to intrinsic differences , including distinct genetic profiles [20 , 21] , or extrinsic factors , such as various oxygenation levels or the extracellular matrix composition [22 , 23] . Tumor organoids have also been a subject to mathematical modeling . Various modeling frameworks were employed to investigate organoid development , metabolic variability and interactions with other components of tumor microenvironment . These include cellular automata models [24–27] , particle-force models [28–30] , Cellular Potts models [31–33] , continuous models [34] and our own immersed boundary model [35 , 36] . Our recent review [37] provides more details on these models and summarizes latest achievements in the mathematical modeling of tumor organoids . The model presented here was first developed in [38] and belongs to the class of particle-force models . The novelty of our studies lies in considering quantitatively a range of biologically relevant features to simulate a library of heterogeneous tumor organoids , and in developing new methods to quantify organoids’ properties and their potential response to drugs . There is a growing interest in using tumor organoids for drug screening in order to design more effective treatment combinations and schedules [39–41] . In a typical drug screening experiment , the drug is dissolved in a medium that surrounds the growing organoid and drug effectiveness is evaluated based on the changes in organoid size that are recorded over time . Usually , the bright field images of tumor organoid cross sections are used to determine two diameter measurements , the longest distance between opposite sides of the tumor ( length , L ) and the distance between tumor sides taken along the perpendicular direction ( width , W ) . Subsequently , these values are used to report either the average tumor diameter ( W+L ) /2 , tumor area W*L , or tumor volume ( L*W*W ) /2 [42] . These measurements reflect well the changes in the tumor mass if the organoids are of spherical shape . However , since organoid morphologies are often irregular and deviate from a sphere , this approach can lead to an inaccurate assessment of the organoid’s area or volume and , consequently , an incorrect measure of the organoid’s response to treatments . To avoid this discrepancy , a perimeter was proposed to better approximate the organoid’s volume from the 2D images [43] . These different morphologic measurements ( diameter , perimeter , area or volume ) were incorporated into the automatic analyses and classifications of tumor morphologies [44 , 45] . However , despite the technical and computational advances engaged in comparing tumor morphological features and their effects on drug exposure , many aspects related to tumor compactness and circularity remain elusive . Therefore , novel methods for assessing how tumor morphophenotypic properties impact drug dose estimation and therapy responses , which depart from evaluations based on organoid’s diameter , would provide a valuable independent prognostic factor . In this study , we generated a large number of in silico organoids , all with similar growth dynamics that fit the same set of tumor diameter measurements . In order to avoid effects of nutrient diffusivity limits , we restricted our simulations only to the initial phase of organoid development when the simulated structures are less that 150 microns in radius . Larger organoids may develop internal regions of hypoxia and necrosis that could influence the efficacy of administered treatments . We showed that the organoids we simulated attained morphologies , which can be classified into four groups ( morphophenotypes ) that describe organoids exposure to and penetration by the drug and , consequently , their potential response to the treatment . Since assessments of tumor diameter are often under- or overestimated , our classification methods are diameter-free but take advantage of organoid compactness and circularity measurements . The in silico organoids were simulated using the 3D framework , MultiCell-LF ( the Multi-Cellular Lattice Free framework ) , by varying three biologically-relevant parameters: cell age at division ( Adiv ) , cell contact inhibition controlled by the number of surrounding neighbor cells ( Nneigh ) and the maximal cell radius ( Rmax ) . The considered ranges of these parameters cover the biological features of distinct cell types [20 , 46 , 47] grown in various microenvironments [19 , 20 , 22 , 46 , 47] . The library of organoid morphologies generated from the sampled parameter space was categorized into morphophenotypic classes , and the impact of each parameter on tumor morphology was evaluated . Additionally , we used two metrics originating from the proteins’ structural biology , i . e . , the radius of gyration ( RGYR ) and the accessible surface area ( ASA ) , [48–51] , as the principal classification criteria . They together provided comprehensive and diameter-free analysis that may help identify the presence of patches , buried niches and extended chains , as well as compact or loose tumor masses that may change how receptive to the treatment the organoid is . The idea behind this approach is to use mathematical modeling and computer simulations to delineate the basic principles of global and local features of tumor organoid organization and , to quantify irregularities in organoid morphology . Our long-term goals are to apply this in silico screening approach to the experimental tumor organoids grown in diverse microenvironmental conditions and to test hypotheses suggesting improvements in cancer treatment . Here , we present the first step in this direction by applying our classification criteria to the computer-generated organoid morphologies .
Our first goal was to examine the diversity of the simulated organoids under the constraint that their diameter measurements fit the test data . We explored three cellular features , which could affect the generated organoid morphology: the radius of a tumor cell Rmax; the cell doubling time defined as the age at which the cell is ready to divide Adiv; and the cell’s sensitivity to contact inhibition , defined as the number of cell neighbors Nneigh that will halt the cell proliferation process . Three sets of simulations were performed for each parameter selection , and only those cases were further analyzed for which all three simulations fitted the test data . In each case , we only used the diameter information of the simulated organoids for comparison to the test data . Only those simulations that fulfilled both criteria: ( i ) were within the average +/- std diameter values of the test set , and ( ii ) the correlation between the diameters of the simulated organoids and the test data satisfied R2>0 . 9 , were accepted for further analysis . Four representative examples characterized by different Rmax , Adiv and Nneigh values are shown in Fig 1A–1D . In our analysis , we considered a wide range of model parameters . Rmax was varied between 5 μm and 10 μm , and these values are consistent with cell sizes reported in the literature [56 , 57] . Adiv was varied between 6 and 32 hours that has also been observed in experiments for different tumor cell lines [57 , 58] . Nneigh was varied between 3 and 22 cells to mimic cell sensitivity to contact inhibition that is directly related to the density of neighboring cells . The resulting heatmaps of the sampled parameter space organized by increasing cell radius are shown in the panels of Fig 2 . The generated parameter space covered an area of heterogeneous phenotypic and genotypic features . For example , different cell sizes ( values of Rmax ) may correspond to cells of a different origin or from distinct cell lines . A cell’s sensitivity to contact inhibition ( controlled by Nneigh ) may be representative of tumor cell invasive properties from epithelial-like to mesenchymal cells . The cell doubling time ( defined by Adiv ) may be associated with a specific tumor type or tumor cell line , and may correspond to a cell’s aggressiveness level or the microenvironmental condition ( acidity , oxygenation , nutrient contents ) in which the cells are growing . The heatmaps in Fig 2 provided an evaluation of how well the diameter measured during in silico organoid growth correlated with the test data . In addition to diameter measurement , our model also generated the organoid morphocharts , which are collections of the final organoid morphologies . The representative morphochart for organoids with a fixed cell radius of Rmax = 7 μm shown in Fig 3 revealed that , within constrains of our correlation criterion , a broad spectrum of organoid morphologies arose when the three cellular parameters selected for our study were varied . Close inspection of this radius-specific morphochart shows that the organoids tend to attain irregular shapes for low values of Nneigh ( strict contact inhibition ) . This effect is more profound for rapidly dividing cells ( small value of Adiv ) . When the values of Nneigh increase ( reduced contact inhibition ) the organoids’ shapes become more regular and reach morphology close to an ideal sphere for the highest values of Nneigh . The Nneigh also has a robust effect on the final organoid size . The organoids grow smaller ( within the average +/- std values ) when Nneigh is reduced compared with those having high Nneigh . This relationship between Nneigh and tumor shape is preserved in all morphocharts ( S1–S6 Figs ) . Additionally , the organoid’s size and shape depend on Adiv , and , for each row with fixed Nneigh , the size of the simulated organoids decreases with an increasing Adiv . Contrasting the contact inhibition parameter , the range of Adiv for which the simulated organoids fit the experimental data ( as observed in the heatmaps in Fig 2 ) shows a strong dependency on the cell radius . For example , for Rmax = 5 μm , the parameter Adiv can be varied between 6 and 17 hours , while Adiv spans from 17 to 32 hours for Rmax = 10 μm ( the range of Adiv values for each cell radius is shown in Fig 4A ) . This confirms that larger cells must be characterized by a slower division rate to fit the experimental organoid dynamics of growth . Interestingly , not only is the range of Adiv different for cells of different sizes , but the total number of simulated organoids that fulfill the correlation criteria also increases with a larger cell size; i . e . , the parameter space consists , on average , of about 50 organoids for Rmax = 5 μm and 100 for Rmax = 10 μm ( Fig 4B ) . Since the organoids simulated for each value of Rmax are fitted and correlated to the same growth dynamics data , it seems that cells characterized by small radii are less likely to adapt to changes in their vicinity and only a narrow valley in model parameter space fits the behavior of in vitro organoids . These parameter combinations are limited by small cell size , which demands faster division times Adiv . At the same time , the simultaneous fluctuations of Adiv and Nneigh provide higher flexibility , allowing the simulated organoids with larger cells to fulfill the correlation criteria . Since the generated organoids attain different morphologies , we examined whether they possess any common structural characteristics . We analyzed the compactness and accessible surface area across all in silico organoids . These two metrics were chosen to assess how effectively the externally supplied chemotherapeutic agent would reach all the organoid cells . Therefore , we tested what portion of the given organoid was directly exposed to the externally supplied drug ( organoid accessible surface area ) , and how easy it would be for the drug to penetrate the organoid ( organoid compactness ) . To assess organoid compactness , we calculated the organoid’s radius of gyration ( RGYR ) , which quantifies the distribution of all cells around the organoid’s center of mass . The larger the RGYR value , the less compact the organoid . The color-coded values of RGYR plotted as a function of cell division age ( Adiv ) and the number of cell neighbors ( Nneigh ) for all organoids with a cell radius of Rmax = 5 , 7 , and 10 μm are shown in Fig 5A–5C , respectively . In all three cases , the organoid compactness decreased when the cells could divide more often ( low Adiv ) or when they were less sensitive to contact inhibition ( larger Nneigh ) to halt their cell cycle progression ( red dots ) . For the same final morphologies , the accessible surface area ( ASA ) was calculated , and three color-coded heatmaps for the corresponding cellular radii are shown in Fig 5D–5F . While each heatmap separately contains a similar pattern of higher ASA values for organoids with faster growing cells and larger Nneigh , there is a visible shift in the ASA values for increasing Rmax , which contrasts with RGYR . This observation indicates that ASA is dependent on the cell size . Organoids with cells with a smaller Rmax have increased surface area , while larger cells tend to “smooth out” ASA . The actual diameter values ( D ) of each generated organoid are shown in Fig 5G–5I . Since there is variability in the sizes of experimental organoids , the generated in silico organoids were accepted if their growth dynamics did not exceed the average +/- std values of the test organoid set . Fig 5G–5I and Fig 5A–5C show good correlation between the patterns of the in silico organoid diameter and RGYR values , which implies that organoid’s compactness depends on its size , with less-compact organoids ( high RGYR ) having a larger diameter . While there is a positive correlation between the values of D and ASA for the whole organoid library shown in Fig 5 , the patterns do not overlap as closely as between D and RGYR . Instead , there are organoids within the same diameter range ( indicated by the same color in Fig 5G–5I ) but with distinct ASA values ( analogous data points have different colors in Fig 5D–5F ) . For example , organoids represented by red dots in Fig 5H ( i . e . , with Rmax = 7 μm , Nneigh = 21 cells , Adiv = 18 hours and Adiv = 16 hours ) have the highest values of D , while their ASA values ( analogous points in Fig 5E ) span a wider range of values: from 30x104 μm2 to above 60x104 μm2 , encompassing a difference of four color bins . Consequently , we observe that significantly different ASA values might be found for organoids with comparable sizes and the same cell radius . This observation directs attention to organoids with comparable diameters but with evident differences in morphology . For example , the organoids represented by blue dots in Fig 5H ( i . e . , Adiv = 17 hours , Nneigh = 16 cells; Adiv = 13 hours , Nneigh = 9 cells; Adiv = 11 hours , Nneigh = 6 cells; all with Rmax = 7 μm ) have diameters between 230–260 μm but have significantly different ASA values ( 31 . 07x104 μm2 , 41 . 35x104 μm2 and 12 . 79x104 μm2 , respectively ) and diverse numbers of cells within each organoid ( 1659 , 935 , and 204 , respectively ) . These organoids also belong to two different morphophenotypic classes ( next section ) , therefore , the differences in ASA calculated for morphologically diverse organoids with similar sizes might provide insights into the importance of correctly estimating the morphological features of the tumor to make proper predictions . Here , under- or overestimation of tumor diameters from 2D images could lead to a highly inaccurate assessment of organoid exposure-related features . Joint analysis of the structural and morphological properties of the generated organoids categorized them into four distinct morphophenotypic classes ( Fig 6 and Table 1 ) . Class 1 is characterized by high values of both RGYR and ASA; consequently , the organoids have a large accessible surface area but are not tightly packed . The representative morphology is of a regular spherical shape ( Fig 6A ) . Class 2 also attains quite regular shapes and large values of ASA , but smaller values of RGYR , resulting in a more compact organoid structure ( Fig 6B ) . In Class 3 , the organoid morphologies are elongated ( Fig 6C ) and are accompanied by smaller ASA and relatively small RGYR values . Finally , in Class 4 , the organoids are characterized by more fragmented , branched shapes ( Fig 6D ) with small RGYR and ASA values . This classification was achieved using the classical k-medians clustering algorithm , and the consensus cluster centroids for each class are listed in Table 1 . These in silico morphophenotypes are similar to the organoid morphologies observed in vitro . Experiments with the organoids derived from 25 different breast tumor cell lines [18] and from 29 different prostate tumor cell lines [19] identified several characteristic organoid shapes , including “Round” spheroids with well differentiated , tightly packed cells similar to our Class 2 “Compact” structures; a “Mass” phenotype characterized by a lower level of cell-cell contacts and cell polarization , resulting in large masses of more loosely packed cells similar to our Class 1 “Spherical” structures; a “Stellate” phenotype characterized by multiple projections and an elongated shape similar to our Class 3 “Elongated” phenotype; and a “Grape-like” phenotype lacking robust cell-cell adhesions and showing a more branched structure similar to our Class 4 “Branched” structures . While there is no complete agreement whether organoid shapes and the level of genetic transformations in the individual cells are correlated , the more aggressive and invasive cell lines tend to have more branched and elongated morphologies and less pronounced cell-cell adhesions , which is also observable in our simulations .
The aberrant morphologies of multicellular in vitro cultures are the first visual indication of cells’ abnormality and potential tumorigenic capabilities . The morphologies of tumor organoids can be quite diverse [18–20 , 55] that poses a question whether a given chemotherapeutic treatment will have the same effect on such heterogeneous tumors . Typically , tumor cell response to the drugs is defined by the IC50 value ( the concentration of an inhibitor where the response is reduced by half ) and examined using the 2D monolayer cultures when all cells are well exposed to the therapeutic compounds . However , tumors grow as 3D multicellular conglomerates and thus the differences in individual cell exposure to the drug and the extent of drug penetration though the whole tumor tissue play a paramount role in treatment efficacy . Even the most potent drug will be ineffective if it cannot reach all tumor cells in efficient quantities . This aspect cannot be examined in the 2D cell cultures , thus there is a growing interest in using tumor organoids for testing anti-cancer treatments . However , current experimental protocols do not take into consideration morphological and structural diversities between tumor organoids . Here , our goal was to draw attention to these overlooked aspects of anti-cancer treatment testing . We used a theoretical approach to evaluate the impact of cells’ intrinsic properties and cell-cell interactions on the emergent tumor shape . To the best of our knowledge , this is the first comprehensive computational study that systematically and quantitatively explores morphological diversity of 3D tumor organoids . The minimalistic set of cell properties considered here included cell size , division time and sensitivity to contact inhibition , all of which can vary from one cell line to another , and between different tumor types . The exploration of these three model parameters gave rise to a library of organoids with diverse morphologies . However , for all these organoids , their average diameters measured during each simulation matched the pre-defined test data ( diameters are the most typical metrics used in laboratory experiments to assess tumor size ) . While we used a specific set of test data for model calibration , this approach can be generalized and applied to different dynamics of organoid development . The method presented here accepts organoids which diameters fit the test data and disregards simulations in which the growth dynamics deviates from the test data . In this way , we can simulate multicellular systems with diverse growth dynamics . By using an analogy to the structural biology of biomolecules , we provided the morphological classification of multicellular organoids using the radius of gyration ( RGYR ) and accessible surface area ( ASA ) . As a common tool for measuring system’s compactness , RGYR applied to organoids informs how dense the tumor mass is and how tightly packed the cells are . The organoid compactness provides a measure of system globularity , such as packing density and intercellular space cavities . Taking this approach , we showed that RGYR helps to identify the classes of organoids characterized by different internal cellularity and external branching that may correspond to cell aggressiveness in vivo . The use of the RGYR values as a sole classification criterion has raised some concerns because organoids within the same range of RGYR attained significantly different morphologies . To overcome this discrepancy , we proposed to also use a concept of the ASA , which determines what fraction of the system is exposed to the external microenvironment . Currently used methods for assessing changes in tumor organoids’ growth rely on measuring their diameters from microscopy images . These measurements , however , may not be accurate , especially if the multicellular structures have irregular shapes . By creating a broad library of organoids morphologies ( the MultiCell-LF morphochart shown in Fig 3 ) , we proposed to use mathematical modeling for reconstruction of experimental organoid structures and to determine their ASA and RGYR in a cost- and time-efficient way . The presented computational methods can be combined with measurements obtained from experimental data , i . e . , the size of tumor cells can be measured directly , the cell proliferation rates can be determined using cell doubling assays and the organoid’s overall size can be calculated using bright field microscopy images . These values can be mapped on the MultiCell-LF morphochart parameter space ( Fig 5 ) , which allows for reconstruction of the organoid’s structure in silico and provides the RGYR and ASA values . This procedure is similar to the acinar morphochart technique we developed previously for determining the functional changes in mutated breast tumor cell lines compared with the parental non-tumorigenic cell line using their 3D morphologies [35 , 59–61] . We hypothesized here that currently used treatment protocols for drug dose selection , scheduling and duration could benefit from additional information provided by the RGYR and ASA values , however further experiments are needed to confirm this postulate . Tumor tissue irregular architecture is considered as one of the barriers in effective drug penetration though the in vivo tumors [62 , 63] . However , similar inefficient penetration has also been observed in tumor organoids and visualized using chemotherapeutic compounds tagged to fluorescent probes [64–66] . We postulate here that ASA and RGYR may provide prospective metrics for assessing the potential of tumor organoids to be exposed and penetrated by the drug . Some recent studies have discussed the relationship between morphologies of in vivo and ex vivo tumors and their response to drugs [67 , 68] showing that there is an interest in developing new measures to assess drug efficacy in tumors and tumor organoids based on their morphology . The organoids considered here ( up to 150 μm in radius ) are of the size that is below the diffusivity limits of oxygen or nutrients ( 200–250 μm ) . However , in larger organoids the gradients of nutrients from spheroid periphery towards its core can develop , that results in the emergence of hypoxic or necrotic regions . As a result , not all cells would have equal access to nutrients and oxygen which may affect their growth , as well as their response to therapeutics . In this study , we limited the number of intracellular and extracellular heterogeneities , however in the future we are planning to examine how the tumor organoids’ morphology changes when cells have individually-regulated cell cycle , cell-cell interactions and variable sensitivity to contact inhibition . We will also incorporate additional cellular processes , such as cell motility , cell death or secretion of autocrine signals , and evaluate their influence of the emergent organoid’s morphology . While in the current study we treated the external medium and the interstitial space between the cells as a homogeneous continuum , experimental media and in vivo stroma can contain proteins and fibers of the extracellular matrix that will further impede drug transport . We plan to investigate these factors in the future . Additionally , the morphologies of the tumor organoids can undergo dynamic changes in response to treatments . Not only can they shrink , but they may also become more irregular , if , for example , cell cycle-specific drugs are administered and target only certain cells within the spheroid . Thus , the numerical estimation of drug doses based on the morphological features of tumor organoids ( including ASA and RGYR ) could provide a more accurate metric for predicting a tumor’s response to chemotherapy , and for adjusting drug schedules leading to more adaptive and personalized treatment . | Primary tumors and tumor metastases grow as three-dimensional ( 3D ) masses of cells . Depending on the surrounding stroma , they may acquire various shapes , more or less irregular . Tumor organoids are the 3D experimental cultures that mimic growth of in vivo tumors , as well as their response to treatments . However , it is difficult to assess experimentally in a reproducible and quantitative way , how tumor morphology influences treatment efficacy . Here , we used mathematical modeling and computer simulations to analyze the structure of the simulated organoids and to classify them with regards to two quantitative features: the tumor accessible surface area ( ASA ) describing organoid exposure to the drug and the extent of drug penetration through the tumor tissue ( organoid compactness ) . We showed that organoids of similar sizes and growth dynamics can , in fact , be characterized by distinct values of compactness and ASA , and thus may respond differently to the drug treatment . We suggest that these tumor features should be taken into consideration in addition to tumor size , when the therapeutic interventions are designed . | [
"Abstract",
"Introduction",
"Results",
"Discussion"
] | [
"organoids",
"medicine",
"and",
"health",
"sciences",
"cell",
"cycle",
"inhibitors",
"cell",
"cycle",
"and",
"cell",
"division",
"cancer",
"treatment",
"biological",
"cultures",
"cell",
"processes",
"radii",
"geometry",
"simulation",
"and",
"modeling",
"oncology",
"... | 2019 | Morphophenotypic classification of tumor organoids as an indicator of drug exposure and penetration potential |
To date , studies on papillary renal-cell carcinoma ( pRCC ) have largely focused on coding alterations in traditional drivers , particularly the tyrosine-kinase , Met . However , for a significant fraction of tumors , researchers have been unable to determine a clear molecular etiology . To address this , we perform the first whole-genome analysis of pRCC . Elaborating on previous results on MET , we find a germline SNP ( rs11762213 ) in this gene predicting prognosis . Surprisingly , we detect no enrichment for small structural variants disrupting MET . Next , we scrutinize noncoding mutations , discovering potentially impactful ones associated with MET . Many of these are in an intron connected to a known , oncogenic alternative-splicing event; moreover , we find methylation dysregulation nearby , leading to a cryptic promoter activation . We also notice an elevation of mutations in the long noncoding RNA NEAT1 , and these mutations are associated with increased expression and unfavorable outcome . Finally , to address the origin of pRCC heterogeneity , we carry out whole-genome analyses of mutational processes . First , we investigate genome-wide mutational patterns , finding they are governed mostly by methylation-associated C-to-T transitions . We also observe significantly more mutations in open chromatin and early-replicating regions in tumors with chromatin-modifier alterations . Finally , we reconstruct cancer-evolutionary trees , which have markedly different topologies and suggested evolutionary trajectories for the different subtypes of pRCC .
Renal cell carcinoma ( RCC ) makes up over 90% of kidney cancers and currently is the most lethal genitourinary malignancy [1] . Papillary RCC ( pRCC ) accounts for 10%-15% of the total RCC cases [2] . Unfortunately , pRCC has been understudied and there is no current form of effective systemic therapy for this disease . pRCC is further subtyped into two major subtypes: type I and type II based on histopathological features . For many years , the only prominent oncogene in pRCC ( specifically , type I ) that physicians were able to identify was MET , a tyrosine-kinase receptor for hepatic growth factor . An amino acid substitution that leads to constitutive activation and/or overexpression are two mechanisms of dysfunction of MET in tumorigenesis . Recently , the Cancer Genome Atlas ( TCGA ) published its first result on pRCC [3] , which greatly improves our understanding of the genomic basis of this disease . Several more genes and pathways were identified to be significantly mutated in pRCC . Nevertheless , a significant portion of pRCC cases remain without any known driver . Therefore , we think it is a good time to explore the noncoding regions of the genome using whole-genome sequencing ( WGS ) . Noncoding regions , often overlooked in cancer studies , have been shown to be actively involved in tumorigenesis [4–6] . Mutations in noncoding regions may cause disruptive changes in both cis- and trans-regulatory elements , affecting gene expression . Understanding noncoding mutations helps fill the missing “dark matter” in cancer research . Meanwhile , it is an open question to the degree which these regions harbor significant driver alterations and here we can address this question in the specific context of pRCC . Multiple endogenous and environmental mutation processes shape the somatic mutational landscape observed in cancers [7] . Analyses of the genomic alterations associated with these processes give information on cancer development and heterogeneity , shed light on the mutational disparity between cancer subtypes and even indicate potential new treatment strategies [8] . Additionally , genomic features such as replication time and chromatin environment govern mutation rate along the genome , contributing to spatial mutational heterogeneity . Last , while studying mutation patterns , landscape and tumor evolution is possible using data from whole exome sequencing ( WXS ) , whole genome sequencing ( WGS ) gives richer information on mutation landscape and minimizes the potential confounding effects of exome capture process and driver selection . In this study , we comprehensively analyzed 35 pRCC cases that were whole genome sequenced along with an extensive set of WXS data on multiple levels . We went from microscopic examination of driver genes to analyses of whole genome sequencing variants , and finally , to an investigation of high-order mutational features . We focused on two aims: exploring potential noncoding drivers and better understanding the heterogeneity of the cancer . First , we focused on MET , an oncogene that plays a central role in pRCC , especially in the type I . We found rs11762213 , a germline exonic single nucleotide polymorphism ( SNP ) inside MET , predicted cancer-specific survival ( CSS ) in type II pRCC . We also discovered several potentially impactful noncoding mutations in MET promoter and its first two introns . The previous TCGA study identified a MET alternate transcript as a driver but without illustrating the etiology [3] . We found that a cryptic promoter from an endogenous long interspersed nuclear element-1 ( L1 ) triggered the alternate isoform expression . Surprisingly , we did not find a significant amount of small structural variations affecting MET . Then we went on to cases not as easily explained as those with MET alterations . We analyzed about 160 , 000 noncoding mutations throughout the whole genome and found several potentially high-impact mutations in the noncoding regions . Further zooming out , we discovered pRCC exhibited mutational heterogeneity in both nucleotide context and genome location , indicating underlying interplay of mutational processes . We found methylation was the leading factor influencing mutation landscape . Methylation status drove the inter-sample mutation variation by promoting more C-to-T mutations in the CpG context . APOBEC activity , although infrequently observed , left an unequivocal mutation signature in a pRCC genome but not in ccRCC . Also , we discovered samples with chromatin remodeler alternations accumulated more mutations in open chromatin and early-replicating regions . Lastly , we derived evolutionary trees based on the whole-genome mutation calls for each individual sample . The tree topologies varied , reflecting tumor heterogeneity and correlating with the known tumor subtypes .
We began with coding variants in the long known driver MET . The TCGA study of 161 pRCC patients found 15 samples carrying somatic , nonsynonymous single nucleotide variants ( SNVs ) in MET . By analyzing 117 extra WXS samples ( see Methods ) , we found six more nonsynonymous somatic mutations in six samples ( S1 Table ) . V1110I and M1268T were two recurrent mutations in this extra set . Both of them were observed in the TCGA study as well . Additionally , we found two samples carrying H112Y and Y1248C respectively . H1112Y has been observed in two patients in the original TCGA study cohort and H1118R is a long-known germline mutation associated with hereditary papillary renal carcinoma ( HPRC ) [9] . Y1248C has been observed in type I pRCC before and the TCGA cohort has a case carrying Y1248H . All mutations occurred in the hypermutated tyrosine kinase catalytic domain of MET . Two out of these six samples were identified as type I pRCC while the subtypes of the rest four were unknown . Although many MET somatic mutations are believed to play a central role in pRCC , some germline MET mutations have also been associated with the disease . In particular , a germline SNP , rs11762213 ( Fig 1A ) , has been discovered to predict recurrence and survival in a mixed RCC cohort [10] . ccRCC predominated the initial discovery RCC cohort . This conclusion was later validated in a ccRCC cohort but never in pRCC [11] . It was not clear whether this SNP has a prognostic effect in pRCC . Using an extensive WXS set of 277 patients ( see Methods; S1 Fig and S1 Table; ) , we found 14 patients carrying one risk allele of rs11762213 ( G/A , minor allele frequency ( MAF ) = 2 . 53% ) . No homozygous A/A was observed . Cancer specific deceases were concentrated in type II pRCC . Among 96 type II pRCC cases , seven patients carried the minor A allele ( MAF = 3 . 65% , Table 1 ) . Cancer-speccific survival was significantly worse in type II patients carrying the risk allele of rs11762213 ( p = 0 . 034 , Fig 1B ) . But we did not find a significant association of this germline SNP with survival in type I patients . We did not observe statistically significant correlation of rs11762213 with MET RNA expression in either tumor samples or normal controls ( p > 0 . 1 , two-sided rank-sum test ) . c-Met pY1235 levels in tumor samples , as measured by Reverse phase protein array ( RPPA ) , were not significantly different between carriers of these two genotypes ( p > 0 . 1 , two-sided rank-sum test ) . The TCGA study identified a MET alternate transcript as driver [3] . However , the etiology of this new isoform was unknown . Here we found this alternate transcript resulted from the activation of a cryptic promoter from an endogenous L1 element ( Fig 1A ) , likely due to a local loss of methylation [12] . This event was reported in several other cancers [13 , 14] . To test its relationship with methylation , we found the closest probe ( cg06985664 , ~3kb downstream ) on the methylation array showed marginally statistically significantly lower methylation level in samples expressing the alternate transcript ( p = 0 . 055 , one-sided rank-sum test ) . Additionally , this event was associated with methylation cluster 1 ( odds ration ( OR ) = 4 . 54 , p<0 . 041 ) , indicating genome-wide methylation dysfunction . This association was stronger in type II pRCC and the alternate transcript was tightly associated with the C2b cluster ( OR = 17 . 5 , p<0 . 007 ) . Despite the fact , MET is the most common driver alteration , about 20% presumably MET-driven yet MET wild-type pRCC samples were still left unexplained [3] . That is , they had a characteristic Met-dysregulated gene expression pattern but no obvious Met-associated alteration . Therefore , we scanned the MET noncoding regions . We observed one mutation in MET promoter region in a type I pRCC sample ( Fig 1A and S2 Table ) . This sample showed no evidence of nonsynonymous mutation in MET gene but had copy number gain of MET . Additionally , we observed 6/35 ( 17 . 1% ) samples carry mutations in the intronic regions between exons 1–3 of MET ( Fig 1A and S2 Table ) . As we describe above we could see that these mutations nearby to regions with methylation dysregulation and the activation of a cryptic promoter . However , we were not able to find a direct statistically significant correlation between the alternative splicing event and these intronic mutations . We further expanded our scope and ran FunSeq [4 , 5] to identify potentially high-impact , noncoding variants in pRCC . First , we identified a high-impact mutation hotspot on chromosome 1 . 6/35 ( 17 . 1% ) samples had mutations within this 6 . 5kb region ( Fig 2A and S2 Table ) . This hotspot located at the 5’-end of ERRFI1 ( ERBB Receptor Feedback Inhibitor 1 ) and overlapped with the predicted regulatory region . ERRFI1 is a negative regulator of EGFR family members , including EGFR , HER2 and HER3; all have been implicated in cancer . However , due to a limited sample size here , our test power was inevitably low . We did not observe statistically significant changes among mutated samples in mRNA expression level , protein level and phosphorylation level of EGFR , HER2 and HER3 . Another potentially impactful mutation hotspot was in NEAT1 . We saw mutations inside this nuclear long noncoding RNA in 6/35 ( 17 . 1% ) samples ( Fig 2B and S2 Table ) . Several studies indicated NEAT1 is associated with various cancers [15 , 16] . It promotes cell proliferation in hypoxia [17] and alters the epigenetic landscape , increasing transcription of target genes [18] . Mutations we found all fell into a putative promoter and its flanking region of NEAT1 . We noticed NEAT1 mutations were associated with higher NEAT1 expression ( Fig 2C , S2A Fig , p < 0 . 032 , two-sided rank sum test ) . We also found NEAT1 mutations were associated with worse prognosis ( Fig 2D , p < 0 . 041 , log-rank test ) . To further investigate the role of NEAT1 in RCCs , we found NEAT1 overexpression is significantly associated with shorter overall survival in ccRCC ( TCGA cohort , p = 0 . 0132 , S2B Fig ) . Moreover , MALAT1 , another noticeable lncRNA in cancer , is tightly co-expressed with NEAT1 in both pRCC and ccRCC ( Spearman’s correlation; 0 . 79 and 0 . 87 respectively ) . MALAT1 is located ~50kb downstream of NEAT1 and might share the same regulatory mechanism with it . The Catalogue of Somatic Mutations in Cancer ( COSMIC ) [19] annotates MALAT1 as a cancer consensus gene , associating it with pediatric RCC and lung cancer . MALAT1 was also reported to be associated with ccRCC [20] . We performed structural variants ( SVs ) discovery using WGS reads ( see Methods and S3 Table , S2C Fig ) . This SV discovery approach has higher sensitivity and resolution than array-based copy number variation methods , which were employed in the TCGA analysis . This was a large-scale , big-compute calculation that involves mapping more than 100 billion reads ( see Methods ) . In the end , we found 424 somatic SV events , including 170 deletions , 53 duplications , 105 inversions and 96 translocations ( S3 Table ) . The samples clearly split into two categories based on the number of SV events ( ranging from 0 to 88 ) : genome unstable ( 6 samples , >40 events/per samples ) and genome stable ( 29 samples , <10 events/per sample ) . The unstable category was significantly associated with type II versus type I pRCC ( p<0 . 015 , two-tailed Fisher exact test ) and enriched in the C2b cluster ( p < 0 . 002 , two-tailed Fisher exact test ) . We overlapped SVs with curated cancer genes from COSMIC [19] . Somewhat surprisingly , we did not find SVs affecting MET except a single example—one genomically highly unstable sample , TCGA-B9-4116 , with deep amplification of MET , showed multiple SVs of various classes hitting MET . To explain this lack of enrichment for small SVs in MET , we postulated trisomy/polysomy 7 is the main mechanism of MET structural alteration rather than small-scale duplication . Moreover , besides duplication , we did not expect to find deletion , inversion or translocation disrupting oncogene MET . These SVs were likely to cause loss-of-function rather than gain-of-function . Indeed , we did not find any breakpoint splitting MET . This was consistent with the putative role of MET as an oncogene , rather than a tumor suppressor . We next looked for other cancer genes affected by somatic SVs . We found two cases with deletions in SDHB . The median SDHB expression was significantly lower ( p<0 . 0034 , one-sided rank sum test ) , only ~50% compared to cases without alternation ( S2D Fig ) . We validated the deletions affecting SDHB with another SV caller , Lumpy-SV . Besides , we confirmed three cases carrying deletions affecting CDKN2A called by the TCGA array-based method but not the other two cases . Notably , three confirmed cases had significantly lower CDKN2A expression ( p<0 . 0013 , one-sided rank sum test ) but the unconfirmed two cases did not ( S2D Fig ) . This suggests SV calling from WGS is accurate and predicts CDKN2A expression better . Lastly , we observed several high-impact sporadic events , including duplications in EGFR and HIF1A , and deletions in DNMT3A and STAG2 ( S2C Fig ) . To further get an overview of the mutation landscape , we summarized the mutation spectra of 35 whole genome sequenced pRCC samples ( Fig 3A ) . C-to-T in CpGs showed the highest mutation rates , which were roughly three to six-fold higher than mutation rates of other nucleotide contexts . We used principle components analysis ( PCA ) to reveal factors that explained the most inter-sample variation . The loadings on the first principle component ( which explained 12 . 5% of the variation ) demonstrated C-to-T in CpGs contributed the most to inter-sample variation ( Fig 3B ) . C-to-T in CpGs is highly associated with methylation . It reflects the spontaneous deamination of cytosines in CpGs , which is much more frequent in 5-methyl-cytosines [21] . So we further explored the association between C-to-T in CpGs and tumor methylation status . First , we validated the TCGA identified methylation cluster 1 showed higher methylation level than cluster 2 in all annotated regions ( S3 Fig , see Methods ) , prominently in CpG Islands ( Odds ratio of sites being differentially hypermethylated: 1 . 29 , 95%CI: 1 . 20–1 . 39 , p<0 . 0001 ) . We confirmed this association by showing samples from methylation cluster 1 had higher PC1 scores as well as higher C-to-T mutation counts and mutation percentages in CpGs ( Fig 3C ) . This trend was further validated using a larger WXS dataset as well . Especially , the most hypermethylated group , CpG island methylation phenotype ( CIMP ) , showed the greatest C-to-T rate in CpGs ( S3C Fig ) . Therefore , methylation status was the most prominent factor shaping the mutation spectra across patients . Furthermore , we explored the functional impacts of the excessive mutations driven by methylation . C-to-T mutations in CpGs we observed in pRCCs were more likely to be in the coding region ( OR = 1 . 54 , 95%CI: 1 . 27–1 . 85 , p<0 . 0001 ) and nonsynonymous ( OR = 1 . 47 , 95%CI: 1 . 17–1 . 84 , p<0 . 001 ) , which indicated they tended to be high-impact mutations . However , C-to-T mutations in CpGs did not show functional bias between the two methylation clusters in noncoding regions ( based on FunSeq score distribution ) . Recently , 30 somatic mutation signatures were identified; many have putative etiology , revealing the underlying mutational processes and helping understand tumor development [7] . We used a LASSO-based approach ( see Methods ) to decompose the observed mutations into a linear combination of these canonical mutation signatures in both WGS and WXS samples ( S4 Fig ) . The leading signature was "signature 5" ( from reference 7 ) . Interestingly , we found one type II pRCC case out of 155 somatic WXS sequenced samples exhibited APOBEC-associated mutation signatures 2 and 13 . APOBEC mutation pattern enrichment analysis ( see Method ) further confirmed the presence of APOBEC activity ( Fig 3D , S4 Table ) . This sample was statistically enriched of APOBEC-induced mutations ( adjusted p-value < 0 . 0003 ) . Prominent APOBEC activities were also incidentally detected in three upper track urothelial cancer ( UC ) samples sequenced and processed in the same pipeline with pRCC samples . UC often carries APOBEC associated mutation signatures and our result is consistent with the TCGA bladder urothelial cancer study [22] . The APOBEC associated signature carrying pRCC case was centrally reviewed by six pathologists in the original study and confirmed to be type II pRCC [3] . Thus , this tumor is likely a special case of type II with genomic alterations sharing some similarities with UC . It had non-silent mutations in ARID1A and MLL2 and a synonymous mutation in RXRA , all are identified as significantly mutated genes in UC but not in pRCC . Potential type II pRCC driver events , for example , low expression of CDKN2A and nonsynonymous alternations in significantly mutated genes of pRCC , were absent in this sample . Noticeably , the four samples with APOBEC activities showed significantly higher APOBEC3A and APOBEC3B mRNA expression level ( p < 0 . 0022 and p < 0 . 0039 respectively , one-sided rank sum test , S5 Fig ) . This is in concordance with previous studies of APOBEC mutagenesis in various types of cancer [23] . Consistent with previous studies [12] , we failed to detect statistically significant APOBEC activities in an extensive WXS dataset of 418 clear cell RCC ( ccRCC ) samples , even after subsampling to avoid p-value adjustment eroding the power . Very low levels of APOBEC signatures ( <15% ) were found in less than 1% ( 4/418 ) samples . With a much larger sample size , this result was unlikely to be confounded by detecting power . Chromatin remodeling genes are frequently mutated in pRCC and many other cancers , including ccRCC [3 , 24 , 25] . Defects in chromatin remodeling cause dysregulation of the chromatin environment . Open chromatin regions usually show a lower mutation rate , presumably due to more effective DNA repair [26] . Thus , chromatin remodeler alterations could possibly alter the mutation landscape , specifically increasing mutation rate in previously open chromatin regions . To test this , we tallied the number of mutations inside DNase I hypersensitive sites ( DHS ) inferred from 11 normal fetal kidney cortex samples ( The NIH Roadmap Epigenomics Mapping Consortium ) [27] , which represent normal tissues under physiological conditions . 9/35 samples with disruptive mutations in ten chromatin remodeling genes , cancer-associated genes showed higher genome-wide mutation counts ( p < 0 . 021 , one-sided rank-sum test; ) , partially driven by higher mutation counts in the DHS regions ( p < 0 . 0023 , one-sided rank-sum test ) . The median number of mutations in DHS regions considerably increased by 60% ( 67 . 5 versus 108 ) in samples carrying chromatin remodeling defects . The effect was still significant after normalizing against the total mutation counts ( p < 0 . 019 , one-sided rank-sum test , Fig 3E ) , indicating a significant shift in mutation landscape . Replication time is known to correlate greatly with mutation rate . Early replicating regions have lower mutation rate compared to late replicating ones . Researchers reason replication errors are more likely to be corrected by DNA repair system in early replicating regions . With defects in chromatin remodeling genes , we observed this trend became less pronounced ( p<0 . 031 , one-sided rank-sum test , S6 Fig ) . This is presumably because dysregulation of the chromatin environment hinders replication error repair by changing the accessibility of newly synthesized DNA chains . With the richness of SNVs in WGS samples , we can further tackle the mutational process heterogeneity of pRCC by constructing evolutionary trees for the 35 tumors ( S7 Fig ) . These trees were derived from the whole-genome mutation calls and were produced individually for each tumor , with their topology suggesting a temporal ordering to the mutations . We could classify the trees into four groups based on their topology ( Fig 4 ) : In addition , three trees were excluded from the analysis since they had a largest population faction <0 . 5 , which was likely due to low mutation number , high sequence error and/or particularly high copy number variation . Both topology groups 3 and 4 showed significant clonal evolution , with more distal subclones , and greater heterogeneity , indicated by substantial mutational divergence between populations . These groups were significantly depleted in type I pRCC ( p < 0 . 0034 , two-tailed fisher exact test ) . In contrast , the short branch group ( #2 ) was significantly enriched in type I pRCC ( p<0 . 011 , two-tailed fisher exact test , Fig 4B ) . This suggested type I tumors were more homogenous and showed less complex evolutionary features compared to type II and unclassified samples .
Our study is the first one that comprehensively looked into the noncoding regions of pRCC . Doing so allowed us to tackle an open question in the field of cancer genomics , whether whole genome sequencing adds additional value over whole exome sequencing . We comprehensively analyzed both WGS and an extensive set of WXS of pRCC , scrutinizing local high-impact events as well as giving an overall view of the mutation landscape and evolution . Our work further completed the genomic alteration landscape of pRCC ( Fig 4B ) . Beyond traditionally driver events , we suggested several novel noncoding alterations potentially drive tumorigenesis . We also provided valuable insights to tumor heterogeneity though investigating the mutational patterns , landscape , and evolutionary profiles . First , we elaborated on previous results of the long known driver MET . In an extended 117 WXS dataset , we found six additional nonsynonymous somatic mutations in the hyper-mutated tyrosine kinase catalytic domain . These somatic mutations were highly recurrent , concentrated on a few critical amino acids . This was in line with MET being an oncogene and supported its central driver status in pRCC . Then we found an exonic SNP in MET , rs11762213 , to be a prognostic germline variance in type II pRCC . Previously , rs11762213 was found to predict outcome in a mixed RCC samples , predominated by ccRCC [10] . Later , the result was confirmed in a large TCGA ccRCC cohort [11] . However , it was never clear whether rs11762213 only predicts the outcome in ccRCC or other histological types as well . In this study , we concluded that the minor alternative allele of rs11762213 also forecasts unfavorable outcome in type II pRCC patients . The mechanism of this exonic germline SNP remains unsettled . A previous study proposed it disrupts a putative enhancer of MET [11] . However , researchers could not find significant association between the SNP and MET expression in either tumor or normal tissues . We noticed there is no other gene within 100 kb in both directions of this SNP . Given the significant role of MET in pRCC , we think rs11762213 is affecting survival through MET , although the mechanism unknown . Similar to ccRCC , type II pRCC is not primarily driven by MET . Not as significantly mutated in ccRCC and type II pRCC , MET nonetheless seems to play a role in cancer development . Our finding on rs11762213 is potentially meaningful in the clinical management of patients with the more aggressive type II pRCC . rs11762213 genotyping could become a reliable , low-cost risk stratification tool for these patients . Also , rs11762213 might become a biomarker for predicting response to Met inhibitors pending further studies . Interestingly , rs11762213 is prevalent mostly in European and American populations but not in African populations and rare in Asian populations . However , the MAF of rs11762213 among African American patients in our cohort is 2 . 73% , higher than MAFs in general for African populations observed in 1000 Genome phase 3 dataset ( 0 . 2% , with 0% in Americans with African ancestry , ASW ) [28] and the ExAC dataset ( 1 . 1% , excluding TCGA cohorts ) [29] . This implies a possible effect of rs11762213 on pRCC incidence among African Americans that is worth further investigation . In MET noncoding regions , we first found a cryptic promoter from a retrotransposon in the second intron initiates the alternate transcript , which was classified as a driver by the TCGA study ( 3 ) . Methylation is a major source of silencing retrotransposon activities in the human genome [12–14] . Indeed , we observed evidence for a local loss of methylation and global methylation dysregulation in samples expressing the alternate transcript . Our finding indicates methylation change might directly drive pRCC growth through MET . We also discovered mutations associated with the MET promoter and first two introns , where the alternate transcript starts . Although the implication is unknown , our analysis suggests there is a mutation hotspot in MET that calls for further research . Expanding our scope from coding to non-coding and using FunSeq to group SNVs by functional elements , we found several potentially significant noncoding mutation hotspots relevant to tumorigenesis throughout the entire genome . A mutation hotspot was found downstream of ERRFI1 , an important regulator of the EGFR pathway , which may serve as a potential tumor suppressor . EGFR inhibitors have been used in papillary kidney cancer with an 11% response rate observed [30] . These mutations potentially disrupt regulatory elements of ERRFI1 and thus play a role in tumorigenesis . However , likely limited by a small sample size , we were not able to detect statistically significant functional changes in ERRFI1 and related pathways . Another noncoding hotpot was in NEAT1 , a long noncoding RNA that has been speculated to involve in cancer . Patients carrying mutations in NEAT1 had significantly higher NEAT1 expression and worse prognosis . High expression of NEAT1 predicted significantly worse survival in ccRCC as well . NEAT1 has been shown to be hypermutated in other cancers and some studies also linked high NEAT1 association with unfavorable prognosis [31 , 32] . Lastly , a downstream lncRNA , MALAT1 , showed tight co-expression pattern with NEAT1 in both pRCC and ccRCC . MALAT1 is on COSMIC consensus cancer gene list and annotated as related with pediatric RCC [19] . It was also reported to be associated with ccRCC [20] . Next , with more than 100 billion carefully remapped reads from WGS , we generated a high-confident SV dataset for 35 pRCC samples . Our method has great accuracy . In fact , we confirmed the well-known deletion of CDKN2A and found that we predicted its down-regulate expression better than the copy number variation analysis in TCGA study [3] . In terms of overall numbers of SVs , we found the pRCCs clearly split into two categories: the stable category had less than 10 events per sample while the unstable category had all above 40 . Moreover , the unstable category was tightly associated with the C2b cluster , which has inferior outcomes [3] . Our SV study also discovered recurrent cases of SDHB deletion and expression data supported our finding . SDHB is a subunit of succinate dehydrogenase . Previous studies indicated the loss of SDHB being a driver event by disturbing tumor metabolic environment [33 , 34] Besides SDHB , we also found some other sporadic events involving known tumor drivers . Somewhat counter-intuitively , we found the absence of MET alterations that involve small deletion or breakage of the MET gene except in one highly unstable sample . Large-scale duplications involving MET , however , have been found ( e . g . trisomy 7 ) . This finding can be rationalized by realizing that the oncogenic activity of MET is encouraged by amplification but not by deletion or disruption . Moreover , we postulated that polysomy 7 might be the major mechanism of MET gain and lack of smaller SVs and breakpoints disrupting MET further supports its oncogene role . WGS provides many times more SNVs compared to WXS . Thus it gives us an opportunity to look into the high-level landscape of mutations in pRCC . Several recent landmark pan-cancer studies lead to the wide recognition of significance and great research interests in cancer mutational processes [7 , 8 , 26 , 35 , 36] . DNA mutation is one of the driving forces of cancer development , and understanding the underlying processes and affecting factors that generate the mutations is vital in cancer studies . In particular , we focused on revealing the underlying sources that fuel tumor heterogeneity , a key feature of pRCC . We identified mutation rate dispersion of C-to-T transitions in CpGs motifs contributed the most to the inter-sample mutation spectra variation . We further pinned down the cause of dispersion by showing the hypermethylated cluster , identified in the previous TCGA study [3] , had a higher C-to-T rate in CpGs . Although increased C-to-T in CpGs is likely the result of hypermethylation , we cannot rule out the possibility the change of mutation landscape plays a role in cancer development . For example , C-to-T in methylated CpGs causes loss of methylation , which could have effects on local chromatin environment , trans-elements recruitment and gene expression regulation . In our study , we observed C-to-Ts in CpGs were enriched in coding regions , which suggested they might have a higher functional impact in the cancer genome . Significant APOBEC activities and consequential mutation signatures were observed in one type II pRCC case . APOBEC activities were known to be prevalent in UCs [22 , 23] . We also successfully detected prominent APOBEC signatures in all three UC samples processed in the same pipeline as pRCCs . Intriguingly , despite being considered to have the same cellular origin with pRCC , we were not able to detect meaningful APOBEC activities in ccRCC . This was in agreement with previous studies [12] . APOBEC mutation signature was also found in a small percentage of chromophobe renal cell carcinoma [37] , although they are believed to have a different cellular origin . APOBEC activities have been linked with genetic predisposition and viral infection [38] . Given a statistically robust signal in our conservative algorithm , it is plausible that a small fraction of type II pRCCs might share some etiologically and gnomically similarities with UC . Standard treatment for UC differs significantly from the one for pRCC . Pending further research , this finding might suggest actionable clinical implications . The chromatin remodeling pathway is highly mutated in pRCC [3 , 24 , 25] . Several chromatin remodelers have been identified as cancer drivers in pRCC . We investigated the relationship between samples with mutated chromatin remodelers and those without such mutations in terms of mutation landscape . We demonstrated pRCCs with defects in chromatin remodeling genes showed higher mutation rate in general , driven by an even stronger mutation rate increase in putative open chromatin regions in normal kidney tissues . This is likely because chromatin remodeling defects disrupt normal open chromatin environment and impede DNA repairing in these regions . It is known that replication time strongly governs local mutation rate . Early replication regions have fewer mutations . But the difference dissipates when DNA mismatch repair becomes defective [21] . In our study , we found this correlation weakened in samples with mutated chromatin remodeling genes , presumably caused by failure of replication error repair in an abnormal chromatin environment . Through defects in chromatin remodeling genes , a tumor alters its mutation rate and landscape , which might provide it advantage in cancer evolution . Yet , high mutation burden in functional important open chromatin regions also raises the chance that tumor antigens activate the host immune system . Researchers found tumors with DNA mismatch repair deficiency responded better to PD-1 blockage [39] . These tumors also accumulate more mutations in early replicating regions [26] . Thus chromatin remodeler alterations might as well correlate with higher response rate of immunotherapy , which is worth further studies . Finally , we constructed individual evolutionary trees for all 35 samples . This is the first study inferring tumor evolutionary trees using a large number of SNVs from WGS in pRCC . Benefited from a large number of SNVs , the tree construction became more statistically robust and revealed more details . In general , evolutionary trees gave us the opportunity to observe how pRCC heterogeneity developed over time . They revealed the history of the tumor and how mutations accumulated . We discovered the trees exhibited four major types of topologies , reflecting different levels of heterogeneity . Type II pRCCs showed distinct evolutionary topologies from type I , perhaps indicating an association with greater heterogeneity and different evolving trajectories . In this first whole genome study of pRCC , we found several novel noncoding alterations that might drive tumor development and we explored the mutational landscape and evolutionary trees to better understand tumor heterogeneity . However , due to a limited sample size , some of our statistical tests were underpowered . As the cost of sequencing keeps dropping and technology for data management and processing continues advancing , we expect to have more whole genome sequenced tumors in the near future [40] . With a larger cohort , we hope to gain enough power to test the hypotheses we formed as well as further explore the noncoding regions of pRCC .
pRCC and ccRCC WXS and pRCC WGS variants calls were downloaded from the TCGA Data Portal ( https://gdc-portal . nci . nih . gov/legacy-archive/search/f ) and TCGA Jamboree ( https://tcga-data-secure . nci . nih . gov/tcgafiles/tcgajamboree ) respectively . pRCC RNAseq , RPPA and methylation data ( under project ID: TCGA-KIRP ) were downloaded from TCGA Data Portal as well . Wavelet-smoothed repli-seq data was obtained as a part of ENCODE project [41–43] and downloaded from UCSC Genome Browser ( Also accessible under GSE34399 in the Gene Expression Omnibus ) . DHS data ( fetal , kidney cortex ) were obtained from Roadmap Epigenomics Project and are accessible from http://www . genboree . org/EdaccData/Current-Release/sample-experiment/Fetal_Renal_Cortex/Chromatin_Accessibility/ . We downloaded pRCC clinical outcomes from TCGA Data Portal ( https://tcga-data . nci . nih . gov/tcga/tcgaDownload . jsp ) . pRCC samples that failed the histopathological review were excluded [3] . In total , we included 277 patients in our analyses ( S1 Fig , S1 Table ) . For germline calls , the majority of samples , 163 out of 277 , were supported by germline SNV callings from two centers ( BCM and BI ) . 100% genotype concordance rate was observed . Also , 162 curated rs11762213 genotypes were in agreement with automated call sets . All calls have alternative allelic fraction of 0 . 42 to 0 . 68 , supporting heterozygous genotype [11] . Calls from BI all have genotype quality scores >125 and all calls in BCM pass the caller filter . With proved high confidence in the accuracy of genotyping rs11762213 in the germline , we recruited additional 114 samples from single-center ( BCM ) , automated calls to form an extensive patient set ( S1 Fig ) . For somatic SNVs in MET , after excluding cases that were recruited in the TCGA study , we formed an additional set encompassing 117 patients . Five callings were supported by two centers . The rest were supported by single-center ( BCM ) automated calls . Cancer-specific survival was defined using the same criteria as described in a ccRCC study [9] . Deaths were considered as cancer-specific if the “Personal Neoplasm Cancer Status” is “With Tumor” . If “Tumor Status” is not available , then the deceased patients were classified as cancer-specific death if they had metastasis ( M1 ) or lymph node involvement ( ≥ N1 ) or died within two years of diagnosis . An R package , “survival” , was used for the survival analysis . We remapped all reads using bwa 0 . 7 . 12 , which supports split read mapping [44] . Then we used DELLY [45] with default parameters for somatic SV calling . To avoid sample contamination or germline SVs , we filtered our call set against the entire TCGA pRCC WGS dataset , regardless of sample match . We discharged all callings that were marked “LowQual” ( PE/SR support below 3 or mapping quality below 20 ) . Last , to further eliminate germline contamination , we filtered out SVs that show at least 0 . 8 reciprocally overlapping with 1000 Genome Phase 3 SV call set ( only 1/425 filtered out ) . For Lumpy-SV [46] , we ran it with default parameters . We also filtered the results using the 1000 Genome Phase 3 call set and required the SV have both paired-end and split reads supports . WGS Mutations were extracted with flanking 5’ and 3’ nucleotide context . The raw mutation counts were normalized by trinucleotide frequencies in the whole mappable genome . To identify signatures in the mutation spectra , we used a robust , objective LASSO-based method . First , 30 known signatures were downloaded from COSMIC ( http://cancer . sanger . ac . uk/cosmic/signatures ) . Then we solved a positive , zero-intercept linear regression problem with L1 regularizer to obtain signatures and corresponding weights for each genome . Specifically , we solved the problem: minW ( ∥SW−M∥2+ λ∥W∥ ) Where M is the mutation matrix , containing the mutations of each sample in 96 nucleotide contexts . S is the 96×30 signature matrix , representing the mutation probability in 96 nucleotide contexts of the 30 signatures . W is the weighting matrix , representing the contribution of 30 signatures to each sample . The penalty parameter lambda ( λ ) was determined empirically using 10-fold cross-validation individually for every sample . λ was chosen to maximize sparsity and constrained to keep mean-square error ( MSE ) within one standard error of its minimum . Last , we discharged signatures that composite less than 5% of the total detectable signatures . In total , we collected HumanMethylation450 BeadChip array data for 139 samples that are either methylation cluster 1 or 2 . We used an R package “IMA” to facilitate analysis [47] . After discharging sites with missing values or on sex chromosomes , we obtained beta-values on 366 , 158 CpG sites in total . Then we tested beta-values of each site by Wilcoxon rank sum test between two methylation clusters . After adjusting p-value using Benjamini-Hochberg procedure , we called 9 , 324 ( 2 . 55% ) hypermethylation sites . These sites had an adjusted p-value of less than 0 . 05 and mean beta-values in methylation cluster 1 were 0 . 2 or higher than in methylation cluster 2 . We used the method described by Roberts et al . [23] . For every C>{T , G} and G>{A , C} mutation we obtained 20bp sequence both upstream and downstream . Then enrichment fold was defined as: Enrichment Fold= MutationTCW/WGA × ContextC/GMutationC/G×ContextTCW/WGA Here TCW/WGA stands for T[C>{T , G}]W and W[G>{A , C}]A . W stands for A or T . p-value for enrichment were calculated using one-sided Fisher-exact test . To adjust for multiple hypothesis testing , p-values were corrected using Benjamini-Hochberg procedure . WXS data for APOBEC enrichment and signature analysis was obtained from a processed somatic call set: hgsc . bcm . edu_KIRP . IlluminaGA_DNASeq . 1 . protected . maf . This dataset includes 155 pRCC samples and three UC samples . We used hgsc . bcm . edu_KIRC . Mixed_DNASeq . 1 . protected . maf for ccRCC analyses . We identified chromatin remodeling genes based on its significance in pRCC and function . Our gene list was the intersection of genes in the original TCGA pRCC study [3] molecular feature table with the chromatin remodeling and SNI/SWF pathway gene lists . Our gene set included ten genes: SETD2 , KDM6A , PBRM1 , SMARCB1 , ARID1A , ARID2 , MLL2 ( KMT2D ) , MLL3 ( KMT2C ) , MLL4 ( KMT2B ) , EP300 . We defined chromatin remodeling defect as nonsynonymous mutations in these genes . For missense mutations , we additionally filtered out mutations with Polyphen score [48] less than 0 . 9 ( benign ) . We noticed BAP1 is not in the gene list . However , adding BAP1 into the list did not change the significance of our key tests ( p<0 . 0115 for mutation counts in DHS and p<0 . 020 for mutation percentage in DHS ) . For replication time , in order to avoid cell type redundancy , we only kept GM12878 as the representative of all lymphoblastoid cell lines . Eleven cell types were included in our analysis: BG02ES , BJ , GM12878 , HeLaS3 , HEPG2 , HUVEC , IMR90 , K562 , MCF7 , NHEK , SK-NSH . Wave smoothed replication time signal was averaged in a±10kb region from every mutation . To avoid potential selection effects , we removed mutations in exome and flanking 2bp . Regions overlapping with reference genome gaps and DAC blacklist ( https://genome . ucsc . edu/ ) were removed as well . Last , we picked the median number from 11 cell types at each mutation position . To test the significance of replication time of noncoding mutations between two groups , we defined the ones have replication time stand above 90 percentile in all pooled mutations as “mutations in early replicating regions” . Then we calculated the percentage of “mutations in early replicating regions” of total mutations for each sample and compared between the two groups using one-sided rank-sum test . We used PhyloWGS [49] to infer the evolutionary trees for each individual tumor . To mitigate the effects of copy number change , we removed all the SNVs inside the copy number change regions as defined by the assay-based method in the original TCGA study [3] . To be prudent , we filtered SNPs in any region with an absolute log2 tumor to normal copy number ratio larger than 0 . 3 . Additionally , we removed all SNVs with allele frequency higher than 0 . 6 as they were likely affected by copy number loss . | Renal cell carcinoma accounts for more than 90% of kidney cancers . Papillary renal cell carcinoma ( pRCC ) is the second most common subtype of renal cell carcinoma . Previous studies , focusing mostly on the protein-coding regions , have identified several key genomic alterations that are critical to cancer initiation and development . However , researchers cannot find any key mutation in a significant portion of pRCC . Therefore , we carry out the first whole-genome study of pRCC to discover triggering DNA changes explaining these cases . By looking at the entire genome , we find additional potentially impactful alterations both in and out of the protein-coding regions . These newly identified critical mutations from scrutinizing the entire genome help complete our understanding of pRCC genomes . Two alterations we find are associated with prognosis , which could aid clinical decisions . We are also able to unveil mutation patterns , signatures and tumor evolutionary structures , which reflect the mutagenesis processes and help understand how heterogeneity arises . Our study provides valuable additional information to facilitate better tumor subtyping , risk stratification and potentially clinical management . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cancer",
"genomics",
"renal",
"cell",
"carcinoma",
"medicine",
"and",
"health",
"sciences",
"carcinomas",
"cancers",
"and",
"neoplasms",
"basic",
"cancer",
"research",
"genitourinary",
"tract",
"tumors",
"oncology",
"mutation",
"dna",
"replication",
"epigenetics",
"d... | 2017 | Whole-genome analysis of papillary kidney cancer finds significant noncoding alterations |
Infectious diseases pose a severe worldwide threat to human and livestock health . While early diagnosis could enable prompt preventive interventions , the majority of diseases are found in rural settings where basic laboratory facilities are scarce . Under such field conditions , point-of-care immunoassays provide an appropriate solution for rapid and reliable diagnosis . The limiting steps in the development of the assay are the identification of a suitable target antigen and the selection of appropriate high affinity capture and detection antibodies . To meet these challenges , we describe the development of a Nanobody ( Nb ) -based antigen detection assay generated from a Nb library directed against the soluble proteome of an infectious agent . In this study , Trypanosoma congolense was chosen as a model system . An alpaca was vaccinated with whole-parasite soluble proteome to generate a Nb library from which the most potent T . congolense specific Nb sandwich immunoassay ( Nb474H-Nb474B ) was selected . First , the Nb474-homologous sandwich ELISA ( Nb474-ELISA ) was shown to detect experimental infections with high Positive Predictive Value ( 98% ) , Sensitivity ( 87% ) and Specificity ( 94% ) . Second , it was demonstrated under experimental conditions that the assay serves as test-of-cure after Berenil treatment . Finally , this assay allowed target antigen identification . The latter was independently purified through immuno-capturing from ( i ) T . congolense soluble proteome , ( ii ) T . congolense secretome preparation and ( iii ) sera of T . congolense infected mice . Subsequent mass spectrometry analysis identified the target as T . congolense glycosomal aldolase . The results show that glycosomal aldolase is a candidate biomarker for active T . congolense infections . In addition , and by proof-of-principle , the data demonstrate that the Nb strategy devised here offers a unique approach to both diagnostic development and target discovery that could be widely applied to other infectious diseases .
Infectious diseases are a leading cause of mortality and morbidity after non-communicable diseases worldwide [1] . Although the majority of these infectious diseases are treatable , the lack of better diagnostic facilities in the developing countries is a major impediment to their control [2] . While disease diagnosis based on clinical signs is relatively cheap , it is not reliable as some infections are latent , mixed , and/or cause pathologies with overlapping symptoms . In some cases , disease diagnosis based on clinical signs may be of little use as the symptoms only manifest themselves when the patient has entered a terminal stage . The diagnosis of infectious diseases is facilitated by the use of sophisticated molecular techniques such as PCR [2 , 3] . While these are reliable and increase the chance of detecting an infection even before the manifestation of clinical signs , their use in resource-constrained settings is untenable as they require automated equipment . To overcome the need for the latter , efforts are ongoing to adapt PCR to field conditions through point-of-care ( POC ) nucleic acid devices [4 , 5] and loop-mediated isothermal amplification [6] . The adoption of these POC nucleic acid techniques in developing countries remains uncertain given the fact that these new devices still operate on expensive reagents , require amplification steps , and sample processing . The majority of the cases of infectious diseases are found in developing countries [1 , 7] , where basic laboratory facilities are scarce . Therefore , most laboratories in these parts of the world have resorted to relatively inexpensive ( but less sensitive ) microscopy-based techniques for routine diagnosis of infectious diseases . Among other factors , the over-reliance on microscopy contributes to the high burden of infections in these regions , as most pathogens escape detection thereby allowing their proliferation and continued transmission . Hence , there is an urgent need for diagnostic tools that can immediately and reliably detect active infections under field conditions . Where nucleic acid-based assays are unaffordable and sensitivity of microscopy becomes limited , enzyme-linked immunosorbent assay ( ELISA ) can be used as an alternative tool for detection of infectious diseases through antibody [8] or antigen [9] detection . ELISA is a robust technique which operates without sophisticated equipments therefore favoring its application in less well-established laboratories such as those found in developing countries . Furthermore , ELISA can be easily adapted to non-laboratory conditions as exemplified by the availability of several immuno-based POC tests for important infectious diseases [10] . Antibody-based ELISA relies on the host immune system to amplify the detection limit . Two issues arise that need to be taken into account , i . e . reduced specificity due to cross-reactivity [11] and the inability to differentiate past from ongoing infections because of the persistence of antibodies in the circulation after clearance of infection [12 , 13] . While both issues are obviously avoided using an antigen detection format , the development of the latter faces its own hurdles . First , the antigen should be present in detectable quantities during an active infection [14] and should not cross-react with diagnostic tests used for the detection of other endemic parasites . Second , the capture and detection antibodies of the ELISA should be able to out-compete the host anti-pathogen antibodies which usually form immune complexes with the circulating antigen [15] . Nanobody ( Nb ) technology [16] offers prospects for the identification of native parasite antigens [17] as well as the detection of pathogens [18 , 19] and these have rejuvenated hope for refining the developments of ELISA and immuno-based POC tests . Nbs are recombinant proteins derived from the gene coding for the antigen binding portion of heavy-chain only IgG antibodies [20] . Nbs possess a relatively high thermostability [21 , 22] and are thereby attractive for the development of immunodiagnostic tests that could be applicable in hot climate such as the sub-Saharan Africa . Moreover , Nbs could be used to overcome certain challenges faced by tests based on monoclonal antibodies ( MAbs ) . It has been documented that host antibodies usually sequester parasite antigens [15] . Given that MAbs are conventional antibodies , the host IgG molecules might conceal the epitopes from the MAbs employed in the diagnostic test [23] . Nb-based diagnostics could circumvent this binding interference caused by the host’s antibody response . Because Nbs possess extensive CDR3 loops [24] , this allows tight binding of epitopes distinct from those recognized by conventional antibodies [25] . In other words , Nbs should detect both free antigens as well as those bound by host antibodies , which would make Nb-based diagnostic tests rather attractive . The exploitation of the Nb technology might expedite the development of potent ELISA and immuno-based POC tests for the endemic tropical diseases . Here , we describe the development of a Nb-based ELISA for the specific detection of active pathogen infections . Trypanosoma congolense , the most important causative agent of African Animal Trypanosomosis ( AAT ) , was chosen as a model system for our studies . Besides vector control , the only way to curb the spread of T . congolense infections is by case detection followed by drug treatment . Unfortunately , current diagnostic tests are not very reliable or user-friendly under field conditions . The buffy coat technique ( BCT ) [26] , which is a general microscopy-based test for trypanosomes , is limited by sensitivity in cases where there is low parasitemia such as in chronic infections or where sample examination is delayed . While sensitive and specific detection of T . congolense infections can be achieved through species specific [27] or pan-trypanosome [28–31] PCRs , these molecular assays are not widely used for diagnosis of animal trypanosomes . Attempts to develop antibody-based ELISAs failed due to lack of specificity [32 , 33] . Even if it would have succeeded , the persistence of antibodies in the circulation several months after clearance of infections would compromise its application for establishing cure as was observed with the test for fascioliasis [12] . An attempt to develop an antigen detection ELISA test for T . congolense based on MAbs was hampered by a lack of sensitivity [34] . Thus , limitations of the current diagnostic tests for trypanosomes require that more research be done to provide assays that meet the expectations of end-users . In this study , a potent Nb ( Nb474 ) was obtained from an immune cDNA library prepared starting from the immunization of an alpaca with T . congolense TC13 soluble proteome . Nb474 was used for development of a T . congolense specific antigen detection homologous sandwich ELISA ( Nb474-ELISA ) . The results presented in this study demonstrate the potential of Nbs as tools for the specific diagnosis of T . congolense infections and of T . congolense glycosomal fructose-1 , 6-bisphosphate aldolase ( TcoALD ) as a candidate biomarker . Moreover , we show that the Nb474-ELISA is able to detect active T . congolense infections under experimental and natural conditions , and that it serves as a test-of-cure in experimental settings .
All the animal experiments were performed according to directive 2010/63/EU of the European parliament for the protection of animals used for scientific purposes and approved by the Ethical Committee for Animal Experiments of the Vrije Universiteit Brussel ( clearance numbers 11-220-6 and 13-220-3 ) . The trypanosomes used in the study were obtained from various sources . T . congolense TC13 was kindly provided by Dr . Henry Tabel ( Canada ) , T . congolense IL3000 was obtained from the laboratory of Microbiologie Fondamentale et Pathogénicité , Université de ( Bordeaux ) , and T . congolense ( STIB68 , IL1180 , TRT55 , J423 , Ruko14 , MF3cl2 , MF5cl4 , Alick339c2 and Kapeya357c1 ) were obtained from the Institute of Tropical Medicine ( Antwerp ) . T . b . brucei AnTat 1 . 1 , T . vivax ILRAD700 and T . evansi STIB816 were also obtained from the Institute of Tropical Medicine ( Antwerp ) . Mice ( C57BL/6 ) approximately 6 weeks old ( Janvier ) were infected with 5000 trypanosomes per animal through intra-peritoneal route . Blood was collected by cardiac puncture ( for one time bleeding ) or peri-orbital route ( for multiple collections at different time points ) . The blood was stored at 4°C for at least 17 h to allow clotting prior to serum collection for immediate examination or storage at -20°C . For quantification of parasites , the mouse tail tip was nipped with scissors and 2 . 5 μl of blood was drawn by pipette . The blood was diluted ( 1/200 ) in PBS and trypanosomes were counted using a Neubauer haemocytometer on a Nikon Eclipse TS 100 inverted microscope ( Japan ) . Cattle sera consisting of T . congolense IL1180 experimentally infected ( n = 45 ) and negative control ( n = 17 ) used for the evaluation of the Nb-ELISA were kindly provided by Prof . Luis Neves of the Faculty of Veterinary Medicine , Universidade Eduardo Mondlane ( Mozambique ) . Hemolysed blood samples collected from domestic animals naturally infected with various endemic haemoparasites that were used to assess the cross-reactivity of the assay were kindly provided by Susan Ndyanabo of the Japanese International Cooperation Agency diagnostic laboratory , COVAB . The clinical samples comprised T . congolense positive cattle ( n = 2 ) , T . brucei positive cattle ( n = 3 ) and dog ( n = 2 ) , T . evansi positive camel ( n = 1 ) , T . theileri positive cattle ( n = 12 ) , Theileri parva positive cattle ( n = 44 ) , Babesia bigemina positive cattle ( n = 1 ) , B . canis positive dog ( n = 1 ) , Anaplasma marginale cattle ( n = 2 ) and goat ( n = 2 ) . The clinical samples were examined for trypanosomes by BCT and 18S-PCR-RFLP [30] . All the samples that were positive for T . brucei by 18S-PCR-RFLP were cross-checked for T . evansi by RoTat PCR [35] and non-RoTat PCR [36] . T . parva , Babesia sp and A . marginale were detected by Giemsa-stained smear . T . congolense IL3000 BSF and PCF parasites were cultured and harvested as described [37] . The trypanosomes grown in mice were purified according to [38] . Briefly , infected blood was centrifuged [805 g , 10 min . , 22°C] , buffy coat was collected and loaded onto a PD-10 column ( GE Healthcare ) packed with Phosphate Saline Glucose pre-equilibrated DE-52 resin ( Whatman ) at pH 7 . 5 for purification of T . congolense and T . vivax; and pH 8 . 0 for purification of T . b . brucei and T . evansi . Trypanosomes eluted from the column were harvested by centrifugation ( 1811 g , 15 min . 22°C ) . The cell pellet was resuspended in 100 mM phosphate buffer pH 7 . 0 containing complete Protease Inhibitor Cocktail ( Roche ) . Then , trypanosome soluble protein was prepared as described [39] and the protein concentration was measured by NanoDrop . An alpaca was injected once a week , for six weeks , with 100 μg T . congolense TC13 soluble proteome potentiated with GERBU adjuvant LQ ( GERBU BIOTECHNIK GmbH ) . On the 7th week , 50 ml of blood was collected and peripheral blood lymphocytes were isolated for RNA extraction . cDNA was synthesized and amplified using Call001 and Call002 primers [39] . The amplicon was resolved on 1% agarose gel ( Lonza ) and eluted from the gel using GenElute Gel extraction Kit ( Sigma ) following the instructions of the manufacturer . The eluted DNA was amplified with nested primers A4short forward ( 5'-GATGTGCAGCTGCAGGAGTCTGGA/GGGAGG-3' ) and 38 reverse ( 5'-GGACTAGTGCGGCCGCTGGAGACGGTGACCTGGGT-3' ) followed by cloning in pHEN4 vector [40] . The library construction was ended with transformation of the ligated pHEN4 construct into fresh electrocompetent Escherichia coli ( TG1 ) cells . For panning , the library was transfected with 1012 M13K07 phage particles ( Invitrogen ) to obtain display phage library . Then 1011of display phages were adsorbed on a 96-well Nunc plate ( Thermo scientific ) coated in parallel with soluble proteome ( 25 μg/well ) of four different T . congolense strains ( STIB68 , TRT17 , TC13 or MF3cl2 ) . Bound phages were eluted with 100mM triethylamine ( pH~10 ) and neutralized with 100 μl 1M Tris ( pH 8 . 2 ) . After three panning cycles , enrichment was assessed by phage ELISA . For phage ELISA , soluble proteins of the four T . congolense strains used in panning were coated in parallel on Nunc plate ( 1μg/well ) and uncoated ( control ) wells were filled with PBS ( 100 μl ) . Display phages ( 1010 ) of panning rounds 0–3 were added to coated and uncoated wells . The fifth pair ( blank ) of wells were filled with blocking buffer only . Thereafter , anti-M13 conjugate horse radish peroxidase ( HRP ) ( 100 μl ) ( GE Healthcare ) diluted ( 1/2000 ) in blocking buffer was added followed by 1-Step ultra 3 , 3′ , 5 , 5′-tetramethylbenzidine ( TMB ) substrate ( Thermo scientific ) added at ( 100 μl/well ) . The reaction was stopped with 1M sulphuric acid ( 50 μl/well ) . OD450nm was read on spectrophotometer ( ELX808 Ultra-microplate reader , Bio-Tek instruments ) using Gen5 software . For screening , single colonies ( of the panning rounds 2 and 3 ) were expressed in Terrific Broth medium as mini culture ( 1ml ) supplemented with ampicillin and 1 mM isopropyl-β-D-thiogalactosidase ( IPTG ) . Cells were harvested , disrupted by osmotic shock and the periplasmic extracts ( 100 μl ) were added to wells coated with soluble proteome ( 1 μg/well ) . A screening strategy was adopted in order to maximize chances of selecting T . congolense cross-reactive Nbs . Thus , clones obtained after panning on TC13 soluble proteome were screened on MF3cl2 soluble protein; clones obtained after panning on MF3cl2 soluble proteome were screened on TC13 soluble proteome; clones obtained after panning on STIB68 soluble proteome were screened on TRT17 soluble proteome; and clones obtained after panning on TRT17 soluble proteome were screened on STIB68 soluble proteome . Negative control wells ( also coated with soluble proteome ) were filled with PBS ( 100 μl ) instead of periplasmic extract . Positive binders were detected using biotinylated mouse Anti-HA IgG ( Eurogentec ) , HRP-conjugated streptavidin ( strep-HRP ) ( Jackson ImmunoResearch laboratories ) and TMB substrate . Colonies were considered positive when the ratio of OD between the test and control wells was ≥ 2 . All the positive binders selected were re-screened on naive mouse serum diluted ( 1/10 ) in PBS in order to identify mouse protein binders . Thereafter the genes of T . congolense specific binders were sequenced and the data was processed using Clone Manager 9 Professional Edition software employing the BLASTN 2 . 2 . 29+ program [41] . The Nb gene in pHEN4 construct was amplified by PCR followed by double digestion with PstI and Eco91I . The fragment was ligated into pHEN6c as well as pBAD expression vectors for incorporation of a hexahistidine tag ( His6-tag ) and biotin acceptor domain ( for in vivo biotinylation ) , respectively . Whereas the Nb ligated pHEN6c construct was transformed alone in E . coli WK6 , the ligated pBAD construct was co-transformed with BirA plasmid which aid in vivo biotinylation . For expression , starter culture ( 1 ml ) was inoculated in 330 ml terrific broth ( TB ) supplemented with ampicillin ( 100 μg/ml ) for cells transformed with pHEN6c or ampicillin ( 100 μg/ml ) and chloramphenicol ( 35 μg/ml ) for cells co-transformed with pBAD/BirA . Cells were grown at 37°C until an OD600nm of ± 0 . 6 and expression induced with 1 mM IPTG . For culture expressing biotinylated Nb , 0 . 05 mM D-biotin ( ACROS Organics ) was added 30 min before induction . After induction , cells were grown overnight at 28°C . The next morning cells were harvested and lysed by osmotic shock . The His6-tag Nb ( Nb474H ) and biotinylated Nb ( Nb474B ) in the perisplasmic extract was affinity purified on His-Select Nickel ( Sigma ) and streptavidin mutein matrix ( Roche ) , respectively , followed by size exclusion chromatography on ÄKTA ( GE healthcare ) . Protein production was analyzed by SDS-PAGE under reducing conditions using 10% Bis-Tris gel ( Novex Life Technologies ) and Western blot . The aim of this study was to develop a Nb-based sandwich test such as described previously [42] . Therefore , the two Nb formats ( Nb474H and Nb474B ) were titrated in a checkerboard fashion in order to arrive at optimal concentration of capturing and detection Nbs required for the highest signal intensity ( OD450nm ) . In this experiment , the capturing Nb ( Nb474H ) was titrated against detecting Nb ( Nb474B ) using T . congolense TC13 soluble protein ( S1 Materials and Methods ) . After checkerboard titration , a protocol for Nb474-based homologous sandwich ELISA was established whereby Nunc plate was coated with 100 μl/well Nb474H ( 0 . 02 μg/ml ) in PBS and incubated overnight at 4°C . The following morning , the non-coated Nb was discarded and wells washed with PBS-T three times . Washed wells were blocked with 300 μl 5% milk for 2 h at 22°C . The blocking buffer was discarded and wells were washed three times . Test samples were added ( 100 μl/well ) and incubated for 1 h at 22°C . Wells were emptied and washed three times . Then , Nb474B ( 100 μl ) diluted to 0 . 02 μg/ml was added per well and incubated for 1h at 22°C . Wells were emptied and washed four times . Strep-HRP diluted to 1 μg/ml in blocking buffer was added ( 100 μl/well ) followed by incubation for 1 h at 22°C . Conjugate was discarded , wells were washed five times and TMB substrate was added at 100μl/well . Color development allowed to progress for at least 25 min and the reaction was stopped using 1M sulphuric acid ( 50 μl/well ) and OD450nm was read . To determine the detection limit of the Nb474-ELISA , T . congolense TC13 infected mouse blood ( with 5 . 86x108 trypanosomes/ml ) was serially diluted ( 4-fold ) in naive mouse blood . The set of samples were incubated overnight at 4°C . The following day , plasma was harvested and then tested by ELISA . To test the ELISA upon prolonged storage at different temperatures , mice infected with T . congolense ( n = 10 ) were bled on day 7 post-infection . Sera were collected and pooled followed by incubation at different temperature ( 4 , 22 or 37°C ) for a specified number of days ( 1 , 3 , 5 or 7 ) . The sera aliquots were tested by Nb474-based ELISA and the OD450nm was compared . To assess if the Nb474-ELISA can be used as a tool for monitoring cure , an infection-treatment assay was performed and OD450nm was monitored alongside blood parasite load . In this experiment , mice were randomly assigned to two groups of 6 animals each . On day 0 post-infection , baseline sera were obtained from all the subjects and pooled by group . Thereafter , all the subjects in each group were infected with T . congolense TC13 . On day 7 , group 2 was treated with 0 . 8 mg Berenil in 50 μl distilled water . During the course of the experiment ( from day 3 until day 21 ) mice were bled on alternate days in order to monitor the level of antigen and the concentration of trypanosomes in the circulation . In addition to microscopy , cure of treated mice was assessed during four weeks by sub-inoculation of naïve mice with blood collected from the treated group on weekly basis . Nb474 was labeled at the lysine residues with Alexa Fluor 488 ( Invitrogen Molecular probes ) following the instructions of the manufacturer . After labeling , the Nb ( 0 . 3 μg ) was incubated with purified motile as well as paraformaldehyde fixed T . congolense BSF cells and no binding was detected . Thereafter , the Nb ( 0 . 3 μg ) was incubated with cells that were fixed and permeabilized using a commercial kit ( BD Biosciences ) following the instructions of the manufacturer followed by incubation with 1 mg/ml 4’ , 6’-diamidino-2-phenylindle ( DAPI ) for 5 min . The preparation was examined using a Zeiss Axio Imager Z1 fluorescence microscope . Based on binding result obtained from the second experiment , co-localization of Alexa Fluor 488 labeled Nb474 with a TcoALD cross-reactive anti T . brucei aldolase ( TbALD ) monoclonal antibody ( Anti-TbALD MAb ) ( ProteoGenix ) against synthetic peptide PEVMIDGTHDIETCQRVSQHVWS was performed on cultured T . congolense IL3000 bloodstream form ( BSF ) and procyclic form ( PCF ) parasites . In this experiment , the cultured cells were resuspended in 320 μl 1% ( v/v ) formaldehyde in PBS for 10 min . Then 1M glycine dissolved in PBS ( 40 μl ) was added and mixed by a pipette . The mixture was incubated for 10 min and diluted with PSG ( 800 μl ) . The dilution ( 20 μl ) was placed in different wells on a microscope slide ( Thermo scientific ) and the liquid on the slide was dried in a fume hood . Wells were washed three times with 50 μl BSA ( 1 mg/ml ) in PBS . An Anti-TbALD MAb ( 20 μl ) diluted ( 1/10 ) in a diluent solution [1mg/ml BSA in PBS containing 0 . 1% ( v/v ) Triton X-100] was added to the wells and incubated for 1 h . Wells were washed three times . Alexa Fluor 594 labelled goat anti-mouse ( Invitrogen ) ( 20 μl ) diluted ( 1/100 ) in a diluent solution containing Alexa Nb474 ( 3 μg/ml ) was added to the washed slide . The slide was then kept in a dark humidified box for 45 min and washed with PBS ( 50 μl/well ) three times . DAPI was added to the washed slide and kept in the dark for 3 min . Wells were washed with PBS ( 50 μl/well ) three times and the preparation was examined immediately . Starting materials for the purification of the antigen from soluble proteome and plasma were 7 mg T . congolense soluble proteome diluted in 20 ml PBS and 30 ml of serum collected at the first peak parasitemia . The purification process was started by concentrating the samples using a 100 , 000 molecular weight cut-off ( MWCO ) concentrator ( Sartorius stedim biotech ) . Thereafter both fractions ( the concentrate and filtrate ) were tested by Nb-ELISA after normalization for protein concentration . A higher OD450nm was recorded in the concentrate compared to filtrate . Then , the concentrate was washed twice by adding 5ml PBS to the concentrator and draining by centrifugation ( 453 g , 4°C ) until 1 ml . Washed concentrate was stored at -20°C until the antigen was purified . The homologous Nb sandwich ELISA was then employed to isolate the native target antigen from the T . congolense soluble proteome , the sera collected from T . congolense infected mice , and the secretome of T . congolense IL3000 . Four Nunc plates were coated with capturing Nb ( Nb474H ) ( 5 μg/well ) overnight at 4°C . The following morning , coating was discarded and the plate was washed three times . Protein-Free T20 Blocking Buffer ( Thermo scientific ) was added ( 300 μl/well ) for 3 h at 22°C . Blocking buffer was discarded and wells were washed three times . Soluble proteome diluted in blocking buffer was added ( 10 μg/well ) to the first pair of the plates and the second pair of the plates was filled with sera ( 100 μl/well ) diluted ( 1/10 ) in blocking buffer . Binding of the target antigen to Nb474H proceeded for 2 h at 22°C and unbound molecules were discarded . Detection Nb ( Nb474B ) was added ( 2 ng/well ) and incubated for 1 h at 22°C . Subsequently , the unbound Nb474B was discarded and the wells were washed four times . Strep-HRP was added to four control wells that were purposely included to serve as reporter for sandwich formation . The rest of the wells were filled with blocking buffer ( 100 μl/well ) for 1 h at 22°C . The plates were emptied and washed five times . Control wells were filled with TMB ( 100 μl/well ) and wells where captured antigen was to be eluted were filled with 0 . 3M glycine HCl pH 3 . 0 ( 100μl/well ) . Elution was proceeded with gentle rocking on bench top shaker over 30 min at 22°C . The eluted samples were collected and pooled according to the starting material ( soluble proteome or sera ) . Solutions obtained from each pair of plates ( about 19 . 2 ml ) was concentrated to 1 ml using a 5000 MWCO concentrator of 15 ml capacity and then to 30 μl using a 3000 MWCO concentrator of 400 μl capacity . The concentrators with low MWCO and capacities were employed in the last steps in order to trap the eluted native target antigen-Nb complex as well as dissociated Nbs . The eluted samples ( 6 μg from lysate plates and 8 μg from serum plates ) were resolved on SDS-PAGE alongside Nbs ( Nb474H and Nb474B ) as control and proteins visualized using silver staining [43] . The purification of the antigen from secretome was performed by QuickPick IMAC ( Bio-Nobile ) following the instructions of the manufacturer . Subsequently , Nb474H , flow through , wash , and eluted samples were resolved on SDS-PAGE under reducing conditions followed by staining with coomassie blue . The proteins in the bands at ± 40 kDa that appeared in all the three eluted samples were identified by mass spectrometry . Bands of interest were excised from the SDS-PAGE gel , digested with trypsin , and analyzed by liquid chromatographic tandem MS as described [44] . Briefly peptides were separated by an acetonitrile gradient on a C18 column and the MS scan routine was set to analyze by MS/MS the five most intense ions of each full MS scan , dynamic exclusion was enabled to assure detection of co-eluting peptides . Protein identification was performed with SequestHT . In details , peak lists were generated using extract-msn ( ThermoScientific ) within Proteome Discoverer 1 . 4 . 1 . From raw files , MS/MS spectra were exported with the following settings: peptide mass range: 350–5000 Da , minimal total ion intensity 500 . The resulting peak lists were searched using SequestHT against a target-decoy T . congolense protein database ( 33144 entries comprising forward and reversed sequences ) obtained from the Welcome Trust Sanger Institute ( http://www . sanger . ac . uk/resources/downloads/protozoa/trypanosoma-congolense . html ) . The following parameters were used: trypsin was selected with proteolytic cleavage only after arginine and lysine , number of internal cleavage sites was set to 1 , mass tolerance for precursors and fragment ions was 1 . 0 Da , considered dynamic modifications were +15 . 99 Da for oxidized methionine . Peptide matches were filtered using the q-value and Posterior Error Probability calculated by the Percolator algorithm ensuring an estimated false positive rate below 5% . The filtered Sequest HT output files for each peptide were grouped according to the protein from which they were derived and their individual number of peptide spectral matches was taken as an indicator of protein abundance . The raw MS data were then searched against a T . brucei /T . congolense database . Of the proteins identified by mass spectrometry , TcoALD had the highest score . Thus , in order to establish if Nb474 was indeed binding to glycosomal aldolase , formation of a sandwich between the Nb474H and Anti-TbALD MAb was assessed in a heterologous sandwich ELISA where T . congolense TC13 soluble proteome was used as the antigen source . In this ELISA , a 96-well Nunc plate was coated ( 5 μg/well ) with Nb474H diluted in PBS overnight at 4°C . The following morning , coated sample was discarded; wells were washed three times and blocked with 300 μl 5% milk for 2 h at 22°C . Blocking was discarded and wells washed three times . The soluble proteome diluted ( 100 μg/ml ) in blocking buffer was added to the plate ( 100 μl/well ) and incubated for 1 h at 22°C . Wells were emptied and washed three times . An Anti-TbALD MAb diluted ( 1/100 ) in blocking buffer was added ( 100 μl/well ) to the plate and incubated for 1 h at 22°C . The antibody was discarded and wells washed three times . Anti-mouse HRP conjugate diluted ( 1/1000 ) in blocking buffer was added ( 100 μl/well ) and incubated for 1 h at 22°C . The conjugate was discarded and wells were washed five times . The ELISA was developed and stopped as described earlier . Nucleotide sequences coding for aldolase of T . congolense ( EMBL-EBI accession no . CCC93713 . 1 ) , T . b . brucei ( EMBL-EBI accession no . M19994 . 1 ) and Leishmania mexicana ( EMBL-EBI accession no . CAB55315 . 1 ) were retrieved from the GenBank and codon-optimized for expression in E . coli . The codon-optimized sequences were further processed by incorporating nucleotides coding for His6-tag and a proximal tomato etch virus cleavage site . Finally , NdeI and XhoI cutting sites were added to the 5′ and 3′ , respectively , followed by in silico construction in pET-21b ( + ) expression vector . Assembled sequences were submitted to a commercial company ( GenScript ) for synthesis and ligation in the expression vector . The constructs were electro-transformed into E . coli BL21 ( DE3 ) and glycerol stocks of the transformed cells were stored at -80°C . The expression of recombinant aldolase was performed as previously described [45] . During the expression , aliquots ( 5 ml ) of the culture were taken 0 h and 18 h post-induction for preparing crude soluble protein . The aliquots were harvested by centrifugation ( 11 , 325 g , 8 min . , 4°C ) and the pellet was completely resuspended in lysis buffer ( 50 mM Tris pH 7 . 0 , 500 mM NaCl ) for 1 h at 4°C followed by sonication while on ice . The crude lysate protein concentration was measured by NanoDrop and expression was checked by SDS-PAGE as well as Western blot . Since Nb474 could not detect aldolase expression in Western blot , ELISA was employed for this purpose whereby the crude soluble proteins ( 10 μg/well ) were probed with homologous ( Nb474H-Nb474B ) as well as heterologous ( Nb474H-Anti-TbALD MAb ) sandwich ELISA . Concurrently , ELISA involving Anti-His IgG and Anti-TbALD MAb were performed on crude soluble proteins that were coated ( 10 μg/well ) as positive controls . Unpaired t-tests used for group comparisons and the performance of the Nb-based ELISA ( sensitivity and specificity ) were analyzed by GraphPad Prism ( version 6 . 03 ) . Values of p ≥ 0 . 05 were considered non-significant ( ns ) . Where * , ** , *** and ****denote p<0 . 05 , p<0 . 01 , p<0 . 001 , and p<0 . 0001 , respectively . The sequences of T . congolense , T . b . brucei and L . mexicana aldolase genes are available from EMBL-EBI nucleotide sequence database with accession numbers CCC93713 . 1 , M19994 . 1 and CAB55315 . 1 , respectively .
Using the approach summarized in Fig 1 , the most potent T . congolense specific binder ( Nb474 ) was obtained . The Nb library , generated against a T . congolense TC13 soluble proteome preparation , was panned in parallel on soluble proteomes prepared from various T . congolense strains ( TC13 , MF3cl2 , STIB68 and TRT17 ) to obtain T . congolense cross-reactive binders . Each of the enriched libraries was screened on heterologous strains whereby 101 positive colonies were identified . The 7 cross-reactive binders interacting with mouse serum components were excluded in the procedure ( S1 Fig ) . Thereafter , the remaining 94 clones were amplified and sequenced . All the in-frame sequences belonged to the same CDR3 sequence family ( S2 Fig ) . The aim of this study was to develop an assay for the specific diagnosis of T . congolense infections . The assay consists of a Nb474-based homologous sandwich ELISA , in which a His-tagged version of the Nb was used for antigen capturing ( Nb474H ) , while a biotin version of the Nb was used for detection ( Nb474B ) . To test the cross-reactivity of the Nb474-ELISA , the latter was validated on soluble proteome of different species of animal infective trypanosomes . As presented in Fig 2A , a significant high optical density ( OD450nm ) was only recorded on T . congolense soluble proteome , whereas extracts from T . b . brucei , T . vivax or T . evansi score negative . Next , the ELISA was evaluated for its capacity to diagnose experimental T . congolense infections in mice . When comparing the sera of T . congolense positive mice to the sera of naive controls , a significant difference in the average OD450nm of about 4 . 0 was observed between both groups ( Fig 2B ) . Having established that the Nb474-ELISA was specific for T . congolense and that it can detect an active infection , the assay was validated in a set-up in which mice were infected with different strains of T . congolense and other species of trypanosomes . While the ELISA identifies T . congolense infected animals irrespective of the parasite strain , the test scores negative on mice infected with trypanosome species other than T . congolense ( Fig 2C ) . Additionally , the detection limit of the Nb474-based assay was also investigated as uneven OD scores on sera collected from various T . congolense strains are attributed to differences in parasite density . To establish the assay’s detection limit , a set of sera samples prepared from blood having different concentrations of T . congolense TC13 parasites were tested in the homologous Nb474 sandwich ELISA . The lowest parasite concentration yielding a signal above the threshold OD450nm ( at ≥2-fold negative sample ) is 1 . 43x105 trypanosomes/ml ( S3 Fig ) . Finally , to alleviate the fear of possible loss of sensitivity of the test in case examination of sample is delayed as often observed with BCT , the ELISA was evaluated on T . congolense infected sera samples that were incubated at different temperatures for specified time periods . The results illustrate that , storing samples at 4°C or 22°C for up to 7 days does not affect the ELISA signal . However , raising storage temperature to 37°C for 3 , 5 and 7 days reduced the signal up to 2 , 10 and 44 folds , respectively ( S4 Fig ) . Given that the ELISA signal is unaffected by sample storage at 37°C within a relatively short time frame ( i . e . 1 day ) , this would enable the applicability of the Nb474-ELISA under realistic ‘field conditions’ where prompt examination of samples is usually delayed by an entire day due to electricity failures or transport to diagnostic laboratories . Next , the Nb474-ELISA was assessed for its capacity to differentiate animals with ongoing infections from those cured from trypanosomes after Berenil treatment in an experimental set-up . During the entire course of the experiment , the antigenemia and the parasitemia were monitored by the Nb474-ELISA and microscopy , respectively . During the early stages of the active infection the antigen profile correlates well with parasitemia ( Fig 3A ) . After the first parasitemia peak , when the parasitemia is drastically reduced due to host immune intervention , the antigen level remains high for 4 days before showing a gradual decline . In contrast , after drug-mediated parasite clearance , no antigen could be detected ( Fig 3B ) . Altogether , these results show that in experimental infections , the Nb474-ELISA can be used as a test-of-cure . The diagnostic performance of the Nb474-ELISA for the specific identification of T . congolense infections in cattle was examined by testing sera collected from cattle experimentally infected with T . congolense IL1180 ( n = 45 ) and T . congolense negative animals ( n = 17 ) . These results , of which the ELISA OD450nm scores of the individual serum samples are displayed in Fig 4A , were employed to determine the i ) Sensitivity ( defined as the proportion of the infected cattle that scored positive ) , ii ) Specificity ( defined as the proportion of uninfected cattle that scored negative ) , iii ) Positive Predictive Value ( PPV; defined as the proportion of the true positive in those samples that scored positive ) and iv ) Negative Predictive Value ( NPV; defined as the proportion of true negative in those sera that scored negative ) of the Nb474-ELISA . The results have been summarized in Table 1 . The test has a sensitivity and specificity of 87% [95% confidence interval ( CI ) , 73% to 95%] and 94% [95% CI , 71% to 100%] , respectively . The PPV is 98% [95% CI , 87% to 100%] , whereas the NPV is 73% ( 95% CI , 50% to 89% ) . The Nb474-ELISA was then tested on clinical field samples to assess the ability of the Nb474-based assay to specifically detect natural T . congolense infections . All clinical field samples were investigated by other techniques ( Giemsa-stained smear , BCT , PCR ) to confirm infection with T . congolense and/or any other endemic tropical parasites . When applying the Nb474-ELISA , the OD450nm observed on T . congolense positive samples is higher than on any other samples tested ( Fig 4B ) , demonstrating that the ELISA can be used to specifically detect T . congolense infections in a natural setting . To identify the target of Nb474 , a localization study was performed whereby cultured T . congolense IL3000 BSF parasites were probed with Alexa Fluor labeled Nb474 . The fluorescent intracellular structures observed ( Fig 5A , panels 3 and 4 ) had the typical appearance of glycosomes [46 , 47] . To investigate the hypothesis that Nb474 targets a glycosomal component , a co-localization experiment was carried out using Alexa Fluor labeled Nb474 and an Anti-TbALD , which is a pan trypanosome marker for the glycosomes and recognizes a linear epitope . This was performed on both T . congolense BSF and PCF parasites . While a clear co-localization is observed for the BSF parasites , Nb474 does not label the PCF parasites ( Fig 5B ) . To verify whether Nb474 indeed binds an antigen that is unique to BSF parasites , the Nb474-ELISA was used to probe the presence of the target antigen in a soluble proteome preparation obtained from both forms of the parasite . As it can be seen from Fig 5C , the ELISA scores positive for the BSF parasites , while no signal is observed for the PCF parasites . Collectively , these experiments demonstrate that Nb474 targets a glycosome-associated soluble protein that is specific to T . congolense BSF parasites . Given that the Nb474-ELISA yields a strong positive score when applied to the T . congolense soluble proteome , the sera of infected mice , or a T . congolense secretome ( S5 Fig ) , it was reasoned that these samples could be used as the source material from which the target could be purified by Nb474-mediated immuno-capturing . To this end , two distinct approaches were employed . In the first immuno-capturing set-up , Nb474H was adsorbed on 96-well ELISA plate to capture the antigen from the soluble proteome and infected serum . In the second experimental setting , Nb474H was immobilized on nickel beads via its His6-tag to capture the antigen from the secretome in an attempt to retrieve the antigen-Nb-bead complex with a Pickpen . The results of both experiments are shown in Fig 6A and 6B , respectively . In both cases , the analysis of the eluted samples via SDS-PAGE reveals a distinct , consistent protein band with an approximate molecular mass of 40 kDa ( Fig 6A , lanes 3 and 4; Fig 6B , lane 4 ) . The three bands were analyzed by mass spectrometry to determine the identity of the target . The mass spectrometry results pointed toward TcoALD ( UniProtKB accession no . G0UWE7 ) as the target antigen ( S6 Fig ) . To support the MS findings , a heterologous sandwich ELISA was carried out with Nb474H and Anti-TbALD as capturing and detecting molecules , respectively . A strong , significant OD450nm signal is observed in the wells coated with Nb474H followed by the subsequent addition of T . congolense TC13 soluble proteome and the Anti-TbALD compared to the control wells ( Fig 6B ) . Because it was unanticipated that Nb474 could bind TcoALD with such a high level of specificity ( given that aldolase is a highly conserved enzyme ) , E . coli lysate containing a recombinant version of the enzyme was probed with the Nb474-ELISA . In this validation experiment , lysates prepared from E . coli producing recombinant TbALD and Leishmania mexicana aldolase ( LmALD ) were included as control samples . While from Fig 7A it is evident that all three recombinant proteins are successfully produced in E . coli , Fig 7B demonstrates that the anti-TbALD MAb effectively recognizes TcoALD and TbALD but not LmALD . The latter can be explained by variation of the linear epitope in LmALD . In accordance with all of the previous experiments , both the homologous ( Nb474H-Nb474B; Fig 7C ) and heterologous ( Nb474H-Anti-TbALD MAb; Fig 7D ) sandwich ELISAs clearly show that Nb474 binds to TcoALD with high specificity .
The continuous monitoring of infectious diseases is crucial for disease control and treatment . Improved methods for the detection of low levels of pathogen infection and transmission are vital for quick and appropriate responses to prevent any outbreaks . Over the years , immuno-based POC tests have become common diagnostic platform because of their ability to rapidly and reliably detect infections under field conditions . These assays can be formatted in two ways . Detection of infection can be antibody- or antigen-based and both formats have their advantages and drawbacks . Antibody-based diagnosis assays are relatively sensitive as the immune response of the host is used as natural amplification mechanism for pathogen detection . However , a recurring issue with these assays is cross-reactivity , which reduces their specificity . For instance , unrelated diseases associated with polyclonal B-cell/antibody amplification can increase the number of false-positive readings [11] . Furthermore , in some cases host antibodies can circulate in the host long after a pathogen has been cleared thereby yielding a false image of the actual status of the infection . In contrast , antigen-based assays can differentiate between ongoing and past infections [48] . The hurdle in antigen-based detection is the selection of a suitable target antigen . Apart from not cross-reacting in diagnostic test for the detection of other endemic parasites , the target antigen should be present in detectable quantities during an active infection [23 , 49] . An additional requirement is that the capture and detection antibodies targeting the antigen should be able to out-compete the host anti-pathogen antibodies induced during infection [15 , 23] . In this paper , we have devised a Nb-based approach that could circumvent the above-mentioned issues hampering the development of antigen-based pathogen detection . Our rationale consists of vaccinating an alpaca with whole-pathogen soluble proteome and applying a stringent panning and screening strategy to deliver the most specific Nbs . While the use of Nb technology allows the selection of specific antigen-binding entities that recognize epitopes different from those bound by conventional antibodies [25] , the library generation and screening rationale are performed such as to allow the acquisition of Nbs that recognize unique antigens . Considering that the activity spectrum of diagnostic tests as well as vaccines based on highly immuno-dominant surface antigens of trypanosomes are narrowed by their variant nature , a non-biased antigen source ( T . congolense s . p . ) was used to ensure selection of potent Nb ( s ) recognizing target ( s ) that are conserved across the different strains of T . congolense . To test our strategy , we employed T . congolense as a case study . The application of our approach led to the identification of a potent T . congolense specific Nb , Nb474 . By making His-tagged and biotinylated versions of this Nb ( Nb474H and Nb474B , respectively ) , we were able to construct a homologous sandwich ELISA that was tested for its diagnostic potential in different scenarios . In an experimental mouse model , the assay is able to detect infections with different T . congolense strains , although there is a significant variation in the measured O . D . values . The observed variation is attributed to uneven parasite densities at the time point when samples were collected . Overall , the fact that all the strains tested show O . D . readings significantly above those of the negative samples or those taken from mice that were infected with trypanosomes other than T . congolense is an indication of the assay’s specificity for the diagnosis of T . congolense infections . The T . congolense infected mice display a typical parasitemia profile , which is characterized by a parasitemia peak around day 7 post infection . This first parasite wave rapidly drops due to the rapid destruction of trypanosomes by the host ( commonly known as ‘trypanolytic crisis’ ) , after which the parasitemia drops below the detection limit of parasitological techniques . Indeed , in accordance with the literature [48 , 50] , the blood parasite load was very scanty by microscopy after the parasite peak despite the presence of trypanosomes in the circulation . Interestingly , the antigenemia detected via the Nb474-ELISA remains significantly high , even after the first parasite wave . Upon clearance of the parasites from the system , the signal of the Nb474-ELISA is reduced to zero , suggesting that the occurrence of the detected antigen in the circulation is linked to the presence of parasites in the host and that the antigen does not persist after parasite clearance . Hence , in experimental mice infections , the Nb474-ELISA has the ability to distinguish active infections from a cured state and can thus be employed as a test-of-cure . By being able to differentiate sick animals from cured , the assay might in part qualify as the first-line of test for drug resistance . This feature of the ELISA , from the clinician’s perspective , is useful for monitoring the efficacy of trypanocidal drugs . The Nb474-ELISA was also able to diagnose T . congolense infections in experimentally and naturally infected cattle . The obtained results showed potential application of the diagnostic tool in the laboratory settings . Additionally , it seems that delayed sample investigation does not affect the assay’s sensitivity provided that the collected samples are kept under cold storage conditions . In the endemic areas , delayed sample investigation is often experienced in case of electricity failure or when there are no efficient means for sample delivery to distant laboratories . Together with its current level of performance , the option of delayed sample investigation would make this assay a suitable tool for monitoring T . congolense infections in remote settings and retrospective analysis of archived samples . We plan to further validate this assay on larger collections of field samples from different geographical locations where the disease is endemic . Finally , we identified the target antigen of the Nb474-ELISA as being TcoALD , which is a glycolytic enzyme found in the glycosomes [51] . Interestingly , TcoALD has been reported to be a part of the so-called T . congolense secretome [52] . However , the exact mechanism by which the enzyme reaches the host’s circulation is yet to be resolved . The findings in this paper demonstrate that TcoALD is a suitable biomarker for the identification of T . congolense infections , despite its high sequence identity with the glycosomal aldolase of other parasites ( S1 Table ) . This is not the first report presenting aldolase as a specific biomarker for pathogen detection . The enzyme variant encoded by Plasmodium vivax serves as a biomarker for the diagnosis of malaria [53] . Based on the structures of aldolases encoded by T . brucei and L . mexicana [54] , Toxoplasma gondii [55] and P . falciparum [56] , it is expected that TcoALD is a homo-tetramer in solution . This also explains why the Nb474 sandwich ELISA can be used in a homologous set-up . In the case of a monomeric target antigen , the same Nb could evidently not be used as a capturing and a detection reagent because the antigen would harbour only a single binding site for the Nb , which necessitates the development of a heterologous sandwich ELISA . However , the principle of a homologous sandwich ELISA is applicable for an oligomeric target antigen [57] . The observation that Nb474 can be used for antigen capturing as well as detection suggests the occurrence of at least two Nb474-binding sites on TcoALD . However , this remains speculative as the structural and biophysical determinants of the specific Nb474-TcoALD interaction remain yet to be investigated . We expect that these studies will help us improve the robustness of the Nb474-ELISA to specifically monitor T . congolense infections in the field . Moreover , the advantage of a selective test for trypanosomosis in the regions of sub-Saharan Africa where animals are infected with human as well as animal infective trypanosomes is that it discriminates between the two parasite groups , hence enabling assessment of the potential risk for human infection [49] . Further development of the ELISA will bring us a step closer towards obtaining an antigen-detection immuno-based POC test that can be used for rapid and reliable detection of pathogen infections . | The lack of diagnostic tests that are sensitive , affordable and user-friendly is the impediment to early detection and containment of infectious diseases . For these reasons , African trypanosomosis continues to pose serious threat to the communities that are unable to access laboratory services . Assays that are able to address above-mentioned challenges would greatly reduce the disease burden . Although few relatively sensitive agglutination assays for some forms of African trypanosomosis have been adapted to field conditions , the tests are not reliable indicators of active infections given the fact that they only detect antibodies . The barrier in the development of alternative tests capable of revealing ongoing infections through antigen detection is the lack of potent monoclonal antibodies that can compete with infection-induced host antibodies for the circulating parasite antigens . Using T . congolense as a model system , we demonstrate that Nanobodies ( Nbs ) targeting the parasite glycosomal aldolase can detect active infections . The strategy described addresses the technical shortcoming of conventional monoclonal antibody-based assay development by adopting an unbiased proteome screening approach combined with a phage panning strategy that is adapted to avoid interference of the infection-induced host antibody response . Hence , this study shows prospect for future development of Nb-based tests for other infectious diseases . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"livestock",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"ruminants",
"immunology",
"vertebrates",
"parasitic",
"diseases",
"animals",
"mammals",
"molecular",
"biology",
"t... | 2016 | An Anti-proteome Nanobody Library Approach Yields a Specific Immunoassay for Trypanosoma congolense Diagnosis Targeting Glycosomal Aldolase |
Bacterial pathogens of plant and animals share a homologous group of virulence factors , referred to as the YopJ effector family , which are translocated by the type III secretion ( T3S ) system into host cells during infection . Recent work indicates that some of these effectors encode acetyltransferases that suppress host immunity . The YopJ-like protein AvrBsT is known to activate effector-triggered immunity ( ETI ) in Arabidopsis thaliana Pi-0 plants; however , the nature of its enzymatic activity and host target ( s ) has remained elusive . Here we report that AvrBsT possesses acetyltransferase activity and acetylates ACIP1 ( for ACETYLATED INTERACTING PROTEIN1 ) , an unknown protein from Arabidopsis . Genetic studies revealed that Arabidopsis ACIP family members are required for both pathogen-associated molecular pattern ( PAMP ) -triggered immunity and AvrBsT-triggered ETI during Pseudomonas syringae pathovar tomato DC3000 ( Pst DC3000 ) infection . Microscopy studies revealed that ACIP1 is associated with punctae on the cell cortex and some of these punctae co-localize with microtubules . These structures were dramatically altered during infection . Pst DC3000 or Pst DC3000 AvrRpt2 infection triggered the formation of numerous , small ACIP1 punctae and rods . By contrast , Pst DC3000 AvrBsT infection primarily triggered the formation of large GFP-ACIP1 aggregates , in an acetyltransferase-dependent manner . Our data reveal that members of the ACIP family are new components of the defense machinery required for anti-bacterial immunity . They also suggest that AvrBsT-dependent acetylation in planta alters ACIP1's defense function , which is linked to the activation of ETI .
It is well established that bacterial pathogens utilize type III secretion ( T3S ) systems to translocate virulence factors ( referred to as T3S effectors ) into eukaryotic hosts to modulate immune signaling during infection [1] . The T3S effector proteome reflects the coevolution of specific host-pathogen interactions as well as microbe-microbe interactions within a given environment . Few T3S effector homologs are conserved among bacterial pathogens that colonize plant or animals hosts . One exception is the YopJ effector family , which is shared by a number of bacterial species in different genera ( e . g . Yersinia , Salmonella , Vibrio , Pseudomonas , Xanthomonas , and Sinorhizobium ) [2] . The YopJ effector family is named after the archetypal protein YopJ , first identified in Yersinia pseudotuberculosis [3] . These effectors belong to the C55 peptidase family because they share putative structural folds characteristic of cysteine proteases and contain the conserved catalytic triad – His , Glu and Cys [4] . Mutation of this catalytic triad destroyed effector-triggered phenotypes in host cells [5] , providing the first clue that enzyme activity is critical for the virulence of the YopJ effector family . Biochemical studies revealed however that YopJ has potent acetyltransferase activity [6] . In subsequent work , several effectors from this family were shown to have acetyltransferase activity important for host-pathogen interactions , including VopA from Vibrio parahemeolyticus [7] , AvrA from Salmonella typhimurium [8] , PopP2 from Ralstonia solanacearum [9] , and HopZ1a from Pseudomonas syringae [10] . These data indicate that a predominant virulence activity for the YopJ effector family is the post-translational acetylation of host proteins . Resistance to YopJ-like effectors ( i . e . AvrBsT , AvrRxv , AvrXv4 , HopZ1a , and PopP2 ) has been reported in several plant hosts [11]–[13]; however , only two disease resistance ( R ) proteins have been characterized to date [14] , [15] . Arabidopsis RRS1-R ( for RESISTANCE TO RALSTONIA SOLANACEARUM1 ) is a Toll-IL-1-receptor-nucleotide binding site-leucine rich repeat-WRKY motif ( TIR-NBS-LRR-WRKY ) -type R protein that recognizes the PopP2 effector from Ralstonia solanacearum [14] . RRS1-R directly interacts with PopP2 in the plant nucleus [16] . Arabidopsis ZAR1 ( for HOPZ ACTIVATED RESISTANCE1 ) is a coiled-coil ( CC ) -NBS-LRR-type disease R protein that recognizes the HopZ1a effector from Pseudomonas syringae and activates immune signaling that is distinct from most R protein pathways and independent of salicylic acid [15] . Neither RRS1-R nor ZAR1 were reported to be acetylated by the corresponding acetyltransferase [9] , [10] suggesting that acetylation of other plant targets is required for recognition and/or initiation of defense signaling by these R proteins . Interestingly , a recent study revealed that HopZ1a acetylates the Arabidopsis ZED1 ( for HOPZ-ETI DEFICIENT1 ) , a pseudokinase that is required for ZAR1-mediated immunity [17] . ZED1 is proposed to act as a decoy in a ZAR1 defense complex . Notably in mammals , YopJ acetylation suppresses innate immune signaling by exclusively targeting kinases in mitogen-activated protein kinase ( MAPK ) and/or NF–κB pathways . For example , YopJ catalyzes the O-acetylation of Ser or Thr residues in the activation loop of MAPKK6 [6] , MEK2 [18] , inhibitor of kappa B kinase [18] , and MAP3K transforming growth factor β-activated kinase 1 ( TAK1 ) [19] . Similarly in flies , AvrA inhibits c-Jun N-terminal kinase signaling by O-acetylation of the Thr residue in the activation loop of the MAPKK JNK-K [8] . In plants , a direct link between YopJ-like effector acetylation and suppression of disease resistance has not been made . HopZ1a was reported to acetylate tubulin in vitro , suggesting that the plant cytoskeleton may be disrupted during infection [10] . Consistent with this hypothesis , P . syringae pathovar tomato strain DC3000 ( Pst DC3000 ) infection reduced microtubule density in a HopZ1a catalytic-dependent manner [10] . Interestingly , the mammalian tubulin acetyltransferase TAT1 acetylates Lys40 in α-tubulin ( Nε-acetylation ) [20] , [21] and this modification is commonly found in less dynamic microtubules . The type of tubulin acetylation mediated by HopZ1a in planta has not yet been reported . In previous work , we exploited the use of the Pseudomonas-Arabidopsis pathosystem to elucidate the biochemical function of the AvrBsT effector from Xanthomonas euvesicatoria . AvrBsT was engineered to be delivered into plant cells by Pst DC3000's T3S system [22] because Arabidopsis is not a host for X . euvesicatoria . Two Arabidopsis ecotypes were identified that differentially respond to Pst DC3000 AvrBsT infection . The Col-0 ecotype is susceptible to Pst DC3000 AvrBsT infection whereas the Pi-0 ecotype is resistant . Pi-0 resistance is due to a recessive , loss of function mutation in SOBER1 ( for SUPPRESSOR OF AVRBST-ELICITED RESISTANCE1 ) . SOBER1 encodes a α/β-hydrolase that negatively regulates the accumulation of phosphatidic acid ( PA ) triggered by AvrBsT activity during bacterial infection [23] . High PA levels in Pst DC3000 AvrBsT-infected Pi-0 leaves correlate with ETI-like defense responses [22] , [23] . These data suggest that AvrBsT interferes with lipid homeostasis during infection and that this interference induces strong immune responses in the absence of SOBER1 activity . Given that PA is a multifunctional stress signal [24] , we hypothesized that AvrBsT-triggered PA bursts may directly lead to the local activation of defense signaling . Moreover , we hypothesized that AvrBsT host targets may be linked to the generation or perception of lipid signals during AvrBsT-triggered immunity . To begin to test these hypotheses , we sought to identify AvrBsT interacting proteins from Arabidopsis and elucidate their function ( s ) in the Pi-0 sober1-1 background [22] . Importantly , the availability of putative host substrates also enabled us to determine if AvrBsT possesses acetyltransferase activity , as reported for other effectors in the YopJ family [6] , [9] , [10] . Here we report that AvrBsT has acetyltransferase activity . We provide evidence that AvrBsT-dependent trans-acetylation activity is required for the activation of ETI in Arabidopsis Pi-0 leaves and that AvrBsT trans-acetylates Arabidopsis ACIP1 ( for ACETYLATED INTERACTING PROTEIN1 ) . ACIP1 is an unknown protein that localizes to punctae on the cell cortex and some of these punctae co-localize with cortical microtubules . We provide evidence that ACIP1 is a new component of the defense machinery required for anti-bacterial immunity . These data support the model that AvrBsT-dependent acetylation in planta alters ACIP1's defense function , which is linked to the activation of ETI .
To identify potential AvrBsT-interacting proteins in Arabidopsis , we performed a yeast two-hybrid screen using the GAL4 DNA-binding domain ( BD ) fused to AvrBsT ( i . e . BD-AvrBsT ) and an Arabidopsis cDNA library fused to the GAL4 activation domain ( AD ) . We screened ∼7 million primary yeast transformants and isolated 11 independent clones with a candidate cDNA encoded by At3g09980 ( Figure 1A and S1A ) . Given that AvrBsT is predicted to encode an acetyltransferase , we named the At3g09980-encoded protein ACIP1 , for putative acetylated-interacting protein 1 ( Figure 1A ) . ACIP1 is predicted to encode a protein with 178 amino acids and molecular weight of ∼20 . 6 kDa . ACIP1's only distinguishing feature is that it is predicted to be a small , α-helical protein [25] that contains the widely conserved domain of unknown function , DUF662 [26] . It was first identified as a tubulin-binding protein [27] . ACIP1 belongs to a small Arabidopsis protein family containing six ACIP-like isoforms ( ACIP-L1 to ACIP-L6 , Figure S1A ) . ACIP-L4 and its wheat ortholog TaSRG are required for salt tolerance [28] , although their biochemical function ( s ) are not known . ACIP1 shares 79% identity and 87% similarity with ACIP-L1 , the closest isoform . A tree for the Arabidopsis ACIP protein family is shown in Figure S1B . None of the ACIP-like isoforms were isolated in the primary AvrBsT interaction screen . A candidate yeast interaction screen comparing AvrBsT binding to ACIP1 or the six ACIP-like isoforms revealed that AvrBsT strongly interacts with ACIP1 but only weakly interacts with ACIP-L1 on selection media containing 1 mM 3-AT ( Figure S1C , D ) . In the presence of 5 mM 3-AT , AvrBsT only interacted with ACIP1 ( data not shown ) . Taken together , these data suggest that AvrBsT preferentially binds to ACIP1 in yeast . Next , we used GST pull-down assays to independently monitor the physical association of AvrBsT and ACIP1 in vitro . Recombinant GST and GST-AvrBsT were expressed in E . coli and then purified using glutathione sepharose . Purified GST-AvrBsT migrated as a doublet in protein gels , suggesting that proteolysis of the full-length polypeptide likely occurred during extraction and/or affinity purification . His-tagged ACIP1 was expressed in E . coli and soluble protein extracts were incubated with the GST or GST-AvrBsT in a standard GST pull-down assay . His6-ACIP1 was affinity purified by GST-AvrBsT but not GST ( Figure 1B ) . These findings are in agreement with the yeast two-hybrid data and provide additional evidence that AvrBsT interacts with Arabidopsis ACIP1 . We attempted to verify AvrBsT-ACIP1 physical interaction in planta; however , the assays were not successful . Transient or inducible expression of AvrBsT in Arabidopsis Pi-0 leaves or Nicotiana benthamiana leaves results in localized cell death . It was difficult to obtain reproducible , conclusive binding data under these cellular conditions . AvrBsT belongs to the YopJ family of T3S effector proteins , some of which have been shown to exhibit acetyltransferase activity [6] , [9] , [10] . To ascertain if AvrBsT acetylates ACIP1 , we first sought to determine if AvrBsT possesses auto-acetylation activity in vitro . Recombinant wild-type GST-AvrBsT , GST ( negative control ) and GST-HopZ1a ( positive control ) [10] were over-expressed in E . coli and then purified using glutathione sepharose . Purified proteins were incubated with 14C-acetyl-coenzyme A ( acetyl-CoA ) ±100 nM inositol hexakisphosphate ( IP6 ) for 30 minutes at room temperature and then separated by SDS-PAGE analysis followed by autoradiography . IP6 is a eukaryotic cofactor that stimulates the acetyltransferase activity of effectors in the YopJ family [9] , [10] , [29] . Auto-acetylation of GST-AvrBsT was detected in the presence of IP6 but not its absence ( Figure 2A and S2 ) . As expected , similar IP6-dependent activation and auto-acetylation of GST-HopZ1a was observed , and GST was not modified ( Figure 2A and S2 ) . Mutation of the conserved catalytic Cys residue ( C222 ) or His residue ( H154 ) to Ala inactivated AvrBsT-dependent acetyltransferase activity but did not affect protein expression levels ( Figure 2A ) . By contrast , mutation of the conserved Lys residue ( K282 ) ( Figure S3A ) to Arg , which has been shown to be an auto-acetylation site for some effectors in the YopJ family [9] , [10] , did not affect AvrBsT's acetylation state or protein accumulation ( Figure 2A ) . The auto-acetylation activity of GST-AvrBsT ( K282R ) was comparable to that of wild-type GST-AvrBsT in reactions with varying concentrations of enzyme ( Figure S3B ) . All GST-AvrBsT protein ( wild type and mutant ) analyzed migrated as a doublet and both of these species were auto-acetylated ( Figure 2 ) . Taken together , these data indicate that AvrBsT possesses auto-acetylation activity in vitro that is dependent on the conserved catalytic residues H154 and C222 , but this activity is independent of K282 . Next , we tested if AvrBsT directly acetylates ACIP1 using similar reaction conditions to those described above . Wild-type GST-AvrBsT activity resulted in auto-acetylation of the enzyme and trans-acetylation of GST-ACIP1 ( Figure 2B ) , whereas the catalytic core mutants GST-AvrBsT ( C222A ) or GST-AvrBsT ( H154A ) exhibited neither activities ( Figure 2B ) . Although the GST-AvrBsT ( K282R ) mutant possessed auto-acetylation activity , trans-acetylation of ACIP1 was not detected under the same reaction conditions ( Figure 2B ) . Importantly , mutation of C222 or K282 did not disrupt AvrBsT binding to ACIP1 in vitro ( Figure S4 ) . To gain insight to the specificity of acetyltransferases in the YopJ effector family , we determined if HopZ1a could acetylate AvrBsT's substrate ACIP1 . Conversely , we determined if AvrBsT could acetylate HopZ1a's substrate tubulin [10] . Incubation of GST-HopZ1a with GST-ACIP1 did not result in detectable acetylation of ACIP1 ( Figure 2C ) . Moreover , neither HopZ1a nor other members of the HopZ family could physically associate with ACIP1 in targeted yeast two-hybrid screens ( Figure S5A , B ) . Similarly , we could not detect AvrBsT-dependent acetylation of tubulin in vitro ( Figure S5C ) or direct physical interaction between AvrBsT and tubulin in yeast ( Figure S5A , B ) . These data suggest that AvrBsT and HopZ1a possess distinct substrate specificity . We assessed the biological activity of the AvrBsT ( K282R ) mutant in Arabidopsis Pi-0 leaves , given that mutation of the analogous Lys residue in PopP2 and HopZ1a inhibits effector auto-acetylation activity and effector-dependent phenotypes in planta [9] , [10] . Bacterial growth curve analyses showed that the K282R mutation attenuated the ability of AvrBsT to activate defense in Pi-0 , similar to that observed for the H154A mutation in the catalytic core ( Figure 3A ) . Furthermore , Pst DC3000 expressing AvrBsT ( K282R ) did not elicit HR in Pi-0 leaves ( Figure 3B ) despite stable protein expression ( Figure S3C ) . These data indicate that the K282R mutation affects AvrBsT's trans-acetylation activity in vitro ( Figure 2B ) and its defense eliciting activity in planta ( Figure 3 ) . Moreover , these data indicate that the auto-acetylation activity of AvrBsT ( K282R ) is not sufficient to activate ETI in Arabidopsis . Given that nothing was known about ACIP1 function , we first sought to elucidate its potential role in immunity . Previously we showed that the Pi-0 ecotype of Arabidopsis is resistant to Pst DC3000 expressing AvrBsT , whereas the Col-0 ecotype is susceptible [22] . Interestingly , ACIP1 mRNA abundance was significantly reduced at 3 and 6 HPI in Pi-0 ( Figure S6A ) and Col-0 ( data not shown ) leaves inoculated with a 2×108 cells/mL suspension of Pst DC3000 or Pst DC3000 AvrBsT compared to leaves inoculated with mock solution of 1 mM MgCl2 . By contrast , endogenous ACIP1 protein levels appeared to remain constant ( Figure S6B ) . These data suggest that ACIP1 may be transcriptionally or post-transcriptionally regulated during pathogen attack and potentially linked to PTI and/or ETI . To explore this further , we first analyzed the growth of virulent Pst DC3000 in a homozygous Col-0 acip1 null mutant ( SALK_028810 ) to determine if ACIP1 is required to limit pathogen growth . Pst DC3000 grew equally well in wild-type Col-0 and acip1 mutant leaves ( data not shown ) . Similar results were observed when the acip1 null allele was crossed into the Pi-0 background ( data not shown ) . We speculated that the lack of a bacterial growth phenotype in the Col-0 acip1 and Pi-0 acip1 mutants may be due to genetic redundancy since ACIP1 belongs to a small gene family in Arabidopsis ( Figure S1A , B ) . Since the nucleotide sequences between ACIP1 and ACIP-like genes are highly conserved ( Figure S7A ) , we engineered RNAi lines to target multiple ACIP family members in attempt to uncover an immune phenotype linked to this gene family . Notably , we silenced the ACIP gene family in the Pi-0 background to be able to monitor both PTI and ETI , considering that AvrBsT induces ETI in the Pi-0 ecotype but not the Col-0 ecotype [22] . A 365-bp hairpin ACIP binary construct ( hp-ACIP ) was designed using the ACIP1 gene , which included the most conserved region shared by the entire gene family ( Figure S7A , B ) , and then it was transformed into Pi-0 plants . Five independent transgenic RNAi lines were characterized . The hp-ACIP construct significantly reduced the mRNA levels for 4 of the 7 family members ( i . e . ACIP1 , ACIP-L1 , ACIP-L2 , and ACIP-L3 ) in two T2 ACIP RNAi lines ( i . e . lines 1 and 29; Figure S7C ) . Of these 4 genes , ACIP1 mRNA was the most abundant transcript in 4-week old Pi-0 leaves ( Figure S7D ) , suggesting that it may be the major isoform expressed in leaves . To monitor ACIP1 protein expression in leaves , we generated rabbit polyclonal antisera using recombinant ACIP1-His6 protein purified from E . coli . The resulting antisera recognized multiple , recombinant purified ACIP isoforms with distinct molecular weights by immunoblot analysis ( data not shown ) . However in wild-type Pi-0 leaf extracts , the antisera only detected a single 20 kDa protein band ( Figure 4A , inset ) . Three of the isoforms have predicted molecular weights in this range: ACIP1 = 20 . 5 kDa , ACIP-L1 = 20 . 2 kDa , and ACIP-L3 = 20 . 9 kDa . The 20 kDa protein band was not detected in the two ACIP RNAi lines ( Figure 4A , inset ) suggesting that ACIP1 , ACIP-L1 and/or ACIP-L2 protein accumulation was significantly reduced . Bacterial growth curves were then performed using a 1×105 cells/mL suspension of Pst DC3000 expressing AvrBsT and the two Pi-0 ACIP RNAi transgenic lines to determine if ACIP expression is required for AvrBsT-triggered ETI . The phenotypes of the ACIP-silenced lines were compared with an unsilenced Pi-0 control plant ( Figure 4 ) . We found that the titer of Pst DC3000 AvrBsT was significantly higher in infected Pi-0 ACIP RNAi leaves compared to that in wild-type Pi-0 leaves ( Figure 4A ) . Notably , the Pi-0 ACIP RNAi leaves were also more susceptible to Pst DC3000 . These data suggested that the silenced ACIP isoforms might function in PTI as well as ETI . To confirm that AvrBsT-triggered ETI is impaired in the RNAi lines , we performed HR and electrolyte leakage assays in leaves challenged with a high titer ( 3×108 cells/mL ) of Pst DC3000 AvrBsT . ETI in the Pi-0 ACIP RNAi lines was delayed but not fully inhibited ( Figure 4B ) . In control Pi-0 leaves , AvrBsT-dependent HR was visible at 9 HPI in many leaves and by 10 HPI , 22/25 leaves exhibited HR . By contrast , HR was not observed in similarly inoculated RNAi leaves at 9 HPI; however , HR started to develop at 10 HPI in 14/25 leaves for line 1 and 12/25 leaves for line 29 . Consistent with these findings , electrolyte leakage was significantly reduced in the Pst DC3000 AvrBsT-inoculated Pi-0 ACIP RNAi leaves relative to the inoculated Pi-0 leaves at 10 HPI ( Figure 4C ) . These data suggest that multiple ACIP isoforms are required for AvrBsT-triggered ETI symptoms in Pi-0 . The Pi-0 ACIP RNAi lines were also examined for their ability to mount ETI in response to two other Pseudomonas effectors – AvrB and AvrRpt2 [30] , [31] . As observed for AvrBsT , HR symptom development was slower in the RNAi lines infected with a high titer of Pst DC3000 AvrB or Pst DC3000 AvrRpt2 ( data not shown ) . Subsequent bacterial growth curve analyses revealed that the Pi-0 ACIP RNAi lines were more susceptible to both Pst DC3000 AvrB and Pst DC3000 AvrRpt2 ( Figure S8 ) . These data suggest that the ACIP isoforms play a general role in ETI and are not specific to defense responses elicited by AvrBsT . Given that the RNAi lines were also more susceptible to Pst DC3000 , we next examined the potential role of the ACIP family in PTI . First , we analyzed the responsiveness of the Pi-0 ACIP RNAi lines to the PAMP elicitor flg22 ( Figure 5 ) . Perception of flg22 by the PRR FLS2 results in the production of reactive oxygen species ( ROS ) [32] , one the first measurable PTI responses , followed by changes in PTI gene induction [33] . Flg22-induced ROS production was significantly reduced in both Pi-0 ACIP RNAi lines ( Figure 5A ) . Similarly , flg22-induced mRNA accumulation for WRKY22 and WRKY29 , two genes encoding transcription factors that positively regulate PTI , was significantly reduced at 3 hr post-treatment in both Pi-0 ACIP RNAi lines ( Figure 5B ) . Consistently , the RNAi line 29 exhibited the least responsiveness to flg22 elicitation ( Figure 5A , B ) . We also examined the responsiveness of the Pi-0 ACIP RNAi lines to Pst DC3000 ΔhrcU , a Pseudomonas strain known to elicit PTI . Pst DC3000 ΔhrcU lacks a functional T3S apparatus [34] and does not suppress PTI because T3S effectors are not secreted . Leaves were infected with a 1×105 cells/mL suspension of bacteria and titers were determined at 4 DPI . Pi-0 ACIP RNAi leaves were significantly more susceptible to Pst DC3000 ΔhrcU ( Figure 5C ) . Consistent with these findings , accumulation of WRKY22 and WRKY29 mRNA was significantly reduced at 6 HPI in Pi-0 ACIP RNAi leaves compared to wild-type Pi-0 leaves inoculated with a high titer ( 2×108 cells/mL suspension ) of Pst DC3000 ΔhrcU ( Figure 5D ) . Taken together , these data suggest that a subset of the ACIP family ( i . e . ACIP1 , ACIP-L1 , ACIP-L2 , and ACIP-L3 ) collectively contribute to anti-bacterial immunity in Arabidopsis . To begin to address ACIP1's function , we examined ACIP1 protein localization in Arabidopsis seedlings and mature plants . We generated homozygous transgenic Pi-0 plants expressing a GFP-ACIP1 protein fusion under the control of the native ACIP1 promoter ( i . e . PACIP1::GFP-ACIP1 ) . Using confocal microscopy , we observed a low level of GFP-ACIP1 fluorescence in 4-day old etiolated seedlings and juvenile leaves . Little or no detectable fluorescence was observed in mature , senescing leaves . In hypocotyl epidermal cells , GFP-ACIP1 was predominantly found as punctae at the cell cortex . A portion of these punctae was aligned , forming transverse cable-like structures ( Figure 6A ) . ACIP1's subcellular localization pattern partially resembled that of cytoskeletal structures . Unlike TaSRG , the ACIP-L4 ortholog , GFP-ACIP1 was not observed in the plant nucleus , indicating that ACIP1 localization is distinct from this predicted transcription factor [28] . Next , we applied drugs to disrupt the cytoskeleton to determine if ACIP1 co-localizes with actin and/or microtubules . Treatment of the Pi-0 PACIP1::GFP-ACIP1 seedlings with oryzalin , a microtubule depolymerizing agent , disrupted the GFP-ACIP1 cable-like structures and caused the formation of numerous GFP-ACIP1 punctae throughout the cell ( Figure 6B ) . By contrast , the actin depolymerizing agent latrunculin B did not appear to significantly disrupt these cables ( Figure 6B ) . To show ACIP co-localization with microtubules , we transformed the Pi-0 PACIP1::GFP-ACIP1 lines with P35S::mCHERRY-TUA5 , a fluorescently tagged isoform of α-tubulin . A large portion of the GFP-ACIP1 punctae co-localized with mCHERRY-TUA5 microtubules ( Figure 6A ) . Some of the cortical GFP-ACIP1 punctae were not associated with microtubules . Inspection of the literature revealed that ACIP1 was identified in the Arabidopsis proteome that co-purified with tubulin by affinity chromatography [27] . We did not detect a direct interaction between ACIP1 and TUA5 in a targeted yeast two-hybrid assay ( Figure S5A , B ) . It is possible that ACIP1 association with tubulin might be indirect or via a weak electrostatic interaction . Or , ACIP1 might interact with another isoform of tubulin . Collectively , our findings indicate that GFP-ACIP1 signal forms punctae on the cell cortex and some of these punctae co-localize with the cortical microtubule network . We speculated that AvrBsT-binding to and acetylation of ACIP1 might interfere with ACIP1's stability and/or localization within plant leaves during pathogen infection . We did not detect by protein gel blot analysis any differences in endogenous ACIP1 protein abundance or mobility using extracts isolated from Pi-0 leaves infected with Pst DC3000 or Pst DC3000 AvrBsT ( Figure S6B ) . However , we did notice that GFP-ACIP1 localization in 4-week old Pi-0 PACIP1::GFP-ACIP1 leaves was dramatically altered in response to both Pst DC3000 and Pst DC3000 AvrBsT ( Figure 7 ) . Unlike the signal in young hypocotyls , GFP-ACIP1 fluorescence at the cortex of epidermal cells in 4-week old leaves inoculated with the 1 mM MgCl2 mock control was diffuse and faint . This signal was difficult to capture in the image projection and varied among plants . By contrast , GFP-ACIP1 punctae were observed at or near the cell periphery of these cells ( Figure 7A ) . Infection with Pst DC3000 for 6 h led to the formation of rod-shaped GFP-ACIP1 structures of various lengths ( Figure 7B ) , which were difficult to detect in the 1 mM MgCl2 mock control ( Figure 7A ) . The GFP-ACIP1 rods were also detected in response to Pst DC3000 ΔhrcU ( Figure 7C ) , indicating that these structures are associated with PTI in a T3S effector-independent manner . Strikingly , Pst DC3000 AvrBsT infection for 6 h led to the formation of large , bright GFP-ACIP1 aggregates and fewer rod-like structures ( Figure 7D ) . This localization pattern was dependent on AvrBsT's catalytic activity . Pst DC3000 AvrBsT ( H154A ) infection resulted in a GFP-ACIP1 signal ( Figure 7E ) similar to that induced by Pst DC3000 alone ( Figure 7B ) . Similarly , Pst DC3000 AvrBsT ( K282R ) infection led to the formation of GFP-ACIP1 rods ( Figure 7F ) , but not large aggregates . To determine if GFP-ACIP1 aggregates are generated specifically by AvrBsT and/or in response to PA production , we tested the phenotype of Pst DC3000 AvrRpt2 . AvrRpt2 elicits ETI in Pi-0 leaves [23] , which is dependent on PA production [23] , [35] . Infection with Pst DC3000 AvrRpt2 induced the formation of GFP-ACIP1 rods ( Figure 7G ) , similar to those formed in response to Pst DC3000 alone , Pst DC3000 AvrBsT ( H154A ) , and Pst DC3000 AvrBsT ( K282R ) ( Figure 7B , E , F ) . However , ACIP1 aggregates were not observed in response to Pst DC3000 AvrRpt2 . Pi-0 PACIP1::GFP-ACIP1 leaves were inoculated with 50 µM PA and then imaged the leaves 1 . 5 hr later . Exogenous PA triggered the formation of several GFP-ACIP1 rods but only a few punctae ( Figure S8C ) , whereas the mock control containing 0 . 2% DMSO did not ( Figure S8D ) . These data suggest that PA exposure is sufficient to promote the formation of ACIP1 punctae and rods , but not the formation of ACIP1 aggregates . Moreover , they indicate that ACIP1 aggregate formation is a specific phenotype linked to AvrBsT acetyltransferase activity in planta .
Pathogen-dependent acetylation of host targets has emerged as a key virulence strategy to alter eukaryotic defense responses . The study of several YopJ and YopJ-like effectors in animals and flies indicates that O-acetylation of Ser/Thr residues or Nε-acetylation of Lys residues in the activation loop of kinases in innate immune pathways directly interferes with residue phosphorylation or ATP binding , respectively [6]–[8] , [18] . Both scenarios prevent the activation of kinases that are required to mediate innate immune signal transduction . In plants , the mechanism ( s ) by which YopJ-like effector acetylation of host substrates modulates immune signaling is less clear . Based on this study , we propose that the YopJ-like effector AvrBsT acetylates Arabidopsis ACIP1 . The role of ACIP1 in planta is not known; however , it is predicted to be a small α-helical protein [25] . ACIP1 emerged from an Arabidopsis screen looking for tubulin-binding proteins [27] , suggesting that it might be a microtubule-associated protein . Our microscopy studies of Arabidopsis Pi-0 lines expressing GFP-ACIP1 revealed that ACIP1 is localized to punctae on the cell cortex and some of these punctae co-localize with the cortical microtubule network ( Figure 6 ) . These data are consistent with ACIP1 being a part of the tubulin proteome [27] . Importantly , we discovered that GFP-ACIP1 organization and accumulation changed significantly during bacterial infection ( Figure 7 ) . Numerous small GFP-ACIP1 punctae and rod-like structures formed throughout the cell in response to Pst DC3000 infection . These structures also formed during Pst DC3000 ΔhrcU infection , indicating that changes in ACIP1 localization are coincident with PTI . Strikingly , Pst DC3000 AvrBsT infection , but not Pst DC3000 AvrRpt2 infection , dramatically altered GFP-ACIP1 localization . AvrBsT activity triggered the accumulation of large GFP-ACIP1 aggregates throughout the plant cell . The aggregates did not appear to be aligned like those observed in leaves infected with Pst DC3000 or uninfected hypocotyl epidermal cells ( Figure 6 ) . Interestingly , a mutation ( K282R ) that disrupted AvrBsT's ability to acetylate ACIP1 in vitro ( Figure 2B ) also prevented the formation of GFP-ACIP1 aggregates ( Figure 7 ) and activation of ETI during Pst DC3000 AvrBsT infection ( Figure 3 ) . Taken together , these data suggest the model that AvrBsT acetyltransferase activity in planta uniquely alters ACIP1's localization , which is linked to AvrBsT-dependent activation of ETI . The nature of the large GFP-ACIP1 aggregates and their function during pathogen infection remains to be determined . Given the requirement for ACIP1 for both PTI ( Figure 5 ) and the formation of ACIP1 punctae and rods during Pst DC3000 infection ( Figure 7 ) , we speculate that ACIP1 association with microtubules and/or the cell cortex is important for plant immunity . We also speculate that AvrBsT acetyltransferase activity either directly or indirectly alters ACIP1 association with microtubules . Association of ACIP1 with microtubules may play a direct role in microtubule organization , or it may be involved in microtubule-dependent processes such as vesicle and protein trafficking . Alternatively , ACIP1 may simply use microtubules to position itself and its interacting proteins at the cell cortex , where plant cells first encounter injected bacterial effectors . Our ACIP RNAi plants ( silenced for ACIP1 , ACIP-L1 , ACIP-L2 , and ACIP-L3 , Figure S7 ) did not show cell shape or cell growth phenotypes , which are caused by microtubule cytoskeleton defects , suggesting that four ACIP family members do not regulate microtubule cytoskeleton structure . Future functional studies will test if ACIP1 and/or other isoforms expressed in leaves play a role in suppressing bacterial growth by regulating microtubule-dependent trafficking or by regulating other processes at the plasma membrane or cell cortex . Notably , AvrBsT catalysis in Arabidopsis Pi-0 leaves leads to the accumulation of PA , a lipid signal associated with plant adaptation to biotic and abiotic stress [24] . Elevated PA levels in Pi-0 leaves inoculated with Pst DC3000 AvrBsT correlate with ETI [22] , [23] . ACIP1 is required for AvrBsT-triggered ETI ( Figure 4 ) ; however , the causal relationship between changes in PA production and ACIP1 localization in response to AvrBsT acetyltransferase activity is not clear . Furthermore , it is not known if PA is required to alter ACIP1 localization and/or function . Exogenous PA treatment ( Figure S8C ) or infection with Pst DC3000 AvrRpt2 ( Figure 7G ) , which triggers a PA burst during ETI [35] , induced the formation of small ACIP1 punctae and rods of various sizes , but large ACIP1 aggregates similar to those formed in response to AvrBsT-dependent catalysis ( Figure 7D ) were not observed . These data further highlight the specificity for AvrBsT in inducing ACIP1 aggregation during infection . They also suggest that a threshold concentration of PA or local production of PA relative to ACIP1 might be required to trigger the formation of large ACIP1 aggregates in planta . PA is known to play a critical role in the regulation of cytoskeletal dynamics [36] . Recent data suggests that PA alters the microtubule network by directly binding to cytoskeletal components , including tubulin and the microtubule bundling protein MAP65-1 [37] , [38] . Interestingly , elevated PA resulting from salt stress recruits Arabidopsis MAP65-1 to the membrane and enhances its ability to stabilize microtubules , which promotes cell survival [38] . How PA directly alters the microtubule network during ETI is not known . The link between PA , ACIP1 , and the microtubule network during pathogen infection established in this study suggests that PA might regulate ACIP1 complex formation and/or association with microtubules . Interestingly , the HopZ1a acetyltransferase was recently shown to disrupt plant cortical microtubule arrays and secretion during bacterial infection [10] . In this case , P . syringae HopZ1a infection led to reduced microtubule density , suggesting that HopZ1a acetylation in planta affects the stability or nucleation of microtubules . Acetylation of mammalian EB1 , a microtubule-associated protein , which promotes microtubule assembly , was recently shown to compromise EB1 binding to other microtubule plus-end tracking proteins [39] . HopZ1a binds and acetylates tubulin in vitro . Whether or not HopZ1a modifies tubulin and/or affects microtubule properties ( i . e . assembly , disassembly , and/or stability ) during infection remains to be determined . In terms of acetylation , our data suggests that AvrBsT trans-acetylation activity , not auto-acetylation activity , triggers ETI in Pi-0 leaves ( Figure 3 ) . Mutation of Lys 282 to Arg in AvrBsT , a conserved residue found in YopJ and YopJ-like effectors [9] , did not affect AvrBsT's auto-acetyltransferase activity in vitro ( Figure 2A ) , although it inhibited its ability to trans-acetylate ACIP1 ( Figure 2B ) . We speculate that K282 is required for enzyme-substrate interactions , although acetyl-CoA docking or direct acetylation [9] is also possible . Importantly , Pst DC3000 expressing the AvrBsT ( K282R ) mutant failed to trigger ACIP1 aggregates ( Figure 7 ) and elicit host resistance ( Figure 3 ) . These data suggest that acetylation is linked to changes in ACIP1 function and immunity . Whether or not acetylation of ACIP1 is directly linked to punctae formation , localization with microtubules , PA production and/or the activation of AvrBsT-triggered ETI awaits characterization of ACIP1's acetylation status in planta . Since acetylation can increase the electronegativity of proteins , it has the potential to disrupt ACIP1 interactions with the negatively charged microtubule lattice . Electrostatic interactions have been shown to play a significant role in microtubule binding of motor proteins and microtubule-associated proteins that often possess domains enriched in positively charged residues [40]–[42] . Future mapping of the ACIP1 microtubule-interaction domain in relation to residues acetylated by AvrBsT will allow us to test the functional significance of ACIP1 acetylation in planta . A growing number of plant targets have been identified for YopJ-like effectors , questioning the specificity of these enzymes as acetyltransferases versus binding partners in immune complexes . Our work indicates that there is selectivity between AvrBsT and HopZ1a in vitro . AvrBsT acetylates ACIP1 whereas HopZ1a acetylates tubulin . In addition to ACIP1 , AvrBsT has been recently shown to bind to arginine decarboxylase ( ADC1 ) , an enzyme proposed to mediate polyamine and γ-aminobutyric acid metabolism and impact cell death responses [43] , and SNF-1 related kinase ( SnRK1 ) , a putative regulator of sugar metabolism [44] . Post-translational acetylation of these plant proteins has not yet been reported . The fact that a number of metabolic enzymes are normally regulated by acetylation warrants further investigation [45] . Similarly , HopZ1a appears to have multiple plant targets . In addition to tubulin , HopZ1a was shown to acetylate Arabidopsis ZED1 , a pseudokinase required for HopZ1a-dependent ETI [17] , and Arabidopsis jasmonate ( JA ) ZIM-domain proteins required to repress JA signaling during PTI [46] . HopZ1a was also shown to bind and destabilize an enzyme involved in isoflavonoid biosynthesis , 2-hydroxyisoflavanone dehydratase ( HID1 ) , by an unknown mechanism [47] . The diverse nature of these targets suggests that HopZ1a is a promiscuous enzyme capable of altering defense signaling at multiple nodes . It is intriguing that HopZ2 , the closest YopJ-like homolog to AvrBsT [48] , was found to directly interact with Arabidopsis MLO2 in planta [49] . Arabidopsis mlo2-7 mutants are compromised for HopZ2-dependent virulence , further supporting the role of MLO2 as a negative regulator of immunity [49] . MLO2 is a plasma membrane protein of unknown function that interferes with vesicular trafficking mediated by the syntaxin PEN1 [50] , [51] . It is too early to tell if there is a common theme for YopJ-like targets in plant cells . However , the identification of ACIP1 , tubulin , and MLO2 as host targets suggests that some YopJ-like effectors might have undergone specialization to interfere with the trafficking function of the microtubule cytoskeleton in infected cells . Does AvrBsT target ACIP1 or an ACIP1 complex to suppress immunity ? This question has been difficult to answer because we have yet to detect an AvrBsT virulence phenotype in Arabidopsis during bacterial infection . This is not so surprising given the potential functional redundancy between AvrBsT and the suite of T3S effectors in Pst DC3000 . Overexpression of AvrBsT in transgenic Arabidopsis lines however was recently shown to enhance susceptibility to Pst DC3000 [52] . In solanaceous plants , AvrBsT is known to suppress PTI in tomato [53] and ETI in pepper [44] during Xanthomonas infection . The study of the ACIP1 ortholog in tomato may provide insight to AvrBsT virulence , by specifically addressing how AvrBsT acetyltransferase activity interferes with ACIP1's function during PTI and/or ETI . In summary , the study of AvrBsT-triggered defense responses in Arabidopsis Pi-0 plants has led to the identification of ACIP1 , a member of a new protein family required for PTI and ETI . We demonstrate that the expression of four Arabidopsis ACIP isoforms ( ACIP1 , ACIP-L1 , ACIP-L2 , and ACIP-L3 ) is required for proper execution of PTI in response to Pst DC3000 ( Figure 5 ) and ETI in response to Pst DC3000 expressing AvrBsT , AvrB or AvrRpt2 ( Figure 4 , S8 ) . In addition , we show that AvrBsT is an acetyltransferase and provide evidence that acetyltransferase activity plays an important role in altering ACIP1 localization within the plant cell during infection and the activation of ETI . This study highlights an important link between ACIP1 and the microtubule network during plant defense .
Escherichia coli DH5 alpha and Agrobacterium tumefaciens strain GV3101 were grown on Luria agar medium at 37 and 28°C , respectively . Pseudomonas syringae pathovar tomato ( Pst ) DC3000 strains were grown on nutrient yeast glycerol agar ( NYGA ) [54] at 28°C . E . coli antibiotic selection was 100 µg/mL carbenicillin and/or 50 µg/mL kanamycin . A . tumefaciens antibiotic selection was 100 µg/mL rifampicin , 50 µg/mL kanamycin , and/or 30 µg/mL gentamicin . Pst antibiotic selection was 100 µg/mL rifampicin , and/or 50 µg/mL kanamycin . Arabidopsis thaliana Col-0 and Pi-0 ecotypes were grown in growth chambers ( 22°C , 60% RH , 125 µE/m2/s fluorescent illumination ) on an 8-h light/16-h dark cycle . Plants were transformed using the floral dip method [55] . Standard DNA cloning methods [56] , PCR , and Gateway technology ( Invitrogen ) were used for plasmid construction . All primer sequences are listed in Table S1 . For GST-AvrBsT , avrBsT ( wild type , H154A , C222A , and K282R ) was amplified by PCR , cloned into pJET1 . 2 , and then sub-cloned into pGEX-5X-3 using BamHI and XhoI . For Gateway constructions , amplified PCR products ( i . e . avrBsT , hopZ1a , hopZ1b , hopZ2 , hopZ3 , ACIP1 , ACIP-like genes ( ACIP-L1 to ACIP-L6 ) , and TUA5 ) were cloned into pCR8 to create donor plasmids . The respective donor plasmids were recombined into: 1 ) pGADT7 to create AD-gene fusions and pGBKT7 or pXDGATcy86 to create BD-gene fusions for two-hybrid analysis; 2 ) pDEST15 for GST-fusions; and/or 3 ) pDEST17 for His6-fusions . For avrBsT mutagenesis , QuikChange Site –Directed Mutagenesis kit ( Stratagene ) was performed with pCR8 ( avrBsT ) and PfuUltra II Fusion HS DNA polymerase ( Agilent ) . Yeast strain AH109 carrying pXDGATcy86 ( avrBsT ) was transformed with pAD-GAL4-2 . 1 containing the Horwitz and MA cDNA library isolated from A . thaliana inflorescence meristem , floral meristem , and floral buds ( obtained from TAIR ) . Approximately 7 million transformants were screened and interaction with At3g09980 cDNA was confirmed . Yeast cells were resuspended in lysis buffer ( 1 . 85 M NaOH and 7% 2-mercaptoethanol ) and then proteins were precipitated in 10% trichloroacetic acid . Protein pellets were washed in 1 M Tris and then resuspended in 8 M urea sample buffer . Protein was extracted from plant cells as described [34] , separated by SDS-PAGE , transferred to nitrocellulose , and then detected by ECL or ECL plus ( GE Healthcare ) using anti-ACIP1 , anti-HA ( Covance ) , anti-Myc ( Covance ) , anti-His ( Qiagen ) , or anti-GST ( Santa Cruz ) sera and horseradish peroxidase conjugated secondary antibodies ( Bio-Rad ) . Membranes were stained with Ponceau S to control for loading . Recombinant His6-ACIP1 was expressed in E . coli BL21 tRNA cells and purified using Ni-NTA agarose under denaturing conditions following manufacturer's protocol ( Qiagen ) . Polyclonal antisera were raised in rabbits using the purified His-ACIP1 protein ( Covance ) . GST or GST-AvrBsT were expressed in E . coli BL21-CodonPlus ( DE3 ) cells ( Stratagene ) . Cells were lysed in buffer ( 1X PBS , pH 8 , 1% Triton X-100 , 0 . 1% 2-mercaptoethanol , and 1 mM phenylmethylsulfonyl fluoride ( Sigma Aldrich ) ) with a sonicator ( Branson ) . GST and GST-AvrBsT supernatants were incubated with 30 µL of pre-equilibrated Glutathione Sepharose 4B ( GE Healthcare ) for 1 h at 4°C with rotation . Sepharose beads were recovered by centrifugation and then washed with buffer for 5 min at 4°C with rotation . GST or GST-AvrBsT ( WT , C222A , or K282R ) bound beads were incubated with soluble E . coli lysates containing His6-ACIP1 for 2 h at 4°C with rotation . The beads were washed with buffer ( 50 mM TrisHCl , pH 7 . 5 , 150 mM NaCl , 10 mM MgCl2 , 0 . 1% Triton X-100 , and 0 . 1% 2-mercaptoethanol ) three times . Protein bound to the beads was separated by SDS-PAGE and analyzed by immunoblot analysis . Anti-GST and anti-His sera were used to detect GST-AvrBsT and His6-ACIP1 . Purified recombinant GST-tagged proteins ( 1 µg each ) were incubated with 100 nM inositol hexakisphosphate ( IP6 ) ( Santa Cruz ) , 0 . 4 µCi 14C-acetyl-CoA ( Perkin Elmer ) in 50 mM TrisHCl pH 8 and 1 mM DTT for 30 min at RT . Urea sample buffer was added to stop the reactions . Proteins assayed included: GST , GST-AvrBsT , GST-AvrBsT ( C222A ) , GST-AvrBsT ( H154A ) , GST-AvrBsT ( K282R ) , GST-HopZ1a , and GST-ACIP1 . Proteins were separated in a 10% SDS-PAGE gel , stained with Coomassie blue , transferred to blotting paper , dried , treated with EN3HANCE ( Perkin Elmer ) , and then exposed to film for 2–3 weeks at 80°C . A 365 bp region of ACIP1 was PCR amplified using primer set JG616/JG617 , and the product was cloned into pKANNIBAL to create pKANNIBAL ( hp-ACIP ) . The NotI fragment was then subcloned into pART27 [57] , creating pAR27 ( hp-ACIP ) . The resulting plasmid was then transformed into A . thaliana ecotypes Col-0 and Pi-0 . Transformants were analyzed by quantitative RT-PCR to measure ACIP isoform mRNA levels using primer sets listed in Table S1 . Fully expanded leaves of 4- to 5-week-old plants were used for bacterial inoculations . A suspension of bacterial cells ( Pst DC3000 , Pst DC3000 AvrBsT , Pst DC3000 AvrB , Pst DC3000 AvrRpt2 , or Pst DC3000 ΔhrcU; 3×108 cells/mL for HR and 1×105 cells/mL for growth curves ) was infiltrated into the extracellular space of fully expanded leaves using a 1-mL syringe . For HR , plants were incubated at RT under lights and phenotypes were recorded 9–12 HPI . For growth curves , plants were incubated at high humidity in a growth chamber for 4 d . Leaf tissue was collected at 0–4 DPI , ground in 1 mM MgCl2 , diluted and then plated on NYGA plates containing appropriate antibiotics and cycloheximide ( 50 µg/mL ) in triplicate to determine bacterial load . Four plants were used and the experiment was repeated at least three times . The average bacterial titer ± SD is reported . Three fully expanded leaves of 4- to 5-week-old plants ( n = 4 ) were inoculated with a 3×108 cells/mL suspension of Pst DC3000 ( vector ) or Pst DC3000 ( AvrBsT ) . Ten HPI , three leaf discs ( 10 mm diameter ) per plant were floated in 20 mL of water in petri dishes for 5 min and then transferred to a test tube containing 3 mL of water . Tubes were incubated for 1 h at RT with shaking . Conductivity of the solution was measured with an EC meter ( Spectrum Technologies ) before and after boiling for 30 min [58] . Percent electrolyte leakage was calculated as conductivity before boiling/conductivity after boiling ×100 . Assay was repeated at least three times . Three leaf discs ( 5 mm diameter ) from the youngest fully expanded leaves of a 4-week-old plant ( n = 9–18 ) were incubated in water in a 96-well plate ( one leaf disc per well ) for 24 h . To measure ROS , leaf discs were treated with ± flg22 ( 100 nM ) in 10 µg/mL horseradish peroxidase and 100 µM Luminol ( Sigma ) , and then luminescence was immediately measured with a 1420 Multilabel Counter ( PerkinElmer ) [32] . Relative luminescence units ( RLU ) are reported . Assay was repeated at least three times . Total RNA was isolated from uninfected or infected leaves using Trizol reagent ( Invitrogen ) according to manufacturer's instructions . For infection , leaves were inoculated with 1 mM MgCl2 or bacterial strains ( 2×108 cells/mL in 1 mM MgCl2 ) and then one leaf from three plants was harvested , pooled , and total RNA was extracted . 2 . 5 µg of RNA were used for cDNA synthesis . Quantitative RT-PCR was performed using the cDNA and gene-specific primers ( Table S1 ) . Each cDNA was amplified by real-time PCR using SensiFAST SYBR Kit ( Bioline ) and the MJ Opticon 2 instrument ( Bio-Rad ) . UBQ5 or ACTIN8 expression was used to normalize the expression value in each sample and relative expression values were determined against the average value of buffer or bacterially infected sample using the comparative Ct method ( 2−ΔΔCt ) . To monitor ACIP1 protein expression and localization , the promoter region ( 1 . 5-kb upstream of start ) was fused with GFP-ACIP1 in the backbone of pMDC43 to create pMDC43 ( PACIP1::GFP-ACIP1 ) . The resulting plasmid was transformed into A . thaliana Pi-0 plants . Transgenic PACIP1::GFP-ACIP1 lines were then transformed with P35S::mCHERRY-HA-TUA5 . This plasmid was construct by modifying pEarleygate 104 [59] . YFP was substituted with mCHERRY and Arabidopsis TUA5 genomic coding region was inserted after mCHERRY . Localization of GFP-ACIP1 and mCHERRY-TUA5 in 5-day old dark grown hypocotyls was determined using a Leica TCS SP5 confocal microscope ( Leica Microsystems ) with Leica LAS AF software and a Leica spinning disc confocal microscope with the Yokogawa CSUX-M1 confocal scanner . Seedlings were treated with 10 µM oryzalin in MeOH for 8 hr at RT or 1 µM latrunculin B in DMSO for 4 hr at RT and then imaged . For infection , Pi-0 PACIP1::GFP-ACIP1 leaves were inoculated with 1 mM MgCl2 or bacterial strains ( 3×108 cells/mL in 1 mM MgCl2 ) for 6 h . For exogenous PA treatment , Pi-0 PACIP1::GFP-ACIP1 leaves were inoculated with 50 µM PA in 0 . 2% DMSO or 0 . 2% DMSO . Images were analyzed using ImageJ [60] . Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: At3G09980 ( ACIP1 ) , At5G03660 ( ACIP-L1 ) , At2G36410 ( ACIP-L2 ) , At3G52920 ( ACIP-L3 ) , At2G27740 ( ACIP-L4 ) , At3G52900 ( ACIP-L5 ) and At2G36355 ( ACIP-L6 ) . | How host disease resistance pathways are activated in response to pathogens remains a fundamental question in host-pathogen interactions . In this work , we used the Pseudomonas-Arabidopsis pathosystem to study how the AvrBsT effector activates plant immune signaling . AvrBsT belongs to the YopJ effector family , a group of virulence proteins shared by bacterial pathogens of plants and animals . Bacteria inject these effectors into plant or animal host cells to promote pathogenesis . Recent biochemical studies show that several members of the YopJ family encode acetyltransferases that acetylate host proteins to suppress immune signaling . How the immune system specifically recognizes this family of effectors and/or monitors host acetylation is poorly understood . In this work , we provide biochemical evidence that AvrBsT is an acetyltransferase . We also report the identification and characterization of ACIP1 , an Arabidopsis protein of unknown function that is an AvrBsT substrate . We provide evidence that ACIP1 is required for plant immunity and its association with microtubules changes during infection . Moreover , our work suggests that AvrBsT acetyltransferase in planta leads to dramatic changes in ACIP1 localization , which coincides with the activation of strong defense responses . This study highlights an important link between ACIP1 and the microtubule network during anti-bacterial immunity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"cell",
"biology",
"plant",
"pathology",
"plant",
"science",
"biology"
] | 2014 | AvrBsT Acetylates Arabidopsis ACIP1, a Protein that Associates with Microtubules and Is Required for Immunity |
Influenza vaccination is the primary approach to prevent influenza annually . WHO/CDC recommendations prioritize vaccinations mainly on the basis of age and co-morbidities , but have never considered influenza infection history of individuals for vaccination targeting . We evaluated such influenza vaccination policies through small-world contact networks simulations . Further , to verify our findings we analyzed , independently , large-scale empirical data of influenza diagnosis from the two largest Health Maintenance Organizations in Israel , together covering more than 74% of the Israeli population . These longitudinal individual-level data include about nine million cases of influenza diagnosed over a decade . Through contact network epidemiology simulations , we found that individuals previously infected with influenza have a disproportionate probability of being highly connected within networks and transmitting to others . Therefore , we showed that prioritizing those previously infected for vaccination would be more effective than a random vaccination policy in reducing infection . The effectiveness of such a policy is robust over a range of epidemiological assumptions , including cross-reactivity between influenza strains conferring partial protection as high as 55% . Empirically , our analysis of the medical records confirms that in every age group , case definition for influenza , clinical diagnosis , and year tested , patients infected in the year prior had a substantially higher risk of becoming infected in the subsequent year . Accordingly , considering individual infection history in targeting and promoting influenza vaccination is predicted to be a highly effective supplement to the current policy . Our approach can also be generalized for other infectious disease , computer viruses , or ecological networks .
Influenza has a long history of causing substantial morbidity , mortality and economic losses annually [1]–[3] . In Israel , influenza is responsible for about 801 , 200 reported infections ( around 10% of the population ) , 4130 hospitalizations , 1140 deaths , and economic costs of 261 million dollars [3] , [4] , while in the US , influenza is responsible for 610 , 600 life years lost and economic loss of $87 . 1 billion annually [1] . Influenza vaccination is the primary approach to reduce the disease burden and is important not only for those vaccinated , but also to reduce transmission [2] . Recommendations by the World Health Organizations ( WHO ) [5] , the U . S . Center for Disease Control and Prevention ( CDC ) [2] , as well as the Israeli Ministry of Health have prioritized vaccination based on age , profession , and co-morbidities . However , these recommendations have not considered individual influenza infection history as an indication of future risk that can be used to supplement current policies . An individual's infection risk is governed by their contacts as manifested by their social interactions . A contact network model captures the patterns of interactions that expose individuals to potential transmission . In the context of contact network epidemiology , centrals , individuals characterized by higher connectivity than average , are more likely both to become infected and to transmit infection [6] , [7] . Thus , prioritizing the vaccination of centrals could be effective in curtailing influenza transmission by reducing the network connectivity . However , identifying centrals is challenging [8] , because the contact network is generally unknown . One study [6] offered a novel way to reach the centrals in a network by randomly choosing individuals and asking them to deliver a vaccination dose to one of their contacts , an approach known as the ‘acquaintance immunization strategy’ , suggests an indirect way to locate the centrals . Although this approach is an effective way to curtail transmission in both computer and population networks , it would be challenging to implement such a policy in the case of influenza vaccination . In the current study we offer a practical way to devise a vaccination policy using the simple logic of targeting previous influenza patients . We propose that even in the absence of information about network structure , centrals are at higher risk of influenza infection and can thus be identified as being disproportionately represented in the pool of individuals who were previously infected . Further , in addition to social interaction , a variety of factors , such as genetics , co-morbidities , demographics , and epidemiological characteristics [2] , may affect the risk and severity of infection , and remain relatively invariable over time . Regardless of whether individuals are predisposed to infection because of these factors or contact connectedness , they can be identified through previous infection . This approach is more straightforward than attempting to target individuals based on all possible risk factors , particularly as some risk factors may be unknown , difficult to identify or politically challenging to implement . Although the effectiveness of a policy that targets those previously infected initially seems to be counter-intuitive , since previously infected are likely to have partial protection against subsequent infection due to cross-reactivate antibodies [9] , we found that previously infected individuals are much more likely to be infected even when taking into account biologically realistic rates of cross-reactivity [10] . Our findings are based on contact network epidemiology simulations and confirmed by empirical clinical data provided by the two largest Health Maintenance Organizations ( HMOs ) in Israel , covering more than 74% of the Israeli population . Our study is the first to address the interplay among previous infection history , immunological cross-reactivity , and social behavior as the basis for an innovative yet feasible supplement to the current influenza vaccination policies .
The surveillance data were analyzed anonymously , and approved to be used by the Clalit health services sub-Helsinky institutional review board , signed and approved by Dr . Eitan Wertheim , protocol number 127/2012 . Our simulations were applied to an epidemiological contact network based on the Portland population [11] . The Portland contact network derives from detailed microscopic simulation-based modeling and integration techniques performed by the Network Dynamic and Simulation Science Laboratory ( NDSSL ) at Virginia Tech with the purpose of creating a contact network reflecting an urban population [11] . The network includes 1 , 575 , 861 nodes , each of which represent an individual and 19 , 681 , 820 edges , each of which represents a contact between individuals . To determine the robustness of our results , we validated our findings on three alternative small-world [12] scale-free networks: the Brightkite location-based network , the Gowalla location-based network , and the Barabási algorithm based network [13]–[15] . These networks vary in terms of the number of contacts , clustering coefficients , and the node-to-node distance [12] . The Brightkite location-based network is based on service providers where users share their locations by checking-in . The network was generated by the Stanford Network Analysis Project using their public Application Programming Interface , which consists of 58 , 228 nodes and 214 , 078 edges [16] . The Gowalla location-based contact network is a website where users share their locations by checking-in . The network was collected by the Stanford Network Analysis . It consists of 196 , 591 nodes and 950 , 327 edges [13] . We also created a network with 100 , 000 nodes and 400 , 000 edges according to the Barabási algorithm [14] . To evaluate centrality for each node in the networks , we calculated two common measures: number of contacts and K-shell decomposition values ( K-shell ) ( Figure S1 ) . Compared with the straightforward number of contacts measure , K-shell also take into account the global connectivity of the nodes to which a node is connected [17] , [18] . We used the Susceptible-Infectious-Recovered ( SIR ) compartmental model [19] to evaluate disease spread within the networks . According to each network configuration , an individual may infect only susceptible neighbors ( i . e . , nodes with whom they have edges ) . Given that not all individuals will be susceptible in the beginning of each season [10] , [20] , [21] , we parameterized transmissibility using the effective reproductive number , Re [21]–[23] , defined as the average number of secondary infections resulting from each infective person [19] ( Table 1 ) . Protection following infection may last longer than a year [10] due to cross-reactivity between years . We considered the entire possible range of cross-reactivity from 0 to 100% , where 0% corresponds to no immunological protection and 100% corresponds to full protection acquired from influenza infection in the prior year . Depending on vaccine efficacy ( Table 1 ) , we assume that a proportion of individuals who are vaccinated are protected for one season [24] . We ran over one million simulations drawing parameter values from distributions that span a biologically realistic range ( Table 1 ) , as well as different vaccination rates and efficacies ( Text S1 ) . To determine whether previously infected individuals are more likely to be centrals , we evaluated the decile of the centrality score for each node in the network ( based on the two measurements of centrality ) and the risk ratio of becoming infected for a range of reproductive ratios and for different levels of cross-reactivity compared with the risk of a random individual . In each iteration of the simulation , we ran two successive influenza seasons . In the first season , we randomly vaccinated 0–40% of the population . In the second season of each simulation , we considered three policies: a Random Vaccination Policy ( RVP ) , an acquaintance immunization policy ( AIP ) , which targets the acquaintances of a random node [6] , and a vaccination policy prioritizing those who were infected in the previous season , Previously Infected Policy ( PIP ) . For the PIP strategy , we assumed previously infected individuals are vaccinated first . If any vaccine doses remain after those prioritized have been vaccinated , the remaining doses are randomly distributed to the rest of the population . Our primary data were provided by Clalit Health Services and Maccabi Health Services and included demographic , socio-economic , and ethnic data [25] . Clalit is the largest HMO in Israel , with membership varying between 3 . 47–3 . 72 million , and constituting about 53% of the Israeli population during the 2003–2012 the study period . The Clalit dataset includes codes of the diagnosis recorded by the physicians in clinics according to the ICD9 protocol ( codes 487 for ‘influenza’ and 465 for ‘acute upper respiratory infections of multiple or unspecified sites’ ) as well as hospitalizations due to influenza or pneumonia ( ICD9 , 486 ‘pneumonia organism unspecified’ as well as codes 487 , 465 ) . Maccabi is the second largest HMO in Israel with 1 . 4–1 . 8 million members , constituting about 23% of the Israeli population , during the study period . Their records include full datasets of 12 years from September 1998 to April 2010 with 380 , 000 records . The data included influenza cases diagnosed in clinics according to the ICD10 protocol ( code J11 ) . Our case definition for influenza-like-illness ( ILI ) in both of the health maintenance organizations is detailed in Text S2 . As influenza infection rates depend on age , we stratified our data analysis by: 0–5 , 6–15 , 16–25 , 26–35 , 36–49 , 50+ . This division also facilitates evaluating the age-specific prioritization of the recommendations of the U . S . CDC as well as the Israeli Health Ministry which currently focus on ages 0–5 and individuals above age 50 . In addition , we considered the relative risk of age group 25–35 which , along with children , are disproportionately responsible for transmission [26] . Rather than a case of influenza , an ILI infection might indicate elevated risk for influenza , because transmission routes of many upper respiratory diseases are similar . Thus , an ILI might serve as a predictor of elevated risk for both ILI and actual influenza . For each age group , in both HMOs , we calculated the relative risk of infection for those previously diagnosed with ILI compared with others in the same age group that had not been diagnosed in prior season . In the Clalit dataset we stratified our ten seasons of data into eleven pairs of two consecutive seasons and calculated the risk of outpatient influenza in season i for influenza outpatient patients diagnosed in season i−1 , the risk of outpatient influenza in season i for patients hospitalized in season i−1 , the risk of becoming hospitalized with influenza in season i for outpatient patients diagnosed in season i−1 , and the risk of becoming hospitalized with influenza in season i for patients hospitalized in season i−1 . In the Maccabi dataset , we stratified our twelve seasons of data into eleven pairs of two consecutive seasons and calculated the risk of outpatient influenza in season i for influenza outpatient patients diagnosed in season i−1 ( Text S2 ) . Not all influenza patients seek medical treatment [1] , [27] , and some people might have higher tendency to seek medical treatment when infected with influenza than others , potentially leading to an overestimation relative risks . In addition , individuals are likely to visit the same physician when infected with influenza , and the latter might not diagnosis the infection as ILI . Thus , we compared evaluations of the policies under the conservative assumption that individuals can be divided into those that either seek medical treatment when infected , or not seek medical treatment . Under this conservative assumption , we calculated an adjusted relative risk of outpatient infection by removing in each age group members who were never an outpatient along the entire period tested .
Given that influenza attack rate ranges epidemiologically between 5–15% [2] , our case definition for influenza may be under-reported in the Maccabi dataset and over-reported in the Clalit dataset . Nevertheless , our analysis of over nine million medical records showed that patients diagnosed with ILI in the previous season have a substantially higher risk to be diagnosed with ILI in the succeeding year ( Table 2 ) . Excepting children below five who had a lower risk of being hospitalized , diagnosed with outpatient influenza in the season prior , our results were robust for both HMOs , for each age group and for almost every year tested . For example , in the Maccabi dataset within the age group of 35–49 , only 1 . 45% of the population was outpatient ILI . However , for individuals within this age group who were diagnosed in the previous year , the risk of becoming infected in the subsequent year was 11 . 35% . In the Clalit dataset , within the age group 35–49 , 9 . 36% were outpatient ILI . The risk was about four times higher in patients who were outpatient ILI , and two fold in patients previously hospitalized compared with patients who did not seek medical treatment of ILI in the year prior ( Table 2 ) . Even when we calculated an adjusted relative risk of outpatient influenza by considering only members who were diagnosed with outpatient influenza at least once in the study period ( i . e . 2003–2012 in Clalit dataset and 1998–2010 in Maccabi dataset ) , the relative risk for individuals previously outpatient was still higher than one in each age group ( Table 2 ) . This finding demonstrates that , in addition to age , an individual's infection history plays an important role in determining their subsequent risk of infection .
Our work shows that considering individual infection history in targeting and promoting influenza vaccination would be an effective supplement to the current policy which prioritizes individuals on the basis of age and co-morbidities . Our findings highlight the fundamental role that an individual's social behavior plays in disease transmission [28] , and reveals that in the interplay between cross-reactivity and individual risk factors the latter dominates in determining the overall effectiveness of targeting influenza vaccination to individuals infected in the prior season . Our simulations demonstrate that targeting individuals previously diagnosed with influenza can be effective even if cross-reactivity is as high as 55–80% . Empirical studies suggest that cross-reactivity for a specific type of influenza is typically below or within this range [10] . Additionally , there can be two or three sub-types of influenza circulating within a single season [29] with dominance shifting among sub-types between successive years . Surveillance systems monitor reporting rates rather than actual prevalence , and thus also include misdiagnoses . Nonetheless , ILI diagnoses , even if from a different etiology , might indicate elevated risk of future infection with influenza , because transmission routes and associated risk factors of many upper respiratory diseases are similar . Given that respiratory infections other than influenza will not elicit cross-reactive antibodies , these misdiagnosed individuals may be at even higher risk for future influenza infection than those who were previously infected with influenza . To evaluate the increased risk of those who were previously infected with influenza versus other respiratory infections , future research should stratify between clinical diagnosis and laboratory confirmation . While social interactions modeled in our contact network simulations are fundamental to influenza transmission , other factors including genetics , co-morbidities , and demography , contribute to determine risk for an individual . Similar to social tendencies , these other factors will also remain relatively invariable for an individual from year to year . Consequently , previous , current , and future infection risk can be even more effectively predicted by prior infection than what we conservatively estimated from social interaction alone . If PIP is implemented every influenza season , individuals with high connectivity might be targeted in the first year , and therefore will reduce their risk of infection in the subsequent year . However , as a result of this reduced risk , in the third year from initiating the PIP policy , they will be less likely to be infected and subsequently targeted . Thus , future studies could evaluate the marginal benefit of considering infection and vaccination history of individuals over several seasons relative to the prior season alone in determining vaccine targets . Targeting previously infected individuals is a relatively straightforward approach to implement in an HMO system with electronic medical records . For instances , individuals previously infected could be flagged within the electronic records for vaccination targeting by mailing pamphlets , telephone reminders or physician recommendations , practices shown to be effective in promoting influenza vaccination [2] . The suggested policy could reach several sub-populations , such as those based on socio-economic status and ethnicity , that correlated with vaccine uptake and infection rates [30] , but which have not been prioritized in the past due to ethical or political reasons . Furthermore , a high level of public adherence to this targeted strategy is likely to be achievable , given that individuals who were recently ill with influenza will probably be responsive to strategies that reduce their personal risk , as has been shown to be a primary motivator in vaccination decisions [23] , [31] , [32] . We demonstrated the potential benefit of targeting last season's patients for influenza , but such policy may also be applicable to other diseases including respiratory syncytial virus , pneumococcal infections and malaria , for which re-infection is common . Our approach may be generalized to networks outside the public health field , such as ecology and computer science . For example , our approach may determine which computers should be prioritized for antivirus software installations . In fact , a previous study on computer networks showed that computers that were attacked in one simulation run are most prone to attack in other simulation runs [33] . The authors even suggested little variation in the number of reinfections experienced by the same computer in different simulation studies , making our approach likely to be highly effective . In another example , our approach may also be helpful in ecological networks to identify and invest efforts to protect species with most essential to community stability [34] . In summary , we modeled the interplay among previous infection history , immunological cross-reactivity , and social behavior as the basis to generate an innovative influenza vaccination policy . Through contact network simulations we showed that individuals infected in the year prior have higher connectivity in the network , and subsequently increased risk of infection and transmission . Empirically , our analysis of the medical records confirms that in every age group , case definition for influenza , clinical diagnosis and year tested , patients infected in the year prior had a substantially higher risk of becoming infected in the subsequent year . Accordingly , the targeting of individuals infected in the prior year is predicted to be a highly effective supplement to the current policy . | WHO/CDC recommendations prioritize influenza vaccinations primarily on the basis of age co-morbidities , but have never considered targeting vaccination for individuals previously infected with influenza . An individual's infection risk is governed by his or her contacts as manifested by his or her social interactions . Thus , through contact network simulations that capture contact patterns , we show here that individuals previously infected with influenza have a disproportionate probability of being highly connected within networks and thus serve as the super-spreaders . Accordingly , targeting them is effective in curtailing transmission . In addition to social interaction , a variety of factors , including genetics , co-morbidities , demographics , and epidemiological characteristics , may affect the risk and severity of influenza infection . Regardless of whether individuals are predisposed to infection because of these factors , or social interactions , we show that they can be identified through previous infection . Empirically , our analysis of medical records of influenza diagnosed in both hospitals and clinics confirms that in every age group , case definition for influenza , clinical diagnosis , and year tested , patients infected in the year prior had a substantially higher risk of becoming infected in the subsequent year . Thus , considering individual infection history in targeting and promoting influenza vaccination is predicted to be a highly effective supplement to the current prioritizations as it focuses on people with greater risk to become infected and transmit . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biology",
"and",
"life",
"sciences",
"population",
"biology",
"medicine",
"and",
"health",
"sciences",
"epidemiology"
] | 2014 | An Innovative Influenza Vaccination Policy: Targeting Last Season's Patients |
Thousands of long intergenic non-coding RNAs ( lincRNAs ) are encoded by the mammalian genome . However , the function of most of these lincRNAs has not been identified in vivo . Here , we demonstrate a role for a novel lincRNA , linc-MYH , in adult fast-type myofiber specialization . Fast myosin heavy chain ( MYH ) genes and linc-MYH share a common enhancer , located in the fast MYH gene locus and regulated by Six1 homeoproteins . linc-MYH in nuclei of fast-type myofibers prevents slow-type and enhances fast-type gene expression . Functional fast-sarcomeric unit formation is achieved by the coordinate expression of fast MYHs and linc-MYH , under the control of a common Six-bound enhancer .
Adult skeletal muscles are composed of slow and fast myofiber subtypes which selectively express the genes required for their specific contraction activity and metabolic properties [1]–[4] . These properties are acquired at the end of fetal development and during the neonatal period , when mixed skeletal myofibers expressing a panel of embryonic , fast and slow genes develop a specific slow or fast phenotype . The formation of efficient fast sarcomeric units , and Ca++ cycling and excitation/contraction/relaxation coupling in fast-myofibers , is achieved through the coordinate control of fast Myhs and associated fast sarcomeric genes ( including Tnnt3 , Tnni2 , Tnnc2 , Atp2a1 and Pvalb ) [2] , [4] . Myofibers can be classified by their MYH expression profile: slow-type myofibers in mice express MYH7 ( also known as MYHCI , β or slow ) , and fast-myofibers express MYH2 ( MYHCIIA ) , MYH1 ( MYHCIIX ) or MYH4 ( MYHCIIB ) . Fast Myh genes found in developmental and adult stages ( Myh3 , Myh2 , Myh1 , Myh4 , Myh8 and Myh13 ) are organized as a cluster within a 300 kb region on mouse chromosome 11 [5] . The spatio-temporal expression of one specific fast Myh gene at the Myh locus is reminiscent of the organization and expression of Globin genes at the beta-globin locus [6] . However , we are yet to investigate potential enhancers or the Myh locus control region ( LCR ) that could be responsible for sequential and specific Myh gene expression in myofibers . The coordination of fast-type and slow-type gene expression in fast myofibers is not currently understood . Distinct intramyofibrillary calcium transients , evoked by slow tonic motor neuron firing , induce a cascade of downstream signaling involving Calcineurin and CamK . This results in the activation of selective transcription activators and repressors in slow myofibres . However , the signaling pathways operating in distinct Myh2 , Myh1 and Myh4 myofiber subtypes , which coordinate the activation of the other fast-type genes and the repression of slow-type genes , is less well understood [1] . Better knowledge of the mechanisms controlling muscle specialization and plasticity is important to enable the understanding and modulation of muscle adaptations in pathophysiological conditions . Six homeoproteins are major myogenic transcription factors which directly bind to DNA sequences ( called MEF3s ) to control myogenesis [7] , [8] and the genesis of fast-type myofibers during embryogenesis [9] , [10] . In adult skeletal muscle , Six1 accumulates at a higher level in the nuclei of adult fast myofibers than in of slow myofibers . Forced expression of Six1 and its Eya1 cofactor in slow myofibers causes adult slow-twitch oxidative fibers toward a fast-twitch glycolytic phenotype [11] . Animals with a Six1 KO present severe muscle hypoplasia and die at birth [12] . This prevents the in vivo analysis of the adult phenotype and the ability to investigate the direct or indirect involvement of Six1 in the spatio-temporal control of the expression of genes in the fast Myh cluster . The mammalian genome encodes thousands of long intergenic non-coding RNAs ( lincRNAs ) which have multiple functions [13] , [14] . Some accumulate in the cytoplam as miRNAs decoys [15] , [16] . Others accumulate in the nucleus and participate to gene regulation through chromatin remodeling and epigenetic modifications [14] , [17] , [18] . Here , they may act as cis [19] or trans [20] transcriptional activators , as transcriptional repressors [21] , [22] or through DNA-RNA triplex formation [23] , [24] . In this study we identify a new lincRNA , linc-MYH , and the mechanism of its control of adult muscle fast fiber-type specification in vivo . We demonstrate a three-element genetic partnership , where an enhancer element under the control of the myogenic homeoprotein Six1 functions as a regulatory hub to control fiber phenotype . In this partnership , the enhancer positively controls the expression of both the adjacent fast Myh gene cluster and linc-MYH , suppressing slow-type gene expression and facilitating fast fiber-type specialization .
Our previous studies suggested that Six1 could be directly involved in the control of the expression of fast Myh genes , since higher levels of this transcription factor accumulate in the nuclei of adult fast myofibers than in slow myofibers [11] . To investigate how Six1 could control the expression of fast Myh isoforms , we used computational analysis to locate MEF3 sites at the fast Myh locus ( see Materials and Methods ) . Six clustered MEF3 sites are conserved across human , rat and mouse genomes in an intergenic region located 50 kb upstream of the Myh2 gene ( Figures 1A and S1 ) and 4 kb upstream of a lincRNA ( 2310065F04Rik ) ; we refer to this lincRNA as linc-MYH ( Figures 1A and S2 ) . Six1 binding at these MEF3 sites was demonstrated in vivo by ChIP ( Chromatin Immunoprecipitation ) experiments with Six1 antibodies on adult fast gastrocnemius plantaris ( GP ) and tibialis anterior ( TA ) muscles ( Figure 1B ) but not on adult slow Soleus ( data not shown ) , and confirmed for five of these sites ( sites 1 , 2 , 3 , 4 and 6 ) by EMSA assays ( Figure S3A ) . We asked whether this Myh intergenic region could constitute an enhancer element , controlling the spatio-temporal expression of Myh genes in this locus . A 2 kb DNA fragment of this region , including the six identified MEF3 sites and 1 kb of DNA fragments upstream of the transcription start site of fast-type Myh2 , Myh1 and Myh4 genes , was isolated . The putative enhancer was ligated to each Myh promoter using luciferase pGL3 basic plasmids to generate pGL3-Enhancer-Myh2/1/4 constructs . To test the involvement of Six binding in enhancer activation of the Myh2 , Myh1 and Myh4 promoters , we mutated all six MEF3 sites present in the enhancer , and named these reporters pGL3-mutEnhancer-Myh2/1/4 . Luciferase activity was tested two weeks after the electroporation of reporter plasmids in adult TA muscles . The luciferase activity of pGL3-Enhancer-Myh2/1/4 was seven to twelve times higher for either of the promoters , than with pGL3-Myh2/1/4 . Enhancer activity was not observed in plasmids with mutated MEF3 sites associated with either of the Myh promoters ( Figure 1C ) . Enhancer activity was neither observed with the promoters of the slow Sln ( Figure S3B ) or Tnni1 genes , or with the promoter of the ubiquitous β-actin gene ( Figure S3C ) . A weak enhancer activity was observed with Myh4 promoter in primary embryonic fibroblasts , in which Six1 is expressed ( Figure S3D and data not shown ) . These data showed that high MYH enhancer activity was only reached in vivo and required specific interactions with MYH promoter elements . To determine in vivo interactions between the enhancer and each Myh gene , we performed chromatin conformation capture ( 3C ) assays of adult fast EDL ( Extensor digitorum longus ) myofibers . These experiments revealed that the enhancer interacts with the promoter of Myh2/1/4 genes in native chromatin of EDL myonuclei ( Figure 1D ) . The strongest interactions were observed with the Myh1 and Myh4 promoters , consistent with the expression profile of these two genes in EDL muscles . The data demonstrates that the identified conserved cis-element acts as an enhancer for the Myh locus and that MEF3 sites are essential for its enhancer activity in vivo . To further characterize the role of Six1 in the control of fast Myh gene expression , we bred Six1flox/flox mice with transgenic mice expressing CRE recombinase under the control of the human skeletal actin ( HSA ) promoter and obtained Six1flox/flox;HSA-CRE conditional knockout mice ( hereafter named cSix1 KO ) [25] , [26] . We analyzed the expression of fiber type specific genes in the back muscles of wild-type control mice and cSix1 KO mice at embryonic day 18 . 5 ( E18 . 5 ) and at several post-natal stages ( two weeks ( P2W ) , four weeks ( P4W ) and eight weeks ( P8W ) ) animals ( Figure 2 ) , as muscle fiber fast-subtype specialization takes place from the end of embryogenesis [9] . Six1 mRNA was not detectable in back muscles of cSix1 KO mice ( Figure 2 ) . The expression of fast-type genes ( Myh4 , Tnnt3 , Tnni2 , Tnnc2 and Pvalb ) increased during postnatal development in control mice but that of slow-type genes ( Myh7 , Tnnt1 , Tnni1 , Tnnc1 and Sln ) decreased . The linc-MYH RNA was detected after birth in muscle samples and its expression increased in line with that of Myh4 ( Figure 2 ) . The induction of fast-type genes and linc-MYH and the suppression of slow-type genes , were impaired in cSix1 KO mice . Expression of linc-MYH was reduced by three to five times in cSix1 KO mice during postnatal development ( Figure 2 ) . These results show that Six1 controls the induction of linc-MYH and fast-type genes during postnatal development , and is required for the downregulation of slow type genes . We next analyzed adult 12 week-old cSix1 KO mice to further characterize the role of Six1 in adult muscle . Six1 mRNA and protein were not detectable in GP enriched with fast-myofibers or soleus ( SOL ) muscle enriched with slow-myofibers ( Figure 3A and B ) , and fatigue resistance of TA muscle was 35% higher ( Figure 3C ) in the cSix1 KO mice . We used immunohistochemistry to analyse the composition of MYH7 , MYH2 and MYH4 in cSix1 mutant myofibers . Mutant TA muscles had a higher percentage of fibers containing MYH7 and MYH2 , but a lower percentage of fibers containing MYH4 ( Figures 3D and S4 ) . We found consistent results during qPCR analysis of Myh mRNA i . e . , higher levels of Myh7 and Myh2 mRNA and lower levels of Myh4 mRNA levels were observed in the fast TA muscles of cSix1 KO ( Figure 3E ) . Expression levels of other specific fast and slow-type genes were also tested . We found in mutant TA muscles a downregulation of fast-type genes ( Tnnt3 , Tnni2 , Tnnc2 and Pvalb ) and a five and to 25 fold increase in the levels of slow-type genes ( Myh7 , Tnnt1 , Tnni1 , Tnnc1 and Sln ) ( Figure 3E ) . Nevertheless expression of slow Myh7 is increased more than ten fold at the mRNA level in cSix1 mutant TA myofibers , while by immunohistochemistry the number of MYH7 positive myofibers is increased less than two fold . This showed that there is no major phenotype switch in cSix1 mutant TA myofibers . This observation could be explained either by the higher amount of Myh7 mRNA accumulating in MYH7 positive fibers or by a general increase of Myh7 mRNA in TA myofibers , mRNAs that would not be translated efficiently and leading to the absence of increase of MYH7 positive fibers . The expression of linc-MYH expression was lower in the adult TA of cSix1 KO mice , than in control mice ( Figure 3E ) . These results indicate that the Six1 homeoprotein can control the phenotype of fast skeletal myofibers in adult animals . We found that linc-MYH is expressed in fast-type skeletal muscles ( GP , TA and EDL ) , but not in SOL , brain , kidney , heart or fat tissues , an expression pattern which parallels that of the fast-fiber Myh4 ( Figure 4A ) . This suggested that linc-MYH is only expressed following robust nuclear accumulation of Six1 , as takes place in the nuclei of MYH4 myofibers [11] , and that the weaker nuclear accumulation of Six1 observed in SOL myonuclei does not allow efficient Six1 binding on the MYH enhancer and linc-MYH expression . We used luciferase reporter transfection assays ( as described above ) to test the requirement for Six binding on the MYH enhancer to activate linc-MYH expression . These transient transfection assays , performed in adult TA , show that the MYH enhancer activates linc-MYH expression in a Six-dependent manner , as measured two weeks after electroporation ( Figure 4B ) . lincRNAs can localize in cytoplasm [16] or as a single focus [19] or multiple foci [20] in nuclei . To analyze linc-MYH localization in skeletal muscle fiber , we performed fluorescent in situ hybridization ( FISH ) , using linc-MYH sense and antisense RNA on isolated myofibers from fast EDL . Intranuclear localization of linc-MYH was observed with the antisense linc-MYH RNA probe , with approximately 10 linc-MYH foci per nucleus ( n = 10 , Figure 4C ) , while the sense linc-MYH RNA probe gave no signal ( data not shown ) . We concluded from these experiments that linc-Myh RNA accumulates at specific sites in the nucleus of fast myofibers . Due to the number of linc-MYH foci observed in fast type nuclei , we hypothesized that linc-MYH could act in trans [17] to control gene expression in fast myofibers . To test this theory , we used electroporation to introduce a shRNA against linc-MYH ( shlinc-MYH ) in TA muscle and analyzed the transfected samples after fourteen days . This method yielded the efficient knockdown of linc-MYH , with a 90% reduction of its expression ( Figure 5A ) . To identify the consequences of linc-MYH knockdown and understand its mode of action , RNA samples from shlinc-MYH transfected adult TA were analyzed by Affymetrix microarrays ( Figure 5D and Table S1 ) , and validated by qPCR experiments ( Figure 5A ) . The expression of linc-MYH was significantly lower in the absence of Six1 , but Six1 expression was not affected by the absence of linc-MYH . Knockdown of linc-MYH led to robust gene expression modification; this knockdown strongly upregulated the expression of numerous slow genes ( such as Sln , Tnni1 , Tnnc1 and Tnnt1 ) and moderately downregulated the expression of several fast genes , ( including Myh4 , Tnnt3 , Tnni2 and Pvalb ) ( Figure 5A ) . We noted that the expression of Sln and other slow genes remained far lower in linc-MYH knockdown TA than in wild type Soleus . This showed that linc-MYH knock down in TA is not sufficient to achieve the full transcription efficiency of these slow genes as is observed in SOL , where calcineurin and several kinases are required to achieve their efficient transcription [27] . In addition , and contrary to what was observed previously in the muscles of cSix1 KO mice , slow Myh7 expression level did not change in linc-MYH knock down TA . We next tested whether the transcription rate of Sln and Tnnt1 slow genes was modified in linc-MYH knock down TA by measuring their pre-mRNA expression level . As can be seen in Figure 5B we observed an upregulation of pre-mRNA of these slow genes after linc-MYH knock down , which is proportional to their respective mRNA accumulation , demonstrating that their transcription is increased after linc-MYH knock down in TA , and showing that linc-Myh can act in trans to decrease their transcription . Since we also observed a moderate down-regulation of the expression of fast genes in linc-MYH knock down TA , among which Myh4 , we next tested whether the activity of the pGL3-Enhancer-Myh4 reporter could be modulated by the absence of linc-MYH . Two weeks after cotransfection of pGL3-Enhancer-Myh4 and shlinc-MYH in TA , we observed a statistically significant decrease of Luciferase activity , demonstrating the requirement of linc-MYH for efficient transcription of MYH4 ( Figure 5C ) . Altogether these experiments show that accumulation of linc-Myh in the nuclei of fast myofibers participates in the transcriptional silencing of slow genes and is required for the full activation of fast genes . We compared the networks of genes under the control of linc-MYH and of Six1 homeoprotein in adult muscles by the transcriptomic analysis of cSix1 and linc-MYH knockdown ( Figure 5D and Table S2 ) . We found that the six genes whose expression was the most increased in the linc-MYH knockdown were also significantly upregulated in cSix1 KO muscles ( Figure S5 ) . Besides slow muscle genes , two genes , Ankrd1 and Peg10 , were more severely upregulated in the linc-MYH knockdown line ( 10 and 8 times , respectively ) than in cSIX1 mutant myofibers ( by 2 . 8 and 1 . 5 times , respectively ) . These non slow-type genes could be exclusively repressed by linc-MYH in adult fast myofibers since there is a stronger downregulation of linc-MYH accumulation after its knockdown than in cSix1 mutant myofibers . Transcriptomic analysis of adult myofibers deprived of either Six1 or of linc-Myh identified a strong qualitative and quantitative correlation in the expression of specific genes between linc-MYH knockdown and cSix1 adult mutant myofibers . The expression of slow muscle genes was 3 to 10 fold higher in linc-MYH knockdown samples , and 5 to 25 fold higher in cSix1KO samples , than in the wild type . We further showed that linc-MYH lies downstream of Six1 in the Six myogenic pathway and helps to repress slow muscle genes in fast myofibers . The downregulation of all fast-type genes ( other than Myh4 ) , and the upregulation of slow-type genes , was weaker in the linc-MYH knockdown than in the Six1cKO line . Six1 may control several inhibitory pathways , including the linc-MYH pathway , to prevent slow-type genes expression in adult fast myofibers . During fetal development , at a stage where linc-MYH expression is not yet activated , Six1/4 increases the nuclear accumulation of the slow muscle repressors Sox6 and HDAC4 to repress slow muscle gene expression [9] , [28] , [29] . In accordance with this , the expression of the slow genes Myh7 , Sln , Tnni1 and Tnnt1 is upregulated in the muscle-specific cSox6 mutant [30] . This demonstrates that linc-MYH and Sox6 , lying both downstream of Six1 , directly participate in the downregulation of Sln , Tnni1 and Tnnt1 in fast myofibers . However , the repression of slow Myh7 in fast myofibers acts by a Six1-Sox6 dependent , but linc-MYH independent , repression mechanism . In this study , we observed that the levels of fast muscle gene expression decreased by 2–3 folds in the linc-MYH knockdown , with the highest decrease found for Myh4 expression . The expression of these genes decreased by a factor of 1 . 3 to 2 . 5 in cSix1KO , with the highest decrease found for Pvalb expression ( Figures 3E and 5A ) . The presence of Six4 and Six5 proteins in adult myofibers [11] , which have the same DNA binding specificity as Six1 , could compensate its absence in cSix1 KO animals and enable the activation of downstream fast muscle targets . In this case , linc-MYH expression could be preferentially dependent upon Six1 , rather than on Six4 or Six5 . Altogether , these experiments suggest that the accumulation of linc-MYH transcripts in the nuclei of fast myofibers facilitates the regulation of a network of genes that drive myofiber specialization via the same pathway as Six1 and downstream of this transcription factor . To test the possibility that linc-MYH could modulate the expression of specific muscle genes , we forced its expression in myogenic C2 cells and in primary myoblasts , where endogenous linc-MYH expression was faintly detectable even in myotubes four days after their differentiation ( data not shown ) . Transfection of a 13 kb genomic fragment encompassing the whole linc-MYH gene lead to efficient linc-MYH RNA accumulation in myotubes , but after this forced expression , we were unable to detect any modification in the expression of slow or fast type genes ( data not shown ) . These results suggest that specific cofactors of linc-MYH required for its appropriate functioning are lacking in cultured myotubes in culture , in agreement with the expression of linc-Myh only in adult fast type fibers . To circumvent the limitations of cultured cells , we turned to in vivo experiments in the Soleus in which linc-Myh is weakly expressed . Two weeks after linc-MYH gene transfection in the Soleus we observed that linc-MYH RNA accumulates up to approximately 80% of its expression level in TA ( Figure 5E ) . We observed a selective upregulation of Myh4 and Pvalb mRNAs which were increased to approximately one-third their expression level observed in TA , while mRNA for slow genes remained unchanged ( Figure 5E ) . To test whether the increase in Myh4 and Pvalb mRNA was due to an increased transcription of their genes , and potentially exclude a mechanism implicating mRNA stabilization , we measured pre-mRNA accumulation . As can be seen in Figure 5F , the transcription of these two fast genes is upregulated in the Soleus samples expressing linc-MYH proportionally to their mRNA accumulation , demonstrating that linc-Myh can work in trans and allow efficient activation of the transcription of specific fast genes . Absence of down regulation of the expression of slow genes in Soleus myofibers expressing linc-Myh suggests that , like in cultured myotubes , linc-Myh RNA needs specific protein-binding partners to achieve its function . Such specific protein-binding partners may be absent in Soleus as well as in cultured myotubes . Nuclear long non coding RNA are known to guide chromatin modifiers to specific gene loci , and by recruiting histone modifiers or DNA methyltransferase to modulate their transcription rate [31] . Potential linc-Myh protein partners expressed differentially in fast and slow adult myofibers may explain how linc-Myh efficiently represses the transcription of slow genes and activates the transcription of fast genes in fast myofibers , while in slow myofibers its forced expression is only able to activate the transcription of fast genes .
The commitment and maintenance of muscle fiber fast sub type specialization relies on the specific expression of one of the fast Myosin heavy chain gene present at the fast Myh locus , and of specific isoforms of sarcomeric genes [1] , [2] , [4] . Myosin heavy chains are the primary determinant of the efficiency of muscle contraction . In this manuscript , we identified a novel mechanism for the specialization of the fast-myofiber subtype . We show that the long intergenic non-coding RNA linc-MYH and fast MYH genes , both of which are essential for myofiber specification , share a common enhancer which is regulated by Six1 homeoproteins . The linc-MYH specifically accumulates in nuclei of adult fast myofibers . Its function , as revealed here by in vivo knockdown and transcriptome-wide analysis , is to prevent slow-type muscle gene transcription and increase fast-type muscle gene expression in fast-type myofibers . We found linc-MYH downregulates the transcription of genes associated with slow muscle contractile properties like the slow genes Tnn and Sln ( a known repressor of Serca1/Atp2a1 protein [32] , [33] involved in Ca++ reuptake by the sarcoplasmic reticulum ) . These genes , which belong to the muscle contractile machinery and are repressed in adult fast myofiber , are positively controlled by Six1 in myogenic C2 cells [34] , where linc-MYH expression is not detected . This suggests that their expression in adult fast myofiber may be restricted by an additional level of regulation involving the Six1-linc-MYH axis . As a result of our study we suggest that Six1 controls the acquisition of fast-type myofiber mechanical properties by binding to a single enhancer region of the fast Myh locus . It promotes the coordinated expression of fast Myhs and that of a strong repressor of genes controlling slow contractile properties . The modulation of Six activity ( depending on fiber-type ) facilitates changes in the expression levels of the fast genes Myh and Tnn; these changes are required for the formation of efficient sarcomeric units and the appropriate Ca++ cycling and excitation/contraction/relaxation coupling [1]–[4] . The Myh enhancer element therefore connects distinct regulatory hubs to achieve ultimate muscle fiber specialization . In this context , linc-MYH functions as an end-of-the-chain control element , conveying information on the state of fast Myh enhancer activity to repress slow-type specific genes and coordinates a finer level of regulation . This genomic organization at the fast Myh locus is reminiscent of the slow Myh7 locus where two microRNA miR-208b and miR-499 involved in fast myofiber program repression are co-regulated with Myh7 [35] . The precise molecular interactions between linc-Myh and higher order chromatin modifying complexes remains to be identified , to explain how linc-Myh coordinates the activation of target genes at specific sites in the nucleus , and the repression of others .
Animals were bred and handled as recommended by European Community guidelines . Experiments were performed in accordance with the guidelines of the French Veterinary Department . cSix1KO mice were obtained by breeding the Six1-LoxP mice [26] and transgenic mice expressing a CRE recombinase under the control of the human skeletal actin promoter ( HSA ) [25] . GP and TA muscles of 2 months old mice were minced with scissors just after sampling and fixed in 1% formaldehyde for 10 minutes . Formaldehyde was quenched by addition of 0 . 125 M glycine , and muscles were washed twice in PBS . The muscles were incubated on ice in lysis buffer ( 10 mM Tris-HCl pH 7 . 9 , 85 mM KCl , 0 . 5% NP40 , protease inhibitors ( cOmplete , Roche ) ) for 10 minutes and homogenized with a mortar and , subsequently with a Dounce homogenizer . The nuclei were obtained by centrifugation , incubated in SDS lysis buffer ( 50 mM Tris-HCl pH 8 , 10 mM EDTA , 1% SDS , protease inhibitors ) for 10 minutes , and sonicated in a Bioruptor apparatus ( Diagenode ) . The debris were removed by centrifugation . The sonicated DNA was incubated with 1 µg of Six1 antibodies ( HPA001893 , Sigma ) under agitation at +4°C overnight . 20 µl of Dynabeads protein G ( Invitrogen ) were added to the samples and incubated under rotation at +4°C for 1 hour . The beads were washed with low-salt buffer ( 2 mM EDTA , 20 mM Tris-HCl pH 8 , 150 mM NaCl , 1% TritonX-100 , 0 . 1% SDS ) , high salt buffer ( 2 mM EDTA , 20 mM Tris-HCl pH 8 , 0 . 5M NaCl , 1% TritonX-100 , 0 . 1% SDS ) , LiCl buffer ( 1 mM EDTA , 10 mM Tris-HCl pH 8 , 0 . 25M LiCl , 1% NP40 , 1% deoxycholate ) and TE buffer ( 1 mM EDTA , 10 mM Tris-HCl pH 8 ) . The DNA was eluted with elution buffer ( 1% SDS , 0 . 1 M NaHCO3 ) containing 0 . 1 mg/ml proteinase K ( Invitrogen ) at 62°C for 2 hours , and , proteinase K was inactivated by incubation at 95°C for 10 minutes . The DNA was finally purified with MinElute PCR purification kit ( Qiagen ) . The amount of specific amplified DNA is normalized by beta-Actin promoter amplification . The sequences of the oligonucleotides used in this study are as follows . Enh 1F , 5′-ATC TCC ACC TCC CTC CAA CT; Enh 1R , 5′-ACC CCC TAG CTT TGA CAG GT; Enh 2F , 5′-AAT CTG ACG ACA GGG TGA GC; Enh 2R , 5′-GGT CGC CTG ACC TGA TAG AG; AldoaF , 5′-CTC TCA AGG CAA ACC AAA GC; AldoaR , 5′-CCA GTG TCC CAG ACC TTC TC; ActbF , 5′-TGT TAC CAA CTG GGA CGA CA; ActbR , 5′-ACC TGG GTC ATC TTT TCA CG , NCF , 5′-ATC CTG CCC CAC TGT GTT AG; NCR , 5′-GCC AGC AAT TTG GTT TGA AT . 3C experiments were performed as described [36] with few modifications . Single myofibers were obtained from adult EDL muscles as previously described [26] , cross-linked in 2% formaldehyde , 10 mM Tris-HCl pH 7 . 9 , 85 mM KCl , 0 . 5% NP40 for 10 min at room temperature . Crosslinking reaction was quenched by 1 M glycine . Cross-linked myofibers were lysed for 10 minutes with lysis buffer ( 10 mM Tris-HCl pH 7 . 9 , 85 mM KCl , 0 . 5% NP40 ) on ice , and the nuclei were harvested . Nuclei were resuspended in appropriate restriction enzyme buffer , 0 . 3% SDS and incubated for 1 hour at 37°C with shaking . Triton X-100 was added to 2% , and samples were incubated for 1 hour at 37°C . Samples were digested with Hind III overnight at 37°C . DNA ligation was performed for 4 hours at 16°C and for 30 minutes at room temperature . Cross-links were reversed , and DNA was then purified by phenol extraction and ethanol precipitation . To correct for the PCR amplification efficiency of different primer sets , a BAC clone containing the mouse Myh locus ( RP23-61C14 ) was digested , ligated and used as control templates . Quantification of the data was performed by quantitative real-time PCR using the Lightcycler 480 probe master ( Roche Diagnostic ) . The sequences of the oligonucleotides used in this study are given in Table S3 . TA , back , soleus and GP muscles were collected from cSix1 KO and control mice . Total RNAs were extracted by Trizol Reagent ( Invitrogen ) according to manufacturer's instruction . RNAs were treated with DNase I ( Turbo DNA-free , Invitrogen ) and were reverse transcribed with Superscript III kit ( Invitrogen ) according to manufacture's instruction . Reverse transcription was performed with 1 µg of total RNA . Quantitative real time PCR ( Light Cycler 480 , Roche ) was performed using Light Cycler 480 SYBR Green I Master Kit ( Roche ) according to the manufacturer's protocols . PCR was performed for 40 cycles of 95°C for 15 seconds , 60°C for 15 seconds , and 72°C for 15 seconds . Genes expression level was normalized by the expression level of the housekeeping gene Actb . The sequences of the oligonucleotides used in this study are given in Table S4 . Pre-mRNA qPCR experiments to measure RNA transcription rate were performed in the same conditions . Reverse oligonucleotides were complementary to intronic sequences , while forward oligonucleotides were complementary to exonic sequences . Samples without reverse transcription were used as controls , and signal due to contaminating DNA was subtracted to the values obtained with cDNA . We noticed that genomic DNA contamination was very low ( less than hundred fold level of qPCR value observed with cDNA ) . Skeletal muscle function was evaluated by measuring in situ muscle contraction , as described previously [37] . 12 week-old male mice were anesthetized ( intraperitoneal injection of pentobarbital sodium , 50 mg/kg ) . Body temperature was maintained at 37°C using radiant heat . The distal tendon of the TA muscle was attached to an isometric transducer ( Harvard Bioscience ) using a silk ligature . The sciatic nerves were proximally crushed and distally stimulated by a bipolar silver electrode using supramaximal square wave pulses of 0 . 1 ms duration . Responses to tetanic stimulation ( pulse frequency 50–143 Hz ) were successively recorded . Maximal forces were determined at optimal length ( length at which maximal force was obtained during the tetanus ) . Fatigue resistance was then determined after a 5-minutes rest period . The muscle was continuously stimulated at 50 Hz for 2 minutes ( sub-maximal continuous tetanus ) , and the duration corresponding to a 20% decrease in force was recorded . Fluorescent-labeled antisense linc-MYH probes were synthesized according to manufacturer's instruction ( FISH Tag RNA kit , Invitrogen ) . FISH experiments were performed on isolated EDL myofibres and images acquired on a Leica SP2 confocal microscope . Five distinct shRNAs targeting mouse linc-MYH were designed , called shlincMYH , and inserted into the psiSTRIKE hMGFP system ( Promega ) . The efficiency of each shRNA was established by determination of linc-MYH transcript levels in TA muscles transfected by each shlincMYH . The shRNA against 5′-TTC TGC TCA CCA CCT ACA ATT-3′ sequence was selected for the knockdown experiment . For knock down experiments using shlincMYH , a plasmid coding for shLacZ was electroporated in the contra-lateral TA as a negative control . In vivo transfections were also carried out on 10-weeks old C57Bl6 mice . For each experimental conditions , three to five Tibialis anterior ( TA ) or Soleus ( Sol ) muscles belonging to different mice were used . Under isoflurane anesthesia , legs were shaved and muscles were pre-treated by injection of a sterile 0 . 9% NaCl solution containing 0 . 4 U of bovine hyaluronidase/µl two hours before plasmid injection . Ten µg of shRNA-expressing vector were introduced into TA muscles of 8 week-old mice by electroporation as previously described [11] . Two weeks following electroporation , TA myofibers expressing GFP were dissected under a Nikon SMZ1500 stereo microscope and frozen in liquid nitrogen before processing for Luciferase assays or RNA purification . TA , soleus and gastrocnemius muscles were embedded in cryomatrix and quickly frozen in isopentane cooled with liquid nitrogen . Cryostat sections ( 10 µm ) were fixed in 4% PFA , washed in PBS , permeabilized with 0 . 1% Triton X-100 and left for 1 hour in blocking solution ( 1× PBS , 1 . 5% goat serum , 0 . 1% Triton X-100 ) . Rabbit poly-clonal antibodies directed against Laminin ( Z0097 , Dako ) ( 1/100 dilution ) , and monoclonal antibodies against MYH7 ( NOQ7 . 5 . 4D , Sigma ) ( 1/1000 dilution ) , MYH2 ( SC-71 , Developmental Studies Hybridoma Bank ) ( 1/20 dilution ) and against MYH4 ( BF-F3 , Developmental Studies Hybridoma Bank ) ( 1/20 dilution ) were applied overnight at 4°C to the treated sections . The next day , after three washes with 1× PBS containing 0 . 05% Tween-20 , cryosections were incubated for 1 h with appropriate fluorescent secondary antibodies ( Alexa Fluor 488 goat anti-rabbit IgG 1/1000 dilution , Alexa Fluor 594 goat anti-mouse IgG 1/1000 dilution , Invitrogen ) . After three washes with 1× PBS containing 0 . 05% Tween 20 , samples were mounted in Vectashield mounting medium . After validation of RNA quality with the Bioanalyzer 2100 ( using Agilent RNA6000 nano chip kit ) , 50 ng of total RNA were reverse transcribed following the Ovation PicoSL WTA System ( Nugen ) . Briefly , the resulting double-strand cDNA was used for amplification based on SPIA technology . After purification according to Nugen protocol , 5 µg of single strand DNA was used for generation of Sens Target DNA using Ovation Exon Module kit ( Nugen ) . 2 . 5 µg of Sens Target DNA were fragmented and labelled with biotin using Encore Biotin Module kit ( Nugen ) . After control of fragmentation using Bioanalyzer 2100 , the cDNA was then hybridized to GeneChip Mouse Gene 1 . 0 ST ( Affymetrix ) at 45°C for 17 hours . After overnight hybridization , the ChIPs were washed using the fluidic station FS450 following specific protocols ( Affymetrix ) and scanned using the GCS3000 7G . The scanned images were then analyzed with Expression Console software ( Affymetrix ) to obtain raw data ( cel files ) and metrics for Quality Controls . The analysis of some of these metrics and the study of the distribution of raw data show no outlier experiment . RMA normalization was performed using R and normalized data was subjected to statistical tests . EMSA was carried out with Six1 full-length mouse cDNA cloned into the pCR3 vector ( Clontech ) as previously described [38] . Recombinant mouse Six1 protein was obtained with a T7 transcription/translation kit ( Promega ) . The oligonucleotide containing double-stranded myogenin MEF3 site was incubated with recombinant proteins . Competition experiments were performed in the presence of a ten-fold and hundred-fold molar excess of unlabeled identified Myh enhancer MEF3 sites ( Enh1 to Enh6 ) or mutated Myh MEF3 sites ( mtEnh1 to mtEnh6 ) , or Myogenin promoter NFI or MEF3 sites . The sequences of the oligonucleotides used are as follows , the MEF3 consensus sequence is underlined; Enh1F 5′-CTC TTG GGT AAC TGG AGC CCC TC-3′ . Enh1R 5′-GAG GGG CTC CAG TTA CCC AAG AG-3′ . Enh2R 5′-GGT TGA CTT AGA TTT CCT TAT GA-3′ . Enh2F 5′-TCA TAA GGA AAT CTA AGT CAA CC-3′ . Enh3F 5′-TGT AAG AGA AAC TGA AAT AAA AT-3′ . Enh3R 5′-ATT TTA TTT CAG TTT CTC TTA CA-3′ . Enh4F 5′-GGG GTA AGA AAT CTG ACG ACA GG-3′ . Enh4R 5′-CCT GTC GTC AGA TTT CTT ACC CC-3′ . Enh5F 5′-CTA TCA GGT CAG GCG ACC TCA GT-3′ . Enh5R 5′-ACT GAG GTC GCC TGA CCT GAT AG-3′ . Enh6F 5′-CGT CAA GGA AAC CTT ATT CCA TC-3′ . Enh6R 5′-GAT GGA ATA AGG TTT CCT TGA CG-3′ . MyogF 5′-TGG GGG GGC TCA GGT TTC TGT GGC GT-3′ . MyogR 5′-ACG CCA CAG AAA CCT GAG CCC CCC CA-3′ . NF1F 5′-TAT CTC TGG GTT CAT GCC AGC AGG G-3′ . NF1R 5′-CCC TGC TGG CAT GAA CCC AGA GAT A-3′ . mtEnh1F 5′-CTC TTG GGT AGG ATC CGC CCC TC-3′ . mtEnh1R 5′-GAG GGG CGG ATC CTA CCC AAG AG-3′ . mtEnh2F 5′-GGT TGA CGA ATT CTT GCT TAT GA-3′ . mtEnh2R 5′-TCA TAA GCA AGA ATT CGT CAA CC-3′ . mtEnh3F 5′-TGT AAG ACC AAC TGA AAT AAA AT-3′ . mtEnh3R 5′-ATT TTA TTT CAG TTG GTC TTA CA-3′ . mtEnh4F 5′-GGG GTA AGA AGG ATC CCG ACA GG-3′ . mtEnh4R 5′-CCT GTC GGG ATC CTT CTT ACC CC-3′ . mtEnh5F 5′-CTA TCA GGT CGG ATC CCC TCA GT-3′ . mtEnh5R 5′-ACT GAG GGG ATC CGA CCT GAT AG-3′ . mtEnh6F 5′-CGT CAA GGA AGG ATC CTT CCA TC-3′ . mtEnh6R 5′-GAT GGA AGG ATC CTT CCT TGA CG-3′ . Western blots were performed with protein extracts of GP and soleus muscles from cSix1KO mice and control mice as previously described [9] . 1∶1000 dilutions of anti-Six1 antibodies ( HPA001893 , Sigma ) or anti-β−tubulin antibodies ( 2128 , Cell Signaling ) were used . All graphs represent mean values ± SEM . Significant differences between mean values were evaluated using two-tailed , unpaired Student's t test ( when two groups were analyzed ) or one-way ANOVA followed by Student Newman-Keuls test ( for three or more groups ) . For the construction of the pGL3-Myh2/1/4 , c57bl6N mouse DNA was first used as a template to clone 1 . 1 kbp promoters of Myh2/1/4 with forward KpnI/SacI 5′-TTC AGA AAC TGC ATC ACT TAA A-3′ and reverse MluI , 5′-GCA GCT CGG GCA GTG GCC AGT GT-3′ , forward KpnI/SacI 5′- CAT ATC TGC ATC TCT AGA TAC C-3′ and reverse MluI , 5′- GGC AGC AGC AGC CAG GAT GTG T-3′ , forward KpnI/SacI 5′- ACC GCT AGC CTT GAG CCT TTG-3′ and reverse MluI , 5′- ATA GCG AGA GCC CTT TGT TCT C-3′ , respectively . Myh2/1/4 promoter fragments were subsequently inserted into an KpnI-MluI digested pGL3 basic plasmid . For the construction of the pGL3-Enhancer-Myh2/1/4 , mouse DNA was first used as a template to clone the enhancer with forward KpnI 5′- GCG TTT CTA ATT CGG CTT GAA C-3′ and reverse SacI , 5′- CAT TTC CTT CCT CTA AAG GCT CTT TATT C-3′ . This enhancer fragment was subsequently inserted into KpnI-SacI digested pGL3-Myh2/1/4 plasmids . For the construction of the pGL3-mutEnhancer-Myh2/1/4 , the six MEF3 sites of the enhancer were mutated as follows; MEF3 1: 5′-GTAACTGGA to 5′-GTAGGATCC; MEF3 2: 5′-TTAGATTTC to 5′-GAATTCTTG; MEF3 3: 5′-GAAACTGAA to 5′-CCAACTGAA; MEF3 4: 5′-GAAATCTGA to 5′-GAAGGATCC; MEF3 5: 5′-GTCAGGCGA to 5′-GTCGGATCC; MEF3 6: 5′-GAAACCTTA to 5′-GAAGGATCC . All plasmids sequence was confirmed by sequencing . For the construction of the pSF-pA-CMVe-linc-MYH , genomic DNA fragment containing linc-MYH was obtained from digestion of a BAC clone containing ( RP23-61C14 ) by BsaWI and AvrII . The 13 . 3 kbp DNA fragment was subsequently inserted into a SpeI-XmaI digested pSF-pA-CMVe plasmid . For linc-Myh gain of function experiments , the empty pSF-pA-CMVe plasmid was electroporated in the contra-lateral Soleus as a negative control . For the construction of the pGL3-Actb , pGL3-Sln and pGL3-Tnni1 , mouse DNA was first used as a template to clone the promoters of Actb , Sln , and Tnni1 with forward 5′- TTT CTC TAT CGA TAG GTA CCT TTG AGC TCC TGA CCC CGT GTG TAG CTC T-3′ and reverse 5′- GAG CCC GGG CTA GCA CGC GTA AGG AGC TGC AAA GAA GCT G-3′ , forward 5′- TTT CTC TAT CGA TAG GTA CCT TTG AGC TCT ACC GAC TAT CAT GCC CAC A-3′ and reverse MluI , 5′- GAG CCC GGG CTA GCA CGC GTC AGG CTA CCA AGG ACC TCA G-3′ , forward 5′- TTT CTC TAT CGA TAG GTA CCT TTG AGC TCC TGG GAT TTG AAC CCA TGA C-3′ and reverse 5′- GAG CCC GGG CTA GCA CGC GTC CTC ACC ACA GAC TGC AGA G-3′ , respectively . Actb , Sln , and Tnni1 promoter fragments were subsequently inserted into a KpnI-MluI site of pGL3 basic plasmid by GeneArt kit ( Life Technologies ) . For the construction of the pGL3-Enhancer-Actb , pGL3-Enhancer-Sln and pGL3-Enhancer-Tnni1 , the enhancer fragment was subsequently inserted into an KpnI-SacI site of pGL3-Actb , pGL3-Sln and pGL3-Tnni1plasmids . All plasmids sequence were confirmed by sequencing . Two µg of Luciferase-expressing vector and one hundred ng of pRL-TK vector ( Promega ) were introduced into TA muscles of 8 week-old mice by electroporation as previously described [11] . Two weeks following electroporation , the TA muscles were dissected and frozen in liquid nitrogen before processing . The TA muscles were homogenized in Passive Lysis Buffer ( Dual-Luciferase Reporter Assay System , Promega ) and rotated for 15 minutes . The homogenate were centrifuged to remove debris , and the supernatant was used for measurement of Luciferase activity according to manufacture's instruction ( Dual-Luciferase Reporter Assay System , Promega ) . In order to computationally identify MEF3 binding sites , we built a PWM ( position-specific weight matrix ) for MEF3 , starting from a list of 15 binding sites ( see Table S5 ) that were previously tested by Electrophoretic Mobility Shift Assay on the basis of their proximity to the MEF3 Myogenin consensus GAAACCTGA [10] . The PWM ( shown in Figure S6 and Table S6 ) was generated by running the de novo motif finder Imogene [39] on small DNA fragments ( see Table S5 ) that contained these binding sites . Imogene used phylogeny to enrich mouse set of DNA fragments with orthologs in 11 other mammalian sequenced genomes and to produce a refined PWM . The information content of the PWM ( the genmot Sg parameter of Imogene ) was set to 8 . 7 bits . Binding sites were predicted using a prediction threshold ( the scangen Ss parameter of Imogene ) of 9 bits and requiring conservation , as explained in [39] . | Adult skeletal muscles are classified into fast-type and slow-type , which display different resistance to muscle atrophy and metabolic protection against obesity . We identify in this manuscript a new mechanism controlling in vivo adult muscle fiber-type specification implicating a long intergenic non-coding RNA , linc-MYH . We demonstrate a three-element genetic partnership , where an enhancer under the control of the myogenic homeoprotein Six1 functions as a regulatory hub to control fibre phenotype . In this partnership , the enhancer controls positively the expression of both the adjacent fast myosin heavy chain ( MYH ) gene cluster and of linc-MYH . linc-MYH is present only in adult fast type skeletal myofibers and controls their phenotype by suppressing slow-type gene expression . The regulation of linc-MYH could provide a lead for new therapeutic approaches or drug development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"cell",
"differentiation",
"dna",
"transcription",
"gene",
"function"
] | 2014 | Six Homeoproteins and a linc-RNA at the Fast MYH Locus Lock Fast Myofiber Terminal Phenotype |
Mathematical models of mitochondrial bioenergetics provide powerful analytical tools to help interpret experimental data and facilitate experimental design for elucidating the supporting biochemical and physical processes . As a next step towards constructing a complete physiologically faithful mitochondrial bioenergetics model , a mathematical model was developed targeting the cardiac mitochondrial bioenergetic based upon previous efforts , and corroborated using both transient and steady state data . The model consists of several modified rate functions of mitochondrial bioenergetics , integrated calcium dynamics and a detailed description of the K+-cycle and its effect on mitochondrial bioenergetics and matrix volume regulation . Model simulations were used to fit 42 adjustable parameters to four independent experimental data sets consisting of 32 data curves . During the model development , a certain network topology had to be in place and some assumptions about uncertain or unobserved experimental factors and conditions were explicitly constrained in order to faithfully reproduce all the data sets . These realizations are discussed , and their necessity helps contribute to the collective understanding of the mitochondrial bioenergetics .
The simulation of mathematical models of mitochondrial bioenergetics provides a powerful analytical alternative to performing numerous exhaustive experiments . Such models aid in the interpretation of experimental data and facilitate experimental design for elucidating the supporting biochemical and physical processes . Current experimental techniques limit the ability to resolve details of the mitochondrial bioenergetic processes in vivo . Specifically , many different chemical species and events must be simultaneously monitored to track the profusion of enzymatic reactions involved in the tricarboxylic acid ( TCA ) cycle , β-oxidation , the electron transport system ( ETS ) , ATP synthesis and electrolyte dynamics . Currently , it is impossible to accurately and simultaneously measure all of these enzymatic processes in vivo with any degree of precision . Nevertheless , a plethora of experimental data is available on mitochondrial bioenergetics that was collected using a variety of techniques , experimental conditions , and tissue sources . No existing experimental data set is complete that consists of measurements of all of the supporting chemical species and events; therefore , the correct interpretation of the available experimental data in isolation or as a collective unit is difficult . This requires careful consideration of all potential data-consistent dynamics of the unobserved species and events . To aid in the interpretation , a quantitative framework established via mathematical model development and parameter identification through experiment simulation is commonly employed in a collective manner so that a data-compatible , semi-mechanistic description of the underlying bioenergetic processes emerges . With the addition of each new experimental data set , the model matures either through corroboration via evidence that supports the hypothesized mechanisms encoded within the mathematical representation or through modification of the model structure and parameters to refine and/or reveal more insights and alter the supporting hypothesized mechanisms underlying bioenergetic processes . Several mathematical models have been developed [1]–[6] to describe aspects of mitochondrial bioenergetics , but none currently capture the complete dynamics of all metabolically relevant ion and substrate regulatory functions observed during physiological and pathophysiological conditions . As the next step towards constructing a complete physiologically faithful mitochondrial bioenergetics model , this manuscript describes the development and corroboration of a mathematical model based on cardiac mitochondria that builds upon these previously published models . The Yugi and Tomita model [6] simulates mitochondrial bioenergetics with the most breadth . They have compiled a large mitochondrial bioenergetics model that qualitatively captures the basic mitochondrial bioenergetic phenomena and included much of the mitochondrial biochemistry from a variety of species and organs . Although the Yugi and Tomita model has been successfully modified to predict the dynamic response of β-oxidation in the context of human disease [7] , it does not incorporate some details required to reproduce a few specific bioenergetic regulatory features of interest herein . The Wu et al . model [1] was chosen as the base model for this work due to its meticulous attention to thermodynamics and inclusion of the various biochemical species present in the mitochondrial milieu . Our model structure and the parameter values were selected so that simulations of the experimental conditions on porcine or rat heart mitochondria simultaneously fit the inorganic phosphate ( Pi ) control over mitochondrial bioenergetics [8] and TCA intermediate dynamics [9] data sets upon which the Wu et al . model was developed as well as additional experimental data on the extra-mitochondrial calcium-dependent steady state matrix calcium concentrations [10] and mitochondrial matrix volume dynamics [11] . Numerous other experimental studies were used to fit constants for the employed rate expressions as described in the Supplemental Material ( Text S1 ) . The resulting extended model is corroborated with additional experimental data on the steady state behavior of the TCA cycle [9] , volume dynamics under various bioenergetic/pharmacological interventions [11] and the bioenergetic/volume responses of mitochondria to variations in buffer osmolarity [12]; a local sensitivity analysis was also used to explore the robustness of the model structure relative to these simulated experiments . This manuscript describes all phases of this process including the model development , parameter estimation and corroboration . It concludes with a discussion of the insights on the bioenergetic processes obtained during the model extension .
The model integrates mitochondrial bioenergetic processes as shown in Figure 1 including oxidative phosphorylation , the ETS , the TCA cycle and related reactions , the Na+/Ca2+ cycle and the K+-cycle . To maintain thermodynamic consistency , all reactions in the model , as shown in Table 1 , are represented as thermodynamically balanced and reversible . ( It should be noted that some of the reactions could have been treated as irreversible reactions since the conditions necessary to reverse them lie very far from physiological conditions . ) The model is primarily based on the mitochondrial bioenergetics model proposed by Wu et al . [1] and extended to address: 1 ) updated formalisms for the calcium-sensitive dehydrogenases , as well as , several modified TCA cycle related rate expressions; 2 ) a two-site model for the adenine nucleotide transporter [13]; 3 ) the Na+/Ca2+ cycle including a non-linear dependence on inner mitochondrial membrane potential difference , ψinside - ψoutside , ( Δψ ) for the calcium uniporter ( CaUNI ) [14] , magnesium inhibition kinetics for the CaUNI , and the proton-regulation of the mitochondrial Na+/H+ exchanger ( mNHE ) ; and 4 ) the K+-cycle including the mitochondrial ATP-dependent K+ channel ( mKATP ) , electrophoretic K+-leak and the mitochondrial K+/H+ exchanger ( mKHE ) with the mitochondrial matrix volume regulation dynamics hypothesized by Garlid [15] . Each addition is described in the following paragraphs . The model proposed in the manuscript is a 73 state system of differential-algebraic equations ( DAEs ) that consists of 65 non-linear ordinary differential equations ( ODEs ) ; five algebraic conservation expressions to compute matrix ATP , guanidine triphosphate ( GTP ) , reduced nicotinamide adenine dinucleotide ( NADH ) , ubiquinol ( UQH2 ) and reduced cytochrome c ( c2+ ) ; one algebraic expression to compute matrix water volume; one algebraic expression to compute inner membrane space ( IMS ) water volume and one algebraic expression to compute matrix chloride content ( Cl− ) . The majority of the experimental data used to parameterize the model proposed in this manuscript were derived from heart tissue of either bovine , porcine or rat with some data obtained from liver tissue . Part S1 of the Supplemental Material ( Text S1 ) lists the state variables and general parameters , Part S2 contains the system of DAEs comprising the model and Part S3 provides a detailed description of the rate expressions and their parameters used to construct the system of DAEs . Part S3 also includes the fitness for some rate expressions calibrated with additional experimental data not explicitly indicated in this manuscript . The TCA and related enzyme rate expressions are structurally identical to Wu et al . except for a few alterations . To include the calcium-dependence of matrix dehydrogenases , the rate expressions for pyruvate dehydrogenase ( PDH ) , isocitrate dehydrogenase ( IDH ) and α-ketoglutarate dehydrogenase ( αKGDH ) were modified . PDH , an important regulatory enzyme involved with mitochondrial bioenergetics , is responsible for the oxidative decarboxylation of pyruvate , transacylation of an acetyl group to CoA and production of reducing equivalents for the ETS . A similar rate expression found in Wu et al . was used with a few notable modifications . The proton , divalent cation and adenine nucleotide regulatory mechanisms were inserted into the expression to reproduce the available data [16] . IDH is responsible for the oxidative decarboxylation of isocitrate to produce α-ketoglutarate and reducing equivalents for the ETS . The rate expression used in the model is from Qi et al . [17] . The key TCA regulatory enzyme , αKGDH , is responsible for the oxidative decarboxylation of α-ketoglutarate transferring a succinyl group to CoA and producing reducing equivalents for the ETS . The consensus hexa-uni-ping-pong mechanism with the appropriate activation and inhibition modifications was used to reproduce a wide variety of data from four independent data sets [18]–[21] . Two additional rate expressions that deviated from Wu et al . are the glutamate-aspartate exchanger ( GAE ) and dicarboxylate carrier ( DCC ) . The GAE is a key enzyme in the malate-aspartate shuttle and is particularly important maintaining state 3 NADH levels when mitochondria respire on glutamate and malate . This electrogenic exchanger is activated by calcium and swaps glutamate and a proton with aspartate taking advantage of the energized state of mitochondria established by the ETS . The enzyme reaction was modeled based on a rapid equilibrium bi-bi mechanism with a third substrate , protons , added to the rate expression; it was fit to data from bovine heart mitochondria [22]–[23] . The DCC exchanges TCA cycle intermediates malate , succinate and hydrogen phosphate . This exchanger was also modeled as forming a ternary complex with its substrates , and the kinetic parameters were fit to rat liver mitochondrial experimental data [24]–[25] . The ANT is the enzyme responsible for exchanging unchealated ATP and ADP across the mitochondrial inner membrane . Previous models used a ping-pong mechanism that employed a single adenine nucleotide binding site whereby Δψ affected only ATP binding [26] . This type of mechanism has been shown to inadequately describe the enzyme kinetics , and studies have identified at least two distinct adenine nucleotide binding sites [27]–[28] . The model presented by Metelkin et al . [13] addresses these issues and was therefore chosen to describe the ANT kinetics . The parameters were refit to the original data to include the effect of Na+ and K+ chelation of adenine nucleotides . The model proposed in this manuscript incorporates mitochondrial calcium dynamics similar to Nguyen et al . [2] , Cortassa et al . [3] and Dash and Beard [14] . The CaUNI is similar to the expression found in Dash and Beard including the nonlinear dependence on Δψ with the addition of explicit magnesium inhibition based on experimental data . In Part S3 of the Supplemental Material ( Text S1 ) , the magnesium inhibition was used to show that a single calcium dissociation constant with magnesium acting as a competitive inhibitor against calcium binding for the channel is capable of reproducing the experimental data ( consisting of both rat heart and liver mitochondria ) . The mNCE is similar to [2]–[3] , [14] with the noted addition of a hypothetical matrix calcium activation mechanism . This calcium activation mechanism resulted in comparable dynamics under the calcium loading experiments used to fit Dash and Beard's matrix calcium inhibition mechanism for the CaUNI but was extended to be analogous with the current experimental evidence regarding the sarcolemmal isoform [29] . The mNHE was slightly modified from Nguyen et al . to include a hill coefficient of 2 for the proton regulation mechanism . The ‘futile’ K+-cycle plays a major role in mitochondrial volume homeostasis [30]–[34] . Potassium influx via the mKATP and electrophoretically driven potassium uptake via leak pathways must be balanced with potassium efflux from the mKHE . Originally , the mKHE was thought to be regulated by the carrier brake hypothesis [35] . This essentially involves some endogenous element in the matrix , such as magnesium , that serves as the “carrier brake” that is reversibly released by matrix swelling . Brierley and Jung call into question this hypothesis noting that under physiological conditions , the known inhibitors of the exchanger are present at concentrations much greater than their respective inhibitory constants [36] . Garlid then proposed that the mKHE is additionally regulated by matrix volume with membrane stretching activating the exchanger [15] . Further evidence for this mechanism is provided by the results of the analysis of the mathematical model proposed in this manuscript discussed below . To parameterize the model , four independent data sets consisting of 32 data curves were used from Bose et al . [8] , LaNoue et al . [9] , Wan et al . [10] and Kowaltowski et al . [11] . In the manuscript these studies will be henceforth referenced as the Bose data set , LaNoue data set , Wan data set and Kowaltowski data set , respectively . For each data set , the model was initialized from a condensed , fully oxidized and de-energized state via initialization simulations that replicated the experimental incubation conditions ( i . e . Pi- and Ca2+-depletion ) prior to parameter estimation . The parameter identification was subsequently conducted using simulation conditions closely mimicking those of the experimental methods . The resulting values of the 42 adjustable parameters , their definitions , their best fit values and their associated normalized local sensitivity coefficients ( LSC ) are provided in Table 2 . The following paragraphs describe the ability of the model with these fitted parameter values to reproduce the experimental results from these four independent experimental data sets; pertinent details of the experimental conditions and their replication through model simulations are described in the Methods section . The NADH-linked respiration components of the model were fitted against the Pi-titration experiments performed by Bose et al . [8] . In their manuscript , the authors reported a rich bioenergetic data set using glutamate/malate energized , Pi-depleted mitochondria under both state 2 and state 3 respiration conditions . State 3 was initiated and maintained with a sufficient bolus addition of 1 . 3 mM ADP . At this concentration , maximum respiration was maintained for at least half a minute before the ANT exerted its control due to limited substrate availability . Mitochondrial Δψ , NAD/NADH redox state , myocardial oxygen consumption ( MVO2 ) , cytochrome c3+/c2+ redox state and matrix pH were reported as the extra-mitochondrial Pi was progressively increased from 0 to 10 mM . Figure 2 shows that the model was capable of fitting this data set . Similar to the Wu et al . model , the model presented in this manuscript produced state 3 Δψ that reproduced the Pi-titration trend but at 10–15 mV lower than the experimentally measured state 3 Δψ . To achieve the low matrix pH levels observed experimentally , the mKHE was temporarily replaced with an expression with sufficient activity to rapidly equilibrate [K+]mtx[H+]ims/[K+]ims/[H+]mtx , and the initial matrix [K+] was adjusted to approximately 125 mM . Without these modifications for this data set , the model produced matrix pH of up to 7 . 2–7 . 5 pH units under the experimental conditions simulated . This issue is further explored in the Discussion . Overall , the experimentally reported Δψ , NAD/NADH redox state and MVO2 Pi-titrations as well as the expected increase in volume with increasing Pi are captured by the model . The TCA cycle intermediate dynamics of the model were fitted to the data set presented by LaNoue et al . [9] . In their experiments , they used pyruvate/malate and pyruvate energized mitochondria in both state 2 and state 3 respiration . State 3 was initiated by the addition of 0 . 5 mM ADP and maintained using a hexokinase trap . They reported detailed time series data of most of the TCA cycle intermediates for each experimental condition . The model was able to capture the salient features of the pyruvate/malate energized mitochondrial TCA cycle intermediate dynamics as shown in Figure 3 . The model simulated pyruvate , citrate , isocitrate , α-ketoglutarate , succinate and malate transients similar to the experimental data . Figure 4 shows that the simulated aspartate and glutamate dynamics using pyruvate energized mitochondria under both state 2 and state 3 respiratory conditions were also consistent with experimental data . In these simulations , the endogenous matrix aspartate content provided sufficient amino acid substrates for glutamate oxaloacetate transaminase ( GOT ) while glutamate efflux via glutamate-H+ cotransporter reduced the total available matrix asparate/glutamate pool . The Na+/Ca2+ cycle was fitted to the steady state data from Wan et al . [10] . In their experiments , they used ATP-energized mitochondria and monitored steady state Ca2+ levels at varying extra-mitochondrial Ca2+ , Na+ and Mg2+ concentrations . The ATP initialized the mitochondrial Δψ to approximately −120 mV in the model simulations corroborating the data reported by Territo et al . [37] . The model's capability of fitting the extra-mitochondrial Ca2+- and Na+-dependence on matrix steady state Ca2+ concentrations is illustrated in Figure 5 . The steady sate matrix free Ca2+ data in the absence of extra-mitochondrial Na+ is not shown nor was used in the parameter estimation , because the Na+-independent calcium efflux was not included in the model structure . ( Note , with an electroneutral Ca2+/2H+ exchange mechanism , the Na+-independent steady state matrix Ca2+ data could be reproduced by the model; however , this mechanism was not included in the current model formulation . The rationale for this choice is described in more detail in the Discussion . ) The volume dynamics were fitted to the transient mitochondrial matrix swelling data published by Kowalowski et al . [11] . They measured the effect of varying matrix ATP levels on the swelling dynamics in K-salt media with succinate-energized mitochondria . Figure 6 shows that the volume dynamics were captured well by the model . When a small amount of ATP was included in the buffer in the presence of oligomycin , a F1FO ATP synthase inhibitor , the mitochondria swelled to approximately 1 . 5 µL/mg in about four minutes . When the ATP was supplemented with ADP , oxidative phosphorylation was activated which lowered the Δψ and generated matrix ATP . This reduced the electrophoretic potassium uptake and inhibited the mKATP channel , respectively , so that the steady state volume reached a lower value of approximately 1 . 2 µL/mg . When ATP was not present , all residual matrix ATP was converted to ADP via reverse activation of the F1FO ATP synthase . This fully activated the mKATP channel , so that the final steady state volume reached a much higher value of about 2 . 0 µL/mg . Model corroboration was necessary to establish confidence in the model resulting from these efforts . Herein , the corroboration considered the robustness of the model to local parameter perturbations , the qualitative agreement of predicted trends with experimental observations , and the ability of the model to reproduce experimental data that was not used in fitting its parameters . A local sensitivity analysis on the model was performed to determine how robust the model simulations were to local perturbations in the parameter values of all the experiments used in the parameter identification . The absolute-value normalized local sensitivity coefficients ( LSC ) were computed using Equation 2 as defined in the Methods section which considered variations for every state variable dynamics uniformly throughout the simulated experiment duration . The average LSC of all 359 parameters was only 7 . 38×10−3 with a variance of 1 . 18×10−3 . This implies on average that a perturbation of 1% for a given parameter results in less than a 0 . 738 +/− 0 . 118% change in the state dynamics of the model for the experiments considered . Of the 42 adjustable parameters , the average of the absolute-value normalized LSC reported in Table 2 was 1 . 26×10−3 with a variance of 3 . 23×10−4 . These low sensitivities for the model parameter values did not reveal any inherent problems with the model structure . The model was also able to reproduce the well known mitochondrial shrinkage/swelling dynamics in the presence of Pi and ADP . Figure 2F shows that as the extra-mitochondrial Pi-titration was increased , mitochondrial matrix water volume increased with the state 3 volume being lower than the state 2 volume . Unfortunately , no volume data for the Bose data set was given; however , the model results do corroborate the qualitative observations presented therein with volume increasing for higher amounts of Pi in the medium . The model was directly corroborated by predicting the steady state accumulation of extra-mitochondrial α-ketoglutarate during state 2 respiration at various extra-mitochondrial malate concentrations as shown in Figure 7 . As the extra-mitochondrial malate concentration was increased , the total amount of α-ketoglutarate generated from the oxidation of pyruvate was also increased and exchanged with the malate in the media via the oxoglutarate-malate exchanger . This exchanger is a critical component of the malate-aspartate shuttle and is particularly important in heart tissue . The model was again directly corroborated by predicting the experimentally observed mitochondrial volume dynamics after various bioenergetic and/or mKATP interventions . Figure 8 shows that when the mKATP channel was manipulated via normal or pharmacological pathways , the model was capable of predicting the matrix volume changes observed . This highlights the interactions between K+-influx via mKATP and mKleak and K+-efflux via mKHE and identifies their role in mitochondrial volume regulation . Upon energization , electrophoretic uptake of K-salts increased matrix volume in an osmotic fashion . The first principles representation of mitochondrial volume dynamics was captured well by the model . Finally , the model's ability to reproduce the expected trends in bioenergetic variables under varying KCl buffer osmolarity conditions was explored . Devin et al . [12] monitored changes in state 2 and state 3 Δψ , matrix pH , proton motive force , MVO2 , NADH level and matrix volume as the buffer osmolarity was changed from a hypoosmolar to a hyperosmolar KCl medium using rat liver mitochondria . Although the model presented in this manuscript was optimized to reproduce experimental data primarily from heart tissue , the effect of varying medium osmolarities on key bioenergetic variables was qualitatively reproduced . Only a partial quantitative comparison to this data was possible due to tissue source differences , experimental limitations , and model structure ( as further described in the Discussion ) . Figure 9A shows that the simulated state 2 MVO2 matched the experimental trend; however , the model simulated the incorrect trend for state 3 MVO2 . This is attributed to the tissue source and model structure detailed below in the Discussion . The simulated Δψ trends matched the reported experimental trends ( state 2: −140 to −160 mV and state 3: −125 to −135 mV ) as shown in Figure 9B . Also , the simulated ΔpH ( pHmtx-pHims ) trends matched the reported experimental trends ( state 2: −30 to −45 mV and state 3: −35 to −50 mV ) as shown in Figure 9C . Although the simulated Δψs and the ΔpHs are over- and underestimated , respectively , the proton motive force ( Δψ + 2 . 303RT/F ΔpH ) matched the experimental data very well as shown in Figure 9D . In other words , as the medium osmolarity varied the total thermodynamic driving force established by the ETS was very similar to that observed experimentally . Devin et al . observed that from hypoosmotic conditions , the NADH levels increased and leveled off as the buffer osmolarity was increased towards hyperosmotic conditions . As shown in Figure 9E , the model was able to reproduce this trend very well . They also reported that in hyperosmotic KCl medium , mitochondrial matrix volume is efficiently regulated , such that the steady state volume is essentially retained , but in hyposomotic KCl medium , the matrix volume dramatically increases relative to isoosomotic conditions [12] . Figure 9F shows that the model volume mechanics adequately reproduced this phenomenon . Overall , the model corroborated the trends of the changes in bioenergetic variables as buffer osmolarity varies .
The model presented in this manuscript is based on previous models [1]–[4] and includes integrated calcium dynamics and a detailed description of the K+-cycle and its effect on mitochondrial bioenergetics and matrix volume regulation . Simulations were used to fit 42 adjustable parameters to four independent experimental data sets consisting of 32 data curves of both transient and steady state data . A sensitivity analysis was performed on the model to reveal the most sensitive components of mitochondrial bioenergetics relative to the experimental conditions modeled herein and revealed no inherent model structural problems . Finally , several simulations were performed to corroborate the model . The mitochondrial volume dynamics and the associated K+-cycle appear to play an important role in cellular and mitochondrial bioenergetics [15] , [38] . Energy transduction , namely adenine nucleotide outer membrane permeability , is regulated under both physiological and pathophysiological conditions [39] . The IMS volume is partly responsible for this regulation by having a direct effect on the cellular bioenergetics in vivo [38] . For example , the adenine nucleotide outer membrane permeability is typically high in freshly isolated mitochondria due to matrix contraction following potassium depletion . During mitochondrial swelling , the increase in matrix volume causes a reciprocal decrease in IMS volume that enables creatine kinase to bind to the voltage-dependent anion channel thus reducing the adenine nucleotide outer membrane permeability . Matrix contraction also occurs in vivo during ischemia . This contraction interferes with the regulation of the adenine nucleotide outer membrane permeability , thereby enabling mitochondria to burn up all the cells available ATP and primes the cell for apoptosis before reperfusion . As a natural defense , potassium influx via mKATP increases matrix volume and helps mitigate the detrimental effects of increased adenine nucleotide outer membrane permeability [38] . As a first step to consider these important volume regulatory events , the model incorporates simple volume dynamics based on osmotic pressures generated by the K+-cycle and other associated processes . Future work studying this intricate energy transduction mechanism is in the beginning stages of development . The hypothesized volume-dependent mKHE by Garid [15] was incorporated into the model . This volume dependence is necessary to maintain sufficient potassium efflux at high Δψ during mKATP opening . Without this dependence , the mKHE would be an ineffective volume regulatory mechanism , and the outer membrane would rupture . Specifically , potassium influx induced upon mKATP opening is maintained and essentially constant at a given Δψ because the current cannot sufficiently depolarize the Δψ and because of thermodynamic considerations [33] . Also , model simulations revealed that matrix free Mg2+ only decreased from 0 . 4 mM to approximately 0 . 15 mM ( depending on total amount of Mg-ligands , such as ATP and Pi , present in the matrix and the matrix volume ) under the experimental conditions . This decrease is insufficient to serve as the primary volume controller required by the carrier-brake hypothesis . These observations require that mKHE posses some sort of volume dependence enabling an effective volume controlling mechanism . This hypothesis was supported by simulation results presented in Figures 6 , 8 and 9F whereby the exquisite control of matrix volume exhibited by mKHE was revealed . To faithfully reproduce all the data sets , a certain network topology had to be in place and some assumptions about uncertain or unobserved experimental factors and conditions were explicitly constrained . It was found that these network features and experimental assumptions described below were necessary to successfully recreate all the experimentally observed data and trends . The necessity of these realizations contributes to our collective understanding of the mitochondrial bioenergetics . The intrinsic thermodynamic dissipation of a system can override or mitigate enzymatic regulation [40] . For example , in glutamate/malate energized mitochondria , αKGDH serves as a key regulatory enzyme responsible for maintaining sufficient NADH levels sustaining MVO2 rates in state 2 and state 3 . This reaction is far from equilibrium making it sensitive to its regulatory mechanisms . In contrast , malate dehydrogenase ( MDH ) is much closer to equilibrium so its regulatory mechanism less effectively controls the dehydrogenase rate . In state 2 , αKGDH is a major enzyme in the pathway responsible for regenerating the matrix ATP that is consumed by the F1FO ATP synthase to help the ETS establish a high Δψ . In state 3 , this enzyme's activity helps dictate which path the carbon substrates flow through the TCA cycle . The regulation for this enzyme helps enable mitochondria to achieve the steady state NADH levels observed with the Bose Pi-titration data seen in Figure 2B . The relative rates identified from the model simulations ( not shown ) predicted that the regulation of MDH played less of a role in the steady state NADH Pi-dependent levels than the regulation of αKGDH . Animal model species specific parameterization may be required for detailed mathematical models of the mitochondrial bioenergetics; however , at this time there is not sufficient data from a single species to fully characterize the dynamics . For example , all the kinetic parameters for the αKGDH expression were found using independent data sets obtained from porcine heart mitochondria as described in Part S3 of the Supplemental Material ( Text S1 ) [18]–[21] . Only , the maximum rate constant , was refit with the fully integrated model simulations that used data from both porcine and rat heart mitochondria . Future compensation for the species specific enzyme kinetic differences between these two isozymes may further improve the fit of the simulated α-ketoglutarate dynamics to the experimental data in Figure 3D . The tissue type used for the supporting experiments is also important for parameterization of a semi-mechanistic mathematical model . For example , in the model , several exchangers and cotransporters are reported to possess low activity in heart tissue compared with other tissues [41] . Although the definition for low activity was ambiguous; herein to reproduce the experimental data , it was necessary that some of these processes possess unexpectedly elevated activities . For example without sufficiently active glutamate-H cotransport , the model was unable to reproduce the observed aspartate/glutamate dynamics reported by LaNoue et al . [9] in Figure 4 . The glutamate-H cotransporter is responsible for the electroneutral transport of the amino acid glutamate and a proton through the inner-mitochondrial membrane down a concentration gradient . This provides a glutamate leak pathway that reduces the matrix aspartate/glutamate pool . Reducing aspartate availability in the matrix prevents the thermodynamically favorable reaction catalyzed by GOT from consuming all of the endogenous aspartate . Alternatively to the proposed elevation in the activity of the cotransporter , it has been argued that there may be two separate aspartate pools [42] in mitochondria . Compartmentalizing the total aspartate pool with slow or volume-dependent transport rates would also enable the model to reproduce the aspartate/glutamate dynamics in the LaNoue data set . As further indirect support for the proposed elevated glutamate-H cotransporter activity , increased glutamate influx helps enable state 2 and state 3 respiration on glutamate and malate for the Bose data set simulations as shown in Figure 2 . At this point , neither mechanism has been proven experimentally . Another exchanger reported to possess low activity in heart tissue is the dicarboxylate carrier [43] . This carrier is responsible for the exchange of primarily Pi , malate and succinate . However , without sufficiently high activity , the succinate-energized mitochondrial matrix volume data reported by Kowaltowski et al . [11] could not be reproduced by the model . Elevated DCC rates were necessary to provide sufficient succinate influx allowing electrophoretic potassium uptake . If the DCC activity was limited to the reported maximum rate [44] , the mitochondrial Δψ would not polarize substantially and mitigate the reported potassium-dependent volume increase . The chosen substrates for the DCC also affected the model simulation capabilities . In the model formulation , only Pi , malate and succinate were assigned as the DCC substrates . Fumarate is also reported to be a substrate for the DCC [45] but was not included in the model due to insufficient data to characterize the kinetics . Including fumarate in the list of the DCC substrates would enable to the model to reproduce the accumulated fumarate data from LaNoue data set ( not shown ) . Also , the omission of fumarate as a substrate for the DCC in addition to the DCC elevated activity contributed to the simulated state 3 malate net oxidation present in Figure 3F . An additional tissue source related phenomena uncovered during model development was the choice of the calcium dissociation constant for the CaUNI . During model development , a single calcium dissociation constant was chosen to model the CaUNI kinetics from rat heart and liver tissue . This was achieved by considering the competitive nature of magnesium inhibition with respect to calcium binding . It is plausible that the major difference between liver and heart CaUNI kinetics is due to different expression levels versus different calcium binding affinities ( as may be plausible between different species ) ; however , more research concerning this matter needs to be done . Mitochondria from specific tissue types are phenotypically different and contain various amounts of electron transport proteins , matrix proteins and lipid types optimized to support their designated function . Specifically , heart mitochondria possess much higher electron transport activity relative to liver mitochondria [46] . Therefore , we partially attribute the discrepancies between the simulated and reported values for the MVO2 , Δψ and ΔpH , and under varying KCl buffer osmolarity conditions to differences in the tissue source . The experiments outlined in Devin et al . [12] were done using mitochondria isolated from rat liver while the model was primarily developed to fit data obtained from heart tissue . This is exemplified in Figure 9A–C . The model simulated much higher state 3 MVO2 rates ( 200 versus 100 nmolO/min/mg ) ; however , the state 2 MVO2 trends ( quantitative data not given ) matched those reported . The measured liver mitochondria Δψs were several 10 s of mVs lower relative to simulated Δψs and is also partially attributed to the model's overestimation of state 3 Δψ seen in Figure 2C . Although the Δψ and ΔpH energetic variables were quantitatively different , it is interesting to note that the thermodynamic variable ( proton motive force ) was consistent across the tissue types as would be expected . Moreover , the authors suggested that the measured hypoosmolar Δψ may have been in error due to volume dependent rhodamine 123 accumulation . This accumulation may partially account for the discrepancy between the measured hypoosmotic state 2 proton motive and that simulated . Thus the model predicts that the cardiac mitochondrial bioenergetic response to varying buffer osmolarity under the conditions presented in Devin et al . would produce Δψs in the range of −200 to −140 mV , ΔpH in the range of 7 to 30 mV ( 0 . 2 to 0 . 5 pH units ) and MVO2 rates in the range of 200 to 15 nmolO/mg/min . The translation of the biochemical processes to appropriate mathematically descriptive expressions plays a large role in the simulated dynamics . For example , the citrate and isocitrate overestimation by the model as shown in Figure 3C is attributed to the simple passive exchange process used to model the tricarboxylate carrier ( TCC ) . Replacing this exchange process with a more kinetically descriptive mechanism [47] would help prevent the simulated early accumulation of citrate and isocitrate; however , there is insufficient data to parameterize such a descriptive mechanism . Partially as a function of the simplified description , the activity of the TCC is elevated even though it is reported to possess low activity in heart tissue compared to that in liver tissue [41] . Additionally , the passive TCC exchange process required higher PDH rates in order to sustain sufficient citrate , isocitrate , α-ketoglutarate and succinate levels during both state 2 and especially state 3 respiration seen in Figure 3 . Together , the TCC and DCC modeling approximations contribute to the simulated higher malate net oxidation rates than those observed in the LaNoue data set ( Figure 3F ) since the TCC also exchanges malate across the inner-mitochondrial membrane [47] . During the model development , it is important to consider any artifacts in the experimental data that may have been inadvertently generated during the mitochondrial isolation . For example , the extraction medium must contain Pi to achieve stable , well-coupled mitochondria [41] . To enable the Bose study with Pi-depleted mitochondria , a special isolation procedure was necessary . This Pi-depletion method may have dramatically changed some of the mitochondrial protein phosphorylation states thus having an unknown regulatory effect [48] . Within the short time scales of the Bose experiments , the slower phosphorylation and dephosphorylation events may not have sufficiently occurred upon the Pi-titration . Specifically , the Pi-depletion method may have altered the proton permeability via cation/proton exchange and/or anion/proton cotransport activity . In these experiments , the buffer pH was fixed at 7 . 1 . With the mKHE rate expression identical to that used with the other data sets , the simulated Δψ is underestimated while the ΔpH is overestimated , but the total energetic contribution from both Δψ and ΔpH was nearly identical to that reported . This compensation results from their thermodynamic equivalence with some important kinetic differences . To achieve the low ΔpH values reported experimentally , the volume-dependent mKHE expression had to be replaced with a high activity K+/H+ exchanger . It is postulated that the Pi-depletion dramatically altered the proton permeability . This is manifested in the model by essentially creating a high activity rapid equilibrium exchanger mechanism that is not supported by the volume-dependent mKHE expression . However , this high activity K+/H+ exchanger is not compatible with the volume dynamics presented in the Kowaltowski data set , so it is only used for the Bose data set simulations since their isolation procedure was done without Pi . Although we chose to address this discrepancy using a high K+/H+ exchanger , there are several alternative mechanisms that could also potentially describe or contribute to the ΔpH discrepancy . For example , the mNHE could be responsible for converting more of the ΔpH into the Δψ than the model predicted . It is possible the mNHE activity is underestimated; however , the models ability to match the reported steady state matrix free calcium concentration at varied buffer sodium concentrations shown in Figure 5 indicate this is not the case . Alternatively , the differences between porcine and rat heart mitochondria may be responsible for the dramatic change in proton permeability . The Bose data set was obtained using porcine heart mitochondria , while rat heart mitochondria were used for all the other datasets . It is not likely that the proton permeability associated with these two species is significantly different . Conversely , the volume-dependent mKHE expression may not sufficiently capture the phenomena . The model fits and corroboration simulations presented in Figure 6 and Figure 8 , respectively , refute this conclusion . From this discussion , it is evident that more experimental data measuring ΔpH and Δψ under various conditions is needed to reduce this uncertainty . To simulate the precise experimental conditions during model development , a few explicit assumptions were necessary . For example , the LaNoue data set reported using 3–4 mg of mitochondrial protein; however , as little as a 25% change in mitochondrial load can dramatically alter the total substrate consumption and product accumulation during high MVO2 rates . Hence for these simulations , the conditions needed to be known with more certainty . The reported malate concentration was used to estimate the mitochondrial load . The initial malate content in state 2 and state 3 experiments with pyruvate was reported to be 1430 nmol/mg . This required that the mitochondrial load be 3 . 5 mg/mL using the stated 5 mM malate concentration . To compute this estimate , it was necessary to also consider the pyruvate concentration . The pyruvate concentration was 2 mM; however after 8 minutes of state 3 respiration , the pyruvate utilization was 848 nmol/mg . Considering the 1 mL chamber volume , this implied that the total initial pyruvate concentration was 2 . 5–3 . 4 mM and not 2 mM . In an attempt to address these potential data inconsistencies , the model simulations were fit to the data using a mitochondrial load of 3 . 5 mg/mL and the reported state 3 pyruvate utilizations were subsequently adjusted to be consistent with an initial pyruvate concentration of 2 mM . All mathematical models are abstractions of the underlying process; the level of detail included in the model is dependent upon the application . This is particularly true for the calcium dynamics associated with mitochondrial bioenergetics . There are known omissions in this and previous models of these calcium dynamics . The mitochondrial Na+/Ca2+ dynamics were simulated using a simplified Na+/Ca2+ cycling mechanisms with only the CaUNI , mNCE and mNHE processes represented . This simplification prohibited a mechanistic representation of the actual physiological event . For example , the omission of the rapid mode of calcium uptake ( RAM ) [49] process necessitated a high CaUNI influx to reach the steady state calcium measurements ( from the Wan data set ) within a few minutes of extra-mitochondrial calcium addition . ( Note , with this higher calcium influx rates , the model still predicts the Na+/Ca2+ cycle consumes less than 1% of the proton electrochemical gradient established by the ETS . ) Also , the Na+-independent calcium efflux mechanism is not included in the model formulation since the underlying process is uncertain ( electroneutral or electrogenic [50]–[52] ) even though it is estimated to contribute up to 33% of total calcium efflux in heart tissue [53] . This omitted calcium efflux mechanism is insensitive to magnesium and prevented adequate fits to the Mg2+-titration data presented in Wan et al . ; however , Mg2+-dependence of the CaUNI and mNCE did enable the simulations to reproduce the reported steady state matrix calcium levels within a few 100 nM ( not shown ) . An explicit , detailed study of the Na+/Ca2+ dynamics should be modeled under various experimental conditions to fully characterize and understand this process at a more mechanistic level . In summary , the model presented in this manuscript proposes an extended mitochondrial bioenergetics model targeted at the cardiac myocyte with the parameters estimated using four independent data sets consisting of 32 data curves . It was capable of fitting the data with good fidelity , had relatively little parameter sensitivity relative to the experimental conditions modeled herein and was capable of adequately modeling metabolic trends during the various conditions simulated . The resulting model simulations reproduce observed mitochondrial volume dynamics lending additional support to the current prevailing theory of mitochondrial volume regulation through the mKHE volume-sensitive exchange rate . The model builds upon previous successes and helps refine and establish a global model framework relating to mitochondrial bioenergetics . During the model development , a certain network topology had to be in place and some assumptions about uncertain or unobserved experimental factors and conditions were explicitly constrained to reproduce all the data sets . Specifically , the effect of intrinsic thermodynamic dissipation of the system on enzymatic regulation , importance of animal species and tissue sources differences , mechanistic detail of the model and potential impact of the experimental environment all help constrain the model formulation contributing to the construction of a successful and physiologically faithful model . The model can serve as a foundation for further extension and refinement efforts . Future work may consider more detailed and mechanistic mathematical abstractions for the ETS and TCC , Ca2+ dynamics including the RAM and Na+-independent Ca2+ efflux pathways , catabolic ( i . e . , glutamate dehydrogenase ) and anabolic ( i . e . , pyruvate carboxyalse ) reactions and β-oxidation pathways enabling integration into whole cell models of cardiac myocytes . Each of these additions will require additional experimental data taken under well controlled and documented conditions in order to be properly reproduced by the model proposed in this manuscript . For example , changing the passive exchange mechanism of the TCC to a more mechanistic , saturable exchange process should enable better fits to the LaNoue data set . This change would keep matrix citrate and isocitrate at sufficient levels to maintain α-ketoglutarate and succinate concentrations experimentally observed allowing fumarate to be included in the list of DCC substrates . With fumarate being removed from the matrix by the DCC , SDH inhibition would be mitigated and the lower branch of the TCA cycle would accelerate and prevent net oxidation of malate observed in the LaNoue experimental data set . Additionally , reproducing the respiratory control ratios as done in silico by Korzeneiwski and Mazat [54] using the experimentally measured respiratory control ratios measure by Rossignol et al . [55] would help constrain , define and corroborate the mathematical abstractions for the ETS , F1FO , ANT and PYRH mechanisms . Experiment design with this model could further reduce parameter uncertainties and help test alternative hypotheses including some postulates made in the Discussion . Although much work is ahead , we feel that this model takes a step towards a more complete physiologically faithful mitochondrial bioenergetics model .
The DAEs describing the model were numerically integrated using MATLAB® ( 2008b ) and the stiff ode solver ode15s ( 10−3 relative tolerance and 10−9 absolute tolerance for matrix and IMS state variables and 10−6 for all others ) . To increase computational efficiency , vectorized functions were used during model development in the MATLAB® environment . Parameter optimizations and sensitivity analyses were done on a cluster of four 8-core Intel Xeon 3 . 4 GHz CPUs each with 16 GB of memory and running the Windows 2003 server platform using the Parallel Computing Toolbox . The results obtained were displayed using MATLAB® . The objective function used for parameterization of the model is defined as ( 1 ) where f is the objective function evaluated at a given parameter point p , yi , j , k is the model output , either a state variable or computed rate , corresponding to the ith experimental data point in the jth experimental data curve for the kth data set evaluated at the parameter point p , Yi , j , k is the ith experimental data point in the jth experimental data curve for the kth data set , σi , j , k is the standard deviation for the ith experimental data point in the jth experimental data curve for the kth data set , Nj , k is the number of data points in the jth experimental data curve for the kth data set and Mk is the number of data curves for the kth data set . When no statistical data were given with the experimental data , a 5–10% relative error was assumed . Fitting such large , non-linear models to data with many unknown parameters and initial conditions requires a robust model structure and many independent data sets to appropriately constrain the parameters . The model presented in this manuscript consists of a total of 359 parameters . These parameters were identified using three methods: i ) 262 parameters were fixed according to previously published values in the literature ( see Part S3 of the Supplemental Material ( Text S1 ) ) , ii ) 55 parameters were found by minimizing the sum of the squares of the difference between simulated rate expressions and the published data using Equation 1 ( see Part S3 of the Supplemental Material ( Text S1 ) ) , and iii ) a custom parallelizable Monte Carlo optimization algorithm based on simulated annealing was used to fit the remaining 42 parameters by minimizing Equation 1 with replicated experiments [8]–[11] . Global candidate points were first identified using the simulated annealing approach and refined with a local search using a gradient-based algorithm ( MATLAB®'s fmincon function ) . The local sensitivity analysis was done using the absolute normalized local sensitivity coefficient ( LSC ) for quantifying how much the model output trajectories for the simulated experimental conditions changed in response to small perturbations about the identified parameter set . This LSC is defined as ( 2 ) where LSC is the normalized local sensitivity coefficient for the ℓth parameter , yi , j , k is the model output of the ith state at the jth time point for the kth experimental condition evaluated at the parameter point p , pℓ is the parameter whose sensitivity is being approximated , Nj , k is the number of states considered for the analysis ( note , all 73 states were included in the sensitivity analysis ) at the jth time point for the kth experimental condition , Mk is the time points used in the analysis ( note , Mk was fixed and defined as 5 equally spaced time points ) for the kth experimental condition and H is the total number of experimental conditions replicated from [8]–[11] . ( Note that although the indices used for the LSC computation are similarly defined , they have different meaning . ) Equation 2 was approximated using a centered finite difference using the numerical methods outlined in Conn et al . [56] . To minimize numerical artifacts when computing the LSCs , the model was integrated with stricter tolerances ( 10−9 relative tolerance and 10−12 absolute tolerance tolerance for matrix and IMS state variables and 10−9 for all others ) . This unit of measure is used in Table 2 to discern the 1st order , one-at-a-time , effects of small parameter perturbations . A parameter that possesses a large LSC is interpreted as having a substantial influence on the model state trajectories and steady state values . To simulate the various data sets used to parameterize and corroborate the model , the appropriate experimental conditions were taken into consideration . The temperature at which the experimental data were obtained , the mitochondrial loads applied in each experiment , the initial state of the mitochondria in the experimental system and the precise nature of the experimental environment , specifically the buffer composition and osmolarity were considered . These points are discussed below . | Mathematically modeling biological systems challenges our current understanding of the physical and biochemical events contributing to the observed dynamics . It requires careful consideration of hypothesized mechanisms , model development assumptions and details regarding the experimental conditions . We have adopted a modeling approach to translate these factors that explicitly considers the thermodynamic constraints , biochemical states and reaction mechanisms during model development . Such models have numerous constant parameters that must be determined . Integrating thermodynamics and detailed mechanistic representation of the principal phenomena help constrain these parameter values; therefore , only a handful of the total number of model parameters ( ∼10% ) must be adjusted during parameter estimation through model simulations . Additionally , all models must undergo some form of corroboration prior to application . In practice , this corroboration should challenge all possible dynamics of the model , but it is recognized that in this data rich world , we are surprisingly data poor . Eventually such developed and corroborated models are capable of supporting current hypotheses , guiding experimental designs and contributing to the overall knowledge base of biological processes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/systems",
"biology",
"biophysics/theory",
"and",
"simulation",
"biochemistry/theory",
"and",
"simulation"
] | 2010 | Modeling Mitochondrial Bioenergetics with Integrated Volume Dynamics |
Many eukaryotic cell-surface proteins are anchored to the membrane via glycosylphosphatidylinositol ( GPI ) . There are at least 26 genes involved in biosynthesis and remodeling of GPI anchors . Hypomorphic coding mutations in seven of these genes have been reported to cause decreased expression of GPI anchored proteins ( GPI-APs ) on the cell surface and to cause autosomal-recessive forms of intellectual disability ( ARID ) . We performed homozygosity mapping and exome sequencing in a family with encephalopathy and non-specific ARID and identified a homozygous 3 bp deletion ( p . Leu197del ) in the GPI remodeling gene PGAP1 . PGAP1 was not described in association with a human phenotype before . PGAP1 is a deacylase that removes an acyl-chain from the inositol of GPI anchors in the endoplasmic reticulum immediately after attachment of GPI to proteins . In silico prediction and molecular modeling strongly suggested a pathogenic effect of the identified deletion . The expression levels of GPI-APs on B lymphoblastoid cells derived from an affected person were normal . However , when those cells were incubated with phosphatidylinositol-specific phospholipase C ( PI-PLC ) , GPI-APs were cleaved and released from B lymphoblastoid cells from healthy individuals whereas GPI-APs on the cells from the affected person were totally resistant . Transfection with wild type PGAP1 cDNA restored the PI-PLC sensitivity . These results indicate that GPI-APs were expressed with abnormal GPI structure due to a null mutation in the remodeling gene PGAP1 . Our results add PGAP1 to the growing list of GPI abnormalities and indicate that not only the cell surface expression levels of GPI-APs but also the fine structure of GPI-anchors is important for the normal neurological development .
Many eukaryotic cell-surface proteins with various functions are anchored to the membrane via glycosylphosphatidylinositol ( GPI ) [1]–[3] . After biosynthesis in the endoplasmic reticulum ( ER ) , GPI-anchors are transferred to the proteins by the GPI transamidase and the structure of the GPI-anchor is then remodeled , which is critical for sorting , regulating and trafficking of the GPI anchored proteins ( GPI-APs ) [3] . This remodeling starts in the ER by eliminating the acyl-chain linked to the inositol in the GPI-anchor by PGAP1 [4] , then a side-chain of ethanolamine-phosphate on the second mannose of the GPI-anchor is removed by MPPE1 ( PGAP5 ) [5] . GPI-APs are then transported from the ER to the plasma membrane through the Golgi apparatus , where further remodeling by PGAP3 and PGAP2 takes place [6] , [7] . Germline mutations in eight genes that are involved in the GPI-anchor biosynthesis and remodeling have been described ( Table 1 ) [8]–[22] . The mutations in all of those , PIGA , PIGL , PIGM , PIGV , PIGN , PIGO , PIGT and PGAP2 , are hypomorphic and lead to partially decreased cell surface expression of various GPI-APs , thus causing a wide phenotypic spectrum ranging from syndromic disorders with various malformations to non-specific forms of intellectual disability . The reported mutations in genes of early steps of the GPI-anchor synthesis such as PIGA ( MIM 311770 ) , PIGL ( MIM 605947 ) , and PIGM ( MIM *610273 ) , or in a gene involved in GPI transfer to proteins such as PIGT ( MIM *610272 ) are supposed to result in a degradation of precursor non-GPI-anchored proteins by ER associated degradation , whereas mutations in genes that are involved in later steps of the pathway , such as PIGV ( MIM *610274 ) , PIGO ( MIM *614730 ) , and PGAP2 ( MIM *615187 ) result in partial secretion of non-GPI-anchored proteins such as alkaline phosphatase ( in case of PIGV or PIGO deficiency ) [23] or of proteins bearing cleaved GPI-anchor ( in case of PGAP2 deficiency ) , and are therefore characterized by hyperphosphatasia . Here we report on the identification of a mutation in PGAP1 that encodes the GPI inositol-deacylase [4] . This leads to a new type of GPI-anchor deficiency manifesting non-specific autosomal recessive intellectual disability ( ARID ) , in which cell surface levels of GPI-APs are not affected whereas the structure of GPI moiety is abnormal .
We undertook clinical characterization , mapping [24] and exome sequencing in a large cohort of families with non-specific ARID . We identified the PGAP1 mutation in the Syrian family MR079 . The parents in family MR079 are the first-degree cousins and the family has one healthy girl and two affected children that carry the mutation in a homozygous status . The affected girl ( III-2 ) was 4 years and 5 months old and the affected boy ( III-3 ) was 2 years and 9 months old at the time of examination ( Figure 1 ) . Pregnancy , delivery , and birth parameters of both children were unremarkable . In the neonatal period , III-2 was hypotonic and III-3 was a floppy baby . Motor development was delayed; III-2 could sit at age of 18 months and at age of 45/12 years first tried to walk independently . At age of 29/12 , III-3 could only roll from back to stomach and back . Both children did not finish potty training and were still partially fed with milk bottles . Both children have a developmental delay and severe intellectual disability with an estimated IQ below 35 . III-2 could only babble a few syllables . While III-2 had major and absence epilepsy , III-3 did not yet have seizures . Sleeping patterns of both children were normal . They showed some stereotypic movements such as hitting on their own mouth and some washing movements of the hands . Both children seemed to see and hear properly , but specific tests could not be done . Brain CT scan of III-2 at age of one year revealed pronounced brain atrophy . At the time of examination , III-2 was 96 cm tall ( 25th percentile ) with a head circumference of 46 cm ( 2 cm below the 5th percentile ) . III-3 was also of normal height and had a head circumference of 47 cm ( 1 . 5 cm below the 5th percentile ) . Their parents had head circumferences of 52 and 53 cm , also in the lower percentiles . Both children have large ears and a flattened nasal root . G-banding , cytogenetic examination and genome wide copy number variants analyses were unremarkable . We did not have information on the levels of alkaline phosphatase and it was not possible to obtain blood probes retrospectively . Autozygosity mapping [24] in family MR079 led to the identification of six candidate regions of a total length of 64 Mb . Subsequently , exome sequencing using DNA from individual III-3 was performed as described in former studies [21] , [25] resulting in an average coverage of 53 . 28 . 66% of the target sequences were covered with a depth of at least 20× , and 80 . 51% were covered with a depth of at least 5× . A total of 42 , 352 SNVs and 2 , 529 indels were identified . 342 SNVs and 64 indels were neither annotated , nor reported in 1000Genomes and Exome Variant Server , nor in in-house controls , and may affect the protein sequence ( non-synonymous , splicing , or UTR ) . Of those , only two , in PGAP1 and SLC40A , were located in a candidate region , conserved , and predicted to be pathogenic by in silico programs . To exclude further candidate mutations , we repeated the exome sequencing using DNAs of both affected siblings . We enriched the exome using a PCR based targeting method ( Ion AmpliSeq Exome Kit ) and sequenced on the Ion Proton . The average coverage of III-3 and III-2 was 149 . 6× and 94 . 6× , respectively . 91 . 1% and 85 . 0% of the target sequences were covered with a depth of at least 20× , 96 . 3% and 93 . 4% with a depth of at least 5× , respectively . A total of 49 , 455 and 47 , 693 SNVs as well as 3 , 343 and 3 , 167 indels were identified . When applying the above mentioned filtering steps , we were by both affected children once again left with the variants in PGAP1 and SLC40A . Since mutations in SLC40A cause hemochromatosis of type 4 and have no effect on cognition ( MIM 606069 ) [26] , [27] , we focused on the variant in PGAP1 , NM_024989 . 3:c . 589_591delCTT , NP_079265 . 2:p . Leu197del . Genotyping the variant in PGAP1 in 372 healthy Syrian adults using Sanger sequencing revealed no further carriers . Taking the minor allele frequency of 0 in the Exome Sequencing Project ( ESP ) data set and in our control sample of 372 healthy Syrian individuals , it seems that the mutation has prevalence far less than 0 . 001 . Molecular modeling using the GeneSilico fold recognition metaserver [28] and Modeler9 . 9 [29] using the closest related hydrolase ( PDB code: 3LP5 ) as template highlighted the detrimental effect of the deletion of leucine 197 on the structure of PGAP1 . Leucine 197 is located in the central strand of a β-sheet and is oriented towards the hydrophobic core of the enzyme where it forms multiple stabilizing interactions with the adjacent helices ( Figure 2A , B ) . Deletion of this amino acid would place Ile198 at the position originally occupied by Leu197 ( Figure 2C ) . The Cβ-branched side-chain of isoleucine cannot be accommodated at this sequence position resulting in several clashes with adjacent amino acids ( Leu184 , Ile194 ) of the hydrophobic core ( Figure 2C ) . This will disrupt the packing of the hydrophobic core and consequently of the entire β-sheet topology , thus leading to a loss of tertiary structure and enzymatic activity . We then ran large scale homozygosity mapping using PLINK in our sample of over 100 consanguineous families [24] and over 600 sporadic cases of ID [30] and identified 7 index patients , 2 from consanguineous families with multiple affected children and 5 from outbreed families with single affected patients , that are homozygous at the PGAP1 . Sequencing all seven individuals using Sanger did not reveal any mutations in PGAP1 . We then screened the exome variant server for functional variants in PGAP1 . 149 variants are reported in this gene , of those 44 were coding or at splice sites . All of those are extremely rare ( 0 . 0077%–0 . 569% , i . e . 1–74 alleles out of ca . 13000 alleles ) . Based on the conservation of the variants and the prediction of in silico programs ( Table S1 ) , we roughly estimate that a maximum of 48 individuals may carry a mutation in PGAP1 ( carrier rate of 48/6500 = 0 . 0073 ) and that the prevalence of the disease would be about 13 per million . If we take more conservative in silico prediction numbers , the prevalence of the disease would be 7 per million inhabitants ( Table S1 ) . The two most frequent variants in the ESP data were p . Lys111Glu and p . Gln585Glu and were observed in a heterozygous form 15 and 74 times out of 12992 and 12932 alleles , respectively . Both sites are well conserved in the mammalian . Molecular modeling showed that the most common variant Gln585Glu is located outside of catalytic active domains and it was not possible to make a prediction for this variant . Lys111Glu is at the C terminus of a helix of the deacylase domain . The charging pattern of the helix is highly conserved so that we expect that the change from Lys to Glu would change the charge of the protein and destabilize the helix . To determine effects of p . Leu197del alteration on cellular GPI-APs , we investigated the surface expression of GPI-APs on B-lymphoblastoid cell lines ( LCLs ) derived from the homozygous individual III-3 ( −/− ) , 2 heterozygous parents ( +/− ) , and the healthy sister ( +/+ ) ( Figure 3 ) , as well as 6 healthy volunteers with a confirmed wild type genotype ( data not shown ) . Using flow cytometry analysis , the respective surface expressions of CD59 , CD55/DAF , and CD48 were quantified . Surface expression of these GPI-APs on LCLs from an affected person , other family members or healthy volunteers showed no significant difference , indicating that the PGAP1 mutation did not affect the surface expression levels of various GPI-APs ( Figure 3A , dotted lines ) . The surface expression of the GPI anchor itself was quantified using fluorochrome conjugated aerolysin ( FLAER , Pinewood Scientific ) , a bacterial toxin that specifically binds GPI anchors , and did not show significant differences between the affected individual , the heterozygous individuals , and the controls ( data not shown ) . We then investigated the expected structural abnormality of GPI-anchors by testing sensitivity of GPI-APs to phosphatidylinositol-specific phospholipase C ( PI-PLC ) [31] . The LCLs were incubated with 10 unit/ml of PI-PLC for 1 . 5 h at 37°C and the remaining surface GPI-APs were determined by flow cytometry . Of GPI-APs , 61% to 90% were removed from the surface of LCLs of the healthy sister with a homozygous wildtype ( Figure 3A , solid line ) and healthy control individuals ( data not shown ) . In contrast , no significant or only slight reduction of the surface GPI-APs was seen with LCLs from the affected person ( Figure 3A ) , indicating that almost all GPI-APs on the affected LCLs had abnormal GPI anchors resistant to PI-PLC [4] . This is a strong indication that the p . Leu197del mutation causes null or almost null activity of the PGAP1 enzyme . GPI-APs on LCLs from heterozygous parents were only partially sensitive to PI-PLC ( Figure 3A ) , indicating that the p . Leu197del mutation causes haplo-insufficiency . These defective sensitivities of affected the person's and parents' GPI-APs to PI-PLC were fully restored by transfection of wild-type PGAP1 cDNA ( Figure 3B , solid lines ) . Finally , the functional effect of the p . Leu197del mutation was tested in the PGAP1 deficient Chinese hamster ovary ( CHO ) cell system [4] . GPI-APs expressed on the PGAP1 deficient CHO cells are resistant to PI-PLC and the activity of PGAP1 cDNA can be assessed by its ability to make PI-PLC-sensitive GPI-APs after transfection . CHO cells defective for PGAP1 were transiently transfected with N-terminally-FLAG-tagged wild-type and p . Leu197del mutant human PGAP1 cDNA in an expression vector with a strong SRα promoter , or an empty vector . Four days after transfection , each transfectant was treated with or without PI-PLC , and the surface expression of CD59 , DAF and urokinase plasminogen activator receptor ( uPAR ) were assessed by flow cytometry . The wild-type PGAP1 cDNA rescued PI-PLC sensitivity ( Figure 4A , left panels ) . In contrast , the transfection of the mutant p . Leu197del cDNA did not increase the sensitivity to PI-PLC , thus indicating functional loss of the mutant PGAP1 cDNA ( Figure 4A , center panels ) . To determine PGAP1 protein levels , lysates were prepared two days after transfection , immunoprecipitated with anti-FLAG beads and analyzed by SDS-PAGE/Western blotting . The p . Leu197del mutant protein was not detected at all , indicating that the deletion of Leu197 caused an unstable protein ( Figure 4B ) . In order to evaluate other known variants in PGAP1 , we screened the public database of ESP ( see above ) . Of listed variants , we chose the two most frequent variants: rs142320636: c . 331A>G ( p . Lys111Glu ) and rs62185645: c . 1753C>G ( p . Gln585Glu ) , and tested the functional effect of these mutations in the PGAP1 deficient Chinese hamster ovary ( CHO ) cell system . Transfection of the mutant p . Lys111Glu cDNA did not increase the sensitivity to PI-PLC , indicating functional loss of the mutant PGAP1 cDNA . Mutant p . Gln585Glu showed an activity comparable to the wild type PGAP1 ( Figure S1 ) . Thus , it is possible that homozygosity of p . Lys111Glu leads to ARID .
Eight GPI deficiencies caused by hypomorphic mutations in the coding regions of GPI biosynthesis genes PIGM , PIGA , PIGL , PIGV , PIGN , PIGO , PIGT , and PGAP2 have been reported . Except PIGM , all lead to a decreased surface expression of GPI-APs and result in intellectual disability , often associated with epilepsy , distinct facial characteristics , and further organ malformations [9]–[22] . We showed here that complete PGAP1 deficiency did not affect the surface expression of GPI-APs but expressed structurally abnormal GPI-APs with the acylated inositol . In previous works , we have reported that Pgap1 knock-out mice had otocephaly , male infertility , growth retardation , and often died right after birth [32] . Also further two mutant mouse strains , otoxray ( oto for otocephaly ) [33] , [34] and beaker [35] were reported to have disrupted Pgap1 . Both mice strains showed developmental abnormalities of the forebrain; the recessive lethal otoxray showed a truncation of the forbrain and the breaker mutant displayed a holoprosencephaly-like phenotype . Both Wnt signaling and Nodal signaling were reported to be affected in these mutant mice . These data emphasize the importance of PGAP1 for vital functions and for brain development . It was also indicated that the Pgap1 mutant mice phenotypes are dependent upon the genetic background since otocephaly and holoprosencephaly are not seen in some mouse strains [34] , [35] . Based on our mapping results , exome sequencing data and functional experiments that proved pathogenicity of the mutation , the previous reports on intellectual disability caused by mutations in the GPI synthesis pathway , and the mouse models that clearly show an association between the disruption of Pgap1 and abnormalities of brain , we consider the deletion of leucine197 to be causative for the severe non-specific autosomal recessive intellectual disability in our examined patients of family MR079 . PGAP1 is the ninth gene of the GPI synthesis pathway that is now associated to a human phenotype ( Table 1 ) . Further mutations in PGPA1 are needed to confirm our findings . Also , describing further patients with different mutations is necessary to delineate the phenotypes of the GPI deficiencies . For example , considering the defect in the modification of the GPI anchors , the alkaline phosphatase would not be elevated in patients with PGAP1 mutations , but this needs to be confirmed . In conclusion , null mutations in PGAP1 lead to severe intellectual disability and encephalopathy with no obvious malformations; we add PGAP1 to the growing number of genes involved in GPI-anchor deficiencies with human phenotypes . PGAP1 deficiency causes a defect in the ER part of the GPI-AP biosynthesis that involves the remodeling of the anchors after attachment to proteins , and it leads to normal protein expression on the cell surface but to abnormal anchor structure .
Genomic DNA was extracted from EDTA blood probes by standard methods and genotyped with the Affymetrix Mapping array 6 . 0 ( Affymetrix , Santa Clara , CA , USA ) . Analysis did not reveal pathogenic deletions or duplications . Mendelian segregation was calculated using PedCheck software and was confirmed in all instances . Autozygosity mapping was performed using HomozygosityMapper [36] . DNA from individual III-3 was enriched using the SureSelect Human All Exon Kit , which targets approximately 50 Mb of human genome ( Agilent , Santa Clara , Ca , USA ) and paired-end sequenced on a SOLiD 5500 xl instrument ( Life Sciences , Carlsbad , CA , U . S . A . ) . Image analysis and base calling was performed using the SOLiD instrument control software with default parameters . Read alignment was performed with LifeScope 2 . 5 using the default parameters with human genome assembly hg19 ( GRCh37 ) as reference . Single-nucleotide variants and small insertions and deletions ( indels ) were detected using LifeScope , GATK 2 and samtools/bcftools [37] , [38] . To replicate the results , DNA from individuals III-2 and III-3 was amplified using the Ion AmpliSeq Exome Kit ( Life Technologies , Carlsbad , CA , U . S . A . ) which targets approximately 58 Mb of the human genome . After quality control on the Bioanalyzer High Sensitivity Chip ( Agilent , Santa Clara , Ca , USA ) and emulsion PCR ( Ion PI Template OT2 200 Kit v3 , Life Technologies , Carlsbad , CA , U . S . A . ) the samples were sequenced on a Proton PI chip Version 2 ( Life Technologies , Carlsbad , CA , U . S . A . ) . Base calling , pre-processing of the reads , short read alignment and variant calling was performed using the Torrent Suite including the Torrent Variant Caller ( TVC , Version 4 . 0 ) with default parameters recommended for the Ampliseq Exome panel ( low stringency calling of germline variants , Version September 2013 ) . Variant annotation was performed using Annovar , integrating data from a variety of public databases [39] , [40] . Additionally , variants were compared to an in-house database containing more than 350 sequenced exomes to identify further common variants which are not present in public databases . Finally , the variants were validated by PCR and Sanger sequencing according to the standard protocols to exclude technical artifacts and to test for segregation . Heparin blood samples were collected from one affected and from all unaffected siblings and parents . Lymphoblastoid Cell lines ( LCLs ) were generated and cultured in RPMI 1640 ( Gibco , Life technologies , Darmstadt , Germany ) that is supplemented with 10% FCS ( PAA Biotech , Cölbe , Germany ) and different other supplements . LCLs from one of the affected siblings ( III-3 ) and the parents were transfected with empty pMEoriP vector or pMEoriP-FLAG-humanPGAP1 . Cells from healthy sister were used without transfection . Cells ( 5×106 ) were suspended in 0 . 8 ml of Opti-MEM and electroporated with 20 µg each of the plasmids at 260 V and 960 µF using a Gene Pulser ( Bio Rad , Hercules , CA ) . Four days after transfection , cells were treated with or without 10 unit/ml of PI-PLC ( Molecular probes , Eugene , OR ) for 1 . 5 h at 37°C . Surface expression of GPI-APs was determined by staining cells with mouse anti-human CD59 ( 5H8 ) , -human DAF ( IA10 ) , -human CD48 ( BJ40 ) antibodies and each isotype IgG followed by a PE-conjugated anti-mouse IgG antibody ( BJ40 , mouse IgG1 and IgG2a , and secondary antibody were purchased from BD Biosciences , Franklin Lakes , NJ ) and analyzed by flow cytometer ( Cant II; BD Biosciences ) using Flowjo software ( Tommy Digital Inc . , Tokyo , Japan ) . pMEFLAG-hPGAP1 mutant ( L197del ) bearing patient's mutation was generated by site directed mutagenesis . PGAP1 deficient CHO cell ( C10 ) [4] were transiently transfected with wild type or mutant pMEFLAG-hPGAP1 by electroporation . Cells ( 107 ) were suspended in 0 . 4 ml of Opti-MEM and electroporated with 20 µg each of the plasmids at 260 V and 960 µF using a Gene Pulser . Four days after transfection , cells were treated with or without 10 unit/ml of PI-PLC for 1 . 5 h at 37°C . Surface expression of GPI-APs was determined by staining cells with mouse anti-human CD59 ( 5H8 ) , -human DAF ( IA10 ) , -hamster uPAR ( 5D6 ) antibodies and each isotype IgG , followed by a PE-conjugated anti-mouse IgG antibody and analyzed by flow cytometer using Flowjo software . Two days after transfection of each PGAP1 construct , lysates were immunoprecipitated with anti-FLAG beads and analyzed by SDS-PAGE/Western blotting . 1000Genomes , http://www . 1000genomes . org/ ABI , L . T . ( 2012 ) . LifeScope . : http://www . lifetechnologies . com/lifescope . ANNOVAR: http://www . openbioinformatics . org/annovar/ GeneTalk: http://www . gene-talk . de BWA , Burrows-Wheeler Aligner; http://bio-bwa . sourceforge . net/ dbSNP , NCBI: http://www . ncbi . nlm . nih . gov/snp/ GATK 2 , Genome Analysis Toolkit: http://www . broadinstitute . org/gatk/index . php Kyoto Encyclopedia of Genes and Genomes , KEGG , http://www . genome . jp/kegg/ MutationTaster: http://www . mutationtaster . org/ELAND , alignment algorithm , Illumina . com NHLBI Exome Sequencing Project ( ESP ) : http://evs . gs . washington . edu/EVS/ Online Mendelian Inheritance in Man ( OMIM ) : http://www . omim . org PolyPhen2: http://genetics . bwh . harvard . edu/pph2/ SIFT: http://sift . jcvi . org/ UCSC Genome Browser: www . genome . ucsc . edu | Glycosylphosphatidylinositols ( GPI ) are glycolipid anchors that anchor various proteins to the cell surface . At least 26 genes are involved in biosynthesis and modification of the GPI anchors . Recently , mutations in eight of those genes have been described . Although those mutations do not fully abolish the functions of encoded enzymes , they lead to a decreased expression of surface GPI-anchored proteins and to different forms of intellectual disability . Here we report a mutation in PGAP1 that encodes a protein that modifies the GPI anchor . We found that the mutation leads to a full loss of PGAP1 enzyme activity , but that the patient cells still express normal levels of surface GPI-anchored proteins . However , the GPI anchors have an abnormal lipid structure that is resistant to cleavage by phosphatidylinositol-specific phospholipase C . Our results add PGAP1 to the growing list of GPI abnormalities that cause intellectual disability and indicate that the fine structure of GPI-anchors is also important for a normal neurological development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"epilepsy",
"medicine",
"and",
"health",
"sciences",
"child",
"growth",
"genetic",
"dominance",
"growth",
"restriction",
"child",
"development",
"autosomal",
"recessive",
"traits",
"neurology",
"neurodegenerative",
"diseases",
"genetics",
"biology",
"and",
"life",
"scie... | 2014 | Null Mutation in PGAP1 Impairing Gpi-Anchor Maturation in Patients with Intellectual Disability and Encephalopathy |
Parkinson's disease ( PD ) , the second most prevalent neurodegenerative disease after Alzheimer's disease , is linked to the gradual loss of dopaminergic neurons in the substantia nigra . Disease loci causing hereditary forms of PD are known , but most cases are attributable to a combination of genetic and environmental risk factors . Increased incidence of PD is associated with rural living and pesticide exposure , and dopaminergic neurodegeneration can be triggered by neurotoxins such as 6-hydroxydopamine ( 6-OHDA ) . In C . elegans , this drug is taken up by the presynaptic dopamine reuptake transporter ( DAT-1 ) and causes selective death of the eight dopaminergic neurons of the adult hermaphrodite . Using a forward genetic approach to find genes that protect against 6-OHDA-mediated neurodegeneration , we identified tsp-17 , which encodes a member of the tetraspanin family of membrane proteins . We show that TSP-17 is expressed in dopaminergic neurons and provide genetic , pharmacological and biochemical evidence that it inhibits DAT-1 , thus leading to increased 6-OHDA uptake in tsp-17 loss-of-function mutants . TSP-17 also protects against toxicity conferred by excessive intracellular dopamine . We provide genetic and biochemical evidence that TSP-17 acts partly via the DOP-2 dopamine receptor to negatively regulate DAT-1 . tsp-17 mutants also have subtle behavioral phenotypes , some of which are conferred by aberrant dopamine signaling . Incubating mutant worms in liquid medium leads to swimming-induced paralysis . In the L1 larval stage , this phenotype is linked to lethality and cannot be rescued by a dop-3 null mutant . In contrast , mild paralysis occurring in the L4 larval stage is suppressed by dop-3 , suggesting defects in dopaminergic signaling . In summary , we show that TSP-17 protects against neurodegeneration and has a role in modulating behaviors linked to dopamine signaling .
Parkinson's Disease ( PD ) is the second most common neurodegenerative disease , after Alzheimer's disease , and affects ∼2% of the population aged over 65 years . Loss of dopaminergic neurons is a pathological hallmark of PD [1] , [2] and aspects of this neurodegeneration have been modeled in C . elegans [3] , [4] . The etiology of PD is largely unknown and its heritability is generally rather low; however ∼5–10% of cases are associated with monogenetically inherited mutations [5] . Approximately 15 disease loci are known , most of which are conserved in C . elegans [6] , [7] . The vast majority of PD cases are ‘sporadic’ with no clear family history . Besides aging , epidemiological studies have shown risk factors for ‘sporadic’ PD to include a long-term history of rural living , farming , well-water drinking and pesticide exposure . The most extreme examples of toxin-induced PD-like symptoms were linked to the accidental exposure to MPTP ( N-methyl-4-phenyl-1 , 2 , 3 , 6-tetrahydropyridine ) . Similar to sporadic PD cases , PD-like symptoms resulting from MPTP exposure could be alleviated by administration of the dopamine precursor L-3 , 4-dihydrooxyphenylalanine ( L-DOPA ) [8] . Exposure to pesticides such as paraquat and rotenone has also been implicated in PD development [9] . The disease is therefore thought to be triggered by a combination of environmental factors and genetic susceptibility [5] . MPTP , paraquat and rotenone all block the mitochondrial electron transport chain , leading to oxidative damage [10] , and have been extensively used to model PD neurodegeneration . 6-Hydroxydopamine ( 6-OHDA ) , an oxidation product of dopamine , is another neurotoxin widely used in mammalian PD models to induce the specific degeneration of dopaminergic neurons [11] . 6-OHDA was initially identified as a metabolite of dopamine [12] , and there is some evidence that 6-OHDA exposure might be linked to PD . 6-OHDA was also identified as a naturally occurring amine in human urine , and has been detected at higher concentrations in PD patients [13] . Furthermore , high 6-OHDA levels were found in postmortem brain samples from PD patients [14] . It has been reported that 6-OHDA interaction with oxygen results in the production of reactive oxygen species ( ROS ) , which in turn trigger free radical-mediated neuronal degeneration [2] , [12] . Other dopamine metabolites may also cause oxidative damage [15] . Nevertheless , the mechanism by which 6-OHDA induces neuronal degeneration remains largely unknown [16] . Although there is no treatment to prevent or halt neuronal loss , L-DOPA administration is still one of the most effective treatments for alleviating PD symptoms [17] , [18] . However , the effectiveness of L-DOPA declines over time . Prolonged L-DOPA treatment is also potentially neurotoxic [11] , [15] . Although not confirmed in a large longitudinal study of L-DOPA use in PD patients ( ELLDOPA trial ) , this nevertheless remains a major concern [19] . C . elegans has been used as a model to study the structure and function of the nervous system , which in hermaphrodite worms consists of 302 neurons [20] , [21] . C . elegans dopaminergic neurons are functionally related to those of humans . The genes driving the biochemical processes involved in dopamine metabolism ( as well as most PD-associated loci ) [6] are also highly conserved in worms [22] . Dopaminergic neurons can be readily visualized in vivo using appropriate GFP markers . Analogous to vertebrate systems , dopaminergic neurons undergo neurodegeneration upon treatment with 6-OHDA . It has been shown that 6-OHDA can enter dopaminergic neurons through the DAT-1 dopamine transporter and thus trigger their degeneration [3] . The exact type of cellular death that occurs following 6-OHDA intoxication is unknown . Electron microscopy has shown apoptotic-like condensed chromatin structures in dying neurons , suggesting that 6-OHDA induces apoptosis . However , 6-OHDA-induced neurodegeneration in C . elegans is independent of CED-4/Apaf1 and CED-3/caspase , two components of the core apoptotic machinery [3] . In an independent study , inactivation of C . elegans autophagy genes partially suppressed 6-OHDA-induced dopaminergic death , suggesting that autophagy might also be involved in this process [23] . During synaptic transmission most of the released dopamine is transported back into the presynaptic terminal by the dopamine reuptake transporter ( DAT1 ) ( for a review , see [24] . Therefore , activity of this transporter affects the duration and extent of dopamine signaling . Mammalian cell experiments led to the identification of several proteins that interact with DAT1 to modulate its activity , cell surface expression and trafficking . These include protein kinase C , dopamine D2 receptors ( discussed below ) , SNCA and parkin [25]–[28] . The physiological actions of dopamine are mediated by conserved seven-transmembrane dopamine receptors , designated D1–5 . Dopamine receptors are coupled to guanosine triphosphate-binding proteins ( G proteins ) and are classified into D1 or D2 type dopamine receptors based on their antagonistic effect on adenylyl cyclase activity [29] , [30] . D1 dopamine receptors , DOP-1 in worms , are solely found in postsynaptic dopamine-receptive cells , whereas in C . elegans the D2 type receptors DOP-2 and DOP-3 are expressed pre and postsynaptically , respectively [31]–[33] . In vertebrates , the dopamine system plays a crucial role in regulating movement , reward and cognition . Dopamine-deficient newborn mice die as a result of severe motor impairments [34] , [35] . In contrast , C . elegans mutants defective in dopamine synthesis are viable , thus facilitating investigations into dopamine-mediated behavior in these animals . Dopaminergic neurons in C . elegans are required for specific , well-described and quantifiable behaviors , often associated with locomotion and feeding . For instance , the basal slowing response allows worms to reduce their speed when encountering a bacterial lawn , which is their food source [36] . Another behavior mediated by dopamine signaling is referred to as “swimming-induced paralysis” ( SWIP ) : dat-1-deficient worms exhibit rapid paralysis in liquid , unlike wild-type controls [37] . Using an unbiased forward genetic approach we identified tsp-17 as a gene that protects dopaminergic neurons from 6–OHDA-mediated neurodegeneration . We provide evidence that TSP-17 regulates DAT-1 transporter activity . Furthermore , our results suggest that DAT-1 regulation by TSP-17 is partly mediated by D2 dopamine receptors .
In order to find genes that protect dopaminergic neurons , we performed a genetic screen for mutants conferring hypersensitivity to 6-OHDA . By adapting procedures initially established by Nass et al . [3] and using the same pdat-1::GFP reporter that highlights dopaminergic neurons , we screened ∼2500 F2 ethyl methanesulfonate ( EMS ) -mutagenized worms at the L1 developmental stage by incubating with 10 mM 6-OHDA for 1 h . This procedure , which is based on reduced , altered , or absent pdat-1::GFP expression , does not lead to neurodegeneration in >95% of wild-type worms , thus allowing the identification of mutants conferring hypersensitivity to 6-OHDA . Of the initial five mutant candidates , only gt1681 maintained a strong hypersensitive phenotype upon backcrossing ( Figure 1A , Figure S1 ) . 6-OHDA-induced degeneration of both wild-type and gt1681 neurons exhibits the same morphological features and pattern of degeneration initially described by Nass et al . [3] . Axonal blebbing becomes apparent ( Figure 1B , inset , arrows ) a feature also consistent with morphological changes previously observed by electron microscopy . Worms were scored 24 , 48 and 72 h after intoxication . Neurons were lost in less than 10% of wild-type worms after 72 h . In contrast , all dopaminergic neurons were lost in ∼40% of gt1681 worms and partial dopaminergic loss was observed in an additional ∼30% of mutant worms after only 24 h ( Figure 1A ) . The extent of neurodegeneration was further increased 72 h after intoxication , with ∼90% of worms displaying total dopaminergic loss at the adult stage ( Figure 1A ) . Enhanced neurodegeneration in the gt1681 background , albeit to a lesser extent , also occurred in L2 , L3 and L4 larvae treated with 6-OHDA; no such enhancement was seen in adults ( Figure 1C ) . To exclude the possibility that neurodegeneration might be caused by increased net 6-OHDA uptake at the organismal level , we took advantage of the partial growth retardation conferred by 6-OHDA treatment . By scoring for progression to ensuing developmental stages , we found the growth of wild-type and gt1681 worms to be similarly retarded upon toxin treatment , suggesting that gt1681 specifically affects dopaminergic neurons ( Figure 1D ) . The gt1681 mutant is recessive in hermaphrodites ( Figure 2A ) . Genetic linkage was established by single nucleotide polymorphism ( SNP ) mapping , which placed gt1681 on the left arm of the X chromosome . Using unc-20 and lon-2 genetic markers to perform three-factor mapping , the locus was further refined to ∼10 map units . A cross between an unc-20 gt1681 lon-2 triple mutant and the CB4856 “Hawaii” mapping strain enabled us to assess the position of single recombination events relative to gt1681 . This analysis localized gt1681 to an interval between nucleotides 3 , 659 , 480 and 3 , 737 , 466 on the physical map . In parallel , next generation sequencing revealed a single exonic mutation within this interval , leading to a guanine to adenine substitution in the C02F12 . 1 open reading frame and resulting in a glycine to glutamic acid change at position 109 of the encoded protein ( Figure 2B ) . C02F12 . 1 encodes a tetraspanin family , integral membrane protein called TSP-17 ( see below ) . Rescue of the phenotype by a fosmid ( WRM0626aC02 ) encompassing tsp-17 and by a tsp-17-encoding transgene ( Figure 2C ) provides further evidence that gt1681 confers 6-OHDA hypersensitivity . Hypersensitivity is also conferred by the vc2026 allele , a substitution obtained via the Million Mutation Project [38] that results in a glycine to arginine change at position 109 ( Figure 2B , 2D ) . Finally , two deletion alleles , generously provided by the Japanese Knockout Consortium , affecting the first exons of tsp-17 also confer hypersensitivity to 6-OHDA-mediated neurotoxicity ( Figure 2B , 2D ) as does the trans-heterozygous gt1681/tm4995 mutant combination ( Figure 2A ) . Tetraspanins constitute a large protein family , with 30 and 21 members encoded in the human and C . elegans genomes , respectively [39]–[41] . Most tetraspanins have not been functionally characterized . In vertebrates , tetraspanins are suggested to be involved in cell–cell fusion , cell adhesion , cell motility and tumor metastasis [42] . In C . elegans , TSP-12 is involved in modulating Notch signaling , and specific hypodermal TSP-15 expression is required to mediate covalent tyrosine–tyrosine cross-linking during cuticle formation [43] , [44] . C . elegans tsp-17 is predicted to encode two isoforms . The large isoform , C02F12 . 1b , encodes a 312 amino acid protein containing four TM domains . The short isoform , C02F12 . 1a , encodes a 243 amino acid protein that , unlike typical tetraspanins , contains only three transmembrane domains and does not have an intracellular N-terminus . The amino acid change at position 109 in gt1681 affects a highly conserved residue in the third transmembrane domain of the long isoform ( Figure 2B , 2E ) . We confirmed expression of mRNAs encoding for both isoforms , and verified the predicted intron–exon structure ( Figure 2B ) . Using BLAST protein analysis of C . elegans TSP-17 , we found the most likely human orthologs of TSP-17 to be CD63 , Tspan5 and CD82 ( Figure 2E ) . A previous phylogenetic analysis placed TSP-17 within the human CD82 subfamily [45] . However , our attempts to firmly establish an orthologous relationship between TSP-17 and a single human tetraspanin or a distinct subfamily of human tetraspanins were unsuccessful . Our phylogenetic analysis included all tetraspanins from several nematodes , arthropods , cnidarians and chordates ( Figure S2 ) . We speculate that the rapid evolution of this protein family , as often occurs with membrane proteins , compromised our ability to firmly identify a human ortholog of C . elegans TSP-17 . To assess the TSP-17 expression pattern , we used biolistic bombardment to generate transgenic worms ( TG2439 ) expressing a tsp-17::GFP gene fusion ( NM001 ) under the control of its own promoter and 3′UTR . A dat-1 ( promoter ) ::mCherry fusion ( PBI001 ) was co-bombarded to mark dopaminergic neurons . The tsp-17::GFP gene fusion largely suppressed the hypersensitivity phenotype conferred by tsp-17 , thus confirming its functionality ( Figure 2C , bar 3 ) . Importantly , fusion protein expression was observed in all dopaminergic neurons: it was uniform along axons and dendrites of both dorsal and ventral pairs of CEP neurons , as well as in ADE neurons ( Figure 3A–I , arrows indicate axons and dendrites ) and in the posterior PDE neurons . Within the cell body , the TSP-17::GFP fusion seems to be excluded from the nucleus , a pattern that is more evident in a “close-up” image of a PDE neuron , where the signal appears to form a ring-like structure around the nucleus ( Figure 3J–L arrowheads ) . mCherry aggregates ( which are not linked to neurodegeneration ) form dot-like structures in dendrites and axons ( arrows ) , and the surrounding TSP-17 fluorescent signal suggests plasma membrane expression ( arrow , Figure 3K ) . TSP-17 enrichment at the plasma membrane can be observed most prominently in the large cells of the vulva and the sheath cells enclosing the spermatheca ( Figure 3M , N ) . In the spermatheca , TSP-17::GFP expression is also clearly enriched around the nucleus ( Figure 3N , arrowheads ) , possibly localizing to the nuclear membrane or endoplasmic reticulum ( Figure 3N , arrowhead ) . Analysis of subcellular localization in the vulva and spermatheca revealed that the TSP-17::GFP ( gt1681 ) mutant protein is uniformly expressed in the cytoplasm , with a loss of enrichment at the plasma membrane and around the nucleus ( Figure S3A ) . Thus , the gt1681 mutation , which leads to an amino acid change in the fourth transmembrane domain , might compromise the membrane localization of TSP-17 and therefore block its function . TSP-17::GFP is also expressed in multiple neurons throughout worm development . For instance , the NSM serotonergic neuron , which is characterized by extensive axon sprouting , shows TSP-17::GFP expression along its entire length ( Figure 3O ) . Prominent expression was also observed in the muscles of early stage larvae ( Figure 3P ) . Finally expression also appears to be apparent in muscles of the adult head ( Figure 3B , C , H , I ) . In summary , the TSP-17::GFP expression indicates that TSP-17 is expressed in dopaminergic neurons . Transgene expression in dopaminergic neurons was also confirmed by analyzing a TSP-17::GFP expressing transgenic strain crossed to a DAT-1 reporter strain ( Figure S3B ) . We cannot rule out expression of TSP-17 not uncovered by the transgene , due to missing regulatory sequences . We next wanted to investigate whether TSP-17 expression in dopaminergic neurons protects them from 6-OHDA-mediated neurodegeneration . By direct injection of transgenes into the gonad , we generated transgenic worms overexpressing TSP-17 under the control of the dat-1 promoter . Consistent with TSP-17 expression in dopaminergic neurons , we found partial rescue of the hypersensitivity conferred by gt1681 ( Figure 2C , compare bars 1 , 2 and 4 ) . Interestingly , overexpression of TSP-17 and TSP-17 ( gt1681 ) under the dat-1 promoter led to spontaneous neurodegeneration ( Figure S4A , B , respectively ) . This phenotype tended to be more severe following TSP-17 ( gt1681 ) overexpression . Taken together , these data indicate that TSP-17 indeed functions in dopaminergic neurons , and that excessive TSP-17 , especially the mutant form , leads to spontaneous neurodegeneration . We next wished to address how TSP-17 protects dopaminergic neurons . We hypothesized that TSP-17 might affect dopamine synthesis , or dopamine and 6-OHDA uptake or degradation . Dopamine metabolism is itself a source of oxidative stress and may initiate ROS-mediated injury to dopaminergic neurons . The link between excessive dopamine exposure and toxicity is controversial , but overexpression of CAT-2 , the rate-limiting enzyme in dopamine synthesis in C . elegans , is reported to lead to age-dependent degeneration of dopaminergic neurons [46] . We repeated these experiments , and indeed found that neurodegeneration conferred by CAT-2 overexpression in dopaminergic neurons is enhanced in the gt1681 mutant background ( Figure 4A ) . In contrast , we found CAT-2 overexpression to confer a strong resistance toward 6-OHDA-dependent neurodegeneration in both wild-type and gt1618 backgrounds ( Figure 4B ) . We consider it likely that 6-OHDA resistance conferred by CAT-2 overexpression can be explained by reduced 6-OHDA uptake into dopaminergic neurons in the presence of excessive levels of intracellular dopamine . Our results indicate that tsp-17 protects against 6-OHDA toxicity and toxicity caused by excessive dopamine . Since these genetic interactions suggest that dopamine levels could be altered in tsp-17 mutants , we next investigated behavioral phenotypes associated with dopamine . Dopamine synthesis and release are required for the basal slowing response , in which worms reduce their speed when encountering a bacterial lawn [36] . We did not observe a defect in this response , indicating that both dopamine synthesis and extracellular dopamine sensing by receptors are intact in tsp-17 mutants ( Figure S5A ) . One of the most accessible phenotypes thought to be associated with excessive extracellular dopamine is the SWIP ( Swimming Induced Paralysis ) phenotype [37] . While wild-type worms placed into a drop of water maintain their thrashing frequency dat-1 mutants become progressively paralyzed . The SWIP phenotype is ascribed to excessive extracellular dopamine as a consequence of the reuptake defect in the dat-1 mutant . Excessive extracellular dopamine triggers paralysis by hyperactivating the DOP-3 receptor expressed on cholinergic neurons and hence blocking acetylcholine release [33] . To perform this experiment , we placed L4 worms into drops of water and scored their ability to swim over a period of 30 minutes . As expected , we found that wild-type but not dat-1 mutant worms can swim for 30 minutes with no change in the speed or pattern of swimming . All four tsp-17 mutants showed a partial SWIP phenotype ( Figure 5A ) . This phenotype is probably caused by dopaminergic signaling because it can be rescued by deletion of the dop-3 dopamine receptor and by deletion of the cat-2 tyrosine hydroxylase ( Figure 5A and Figure 5B ) . It was surprising to find a SWIP phenotype in tsp-17 mutants as we argue that tsp-17 inhibits dat-1 function ( see below ) . While elucidating the exact mechanism of how TSP-17 affects behavioral phenotypes will require further investigation we speculate that hyper-activation of DAT-1 in tsp-17 strains could trigger a feedback loop that transiently enhance extracellular dopamine levels inducing the weak SWIP phenotype we observe . We also tested for a SWIP phenotype in L1 stage worms , and found that all tsp-17 mutants tested , except the gt1681 allele , behaved similarly to dat-1 mutants ( Figure 5C , D ) . This phenotype , however , is not suppressed by a dop-3 mutation or blocked by a cat-2 mutation ( Figure 5D and Figure S5B ) . We discovered that the “L1 SWIP phenotype” is linked to lethality because worms placed onto agar plates after SWIP assay show reduced viability ( Figure S5C , D ) . Thus , the L1 “swimming-induced lethality” phenotype is unlikely to be related to dopamine levels . Given that TSP-17 is expressed in body wall muscles in L1 larvae , we speculate that swimming-induced lethality might be caused by a muscle defect . To systematically test whether TSP-17 protects dopaminergic neurons by modulating dopamine metabolism , catabolism , reuptake or signaling , we performed a genetic epistasis analysis . As expected , tsp-17 dat-1 double mutants were completely resistant to 6-OHDA-induced neurodegeneration , consistent with the notion that TSP-17 does not bypass 6-OHDA uptake by the DAT-1 dopamine transporter ( Figure 6A ) . We observed no alterations in 6-OHDA sensitivity in cat-2 ( tyrosine hydroxylase ) , bas-1 ( aromatic amino acid decarboxylase/AAADC ) and cat-1 ( VMAT ortholog required for dopamine packaging ) tsp-17 double mutants , indicating that TSP-17 is unlikely to affect levels of dopamine synthesis or packaging ( Figure S6 ) . As 6-OHDA can enter dopaminergic neurons through the DAT-1 transporter owing to its structural similarity to dopamine [3] , [47] , we wondered whether DAT-1 localization or activity is modified in a tsp-17 mutant background . Having established that 6-OHDA hypersensitivity in tsp-17 worms depends on the DAT-1 transporter ( Figure 6A ) , we tested the hypothesis that enhanced DAT-1 transporter activity may contribute to enhanced 6-OHDA-mediated neurotoxicity . Using a functional pdat-1::dat-1::YFP translational fusion , we found that overexpression of this transgene generated by bombardment does not confer overt 6-OHDA hypersensitivity ( Figure 6A , Figure S7 ) . Furthermore , the localization of DAT-1::YFP was similar between wild-type and tsp-17 mutants worms ( Figure S8A ) , a notion further confirmed by Structural Illumination ‘super resolution’ images of CEP dendrites ( Figure S8B ) . Additionally , photobleaching experiments indicated that ∼half of DAT-1::YFP is in the mobile fraction and that the t1/2 is around 30 seconds in both wild-type and tsp-17 ( gt1681 ) worms ( Figure S8C–E ) . We thus aimed to test whether TSP-17 negatively regulates DAT-1 activity using a pharmacological approach . We confirmed previous reports that imipramine specifically inhibits the DAT-1 transporter in the worm [3] ( Figure 6B , left panels , wild-type 0 . 25 mM and 1 mM ) . We reasoned that if DAT-1 is hyperactive in tsp-17 ( gt1681 ) , relatively more imipramine should be needed to inhibit DAT-1 activity and prevent neurodegeneration . We thus treated wild-type , tsp-17 ( gt1681 ) worms and wild-type worms overexpressing DAT-1::YFP with 10 mM 6-OHDA and increasing doses of imipramine ( Figure 6B , middle and right panels ) . We indeed found that higher levels of imipramine are needed to reduce neurodegeneration in DAT-1::YFP overexpressing worms and in tsp-17 ( gt1681 ) worms , and that the effect being stronger in the tsp-17 ( gt1681 ) mutant . Reduced levels of neurodegeneration levels were most clearly observed when concentrations of 0 . 125 mM and 0 . 25 mM imipramine were used ( Figure 6B ) . This result provides evidence that DAT-1 activity may be higher in the tsp-17 mutant background . We aimed to provide further support for this hypothesis by directly measuring dopamine uptake , following previously described procedures . We macerated C . elegans embryos to establish primary embryonic cell cultures , and used these for dopamine uptake assays [48] , [49] . Using two concentrations of tritiated dopamine , we indeed found increased dopamine uptake in tsp-17 mutants ( Figure 6C , D ) . We note that we found this in 7/8 repeat experiments . However , we also note that only a very small proportion of tissue culture cells are dopaminergic neurons and that the absolute amount of dopamine uptake is low especially in the wild-type background . Our combined genetic , pharmacological and biochemical analysis suggests that TSP-17 modulates DAT-1 activity . Previous studies using tissue culture-based assays demonstrated that dopamine receptor activation might promote DAT-1 activity [25] , [50] , [51] . Consistent with these results , we found dop-2 and dop-3 mutant worms to be partially resistant to high doses of 6-OHDA compared to wild-type ( Figure 7A ) . We therefore investigated whether tsp-17 genetically interacts with dopamine receptors to modify DAT-1 activity and confer differential 6-OHDA sensitivity . This was done by assessing the sensitivity of tsp-17 mutants in the absence of the C . elegans DOP-1 D1-like receptor and/or in the absence of the DOP-2 and/or DOP-3 D2-like receptors . C . elegans DOP-1 is expressed in a variety of cells , including cholinergic neurons , mechanosensory neurons , head muscles and neuronal support cells . DOP-3 is expressed postsynaptically and its antagonism of DOP-1 in cholinergic neurons is required for the regulation of locomotion [33] . The DOP-2 receptor is expressed both postsynaptically and presynaptically . When expressed presynaptically , it acts as an autoreceptor on the plasma membrane of dopaminergic neurons . We found that dop-1; tsp-17 ( gt1681 ) was as sensitive to 6-OHDA as the respective tsp-17 single mutant . In contrast , 6-OHDA hypersensitivity was reduced in dop-2; tsp-17 ( gt1681 ) and dop-2; tsp-17 ( tm4994 ) and in dop-3; tsp-17 ( gt1681 ) and dop-3; tsp-17 ( tm4994 ) double mutant worms ( Figure 7B , C and Figure S9 ) Our genetic data thus argue that TSP-17 might inhibit DOP-2 and DOP-3 function , which in turn might be required for full DAT-1 transporter activity ( Figure 7A , E ) . Given that deletion of dop-2 and dop-3 only partially rescues 6-OHDA hypersensitivity in tsp-17 mutants , we speculate that TSP-17 also inhibits DAT-1 activity independently of DOP-2 and DOP-3 . We next aimed to investigate how TSP-17 might regulate DAT-1 or D2-like receptors to modulate DAT-1 activity . Given that these are integral membrane proteins , we employed the split-ubiquitin membrane-based yeast two-hybrid system [52] . In this system , a C-terminal ubiquitin moiety fused to a transmembrane protein and a transcription factor is used a bait . An N-terminal ubiquitin moiety is used as the “prey . ” Upon “reconstruction” of the split ubiquitin , this molecule is recognized by a protease , which cleaves the transcription factor , thus promoting reporter gene activation . By employing various bait and prey fusions with TSP-17 , DAT-1 and DOP-2 , we could not find a direct interaction between TSP-17 and DAT-1 using the split-ubiquitin system ( Figure 7D ) . In contrast , we found that DOP-2 and TSP-17 may indeed interact . The specificity of this interaction was clearly revealed when the beta-galactosidase reporter assay was used as an output . In addition , yeast colony formation on his-3 or his-3 ade-2 plates was enhanced when the corresponding reporters where used ( Figure 7D ) . Thus , TSP-17 might modulate DOP-2 activity by a direct physical interaction , consistent with TSP-17 affecting ligand binding , downstream signaling or membrane trafficking of DOP-2-like receptors . Our genetic data also suggest that TSP-17 might also act via other factors to dampen DAT-1 activity ( Figure 7B ) .
Using C . elegans as a model and employing unbiased genetic approaches , we aimed to find neuroprotective genes that alleviate the 6-OHDA-induced degeneration of dopaminergic neurons . Based on our genetic data , which is supported by the characterization of several alleles and transgenic rescue experiments , we provide compelling evidence that TSP-17 protects dopaminergic neurons from 6-OHDA-mediated toxicity . TSP-17 appears to function in dopaminergic neurons , and our combined genetic , pharmacological and biochemical evidence suggests that it might act by antagonizing DAT-1 dopamine transporter activity . We do not know how TSP-17 regulates DAT-1 at a mechanistic level . TSP-17 is a member of the evolutionarily conserved family of tetraspanins , comprising 20–50 kDa membrane proteins that contain four transmembrane domains . A characteristic feature of tetraspanins is their ability to form lateral associations with each other and with other proteins . Such interactions are thought to lead to a dynamic assembly , resulting in the formation of a network of molecular interactions referred to as the tetraspanin web [41] , [53] . Tetraspanins are thought to have regulatory functions in the ligand binding , downstream signaling , protein trafficking and proteolytic activities of associated proteins [42] , [54] . In C . elegans , only two tetraspanins have known functions . TSP-15 appears to be required to activate the BLI-3 dual oxidase to regulate H202 production at the plasma membrane and thus alter dityrosine cross-linkage of extracellular matrix proteins [44] , [55] . Genetic evidence suggests that TSP-12 , most closely related to human TSPAN33 , appears to facilitate Notch signaling redundantly with TSP-14 . Thus conserved tetraspanins likely function by facilitating γ-secretase cleavage of the membrane-bound form of Notch , thus promoting nuclear localization of this transcription factor [43] . DAT-1 hyperactivity in the tsp-17 mutants could result from altered DAT-1 localization or abundance at the cell membrane; alternatively , TSP-17 might indirectly regulate DAT-1 activity . Using a functional DAT-1::YFP construct , we did not see any obvious change in DAT-1 expression , localization , or change in half life in tsp-17 mutants and we thus favor the idea that TSP-17 regulates DAT-1 activity . Our finding that TSP-17 genetically and biochemically interacts with the DOP-2 D2-like dopamine receptor , suggests an indirect mode of DAT-1 regulation by TSP-17 ( Figure 7E ) . Our genetic analysis provides evidence that TSP-17 might in part regulate DAT-1 via DOP-2 and DOP-3 dopamine receptors ( Figure 7E ) . We found that depletion of the D2-like dopamine receptors , DOP-2 and/or DOP-3 , in tsp-17 mutants leads to a moderate reduction in the 6-OHDA hypersensitivity conferred by tsp-17 , while D2-like dopamine receptor single knockout strains show the same 6-OHDA sensitivity as wild-type worms . Thus , our analysis suggests that tsp-17 genetically interacts with D2-like dopamine receptors , in line with our observation that TSP-17 directly binds to DOP-2 . In mammalian systems , dopamine autoreceptors are reported to have a major role in providing inhibitory feedback to adjust the rate of neuronal firing , dopamine synthesis and dopamine release in response to the dopamine level in the synaptic cleft [30] , [32] . Several studies suggest that vertebrate D2 dopamine receptors also modulate DAT-1 activity to regulate the dopamine level in the synaptic cleft . Cass and Gerhardt used pharmacological approaches to demonstrate that inhibition of D2 class dopamine receptors significantly inhibits DAT function [50] . Two independent studies provided evidence that D2 receptors regulate both the activity and cell surface expression of DAT-1 [25] , [51] . Nevertheless , further investigations are required to establish functional links between C . elegans DOP-2 receptors and DAT-1 activity . The ability of TSP-17 to inhibit DAT-1 both via DOP-2 and independent of D2-like receptors ( Figure 7E ) suggests that TSP-17 modulates the activity of multiple signaling proteins . Indeed , our observation of excessive neurodegeneration following wild-type , and especially mutant , TSP-17 overexpression in dopaminergic neurons hints that malfunctioning and/or excessive TSP-17 blocks pathways needed to maintain the integrity of dopaminergic neurons . The enhanced defect associated with overexpression of mutant TSP-17 that fails to show the correct cytoplasmic localization hints the neurotoxicity might be conferred by the sequestration of TSP-17 interacting proteins essential for neuronal survival . Dopamine neuronal dysfunction has been associated with several common neurobehavioral disorders , including drug addiction , schizophrenia and attention-deficit hyperactivity disorder [32] , [56]–[58] . The DAT-1 dopamine transporter plays a central role in dopamine signaling , and it is likely to be subjected to complex modes of regulation . DAT-1 is the target of psychoactive addictive drugs such as cocaine and amphetamine , and DAT1 overexpression leads to increased amphetamine sensitivity [59]–[63] . Mechanisms related to dopamine signaling tend to be evolutionarily conserved . Thus , studies aimed to genetically define modulators of dopamine signaling and 6-OHDA-mediated toxicity will provide important insights into the mechanisms of dopamine signaling in health and disease . Idiopathic PD is thought to be triggered by a combination of environmental factors and genetic susceptibility , and a case has been made that exposure to environmental toxins such as the pesticides paraquat and rotenone leads to increased PD [9] . Indeed , chemical and tissue culture studies have provided evidence that increased dopamine levels may lead to enhanced neurodegeneration , probably through the generation of toxic intermediates such as the neurotoxic product of dopamine oxidation , 6-OHDA [13] , [15] , [64]–[68] . The specificity of 6-OHDA entry into dopamine neurons depends on DAT , and DAT antagonists can block uptake [3] , [4] , [11] , [47] . Interestingly , DAT-1 hyperactivity in tsp-17 mutants further enhances the neurodegeneration conferred by elevated dopamine synthesis in CAT2 tyrosine hydroxylase-overexpressing worm strains . Thus , DAT-1 hyperactivity might enhance neurodegeneration by further increasing the intracellular concentration of dopamine and/or toxic metabolites . DAT1 expression or activity has not been linked to PD , but it is intriguing that among dopamine neurons those residing in the substantia nigra express the highest DAT levels in vivo and are most strongly affected in PD [4] , [60] .
Strains were grown at 20°C under standard conditions , unless indicated otherwise . N2 Bristol was used as the wild-type strain . The tsp-17 ( tm4994 ) and tsp-17 ( tm5169 ) mutants were generated and kindly provided by Shohei Mitani of the National Bioresource Project for the Nematode ( http://www . shigen . nig . ac . jp/c . elegans/ ) . Details of the respective alleles are described by the National Bioresource Project for the Nematode and by WormBase ( www . wormbase . org ) . All mutants were outcrossed a minimum of four times to the TG2435 vtIs1[pdat-1::gfp] strain originally generated by the Blakely laboratory ( BY200 ) and repeatedly crossed into the N2 background . TG2435 vtIs1[pdat-1::gfp; rol-6] V , TG1681 vtIs1 V; tsp-17 ( gt1681 ) X , TG2436 vtIs1 V; tsp-17 ( tm4994 ) X , TG2437 vtIs1 V; tsp-17 ( tm5169 ) X , TG2438 vtIs1 V; tsp-17 ( gk276386 ) X , TG2462 vtIs1 V; CB4856 , TG2463 vtIs1 V; lon-2 ( e678 ) unc-20 ( e112 ) X , TG2464 vtIs1 V; tsp-17 ( gt1681 ) unc-20 ( e112 ) X , TG2465 vtIs1 V; tsp-17 ( gt1681 ) lon-2 ( e678 ) X , TG2395 cat-2 ( e1112 ) II; vtIs1 V , TG2394 cat-2 ( e1112 ) II; vtIs1 V; tsp-17 ( gt1681 ) X , TG2396 bas-1 ( tm351 ) III; vtIs1 V , TG2397 bas-1 ( tm351 ) III; vtIs1 V; tsp-17 ( gt1681 ) X , TG2399 vtIs1 V; cat-1 ( e1111 ) X , TG2398 vtIs1 V; cat-1 ( e1111 ) tsp-17 ( gt1681 ) X , TG2400 dat-1 ( ok157 ) III; vtIs1 V , TG2401 dat-1 ( ok157 ) III; vtIs1 V; tsp-17 ( gt1681 ) X , TG2404 amx-1 ( ok659 ) III; vtIs1 V , TG2403 amx-1 ( ok659 ) III; vtIs1 V; tsp-17 ( gt1681 ) X , TG2406 amx-2 ( ok1235 ) I; vtIs1 V , TG2405 amx-2 ( ok1235 ) I; vtIs1 V; tsp-17 ( gt1681 ) X , TG2408 amx-2 ( ok1235 ) I; amx-1 ( ok659 ) III; vtIs1 V , TG2407amx-2 ( ok1235 ) I; amx-1 ( ok659 ) III; vtIs1 V; tsp-17 ( gt1681 ) X , TG2410 vtIs1 V; dop-1 ( vs100 ) X , TG2409 vtIs1 V; dop-1 ( vs100 ) tsp-17 ( gt1681 ) X , TG2412 vtIs1 dop-2 ( vs105 ) V , TG2411 vtIs1 dop-2 ( vs105 ) V; tsp-17 ( gt1681 ) X , TG2414 vtIs1 V; dop-3 ( vs106 ) X , TG2413 vtIs1 V; dop-3 ( vs106 ) tsp-17 ( gt1681 ) X , TG2466 vtIs1 dop-2 ( vs105 ) V; dop-3 ( vs106 ) X , TG2467 vtIs1 dop-2 ( vs105 ) V; dop-3 ( vs106 ) tsp-17 ( gt1681 ) X , TG2415 vtIs1 dop-2 ( vs105 ) V; dop-1 ( vs100 ) dop-3 ( vs106 ) X , TG2416 vtIs1 dop-2 ( vs105 ) V; dop-1 ( vs100 ) dop-3 ( vs106 ) tsp-17 ( gt1681 ) X , UA57 baIn4[pdat-1::gfp pdat-1::cat-2] , TG2402 baIn4[pdat-1::gfp pdat-1::cat-2] , ; tsp-17 ( gt1681 ) X , TG2470 gtIn2469[pdat-1::dat-1::yfp::let-858 3′UTR , unc-119 ( + ) ]; gtIn2468[pdat-1::mcherry::let858 3′UTR , unc-119 ( + ) ]; unc-119 ( ed3 ) III , TG2471 gtIn2469[pdat-1::dat-1::yfp::let-858 3′UTR , unc-119 ( + ) ]; gtIn2468[pdat-1::mcherry::let858 3′UTR , unc-119 ( + ) ]; unc-119 ( ed3 ) III; tsp-17 ( gt1681 ) X , TG2439 gtIn2439[ptsp-17::tsp-17::gfp::tsp-17 3′UTR , pdat-1::mcherry::let858 3′UTR , unc-119 ( + ) ]; unc-119 ( ed3 ) III , TG2472 tsp-17 ( gt1681 ) X; gtIn2439[ptsp-17::tsp-17::gfp::tsp-17 3′UTR , pdat-1::mcherry::let858 3′UTR , unc-119 ( + ) ]; unc-119 ( ed3 ) III , TG2440 gtEx2440[pdat-1::tsp-17::cfp:: let-858 3′UTR , unc-119 ( + ) ]; unc-119 ( ed3 ) III; vtIs1 [pdat-1::gfp; rol-6] V , TG2473 vtIs1 [pdat-1::gfp; rol-6] V; tsp-17 ( gt1681 ) X; gtEx2440 [pdat-1::tsp-17::cfp:: let-858 3′ UTR , unc-119 ( + ) ] , TG2474 vtIs1 [pdat-1::gfp; rol-6] V; unc-119 ( ed3 ) III; gtEx2474[pdat-1::tsp-17 ( G74E ) ::cfp:: let-858 3′UTR , unc-119 ( + ) ] , TG2478 cat-2 ( e1112 ) II; vtIs1V; tsp-17 ( tm4994 ) X , TG2475 dat-1 ( ok157 ) III; vtIs1V; tsp-17 ( tm4995 ) X , TG2477 vtIs1; dop-3 ( vs106 ) tsp-17 ( tm4995 ) X , TG2476 dat-1 ( ok157 ) III; vtIs1V; dop-3 ( vs106 ) X , EMS was added to 4 ml synchronized young adult worms in M9 buffer to a final concentration of 25 mM and incubated for 4 h at 20°C . Mutagenized worms were washed in M9 buffer and incubated at 15°C . Synchronous F1-generation L1 larvae were used for screening . F2-generation L1 larvae from mutagenized TG2435 dat-1::gfp ( BY200 ) worms were used for the mutagenesis screen . L1 larvae were intoxicated with 10 mM 6-OHDA . After 72 h , worms with the highest incidence of neurodegeneration were isolated and scored as hypersensitive . SNP mapping of mutants was done as previously described [69] . To obtain synchronized L1 larvae , 1–10 adult worms ( 24 h post-L4 stage ) were incubated in 70 µl M9 without food on at 20°C , with shaking at 500 rpm for 27–40 h to lay eggs . After hatching , all L1 larvae were collected . Approximately 50 L1 larvae were added to an assay mix ( 50 µl ) containing 10 mM 6-OHDA and 40 mM ascorbic acid , and incubated for 1 h at 20°C , with shaking at 500 rpm . For co-treatment with imipramine or haloperidol , the respective compounds were added to the assay mix at the same time as 6-OHDA . After a 1-h incubation , M9 buffer ( 100 µl ) was added to the assay mix , and the solution containing L1 worms was then transferred to an unseeded NGM plate . After 30 min , L1 worms were individually picked and transferred onto a fresh NGM plate seeded with a line of OP50 bacteria to ease subsequent scoring . Intoxicated worms were incubated at 20°C and scored for dopaminergic neurodegeneration every 24 h for 3 days . All 6-OHDA treatments were done in triplicate and at least 80–100 worms were tested for each strain and condition . All worms used for SWIP analysis were grown on NGM plates seeded with E . coli OP50 bacteria . For each test , 5–10 L4 hermaphrodites or 10 L1 worms were placed into 40 µl water in a single well of a Pyrex Spot Plate . Paralyzed worms were counted at 1-min intervals using a Leica dissecting microscope [70] . L1 worms were hand picked from seeded plates , 12 hours after the addition of embryos , obtained by bleaching . For semi-quantitative analyses of 6-OHDA-induced degeneration , worms were examined using a Leica fluorescent dissecting microscope . The absence of all eight dopaminergic neurons in worms was scored as “complete loss . ” The presence of a complete , intact set of eight dopaminergic neurons was scored as “no loss . ” Any intermediate situation , for example a damaged or absent subset of dopaminergic neurons or missing dendrite portions , was scored as a “partial loss . ” Neurodegeneration resulting from cat-2 overexpression was scored using developmentally synchronized worms , as indicated . A DeltaVision microscope ( Applied Precision ) was used to acquire images . All images were analyzed using softWoRx Suite and softWoRx Explorer software ( Applied Precision ) . Total RNA was isolated and reverse transcribed from wild-type C . elegans ( N2 ) using an RNeasy mini kit ( QIAGEN ) . Coding regions of dop-2c ( K09G1 . 4c ) and dat-1 ( T23G5 . 5 ) were amplified and cloned into pBT3-STE vectors ( Dual Systems Schlieren ) for expression of a fusion protein containing the C-terminal half of ubiquitin ( Cub ) and the artificial transcription factor LexA-VP16 . tsp-17b ( C02F12 . 1b ) cDNA was amplified and cloned into prey vector pPR3-STE for expression of a fusion protein containing a mutated version of the N-terminal half of ubiquitin ( NubG ) . Constructs were verified by DNA sequencing , and sequences of the respective constructs can be provided upon request . Yeast transformations and pairwise interaction assays were done according to the protocol of Dualsystems Schlieren . Embryonic cells were prepared as described previously ( Christensen , M , et al 2002 , Neuron ) . The uptake assay was done according to Carvelli et al . ( 2004 ) . Briefly , C . elegans cells cultured for 2 days were washed twice with KRH buffer ( 120 mM NaCl , 4 . 7 mM KCl , 1 . 2 mM KH2P04 , 10 mM Hepes , 2 . 2 CaC12 , 10 mM glucose , 0 . 1 mM ascorbic acid and 0 . 1 mM tropolone and 0 . 1 mM pargyline mono amine oxidase inhibitors ) and incubated with 50 or 250 nM [3H]-dopamine for 20 min at room temperature . Uptake was terminated by three washes of ice-cold KRH buffer , and cells were lysed by incubation with 1% SDS for 20 min . [3H]-dopamine uptake was measured in each genetic background , based on radioactive counts , using a scintillation counter ( PerkinElmer Liquid Scintillation Analyzer Tri-Carb 1800TR ) . Total cell numbers were determined with a hemocytometer and were used to normalize radioactive counts . Cell numbers varied between experiments but were not biased towards mutant or control strain: There were 400 , 000/400 , 000 , 75 , 000/150 , 000 and 1 , 000 , 000/400 , 000 cells for control/mutant strain , respectively . Cell extraction and uptake assays were always done simultaneously for both strains . The error bars depict the standard error of the means ( SEM ) . Neurodegeneration and SWIP assay data are presented as the average of three biological replicates , and error bars represent the standard error of the mean , unless otherwise indicated . When assaying neurodegeneration statistical significance was calculated using the Chi-Sqare test using Yates p-values . http://www . quantpsy . org/chisq/chisq . htm . The statistical significance of differences in the SWIP assays ( Figure 5 ) was calculated using the two-tailed t-test . | Parkinson's disease ( PD ) is characterized by the progressive loss of dopaminergic neurons . While hereditary forms are known , most cases are attributable to a combination of genetic and environmental risk factors . In PD models , dopaminergic neurodegeneration can be triggered by neurotoxins such as 6-hydroxydopamine ( 6-OHDA ) . This drug , which is taken up by the presynaptic dopamine reuptake transporter ( DAT-1 ) , also causes the selective death of C . elegans dopaminergic neurons . We found that TSP-17 , a member of the tetraspanin family of membrane proteins , protects dopaminergic neurons from 6-OHDA-induced degeneration . We provide evidence that TSP-17 inhibits the C . elegans dopamine transporter DAT-1 , leading to increased neuronal 6-OHDA uptake in tsp-17 mutants . TSP-17 also protects against toxicity conferred by excessive intracellular dopamine . TSP-17 interacts with the DOP-2 dopamine receptor , possibly as part of a pathway that negatively regulates DAT-1 . tsp-17 mutants have subtle behavioral phenotypes that are partly conferred by aberrant dopamine signaling . In summary , we have used C . elegans genetics to model key aspects of PD . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"model",
"organisms",
"genetics",
"biology",
"and",
"life",
"sciences",
"neuroscience",
"research",
"and",
"analysis",
"methods"
] | 2014 | Tetraspanin (TSP-17) Protects Dopaminergic Neurons against 6-OHDA-Induced Neurodegeneration in C. elegans |
Natural variation separates Epstein-Barr virus ( EBV ) into type 1 and type 2 strains . Type 2 EBV is less transforming in vitro due to sequence differences in the EBV transcription factor EBNA2 . This correlates with reduced activation of the EBV oncogene LMP1 and some cell genes . Transcriptional activation by type 1 EBNA2 can be suppressed through the binding of two PXLXP motifs in its transactivation domain ( TAD ) to the dimeric coiled-coil MYND domain ( CC-MYND ) of the BS69 repressor protein ( ZMYND11 ) . We identified a third conserved PXLXP motif in type 2 EBNA2 . We found that type 2 EBNA2 peptides containing this motif bound BS69CC-MYND efficiently and that the type 2 EBNA2TAD bound an additional BS69CC-MYND molecule . Full-length type 2 EBNA2 also bound BS69 more efficiently in pull-down assays . Molecular weight analysis and low-resolution structures obtained using small-angle X-ray scattering showed that three BS69CC-MYND dimers bound two molecules of type 2 EBNA2TAD , in line with the dimeric state of full-length EBNA2 in vivo . Importantly , mutation of the third BS69 binding motif in type 2 EBNA2 improved B-cell growth maintenance and the transcriptional activation of the LMP1 and CXCR7 genes . Our data indicate that increased association with BS69 restricts the function of type 2 EBNA2 as a transcriptional activator and driver of B cell growth and may contribute to reduced B-cell transformation by type 2 EBV .
Epstein-Barr virus ( EBV ) is a ubiquitous γ-herpesvirus that immortalises human B lymphocytes to establish a lifelong persistent infection that is usually harmless . Primary infection can however give rise to infectious mononucleosis and this is more common when infection is delayed until adolescence or adulthood . EBV is also associated with the development of malignancies that include Burkitt’s ( BL ) , Hodgkin’s , diffuse large B cell and post-transplant lymphoma and nasopharyngeal or gastric carcinoma . EBV expresses nine latent proteins in in vitro infected lymphoblastoid cell lines ( LCLs ) , including 6 Epstein-Barr nuclear antigens ( EBNA1 , 2 , 3A , 3B , 3C and leader protein ) and 3 latent membrane proteins ( LMP1 , 2A , 2B ) . The EBNA2 transcription factor is one of five of these latent genes essential for B cell transformation [1] . EBNA2 functions as the master regulator of EBV latent gene transcription and activates numerous cell genes that control B cell growth and survival [2] . It cannot however bind to DNA directly and hijacks cell DNA binding proteins e . g . RBPJ ( RBPJκ , CBF1 ) and EBF1 to target viral and cell gene regulatory elements [2 , 3] . Although EBNA2 binding sites are close to gene promoters in the viral genome , in the B cell genome they are mostly found at enhancer elements and EBNA2 has been shown to promote enhancer-promoter interactions [4–6] . EBNA2 activates transcription through interactions between its acidic transactivation domain ( TAD ) and histone acetyl transferases , ATP-dependent remodellers and components of the preinitiation complex [7–13] . EBV genome sequences worldwide separate into two main strains ( type 1 and type 2 ) based on differences in the EBNA2 and EBNA3A , 3B and 3C genes [14–18] . Type 2 strains are less efficient at immortalising resting B cells in vitro than type 1 strains [19] . This phenotype is determined by sequence variation in EBNA2 since complementation of an EBNA2 defective virus with type 1 EBNA2 but not type 2 EBNA2 supports efficient primary B cell immortalisation [1] . Consistent with its reduced primary B cell transforming function , type 2 EBNA2 cannot complement loss of type 1 EBNA2 function to maintain the growth of lymphoblastoid cell lines [20] . Amino acids responsible for the differences in B cell growth maintenance between type 1 and type 2 EBNA2 were mapped to the C-terminal region of EBNA2 [20] . Surprisingly , a single amino acid ( aspartate 442 of type 1 EBNA2 ) in the TAD appears to be a key determinant of B cell growth maintenance by type 1 EBNA2 . Replacing the serine that occurs at the equivalent position in type 2 EBNA2 ( amino acid 409 of type 2 EBNA2 ) with aspartate ( mutant S442D ) confers efficient growth maintenance function [21] . Type 2 EBNA2 has reduced ability to activate expression of the EBV oncogene LMP1 and a small number of cellular genes e . g . CXCR7 [22] . These differences in gene activation could underlie the reduced B cell growth maintenance and transforming function of type 2 EBNA2 , although the mechanism involved and the role played by the single aspartate residue is unclear . Part of the mechanism may involve ( or result in ) reduced binding of type 2 EBNA2 to the LMP1 promoter and cell gene regulatory elements [21] . EBNA2 binding sites at genes activated less efficiently by type 2 EBNA2 are enriched for composite binding motifs for ETS and IRF transcription factors ( ETS and IRF composite element; EICE ) , implicating ETS/IRF family members in the gene specificity of the observed effects [21] . Despite a clear deficiency in the immortalising and B cell growth maintenance properties of type 2 EBNA2 in vitro , no specific differences in disease association have been reported to date for type 1 and type 2 EBV . Interestingly , although the outgrowth of immortalised cells is less efficient and much slower in primary B cell cultures infected with type 2 EBV [19 , 20] , the LCLs that are eventually established from type 2 viruses proliferate at similar rates to type 1 LCLs . Type 1 and 2 LCLs also show equivalent expression of LMP1 and CXCR7 [20] . Over extended periods of time , it is therefore possible to select for immortalised cells infected with type 2 EBV that have the required levels of expression of these genes to support their long term proliferation . In vivo other factors may create an environment that helps support B cell immortalisation by type 2 EBV . New research also suggests that type 2 EBV may use alternative approaches to persist in vivo . Type 2 EBV has the unique capacity to infect T cells in culture and is detected in T cells from healthy infants from Kenya , indicating that T cell infection may form part of a natural type 2 EBV infection [23 , 24] . Recent work also showed that a type 2 EBV strain was able to infect both B cells and T cells in humanised mice [25] . Mice infected with type 2 EBV developed tumours that resembled the diffuse large B-cell lymphomas that also developed in mice infected with type 1 EBV , confirming the tumorigenic potential of a persistent type 2 EBV infection , once established [25] . BS69 ( ZMYND11 ) is a multi-domain chromatin-associated repressor protein that suppresses transcription elongation , regulates pre-mRNA processing and has tumour suppressor function [26 , 27] . The BS69 gene undergoes chromosomal translocation in minimally differentiated myeloid leukaemia leading to the expression of a BS69-MBTD1 fusion protein [28] . BS69 contains three histone reader domains in its N-terminal region; a plant homeodomain , a bromodomain and a PWWP domain . The tandemly-arranged bromodomain and PWWP domain bind to histone H3 or the variant histone H3 . 3 when trimethylated on lysine K36 [27 , 29] . BS69 also contains a coiled-coil ( CC ) dimerisation domain adjacent to a MYND domain in its C terminus . BS69 binds to a number of chromatin modifying enzymes ( BRG1 , HDAC1 , EZH2 ) and transcription factors ( adenovirus E1a , c-Myb , ETS2 , E2F6 , and the Myc-associated MGA protein ) and inhibits transcription factor activation function [30–34] . The BS69 MYND domain binds to E1a and MGA through a PXLXP motif [34] . BS69 has also been shown to interact with the TAD of type 1 EBNA2 through two PXLXP motifs and to restrict EBNA2 transcriptional activation function [34 , 35] . The structure of the dimeric CC-MYND domain of BS69 bound to two peptides encompassing sequences from one of the EBNA2 PXLXP motifs ( motif 1 ) has been solved [35] . Based on this three-dimensional structure , a BS69 dimer was predicted to interact with the two adjacent PXLXP motifs in type 1 EBNA2 . Interestingly the amino acid implicated in the type-specific differences in growth maintenance observed for type 1 and type 2 EBNA2 ( amino acid 442 in type 1 EBNA2 ) [21] lies immediately adjacent to the second BS69 binding motif . We set out to determine whether sequence differences between type 1 and type 2 EBNA2 affect BS69 binding . We hypothesised that type 2 EBNA2 would show increased binding to BS69 and that this would impair its gene activation and growth maintenance function . We initially examined the impact of type-specific differences in EBNA2 amino acid 442 on BS69 binding . We also identified a third PXLXP BS69 binding motif in type 2 EBNA2 , so we examined whether the presence of this extra motif resulted in the interaction of additional molecules of BS69 with type 2 EBNA2 . We found that amino acid 442 did not affect BS69 binding , but influenced the conformation of the TAD , potentially affecting binding of other transcriptional regulators . We also demonstrated that the third PXLXP motif in type 2 EBNA2 was responsible for the binding of an additional BS69 dimer . Importantly , mutation of the third PXLXP BS69 binding motif in full length type 2 EBNA2 restored its ability to maintain B cell growth and activate transcription indicating that increased BS69 binding contributes to the impaired activity of type 2 EBNA2 .
Previous studies demonstrated that differences in a single amino acid in the TAD between type 1 and type 2 EBNA2 determined the ability of EBNA2 to maintain the growth of an EBV-infected LCL [21] . This amino acid ( located at position 442 in the EBNA2 sequence from the prototypical type 1 B95-8 strain of EBV ) is conserved as aspartate in type 1 strains and as serine ( at the corresponding position of 409 ) in type 2 strains . In type 1 EBNA2 , aspartate 442 is located immediately adjacent to a previously identified binding motif for the cell transcriptional repressor BS69 ( motif 2 ) that fits the PXLXP consensus ( PILFP437-441 ) [34] . In the TAD of type 2 EBNA2 , the PXLXP motif is conserved ( PFLFP404-408 ) and is flanked by serine 409 ( Fig 1A ) . We hypothesised that the impaired gene activation and growth maintenance properties of type 2 EBNA2 may be the result of increased binding to BS69 as a result of the aspartate to serine amino acid difference adjacent to BS69 binding motif 2 . We therefore tested whether a BS69 binding motif 2 peptide from type 2 EBNA2 showed enhanced binding to BS69 compared to a motif 2 peptide from type 1 EBNA2 . We used isothermal titration calorimetry ( ITC ) to determine the affinity of peptide binding to the C-terminal region of BS69 ( amino acid 480–602 ) comprising the CC-MYND domain that we expressed and purified from E . coli ( Fig 1B and 1C ) . In contrast to our hypothesis , we found that the type 2 EBNA2 motif 2 peptide bound to BS69CC-MYND with reduced affinity ( KD = 176 μM ) compared to the corresponding peptide from type 1 EBNA 2 ( KD = 47 . 7 μM ) ( Fig 1B and 1C and S1 Table ) . The affinity of binding of the type 1 EBNA 2 motif 2 peptide to BS69CC-MYND was very similar to the previously reported KD of 35 μM [35] . The difference in binding between type 1 and type 2 EBNA2 motif 2 peptides could be influenced by both the aspartate to serine change and differences in two other amino acids present in the sequence ( Fig 1B and 1C ) . An additional BS69 PXLXP binding motif previously identified in type 1 EBNA2 ( motif 1 ) located N-terminal to motif 2 is also present in type 2 EBNA 2 ( Fig 1A ) . The BS69CC-MYND dimer binds a type 1 EBNA2 polypeptide containing both motif 1 and motif 2 with high affinity and the structure of BS69 dimer could accommodate binding to both motifs simultaneously [35] . We therefore tested whether sequence differences in type 2 EBNA2 ( including the aspartate to serine change ) affected the binding of a region of EBNA2 containing both motif 1 and motif 2 to BS69 . Type 1 EBNA2381-445 and type 2 EBNA2348-412 polypeptides were expressed and purified from E . coli and their interaction with BS69CC-MYND examined using ITC . Consistent with previous reports [35] we found that type 1 EBNA2381-445 bound to BS69CC-MYND with high affinity ( KD = 0 . 95 μM ) likely due to the high avidity of interaction with two binding sites ( Fig 1D ) . In the context of this larger region of EBNA2 we found very little difference in the affinity of type 2 EBNA2 binding to BS69CC-MYND ( KD = 1 . 21 μM ) ( Fig 1E ) . In addition to binding affinities , ITC data can also be used to calculate thermodynamic parameters ( ΔG , ΔH and–TΔS ) thus providing insight into the noncovalent forces involved in binding . We observed very similar thermodynamic profiles on binding of type 1 EBNA2381-445 and type 2 EBNA2348-412 to BS69CC-MYND , with negative ( favourable ) binding enthalpy ( ΔH ) and positive ( unfavourable ) entropy factors ( S1 Table ) . ITC data can also be used to calculate binding stoichiometry ( n ) which can be visualised as the molar ratio at the mid ( inflection ) point of the sigmoidal binding curve . We titrated EBNA2 polypeptides into a cell containing BS69CC-MYND , so the stoichiometry values we obtained indicate the molar ratio at which the EBNA2 polypeptide saturates the available sites in BS69CC-MYND monomers . Consistent with the presence of two BS69 binding sites in the EBNA2 polypeptides , we obtained n values of 0 . 42 and 0 . 33 for type 1 and type 2 EBNA2: BS69CC-MYND binding , respectively ( Fig 1D and 1E ) . These approximate to the expected molar ratio of 0 . 5 taking into consideration some margin of error in n value determination by ITC , which is heavily influenced by the accuracy of protein concentrations and the proportion of ‘active’ protein in the sample . We conclude that the aspartate 442 to serine amino acid difference between type 1 and type 2 EBNA2 does not affect the binding of BS69 to the TAD of type 2 EBNA2 in these assays . Our data also indicate that additional sequence differences in and around BS69 binding motifs 1 and 2 in type 2 EBNA2 do not influence the binding of the BS69CC-MYND dimer to this region of the protein . During the course of our study we also identified a third potential BS69 binding site in the type 2 EBNA2 TAD ( Fig 2A ) . In type 2 EBNA2 , a sequence that is an exact match to the PXLXP BS69 consensus binding motif is present C-terminal to motif 2 ( PTLEP414-418 ) . In type 1 EBNA2 the corresponding region has an isoleucine in place of the leucine residue ( PSIDP447-451 ) . To determine whether these regions of EBNA2 also interact with BS69 , we performed ITC experiments using type 1 and type 2 EBNA2 peptides ( Fig 2B and 2C ) . We were not able to detect any binding of the type 1 EBNA2 peptide encompassing this region ( T1 EBNA2445-455 ) to BS69CC-MYND , underscoring the importance of the central leucine in the PXLXP motif for the BS69 interaction ( Fig 2B ) . In contrast , a peptide from the corresponding region of type 2 EBNA2 ( T2 EBNA2412-422 ) interacted with BS69 with a KD = 219 μM ( Fig 2C ) . The affinity of interaction with this new motif ( that we named motif 3 ) is weaker than the interaction we observed for type 1 or type 2 EBNA2 motif 2 ( Fig 1 ) . To determine the impact of motif 3 on the interaction of type 2 EBNA2 with BS69CC-MYND in the presence of the two other BS69 binding motifs , we expressed and purified a larger type 2 EBNA2 polypeptide containing motif 1 , 2 and 3 ( T2 EBNA2348-422 ) for use in ITC . For comparison , we also analysed the binding of the corresponding larger region of type 1 EBNA2 ( T1 EBNA2381-455 ) . We found that inclusion of the additional C-terminal amino acids had little impact on the affinity , stoichiometry or thermodynamic profile of binding of type 1 EBNA2 to BS69CC-MYND ( compare Fig 2D and Fig 1D ) ( S1 Table ) . In contrast , for type 2 EBNA2 , we observed a change in the stoichiometry of binding from 0 . 33 when motif 1 and 2 were present ( T2 EBNA2348-412 ) to 0 . 15 when motif 1 , 2 and 3 were present ( T2 EBNA2348-422 ) ( compare Figs 2E and 1E ) . We also observed an increase in ΔH from -22 . 5 to -53 . 2 kcal/mol and increase in -TΔS from 14 . 4 to 45 kcal/mol on inclusion of the additional region of type 2 EBNA2 containing motif 3 ( S1 Table ) . These data are therefore consistent with the presence of an additional BS69 binding site in type 2 EBNA2 that supports the interaction of an additional BS69CC-MYND molecule . Perhaps surprisingly , we did not observe an increase in the affinity of binding of the longer type 2 EBNA2 polypeptide to BS69CC-MYND ( compare Figs 2E and 1E ) . Nonetheless the recruitment of more BS69 to type 2 EBNA2 could be physiologically relevant for the function of type 2 EBNA2 as a transcriptional activator . To confirm that the observed change in binding stoichiometry was due to the presence of motif 3 in the type 2 EBNA2 polypeptide , we analysed the binding of T2 EBNA2348-422 with motif 3 mutated from PTLEP to ATAEA to substitute both prolines and the central leucine residue to alanines ( T2 EBNA2348-422 motif 3 mt ) . We found that this mutation in motif 3 altered the stoichiometry of binding to BS69CC-MYND from 0 . 15 to 0 . 30 , reduced ΔH to -25 . 7 kcal/mol and reduced -TΔS to 17 . 2 kcal/mol consistent with the loss of a BS69 binding motif ( Fig 2F and S1 Table ) . To confirm these results , we also generated T2 EBNA2348-422 containing more conservative changes in motif 3 where PTLEP was replaced with the equivalent type 1 EBNA2 sequence PSIDP lacking the central leucine . ITC analysis confirmed that this mutation also led to reduced BS69 binding as indicated by altered binding stoichiometry and thermodynamic parameters ( S1 Fig and S1 Table ) . Binding parameters for both motif 3 mutants were similar to those obtained for the type 2 EBNA2 polypeptide containing only motif 1 and motif 2 ( T2 EBNA2348-412 ) ( Fig 1E and S1 Table ) . We conclude that type 2 EBNA2 contains an additional binding site for BS69 that is not present in type 1 EBNA2 . To further examine whether type 2 EBNA2 could form higher-order complexes with the BS69 CC-MYND domain that are larger than type 1 EBNA2 , we examined the properties of BS69-EBNA2 complexes using size exclusion chromatography ( SEC ) . Consistent with complex formation , when pre-incubated with BS69CC-MYND , both T1 EBNA2381-455 and T2 EBNA2348-422 polypeptides migrated through the size exclusion column faster and eluted at a lower elution volume compared to the migration of each component individually ( Fig 3A ) . In line with the binding of additional BS69CC-MYND molecules to T2 EBNA2348-422 and the formation of higher molecular weight complexes , we found that type 2 EBNA2 complexes eluted at a lower volume than type 1 EBNA2 complexes ( Fig 3A ) . SDS-PAGE of SEC column fractions confirmed the presence of BS69CC-MYND and EBNA2 in the higher molecular weight complexes ( Fig 3B ) . Note that both type 1 and type 2 EBNA2 polypeptides migrate anomalously on SDS-PAGE gels and not at their predicted molecular weights ( MW ) of 7 . 9 and 8 . 1 kDa respectively , likely due to their high proline content ( Fig 3B ) . They are however pure and resolve as single species on gel filtration columns ( Fig 3A ) . Because migration in SEC is influenced by both size and shape and BS69CC-MYND has an elongated structure due to the CC domain , we were unable to determine the MW of BS69-EBNA2 complexes accurately using SEC . In order to obtain more accurate MW information that would allow us to determine the number of molecules of BS69CC-MYND and EBNA2 present in type 1 and type 2 complexes , we used SEC with multi-angle light scattering ( SEC-MALS ) ( Table 1 ) . SEC-MALS gave MWs for T1 EBNA2381-455 and T2 EBNA2348-422 that matched the theoretical MW of their monomeric forms and gave a MW for BS69CC-MYND consistent with its dimeric state ( Table 1 ) . For the T1 EBNA2381-455-BS69CC-MYND complex , SEC-MALS gave a MW of 62 . 3 kDa . Given that there are two binding sites for BS69 in the T1 EBNA2381-455 polypeptide , this figure most closely matches the MW of a complex containing two type 1 EBNA2 polypeptides and two BS69CC-MYND dimers ( theoretical MW of 76 . 7 kDa ) rather than a single type 1 EBNA2 polypeptide with one dimer of EBNA2 BS69CC-MYND ( theoretical MW of 38 . 3 kDa ) ( Table 1 ) . For the BS69CC-MYND-T2 EBNA2348-422 complex , SEC-MALs gave a MW of 135 kDa consistent with the larger complex size observed in SEC ( Table 1 and Fig 3A ) . Given the presence of three BS69 binding motifs in type 2 EBNA2 , this MW most closely matches that of a complex containing three BS69CC-MYND dimers and two type 2 EBNA2 polypeptides ( theoretical MW of 107 . 5 kDa ) ( Table 1 ) . Because of the discrepancies in the theoretical and experimentally determined MWs for BS69CC-MYND-EBNA2 complexes , we also used small-angle-X-ray scattering ( SAXS ) to obtain information on the shape and size of these complexes in solution . This technique is routinely applied to directly reconstruct low-resolution shapes of proteins and to generate models of macromolecular assemblies . Initially we used SEC-SAXS to analyse each polypeptide individually . We used a Kratky representation to visualize features of the scattering profiles obtained for T1 EBNA2381-45 , T2 EBNA2348-422 and BS69CC-MYND individually to identify the folding state of the polypeptides in solution . The absence of a bell-shaped curve with a well-defined maximum for both EBNA2 polypeptides indicates that they are natively unfolded in solution ( S2 Fig ) . The bell-shaped curve obtained for the BS69CC-MYND dimer indicates that it is folded in solution as expected from the crystal structure [35] . Three-dimensional models were created for the individual polypeptides by ab initio shape determination . For BS69CC-MYND a solution structure consistent with the coiled-coil dimer structure determined by X-ray crystallography was obtained [35] ( S3 Fig ) . For the EBNA2 polypeptides , solution structures consistent with flexible unfolded peptide chains were obtained ( S3 Fig ) . SAXS analysis of BS69CC-MYND pre-mixed with either type 1 or type 2 EBNA2 polypeptides gave a larger Porod volume ( directly related to MW ) compared to the individual proteins , consistent with complex formation ( S3 Table ) . An ab initio dummy atom model was generated for the type 1 EBNA2-BS69cc-MYND complex and this fitted well to the experimental SAXS data ( χ2 of 1 . 4 ) ( Fig 4A ) . The three-dimensional model generated by ab initio shape determination for the type 1 EBNA2-BS69cc-MYND complex indicated that the complex has a large elongated shape with a volume of 166 nm3 and a maximum dimension ( Dmax ) of 138 Å ( Fig 4B ) . This three-dimensional model could accommodate two BS69CC-MYND dimer structures that were manually docked into the SAXS envelope . The additional space at the bottom of the model was allocated to the model solution structures of two type 1 EBNA2 polypeptides ( Fig 4B ) . This docked structural model for the type 1 EBNA2-BS69CC-MYND complex was then fitted to the experimental scattering data and gave a reasonable χ2 value of 2 . 56 . For comparison a structural model where only one BS69CC-MYND dimer and a single type 1 EBNA2 polypeptide were docked into the SAXS envelope was created but this alternative model gave a worse fit to the experimental data ( S4 Fig ) . An ab initio dummy atom model was then generated for the type 2 EBNA2-BS69CC-MYND complex and this fitted well to the experimental SAXS data ( χ2 of 1 . 0 ) ( Fig 4C ) . The three-dimensional model created by ab initio shape determination for the type 2 EBNA2 complex had a larger volume ( 239 nm3 ) and maximum dimension ( 145 Å ) than the type 1 EBNA2 complex model ( Fig 4D ) . The type 2 EBNA2 model could accommodate the docking of three BS69CC-MYND dimer structures along with two type 2 EBNA2 polypeptides and this structural model gave a good fit to the experimental data ( χ2 of 1 . 44 ) ( Fig 4C and 4D ) . In comparison a docked model containing two BS69CC-MYND dimers and a single type 2 EBNA2 polypeptide gave a worse fit to the experimental data ( S4 Fig ) . Taken together our data indicate that BS69 forms higher order complexes with EBNA2 that involve the interaction of each MYND domain of the BS69 dimer with binding sites in two separate EBNA2 molecules . Rather than an in vitro artefact , this intermolecular ‘bridging’ interaction is consistent with the fact that EBNA2 forms dimers in vivo . Although the N-terminal regions of EBNA2 that mediate dimerisation [36] are absent in the EBNA2 polypeptides we examined in our interaction studies , our data indicate that BS69 may have the capacity to stabilise or enhance dimerisation between two EBNA2 molecules held together through their N-termini . Importantly , using multiple independent techniques , we also demonstrate that type 2 EBNA 2 interacts with an additional BS69CC-MYND dimer . During the course of our ITC experiments we observed that binding data obtained using the longer EBNA2 polypeptides ( T1 EBNA2381-455 and T2 EBNA2348-422 ) showed some deviation from curves fitted using the single binding event ( ‘one set of sites’ ) model ( where binding to multiple sites cannot be detected as separate heat change events ) ( Fig 2D and 2E ) . This suggested that the mode of binding of these polypeptides to BS69CC-MYND could involve more than one distinguishable binding event . To determine whether this was the case , we performed ITC experiments using an increased number of smaller injections of the EBNA2 polypeptide to obtain more data points for curve fitting ( Fig 5 ) . For T1 EBNA2381-455 the binding data did not fit well to curves generated using an alternative two binding event ( ‘two sets of sites’ ) model ( χ2/degrees of freedom = 0 . 56 ) ( Fig 5A and S1 Table ) . This indicates that the deviation of T1 EBNA2381-455-BS69CC-MYND binding data from fitted curves at low molar ratios was unlikely to be the result of a separate binding event ( Figs 2D and 5A ) . In contrast , for T2 EBNA2348-422 the binding profiles obtained fitted well to curves generated using the two binding event model ( χ2/degrees of freedom = 0 . 23 ) ( Fig 5B and S1 Table ) . This enabled the affinity and thermodynamic parameters of the two separate binding events to be determined ( Fig 5B and S1 Table ) . Dissociation constants for these two binding events were both in the nanomolar range ( KD1 = 0 . 009 μM and KD2 = 0 . 091 μM ) . These data indicate that this region of type 2 EBNA2 may adopt a different conformation to type 1 EBNA2 when binding to BS69CC-MYND . In further experiments we addressed the impact of changing the serine at position 409 in type 2 EBNA2 to the aspartate present at the equivalent position ( aspartate 442 ) in type 1 EBNA2 on BS69 binding in the context of the longer type 2 EBNA2 polypeptide containing three BS69 binding sites . To do this we expressed and purified a type 2 EBNA2 polypeptide with an S409D mutation ( T2 EBNA2348-422 SD mutant ) . Interestingly , we found that the SD substitution enhanced the detection of the second binding event on interaction with BS69CC-MYND ( Fig 5C ) . The second binding event for T2 EBNA2348-422 SD was associated with a larger change in enthalpy ( ΔH -7 . 65 kcal/mol ) than the second binding event detected for T2 EBNA2348-422 ( ΔH -1 . 59 kcal/mol ) ( S1 Table ) . Importantly we found that the affinities of the two binding events remained largely unaffected ( Fig 5B and 5C ) , consistent with our earlier observations that the presence of serine at position 409 does not affect the ability of a type 2 polypeptide containing motif 1 and motif 2 to bind BS69 ( Fig 1 ) . To determine whether the impact of the SD mutation was dependent on the presence of BS69 binding motif 3 , we also produced polypeptides containing both the SD and motif 3 mutations ( EBNA2348-422 SD + m3 mt and EBNA2348-422 SD + m3 T1 ) . We found that two binding events were still clearly detectable on interaction of the double mutant forms of type 2 EBNA2 with BS69CC-MYND and that the enthalpy changes on the second binding event were similar to that of the single SD mutant ( Fig 5D , S1 Fig and S1 Table ) , indicating that the impact of the SD change was still evident . We conclude that the SD mutation previously shown to enhance the growth maintenance properties of type 2 EBNA2 [21] does not affect BS69 binding but likely alters the conformation of the type 2 EBNA2 TAD . This may therefore impact on the binding of other transcriptional regulators that influence type 2 EBNA2 function . To confirm our in vitro observations that a type 2 EBNA2 polypeptide binds an additional BS69 dimer , we examined the interaction of BS69CC-MYND with full-length EBNA2 proteins stably expressed in B cells . Lysates from cells expressing type 1 or type 2 EBNA2 or the type 2 SD mutant were incubated with recombinant GST-BS69CC-MYND immobilised on glutathione beads and the amount of EBNA2 precipitated determined by Western Blotting ( Fig 6 ) . During the course of these experiments we observed that incubating GST- BS69CC-MYND with lysates for shorter times allowed us to distinguish differences in binding between type 1 and type 2 EBNA2 ( Fig 6 ) . Consistent with the presence of an additional BS69 binding site in type 2 EBNA2 , we found that GST-BS69CC-MYND pulled down type 2 EBNA2 more efficiently than type 1 EBNA2 at short incubation times ( Fig 6 ) . In agreement with our in vitro observations using the type 2 EBNA2 SD mutant , we found that this protein interacted with BS69CC-MYND with the same efficiency as type 2 EBNA2 ( Fig 6 ) . After 30 minutes incubation , GST-BS69CC-MYND became saturated with EBNA2 and differences in association were no longer evident . A control GST fusion protein ( GST-Rab11 ) did not precipitate EBNA2 , confirming the specificity of the interactions . These data therefore confirm the increased association of BS69CC-MYND with type 2 EBNA2 . To determine whether the presence of the additional BS69 binding motif in type 2 EBNA2 ( motif 3 ) had functional consequences for the activity of type 2 EBNA2 , we first examined the ability of the type 2 EBNA2 motif 3 mutant containing the type 1 sequence substitution ( T2 E2 m3 T1 ) to maintain B cell growth . We utilised a previously described assay using an EBV-infected LCL ( EREB2 . 5 ) in which the activity of a type 1 estrogen receptor-EBNA2 fusion protein can be switched off by estrogen withdrawal [37] . Loss of EBNA2 activity leads to growth arrest , but transfection of a stably-maintained plasmid expressing type 1 EBNA2 into these cells supports their survival [20] . In contrast , the expression of type 2 EBNA2 cannot maintain the growth of these cells [20] . This assay reproduces the growth impairment observed in the early stages of infection with type 2 EBV . We found that mutation of BS69 binding motif 3 created a type 2 EBNA2 protein that was able to support the recovery of these cells from the loss of type 1 EBNA2 activity , with cells growing well by 4 weeks post estrogen withdrawal ( Fig 7A and 7B ) . The type 2 EBNA2 motif 3 mutant behaved similarly to the type 2 EBNA2 SD mutant that was previously shown to support B cell growth in this assay [21] . We conclude that the presence of the additional BS69 binding motif in type 2 EBNA2 impairs the ability of type 2 EBNA2 to maintain B cell growth . We next studied the impact of mutation of the third BS69 binding motif in type 2 EBNA2 on the regulation of gene expression . Type 2 EBNA2 is impaired in its ability to activate transcription of a small number of genes including the viral LMP1 and host CXCR7 genes [20 , 22] . Much weaker and delayed activation of these genes by type 2 EBNA2 is observed during the early stages of primary B cell infection and this impaired activation is also evident on induction of EBNA2 expression in B cell lines infected with EBNA2 deleted viruses e . g . Daudi [20 , 22] . We created a new set of Daudi cell lines stably transfected with plasmids that express type 1 , type 2 , type 2 SD mutant or type 2 m3 T1 mutant EBNA2 proteins from a cadmium-inducible metallothionein promoter , as previously described [21] . Consistent with previous studies , we found that activation of LMP1 and CXCR7 expression by type 2 EBNA2 was virtually undetectable , but that 48 hrs after cadmium addition type 1 EBNA2 efficiently induced expression of these genes [21] ( Fig 7C and 7D ) . Consistent with previous observations [21] , we found that the type 2 EBNA2 SD mutant was also able to induce LMP1 and CXCR7 expression ( Fig 7C and 7D ) . Importantly , we observed that mutation of the third BS69 binding motif in type 2 EBNA2 also rescued the ability of type 2 EBNA2 to activate these genes effectively ( Fig 7C and 7D ) . In conclusion , our functional assays correlate the presence of an additional BS69 binding site in type 2 EBNA2 with a reduced ability to maintain B cell growth and activate gene transcription and show that the mutation of this motif is sufficient to enhance type 2 EBNA2 function . BS69 functions as a negative regulator of EBNA2 transcription activity in reporter assays [34 , 35] , but previous studies have reported that BS69 expression is downregulated on infection of resting B cells by EBV and is low in the resulting immortalised LCLs [35] . Transcriptional repression of BS69 by EBNA2 was implicated in BS69 downregulation indicating that EBNA2 may act to restrict expression of its own negative regulator [35] . The cell lines examined in this previous study all harboured type 1 EBV or type 1 EBNA2 , so we next addressed whether BS69 was expressed at similar levels in cells infected with type 1 and type 2 EBV . We examined BS69 protein levels in type 1 and type 2 LCLs using an anti-BS69 antibody raised against a region within the MYND domain of BS69 . We found that BS69 was expressed at similar levels in type 1 and type 2 LCLs , but surprisingly levels in LCLs were similar to those in an EBV negative B cell line ( AK31 ) ( Fig 8A ) . We expanded our analysis to include additional EBV negative B cell lines ( BJAB and DG75 ) , EBV infected cell lines displaying the EBNA1 only latency I pattern of EBV gene expression ( Akata and Mutu I ) , an additional type 1 LCL ( IB4 ) and a BL cell line expressing all EBV latent proteins including EBNA2 ( Mutu III ) ( both latency III cell lines ) ( Fig 8B ) . We found no correlation between BS69 expression and EBV infection or EBNA2 expression ( Fig 8B ) . BS69 did not therefore appear to be downregulated as a result of EBNA2 expression . We also examined BS69 expression over the course of a primary B cell infection and found that BS69 was not downregulated as previously reported ( Fig 8C ) . We therefore explored the possibility that we were detecting a different isoform of BS69 . Alternative splicing has been reported to give rise to different BS69 isoforms and four have been experimentally verified [33] ( Fig 9A ) . The canonical isoform ( isoform 1 , UniProt identifier: Q15326-1 ) contains 15 exons and encodes a protein of 71 kD ( 602 amino acids ) . Isoform 2 ( Q15326-2 ) lacks amino acids 93–146 encoding the PHD domain ( exon 4 ) and encodes a protein of 64 . 4 kD . Isoform 3 ( Q15326-3 ) lacks amino acids 563–602 encoding the MYND domain ( exon 15 ) and has a unique C-terminus encoded by an extended exon 14 sequence . Isoform 3 encodes a protein of 66 . 6 kD . Isoform 4 ( Q15326-3 ) lacks exon 4 and exon 15 ( and thus both the PHD and MYND domains ) and encodes a protein of 60 kD . These isoforms were previously described as full length ( FL ) , ΔPHD , ΔMD and ΔPHD plus ΔMD respectively [33] , but the exon numbering used in this previous study differed . The BS69 protein detected in Fig 8A and 8B has a molecular weight of approximately 64 kD consistent with that expected for isoform 2 . This was the only protein detected by this antibody ( against the MYND domain ) , indicating that isoform 1 was not expressed in the cell lines examined . Since the antibody we used would not detect BS69 isoforms 3 and 4 , we could not exclude the possibility that one or more of these isoforms was also expressed and that an alternative BS69 isoform was detected previously [35] . In line with this possibility , we noted that the QPCR analysis carried out by Harter et al used primers located in exon 4 , which is absent from isoform 2 . No detail was provided on the anti-BS69 antibody used previously [35] and we were not able to find another antibody that detected isoform 3 and 4 in Western blotting . We therefore took a non-quantitative PCR approach to screen for different BS69 isoforms using cDNA prepared from LCLs and from B cells during a primary EBV infection . PCR using a forward primer in exon 3 and a reverse primer in exon 13 amplified two products indicating the presence of at least two different isoforms , one containing exon 4 ( 1139 bps ) and one lacking exon 4 ( 977 bps ) ( Fig 9B ) . This would be consistent with the presence of isoform 3 ( which contains exon 4 ) and isoform 2 ( which lacks exon 4 and was detected by Western blotting ( Fig 8 ) ) . PCR products were sequenced and their identity confirmed . However , since isoform 4 also lacks exon 4 , this PCR analysis could not rule out the additional presence of isoform 4 . Since in isoforms 3 and 4 exon 15 is replaced by a short unique 3’ sequence from exon 14 , we designed reverse PCR primers in this unique 3’ region . PCR using these primers amplified only one product of 1578 bps consistent with presence of exon 4 and the unique 3’ region ( isoform 3 ) ( Fig 9B ) . The identity of this PCR product was again confirmed by sequencing . Importantly , we did not detect a smaller product ( 1416 bps ) that would indicate the presence of isoform 4 ( lacks exon 4 and exon 15 ) . Our data therefore indicate that LCLs infected with either type 1 or type 2 EBV express both isoform 2 and isoform 3 of BS69 . To quantitatively examine whether either BS69 isoform 2 or isoform 3 were downregulated on EBV infection and in cells expressing EBNA2 as previously described [35] , we used QPCR to analyse BS69 mRNA expression in primary B cells infected by EBV and in a panel of EBV negative and positive B cell lines . QPCR using primers that spanned exon 14 and exon 15 ( present in isoform 1 and 2 ) detected variable levels of BS69 across the cell lines examined , with no obvious correlation with EBV positivity or EBNA2 expression ( present in latency III EBV infected cell lines ) . This is consistent with the variability in BS69 protein expression detected in Western blot analysis of isoform 2 expression ( Fig 8 ) . Although , in one experiment ( #2 ) primary B cells expressed high levels of BS69 isoform 2 mRNA that were reduced on EBV infection , the second primary infection experiment did not reproduce this observation . In fact , primary infection #2 was the same infection analysed by Western blotting in Fig 8C so this change in BS69 RNA expression did not result in decreased expression of BS69 isoform 2 protein . It is most likely therefore that BS69 isoform 2 expression varies in an EBV and EBNA2 independent manner . Analysis of BS69 mRNA expression using QPCR primers that specifically amplify BS69 isoforms containing the long form of exon 14 ( isoforms 3 and 4 ) also detected variable expression of BS69 that did not correlate with EBV positivity or EBNA2 expression indicating that isoform 3 expression is also EBV independent ( Fig 9D ) . We conclude that B cells infected with type 1 or type 2 EBV do not consistently display reduced expression of any detectable isoform of BS69 compared to uninfected B cells . Since BS69 isoform 2 contains the MYND domain that binds EBNA2 ( that is absent in isoform 3 ) , the continued expression of isoform 2 in EBV infected B cells would be expected to restrict the gene activation function of EBNA2 . To determine whether inhibition of BS69 function increased EBNA2 transactivation function , we carried out EBNA2 transactivation assays in an EBV negative B cell line ( BJAB ) in which we overexpressed isoform 3 of BS69 lacking the MYND domain ( ΔMYND ) ( but containing the coiled-coil dimerisation domain ) . This form of BS69 has been proposed to act as a dominant negative inhibitor of the MYND-domain dependent functions of BS69 [33] . We performed transactivation assays using EBNA2-GAL4-DNA binding domain ( DBD ) fusion proteins and a Firefly luciferase reporter plasmid containing a synthetic promoter with 4 GAL4 binding sites . Plasmids expressing GAL4-DBD fusion proteins containing regions of type 1 EBNA2 ( 334–487 ) and type 2 EBNA2 ( 301–454 ) encompassing all BS69 binding motifs were transfected into BJAB cells in the presence or absence of plasmids expressing either full length BS69 ( isoform 1 ) or isoform 3 ( ΔMYND ) . In agreement with previous reports , we found that overexpression of full length BS69 inhibited transactivation by type 1 EBNA2 [34 , 35] ( Fig 10 ) . BS69 also inhibited transactivation by type 2 EBNA2 ( Fig 10 ) . Consistent with its function as a dominant negative inhibitor , we found that expression of BS69 ΔMYND increased transactivation by both type 1 and type 2 EBNA2 ( Fig 10 ) . To determine whether this was a non-specific or EBNA2-dependent effect , we expressed BS69 ΔMYND in the absence of any GAL4-DBD-EBNA2 expressing constructs . In the absence of EBNA2 fusion protein expression , BS69 ΔMYND had no effect on the activity of the GAL4 reporter ( Fig 10 ) . These data therefore demonstrate that inhibition of the MYND-domain dependent function of BS69 in B cells relieves repression of the transactivation function of type 1 and type 2 EBNA2 . These data support our hypothesis that the expression of MYND-domain containing BS69 isoforms in B cells impedes EBNA2 gene activation function . We note that these GAL4 fusion protein reporter assays do not appear to fully recapitulate the function of full length EBNA2 since we observe very little difference in gene activation between type 1 and type 2 EBNA2 , despite the presence of the additional BS69 binding motif in the type 2 EBNA2 sequence present in the fusion protein . We postulate that this is because the fusion proteins are not dimeric and that the stable interaction of BS69 dimers with the low affinity third motif in type 2 EBNA2 requires a bridging interaction between two molecules of EBNA2 . Taken together our in vitro and cell-based assays suggest that during initial B cell infection the increased association of BS69 with type 2 EBNA2 may impede key gene activation events that are required for the efficient outgrowth of immortalised cell lines .
Type 2 EBV strains have reduced B cell transformation capacity and type 2 EBNA2 activates some viral and cell genes less efficiently than type 1 EBNA2 , a feature that may underlie the impaired transformation phenotype . We have identified an additional binding site for the transcriptional repressor BS69 in the EBNA2 protein encoded by type 2 strains of EBV and show that mutation of this additional binding site improves the B cell growth maintenance and gene activation properties of type 2 EBNA2 . Our data therefore implicate increased BS69 association in the impaired function of type 2 EBNA2 . Type 2 EBV transforms resting B cells more slowly and results in the outgrowth of less immortalised cell clones than type 1 EBV [1 , 19] . Although early type 2 EBV transformants show reduced cell growth , the immortalised LCLs that eventually arise from a type 2 EBV infection grow similarly to those infected with type 1 EBV . Type 2 LCLs also maintain similar levels of expression of key EBNA2 target genes [20] . This indicates that the impaired function of type 2 EBNA2 restricts an early stage in the B cell transformation process in vitro . Indeed two EBNA2 target genes that are only weakly activated by type 2 EBNA2 compared to type 1 EBNA2 [22] , the viral oncogene LMP1 and the cell gene CXCR7 , display slower and weaker induction during primary infection with type 2 EBV [20] . Although it was previously reported that BS69 is downregulated on EBV infection , we found that there is continued expression of BS69 isoform 2 in EBV-infected cells . Since this isoform contains the MYND domain that mediates the BS69-EBNA2 interaction the expression of BS69 would be expected to restrict EBNA2 activation function . Consistent with this prediction we found that the expression of a dominant negative form of BS69 ( isoform 3 ) lacking the MYND domain enhances EBNA2 activation function in B cells . Our data are consistent with a model where BS69 acts as a restriction factor for both type 1 and type 2 EBNA2 but the association of type 2 EBNA2 with more molecules of the BS69 repressor protein further restricts the activation of growth and survival genes important in early transformation . Why a small number of specific genes are activated less well by type 2 EBNA2 is as yet not fully clear , but sequences resembling EICEs ( bound by ETS and IRF transcription factors ) are found at EBNA2 binding sites in the LMP1 promoter and binding sites closest to the cell genes that show reduced activation by type 2 EBNA2 . Interestingly BS69 binding to PXLXP motifs in ETS2 has been shown to inhibit its transactivation activity [32] . Since the ETS family member PU . 1 is known to bind to the putative EICE in the LMP1 promoter and plays a role in LMP1 promoter activation [38] , it is possible that type 2 EBNA2 functions less well in the context of PU . 1 binding sites . Interestingly , PU . 1 also contains PXLXP motifs that would be predicted to bind BS69 , so enhanced tertiary complex formation between type 2 EBNA2 , BS69 and PU . 1 at the regulatory elements of specific genes may function to stabilise BS69 binding and further restrict gene activation by type 2 EBNA2 . Our data also provide important new molecular information on the nature of the complexes formed between EBNA2 and BS69 that may be applicable to the way BS69 interacts with other cellular and viral transcription factors via its MYND domain . Single and multiple PXLXP BS69 binding motifs have been identified in the cell and viral binding partners of BS69 , but the elucidation of the structure of the dimeric coiled-coil-MYND domain of BS69 led to a model that proposed that a BS69 dimer bound to the two adjacent PXLXP motifs ( motif 1 and motif 2 ) in the same type 1 EBNA2 molecule [35] . However , although the authors found that BS69 bound with increased affinity when two PXLXP motifs were present in EBNA2 polypeptides , the three-dimensional structure obtained comprised a BS69CC-MYND dimer bound to two separate type 1 EBNA2 motif 1 peptides . Although the binding of motif 1 and motif 2 could be accommodated in the BS69CC-MYND structure if the intervening 52 amino acids were looped out , the formation of this complex has not been formally demonstrated [35] . Our SEC-MALS and SAXS analysis provides the first evidence that the BS69CC-MYND dimer preferentially forms an intermolecular bridge between PXLXP motifs located on different EBNA2 molecules . This mode of binding is consistent with the fact that EBNA2 is a dimeric protein , with dimerisation mediated by the N-terminal END domain comprising amino acids 1–58 [36] . Additional self-associating regions have also been mapped elsewhere in EBNA2 and include amino acids 97–121 and 122–344 [39 , 40] , although no molecular information is available on how these regions may contribute to dimerisation . Our data indicate that BS69 binding to sites in the C-terminal transactivation domain may contribute to the formation or stabilisation of EBNA2 dimers . Interestingly , although SEC analysis clearly demonstrated complex formation between both type 1 and type 2 EBNA 2 polypeptides and BS69CC-MYND , the elution profiles of both complexes were broad . The type 1 EBNA2-BS69CC-MYND elution profile had a clear shoulder indicating the presence of smaller MW complexes ( Fig 3A ) . This would explain why the average MW determined by SEC-MALs was smaller than expected for a complex that contained two molecules of type 1 EBNA2 and two BS69CC-MYND dimers . It is possible that in solution in vitro there is a mixed population of dimeric type 1 EBNA2 and monomeric type 1 EBNA2 complexes ( where a single EBNA2 polypeptide is bound by one BS69CC-MYND dimer as previously proposed ) . We were not able to investigate this further using SAXS as this ‘shoulder’ was not clearly defined , so SAXS analysis for both type 1 and type 2 EBNA2-BS69 complexes focused on the major elution peak of the large complex . Given that full length EBNA2 expressed in EBV-infected cells is a dimer , complexes involving two EBNA2 molecules are more likely to be physiologically relevant . Surprisingly , in our GAL4-EBNA2 fusion protein assays we did not see weaker transactivation by the type 2 EBNA2 fusion protein compared to the type 1 EBNA2 fusion protein as reported previously [21] . We used a longer region of EBNA2 compared to this previous study that encompassed all three BS69 binding sites for type 2 EBNA2 and the corresponding region of type 2 EBNA2 ( with only two functional BS69 binding sites ) . Previously GAL4-EBNA2 fusion protein constructs were used that expressed a type 1 EBNA2 protein containing only BS69 binding motif 2 or the corresponding region of type 2 EBNA2 that contained BS69 binding motif 2 and 3 [21] . It is not completely clear why the increased association of BS69 with type 2 EBNA2 is not associated with weaker transactivation in our assays in the context of a longer region of EBNA2 , but it could point to the importance of the dimerisation that occurs in the context of the full-length protein in the assembly of larger BS69-EBNA2 complexes that involve the low affinity third BS69 binding site . When considering the nature of assembly of BS69-EBNA2 complexes , it is likely that binding to motif 1 ( which in type 1 EBNA2 has the highest affinity for BS69CC-MYND ) would drive the initial interaction between EBNA2 and BS69 . Binding to motif 1 probably constitutes the first binding event that can be distinguished in our detailed ITC analysis . For type 2 EBNA2 , since both motif 2 and 3 bind BS69 with similar affinity , binding to both of these motifs probably occurs with similar kinetics and is detectable as a single second binding event by ITC . Given the fact that BS69CC-MYND dimers are predicted in the solution structure of the BS69-EBNA2 complex to be located side by side along a dimeric EBNA2 molecule , it is possible that interactions between BS69 coiled-coil dimers play a role in stabilising the oligomeric complex . Our initial interest in examining type-specific binding of EBNA2 to BS69 centred around the influence of a serine residue in the TAD of type 2 EBNA2 that plays a key role in restricting B cell growth maintenance by type 2 EBNA2 [21] . Although this residue is located immediately adjacent to BS69 binding motif 2 in type 2 EBNA2 , we found that it did not increase BS69 binding ( as might have been expected ) when binding was compared to the corresponding region of type 1 EBNA2 where there is an aspartate residue in its place . It does not appear therefore that the influence of serine 409 on growth maintenance and gene activation is mediated through alterations in BS69 binding affinity . Our ITC analysis however did find that a serine to aspartate change at this position in type 2 EBNA2 altered the nature of BS69 binding indicating that it may induce a conformational change in this region of EBNA2 . This could result in differences in the binding of other transcription regulators to the type 2 EBNA2 TAD compared to the type 1 EBNA2 TAD . Possibilities could include increased binding of a repressor or co-repressor to the type 2 EBNA2 TAD or decreased binding of an activator or co-activator . BS69 may have a wider role in regulating B cell transformation and the growth of EBV-infected cells in addition to its modulation of EBNA2 transactivation . BS69 localised to the cell membrane has also been implicated as an adaptor in signalling mediated by the EBV oncogene LMP1 . The MYND domain of BS69 was reported to bridge an interaction between the carboxy terminal cytoplasmic domain of LMP1 and the TRAF6 signalling protein to activate the JNK signalling pathway [41] . Conversely , BS69 has also been implicated as a negative regulator of LMP1-mediated NF-κB signalling by decreasing the association between C-terminal activation region ( CTAR ) 2 of LMP1 and the signalling adaptor TRADD [42] and by binding to CTAR1 and bringing in the negative regulator of NF-κB signalling , TRAF3 [43] . Although further work appears to be required to fully understand the role of BS69 in LMP1 signalling and the relative proportions of nuclear and membrane-associated BS69 , it is possible that BS69 is a key modulator of growth promoting events in EBV-infected cells . In this context , our work now sheds new light on how transformation by type 2 strains of EBV may be specifically curbed as a result of sequence variation that results in the creation of an additional binding site for BS69 .
All cell lines were passaged twice weekly in RPMI 1640 media ( Invitrogen ) supplemented with 10% Fetal Bovine serum ( Gibco ) , 1 U/ml penicillin G , 1 μg/ml streptomycin sulphate and 292 μg/ml L-glutamine at 37°C in 5% CO2 . DG75 [44] and AK31 [45] are EBV negative BL cell lines and BJAB [46] is an EBV negative B cell lymphoma line . Akata [47] and Mutu I are EBV positive latency I BL cell lines and Mutu III is a cell line derived from Mutu I cells that drifted in culture to express all EBV latent proteins ( latency III ) [48] . Daudi is a latency III BL cell line carrying EBV with a deletion in EBNA2 [49] . All LCLs also display the latency III pattern of EBV gene expression and were described previously [50]; IB4 , spLCL , LCL3 , C2 + Obaji , JAC-B2 , BM + Akata LCLs are infected with type 1 EBV and C2 + BL16 , WEI-B1 , Jijoye and AFB1 LCLs are infected with type 2 EBV . The ER-EB 2 . 5 LCL , expressing a conditionally active oestrogen receptor ( ER ) -EBNA2 fusion protein , was provided by Prof B . Kempkes and was cultured in the presence of β-estradiol [37] . B cell infection samples were described previously [51] . For the LMP1 and CXCR7 gene expression analysis , a new set of Daudi cell lines were established as described previously [21] . 2 x 106 Daudi cells were transfected with 6 μg pHEBo-MT constructs expressing full length type 1 , type 2 or type 2 mutant EBNA2 using Neon transfection with 1 pulse of 1400 V for 30 msec . Cells were left to recover for 48 hrs before initial selection in 100 μg/ml hygromycin B for a further 48 hrs . Cultures were then diluted 10–25 fold and the hygromycin B concentration increased to 300 μg/ml . Cell lines were selected for 4 weeks prior to induction of EBNA2 expression by the addition of 2–10 μM cadmium chloride for time points up to 48 hrs . Constructs expressing N-terminal 6 x histidine tagged EBNA2 polypeptides were generated using the Sequence and Ligation Independent Cloning ( SLIC ) technique using type 1 EBNA2 ( B95-8 ) , type 2 EBNA2 ( AG876 ) and type 2 EBNA2 SD ( serine to aspartate at position 409 ) pSG5 expression plasmids as templates to amplify the regions of interest . DNA was PCR amplified using primers containing 20–30 bp of additional sequence from the regions 5’ and 3’ to the multiple cloning site of pET47b+ ( S2 Table ) . pET47b+ was digested using SmaI and HindIII and the PCR product and double-digested vector were then partially digested using the 3’ to 5’ exonuclease activity of T4 DNA Polymerase in the absence of dNTPs to generate long complementary 5’ overhangs . The PCR products and pET47b+ were then annealed on ice . The ligated DNA fragments obtained contain four nicks that are repaired by E . coli after transformation . For type 1 EBNA2 , constructs encoded amino acids 381–445 or 381–455 to generate pET47b+ T1 EBNA2381-445 and pET47b+ T1 EBNA2381-455 . Type 2 EBNA2 constructs encoded amino acids 348–412 or 348–422 to generate pET47b+ T2 EBNA2348-412 , pET47b+ T2 EBNA2348-422 and pET47b+ T2 EBNA2348-422 SD . The BS69 CC-MYND domain ( amino acids 480–602 ) was amplified from pCI-BS69 containing the full length human BS69 sequence ( gift from Dr Stéphane Ansieau ) and cloned using SLIC into the SmaI and HindIII sites of pET49b+ to generate a construct expressing an N-terminal GST-6x Histidine tag BS69CC-MYND fusion protein ( S2 Table ) . To create GAL4-DNA-binding domain-EBNA2 TAD fusion protein expressing constructs pBlueScript plasmids carrying EBNA2 sequences were used as the template to PCR amplify the type 1 EBNA2 TAD ( amino acids 426–463 ) and the type 2 EBNA2 TAD ( amino acids 334–487 ) using Taq DNA polymerase . Primers contained BamHI or NotI restriction sites at their 5’ ends ( S2 Table ) . PCR products were first cloned into pCR2 . 1 using the TA cloning kit ( Invitrogen ) according to the manufacturer’s instructions . The pCR2 . 1 vector carrying the cloned PCR product was then digested with BamHI and NotI and the EBNA2 TAD fragment then cloned into the BamHI and NotI sites of pcDNA3 . 1-GAL4-DBD . The Q5 Site-Directed Mutagenesis kit ( NEB ) and primers carrying specific mutations ( S2 Table ) were used to mutate BS69 binding motif 3 in the type 2 EBNA2 constructs pET47b+ T2 EBNA2348-422 , pET47b+ T2 EBNA2348-422 SD , OriP-p294 T2 E2 and pHEBo-MT:E2T2 Motif 3 ( PTLEP ) was mutated to ATAEA to generate T2 E2 m3 mt constructs or to the corresponding type 1 sequence ( PSIDP ) to generate T2 E2 m3 T1 constructs . The EREB2 . 5 growth assay was performed as described previously [20] . Briefly , 5 μg of OriP-p294 plasmids expressing type 1 EBNA2 , type 2 EBNA2 or type 2 EBNA2 mutants were transfected into 5 × 106 EREB 2 . 5 cells resuspended in 110 μl of buffer T using Neon transfection with 1 pulse of 1300 V for 30 msec . Following transfection , cell suspensions were added to 2 ml of media supplemented with 10% FBS and antibiotics but without β-oestradiol and incubated overnight in 12-well plates . The following day each transfected sample was made up to 10 mls with media and divided into 5 x 2 ml aliquots in a 12-well plate . Samples were harvested for cell counting and protein analysis at time points up to 4 weeks . Cells were diluted 1:3 in fresh culture medium one day before transfection . 2 x 106 BJAB cells were used for each individual transfection . Cells were pelleted by centrifugation at 335g for 5 mins at 4ºC and washed twice with pre-warmed PBS . Cells were resuspended in 100 μl of Neon resuspension solution R ( Invitrogen ) . Cell suspensions were then mixed with plasmid DNA ( 2–12 μg in TE buffer ) . Cells were co-transfected with 300 ng of either type 1 GAL4-DBD:EBNA2 ( aa 334–487 ) or type 2 GAL4-DBD:EBNA2 ( 301–454 ) constructs , 500 ng of pFRLuc ( Agilent technologies ) , 10 ng of pRL-CMV ( Promega ) and 1 μg of BS69 ( pCI-BS69 ) or BS69 ΔMYND ( pCI-BS69-ΔMYND ) expressing plasmids ( gift from Dr Stéphane Ansieau ) . The DNA and cell mixture was transferred to a 100 μl Neon transfection pipette tip ( Invitrogen ) . Cells were electroporated using Neon transfection protocol 14 ( 2 pulses of 1200 V for 20 ms ) and then transferred into 2 ml of pre-warmed media in a 6-well plate and incubated at 37ºC for 24 hrs . Cell pellets were then lysed using 100 μl of 1X Passive Lysis buffer ( Promega ) . Two freeze-thaw cycles were performed to achieve efficient lysis ( 20 sec on dry ice and thawing at room temperature for 2 mins ) . Cell debris was removed by centrifugation at 25 , 000g , for 1 min at 4ºC and the clear supernatant was then transferred to a fresh tube . 20 μl of lysate was assayed for firefly and Renilla luciferase activity using 20 μl of each dual luciferase assay kit reagent ( Promega ) and a microplate luminometer ( LUMIstar Omega , BMG Labtech ) . pGEX4T1-BS69 ( kindly provided by Dr Stéphane Ansieau ) was used to express a GST-BS69 fusion protein containing amino acids 452–602 of BS69 encompassing the CC-MYND domain ( numbered according to the canonical isoform ) [34] . pGEX4T1-RAB11B expressing GST-tagged RAB11B ( gift from Prof Gill Elliott ) was used to produce a negative control protein for the GST pull-down assays . For production of BS69CC-MYND and EBNA2 polypeptides , the relevant plasmids were transformed into the Rosetta 2 ( DE3 ) pLysS E . coli strain and protein expression induced by adding 0 . 4 mM of isopropyl βD-1-thiogalactopyranoside ( IPTG ) to 3 litre cultures at an OD600nm of 0 . 6 . The bacteria were then grown at 20°C overnight before harvesting for protein purification . Cell pellets from a 3 litre induced culture were lysed for 30 mins on ice with constant stirring in 100 ml of lysis buffer ( 25 mM Tris-HCL pH 7 . 5 , 500 mM NaCl , 5% Glycerol ) . The lysis buffer was supplemented with 0 . 25 mg/ml lysozyme , 2 mM MgCl2 , 1 mM TCEP ( tris ( 2-carboxyethyl ) phosphine ) , two protease-inhibitor complete tablets ( Roche ) and DNase and 0 . 2 mg/ml of DNase I . Lysates were then sonicated at 37% amplitude for 5 mins with 10 seconds pulses using a Vibra-cell sonicator ( SONICS ) . The cell debris was pelleted at 15000 rpm for 45 mins at 4°C ( Biofuge Stratos , Heraeus ) . Beads from 3 ml of HisPur Cobalt Resin slurry ( Thermo Scientific ) were added to the cleared lysate along with 2 mM imidazole and the sample incubated for 1 . 5 hrs at 4°C with rolling . Samples were decanted into a centrifuge column ( Thermo Scientific Pierce ) and washed with buffer ( 25mM Tris-HCl , 500mM NaCl and 1mM TCEP , 2 mM of imidazole , pH 7 . 5 ) . The protein was eluted from the beads using buffer containing increasing concentrations of imidazole ( 5 mM , 10 mM , 20 mM , 50 mM , 100 mM , 200 mM , 300 mM and 500 mM ) . Fraction samples were analysed by SDS-PAGE and the fractions containing recombinant protein were pooled and incubated with 3C protease ( 200 μl of 2 mg/ml ) in the presence of 2 mM DTT overnight at 4°C to cleave the Histidine tag . Cleaved proteins were then separated from the 3C protease by passing the protein sample through a GSTrap HP column ( Amersham ) using a peristaltic pump to capture the GST-tagged 3C protease . Untagged recombinant proteins were then concentrated and injected into a HiLoad 16/600 Superdex 75 pg column ( GE Healthcare ) pre-equilibrated in 25 mM Tris-HCL , 200 mM NaCl , 1 mM TCEP , pH 7 . 5 purified at 0 . 5 ml/min . Protein fractions containing purified protein were then pooled , concentrated and stored at -80°C until required . Approximately 1 mg of purified EBNA2 polypeptides or 4 mg of BS69CC-MYND was obtained from 1 litre of culture . Four commercially synthesized peptides ( Peptide Synthetics ) were used for ITC . These included BS69 binding motif 2 of type 1 EBNA2 ( 435–445 ) or type 2 EBNA2 ( 402–412 ) and putative BS69 binding motif 3 of type 1 EBNA2 ( 445–455 ) or type 2 EBNA2 ( 412–422 ) . Frozen protein was quickly thawed using running water and dialysed overnight at 4°C using Slide-A-Lyzer MINI Dialysis Units ( Thermo Scientific ) against ITC buffer ( 20mM Tris-HCl , 100mM NaCl and 1mM TCEP , pH 7 . 5 ) . The next day , protein samples were centrifuged at 13000 rpm for 10 mins at 4ºC and the concentration was determined by NanoDrop spectroscopy ( NanoDrop Technologies ) with their respective molecular weights and extinction coefficients . EBNA2 peptides ( 1mM ) and polypeptides ( type 1 , 0 . 3 mM and type 2 , 0 . 6 mM ) were titrated against BS69CC-MYND ( 0 . 1mM ) at 25°C using a MicroCalTM iTC200 instrument ( Malvern ) . For peptides , 13 x 3 . 0 μl injections were used for titration . For EBNA2 polypeptides 19 x 2 . 0 μl or 29 x 1 . 3 μl injections were used for titration . ITC data were corrected for non-specific heat and analysed using MicroCal Origin 7 . 0 . The experiments were performed in triplicate alongside a control experiment with no BS69CC-MYND ( buffer only in the cell ) . All polypeptides were used within 24 hrs of dialysis into ITC buffer . An S200 10/300 GL gel filtration column ( GE Healthcare ) was equilibrated with buffer containing 20mM Tris-HCl , 100mM NaCl and 1mM TCEP , pH 7 . 5 . Individual EBNA2 or BS69CC-MYND polypeptides or complexes were applied to the column and analysed at a flow rate of 0 . 5ml/min . The eluted fractions were then analysed by SDS-PAGE and Quick Coomassie staining . EBNA2-BS69 complexes were prepared by pre-incubating proteins in a 1:3 molar ratio for at least 30 mins at 4°C . Purified samples ( 45 μl ) at a concentration of 5 mg/ml were loaded onto a Shodex KW403-4F column at 25°C pre-equilibrated in 20 mM Tris-HCl , 100mM NaCl and 1mM TCEP , pH 7 . 5 . Elution fractions were monitored using a DAWN HELEOS II MALS detector followed by a refractive index detector Optilab T-rEX ( Wyatt Technology ) . Molecular masses of each individual peak were determined using ASTRA 6 software ( Wyatt Technology ) . For normalization of the light scattering and data quality , BSA was used as a calibration standard . Synchrotron radiation X-ray scattering data from solutions of individual proteins or complexes prepared as for SEC-MALS were collected on beamline B21 at Diamond Light Source ( Didcot , United Kingdom ) , with an inline HPLC system . X-ray scattering patterns were recorded on a Pilatus detector after injection of 45 μl of protein sample ( 5–10 mg/ml ) in a Superdex 200 3 . 2/300 column equilibrated in 20mM Tris-HCl , 100mM NaCl , 2% Sucrose and 1mM TCEP , pH 7 . 5 . Samples were analysed at 20°C using a flow-rate of 0 . 25 ml/min . Initial data processing ( background subtraction and radius of gyration Rg calculation ) was performed using ScÅtter ( v3 . 0 by Robert P . Rambo; Diamond Light Source ) . Ab initio beads model for the complex were prepared using DAMMIF [52] . 23 independent dummy atom models were obtained by running the program in ‘slow’ mode . DAMAVER was then used to align and average the models [53] . The ab initio generated beads models were refined using DAMMIN and compared to the experimental scattering data to derive χ2 values [54] . The goodness-of-fit χ2values for the docked structure compared to the experimental scattering data were determined with FoXS [55] . Nuclear extracts were prepared from control or EBNA2 expressing Daudi cell lines . EBNA2 expression in Daudi:pHEBo-MT:EBNA2 cells was induced with 5 μM CdCl2 for 24 hrs . At least 4x107 cells were then harvested and resuspended in 1 ml of buffer A ( 10 mM HEPES pH 7 . 9 , 1 . 5 mM MgCl2 , 10 mM KCl , 0 . 5 mM 1 , 4-dithiothreitol ( DTT ) ( Sigma ) , 1 mM PMSF ( Sigma ) and 1x complete protease inhibitor cocktail ( Roche ) ) . Cells were pelleted by centrifugation at 1000g for 5 mins at 4ºC and lysed in 100 μl of buffer A supplemented with 0 . 1% ( v/v ) NP-40 and incubated on ice for 5 mins . Cell lysates were centrifuged at 2700g for 30 sec at 4ºC and the nuclei resuspended in 50 μl of buffer B ( 20 mM HEPES pH 7 . 9 , 420 mM NaCl , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 1 mM PMSF , 25% ( v/v ) glycerol , 1 mM DTT and 1x complete protease inhibitor cocktail ) at 4ºC for 20 mins with rotation . Samples were finally centrifuged at 11 , 600g for 10 mins at 4ºC and the supernatants/nuclear extracts were transferred to fresh eppendorf tubes and the protein concentration was determined before storage at -80ºC . Lysates containing GST-tagged BS69CC-MYND or GST-RAB11B were prepared from 100 ml cultures of E . Coli BL21 ( DE3 ) . Transformed cells were cultured at 30ºC until they reached an OD600nm of 0 . 6 and protein expression was induced at 25ºC with 0 . 5 mM IPTG for 3–4 h . Cells were pelleted at 2 , 800g for 20 mins at 4ºC and then resuspended in 10 ml of Lysis buffer ( 20 mM Tris-Cl pH 8 . 0 , 150 mM NaCl and 1 mM DTT ) supplemented with 120 μl of lysozyme ( 10 mg/ml ) and lysates incubated on ice for 20 mins . Lysates were sonicated at high speed for 3 x 15 sec pulses in ice water using an Ultrasonic XL2020 Processor ( Heat Systems ) and cell debris pelleted at 17 , 900g for 30 min at 4ºC . Lysates were stored at -80ºC until required . For pull-down assays , 50 μl of 50% Glutathione-Sepharose 4B Bead slurry ( GE Healthcare ) was washed three times in ice-cold binding buffer ( 20 mM Tris-Cl pH 8 . 0 , 150 mM NaCl , 1 mM DTT and 0 . 1 mg/ml BSA ) . Beads were pelleted by centrifugation at 25 , 000g for 1 min and 100 μl of bacterial lysate containing the GST-tagged protein was incubated with the washed beads for 1 hr at 4ºC with rotation . Glutathione-Sepharose Beads bound to the GST-tagged protein were then washed with ice-cold binding buffer six times and pelleted by centrifugation at 25 , 000g for 1 min . Loaded GST- BS69CC-MYND beads were then incubated with nuclear extracts containing EBNA2 at 4ºC for different times ( 5 , 10 and 30 mins ) . Loaded GST-RAB11B beads were incubated with lysates for 30 mins . Beads were then washed six times with ice-cold binding buffer and pelleted by centrifugation at 25 , 000g for 1 min . Beads were then resuspended in 25 μl of 2x SDS sample buffer ( 120 mM Tris-Cl pH 6 . 8 , 4% ( w/v ) SDS , 2% ( v/v ) β-mercaptoethanol , 20% ( v/v ) glycerol and 0 . 01% ( w/v ) bromophenol blue ) and incubated at 95ºC for 5 mins and analysed for EBNA2 levels by SDS-PAGE and Western blotting . SDS-PAGE and Western blotting was carried out as described previously [56 , 57] using the anti-EBNA2 monoclonal antibody PE2 1/300 , anti-LMP1 monoclonal antibody CS1-4 1/300 ( gifts from Prof M . Rowe ) , anti-actin 1/5000 ( A-2066 , Sigma ) and anti-BS69 1/1000 ( ab190890 , Abcam ) . Western blot visualisation and signal quantification was carried out using a Li-COR Imager . Gels were stained using Quick Coomassie stain ( Generon Ltd ) . Total RNA was extracted from cells using TriReagent ( Sigma ) , further purified using the RNeasy kit ( Qiagen ) and cDNA synthesised using random primers and the ImProm II reverse transcription kit ( Promega ) . Standard PCR reactions were performed with Phusion DNA polymerase ( New England Biolabs ) using the relevant BS69 primers listed in S2 Table . Quantitative PCR was performed in duplicate using the standard curve absolute quantification method on an Applied Biosystems StepOnePlus real-time PCR machine as described previously [58] using the relevant primers listed in S2 Table previously described control primers: GAPDH ( MW84 and MW85 ) [59] and β2 microglobulin ( MW1447 and MW1448 ) [60] . The efficiency of all primers was determined prior to use and in each experiment and all had amplification efficiencies within the recommended range ( 90–105% ) . | Epstein-Barr virus ( EBV ) drives the development of many human cancers worldwide including specific types of lymphoma and carcinoma . EBV infects B lymphocytes and immortalises them , thus contributing to lymphoma development . The virus promotes B lymphocyte growth and survival by altering the level at which hundreds of genes are expressed . The EBV protein EBNA2 is known to activate many growth-promoting genes . Natural variation in the sequence of EBNA2 defines the two main EBV strains: type 1 and type 2 . Type 2 strains immortalise B lymphocytes less efficiency and activate some growth genes poorly , although the mechanism of this difference is unclear . We now show that sequence variation in type 2 EBNA2 creates a third site of interaction for the repressor protein ( BS69 , ZMYND11 ) . We have characterised the complex formed between type 2 EBNA2 and BS69 and show that three dimers of BS69 form a bridged complex with two molecules of type 2 EBNA2 . We demonstrate that mutation of the additional BS69 interaction site in type 2 EBNA2 improves its growth-promoting and gene induction function . Our results therefore highlight a molecular mechanism that may contribute to the different B lymphocyte growth promoting activities of EBV strains . This aids our understanding of immortalisation by EBV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"chemical",
"characterization",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"organisms",
"viruses",
"materials",
"science",
"dna",
"viruses",
"sequ... | 2019 | Increased association between Epstein-Barr virus EBNA2 from type 2 strains and the transcriptional repressor BS69 restricts EBNA2 activity |
Selective pressures between hosts and their parasites can result in reciprocal evolution or adaptation of specific life history traits . Local adaptation of resident hosts and parasites should lead to increase parasite infectivity/virulence ( higher compatibility ) when infecting hosts from the same location ( in sympatry ) than from a foreign location ( in allopatry ) . Analysis of geographic variations in compatibility phenotypes is the most common proxy used to infer local adaptation . However , in some cases , allopatric host-parasite systems demonstrate similar or greater compatibility than in sympatry . In such cases , the potential for local adaptation remains unclear . Here , we study the interaction between Schistosoma and its vector snail Biomphalaria in which such discrepancy in local versus foreign compatibility phenotype has been reported . Herein , we aim at bridging this gap of knowledge by comparing life history traits ( immune cellular response , host mortality , and parasite growth ) and molecular responses in highly compatible sympatric and allopatric Schistosoma/Biomphalaria interactions originating from different geographic localities ( Brazil , Venezuela and Burundi ) . We found that despite displaying similar prevalence phenotypes , sympatric schistosomes triggered a rapid immune suppression ( dual-RNAseq analyses ) in the snails within 24h post infection , whereas infection by allopatric schistosomes ( regardless of the species ) was associated with immune cell proliferation and triggered a non-specific generalized immune response after 96h . We observed that , sympatric schistosomes grow more rapidly . Finally , we identify miRNAs differentially expressed by Schistosoma mansoni that target host immune genes and could be responsible for hijacking the host immune response during the sympatric interaction . We show that despite having similar prevalence phenotypes , sympatric and allopatric snail-Schistosoma interactions displayed strong differences in their immunobiological molecular dialogue . Understanding the mechanisms allowing parasites to adapt rapidly and efficiently to new hosts is critical to control disease emergence and risks of Schistosomiasis outbreaks .
Schistosomiasis is the second most widespread human parasitic disease after malaria and affects over 200 million people worldwide [1] . Schistosoma mansoni ( Platyhelminthes , Lophotrochozoa ) causes intestinal schistosomiasis . Schistosoma needs a fresh water snail acting as its first intermediate host to undergo part of its life cycle before infecting humans . Patently infected snails support the continuous production of thousands of cercariae , infective for humans . Vector snails are central actors of the parasite transmission and obvious targets for schistosomiasis control that deserve more attention . It is therefore necessary to understand snail-parasite immunobiological interactions and to characterize the molecular mechanisms of successful snails and Schistosoma interactions . The compatibility of numerous strains of Biomphalaria glabrata and Schistosoma sp . has been extensively tested , revealing that ( i ) different B . glabrata laboratory strains ( or isolates ) show various degrees of susceptibility to S . mansoni infection and ( ii ) different strains of S . mansoni display different levels of infectivity towards a particular strain of snail host [2–6] . Compatibility is defined as the ability for the miracidia to infect snail and become a living primary sporocyst in snail tissue . Incompatibility refers to miracidia that are recognized by the snail immune system and encapsulated and killed by the hemocytes ( the snail immune cells ) . Thus , the success or failure of the infection of B . glabrata by S . mansoni reflects a complex interplay between the host’s defense mechanisms and the parasite’s infective strategies , based on a complex phenotype-to-phenotype or matching-phenotype model [2–4 , 7–9] . In the past 15 years , the molecular basis of this compatibility polymorphism has been investigated at the genomic [10–12] , transcriptomic [8 , 13–17] , proteomic/biochemical [18–23] and epigenomic levels [24–29] . These studies have revealed that various molecules and pathways involved in immune recognition ( snail immune receptors versus parasite antigens ) , immune effector/anti-effector systems , and immune regulation/activation participate in a complex interplay that governs the match or mismatch of host and parasite phenotypes [30] . This complex phenotype-by-phenotype interaction or compatibility polymorphism varies between populations and individuals resulting in a "multi-parasite susceptibility" or "multi-host infectivity" phenotypes [4] that reflect between-population variations in parasite infectivity/virulence and host defense/resistance [31 , 32] . Most of the time , interaction in B . glabrata/Schistosoma models has been investigated by comparing , ( i ) sympatric/compatible and ( ii ) allopatric/incompatible host-parasite associations . The general assumption is that the parasites thanks to their shorter generation times , larger population sizes and higher reproductive outputs , are ahead in the co-evolutionary race against their host and are therefore more likely to locally adapt and perform better when infecting local hosts [33 , 34] , than allopatric hosts [34–37] . However , in many instances , Schistosomes are highly compatible to hosts from other localities , showing the same or even greater infection success when exposed to allopatric hosts . Thus they do not fulfil the "local versus foreign" main criterion of the local adaptation between a host and its parasite [5 , 38–40] . Very few studies have investigated the molecular basis of allopatric compatible interactions from the perspective of both side of the interaction , the host and the parasite [41 , 42] . Hence , in order to bridge this gap , we herein study sympatric/allopatric interactions displaying similar compatibilities using an integrative approach that links the underlying molecular mechanisms to the resulting phenotypes , based on comparative molecular approaches on both host snails and Schistosoma parasites . We characterize the underlying cellular and molecular mechanisms of the interaction between South American snail strains ( from Recife Brazil and Guacara Venezuela ) and three different highly compatible parasite isolates: ( i ) the sympatric strains of S . mansoni from Recife Brazil , ( ii ) the allopatric S . mansoni from Guacara Venezuela ( narrow geographic scale ) , and ( iii ) the allopatric S . rodhaini from Burundi Africa ( large geographic and phylogenetic scales ) . Our results clearly show that even though the compatibility phenotypes among these strains is similar , a very different immunobiological dialogue is taking place between B . glabrata vector snails and their sympatric or allopatric Schistosoma parasites at the cellular and molecular levels .
The B . glabrata transcriptome was analyzed using the previously described RNAseq pipeline developed in our laboratory [8 , 43 , 44] . Of the 159 , 711 transcripts of the BgBRE transcriptome , 3 , 865 ( 2 . 4% ) were differentially represented in all sympatric and allopatric conditions compared to naive snails ( Table 1 , S1 Fig ) . We performed automatic Blast2GO annotation , discarded the non-annotated transcripts , and retained 1 , 017 annotated transcripts ( 26 . 3% of the differentially expressed ( DE ) transcripts , S1 Fig ) . In the following analysis , we focused on the 336 transcripts known to have immune-related functions ( 8 . 7% of DE transcripts , S1 Fig ) . Of these immune related transcripts , 189 , 180 , and 164 DE transcripts were identified in the BB ( BgBRE/SmBRE , sympatric ) , BV ( BgBRE/SmVEN , allopatric ) , and BR ( BgBRE/Srod , allopatric ) interactions , respectively ( Fig 1A ) . Among those , 40 transcripts were consistently differentially expressed in response to infection ( Fig 1A ) . They also displayed similar expression profile ( Fig 1B , cluster 1 ) . Most ( 74 . 1% ) of the transcripts differentially expressed in response to infection by the sympatric parasite ( BB ) were not differentially expressed in response to either one of the two other parasites . Most importantly , all of the sympatric-specific transcripts were under-represented at 24 h post-infection , and 74 . 6% of these transcripts were differentially expressed exclusively at this time point ( Fig 1B , cluster 5 ) , suggesting a parasite-induced immunosuppression . In contrast , very similar transcript expression patterns were observed in response to infection by the two different species of allopatric parasites: S . mansoni ( BV ) and S . rodhaini ( BR ) and most of the variations in gene expression occurred 96h after infection . Of the 108 transcripts consistently differentially expressed in allopatric response ( Fig 1A; Fig 1B , cluster 3 ) , 98 . 1% were differentially expressed at 96 h post-infection , and 28 . 2% were more abundant following infection ( Fig 1B , cluster 3 ) . Transcripts differentially expressed exclusively in response to SmVEN or Srod were groupd in Clusters 6 ( 28 transcripts ) and 7 ( 11 transcripts ) , respectively . In response to SmVEN ( BV , Fig 1B , cluster 6 ) , 96 . 5% of the transcripts were differentially abundant 96 h after infection ( 22% over-represented ) and in response to Srod ( BR , Fig 1B , cluster 7 ) , 100% of the transcripts were differentially abundant 96 h after infection ( 82% over-represented ) . We explored the function of DE transcripts in response to the three different parasites . We initially distributed the relevant differentially expressed immune transcripts into three groups: ( i ) immune recognition molecules , ( ii ) immune effectors , and ( iii ) immune signalling molecules ( Fig 1C , S2 Table ) , that were then subdivided into functional categories ( Fig 2 ) . When we compared the percentage of each immunological group in the sympatric and allopatric interactions , no specific functional subset was particularly repressed in the BB sympatric interaction ( Fig 1C; Fig 2 ) . The same immune functions were affected in response to infections by sympatric or allopatric parasites but different immune transcripts ( grey and black diamond in Fig 2 ) showed differential regulation following infections ( Fig 2 ) . The differentially regulated transcripts belonging to the three immunological groups ( Fig 2 ) were largely involved in immune cellular responses , cell adhesion , extra cellular matrix component , cell migration , cell differentiation and cell proliferation . These functions were consistently reduced at the 24h time point in sympatric interaction ( 76% ) , whereas many transcripts involved in the same molecular processes were over-represented in allopatric interactions ( 39% ) ( Fig 2 ) . Hemocytes , the snail immune cells , participate directly in the immune response against the parasites , and immune cell activation under an immunological challenge can translate into cell proliferation and/or cell morphology modifications . Thus , cell proliferation was quantified using in vitro ( Fig 3 ) and in vivo ( Fig 4 ) EdU nuclear labelling . EdU is a nucleoside analogue of thymine incorporated into DNA during DNA synthesis . Its incorporation reflects the mitotic activity of hemocytes . In vitro labelling was used on circulating hemocytes recovered from BgBRE 24h after infection with SmBRE and SmVEN to compare the proportion of mitotic circulating hemocytes in sympatric and allopatric interaction , respectively ( Fig 3A ) . Quantification of Edu-positive hemocytes using confoncal microscopy showed that 24h after infection , hemocyte proliferation was 3 times more important following infection of BgBRE by SmVEN ( 5 . 2% of proliferative cells in BV ) than SmBRE ( 2 . 6% of proliferative cells in BB ) ( Fisher exact test two-tailed p = 7 . 6 e10-6 ) ( Fig 3B ) . Moreover , this result demonstrates for the first time that “circulating” hemocytes are able to proliferate following Schistosoma infections . Hemocyte proliferation 24h after infection was then further assessed using flow cytometry after in-vivo EdU-labelling ( Fig 4A and 4B ) . Here , we performed the same experiments using another Biomphalaria glabrata strain , BgVEN as the host and SmVEN and SmBRE as the sympatric and the allopatric parasite , respectively ( Fig 4B ) . The rate of proliferating cells was significantly higher in allopatric than sympatric interaction in both BgBRE and BgVEN ( BgBRE Mann Whitney U test: U = 36; z = -2 . 8; p = 0 . 0022; BgVEN , Mann Whitney U test: U = 36; z = -2 . 8; p = 0 . 0022 ) . In BgBRE , allopatric interaction ( BV ) was associated with 4 . 2% of proliferative cells whereas sympatric interaction resulted in 1 . 8% of proliferative cells ( Fig 4A ) . In BgVEN , allopatric interaction ( BgVEN/SmBRE , VB ) was associated with 6 . 8% of proliferative cells whereas sympatric interaction ( BgVEN/SmVEN , VV ) resulted in 2 . 0% of proliferative cells ( Fig 4B ) . At 96 h after infection , there were fewer proliferating cells: the percentage of proliferating hemocytes in sympatric BB and VV interactions were similar to the non-infected controls ( BB , 1% , Mann Whitney U test: U = 17; z = -0 . 27; p = 0 . 3936; VV , 0 . 1% , Mann Whitney U test: U = 2; z = 2 . 48; p = 0 . 013 ) , while remaining somewhat higher in both allopatric interactions ( BV , 2 . 3% , Mann Whitney U test: U = 0; z = 2 . 65; p = 0 . 009; . VB , 2 . 7% , Mann Whitney U test: U = 36; z = 2 . 8; p = 0 . 0022 ) . These results confirm that the reduced cell proliferation is associated with sympatric interaction regardless of the strain used . The morphology of hemocytes ( size and granularity ) from non-infected and infected B . glabrata snails ( BgBRE and BgVEN ) in sympatric and allopatric interactions with the parasites SmBRE and SmVEN was observed using flow cytometry ( Fig 4C and 4D ) . Morphology and heterogeneity of circulating hemocytes varied similarly in BgBRE and BgVEN snails ( Fig 4C and 4D ) . In non-infected snails , the content of circulating hemocytes was very heterogeneous , but represented a single population with continuous gradient of size and granularity typical of B . glabrata hemocytes ( Fig 4C and 4D ) [45] . Hemocyte population heterogeneity changed quickly after infection . In allopatric interactions , 24 h after infection ( Fig 4C , BV24 , and 4D , VB24 ) two populations could be distinguished: a population P1 ( corresponding to that seen in non-infected snails ) and a population P2 ( a new population ) . P2 cells exhibited increased granularity , retained a high degree of size variability , and showed a mitotic activity , as indicated by EdU labeling ( Fig 4C and 4D , red dots ) . This profile was transitory , as the P2 population had disappeared 96 h after infection ( Fig 4C , BV96 and 4D , VB96 ) . Altogether , these results show that , upon infection , the snail circulating immune cells exhibit a particular population dynamic with transient increase of the mitotic activity associated with morphology modifications . Moreover , this cellular response appears to be inhibited by sympatric parasites . To examine the potential impact of allopatric or sympatric parasites on snail survival , we investigated the mortality rates of infected snails over 4 months . The survival rate was significantly higher for non-infected snails compared to infected snails ( sympatric interaction Kaplan-Meier Log Rank test p = 1 . 39 e10-5 and allopatric interaction p = 0 . 0005 ) . However , there was no significant difference in the mortality rates of snails subjected to sympatric versus allopatric interactions: at the end of the experiment , the survival rates were 72% and 65% for the allopatric and sympatric interactions , respectively ( Kaplan-Meier Log Rank test p = 0 . 243 ) ( S2 Fig ) .
In the natural environment , it is assumed that the parasitic genes responsible for infectivity will evolve alongside the host defence genes , resulting in adaptation of the interactions between local host and parasite populations [52 , 53] . In this context , local/sympatric parasites were expected to display a greater infectiveness , reproductive success , and virulence in host populations compared to foreign/allopatric parasites [36 , 37 , 54 , 55] . However , in some cases this rule may be contradicted , as certain allopatric parasite-host interactions have been reported to be significantly more compatibles than the corresponding sympatric combinations [56 , 57] , it appears that certain Biomphalaria/Schistosoma interactions do not fulfil at the local adaptation between the host and the parasite , in which the sympatric parasite is expected to perform better than the allopatric one [36 , 37 , 54 , 55] . Using field data , Morand et al . ( 1996 ) [38] , Prugnolle et al . ( 2006 ) [5] and Mutuku et al . ( 2014 ) [39] showed that although sympatric parasite-host combinations of schistosomes and snails do tend to be more compatible , exceptions exist wherein particular allopatric combinations are equally or significantly more compatibles . Similar results were obtained when comparing the interactions of Brazilian and Guadeloupean snails versus Schistosoma infections [41] . The authors found that allopatric Guadeloupean parasites were not able to infect Brazilian snails; but Brazilian parasites were able to infect the allopatric Guadeloupean snails . Furthermore , this work demonstrated the presence of local adaptation between reactive oxygen species ( ROS ) and ROS scavengers in this system [41] . Based on these observations , we propose that it would be important to develop integrative analysis to depict and understand the precise molecular crosstalk ( immunobiological interactions ) occurring in such highly compatible sympatric and allopatric systems . Thus , dual-comparative approaches were used herein to simultaneously analyze the responses of Biomphalaria snails and Schistosoma parasites into sympatric or allopatric interactions displaying similar compatibilities . The present RNAseq analysis demonstrated that in sympatric interaction ( BB ) a huge immunosuppression occurs . Twenty-four hours after the infection , the three immunological processes: ( i ) immune recognition , ( ii ) effector and ( iii ) signaling pathways ( Figs 1 and 2 ) were down regulated . Conversely , in allopatric interactions ( BV and BR ) , host immune response was activated after 96 hours ( Figs 1 and 2 ) . Differentially regulated transcripts mostly belong to immune cellular activation , migration , proliferation , or differentiation ( Fig 2 ) . An EdU labelling was used to detect proliferation and confirmed that hemocyte proliferation is inhibited during interaction with two different strains from Brazil and Venezuela ( Figs 3 , 4A and 4B ) . In addition , we discovered that a new subpopulation of proliferating hemocytes ( named P2 ) , exclusively differentiate 24h following allopatric infections ( Fig 4C and 4D ) . P2 was EdU-positive and characterized by an increased in granularity , indicating that the new P2 cell subtype could proliferate ( Fig 4 ) . However , in absence of specific hemocyte markers , it is difficult to analyze precisely which hemocyte morphotype are proliferating ( Fig 4C and 4D ) . The P2 subpopulation would thus originates from either a morphological change in an existing subset ( correlating potentially with a decline in the P1 population ) , or represents cells that are migrating from tissues or hematopoietic organ to reach the hemolymph . Indeed , P2 population reflects newly proliferating cells that present higher EdU positive cells than the P1 population ( Fig 4C and 4D ) . Further investigations will be necessary to conclude on the origin of P2 population . In Biomphalaria snails , we know 3 main hemocyte morphotypes , the blast-like cells , the type I hyalinocytes and the granulocytes [58] . Based on the flow cytometry and Edu labelling approaches we can supposed that bigger and granular cells ( granulocytes and hyalinocytes ) are the ones that proliferates . This is demonstrated in S3 Fig in which Edu labelling was observed for hyalinocytes and granulocytes but never for blast-like cells ( S3 Fig ) . These results seem to demonstrate a differentiation or sub-functionalization in hemocyte subtypes following infection . This differentiation or sub-functionalization is different comparing sympatric and allopatric interactions , i . e . , hemocyte proliferation decreased more rapidly in sympatric rather than in allopatric interactions ( Figs 3 and 4 ) , P2 population observed solely in allopatric interactions ( Fig 4 ) . Using reciprocal sympatric and allopatric interactions , we demonstrate that the cellular or molecular phenotype observed refers to potential co-evolution or adaptation rather to a simple host or parasite strain effect ( Figs 3 and 4 ) . The strong immunosuppression observed within 24h of infection by a sympatric parasite , and the inhibition of hemocyte proliferation can certainly explain the differences in the growth of sympatric and allopatric parasites . Indeed , we observed a significant difference in sporocyst size 24h after infection ( Fig 5 ) , with sympatric sporocysts that were one-third bigger than allopatric sporocysts . But , 96h after infection , there was no more significant size difference between sympatric and allopatric parasites ( Fig 5 ) . This difference in size between the sympatric and the allopatric parasites at the beginning of the interaction can be explained by several hypotheses , ( i ) a delay in development of the allopatric parasite due to the necessity to circumvent the host immune response , ( ii ) the intrinsic ontogenesis or morphogenesis of post-miracidial intramolluscan stages that can be longer for allopatric SmVEN parasite compared to sympatric SmBRE parasite , finally ( iii ) the miracidial binding and penetration into the tissues of the host may take longer for the allopatric parasite than for the sympatric parasite . The consequences of this delay in terms of secondary sporocyst development , number of cercariae produced , or cercariae infectivity and pathogenicity for the vertebrate host , will deserve further investigation to conclude about a potential fitness cost between sympatric and allopatric parasites . To find new clues as to how sympatric parasites immunosuppress the host or circumvent the host immune system , we used a dual-RNAseq approach to compare transcripts expression of the sympatric and allopatric parasite intra-molluscal stages ( Fig 6 ) . As the histological differences were solely observed at 24h after infection , we used dual-RNAseq to investigate the parasite expression patterns at the same time point of infection . Most of the parasite transcripts belonged to the processes of nucleotide metabolism , transcription , translation and cell differentiation , development , and growth . We also identified some transcripts with GO terms or functions related to immuno-modulation or immuno-suppression ( Fig 6 and S4 Table ) . Nearly all of the identified transcripts were over-represented in the sympatric interaction compared to the allopatric ones . Our results therefore suggest that the installation , development and growth of the parasite occurred much more rapidly in the BgBRE/SmBRE combination , as sympatric parasites seemed to interfere more efficiently with the host immune system . However , RNAseq data did not give any clear information about how allopatric parasites succeed in circumventing the host immune system . We thus next examined the generated dual-RNAseq libraries in an effort to identify whether sympatric and/or allopatric schistosomes could hijack the host immune system using microRNAs . To begin testing this hypothesis , we confronted the dual-RNAseq data to the Schistosoma mansoni subset of miRBase to identify the presence of parasite microRNAs ( pmiRNAs ) in our datasets . Even if we don’t know whether pmiRNAs were present in contact with the host immune system or simply endogenic , this in-silico exploration may ask the question to a potential molecular discussion between metazoan organisms in a host-parasite system , based on nucleic acid weapons . miRNAs are known to regulate numerous biological processes , including key immune response genes [59 , 60] . Recent work has demonstrated that circulating small non-coding RNAs from parasites have hijack roles against host metabolism , notably in the interaction of schistosomes with their vertebrate hosts [61–63] . Such non-coding RNAs could act as exogenous miRNAs to interfere with or circumvent the host immune system . In the present study , 24h after infection , several differentially expressed pmiRNAs were identified . We predicted targets of such pmiRNAs in the Biomphalaria immune reference transcriptome and found that they may target 43 . 5% of the differentially regulated immune transcripts identified in the RNAseq approach ( Fig 7 ) . In contrast , far fewer correspondences were identified for the allopatric interactions ( Fig 7 ) . The higher proportion of targeted genes in the sympatric interaction may be responsible for the observed efficient immunosuppression . If confirmed , such mechanism would reveal a specific co-evolution or adaptation in the transcriptional regulation between sympatric host and parasite . However , even if more host immune genes appeared to be targeted in the sympatric combination compared to the allopatric one’s ( Fig 7 ) , both sympatric and allopatric interactions displayed the same ability to succeed to infect the host . This similarity in compatibility phenotype between sympatric and allopatric parasites could potentially results from their ability to target host immune weapons or host genes that regulate innate cellular response using miRNAs . A unique miRNA was found in all allopatric and sympatric parasites , sma-miR-190-3p . It is predicted to bind various targets including Fibrinogen Related Protein ( FREP ) and biomphalysin . The FREP family members are known as pathogen recognition receptors [64 , 65] and FREP knockdown is associated with an increase of snail compatibility toward Schistosoma infections [66 , 67] . The biomphalysins belong to beta pore forming toxins and are key humoral factors of biomphalaria snails involved in cytotoxic/cytolytic activities against Schistosoma parasites with the ability to bind miracidia and sporocyst surfaces [68 , 69] . Moreover , transcription of these molecules was mostly reduced in sympatric and allopatric interactions ( Figs 1 and 2 ) supporting the hypothesis that sma-miR-190-3p or other pmiRNA members could play an essential role in parasite compatibility . Parasites expressing such miRNAs would thus be considered as highly virulent parasites with strong infecting capabilities . By producing dedicated miRNAs , the parasites were potentially able to regulate transcriptional , post-transcriptional , translational and protein stability processes that might help them to subvert the snail’s immune defences . Even if these results are particularly interesting , a dedicated small RNAs ( <30nt ) sequencing is now mandatory to validate or not the miRNA molecular cross talk occurring between Schistosome larval stages and their snail intermediate hosts as it has been shown for the interaction with their vertebrate definitive hosts . Compatibility reflects the outcome of complex immunobiological interactions and depends on: ( i ) the ability of the snail immune system to recognize and kill the parasite; and ( ii ) the ability of the parasite to circumvent or evade the host immune response [20 , 46 , 70] . Based on the present observations , we propose that sympatric and allopatric interactions trigger totally different responses . In the sympatric interaction , the parasite is able to induce a host immunosuppression within the first day of infection enabling it to quickly infect the host and readily begins its development . In the allopatric interaction , the parasite is not able to quickly neutralize the host immune system , and as a consequence the parasite is recognized by host defense system that mounts a potent immune response . In allopatric parasite , the disruption of the activation of their developmental program during the first day of infection could results from the need to resist to the snail immune system . However , they seemed to be able to quickly protect themselves against the host immune response and develop normally in snail tissues as soon as 96h post-infection . Thereafter , in the medium- or long-term , there are no observable differences in the prevalence , intensity , or snail survival comparing sympatric and allopatric interactions ( S1 Table , S2 Fig ) . Thus , we show that despite having similar prevalence phenotypes , sympatric and allopatric snail-Schistosoma interactions displayed a very different immunobiological dialogue at the molecular level . Intriguingly , these different immunobiological interactions seem to have no repercussions upon parasite growth at longer term or to host survival . As differences at the molecular level do not correspond apparently to any ecologically meaningful changes in term of fitness , it is not straightforward to demonstrate local adaptation in such systems . However , we do not know if fitness costs could affect other biological traits in sympatric and allopatric interactions , as for example secondary sporocysts production and growth , number of cercariae shedding , or cercariae infectivity and pathogenicity towards the vertebrate host . Demonstrating local adaptation would thus appear extremely complex and would indeed deserve further investigation . It is hard to draw the line as to when local adaptation is or is not present . However , our results argue that the differences find at the molecular level may ultimately contribute to the evolution of local adaptation at an ecological level . Nevertheless , the ability for allopatric pathogens to adapt rapidly and efficiently to new hosts could have critical consequences on disease emergence and risk of schistosomiasis outbreaks . Past events of allopatric parasites reaching new areas of transmission , even in large-geographic scale dispersion , have been largely documented . The most famous example being the schistosomiasis colonization of South America since the slave trade of the 16th-19th Centuries [71 , 72] . Schistosoma originated in Asia , reached Africa 12 to 19 million years ago ( MYA ) , and gave rise to all Schistosoma species known in Africa [72] . S . mansoni diverged from S . rodhaini around 2 . 8MYA [71 , 73] , and thereafter , 400 to 500 years ago , colonized South America [71 , 72] . This colonization of South America by S . mansoni from Africa was rendered possible by the presence of the snail host: Biomphalaria glabrata . All African species of Biomphalaria are monophyletic and seem to have originated from paraphyletic South American clade [74–76] . The ancestor of B . glabrata appears to have colonized Africa 1 to 5 MYA , giving rise to all 12 species of Biomphalaria known today in Africa [77] . In South America and Caribbean Island , S . mansoni infects B . glabrata; in Africa , it infects mostly B . pfeifferi and B . alexandrina . We found that South American S . mansoni parasites are highly compatible with their sympatric South American snail hosts , whereas African S . mansoni parasites display low compatibility phenotype with South American snail hosts ( S1 Table ) . Interestingly , the South American parasites did not lose their compatibility for African snail hosts; i . e . , the prevalences are similar to African parasites when confronted to African snails ( S1 Table ) . The recent African origin of South American Schistosoma parasites ( introduction in South America 400 to 500 years ago ) may explain why they have not diverged sufficiently in South America to lose their compatibility for African snail hosts . In this case , the transfer of allopatric parasites from Africa to South American snail hosts have be successful and result in the emergence of schistosomiasis in South America . More recently another case of compatible allopatric parasite emergence have been observed when schistosomiasis have reach Europe [78 , 79] . Here , humans infected in Senegal have imported a hybrid between Schistosoma haematobium and Schistosoma bovis into Corsica . In this case urogenital schistosomiasis could be introduced and easily and rapidly spread into this novel area of south Corsica because Bulinus truncatus the vector snail of S . haematobium was endemic in the Corsica Cavu River [78 , 79] . However , this allopatric African hybrid parasite was able to adapt efficiently to the Corsican new B . truncatus host . If parasite hybridization can potentially have a putative role in increasing the colonization potential of such S . haematobium , it would be particularly interesting to analyze and depict the molecular support of such allopatric interactions to predict the potential risk of schistosomiasis outbreaks in other European areas , or other potential transmission foci . If we hope to draw conclusions regarding the existence of emerging or outbreak risks , we need to develop integrative approaches to explore fine-scale patterns of host-parasite interactions . We must consider the spatial scale at which comparisons are conducted , the patterns of disease occurrence , the population genetics , and the involvement of physiological , immunological , and molecular processes . Studying the relevant factors at the relevant timing would be of critical importance in terms of schistosomiasis control . Understanding further , how these allopatric parasites efficiently infect host snails would be mandatory to identify markers and develop new tools to predict or to quantify risks of schistosomiasis outbreaks . Now it would be particularly relevant to go back to the field to see how translatable are our results in a more dynamic field situations with genetically diverse hosts and parasites witch evolved under complex abiotic and biotic interactions , with newly encountered allopatric hosts and potentially on quite different spatial scales . For this we have a wonderful playground in Corsica .
Our laboratory holds permit # A66040 for experiments on animals from both the French Ministry of Agriculture and Fisheries , and the French Ministry of National Education , Research , and Technology . The housing , breeding and animal care of the utilized animals followed the ethical requirements of our country . The researchers also possess an official certificate for animal experimentation from both French ministries ( Decree # 87–848 , October 19 , 1987 ) . Animal experimentation followed the guidelines of the French CNRS . The different protocols used in this study had been approved by the French veterinary agency from the DRAAF Languedoc-Roussillon ( Direction Régionale de l'Alimentation , de l'Agriculture et de la Forêt ) , Montpellier , France ( authorization # 007083 ) . The two studied strains of S . mansoni ( the Brazilian ( SmBRE ) or the Venezuelan ( SmVEN ) strains ) and the strain of S . rodhaini ( Srod ) had been maintained in the laboratory using Swiss OF1 mice ( Charles River Laboratories , France ) as the definitive host . Two snail strains of Biomphalaria glabrata were used in this study: the albino Brazilian strain , ( BgBRE ) and the Venezuelan strain , ( BgVEN ) . All host and parasite strains of each different geographical origin were recovered in their native locality and parasite strains were maintain in the laboratory always on their sympatric snail hosts to maintain the same selective pressure and sympatric adaptation on parasite . We housed snails in tanks filled with pond water at 25°C with a 12:12 hour light:dark cycle and supplied ad libitum with fresh lettuce . The Brazilian strain originates from the locality of Recife ( east Brazil , recovered in the field in 1975 ) , the Venezuelan strains of snail and parasite were recovered from the locality of Guacara ( north Venezuela , recovered in the field in 1975 ) and the African species Schistosoma rodhaini originates from Burundi and was obtained from the British Museum National History ( recovered in 1984 ) . These Schistosoma isolates/species have been selected because they exhibited similar infectivity toward BgBRE or BgVEN strains ( see prevalence and intensity in S1 Table ) . These high compatibilities were followed-up by the cercariae emissions . For all these interactions we observed comparable cercariae shedding ( S1 Table ) . Prevalence of SmBRE and SmVEN for the African vector snail Biomphalaria pfeifferi from Senegal ( BpSEN ) , and prevalence of the corresponding parasite SmSEN on South American snails were also tested ( S1 Table ) . In order to investigate the molecular response of snails against sympatric and allopatric parasites , a global comparative transcriptomic approach was conducted . One hundred and twenty BgBRE snails were infected with SmBRE , SmVEN or Srod . Each snail was individually exposed for 12 h to 10 miracidia in 5mL of pond water . For each experimental infection , 30 snails were recovered at 24h and 96h after infection . Pools of 30 snails were composed of 10 juvenile snails ( shell diameter from 3 to 5 mm ) , 10 mature adult snails ( shell diameter from 7 to 9 mm ) and 10 old adult snails ( shell diameter from 11 to 13 mm ) . The samples were named as follows: BB24 , BB96 for BgBRE infected with SmBRE; BV24 , BV96 for BgBRE infected with SmVEN; and BR24 , BR96 for BgBRE infected with Srod . We realised 2 pools of 30 uninfected BgBRE snails ( pool of immature , mature and old snails ) named Bre1 and Bre2 , that were used as control conditions for all downstream comparative analyses . A dual RNA-seq approach was conducted to gain in a broader understanding of sympatric and allopatric host/parasite interactions . Hemocytes appeared as the main cells supporting Biomphalaria snail immune response . Thus , to go further in the description of snail response against parasites , quantitative and qualitative changes in hemocyte populations were investigated . For this purpose , BgBRE and BgVEN snails were used . Snails were infected as described above , using either SmBRE or SmVEN parasites . For each experimental infection , snails were recovered at 24 and 96 h after infection , and designated as follows: BB24 and BB96 for BgBRE infected with SmBRE; BV24 and BV96 for BgBRE infected with SmVEN; VV24 and VV96 for BgVEN infected with SmVEN; and VB24 and VB96 for BgVEN infected with SmBRE . Snails of each strain , BgBRE and BgVEN , were recovered and used as controls . A histological approach was conducted in order to investigate differences in miracidia to sporocyst development , while comparing sympatric and allopatric parasite growth , development and maturation into snail tissues . BgBRE snails were infected as described above with either 10 mi of SmBRE ( sympatric ) ( n = 6 snails ) or 10 mi of SmVEN ( allopatric ) parasite ( n = 6 snails ) . At each time point , 24 and 96 h after infection , three snails were fixed in Halmi’s fixative ( 4 . 5% mercuric chloride , 0 . 5% sodium chloride , 2% trichloroacetic acid , 20% formol , 4% acetic acid and 10% picric acid-saturated aqueous solution ) . Embedding in paraffin and transverse histological sections ( 3-μm ) were performed using the RHEM platform ( Montpellier , France ) facilities . The slides were stained using Heidenhain’s azan trichromatic staining solution as follows: ( i ) serial re-hydration was performed in toluene followed by 95% , 70% , and 30% ethanol and then distilled water; ( ii ) coloration was performed using azocarmine G ( 70% ethanol , 1% aniline , 1% acetic alcohol , distilled water , 5% phosphotungstic acid , distilled water , Heidenhain’s azan ) and ( iii ) serial dehydration was performed using 95% ethanol , absolute ethanol , and toluene . The preparations were then mounted with Entellan ( Sigma Life Science , St . Louis Missouri , USA ) and subjected to microscopic examination . When a parasite is observed in snail tissue , the parasite size was measured using the imaging analysis software ImageJ ( v2 . 0 . 0 ) for each adjacent histological section in which the parasite is observed . The contour of the parasite is detailed very precisely using ImageJ and the pixel number is reported on a size scale analyzed in the same manner to calculate parasite size . Size is expressed as parasite surface in μm2 as the mean of the 3 bigger parasite sections recorded . At 24h , n = 9 sympatric and n = 8 allopatric parasites were measured and at 96h , n = 7 sympatric and n = 10 allopatric parasites were measured . The size differences between sympatric and allopatric parasite groups were tested using the Mann-Whitney U-test with statistical significance accepted at a p-value < 0 . 05 . Parasites may communicate or interfere with their host using different strategies based mainly on excreted/secreted products released into hemolymph . In this context , miRNAs appeared as the most relevant mean of communication that can be used by parasites . To test for such hypothesis S . mansoni miRNAs were analyzed in-silico by comparing the relevant miRNA database ( miRBase ) to our RNAseq libraries generated at the 24h following sympatric or allopatric infections . S . mansoni precursor sequences were downloaded from miRBase ( http://www . mirbase . org , 03/09/2017 ) , and high-quality reads from naive ( BgBRE ) and 24 h post-infection samples ( BB24 , BV24 , BR24 ) were aligned against a S . mansoni sub-database of miRBase , as previously described [81] . The identified precursors were confirmed by alignment of high-scoring reads onto precursor and mature miRNAs from miRBase . Solely reads with 100% identity were retained for analysis . The localization of each read against miRNA sequence allowed us to identify either the precursor or just the mature miRNA . Precursors found under both naive and infected conditions were excluded to retain exclusively the miRNAs present in samples from infected snails and avoid cross-species contamination because of the potential conserved features of miRNAs from B . glabrata and S . mansoni . Putative miRNA targets were predicted from among the differentially represented immune-related transcripts ( Fig 1 ) using Miranda tools ( using parameters: Miranda input_miRinput_Transcriptome -out results . txt -quiet -sc 140 -en -15 ) [82] . Because mature miRNAs may exist in two forms depending on which strand ( 5'-3' ) of the precursor stem-loop is maturated the predicted interactions could involve the 5' and/or 3' forms , as noted . The results were extracted using the awk tool , listed in S4 Table , and used to generate a Venn diagram . To confirm the ability of a selected pre-miRNA to produce the stem-loop necessary to produce the mature form , the secondary structures of precursor were predicted using RNA structure Web tool ( http://rna . urmc . rochester . edu/RNAstructureWeb , 03/09/2017 ) using default parameters . Allopatric or sympatric parasites could have different levels of virulence or impacts on their host that could impair snail survival . To test for such discrepancy we investigated the mortality rates of infected snails over the course of sympatric or allopatric infections . One hundred and sixty BgBRE snails were infected as described above with SmBRE or SmVEN strains ( n = 50 ) , and 60 non-infected BgBRE snails were retained as controls . The numbers of dead snails were compiled weekly for 14 weeks . A Kaplan-Meier estimator was used to estimate the survival function from lifetime data . Survival curves were generated using the xlstats Mac software and the log-rank test was applied with significance accepted at p<0 . 05 . | Schistosomiasis , the second most widespread human parasitic disease after malaria , is caused by helminth parasites of the genus Schistosoma . More than 200 million people in 74 countries suffer from the pathological , and societal consequences of this disease . To complete its life cycle , the parasite requires an intermediate host , a freshwater snail of the genus Biomphalaria for its transmission . Given the limited options for treating Schistosoma mansoni infections in humans , much research has focused on developing methods to control transmission by its intermediate snail host . Biomphalaria glabrata . Comparative studies have shown that infection of the snail triggers complex cellular and humoral immune responses resulting in significant variations in parasite infectivity and snail susceptibility , known as the so-called polymorphism of compatibility . However , studies have mostly focused on characterizing the immunobiological mechanisms in sympatric interactions . Herein we used a combination of molecular and phenotypic approaches to compare the effect of infection in various sympatric and allopatric evolutionary contexts , allowing us to better understand the mechanisms of host-parasite local adaptation . Learning more about the immunobiological interactions between B . glabrata and S . mansoni could have important socioeconomic and public health impacts by changing the way we attempt to eradicate parasitic diseases and prevent or control schistosomiasis in the field . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"biomphalaria",
"helminths",
"natural",
"antisense",
"transcripts",
"gene",
"regulation",
... | 2019 | Sympatric versus allopatric evolutionary contexts shape differential immune response in Biomphalaria / Schistosoma interaction |
Metabolic rate , heart rate , lifespan , and many other physiological properties vary with body mass in systematic and interrelated ways . Present empirical data suggest that these scaling relationships take the form of power laws with exponents that are simple multiples of one quarter . A compelling explanation of this observation was put forward a decade ago by West , Brown , and Enquist ( WBE ) . Their framework elucidates the link between metabolic rate and body mass by focusing on the dynamics and structure of resource distribution networks—the cardiovascular system in the case of mammals . Within this framework the WBE model is based on eight assumptions from which it derives the well-known observed scaling exponent of 3/4 . In this paper we clarify that this result only holds in the limit of infinite network size ( body mass ) and that the actual exponent predicted by the model depends on the sizes of the organisms being studied . Failure to clarify and to explore the nature of this approximation has led to debates about the WBE model that were at cross purposes . We compute analytical expressions for the finite-size corrections to the 3/4 exponent , resulting in a spectrum of scaling exponents as a function of absolute network size . When accounting for these corrections over a size range spanning the eight orders of magnitude observed in mammals , the WBE model predicts a scaling exponent of 0 . 81 , seemingly at odds with data . We then proceed to study the sensitivity of the scaling exponent with respect to variations in several assumptions that underlie the WBE model , always in the context of finite-size corrections . Here too , the trends we derive from the model seem at odds with trends detectable in empirical data . Our work illustrates the utility of the WBE framework in reasoning about allometric scaling , while at the same time suggesting that the current canonical model may need amendments to bring its predictions fully in line with available datasets .
Whole-organism metabolic rate , B , scales with body mass , M , across species as [1] ( 1 ) where B0 is a normalization constant and α is the allometric scaling exponent , typically measured to be very close to 3/4 [2] . The empirical regularity expressed in Equation 1 with α = 3/4 is known as Kleiber's Law [3] , [4] . Many other biological rates and times scale with simple multiples of 1/4 . For example , cellular or mass-specific metabolic rates , heart and respiratory rates , and ontogenetic growth rates scale as M−1/4 , whereas blood circulation time , development time , and lifespan scale close to M1/4 [5]–[9] . Quarter-power scaling is also observed in ecology ( e . g . , population growth rates ) and evolution ( e . g . , mutation rates ) [2] , [10] , [11] . The occurrence of quarter-power scaling at such diverse levels of biological organization suggests that all these rates are closely linked . Metabolic rate seems to be the most fundamental because it is the rate at which energy and materials are taken up from the environment , transformed in biochemical reactions , and allocated to maintenance , growth , and reproduction . In a series of papers starting in 1997 , West , Brown , and Enquist ( WBE ) published a model to account for the 3/4-power scaling of metabolic rate with body mass across species [1] , [12]–[14] . The broad theory of biological allometry developed by WBE and collaborators attributes such quarter-power scaling to near-optimal fractal-like designs of resource distribution networks and exchange surfaces . There is some evidence that such designs are realized at molecular , organelle , cellular , and organismal levels for a wide variety of plants and animals [2] , [14] . Intensifying controversy has surrounded the WBE model since its original publication , even extending to a debate about the quality and analysis of the data [15]–[28] . One of the most frequently raised objections is that the WBE model cannot predict scaling exponents for metabolic rate that deviate from 3/4 [16] , [29] , even though the potential for such deviations was appreciated by WBE themselves [1] . If this criticism were true , WBE could not in principle explain data for taxa whose scaling exponents have been reported to be above or below 3/4 [29]–[34] , or deviations from 3/4 that have been observed for small mammals [35] . Likewise , the WBE model would be unable to account for the scaling of maximal metabolic rate with body mass , which appears to have an exponent of 0 . 88 [36] . It is important to note that the actual nature of maximal metabolic rate scaling is , however , not without its own controversy; see [37] for an argument that maximal metabolic rate scales closer to 3/4 when body temperature is taken into consideration . Much of the work aimed at answering these criticisms has relied on alteration of the WBE model itself . Enquist and collaborators account for different scaling exponents among taxonomic groups by emphasizing differences in the normalization constant B0 of Equation 1 and deviations from the WBE assumptions regarding network geometry [26] , [38]–[40] . While these results are suggestive , it remains unclear whether or not WBE can predict exponents significantly different from 3/4 and measurable deviations from a pure power law even in the absence of any variation in B0 and with networks following exactly the geometry required by the theory . Although WBE has been frequently tested and applied [40]–[53] , it is remarkable that no theoretical work has been published that provides more detailed predictions from the original theory . Also , work aimed at extending WBE by relaxing or modifying some of its assumptions has hardly been complete; many variations in network structure might have important and far-reaching consequences once properly analyzed . This is what we set out to do in the present contribution . We show that a misunderstanding of the original model has led to the claim that WBE can only predict a 3/4 exponent . This is because many of the predictions and tests of the original model are derived from leading-order approximations . In this paper we derive more precise predictions and tests . For the purpose of stating our conclusions succinctly , we refer to the “WBE framework” as an approach to explaining allometric scaling phenomena in terms of resource distribution networks ( such as the vascular system ) and to the “WBE model” as an instance of the WBE framework that employs particular parameters specifying geometry and ( hydro ) dynamics of these networks [1] , [14] . ( We shall detail these assumptions and define terminology more accurately in section “Assumptions of the WBE model” . ) Our main findings are: 1 . The 3/4 exponent only holds exactly in the limit of organisms of infinite size . 2 . For finite-sized organisms we show that the WBE model does not predict a pure power-law but rather a curvilinear relationship between the logarithm of metabolic rate and the logarithm of body mass . 3 . Although WBE recognized that finite size effects would produce deviations from pure 3/4 power scaling for small mammals and that the infinite size limit constitutes an idealization [1] , the magnitude and importance of finite-size effects were unclear . We show that , when emulating current practice by calculating the scaling exponent of a straight line regressed on this curvilinear relationship over the entire range of body masses , the exponent predicted by the WBE model can differ significantly from 3/4 without any modifications to its assumptions or framework . 4 . When realistic parameter values are employed to construct the network , we find that the exponent resulting from finite-size corrections comes in at 0 . 81 , significantly higher than the 3/4 figure based on current data analysis . 5 . Our data analysis indeed detects a curvilinearity in the relationship between the logarithm of metabolic rate and the logarithm of body mass . However , that curvilinearity is opposite to what we observe in the WBE model . This implies that the WBE model needs amendment and/or the data analysis needs reassessment . Beyond finite-size corrections we examine the original assumptions of WBE in two ways . First , we vary the predicted switch-over point above which the vascular network architecture preserves the total cross-sectional area of vessels at branchings and below which it increases the total cross-sectional area at branchings . These two regimes translate into different ratios of daughter to parent radii at vessel branch points . Second , we allow network branching ratios ( i . e . , the number of daughter vessels branching off a parent vessel ) to differ for large and small vessels . We analyze the sensitivity of the scaling exponent with respect to each of these changes in the context of networks of finite size . This approach is similar in spirit to Price et al . [40] , who relaxed network geometry and other assumptions of WBE in the context of plants . In the supplementary online material Text S1 , we also argue that data analysis should account for the log-normal distribution of body mass abundance , thus correcting for the fact that there are more small mammals than large ones . Despite differences in the structure and hydrodynamics of the vascular systems of plants and animals [1] , [13] , detailed models of each yield a scaling exponent of 3/4 to leading-order . In the present paper , we focus on the WBE model of the cardiovascular system of mammals . All of our assumptions , derivations , and calculations should be interpreted within that context . Finite-size corrections and departures from the basic WBE assumptions are important in the context of plants as well , as shown in recent studies by Enquist and collaborators [26] , [38]–[40] . In final analysis , we are led to the seemingly incongruent conclusions that ( 1 ) many of the critiques of the WBE framework are misguided and ( 2 ) the exact ( i . e . , finite-size corrected ) predictions of the WBE model are not fully supported by empirical data . The former means that the WBE framework remains , once properly understood , a powerful perspective for elucidating allometric scaling principles . The latter means that the WBE model must become more respectful of biological detail whereupon it may yield predictions that more closely match empirical data . Our work explores how such details can be added to the model and what effects they can have . The paper is organized as follows . For the sake of a self-contained presentation , we start with a systematic overview of the assumptions , both explicit and implicit , underlying the WBE theory ( section “Assumptions of the WBE model” ) . In Text S1 , we provide a detailed exposition of the hydrodynamic derivations that the model rests upon . These calculations are not original , but they have not appeared to a full extent before in the literature . While nothing in section “Assumptions of the WBE model” is novel , there seems to be no single “go to” place in the WBE literature that lays out all components of the WBE theory . Our paper then proceeds with a brief derivation of the exact , rather than approximate , relationship between metabolic rate and body mass ( section “Derivation of the 3/4 scaling exponent” ) . We then calculate the exact predictions for scaling exponents for networks of finite size ( section “Finite-size corrections to 3/4 allometric scaling” ) and revisit certain assumptions of the theory ( section “Making the WBE model more biologically realistic” ) . In section “Comparison to empirical data” we compare our results to trends detectable in empirical data . We put forward our conclusions in the Discussion section .
The WBE model rests on eight assumptions . Some of these assumptions posit the homogeneity of certain parameters throughout the resource distribution network . Any actual instance of such a network in a particular organism will presumably exhibit some heterogeneity in these parameters . The object of the theory is a network whose parameters are considered to be averages over the variation that might occur in any given biological instance . For the sake of brevity , we refer to such a network as an “averaged network” . The impact of parameter heterogenity on the scaling exponent is very difficult to determine analytically . ( Section “Changing branching ratio across levels” addresses a modest version of this issue numerically . ) Using the above assumptions , we can derive how metabolic rate , B , varies with body mass , M , which is the fundamental result of WBE . The key insight is that body mass is proportional to blood volume ( following from Assumption 6 ) and that blood volume is the sum of the volumes of the vessels over all the levels of the network . Using Assumptions 2–6 , this sum can be expressed in terms of properties of capillaries , providing a direct link to metabolic rate ( owing to Assumption 8 ) . Upon expressing blood volume in terms of capillary properties , we can separate terms of the sum that are invariant ( by Assumption 7 ) from others that vary with the total number of capillaries . The total number of capillaries is directly proportional to the whole-oganism metabolic rate , because each capillary supplies resources at the same rate regardless of organism size ( Assumption 7 ) . This ties body mass to metabolic rate . We now provide the formal derivation . Using Assumptions 2–4 the total blood volume or total network volume ( assuming the network is completely filled with blood and ignoring the factor of 2 that may arise from blood in the venous system , which returns blood to the heart ) can be expressed as the sum ( 5 ) where the volume of a vessel is that of a cylinder . Next , we use the scale-free ratios γ , β< , and β> , defined by Equations 2–4 resulting from Assumptions 5 and 6 , to connect level k to successively higher levels and all the way to the capillary level N: ( 6 ) where is the volume of a capillary and Ncap = NN = nN the number of capillaries . The first sum ranges over the area-preserving regime and the second sum is over the area-increasing regime . The first sum is a standard geometric series . Observing that then yields ( 7 ) where N̅ is the fixed number of levels from the capillaries to the level where the transition from area-increasing to area-preserving branching occurs . Since , we have , and the second sum in Equation 6 is simply ( 8 ) Combining these results we have ( 9 ) This equation can be re-expressed as ( 10 ) where ( 11 ) are both constant with respect to body mass . Equation 10 will play a fundamental role in the following sections . Given this simple relation between total blood volume ( or network volume ) and the number of capillaries , it is straightforward to relate metabolic rate , B , to body mass , M . Using Assumption 8 , the whole-body metabolic rate is just the sum total of the metabolic rates enabled by the resources delivered through each capillary . Let the contribution to total metabolic rate enabled by a capillary be Bcap . By Assumption 7 Bcap is constant across organisms . Thus , B = NcapBcap , or simply B∝Ncap . Inserting this into Equation 10 , invoking Assumption 7 that Vcap is independent of body mass , and using Assumption 6 to recognize that M∝Vblood yields ( 12 ) with C2 a constant , , , and are new constants . Letting the number of levels in the cardiovascular system , N , tend to infinity—which necessarily means that body mass , M , and metabolic rate , B∝Ncap = nN , become infinitely large—we conclude that ( 13 ) This is the celebrated result that has been empirically observed for nearly a century . Equation 13 is approximately true as long as ( 14 ) It is essential to realize that the prediction of a 3/4 scaling relationship only holds in the infinite M-limit . The approximation becomes less accurate as organisms become smaller , corresponding to smaller metabolic rate B . The exact relationship is Equation 12 , or 10 , which does not predict a pure power law but a curvilinear graph of ln B versus ln M . Forcing such a curve to fit a straight line will therefore not produce an exact value of 3/4 , except when the magnitude of the correction term is small compared with 1 ( see Equation 14 ) , the measurement error , or the residual variation in the empirical data . Given that the importance of these deviations will be larger for smaller organisms , it would in principle be interesting to look more carefully at finite-size effects for small fish or plants , because the smallest mammals are considerably larger than the smallest fish or plants [25] , [26] , [29] , [33] , although we do not perform such an analysis here . Different taxa often span different ranges of body size and exhibit a particular relative proportion of small to large organisms . These characteristics will likely lead to different measured scaling exponents . We conclude that the WBE model actually predicts variation in scaling exponents due to finite-size terms whose magnitude depends on the absolute range of body masses for a given taxonomic group . These predictions can be tested against the allometric exponents reported in the empirical literature .
To quantify finite-size corrections , we focus on Equation 10 because the blood volume , Vblood ( ∝M ) , and the number of capillaries , Ncap ( ∝B ) , are really the fundamental parameters of the theory . Proceeding in this way , we avoid the additional constants C2 and Bcap . By inspecting Equation 10 , we see that finite-size effects can become manifest in two different ways . First , even in the absence of network regions with area-increasing branching ( N̅ = 0 ) , there are corrections to the 3/4 scaling exponent . Second , the switch-over point N̅ from area-preserving branching to area-increasing branching determines the relative contributions of these two network regimes , and has the potential to considerably influence the scaling exponent . To quantify these effects , we consider three cases: ( i ) a network with only area-preserving branching ( section “Networks with only area-preserving branching” ) , ( ii ) a network with only area-increasing branching ( section “Networks with only area-increasing branching” ) , and ( iii ) a mixture of the two with a transition level ( section “Networks with a transition from area-preserving to area-increasing branching” ) . The discrepancies between the WBE model and data might be addressed in several ways: ( i ) by correcting for biases in the empirical distributions of species masses; ( ii ) by adding more detail to any of the WBE assumptions; ( iii ) by relaxing the assumptions . In Text S1 ( Figure S4 ) , we exemplify case ( i ) by accounting for the fact that most mammals , in particular those that have been measured , are of small mass . The body-size distribution across species is approximately log-normal . By sampling body sizes according to such a distribution and using the same numerical methods as in section “Finite-size corrections to 3/4 allometric scaling” above , we determined that the overall effect on the scaling exponent is essentially negligible ( the exponent is slightly lowered ) . In section “Modifying the transition level between area-preserving and area-increasing regimes” , we illustrate approach ( ii ) by altering the level at which the transition from area-preserving to area-increasing branching occurs , as well as the width of the region over which it extends , as motivated by complexities in the hydrodynamics of blood flow . These considerations affect the scaling exponent , but the change is too small to restore the 3/4 figure . In section “Changing branching ratio across levels” , we illustrate approach ( iii ) by relaxing the assumption of a constant branching ratio ( Assumption 4 ) . We show that systematic changes in the branching ratio can significantly lower the measured scaling exponent and lead to intriguing non-linear effects that depend on where the transition from one branching ratio to another occurs . Savage et al . [23] published an extensive compilation of empirical data for basal metabolic rate and body mass of 626 mammals . In this section we compare the dependency of scaling exponents on body mass as obtained from this dataset to our predictions for scaling exponents with finite-size corrections . We sorted organisms according to body mass and grouped them , starting with the smallest exemplar , into disjoint bins spanning one order of magnitude each . We then analyzed this data compilation in three ways . First , we determined the scaling exponents for successive cumulations of bins . At each addition of a bin , we computed a linear regression on the entire cumulated data , plotting the resultant scaling exponent against the range of sizes . In other words , the first scaling exponent is determined for the first order of magnitude in body mass , the second exponent is determined for the first two orders of magnitude , and so on . This is similar in spirit to the procedure used for analyzing and presenting the numerical data in section “Finite-size corrections to 3/4 allometric scaling” . The result is shown in Figure 10A . In a second approach we proceeded similarly , but starting with the largest order of magnitude in body mass , then successively adding bins of smaller orders ( Figure 10B ) . Lastly , we computed the scaling exponent for each bin separately ( Figure 10C ) . The panels of Figure 10 show the results with error bars based on the 95% confidence intervals obtained from ordinary least squares ( OLS ) . In panels 10A and 10B , the exponents exhibit an increasing trend with body mass . Panel 10C shows a similar trend for bins that correspond to intermediate mass ranges . These are the bins that contain most of the data points . There is too much scatter at either end of the body mass distribution to make a statement about the entire range for panel 10C . We find that for those ranges and aggregations with smallest scatter ( as determined from error bars ) , the scaling exponent approaches the 3/4 figure from below . Although these data are suggestive , it would be incautious at this point to assert that the data flatten out at 3/4 for some maximum mammalian size . Given the current dataset , however , an “asymptotic” 3/4 scaling seems a reasonable guide . The concave increase of the scaling exponent with body mass is most consistent with a finite-size WBE model based on pure area-preserving branching throughout the network , see section “Networks with only area-preserving branching” . ( The concave increase of the scaling exponent , Figure 3B , corresponds to a convex relationship between metabolic rate and body mass , see the schematic in Figure 2 . ) Recall that in our numerical studies of section “Networks with only area-preserving branching” the scaling exponent approached 3/4 in a concave fashion from below , while networks built entirely with area-increasing branching ( section “Networks with only area-increasing branching” ) have scaling exponents that always lie above 3/4 , converging to an accumulation point at 1 . Networks built with a mixture of these branchings ( section “Networks with a transition from area-preserving to area-increasing branching” ) , approach 3/4 scaling in a convex fashion from above , opposite to the trends seen in Figure 10 . ( The convex decrease of the scaling exponent , Figure 6A , corresponds to a concave relationship between metabolic rate and body mass , see the schematic in Figure 5 . ) A similar analysis of a more limited dataset for heart rate ( 26 data points ) and respiratory rate ( 22 data points ) [23] also shows a trend that is not easily reconciled with our finite-size corrections for networks with a mixture of area-preserving and area-increasing branching . In WBE , heart rate ω and respiratory rate R are both predicted to scale as ω∝R∝M−α/3 ( see Table S1 and related text in section “Impact of finite-size corrections on additional WBE predictions” of Text S1 ) . Since our calculations in section “Networks with a transition from area-preserving to area-increasing branching” yield scaling exponents , α , that approach 3/4 from above as body mass increases , we expect the scaling of heart and respiratory rates to both have exponents that are bounded by the maximum value of −1/4 . The WBE model with finite-size corrections predicts α≈0 . 81 . Hence , heart and respiratory rates should scale as M−0 . 27 and asymptote to −1/4 with increasing mass . That is , there should be very little change in the scaling exponent when analyzing data for either small or large mammals . This does not match empirical heart rate data well . Regressing on the first three , four , and six orders of magnitude in body mass yields exponents of −0 . 33 , −0 . 27 , and −0 . 25 , respectively . The match is worse for respiratory rate data . Regressing on the first two , three , five , and seven orders of magnitude in body mass gives exponents of −0 . 64 , −0 . 44 , −0 . 34 , and −0 . 26 , respectively . We observe a convergence to −1/4 , but over a much larger range of scaling exponents than expected . While the WBE model has been predominantly interpreted in the context of interspecific scaling [9] , [23] , metabolic rate also varies with body mass during development . Such intraspecific data [29] , [69] sometimes exhibits a concave curvature across growth stages ranging from young to adult mammals . Indeed , our finite-size corrections for the canonical WBE model predict a concave curvature of ln B versus ln M . However , they also predict an asymptotic approach to a slope of 3/4 for large mammals , and thus a fitted OLS slope for the entire body mass range that is greater than 3/4 , as schematically shown in Figure 5 . In his Table 5 , Glazier [29] reports slopes from 29 intraspecific regressions for 14 species of mammals . From these , we compute an average slope α = 0 . 70; in this dataset , 20 of the slopes are smaller than 3/4 and only 9 of the slopes are larger than 3/4 . This is inconsistent with our predictions . Moreover , the average body mass range of mammals , for which Glazier reports intraspecific regressions , spans only half an order of magnitude . Yet , our calculations show that several orders of magnitude in body mass are required to detect curvilinearity from finite-size effects , as seen in Figure 6A . We thus conclude that the curvature revealed by these intraspecific datasets is either unrelated to finite-size effects or fails to support the finite-size corrected canonical WBE model . It is important to note that empirical data for the inter- and intraspecific case ( especially for restricted size classes ) are rather limited . We therefore do not wish to overstate the strength of our conclusions . We merely report discrepancies between the predictions of the canonical WBE model and limited sets of data . We anticipate that further data acquisition , statistical analysis , and model refinement will bring theory and data into agreement .
Over the past decade , the WBE model has initiated a paradigm shift in allometric scaling that has led to new applications ( e . g . , [2] , [70] , [71] ) , new measurements and the refinement of data analysis ( e . g . , [41]–[53] ) , and the recognition of connections between several variables that describe organismic physiology [1] , [23] . However , WBE has also drawn intense criticism and sparked a heated debate [15]–[28] . In section “Assumptions of the WBE model” , we provide a detailed presentation of the complete set of assumptions and calculations defining the WBE model . While none of these originated with us , the literature lacked , surprisingly , an exhaustive exposition . ( In particular , the consequences of Assumption 6 are a distillation of hydrodynamical calculations that we summarize in Text S1 . ) In section “Derivation of the 3/4 scaling exponent” , we connect each step in the derivation of the main WBE result to the assumptions it invokes . In this way , we provide a self-contained platform for motivating , deriving , and interpreting our results . One of our main objectives is to clarify that the WBE model predicts ( and thus “explains” ) the 3/4 exponent of the scaling law relating whole-organism metabolic rate to body mass only as the limit of infinite network size , body mass , and metabolic rate is approached . Although this fact was appreciated by WBE in their original work [1] the nature of this approximation has been broadly misunderstood in the subsequent literature , e . g . , [16] , [29] . In this work , we conduct a systematic exploration of finite-size effects in the WBE framework and find that these effects yield measurable deviations from the canonical 3/4 scaling exponent , shifting the actual prediction to a value closer to 0 . 81 when published parameters are employed [1] , [14] . This finding has major implications and immediately clarifies some contentious issues . On the one hand , the common criticism that the WBE model can only predict a scaling exponent of 3/4 is incorrect . As we show in section “Finite-size corrections to 3/4 allometric scaling” , a continuum of exponents can be obtained as a function of body-mass . On the other hand , the 0 . 81 figure ( obtained for N̅ = 24 and n = 2 ) shifts the predicted exponent for mammals away from the canonical figure of 3/4 that reflects current data analysis . In section “Impact of finite-size corrections on additional WBE predictions” of Text S1 we report the finite-size corrections for several variables related to vascular physiology that were documented in the original WBE paper [1] . A major consequence of the curvilinear relationship between ln B and ln M predicted by the model is the fact that the scaling exponent , as measured by a simple power law regression , will show a dependence on the absolute masses of the organisms in question . Notably , our numerical calculations for area-increasing branching in Figure 4 are consistent with the linear scaling of metabolic rate versus body mass that has been observed for small fish [32] . Indeed , with minor modifications , our Equations 19 and 20 could be used to test the form of isometric scaling observed in young and small fish . It should be noted , however , that the magnitude of these finite-size corrections depends strongly on certain network properties , such as N̅ . Furthermore , we find evidence for size-dependent relationships in the available empirical data for mammals ( section “Comparison to empirical data” ) . Specifically , we find that the measured scaling exponent tends to increase with body mass , indicating that the empirical data ( of log metabolic rate versus log body mass , or , equivalently , ln Ncap versus ln Vblood ) exhibits convex curvature ( i . e . , the type of relationship dramatized in Figure 2 ) . However , networks constructed with a mixture of area-increasing and area-preserving branching can never produce scaling relationships with exponents less than 3/4 and , although 3/4 scaling is approached in the limit of networks of infinite size , the exponents always approach 3/4 from above ( unlike in Figure 10 ) . Mixed networks of this type display inherently concave curvature of the log metabolic rate versus log body mass relationship ( i . e . , the type of relationship dramatized in Figure 5 ) . That is , a group of organisms of larger sizes will yield smaller fitted exponents than a group of organisms of smaller sizes . Yet , empirical data are best fit by a power law with an exponent less than 3/4 and demonstrate convex curvature in several datasets of log metabolic rate versus log body mass . Thus , assuming that this represents the actual curvature in nature , either ( i ) a transition between radial scaling regimes does not occur , potentially contradicting Assumption 6 of the WBE model , or ( ii ) at least one assumption of the WBE model must be modified . The case for pure area-increasing branching ( hypothesis ( i ) above ) within the WBE model is somewhat problematic . The only way for such a network to be consistent with Assumption 6 would be to posit that the transition from area-preserving to area-increasing regimes occurs at a vessel radius smaller than a capillary; in this case , this transition would in principle exist but would simply never actually be observed in nature . A number of facts contradict this explanation . For one , estimates place the transition at vessel radii of about 1 mm . Despite the fact that predictions of where the transition might occur are problematic ( see section “Modifying the transition level between area-preserving and area-increasing regimes” ) , the estimate is unlikely to be 3 orders of magnitude larger than the actual value ( since capillary radii are on the order of 1 µm in radius ) . A further complication is that a pure area-preserving network would theoretically not be able to “slow down” blood flow due to the conservation of volume flow rate for an incompressible fluid . The fact that blood flows much more quickly in the aorta than it does in the capillaries would tend to argue that area-increasing branching must occur somewhere in the network . Finally , there is the simple fact of Murray's Law; empirical findings squarely place β for small vessels in the neighborhood of n−1/3 , strongly implying that area-increasing branching is in fact dominant when vessel radii are small [54] , [57] . In our hands , empirical data seem most consistent with networks built with purely area-preserving branching , although the lack of very high-quality data for both metabolic rate and body mass makes it difficult to be absolutely certain of this trend . The reasoning outlined above makes hypothesis ( i ) appear somewhat unlikely . This leaves us with a riddle: cardiovascular networks with architectures that support the scaling trends observed for real organisms would seem to violate Assumption 6 of the WBE model . We are thus led to believe that some modification of assumptions 2–8 is necessary to explain the concavity in the data and an empirical scaling exponent less than 3/4 . While a model that aligns with the empirical evidence might differ from the canonical WBE model ( assumptions 2–8 plus specific values for the parameters N̅ and n ) , we believe such a model will squarely remain within the WBE framework ( assumption 1 , that is , the exploration of allometric scaling in the context of resource distribution networks ) . Resolving this paradox will likely require intensive further data analysis and extension of the canonical WBE model . It is clear that work in this area would benefit from a more detailed empirical understanding of cardiovascular networks themselves . Although data for the coronary artery in humans , rats , and pigs exist [51]–[54] , [72] , [73] , along with measurements for the vascular system in the lungs of armadillos [74] , stringent tests of the core WBE assumptions require measurements throughout the body , in a larger variety of species , and for vessels farther away from the heart . Measurements are needed especially for the number of levels from the heart to the capillaries for different species , the scaling ratios of vessel radii ( β = rk+1/rk ) and vessel lengths ( γ = lk+1/lk ) , vessel blood flow rates , and branching ratios ( n = Nk+1/Nk ) . Such data will help to assess the extent to which mammalian vascular systems are space filling ( Assumption 5 ) , the scope of area-preserving and area-increasing branching ( Assumption 6 ) , the value ( s ) of the branching ratio throughout the network ( Assumption 4 ) , and the degree of symmetry or asymmetry in branchings and scaling ratios ( Assumption 3 ) . Analyzing intraspecific variation in network geometry may also enable a quantification of selection pressures for optimality with respect to energy loss , as implied by Assumption 6 ( see Figures S1 and S2 in Text S1 ) . Advances in fluorescent microspheres [74] , plasticene casting , imaging , and image analysis all hold promise for a careful gauging of the vascular system . In this paper we have begun the process of relaxing some assumptions of the canonical model . Although these modifications produce interesting results , they do not fully address the riddles discussed above . Addition of further biological realism , such as asymmetric branching or the flow characteristics of the slurry of blood cells at small vessel sizes , may generalize the WBE model from an asymptotic predictor of metabolic scaling into a universal theory that provides an understanding of which properties of resource distribution networks are most relevant for metabolic scaling in any given biological context . This will enable testing the very soundness of the WBE framework ( Assumption 1 ) and the extent to which the cardiovascular system shapes one of the most wide ranging regularities across animal diversity . | The rate at which an organism produces energy to live increases with body mass to the 3/4 power . Ten years ago West , Brown , and Enquist posited that this empirical relationship arises from the structure and dynamics of resource distribution networks such as the cardiovascular system . Using assumptions that capture physical and biological constraints , they defined a vascular network model that predicts a 3/4 scaling exponent . In our paper we clarify that this model generates the 3/4 exponent only in the limit of infinitely large organisms . Our calculations indicate that in the finite-size version of the model metabolic rate and body mass are not related by a pure power law , which we show is consistent with available data . We also show that this causes the model to produce scaling exponents significantly larger than the observed 3/4 . We investigate how changes in certain assumptions about network structure affect the scaling exponent , leading us to identify discrepancies between available data and the predictions of the finite-size model . This suggests that the model , the data , or both , need reassessment . The challenge lies in pinpointing the physiological and evolutionary factors that constrain the shape of networks driving metabolic scaling . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
] | [
"physiology/cardiovascular",
"physiology",
"and",
"circulation",
"biophysics/theory",
"and",
"simulation",
"physics/general",
"physics",
"mathematics/statistics",
"computational",
"biology/systems",
"biology"
] | 2008 | Sizing Up Allometric Scaling Theory |
Tropical diseases caused by parasites continue to cause socioeconomic devastation that reverberates worldwide . There is a growing need for new control measures for many of these diseases due to increasing drug resistance exhibited by the parasites and problems with drug toxicity . One new approach is to apply host defense peptides ( HDP; formerly called antimicrobial peptides ) to disease control , either to treat infected hosts , or to prevent disease transmission by interfering with parasites in their insect vectors . A potent anti-parasite effector is bovine myeloid antimicrobial peptide-27 ( BMAP-27 ) , a member of the cathelicidin family . Although BMAP-27 is a potent inhibitor of microbial growth , at higher concentrations it also exhibits cytotoxicity to mammalian cells . We tested the anti-parasite activity of BMAP-18 , a truncated peptide that lacks the hydrophobic C-terminal sequence of the BMAP-27 parent molecule , an alteration that confers reduced toxicity to mammalian cells . BMAP-18 showed strong growth inhibitory activity against several species and life cycle stages of African trypanosomes , fish trypanosomes and Leishmania parasites in vitro . When compared to native BMAP-27 , the truncated BMAP-18 peptide showed reduced cytotoxicity on a wide variety of mammalian and insect cells and on Sodalis glossindius , a bacterial symbiont of the tsetse vector . The fluorescent stain rhodamine 123 was used in immunofluorescence microscopy and flow cytometry experiments to show that BMAP-18 at low concentrations rapidly disrupted mitochondrial potential without obvious alteration of parasite plasma membranes , thus inducing death by apoptosis . Scanning electron microscopy revealed that higher concentrations of BMAP-18 induced membrane lesions in the parasites as early as 15 minutes after exposure , thus killing them by necrosis . In addition to direct killing of parasites , BMAP-18 was shown to inhibit LPS-induced secretion of tumour necrosis factor alpha ( TNF-α ) , a cytokine that is associated with inflammation and cachexia ( wasting ) in sleeping sickness patients . As a prelude to in vivo applications , high affinity antibodies to BMAP-18 were produced in rabbits and used in immuno-mass spectrometry assays to detect the intact peptide in human blood and plasma . BMAP-18 , a truncated form of the potent antimicrobial BMAP-27 , showed low toxicity to mammalian cells , insect cells and the tsetse bacterial symbiont Sodalis glossinidius while retaining an ability to kill a variety of species and life cycle stages of pathogenic kinetoplastid parasites in vitro . BMAP-18 also inhibited secretion of TNF-α , an inflammatory cytokine that plays a role in the cachexia associated with African sleeping sickness . These findings support the idea that BMAP-18 should be explored as a candidate for therapy of economically important trypanosome-infected hosts , such as cattle , fish and humans , and for paratransgenic expression in Sodalis glossinidius , a bacterial symbiont in the tsetse vector , as a strategy for interference with trypanosome transmission .
Trypanosomatid protozoan parasites cause a variety of diseases that affect the livelihood of people in vast areas of the world . Control of such diseases depends , to a large extent , on a small set of prophylactic and therapeutic anti-parasite drugs [1] . However , as with many microbes , inappropriate use of these agents has led to an alarming increase in parasite resistance , for example with African trypanosomes and Leishmania [2] , [3] . This has created a need for novel compounds to prevent and cure diseases caused by these predominantly tropical parasites . An ideal anti-parasite agent would have a broad spectrum of activity , would be largely unaffected by mutations in the target microbe and be non-toxic in vivo . Effectors with such potential are the host defense peptides ( HDP ) , ancient and highly successful molecules that remain functional over long periods of evolution , suggesting that microbial resistance to them is not inevitable . HDP are widely distributed elements of immunity that show a broad spectrum of activity against a variety of bacteria , fungi , parasites , enveloped viruses and transformed cells [4] . HDP are evolutionarily conserved peptides that have received much attention from the medical research community for their versatility and potential as alternatives to existing antimicrobials , many of which have become less effective due to increasing resistance in the target populations [2] , [3] . In some cases , HDP have advantages over traditional drugs and antibiotics , including having a much wider spectrum of activity and an apparent robustness against resistance . Most importantly , HDP naturally occupy positions in immunity as innate immune mediators [4] and thus may show pleiotropic activity in controlling infections . HDP are expressed by a variety of cells as holoproteins that are cleaved proteolytically to release the active peptides . An impressive repertoire of more than 1000 HDP have been reported in the antimicrobial peptide database [5] , [6] . Many of these molecules target pathogens directly by disturbing the pathogen's membrane potential or by interfering with cell functioning internally , thus causing cell death by either necrosis or apoptosis . Perhaps as important , because many HDP selectively modulate host innate immune responses , they may exhibit multi-pronged effector functions . HDP , many of which were first detected by their direct antimicrobial properties , show variable chemotactic influences on a variety of immune cells including monocytes , neutrophils , eosinophils and T cells [7] , [8] . In addition , HDP can be potent and selective anti-inflammatory agents , that suppress the production of pro-inflammatory cytokines in response to lipopolysaccharide , and can enhance protective immunity and promote the transition to adaptive immunity [4] . The local antimicrobial protection conferred by peptides released at the site of challenge is enhanced by the complex interaction of these peptides with innate and adaptive elements of the immune system [8] . The cathelicidins are a family of HDP , some of which are produced and stored in the neutrophils of many mammalian species [9] . All members of the cathelicidin family contain a C-terminal cationic domain and an N-terminal cathelin portion that must be cleaved to release the active C-terminal peptide . Bovine myeloid antimicrobial peptide-27 ( BMAP-27 ) [10] is one such active peptide ( not to be confused with the very different BMAP-28 molecule [10] ) , with a predicted cationic N-terminal amphipathic helix and a hydrophobic C-terminal tail . BMAP-27 has been shown to have potent antimicrobial activity on bacteria [10] , fungi [11] , chlamydia [12] viruses [13] and parasites [14] , [15] . However , BMAP-27 showed some cytotoxic activity on human cells ( erythrocytes and neutrophils [10] ) and thus was subsequently modified by cleaving off the C-terminal hydrophobic tail to reveal a potent antimicrobial peptide , BMAP-27 [1]–[18] ( hereafter called BMAP-18 ) , which demonstrated reduced toxicity to human cells . Previous work in our lab showed that African trypanosomes were effectively killed in vitro with low concentrations of the BMAP-27 peptide , although cytotoxicity to mammalian cells was also observed [15] . The strong anti-parasite activity observed with BMAP-27 inspired us to further investigate BMAP-27 and its truncated form , BMAP-18 , for their effects on a variety of species and life cycle stages of kinetoplastid parasites and on a range of mammalian and insect cell lines . Our goal was to qualify BMAP-18 as an effective in vitro anti-parasite effector and to lay the groundwork for further investigations into using BMAP-18 as a novel therapeutic in vivo . Our results support the idea that BMAP-18 is a strong candidate for possible use as a therapeutic in animals for treating trypanosome infections or in the tsetse vector as a parasite transmission blocking agent .
Bloodstream forms ( BSF ) of T . b . brucei 427 . 01 [16] were obtained from Dr . Sam Black ( Amherst , MA ) and adapted to growth in vitro . In brief , upon thaw , BSF trypanosomes ( a dilution series from 106–103/mL ) were grown at 37°C in flat-bottomed 24-well tissue culture plates ( Falcon 3047 , Becton-Dickinson , Lincoln Park , NJ ) in modified Baltz's medium supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) [17] . Trypanosomes were examined daily in each well and only those wells showing healthy organisms in log-phase of growth were chosen for further subculture and use in minimal inhibitory concentration ( MIC ) assays ( see below ) . Procyclic culture forms ( PCF ) of T . b . brucei 427 . 01 , T . b . gambiense U2 [18] , T . b . rhodesiense ViTat 1 . 1 [19] , and T . congolense IL3000 [20] were derived from their corresponding BSF by transformation at 27°C [21] . Tsetse transmissible T . b . brucei TSW 196 PCF [22] were obtained from the Liverpool School of Tropical Medicine . All PCF were cultured in MEM/10% FBS as described [20] , [23] . The fish pathogen T . danilewskyi strain TrCa was obtained from Mike Belosevic and Debbie Plouffe ( University of Alberta ) and were grown in vitro as described [24] . Promastigotes of Leishmania donovani LD3 [25] were obtained from Dr . Sam Turco ( Lexington , KY ) . Leishmania promastigotes were adapted to a modified minimal essential medium ( MEM ) containing 10% FBS [20] , [23] and maintained in culture in vitro at 27°C . B lymphocyte hybridomas 2A11 ( anti-hypoderma C ) and 10B1 ( anti-trypanosome tubulin ) were of murine origin ( BALB/c ) and secrete IgG1 mAbs ( Pearson lab , unpublished ) . EL4 thymoma cells were also of murine origin [26] . NIH 3T3 cells were derived from an embryonic fibroblast cell line originally isolated from Swiss mice [27] . NR8383 alveolar cells were derived from a Sprague-Dawley rat lung [28] . HeLa cells were originally established from a cervical carcinoma of human origin [29] . All mammalian cell lines were grown in Dulbecco's Modified Eagles Medium ( Gibco BRL cat no . 23700-057 ) supplemented with 10% FBS , L-glutamine , sodium pyruvate , 5×10−5 M 2-mercaptoethanol and penicillin/streptomycin . CF70 cells were derived from ovarian cultures of the eastern spruce budworm Choristoneura fumiferana [30] . Sf9 cells were derived from pupal ovarian tissue of the fall armyworm Spodoptera frugipera [31] . FPMI-NL-18 cells were established from the embryonic tissues of the red-headed pine sawfly , Neodiprion lecontei [32] . Stocks of all insect cells were obtained from Beatrixe Whittome [Dept . of Biology , UVic] . Insect cell lines were grown in Grace's insect medium [33] containing 15% FBS and 0 . 25% ( w/v ) bacto tryptose broth ( Difco; Becton Dickinson , Oakville , ON ) . Sodalis glossinidius , a secondary symbiont of tsetse flies ( Glossina ) [34] , were isolated at the University of Alberta according to the protocol of Welburn and Maudlin [35] , modified by changing the method used to maintain asepsis prior to initial culture [36] . In brief , S . glossinidius from tsetse hemolymph were initially grown on monolayers of Aedes albopictus cells and after 20 days in culture were adapted to axenic growth in serum-free medium [36] . The relatively slow growing bacteria were maintained by passaging every 5 days . Clones were isolated by spreading 2 ml of culture aseptically onto Mitsuhashi and Maramorosch ( MMI ) agar plates ( 1% ( w/v ) agar ( Difco , Detroit , MI ) followed by a 15 min incubation at room temperature to allow the culture to absorb into the agar . The plates were then placed in an unvented BBL GasPak™ anaerobic jar with a BBL CampyPak Plus™ system ( Becton Dickinson , Sparks , MD ) to generate microaerophilic conditions . Incubation at 27°C for 7 days resulted in small , white , papillated colonies . Culture purity was monitored by Coomassie Brilliant Blue staining of SDS-PAGE-separated proteins that yield a unique banding pattern [36] and by polymerase chain reaction using S . glossinidius-specific primer sets [34] . The cathelicidins BMAP-27 and BMAP-27 [1]–[18] ( BMAP-18 ) were synthesized by Fmoc ( N- ( 9-fluorenyl ) methoxycarbonyl ) chemistry at the University of British Columbia's Nucleic Acid/Protein Services Facility . A comparison of the BMAP-27 and BMAP-18 sequences is shown in Figure 1 . BMAP-18 was also synthesized with a C-terminal cysteine for coupling of the peptide to the protein carrier keyhole limpet hemocyanin ( KLH ) for polyclonal antibody production and for affinity purification of the resultant antibodies ( see below ) . The peptides were analysed for purity by high-performance liquid chromatography and for identity ( and purity ) by matrix-assisted laser desorption ionization time of flight ( MALDI-TOF ) mass spectrometry . All peptides were of greater than 80% purity . Lyophilized peptides were dissolved in sterile distilled water to a final concentration of 2 . 0 mg/mL and dilutions were prepared for inhibition assays , for enzyme-linked immunosorbent assays ( ELISA ) , for affinity purification of antibodies and for immuno-mass spectrometry assays as described below . Minimal Inhibitory Concentration ( MIC ) assays were performed on trypanosomes , Leishmania , mammalian cell lines , insect cell lines and the bacterium Sodalis glossinidius using the chromogenic/fluorogenic substrate alamarBlue [37] as previously described [15] with minor changes . In brief , healthy , log phase , BSF and PCF trypanosomes , Leishmania promastigotes , mammalian cells and insect cells were adjusted to 2 . 0×104 cells/mL in fresh medium and 100 µL ( 2 . 0×103 cell/well ) were dispensed into wells of round-bottom 96-well polypropylene microtitre plates ( COSTAR/ Corning Inc , Corning , New York: Cat . No . 3790 ) containing serial dilutions ( 10 µL in 0 . 2% FBS ) of the BMAP-27 and BMAP-18 peptides . One exception was that T . danilewskyi BSF were adjusted to 5×104 cells/mL and 100 µL ( 5×103 ) were seeded per well . After 66 hours of incubation at 37°C ( for BSF and mammalian cells ) or 27°C ( for PCF , T . danilewskyi , Leishmania promastigotes , insect cells and S . glossinidius ) , 10 µL of alamarBlue ( BioSource International , Inc . , Camarillo , CA ) were added to each well and the plates were incubated for another 6 h ( final incubation of 72 h ) . Seventy µL of cell free supernatant from each well were then transferred into 96-well flat-bottomed black/white microplates ( Greiner Bio-One , CellStar , Cat no . 655079 , MJS Biolynx , Brockville ON ) for fluorescence measurement . Fluorescence was measured using a Cytofluor 2300 microplate reader ( Millipore , Bedford , MA ) set to an excitation wavelength of 540 nm and an emission wavelength of 590 nm . The alamarBlue assay measures cell viability and proliferation based on detection of metabolic activity [38] . The alamarBlue colorimeteric/fluorometric growth indicator incorporates an oxidation-reduction indicator that changes colour in response to chemical reduction by metabolically active cells . Growth related reduction causes the indicator to change from oxidized ( blue , non-fluorescent ) to reduced ( red , fluorescent ) . S . glossinidius were cultured in Mitsuhashi and Maramorosch broth ( Sigma-Aldrich Canada , Oakville , ON ) for 3 days at 27°C in an atmosphere of 5% CO2 in air [36] . Cultures were diluted in fresh medium to 5×105 bacteria per mL using the following conversion: OD600 of 0 . 8 = 1×109 bacteria/ml . Diluted bacterial suspensions ( 100 µL ) and peptides ( 10 µL of each dilution ) were loaded into the wells of round-bottom microtitre plates as described above . Because of the slower growth rate of S . glossidinius , alamarBlue was added after 72 hours and 6 hours later , the plate contents were transferred into flat-bottomed black/white microplates for fluorescence determination as described above . Suspensions of parasites ( 106 in 500 µL PBS containing 25 µg BMAP-18 ) were incubated for 15 min at 37°C ( BSF trypanosomes ) or 27°C ( PCF trypanosomes or Leishmania promastigotes ) . The treated cell suspensions were mixed with an equal volume of fixative containing 3% formaldehyde and 3% glutaraldehyde in 0 . 1 M sodium cacodylate buffer and placed in a fridge at 4°C for overnight fixation . The next day , after warming to room temperature , the fixed parasites were centrifuged in 1 . 5 ml Eppendorf microcentrifuge tubes , the supernatant decanted and replaced with new fixative . This procedure was repeated , to ensure the complete removal of serum from the parasites thus allowing them to bind to polylysine coated slides . An aliquot of each sample was placed onto separate 0 . 1% poly-l-lysine ( Sigma Chemical Company , St . Louis , MO ) treated 10 mm circular glass cover slips and these were incubated in a humid chamber for 1 hour . After washing in Karnovsky's cacodylate buffer , the samples were dehydrated in a graded ethanol series and critical point dried using liquid carbon dioxide . The cover slips were glued with silver paste onto labeled aluminum scanning electron microscopy stubs and sputter coated with gold in an Edwards S150B Sputter Coater . Digital images were collected using a Hitachi S-3500N Scanning Electron Microscope operating at 10 kV at 5000× magnification . Trypanosomal mitochondrial activity was measured using the fluorescent , cell-permeant , cationic fluorescent dye rhodamine 123 ( R-302; Molecular Probes , Eugene , OR ) that is sequestered by active mitochondria [39] . Wild type T . b . brucei 427 . 01 PCF ( 1×107/mL ) were incubated at 27°C for 10 minutes with 250 nM ( final concentration ) rhodamine 123 ( Cat . No . R8004 , Sigma St Louis , MO ) and then BMAP-18 peptide ( 5 µL of 1 . 0 mg/mL = 5 µg ) was added . At intervals , trypanosomes were analyzed by immunofluorescence microscopy and by flow cytometry . Trypanosomes were first examined by light- and fluorescence microscopy ( for observation of rhodamine 123 fluorescence ) using a Zeiss Standard binocular microscope fitted with an epifluorescence attachment and a Zeiss NeoFluor 63/1 . 25 oil immersion objective . Digital photographs were taken with a Nikon CoolPix 2700 camera at high resolution . The images were stored and manipulated in TIFF format using PhotoshopTM 5 . 5 graphic software ( Adobe Systems Inc . , San Jose , CA ) . The same populations of trypanosomes were analyzed for forward scatter ( size ) , side scatter ( granularity ) and fluorescence ( rhodamine 123 ) using a FACSCalibur flow cytometer ( Becton-Dickinson , San Jose , CA ) . For each sample , a minimum of 10 , 000 events were analyzed . Human venous blood ( 100 mL ) was collected from healthy volunteers in Vacutainer collection tubes containing heparin ( Cat No . 362753 , BD , Franklin Lakes , NJ ) according to the University of British Columbia Clinical Research Ethics Board guidelines and approval . The blood was mixed at a 1∶1 ratio with RPMI 1640 medium ( supplemented with 10% ( v/v ) FBS , 2 mM L-glutamine , and 1 mM sodium pyruvate ) and the peripheral blood mononuclear cells ( PBMC ) were separated by centrifugation in Ficoll-Paque™ PLUS ( Cat No . 17-1440-02 , GE Healthcare ) . PBMCs were isolated from the buffy coat , washed twice in sterile PBS and the numbers of live cells were determined by trypan blue exclusion . PBMCs ( 0 . 1 mL at 1×106 cells/mL ) were seeded into 96-well tissue culture dishes ( Sarstedt ) , incubated at 37°C in 5% CO2 in air and rested for 2 h before experimental treatment . To measure cytokine stimulation , PBMCs were treated for 24 h with doubling dilutions ( 100 µg/mL-1 . 56 µg/mL ) of BMAP-18 and BMAP-27 . Monocyte chemotactic protein-1 ( MCP-1 ) , the chemotactic cytokine Gro-alpha ( Gro-∝ ) , and tumour necrosis factor alpha ( TNF-a ) were measured in culture supernatants using ELISAs ( see below ) . To test for anti-endotoxin activity , PBMCs were pretreated with 5 µg/mL BMAP-18 or BMAP-27 for 30 minutes prior to the addition of 10 ng/mL of purified Pseudomonas aeruginosa LPS for 4 hours . Pseudomonas aeruginosa LPS was purified as described previously [40] . After 4 hours , secreted TNF-∝ was measured by ELISA ( see below ) . All tissue culture supernatants were centrifuged at 1000× g for 10 min to remove cells , aliquoted and then stored at −20°C before assaying for cytokines by ELISA , according to the manufacturer's instructions ( MCP-1 and TNF-α , eBioscience Inc . , San Diego , CA; Gro-α , R&D Systems , Minneapolis , MN ) . Immunization of rabbits was performed to obtain high affinity polyclonal antibodies specific for BMAP-18 . Two female New Zealand White rabbits were bled to obtain preimmunization sera and after one day of rest were immunized with BMAP-18-KLH conjugate ( subcutaneous injections of 0 . 25 mg in complete Freund's adjuvant ) . Four subsequent subcutaneous injections containing 0 . 25 mg of peptide-KLH conjugate in incomplete Freund's adjuvant were given using standard procedures . At intervals , to determine if an adequate immune response had been achieved , test bleed sera were prepared and titrated using the unconjugated , “immunizing peptides” in an ELISA where the peptide antigens are dried onto assay plates ( see “peptide ELISA” , below ) . Once adequate antibody responses were achieved , antisera were prepared from several bleeds from each rabbit . Affinity purified anti-BMAP antibodies were obtained by elution from an immunoadsorbent made by covalent coupling of BMAP-18 peptide through an added C-terminal cysteine to SulfoLink® Coupling Gel ( Cat No . 20401; Pierce , Rockford IL ) . The antibodies were used in ELISA and in an immuno-mass spectrometry assay ( see below ) to detect BMAP-18 peptide in human blood or in plasma . A modified ELISA was used to measure specific anti-BMAP-18 antibody titres and to detect BMAP-18 in human serum samples . First , a standard ELISA method [41] was modified to use anti-rabbit alkaline phosphatase second antibodies and free peptide antigens ( i . e . not coupled to protein carriers ) to coat the polystyrene microtitre ELISA plates . In this specialized “peptide ELISA” , peptide was dissolved in distilled water to 5 . 0 µg/mL and 100 µL ( 0 . 5 µg/well ) of this solution were dried onto each well by overnight incubation at 37°C in a dry incubator . Dilutions of the rabbit antisera were then used as a source of primary antibodies to determine the titres . To determine if BMAP-18 could be detected in complex protein mixtures , 0 . 5 µg amounts of BMAP-18 were spiked into 100 µL of serially diluted human serum and dried onto ELISA plates overnight , prior to adding a 1/400 dilution ( predetermined from the titration data ) of first antibody , followed by second antibody and substrate . An immunoenrichment technique coupled with peptide detection by mass spectrometry was also used to detect BMAP-18 in human blood and plasma . We used immuno-matrix assisted laser desorption ionization ( iMALDI ) time of flight mass spectrometry to detect BMAP-18 by its characteristic mass . To do this , 1 µg of BMAP-18 was spiked into 50 µL aliquots of freshly prepared undiluted human blood , plasma or serum , 2 µg of affinity purified rabbit antibody were added to each sample and the mixtures were incubated ( with shaking ) for 18 hours at 4°C to allow antibody-peptide binding . Negative controls consisted of all three samples without added peptide . Another negative control consisted of 50 µL of PBS containing 2 µg of a “wrong” affinity purified antibody and 1 µg of BMAP-18 peptide . A positive control consisted of 50 µL of PBS containing 2 µg of anti-BMAP-18 affinity purified antibody with 1 µg BMAP-18 peptide . After 18 hours of incubation , 10 µL of a suspension of washed Dynabeads ( M-280; sheep anti-rabbit IgG; cat no . 112 . 03D ) were added to all of the experimental samples , the mixtures were incubated for 2 hours at room temperature and the beads from each were washed three times with PBS before eluting the peptides with 25 µL of 5% acetic acid . The 25 µL samples were Zip-Tipped ( C18; P10 tip size; Cat No . ZTC18S960; Millipore Corporation , Billerica , MA ) to concentrate the eluted peptides and to remove salts and then the peptides were eluted in 1 µL of 50% acetonitrile/0 . 1% trifluoroacetic acid . MALDI-TOF analysis was performed by spotting the eluted peptides ( all of the 1 µL ) onto the wells of a Voyager , 100 position , stainless steel MALDI plate ( Applied Biosystems , Foster City , CA ) . After drying , the peptide spots were covered with 1 µL of matrix ( 0 . 5% alpha-cyano-4-hyrdoxycinnamic acid/0 . 18% ammonium citrate/70% acetonitrile/0 . 1% trifluoroacetic acid ) . An Applied Biosystems Voyager DE-STR mass spectrometer ( Applied Biosystems , Foster City , CA ) running in delayed extraction , reflectron mode was used to acquire MALDI-TOF data .
To compare the growth inhibitory activity of the full-length bovine cathelicidin BMAP-27 with its truncated form , BMAP-18 , both BSF and PCF of T . b . brucei parasites were cultured in the presence of varying concentrations of the peptides . After incubation for 66–72 hours , the metabolic activity of the organisms was determined using alamarBlue substrate . The results are shown in Figure 2A . Both life cycle stages of the trypanosomes were inhibited by BMAP-27 at extremely low concentrations ( 50% inhibition at <2 µg/mL for BSF , solid black line; and <10 µg/mL for PCF , solid gray line ) . The truncated BMAP-18 exhibited similar effects , with strong inhibition of both BSF and PCF trypanosomes observed at low doses of the peptide ( 50% inhibition at 8 and 12 µg/mL respectively ) . In comparison , a bacterial symbiont of tsetse , Sodalis glossinidius was not killed as effectively ( green lines ) , especially with BMAP-18 ( dashed green line ) that required at least a ten-fold higher concentration ( 50% inhibition at 100 µg/mL ) of peptide to effect similar levels of growth inhibition . To test the parasite specificity of BMAP-18 , several species of kinetoplastid parasites were tested in inhibition assays . The results are shown in Figure 2B . T . b . brucei , T . b . gambiense and T . b . rhodesiense PCF , members of the subgenus genus trypanozoon and species that are pathogens of domestic animals and humans , showed strong growth inhibition by low doses ( 50% inhibition at <10 µg/mL to a maximum of 25 µg/mL ) of BMAP-18 . T . congolense , a member of the subgenus Nannomonas , was also strongly inhibited by BMAP-18 . Promastigote forms of Leishmania donovani parasites and the non-African trypanosome , T . danilewskyi , a pathogen of fish , were also highly sensitive to killing by BMAP-18 . Both mammalian cells and insect cells were tested in the alamarBlue assay with both forms of BMAP . The results are shown in Figure 3 . All of the mammalian cell lines tested were less sensitive to BMAP-18 than to BMAP-27 ( Panel A ) . Even the rapidly growing hybridoma cell line was less sensitive than T . b . brucei ( shown on the left side of each of the panels ) used as a comparative positive control . Similarly , BMAP-18 showed reduced cytotoxicity ( approximately 2–8 fold ) on all insect cell lines ( Panel B ) when compared with the parental BMAP-27 peptide . T . b . brucei 427 . 01 PCF , T . b . rhodesiense BSF and L . donovani populations were treated with high doses ( 50 µg/mL ) of BMAP-18 , fixed and examined by scanning electron microscopy at 5000× magnification ( Figure 4 ) . All three species of parasites showed membrane damage after treatment with BMAP-18 ( Panels 2 , 4 , 6 ) whereas membranes of untreated parasites remained intact ( Panels 1 , 3 , 5 ) . Three individual T . b . brucei 427 . 01 PCF trypanosomes were photographed in the same microscopic frame in various stages of BMAP-18-based destruction after 15 min incubation ( Panel 7 ) . After 30 min incubation , no parasites had survived ( not shown ) . Rhodamine123 is a cell permeant fluorescent dye that accumulates in mitochondria , which run the entire length of the trypanosome body . It is thought that an attraction of cationic rhodamine 123 molecules by the relatively high negative electric potential across the mitochondrial membrane may be the basis for the selective staining of mitochondria by rhodamine 123 in living cells . The dye at low concentrations exhibits no cytotoxicity [40] . It is used , among other things , to indicate mitochondrial membrane disruption . After incubating T . b . brucei 427 . 01 PCF with rhodamine123 and subsequently treating the same populations with low-doses ( 5 µg/mL ) of BMAP-18 , we assessed the dye retention of the cells by fluorescence microscopy ( Figure 5 ) . Trypanosomes were first treated with rhodamine 123 and photographed . These control ( untreated ) parasites showed bright red fluorescence distributed along the mitochondria ( Panel A1 ) . Rhodamine 123 labeled trypanosomes were then treated with BMAP-18 for 10 min and photographed again . This time no fluorescence was observed ( Panel A2 ) , indicating that the mitochondrial potential was disrupted , releasing the dye . These results were confirmed by flow cytometry over a 0–30 minute time course ( Figure 5B ) . The plot clearly shows decreasing rhodamine 123 fluorescence as the incubation period of BMAP-18 with the trypanosomes was extended . A dot-plot comparison shows the effects of BMAP-18 treatment on mitochondrial fluorescence vs . side scatter ( granularity ) over time ( Figure 5C ) . The diluent control ( untreated trypanosomes after 30 min; far right panel ) showed that approximately 83% of the parasites were alive and healthy as they had strong rhodamine 123-stained mitochondria and low granularity . After incubation with BMAP-18 , the parasites remained intact and appeared to have a normal plasma membrane but showed decreasing rhodamine fluorescence and increasing granularity as the cells rounded up and underwent an apoptosis-like death . In other experiments ( not shown ) the vital dye fluorescein diacetate was retained by the parasites incubated with low levels ( 5 µg/mL ) of BMAP-18 , indicating that the integrity of the plasma membrane was maintained . Both BMAP-27 and BMAP-18 were tested for their ability to induce release of the cytokines MCP-1 , Gro-α and TNF-α from human PBMC . In four separate experiments , BMAP-27 , at a range of concentrations , stimulated release of MCP-1 and Gro-α cytokines , whereas BMAP-18 at the same range of concentrations did not . Neither BMAP-27 nor BMAP-18 directly stimulated the release of TNF-α from PBMC ( data not shown ) . However , in three separate experiments , both BMAP-27 and BMAP-18 , at physiologically relevant concentrations , strongly inhibited LPS-induced TNF-α secretion from PBMC ( Figure 6 ) . A modified “peptide ELISA” ( see Methods ) was used to measure the titres of rabbit anti-BMAP-18 antibodies without interference from the linker moieties used to couple peptides to the KLH carrier used for immunization . Titres of rabbit antisera in excess of 1∶6400 were obtained . To test the utility of these antisera for detection of BMAP-18 in complex antigenic mixtures , a dilution series of human serum was made and 0 . 5 µg BMAP-18 peptide was spiked into 100 µL of each dilution for antigen coating of ELISA plates . The rabbit antibodies detected the BMAP-18 peptide even in undiluted human serum ( data not shown ) thus affinity-purifed anti-BMAP-18 antibodies were prepared for use in a novel immuno-mass spectrometry assay . Enrichment of BMAP-18 from spiked human whole blood , plasma or serum was attempted using affinity-purified anti-BMAP antibodies and capture of the antibody-peptide complexes by magnetic Dynabeads , followed by detection of the eluted peptide by MALDI-TOF mass spectrometry . The results are shown in Figure 7 . We easily detected intact ( non-degraded ) BMAP-18 ( 2342 . 57 m/z; Panel A ) after 18 hours of incubation in human blood or plasma , indicating that it is remarkably stable and suggesting that it would remain intact for a reasonable time in the blood of humans if it were used as a potential therapeutic . In contrast , we were unable to detect intact BMAP-18 in human serum ( data not shown ) , perhaps because of the increased proteolytic activity in serum after activation of the clotting cascade .
Host defense peptides show great promise as therapeutics for a variety of infectious diseases . Here we have shown that BMAP-18 exhibits potent killing activity in vitro against a broad range of socioeconomically important parasites that infect humans , cattle and fish . However , for therapeutic applications in vivo , there are ongoing concerns about HDP toxicity , serum sensitivity and stability of the effector peptides in the blood of treated animals . In addition , there is evidence that resistance to direct anti-microbial killing by cationic peptides may arise , at least in some bacteria ( 50 ) , thus any immunomodulatory effects of HDP are a welcome adjunct to direct killing . We have , in part , addressed these concerns for BMAP-18 . First , BMAP-18 exhibited reduced toxicity on mammalian and insect cell lines . Second , BMAP-18 killing activity against several species of free-living parasites was potent in the presence of 10% fetal bovine serum in the culture medium used . Third , BMAP-18 could be detected by mass spectrometry in an unaltered form after enrichment from both whole blood and plasma using anti-peptide antibodies . Fourth , it was shown that BMAP-18 strongly inhibited LPS-induced release of TNF-α from leukocytes . Since TNF-alpha is one of the causes of cachexia ( wasting ) associated with African sleeping sickness ( 51 ) and is involved in immunosuppression in trypanosome infected animals ( 52 ) , the data suggest that BMAP-18 is an excellent candidate for testing in vivo as a therapeutic that would directly kill trypanosomes while helping to maintain the overall health of the infected host . | Protozoan parasites cause serious diseases in large areas of the tropics . Control of these diseases depends to a great extent on the use of therapeutic drugs , many of which are highly toxic . In addition , parasite resistance to several of the front-line drugs is increasing . Host defense peptides ( HDP; formerly called antimicrobial peptides ) have recently received attention as potential anti-parasite effector molecules . We earlier reported that one such peptide , bovine myeloid antimicrobial peptide ( BMAP-27 ) , is a potent inhibitor of the growth of trypanosomes and Leishmania in vitro . Here we report our studies on BMAP-18 , a truncated form of BMAP-27 , which showed reduced toxicity to mammalian and insect cells and yet retained its direct toxicity to parasites in vitro . BMAP-18 also strongly inhibited LPS-induced release of tumour-necrosis factor alpha ( TNF-α ) from human leukocytes , and thus has immunomodulatory activity . These findings suggest that BMAP-18 has potential as a therapeutic agent for treatment of infected animals or as an inhibitor of parasite transmission by their insect vectors . In anticipation of using BMAP-18 in vivo , we have also developed high affinity antibodies to BMAP-18 and have shown that these can be used , in conjunction with mass spectrometry , to detect the peptide in whole blood or plasma . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance",
"cell",
"biology/cell",
"growth",
"and",
"division",
"biochemistry/protein",
"chemistry",
"immunology/immunity",
"to",
"infections"
] | 2009 | Killing of Trypanosomatid Parasites by a Modified Bovine Host Defense Peptide, BMAP-18 |
An important problem in neuronal computation is to discern how features of stimuli control the timing of action potentials . One aspect of this problem is to determine how an action potential , or spike , can be elicited with the least energy cost , e . g . , a minimal amount of applied current . Here we show in the Hodgkin & Huxley model of the action potential and in experiments on squid giant axons that: 1 ) spike generation in a neuron can be highly discriminatory for stimulus shape and 2 ) the optimal stimulus shape is dependent upon inputs to the neuron . We show how polarity and time course of post-synaptic currents determine which of these optimal stimulus shapes best excites the neuron . These results are obtained mathematically using the calculus of variations and experimentally using a stochastic search methodology . Our findings reveal a surprising complexity of computation at the single cell level that may be relevant for understanding optimization of signaling in neurons and neuronal networks .
A central question in neuronal computation is to determine the features of neural stimuli that cause action potentials [1] , [2] . One aspect of this problem is a study of how an action potential , or spike , can be elicited by a signal with the least energy cost , e . g . , a minimal amount of applied current [3] , [4] . This problem is relevant to a number of questions in neuroscience , e . g . , what mechanisms enable sensory neurons to optimally discriminate between different percepts [5] , [6] , and what are the optimal shapes of exogenous current stimulations that cause excitation in a neuronal network for therapeutic purpose [7]–[9] . Here we investigate stimulus optimization in a well-studied neuronal preparation using computational and experimental methods . One method for determining optimal signals is the calculus of variations [10] . The rationale of this approach is that if a particular signal is optimal , small changes in signal shape cannot lead to a more effective signal for eliciting a desired response . This requirement allows a determination of relative optimum shapes . Another approach for finding optimal stimuli uses a stochastic search methodology [11] . In this method an array of stochastically determined stimulus shapes is considered , including those that displace the membrane from rest to firing . When the overall intensity of the stimulus array is reduced to a level at which action potentials rarely occur , then such rarely supra-threshold stimuli are candidate optimal shapes for eliciting an action potential . Comparison of these methods has yielded similar optimal stimulus shapes in models of biological oscillators [11] . An important step in addressing these questions is the development of a theory of optimality in single neurons . This theory should account for the complex , multi-scale and nonlinear behavior of a neuron . For example , several mechanisms are known to generate an action potential including membrane depolarization and post-inhibitory rebound excitation [2] , as illustrated in Figure 1 . A family of neighboring trajectories exists for each mechanism that takes the neuron from rest to an action potential . We seek for each mechanism the optimum trajectory that triggers an action potential with the least energy cost , for example the total current delivered . The signals a neuron receives are combinations of post-synaptic currents ( PSCs ) , which can be either excitatory or inhibitory . The duration of PSCs can vary considerably depending on cell type [12] . Moreover , the timing , number , and amplitude of PSCs also vary significantly . Consequently , PSCs can , in principle , generate a wide range of signals in the post-synaptic cell , although the properties of the post-synaptic cell limit the output that the cell can actually produce . A theory of neuronal optimality should account for these physiological constraints . In the present study we investigate stimulus optimization principles using one of the best characterized experimental preparations - the squid giant axon - and its mathematical representation , the Hodgkin & Huxley model and a recent modification of the model [13] , [14] . A major finding is that the excitatory properties of this preparation are , as suggested above , exquisitely sensitive to stimulus shape . Moreover , the neuron uses different mechanisms for generating an action potential depending on the physiological context in which it finds itself thereby requiring context dependent optimal shapes . These results on stimulus optimization in single neurons may be important for considering optimization within and across neural circuits throughout the nervous system .
The Hodgkin & Huxley model [13] consists of four state variables , V , m , h , and n , where V is membrane potential , m and h are associated with the sodium ion current , INa , and n is associated with the potassium ion current , IK ( Methods ) . The model provides an excellent description of the action potential response of squid giant axons to suprathreshold depolarizing current pulses having brief duration . It is less successful for longer duration pulses . In particular , it predicts repetitive firing for these conditions over a large range of pulse amplitudes . The axon itself fires once and only once regardless of pulse amplitude or duration [14] . This discrepancy between theory and experiment is accounted for by changing a single parameter in the equation for n , the IK gating variable [14] . Both versions of the model provide comparable descriptions of the response of the axon to brief duration pulses , which is the focus of this work , i . e . , the optimal stimulus for eliciting a single spike rather than a train of spikes . Consequently we begin our analysis with the original version of the model and compare those results with results obtained from the revised version . All simulations were carried out with the full model ( either version ) including results obtained using calculus of variations ( Methods ) . We elicited an action potential in the usual way , i . e . , with a rectangular depolarizing current pulse Istim ( t ) having slightly suprathreshold amplitude ( Figure 2B , blue tracings ) . The pulse takes the model from the rest state α0 to threshold β1 along the trajectory in V , n , and h space illustrated in Figure 2A ( blue tracing ) . We used the calculus of variations to find a neighboring Istim ( t ) trajectory that also takes the model from α0 to β1 with a minimum amount of applied root mean square ( RMS ) current ( red tracing in Figure 2A ) . The V vs t result obtained is overlaid on the rectangular pulse result in Figure 2B . The RMS current of the calculus of variations stimulus over its 20 msec duration is approximately 40% less than that of the 4 msec duration rectangular pulse . We note that the stimulus obtained from the calculus of variations contains an oscillatory component , seen as a loop around α0 in Figure 2A ( arrow ) coinciding with the oscillations in stimulus current and membrane potential shown in Figure 2B . As noted above , action potentials are also elicited following a hyperpolarizing current pulse - anode break excitation , a result referred to as post-inhibitory rebound ( PIR ) . These conditions partially remove the resting level of INa inactivation by increasing the h state variable from its resting level . The effect of a hyperpolarizing pulse on the h variable is the mechanism underlying PIR in the Hodgkin & Huxley model . We adjusted the current amplitude of a 10 msec hyperpolarizing pulse until threshold was achieved , state β2 in Figure 2A . Note that the membrane potential of β2 at the end of the hyperpolarizing pulse is below the resting level ( Figure 2C ) . Referring to this point as a threshold for spike initiation may seem counterintuitive but is consistent with the behavior of both the Hodgkin & Huxley model and squid giant axons . A hyperpolarizing pulse of insufficient amplitude or duration will fail to elicit an action potential following the pulse . Increasing both , or either , pulse parameter will generate a spike . We fixed the pulse duration at 10 msec and increased its amplitude until a spike was elicited . The V , n , and h trajectory of this result connecting α0 and β2 is illustrated in Figure 2A ( blue tracing ) . The calculus of variations was used to identify a nearby trajectory ( red tracing in Figure 2A connecting α0 and β2 ) that minimized the amount of current required for the anode break result . The V vs t tracings for both results are overlaid in Figure 2C . In this case the RMS current throughout its 20 msec duration is ∼22% less than that of the 10 msec duration rectangular pulse . Note that the timing of the action potential elicited by the pulse in Figure 2C does not exactly match that of the spike elicited by the calculus of variations signal even though both waveforms do closely overlap for some time following each respective stimulus . This result is attributable to the non-linear character of the Hodgkin & Huxley model . ( The blue and red voltage waveforms more nearly superimpose in Figure 2B . ) For both sets of results in Figure 2 the calculus of variation trajectory was optimal relative to the trajectory corresponding to a spike elicited by an excitatory or inhibitory rectangular pulse . In the above analysis the only restrictions placed on the current Istim ( t ) using the calculus of variations is that it takes the Hodgkin & Huxley model from point α0 to β1 ( or β2 ) in 20 msec with minimal RMS current . This approach is relevant for exogenous stimulation of a neuron that occurs , for example , during deep brain stimulation [7]–[9] in which Istim ( t ) is unconstrained by the intrinsic properties of the membrane . Neuronal PSCs generated endogenously are constrained by the ionic mechanisms of excitability expressed generically for a synapse by the relationship Istim ( t ) = gsyn ( t ) ( V ( t ) -Esyn ) . We used calculus of variations to find the optimal pathway to a spike - optimal gsyn ( t ) - with either excitatory or inhibitory PSCs . These results ( supporting material: Text S1 , Figures S1 and S2 ) are not substantially different from the results in Figure 2 at least when Esyn is far from the membrane potential V ( t ) . We note that more complex stimuli can be seen when V ( t ) is close to Esyn . In the remainder of this study we optimized exogenous Istim ( t ) since our experimental protocol explicitly tests for candidate optimal stimuli applied to the membrane . The calculations below ( supporting material ) suggest that the approach we are using is relevant to at least a range of endogenous synaptic currents . As noted above , stochastic perturbations can also be used to determine stimulus optimization without requiring a mathematical description of the underlying dynamics [11] . We implemented the stochastic approach experimentally using squid giant axons . We used noise that consisted of excitatory and inhibitory model PSCs having rise and decay time constants based on experimental data [12] . In any given experimental run PSC shapes were kept the same . The times at which PSCs were added to the input signal were determined using a random number generator . Figure 3 illustrates an example of our experimental protocol along with results obtained from a single axon . A 100 sec stimulus was applied that consisted of PSCs having a one millisecond decay constant where excitatory and inhibitory PSCs were generated with equal probability . The details of the stimulus are illustrated in the bottom trace of Figure 3A which is a one second portion of the signal shown on an expanded time scale . The intensity of stimulation was adjusted to so that spikes were elicited infrequently ( 0 . 05–1 Hz ) as required by the stochastic search methodology [11] . We analyzed the portions of the run during which spikes were elicited to determine the specific attributes of the stimulus that preceded the action potentials . All spikes were aligned at the time of their peak voltage ( Figure 3B , top panel ) . The underlying stimulus currents were similarly aligned so that a spike-triggered average of the stimulus could be obtained ( Figure 3B , middle panel ) . The average values of the current ( ±2 SEM ) are shown in the bottom panel of Figure 3B . Based on previous work [11] , we hypothesized that the average stimulus prior to the spike is an optimal stimulus shape , i . e . , this signal should elicit an action potential with minimal current . To test this hypothesis , we applied this stimulus to the same axon from which the results in Figure 3A were obtained and found that it did , in fact , elicit an action potential ( Figure 3C ) . Note that the candidate optimal stimulus in the bottom panel of Figure 3B is shown in Figure 3C on a different time scale below the action potential elicited by the stimulus . A rectangular depolarizing current pulse having the same RMS current amplitude as the optimal stimulus failed to elicit an action potential ( Figure 3C ) . The comparison of the effects of the experimentally determined optimal stimulus with rectangular pulses is further illustrated in the bottom tracings of Figure 3C shown on a compressed time scale relative to the results in the top panels of Figure 3C . Rectangular pulses having the same RMS current amplitude as the optimal stimulus and with durations ranging between 1 and 10 msec were applied to the axon . None of the pulses elicited a spike . Next we asked how close any 20 msec portion of the input signal ( Figure 3A ) was to the shape of the optimal stimulus ( similarity index ) . To do this we considered the 20 msec of signal prior to each time point and convolved these signals with our time-reversed candidate optimal signal . The results are given by the histogram in Figure 3D which is very close to a Gaussian distribution . We then considered the 20 msec signals that preceded each action potential . Every one of these signals that elicited an action potential ( shown by arrows in Figure 3D ) were greater than 2 standard deviations from the mean in this histogram , indicating a high correlation with the optimal stimulus . This result also indicates that the optimal stimulus has strong predictive value in determining when the axon will fire . The experimental protocol and analysis illustrated in Figure 3 was carried out on a total of seven axon preparations . In all seven we confirmed the results shown in Figure 3B–D . The optimal noise stimuli obtained from each experiment ( including the result in Figure 3C ) are shown superimposed in Figure 4 . A visual comparison of the noise-derived optimal signal in Figure 3C with the calculus of variation waveforms in Figure 2B & C reveals important differences in stimulus shape . The result in Figure 2B has a marked depolarization phase early in the signal that is not clearly apparent in the experimental results obtained with the stochastic approach ( Figures 3C & 4 ) . The experimental results have two clear phases: a marked hyperpolarization followed by depolarizing phase just prior to spike initiation . Not surprisingly , therefore , neither of the waveforms in Figure 2 elicited a spike from the axon preparation described in Figure 3 when the RMS amplitude was adjusted to match the RMS level of the noise stimulus ( results not shown ) . In other words , the noise-derived optimal shape was superior to the shape derived from the calculus of variations . We hypothesized that these differences between the experimental and theoretical results might be attributable to our observation that rectangular pulses do not optimally elicit spikes ( Figure 3C ) . The initial conditions used for the waveforms in Figure 2B & C were the set of values for V , m , h , and n corresponding to rest - the starting point for the calculus of variations - and the set of values for V , m , h , and n corresponding to the end of rectangular pulses - the end point for the calculations . Since rectangular pulses do not themselves optimally elicit spikes , the observation that the values of the Hodgkin & Huxley model obtained from similar pulses do not yield optimal stimuli using calculus of variations is not surprising . There are many other final conditions ( combinations of V , m , h , and n ) that also lead to a spike . Thus we used the results for V , m , h , and n at the end of the depolarizing and hyperpolarizing pulses in Figure 2 as a starting point for additional simulations to determine waveforms that were optimal based on the RMS current metric . Specifically , we made small changes in one or more of the four parameters from their initial conditions for both the depolarizing and hyperpolarizing pulses and determined if these new values resulted in an Istim ( t ) waveform having a lower RMS current . This procedure was iterated repeatedly ( a coordinate search ) until we found local minima that we hypothesized do correspond to separate , optimal pathways for firing . The results of the analysis described above are illustrated in Figure 5A . The shapes of the new optimals depicted in blue and green are similar to their counterpoints in Figure 2B & C , respectively . The relative RMS currents of the curves in Figure 5A are different . Specifically , the RMS current of the blue curve is 38% less than that of the green curve , which suggests that it is the more optimal result . This waveform compares favorably with the optimal stimulus determined from the noise analysis in Figure 3 , as shown in Figure 5B . Both of these results have a slight depolarizing phase in the early portion of each respective signal , a feature not apparent in all experimental results ( Figure 4 ) . Analysis comparable to that of Figure 5A on our modified version of the Hodgkin & Huxley model noted above [14] produced a waveform without the initial depolarizing phase ( Figure 5C ) . The revised model provides an excellent description of some of our results ( Figure 5C & Discussion ) . We note that the theoretically derived waveforms in Figure 5B & C have not been tested experimentally to see if they optimally elicit spikes from the axon . We have determined waveforms experimentally that do elicit spikes , optimally , using the stochastic approach described in Figure 3 . Those waveforms are very similar to our theoretical results as shown in Figure 5B & C , which suggests that the latter would also elicit spikes , optimally , from the experimental preparation . The similarity of results obtained from two very different approaches - one theoretical , the other experiments – provides a testable prediction of our theoretical work , a prediction that is well met based on the results in Figure 5B & C . The analysis of Figure 5A demonstrates two local minima of stimulus optimality . The more optimal of the two ( blue curve ) is consistent with the optimal stimulus obtained from the noise analysis in which both excitatory and inhibitory PSCs were used ( Figure 5B & C ) . We hypothesized that the less optimal result ( Figure 5A , green curve ) might correspond to conditions in which only inhibitory PSCs were used . The results in Figure 6 describe an experimental test of this idea . The spike-triggered average current waveform for these conditions is illustrated in Figure 6A . We note that the depth of the hyperpolarizing phase is approximately twice as large as the depolarizing phase just prior to spike initiation . By contrast the amplitudes of these phases are approximately the same when mixed inhibitory and excitatory PSCs are used ( Figure 4 ) . These results provide evidence that the optimal stimulus for spike initiation depends upon overall state of inputs to the axon . We also note that the result is Figure 6A is qualitatively similar to the green curve in Figure 5A ( also shown in Figure 6B ) in that both have a slight depolarizing phase followed by a strong hyperpolarizing phase . Consequently experiments in which only inhibitory PSCs are used do appear to favor the less optimal of the two waveforms in Figure 5A . As noted above ( Introduction ) PSC duration can vary according to neuronal cell types ( [12] , and references therein ) . Since we have shown that the optimal input signal depends upon the type of inputs to the neuron ( inhibitory vs excitatory ) and that these optimal signals have different time scales ( Figure 2 ) , we hypothesized that short and long PSCs would optimally excite the neuron with different stimulus shapes . We repeated the experiments described in Figure 3 using a balanced combination of excitatory and inhibitory PSCs as in those experiments but with either a short ( 1 msec ) or a long ( 20 msec ) decay time constant . Figure 7A shows spike-triggered stimulus averages for the short ( blue ) and long ( green ) PSCs . Note that 5–10 milliseconds before a spike , the short PSC signal is excitatory ( Figure 7A , blue ) , whereas the long PSC signal is inhibitory ( Figure 7A , green ) . We tested the significance of the difference between the two results in Figure 7A using correlation analysis ( Methods ) . The 20 msec portion of the noise signal preceding each spike in the short PSC experiment was correlated with the spike-triggered averaged signal from these experiments ( blue trace in Figure 7A ) . These results shown in Figure 7B ( panel a ) are , not surprisingly , clustered close to a correlation value of one . A correlation of the 20 millisecond portion of the noise signal preceding each spike in the short PSC case with the spike-triggered average from the long PSC experiment ( Figure 7B , panel c ) gave correlation values between 0 and 1 , indicating a poor correlation . A similar analysis of the 20 millisecond portion of the noise signal preceding each spike in the long PSC experiment correlated with the spike-triggered average determined from the long and short PSC cases are illustrated in Figure 7B , panels b and d , respectively . To further explore the difference between short and long PSCs we increased the excitability of the axon , as demonstrated previously [15] , by raising the internal pH and repeating the experiments described in Figure 3 . We found that the optimal shape with short PSCs consisted of a growing sinusoidal stimulus with alternating periods of excitation and inhibition ( Figure 7C ) . The long PSC signal consisted mainly of inhibition , with a less prominent superimposed sinusoidal fluctuation . Thus , the difference between short versus long PSCs appears to be more pronounced when the neuron has increased intrinsic excitability .
We have shown that a single neuron can be highly discriminatory for the shape of low amplitude stimuli that elicit an action potential and that the shape of the optimal stimulus is dependent upon input context , i . e . , the optimal stimulus for eliciting a spike is determined by the nature and the type of all inputs to the neuron . Our results validate two methodologies to study optimality in neuronal systems . Using the calculus of variations , we determined optimal signals for the Hodgkin & Huxley model . This theory predicts that our stochastic search methodology derived from experiments should converge to the optimal stimulus derived from the theoretical approach , a prediction that is supported by the results in Figure 5B & C . Although optimality has been explored previously in simplified models [4] , [11] , [16] , these are the first results using a complete ionic model of a neuron , which enabled us to demonstrate multiple mechanisms to elicit an action potential . Using a stochastic search methodology , we determined optimal signals in the squid giant axon . Unlike other studies that use spike-triggered averaging , we used minimally supra-threshold stimulation that is required to accurately determine optimal stimulus shapes [11] , [17] . Careful titration of stimulus intensity to minimally suprathreshold levels enabled us to show that optimal shapes depend on the physiological context in which stimuli are presented . A novel feature of our analysis concerns two versions of the Hodgkin and Huxley model [13] , [14] . The original version [13] predicts sustained firing of action potentials in response to a sustained , suprathreshold depolarizing current pulse . The axon preparation fires only once for these conditions , a result that is mimicked by our revised version of the model [14] . We applied the calculus of variations approach to both and found similar results ( Figure 5B & C ) , which is not surprising since both models provide a good description of responses to brief duration pulses . The revised model provides a slightly improved description of our results compared to the original model ( Figures 4 & 5 ) and the reduction in the oscillatory component of the theoretical results ( Figure 5B & C ) is consistent with the change from repetitive firing in the original version of the model in the response to long duration pulses ( oscillatory behavior ) compared to a single spike in the revised version for these conditions ( absence of oscillations ) . Our results indicate that questions of optimality are more complex than the one model-one optimal view that is widely found in the discussions of neuronal excitation . While simpler qualitative models which are more amenable to mathematical analysis than ionic models can also be used to qualitatively predict optimal signal , they may miss the multiple locally optimal signals that are needed to understand the full landscape of neuronal signaling . For example , an integrate-and-fire model does not predict post-inhibitory rebound excitation nor does it predict neuronal firing with inputs consisting solely of inhibitory PSCs . Multiple optimal signals could allow a neuron to be responsive to a wider range of stimuli , where stimulus context is key to understanding neuronal optimality . As further details of this context are considered [18] , e . g . synaptic placement along a dendritic tree , both active and passive dendritic processing , synaptic facilitation/depression , all of which affect the temporal dynamics and polarity of the input stimulus to the soma , the role of separate firing mechanisms and multiple optimal signals will likely become even more important . We have shown that PSC duration is an important factor in stimulus optimization ( Figure 7 ) . Further experiments could be carried out in which the duration of inhibitory PSCs are different than those of excitatory PSCs or the duration of either PSC type is itself a variable factor in the experiments . Additionally , we have relied on RMS current minimization as our criterion for stimulus optimization . Other minimization strategies could be used in future experiments such as one in which the rate of change of input current is used in conjunction with RMS current . Minimization of RMS current is directly relevant to deep brain stimulation protocols [7]–[9] . Its significance in other contexts such as sensory processing is less clear . For example , in the visual system quantum efficiency is perhaps the most relevant measure of optimality , i . e . , the ability of an observer to detect a visual input with the fewest number of photons possible [5] . The relationship of RMS current input to this type of optimization is itself of topic of further research , as it the relationship of RMS current to stimulus optimization in other sensory modalities . Those studies , which have yet to be carried out , may demonstrate the relevance of the optimization of current shapes and current amplitude in the behavior of neural networks during information processing .
Experiments were carried out on squid giant axons using methods previously described [15] . Stochastically varying current was administered to the axon for 10 sec periods using stimulus profiles generated by computer ( MatLab ) of a simple model of stochastically summated polysynaptic currents ( PSCs ) . Excitatory and inhibitory PSCs were generated independently , each with a Poisson rate having a mean of 10 events per msec . Each PSC had an exponential rise time constant of 0 . 25 msec and decay time constant of 1 msec [15] . These parameters were used in all runs unless otherwise noted . The stimulus profile was the sum at any moment of all PSCs . The overall intensity of the stimulus was varied by changing the amplitude of all PSCs . The computed stimulus profiles were converted to an analog stimulus using a D-A converter ( National Instruments , Austin , TX ) controlled by software ( LabView 6 , National Instruments ) . The mean current for any run was zero because the excitatory and inhibitory PSCs had identical profiles and Poisson distributions . The exception was the experiment described in Figure 6 for which only inhibitory PSCs were used . | Computational neuroscience seeks to understand the mechanisms by which signals excite a neuron or a neuronal network . An important consideration in these studies is optimality , i . e . , what signal most effectively causes excitation . Optimization of neuronal signaling is important for networks that need to minimize energy costs , for sensory neurons to selectively respond to specific stimulus features , and for therapeutic deep brain stimulators to maximize battery life . Here we show in a classic mathematical model of the action potential and in experiments on a single cell preparation that: 1 ) a single neuron can be highly discriminatory for the shape of low amplitude stimuli that elicit an action potential and 2 ) the shape of the optimal stimulus depends upon the overall state of inputs to the neuron . Our findings reveal a surprising complexity of computation at the single cell level that may be important for understanding physiological function of the nervous system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"mathematics",
"biology"
] | 2011 | Optimal Stimulus Shapes for Neuronal Excitation |
The Josephin Domain ( JD ) , i . e . the N-terminal domain of Ataxin 3 ( At3 ) protein , is an interesting example of competition between physiological function and aggregation risk . In fact , the fibrillogenesis of Ataxin 3 , responsible for the spinocerebbellar ataxia 3 , is strictly related to the JD thermodynamic stability . Whereas recent NMR studies have demonstrated that different JD conformations exist , the likelihood of JD achievable conformational states in solution is still an open issue . Marked differences in the available NMR models are located in the hairpin region , supporting the idea that JD has a flexible hairpin in dynamic equilibrium between open and closed states . In this work we have carried out an investigation on the JD conformational arrangement by means of both classical molecular dynamics ( MD ) and Metadynamics employing essential coordinates as collective variables . We provide a representation of the free energy landscape characterizing the transition pathway from a JD open-like structure to a closed-like conformation . Findings of our in silico study strongly point to the closed-like conformation as the most likely for a Josephin Domain in water .
Proteins are fascinating molecular machines capable of organizing themselves into well-defined hierarchical structures through a huge number of conformational changes , in order to accomplish a wide range of cellular physiological functions . From an energy landscape point of view , protein conformational changes may be characterized by transitions from a low-energy conformation to another . In this connection , computational approaches have widely demonstrated their utility by providing important insights into the protein conformational features [1–5] . Molecular Dynamics simulations , and in particular enhanced sampling techniques , are able not only to predict protein transition pathways , but also to quantify the free-energy landscape along selected reaction coordinates , thus playing a key role in describing protein tendencies towards specific conformational rearrangements . Approaching this problem from an energetic point of view is of great importance especially in case of amyloidogenic proteins , given the intimate interconnection between the functional energy landscape and aggregation risk [6] . The Josephin Domain ( JD ) , i . e . the N-terminal domain of Ataxin 3 ( At3 ) protein , is an interesting example of competition between physiological function and aggregation risk [6 , 7] . In fact , the fibrillogenesis of Ataxin 3 is responsible for the spinocerebbellar ataxia 3 , also called Machado Joseph Disease ( MJD ) . Structurally , At3 is composed of a structured globular N-terminal region ( i . e . the JD , residues Met1-Arg182 in the human protein ) , combined with a more flexible C-terminal tail that contains the polyQ tract and the Ubiquitin Interacting Motifs ( UIM ) [8 , 9] . The expansion of polyglutamine ( polyQ ) tract in Ataxin 3 ( so-called expanded At3 ) is considered a cause for protein misfolding and aggregation , but the underlying mechanism remains to be elucidated . Although it is known that the polyQ tract is necessary for kinetic instigation of an aggregation mechanism [10–14] , several experimental studies support the hypothesis that JD structural stability could play a major role in determining the aggregation features and toxicity of polyQ proteins [7 , 15–21] . In this regard , experimental evidences have suggested a two-stage pathway for At3 fibrillogenesis: the first , JD-mediated and the second , polyQ-dependent [19 , 22 , 23] . Fibrillar aggregates of both not-expanded At3 and isolated JD have shown markedly similar morphological and mechanical properties , suggesting a leading role for the JD in the mechanism of fiber formation [17] . Moreover , inhibition of JD self-association by a small heat-shock protein significantly slows down expanded At3 aggregation [24] . For these reasons , the role of JD has been the subject of a robust debate in the past [6 , 7 , 18–21 , 25 , 26] . To date , several JD models solved by NMR are available in the literature ( PDB entry 2JRI [27] , 1YZB [28] , 2AGA [29] and 2DOS [30]—UNIPROTID: P54252 ) . Differences in the available models are located in the hairpin region ( region α2-α3 , residues Val31-Leu62 ) . In particular , the 1YZB and 2JRI models are characterized by a “half-open” and “open” L-shape hairpin conformation , respectively . On the other hand , the “closed” 2AGA and “half-closed” 2DOS models exhibit the hairpin region packed against the rest of the globular structure [30] . Whereas all the above-mentioned NMR data have demonstrated the existence of several different available conformations for the JD , issues concerning i ) the likelihood of JD achievable conformational states in solution , and ii ) the role played by environmental conditions ( such as the solution’s pH ) and interacting physiological partners ( such as ubiquitin ) in JD conformational arrangement are still unresolved . Specifically , results from a recent characterization of the JD free energy landscape using MD simulations suggested the open-like model as the most representative of the JD structure in solution [31] . Nevertheless , other previous experimental and computational studies strongly support the idea that JD has a highly flexible extended hairpin in dynamic equilibrium between open and closed states [1 , 30] . In a very recent in silico study , the early stage of the JD-JD dimerization mechanism [1] has been investigated by MD and indicates that the JD-JD binding might play a role in determining the kinetics of hairpin opening/closure . However , the previous computational investigation is limited to a classical MD approach with a relatively short simulation timescale [1] . In principle , to prove the JD open-like or the JD closed-like configuration as favored , it would be necessary to show not only that i ) there is more sampling in one state during a classical MD , but also that ii ) several transitions between states are sampled during the simulation . Hence , an accurate evaluation of the JD conformational changes requires a longer simulation time-scale and robust sampling methods . In this regard , enhanced sampling methods represent a powerful tool to improve the sampling efficiency of classical MD [32–40] , including those that artificially add an external driving force to guide the protein from one structure to another [38 , 41] . Moreover , reducing the dimensionality of the trajectory obtained from MD simulations can help identify the dominant modes in the motion of the molecule [38 , 41] . Motivated by the still open debate regarding the most representative JD structural arrangement [1 , 30 , 31 , 42 , 43] , we have carried out an investigation on JD conformational changes using both unbiased MD and Metadynamics guided by essential coordinates . In this work , we provide an estimation of the free energy landscape characterizing the transition pathway from a JD open-like to a closed-like structure ( which is henceforth called the folding pathway ) . The findings of our in silico study strongly suggest the closed conformation as the most likely for a Josephin Domain in water .
The 1YZB model [28 , 42] of JD was selected as the starting point for the present work . The rationale for this choice is in the experimental work of Nicastro et al . [42] indicating a satisfactory validation of the 1YZB model through the application of an arsenal of tools for checking the quality , accuracy and mutual consistency of the structures available . The 1YZB model was solvated in a dodecahedron box where the minimum distance between the protein and the edge of the box was 1 nm , resulting in a molecular system of about 50 , 000 interacting particles . The net charge of the system was neutralized by the addition of Cl− and Na+ ions . AMBER99-ILDN force-field [44–46] and water TIP3P model [47] were chosen to describe the system’s topology . Particle-Mesh Ewald method with a short-range cut off of 1 . 2 nm was applied to treat electrostatics . A cut-off of 1 . 2 nm was also applied to Lennard-Jones interactions . The system was minimized by the steepest descent energy minimization algorithm ( 1000 steps ) . Then , in order to increase the statistics of MD data , five replicas , differing in initial atom velocities , were created from the minimized system . In particular , for each replica , a random velocity taken from a Maxwell-Boltzmann distribution at 310 K was assigned to every atom of the system ( i . e . JD , water and ions ) . A position-restrained and production MD simulations were carried out as described in the following . Two subsequent MD simulations ( 500 ps , and 100 ps , respectively ) were run in the NPT ensemble , applying position restraints of 1000 , and 100 kJ/mol/nm2 , respectively , to the JD Cα atoms . System temperature was set at 310 K by using the v-rescale [48] thermostat with a coupling time step of 0 . 1 ps . Moreover , in NPT simulations a Berendsen barostat [49] was also employed with a reference pressure of 1 atm and a coupling time step of 1 . 0 ps . A third position restrained dynamics simulation ( 100 ps ) was carried out by applying a force constant of 10 kJ/mol/nm2in the NVT ensemble at 310 K . Finally , an unrestrained production MD of 500 ns was run in the NVT ensemble at 310 K , as done in several previous Molecular Dynamics studies [50–52] . GROMACS 4 . 6 package was employed for all MD simulations and data analysis [53] . Ensemble data taken from all production MD trajectories of the above-mentioned five replicas ( each simulated for 500 ns ) were used for JD conformational analysis . The Visual Molecular Dynamics ( VMD ) package [54] provided the visual inspection of the simulated systems . Dedicated GROMACS tools were used for quantitative analyses in terms of Root-Mean-Square Deviation ( RMSD ) and Root-Mean-Square Fluctuation ( RMSF ) . The secondary structure of the protein has been calculated by the STRIDE software [55] on several snapshots along the simulation time . The identification of JD conformational transitions from open to closed JD conformations has been carried out by employing quantities which have already been demonstrated to be meaningful in describing the JD transition pathway: the radius of gyration ( RG ) and the hairpin angle [1 , 31] . Given that the NMR models ( 1YZB , 2JRI , 2DOS , and 2AGA ) considered in this work present a different number of residues , the RG has been calculated by considering all the residues in common among the above mentioned PDB models . In detail , the residue range 1Met-171Asp ( according to 1YZB numbering ) has been chosen . The hairpin angle was calculated from the centers of mass of the Cα atoms from three distinct JD regions: globular subdomain ( residues 111–113 , 122–125 and 162–165 ) , hinge ( residues 32–35 ) and loop ( residues 45–48 and 58–61 ) [31] ( Section 3 in S1 Text ) . Principal Component Analysis ( PCA ) was applied to classical MD trajectories . PCA is an established method which allows to elucidate large-scale and low-frequency modes , respectively , yielding collective motions directly related to a specific molecular event [56] . In detail , after the alignment of the JD Cα Cartesian coordinates , the covariance matrix was calculated and diagonalized ( Section 2 in S1 Text ) . The free energy landscape representing the JD folding pathway was investigated by means of Metadynamics [58 , 59] , a powerful technique for enhancing the sampling in MD simulations and reconstructing the free-energy surface as a function of few selected collective variables ( CVs ) . The first eigenvector derived from the PCA was used as CV for a well-tempered Metadynamics simulation of 500 ns starting from the open-like 1YZB model [16] . The JD model was prepared for Metadynamics by applying system minimization and position restraint dynamics , as described above for the classical MD . To perform Metadynamics simulations , a Gaussian width of 0 . 1 was used . Along the simulation , the initially prescribed Gaussian deposition rate value of 0 . 2 kJ/mol·ps was used and it gradually decreased on the basis of an adaptive scheme , with a bias factor of 20 . The setting of Gaussian width and deposition rate was done on the basis of a well-established scheme [37 , 40] . In particular , the Gaussian width value was of the same order of magnitude as the standard deviation of the collective variable , calculated during unbiased simulations ( production MD ) . The authors have also verified that the maximum force introduced by a single Gaussian distribution is smaller than the typical derivative of the free energy . The estimation of the free energy profile was performed by employing the reweighted-histogram procedure [60 , 61] , taking into account for the following collective variables: the projection along the first PCA eigenvector , the JD’s RG , the hairpin angle and the alphaRMSD variable . More specific information about the definition of the CVs , the convergence of the Metadynamics simulations and the free energy reconstruction is reported in Section 3 in S1 Text . GROMACS 4 . 6 package patched with PLUMED was employed for metadynamics simulations and data analysis [57] .
As stated above , five independent replicas of a single JD in explicitly modeled water and ions were simulated for 500 ns . Structural conformational properties and stability were initially checked by monitoring the time evolution of the RMSD and secondary structure ( Section 1 in S1 Text ) . The data generated indicated that a reasonable stability of the above-mentioned quantities has been reached in all cases in the last 100 ns of the production run of the MD simulations . Moreover , the JD secondary structure showed to be highly conserved throughout the whole simulation time ( Section 1 in S1 Text ) . The time evolution of the RG calculated over the classical MD trajectories ( Fig 1A ) reveals the JD transition for all replicas , from a half-open ( starting configuration 1YZB ) to a closed or half-closed conformation , characterized by RG lower than 1 . 6 nm ( Fig 1A ) and a hairpin angle lower than 80° ( Fig 1B ) . Several intermediate half-open and half-closed conformations are explored during the MD simulation ( Fig 1C ) . Moreover , no transition from the reached JD closed-like to the open-like structure has been detected during the simulated time . By analyzing the same data in the form of a distribution plot ( Fig 2 ) , it is possible to observe that sampled structures in the stability region of the simulation ( 400–500 ns ) are far from open like JD arrangement . Secondly , it is worth mentioning the different distribution shape when considering the whole trajectory ( 0–500 ns , black dashed line in Fig 2 ) and the trajectory in the stability region ( 400–500 ns , red line in Fig 2 ) . Specifically , a curve comparison of both the RG and hairpin angle distribution indicates how intermediate states have a tendency to converge toward closed like arrangements . The importance of using both the RG and hairpin angle to analyze the JD arrangement is demonstrated by looking at the NMR range values labeled in Fig 2 . Namely , whereas the RG helps in discerning between half-open ( 1YZB ) and open ( 2JRI ) JD arrangement ( Fig 2A ) the hairpin angle perfectly distinguish between closed ( 2AGA ) and half-closed ( 2DOS ) JD ( Fig 2B ) . To reduce the high-dimensionality of the MD trajectory and to identify the dominant molecular phenomena related to the hairpin closure , PCA was applied . After the alignment of the JD Cα atoms , the MD trajectory was filtered to show only the motion along the first eigenvector , calculated by covariance matrix diagonalization . More information on PCA and the eigenvector values is reported in Section 2 in S1 Text . The JD Root Mean Square Fluctuation ( RMSF ) calculated over the filtered trajectory ( Fig 3 ) shows that , as expected , the first PCA eigenvector effectively captures the hairpin motion ( RMSFhairpin>0 . 5 nm ) . The first eigenvector derived from the PCA was used as collective variable ( CV ) for a well-tempered Metadynamics simulation . Analyzing the free energy profiles reported in Fig 4A as a function of the RG , two energy wells of 36 kJ/mol and 4 kJ/mol , located at RG values of 1 . 55 nm and 1 . 78 nm , respectively , can be identified . This result is also confirmed by reweighting the free energy profile as function of the hairpin angle ( Fig 4B ) . In this case , the deepest free energy minimum ( 36 kJ/mol ) is found to be located at a value of the hairpin angle equal to 63° . A second minimum ( 5 kJ/mol ) is found to be located at a value of the hairpin angle equal to 100° . As expected , the RG and hairpin angle values corresponding to the free energy wells ( Fig 4A ) are in agreement with the distribution peaks obtained from the unbiased MD simulations in the stability region ( 400–500 ns ) shown in Fig 2 . This finding confirms the reliability of our Metadynamics results given that the free energy minima are expected to identify the most energetically favorable configuration . An overall picture of the JD free energy landscape is provided in Fig 5B , showing the 2D color map of the free energy profile as a function of the RG and hairpin angle . Again , the free energy minima are expected to match the most energetically favorable JD configurations . Hence , it is interesting to compare the free energy map with the JD configurations sampled by classical MD ( Fig 5A ) . Fig 5A also provides the snapshots derived from the JD models available in the literature . Interestingly , 2AGA , 2DOS and 2JRI models lie in regions regularly sampled by classical MD , and characterized by absolute or relative free energy minima . The most sampled configurations by classical MD , corresponding in term of RG and hairpin angle to the 2AGA model , is also the deepest energy minimum in the free energy landscape . Similar characteristics between metadynamics lowest energy state and the 2AGA model are also highlighted by contact maps reported in Section 2 in S1 Text . On the contrary , values of the RG and hairpin angle corresponding to the 1YZB , i . e . , starting structure of our simulations , are merely sampled by classical MD and far from the absolute energy minimum in Metadynamics .
The characterization of the free energy landscape in a protein folding pathway represents a significant contribution to both experimental and theoretical approaches , given the intimate interconnection between the functional energy landscape and aggregation risk [6] . The JD folding pathway is an issue still debated in the literature [30 , 31 , 42] since a decisive proof of the most likely JD conformation has not been provided yet . Several JD models , solved by NMR , are available in the literature: open ( 2JRI ) [27] , half-open ( 1YZB [28] , closed ( 2AGA ) [29] and half-closed ( 2DOS ) [30] . The above mentioned models have been questioned and debated in the recent literature on both computational and experimental studies . In agreement with previous research in this area [30] we have recently observed the metastable behavior of the JD , which is dynamically able to switch between an open-like and closed-like structure during the dimerization [1] . However , in order to predict the JD free energy landscape , the conformational ensemble provided by the classical MD simulation is not adequate . In particular , data from MD with a timescale of hundreds of ns show that the JD closed-like state is achievable starting from an open-like state whereas a transition from a completely closed-like to an open-like state has never be detected ( Fig 1 and Section 1 in S1 Text ) . The classical MD simulation sampling is in this case insufficient , because when the system is trapped in the energy minimum characterized by the closed-like structure , thermal fluctuations are not enough to overcome the energy barrier ( around 36 kJ/mol ) needed to switch to an open-like configuration . A classical MD simulation might never be able to get out of such a deep energy minimum . To overcome this limitation , inherent to the classic MD , Metadynamics can be used to sample large-scale protein transitions as demonstrated by some relevant and pioneering papers in the field [62–64] . Our work brings together the efficient sampling of Metadynamics with a MD-PCA-based dimensionality reduction method . In particular , PCA was used to elucidate the transition pathway between the JD open-like and closed-like models , and Metadynamics was performed to estimate the corresponding free-energy landscape . Nevertheless , this computational approach was already successfully applied earlier [38 , 65–69] , thus confirming its promise as a successful strategy for investigating conformational changes in complex biomacromolecules . A limitation of the presented approach is , in fact , that preliminary information regarding the molecular transition is needed in order to calculate the essential coordinates: the transition from a JD open-like to a closed-like conformation allowed us to obtain the CV to guide the Metadynamics method . Our findings confirm that the JD hairpin region , which protrudes out into the solvent , can be responsible for an extensive conformational change , switching between the open-like conformation and the closed-like one . As suggested in previous works [1 , 30] , the hairpin mode of motion mainly consists of movement of region α3 ( Asp57-Leu62 ) toward α6 ( Asp145-Glu158 ) ( Figs 1C and 3 ) . Interestingly , in our simulations , when the JD is alone in water environment the most stable configuration is characterized by the hairpin packed against the globular core , in agreement with models 2AGA [29] ( Figs 4 and 5 ) . However , an energy well of about -5 kJ/mol has been also detected corresponding to the open conformation , in agreement with the 2JRI [27] model . In general , such free energy minima should predict the most favorable conformational state . In fact , the RG values corresponding to the free energy minima are consistent with the peaks of the RG distribution of the unbiased simulation at equilibrium ( Figs 1 , 4 and 5 ) , indicating the reliability of the presented approach in describing the JD free energy landscape . Surprisingly the JD half-open structure , with the RG and hairpin angle mainly corresponding to the 1YZB is the less sampled structure even during classical MD simulations . In this connection , it is important to clarify that our work is not oriented to evaluate the “quality” of produced NMR which has already been checked with proper methodologies [42] . Instead , a first novel aspect of our work is that the employed approach has demonstrated to explore the state space by granting several transitions among closed and open JD conformations . Data reported in recent literature [31] indicated the JD open structure ( particularly referring to the 1YZB model ) as the most likely JD conformation in water . Moreover , it has been emphasized how the 2AGA ensemble data were far away from the calculated free energy minima [31] . However , it is worth mentioning that , all the previous computational works [1 , 31 , 43] did not show transitions between open and closed JD and viceversa . In our opinion , such transitions are required to claim that a specific protein conformational state is characterized by lower free energy minima than another one . In fact , in the partial section of the free energy profile corresponding to the JD open-like structure , our results are in agreement with the above mentioned recent work [31] . However , the representation of the whole free energy profile describing both open and closed structures demonstrates the closed arrangement as the most likely for a Josephin Domain in water environment . Nonetheless , there may be conditions under which the open-like state is stabilized ( e . g . in the context of the full-length protein or in the presence of a physiological partner ) . For example , we have recently shown how the JD conformational state might be affected by the presence of another interacting JD [1] as well as by an inorganic surface [43] . In addition , as already suggested [30] , the JD free energy landscape could be influenced by other environmental conditions , such as temperature and pH . Further investigations are planned and will help in clarifying the influence of JD functional partners and environmental factors affecting the JD conformational arrangement . This information may be relevant not only to better understand the physiological function of the Josephin Domain , but also to provide insight into molecular phenomena characterizing the pathological nature of spinocerebellar ataxia 3 . | Proteins are fascinating molecular machines capable of organizing themselves into well-defined hierarchical structures through a huge number of conformational changes to accomplish a wide range of cellular functions . Protein conformational changes may be characterized by transitions from a low-energy conformation to another . Computer simulations and in particular molecular modelling may be able to predict protein transition dynamics and kinetics , thus playing a key role in describing protein tendencies towards specific conformational rearrangements . Approaching this problem from an energetic point of view is of great importance especially in case of amyloidogenic proteins , given the intimate interconnection between the functional energy landscape and aggregation risk . In this work we have employed molecular modelling techniques to shed light into conformational dynamics and kinetics of the Josephin Domain , part of the protein Ataxin 3 , which is responsible for the spinocerebbellar ataxia 3 , also called Machado Joseph Disease . In greater detail , we have employed enhanced sampling approaches to provide an estimation of the free energy landscape characterizing the transition pathway among several known molecular arrangements of the Josephin Domain . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2016 | Josephin Domain Structural Conformations Explored by Metadynamics in Essential Coordinates |
Kato-Katz is a widely used method for the diagnosis of soil-transmitted helminth infection . Fecal samples cannot be preserved , and hence , should be processed on the day of collection and examined under a microscope within 60 min of slide preparation . Mini-FLOTAC is a technique that allows examining fixed fecal samples . We assessed the performance of Mini-FLOTAC using formalin-fixed stool samples compared to Kato-Katz and determined the dynamics of prevalence and intensity estimates of soil-transmitted helminth infection over a 31-day time period . The study was carried out in late 2013 on Pemba Island , Tanzania . Forty-one children were enrolled and stool samples were subjected on the day of collection to a single Kato-Katz thick smear and Mini-FLOTAC examination; 12 aliquots of stool were fixed in 5% formalin and subsequently examined by Mini-FLOTAC up to 31 days after collection . The combined results from Kato-Katz and Mini-FLOTAC revealed that 100% of children were positive for Trichuris trichiura , 85% for Ascaris lumbricoides , and 54% for hookworm . Kato-Katz and Mini-FLOTAC techniques found similar prevalence estimates for A . lumbricoides ( 85% versus 76% ) , T . trichiura ( 98% versus 100% ) , and hookworm ( 42% versus 51% ) . The mean eggs per gram of stool ( EPG ) according to Kato-Katz and Mini-FLOTAC was 12 , 075 and 11 , 679 for A . lumbricoides , 1 , 074 and 1 , 592 for T . trichiura , and 255 and 220 for hookworm , respectively . The mean EPG from day 1 to 31 of fixation was stable for A . lumbricoides and T . trichiura , but gradually declined for hookworm , starting at day 15 . The findings of our study suggest that for a qualitative diagnosis of soil-transmitted helminth infection , stool samples can be fixed in 5% formalin for at least 30 days . However , for an accurate quantitative diagnosis of hookworm , we suggest a limit of 15 days of preservation . Our results have direct implication for integrating soil-transmitted helminthiasis into transmission assessment surveys for lymphatic filariasis .
Kato-Katz technique is the standard method recommended by the World Health Organization ( WHO ) for the diagnosis of intestinal schistosomiasis and soil-transmitted helminthiasis [1 , 2] . By counting helminth eggs in a given amount of stool , this method allows not only determining the presence of infection , but also its intensity , as expressed in eggs per gram of feces ( EPG ) [3 , 4] . Kato-Katz technique requires processing fecal specimens preferentially within 24 hours from production and collection in the field , in order to minimize degradation of hookworm eggs [1 , 4–6] . This strict time requirement entails that , during field surveys , the team collecting fecal samples either performs the microscopic examination on the spot or transfers the samples to a nearby laboratory for work-up the same day [7] . FLOTAC technique [8] has been developed in veterinary parasitology for the diagnosis of intestinal parasites , and it has been adopted in human parasitology due to its high sensitivity [8–10] . However , FLOTAC technique is more time-consuming than Kato-Katz method and requires some specific laboratory equipment ( e . g . , large bucket centrifuge with special adaptors ) [11] . Recently , FLOTAC has been simplified and Mini-FLOTAC has been developed in order to meet the needs of resource-limited settings . Indeed , Mini-FLOTAC is simple to apply , it allows the analysis of fixed fecal samples , and helminth eggs can be quantified , as with Kato-Katz technique [12] . The possibility of collecting fecal specimens in the field , adding a fixative , and analyzing the samples several days later in a central laboratory could overcome the time limitation of working on fresh samples , and hence improve the easiness and the quality of soil-transmitted helminthiasis diagnosis . For example , it has been suggested to integrate soil-transmitted helminthiasis within transmission assessment surveys ( TAS ) that are conducted in the context of the program to eliminate lymphatic filariasis [13–15] . Usually , the TAS team remains in a school ( or a village ) only for few hours for collecting and analyzing blood samples [16] . While this short time frame would allow collection and fixing of stool samples with formalin for subsequent soil-transmitted helminth diagnosis in the laboratory , it would not suffice to prepare and microscopically examine Kato-Katz thick smears on the spot . Hence , the use of a method that does not limit the time-to-process the stool specimens in the laboratory holds promise to be included in TAS in areas where soil-transmitted helminthiasis and lymphatic filariasis are co-endemic . Additionally , Mini-FLOTAC is a closed system , and the safe handling , together with fixing of stool samples , protect the operator from potential contamination [12 , 17] . However , the length of storage time of fixed fecal samples to maintain reliable and accurate diagnostic performance of the Mini-FLOTAC for the detection and quantification of soil-transmitted helminth eggs has not been evaluated before . The aim of this study was to assess the accuracy of the Mini-FLOTAC technique on fecal samples which had been fixed for up to 31 days ( from stool collection to microscope examination ) in maintaining good/optimal correlation in terms of prevalence and intensity of infections and a good/robust “microscope readability” ( the ease by which the different soil-transmitted helminth eggs can be identified over time ) .
This study was embedded in a randomized controlled trial to assess the efficacy and safety of different anthelmintic drugs against Trichuris trichiura and concurrent soil-transmitted helminth infections [18] . In brief , the study was approved by the ethics committees of Basel , Switzerland ( EKBB; reference no . 123/13 ) and the Ministry of Health and Social Welfare of Zanzibar , United Republic of Tanzania ( ZAMREC; reference no . 0001/June/13 ) . The trial is registered at controlled-trials . com ( identifier: ISRCTN80245406 ) . Written informed consent was obtained from the parents/guardians of the children before enrolment . Data were anonymized and confidentiality assured throughout the study . Data files were stored in a safe cabinet within the Public Health Laboratory-Ivo de Carneri ( PHL-IdC ) and children were identified by code . At the end of the study , all children were treated with a single oral dose of albendazole ( 400 mg ) as part of the mass drug administration intervention of the national neglected tropical diseases ( NTD ) control program , implemented in January 2014 . The study was carried out on Pemba Island , United Republic of Tanzania , in October and November 2013 . Pemba is part of the Zanzibar archipelago , together with the main island of Unguja and located few degrees south of the Equator , about 50 km off the coast of mainland Tanzania . Pemba is an island where soil-transmitted helminth infections are still widespread despite deworming activities that have been implemented over the past 20 years [19] . Indeed , numerous epidemiological surveys and clinical trials have been conducted , which unanimously report a high prevalence and intensity of soil-transmitted helminth infection [19–23] . Embedded in a clinical trial which assessed the efficacy and safety of different drug combinations against T . trichiura and concomitant soil-transmitted helminth infections [18] , 41 children from the primary schools of Mchamgandogo and Shungi were selected for our study . The inclusion criteria were ( i ) double or triple infection with T . trichiura , A . lumbricoides , and/or hookworm in order to obtain at least 20 infections for each of the three soil-transmitted helminth species; and ( ii ) no recent ( within the last 6 months ) anthelmintic treatment . In view of the second criterion , we chose those children who had initially been screened for soil-transmitted helminth infection in September 2013 , but were not subjected to treatment in any of the trial arms and had not yet been treated by the national NTD control program ( done only after the clinical study in November 2013 , namely in January 2014 ) . We collected one fecal sample of about 50 g from each child . Samples were transferred to the PHL-IdC . On the same day ( D1 ) , each sample was divided into 12 Fill-FLOTAC ( reusable plastic containers used to collect , homogenize , and filter stool samples ) . Each Fill-FLOTAC contained 2 g of stool , weighted with a digital scale ( CS 200 Compact Scale; People’s Republic of China , precision 0 . 1 g ) , and 2 ml of 5% formalin ( dilution 1:1 ) . Each Fill-FLOTAC was used to perform Mini-FLOTAC at different time intervals post stool collection and fixing until day 31 ( D31 ) . On the collection day ( D1 ) , each sample was subjected to a single Kato-Katz thick smear and to a single Mini-FLOTAC . Briefly , Kato-Katz thick smear was performed using a 41 . 7 mg template following a standard protocol and fecal egg counts were multiplied by a factor 24 to obtain an estimate of EPG [2 , 24] . Mini-FLOTAC technique was performed using saturated saline as flotation solution ( FS no . 2 ) , with a sample dilution ratio of 1:20; Mini-FLOTAC chambers were filled with 1 ml of sample solution per chamber [12] . The fecal egg count of each sample was multiplied by a factor 10 to obtain an estimate of EPG , therefore the sensitivity detection limit of Mini-FLOTAC was 10 EPG . From D1 until day 15 ( D15 ) the samples were analyzed every other day , and from day 16 ( D16 ) until day 31 ( D31 ) the samples were analyzed every third day with a single Mini-FLOTAC . All Fill-FLOTAC containers were stored at room temperature ( between 20 and 30°C ) at PHL-IdC throughout the study . Quality control was carried out on 10% of Kato-Katz [25] and Mini-FLOTAC slides that have been re-checked by skilled microscopists with experience on both methods . The classification of light , moderate , and heavy infection was done according to WHO recommendations: for A . lumbricoides: light 1–4 , 999 EPG , moderate 5 , 000–49 , 999 EPG , and heavy infection ≥50 , 000 EPG; for T . trichiura: light 1–999 EPG , moderate 1 , 000–9 , 999 EPG , and heavy infection ≥10 , 000; for hookworm: light 1–1 , 999 EPG , moderate 2 , 000–3 , 999 EPG , and heavy infection ≥4 , 000 EPG [26] . In order to assess the diagnostic accuracy over time , we evaluated the shape and contrast from the background of the different soil-transmitted helminth eggs by taking sample photographs of the different eggs during the study , randomly chosen among the samples . Moreover , readability of helminth eggs on the slides was assessed by asking laboratory technicians about the ease in recognizing the different eggs under the microscope at each time point on all samples . Data were entered into an Excel file . Analysis was performed using SPSS 16 . 0 EV ( WinWrap Basic , 1993–2007 ) . The results were analyzed by 2x2 contingency tables . Pearson index was calculated to assess the accuracy of the two diagnostic methods and their agreement . The strength of agreement criteria were: ≤0 indicating poor , 0 . 01–0 . 20 indicating slight , 0 . 21–0 . 40 indicating fair , 0 . 41–0 . 60 indicating moderate , 0 . 61–0 . 80 indicating substantial , and 0 . 81–1 . 00 indicating almost perfect agreement [27] . The comparison between arithmetic mean EPGs was calculate with Student’s t test for paired samples; the level of significance was set at p value <0 . 05 , and 95% confidence intervals ( CIs ) were calculated . We used an estimated ‘gold’ standard that considered any positive sample for each soil-transmitted helminth infection detected by any method at any time of examination . Accuracy and repeatability of the method was calculated throughout the 31-day period of stool preservation .
The mean age of the 41 children was 11 years ( range: 8–14 years ) ; one third ( n = 14 ) were girls . At D1 , the number of positive samples according to Kato-Katz and the initial Mini-FLOTAC reading was 41/41 ( 100% ) for T . trichiura , 36/41 ( 88% ) for A . lumbricoides , and 22/41 ( 54% ) for hookworm . The prevalence and intensity of infections according to our estimated ‘gold’ standard ( a single Kato-Katz thick smear plus multiple Mini-FLOTAC ) are shown in Tables 1 and 2 . Kato-Katz was more sensitive than Mini-FLOTAC for the diagnosis of A . lumbricoides ( 88% vs . 78% ) at D1 examination . Conversely , Mini-FLOTAC showed higher sensitivity for T . trichiura ( 100% vs . 98% ) and hookworm diagnosis ( 66% vs . 53% ) than Kato-Katz . Table 3 shows the sensitivity of Mini-FLOTAC for detection of species-specific soil-transmitted helminth diagnosis over the 31 days of stool preservation . The sensitivity was statistically different only for hookworm for all detections except for D3 and D5 compared to our estimated ‘gold’ standard . We found a substantial agreement between the two methods for the diagnosis of A . lumbricoides ( κ = 0 . 78 ) and hookworm ( κ = 0 . 65 ) , and a perfect agreement for the diagnosis of T . trichiura ( κ = 1 . 0 ) . The accordance between EPG was substantial for A . lumbricoides ( κ = 0 . 81 ) and hookworm ( κ = 0 . 73 ) , and moderate for T . trichiura ( κ = 0 . 45 ) . The arithmetic mean fecal egg counts using Kato-Katz was 12 , 075 EPG for A . lumbricoides , 1 , 074 EPG for T . trichiura , and 255 EPG for hookworm . Mini-FLOTAC revealed arithmetic mean fecal egg counts of 11 , 679 EPG for A . lumbricoides , 1 , 592 EPG for T . trichiura , and 220 EPG for hookworm . There was no statistically significant difference between the mean fecal egg counts of any soil-transmitted helminth infection detected by either Kato-Katz or Mini-FLOTAC at D1 . Similar proportions of light , moderate , and heavy infections were detected by the two techniques for A . lumbricoides ( light n = 15 using Kato-Katz , and n = 9 using Mini-FLOTAC; moderate , n = 20 using either Kato-Katz or Mini-FLOTAC; heavy , n = 0 using Kato-Katz and n = 2 using Mini-FLOTAC ) , T . trichiura ( light n = 23 using Kato-Katz , and n = 22 using Mini-FLOTAC; moderate n = 17 using Kato-Katz , and n = 19 using Mini-FLOTAC; heavy n = 0 using either Kato-Katz or Mini-FLOTAC ) , and hookworm ( light n = 17 using Kato-Katz , and n = 21 using Mini-FLOTAC; moderate and heavy n = 0 using either Kato-Katz or Mini-FLOTAC ) . The mean prevalence from D1 until D31 was 77 . 4% for A . lumbricoides , 99 . 8% for T . trichiura , and 53 . 5% for hookworm . The trend of infection prevalence is shown in Fig 1 . The agreement among the soil-transmitted helminth species-specific infection prevalence estimates over time was stable for the three soil-transmitted helminths . For A . lumbricoides , Pearson index was above 0 . 6 for all detections up to D19 and then 0 . 5 from D23 to D31 . For T . trichiura the Pearson index was 1 throughout the study , while it was above 0 . 7 for hookworm , apart from detection at D7 that was 0 . 6 compared with D3 . The mean intensity of fecal egg counts from D1 until D31 was 10 , 582 EPG for A . lumbricoides , 1 , 448 EPG for T . trichiura , and 144 EPG for hookworm . The mean intensity of infection according to samples analyzed from D1 to D31 is shown in Fig 2 . The calculated agreement by Pearson index between fecal egg counts was almost always above 0 . 9 apart from a couple of detections ( 0 . 8 between D1 and D31 and between D3 and D5 ) for A . lumbricoides , above 0 . 8 until D13 and 0 . 7 from D15 for T . trichiura , and above 0 . 8 for hookworm apart from two detections ( 0 . 7 between D1 and D31 and between D5 and D19 ) . The major change in the estimated prevalence was associated with low intensity infections , as shown in Fig 3 . From D1 until D31 Mini-FLOTAC detections for hookworm were statistically different compared to our estimated ‘gold’ standard , except for D3 and D5 , but not significantly different among each other . The Student’s t test showed no significant difference among the mean fecal egg counts of any soil-transmitted helminth infection from D1 until D31 detected with Mini-FLOTAC technique . For T . trichiura , the observed prevalence remained constantly 100% until the last detection ( 98% ) , for A . lumbricoides it was always above 77% , and for hookworm ranged between 63% and 84% ( Table 2 ) . Over the course of our study , the shape and unique identification of hookworm eggs gradually deteriorated , and hence , hookworm eggs became progressively less readable . Meanwhile , eggs of A . lumbricoides and T . trichiura were clearly visible and readable at the same ease throughout the study . Sequential microphotographs are presented in Fig 4 . The external membrane of the hookworm eggs started to fade and be less recognizable from D11 onwards , hence 12 days after stool production and fixation in 5% formalin . It became increasingly difficult to read hookworm eggs until the end of the study ( D31 ) when the eggs were very hard to recognize . Progressive degradation was not observed for the other soil-transmitted helminth eggs , as shown in Fig 4 . The A . lumbricoides eggs sometime became decorticated and T . trichiura eggs showed larvae inside , but for both these two species the eggs were still perfectly recognizable at the final observation time point at D31 .
The possibility to fix and examine fecal samples after days and weeks from collection is an important feature to facilitate the integration of soil-transmitted helminthiasis surveys with other NTDs , such as lymphatic filariasis and trachoma [13 , 14 , 16 , 28 , 29] . Only few studies , however , have been conducted on preserved stool samples [16 , 30–32] , especially on the effect of preservation of samples over time either on eggs detection or fecal egg counts . Therefore the time limit , beyond which qualitative and quantitative diagnosis for soil-transmitted helminth infections becomes unreliable , needs to be examined . In the current study , we assessed the effect of 5% formalin ( ratio 1:1 ) preservation on helminth eggs diagnostic accuracy over a 31-day time period using Mini-FLOTAC . The 1:1 dilution has been used before in studies on comparison among different copromicroscopic techniques [17 , 32 , 33] , but never as a preservative concentration; usually the storage was carried out with 10% or 5% formalin at 1:4 dilution . The performance of Mini-FLOTAC was also compared with Kato-Katz , using fresh stool samples . Of note , Kato-Katz is the WHO-recommended diagnostic method for soil-transmitted helminth infection ( and intestinal schistosomiasis ) [26] and is indeed widely used [11] . Our study provides new insight into the timing of preservation for an accurate estimate of prevalence and intensity of soil-transmitted helminth infections over the course of stool preservation . Importantly , prevalence estimate for each of the three soil-transmitted helminth species did not change significantly over the 31-day observation period . For T . trichiura , the observed prevalence remained 100% throughout . With regard to A . lumbricoides , the prevalence did not vary greatly over the examination period . The greatest variation was noted for hookworm: over the first 8 days the steepest drop of prevalence ( from 66% to 51% ) was observed , which then remained stable at 51% until the final day of analysis ( D31 ) . The Pearson index of correlation resulted to be stable throughout the study , which suggests a good effect on preservation and accordance among detections , and only slight changes in the prevalence and intensity for each observation during the 31 days of the study . The major change in prevalence was linked to low-intensity infections . T . trichiura and A . lumbricoides were mainly of light and moderate infection intensities and only two children had heavy A . lumbricoides infection . With regard to hookworm , all infections were of light intensity . Furthermore , since many children did have very low infections approaching 10 or lower EPGs , being so close to Mini-FLOTAC detection limit , it considerably affected the variation of prevalence estimates . The mean fecal egg counts of hookworm infections remained stable until 2 weeks from collection , whilst for T . trichiura and A . lumbricoides it did not change throughout the 31-day observation period . As shown in Fig 3 the classes of intensity of infection were consistent throughout the study for all the children . Although the main aim of this study was not to compare the performance of Mini-FLOTAC and Kato-Katz technique , which had been previously assessed in a larger studies [32 , 33] , we tested fresh samples with both techniques at baseline to validate the prevalence data and to confirm whether the accuracy of the two techniques was comparable . In fact , we noted that Mini-FLOTAC technique resulted to be as sensitive as a single Kato-Katz thick smear with no statistical differences among detections for the three soil-transmitted helminth species . There was a variation in sensitivity for hookworm among Mini-FLOTAC detections and this could be explained by the aforementioned limit of the current study , as for hookworm many infections were light and close to the sensitivity threshold . It is to be noted that the flotation solution used in this study is not the most suitable to detect A . lumbricoides . As reported from other studies [32] the most appropriate flotation solution to for the diagnosis of A . lumbricoides is zinc sulphate; but even if the sensitivity of the latter was higher compared to the flotation solution no . 2 , the mean fecal egg counts were lower [33] . Moreover , the zinc sulphate solution is more expensive and less easy to supply in low-resource setting . In conclusion , the findings of our study suggest that for a qualitative diagnosis with Mini-FLOTAC , stool samples fixed with 5% formalin can be preserved at least one month without impairing the quality of the data on prevalence of soil-transmitted helminth infections . However , for an accurate quantitative diagnosis for hookworm , we suggest a maximum of 15 days of preservation; after this time , hookworm eggs start to deteriorate and the consistence of microscope reading decreases ( unless the reader places additional attention to detect hookworm eggs ) , and the fecal egg counts progressively decline . As for A . lumbricoides and T . trichiura , eggs remain stable over one month and therefore a longer preservation might still give accurate data on intensity of infections . Further studies are needed to explore the performance of stool preserved with formalin at different concentrations and dilutions , or with other preservatives , and possibly these studies should be carried out in areas where hookworm infections are moderate and/or heavy . Additionally , studies should determine the effect longer fixation periods ( perhaps up to 2 or 3 months of stool preservation in formalin ) to evaluate the durability of A . lumbricoides and T . trichiura eggs . Data from this study are pivotal for the use of Mini-FLOTAC as an alternative to Kato-Katz , which would allow the integration of soil-transmitted helminthiasis into TAS surveys , and hence , integrated monitoring and evaluation of lymphatic filariasis with soil-transmitted helminthiasis , as recommended recently by WHO [34] . | Soil-transmitted helminths are parasitic worms ( hookworm , roundworm , and whipworm ) that affect hundreds of millions of people . Kato-Katz is the most widely used technique for the diagnosis of soil-transmitted helminth infection . It requires the collection , processing , and microscopic examination of stool samples within the same day . In remote areas , laboratories are often far away from where stool samples are being collected , which makes it difficult to examine the samples on the same day . Mini-FLOTAC is an alternative to Kato-Katz technique that allows analysis of fixed stool samples several days after collection . We assessed the accuracy of the Mini-FLOTAC with fixed stool samples . The study was carried out in late 2013 on Pemba Island and 41 children participated . Fresh stool samples were first examined by Kato-Katz and then fixed in formalin and examined by Mini-FLOTAC ( 12 examinations within one month ) . We found that for a qualitative diagnosis ( presence or absence of infection ) stool samples can be fixed for 31 days . However , for an accurate quantitative diagnosis of hookworm eggs , stool samples should not be stored for more than 15 days , as egg counts decline . Our results have implications for integrating soil-transmitted helminth surveys with other neglected tropical diseases . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | How Long Can Stool Samples Be Fixed for an Accurate Diagnosis of Soil-Transmitted Helminth Infection Using Mini-FLOTAC? |
Adipose tissue grows by two mechanisms: hyperplasia ( cell number increase ) and hypertrophy ( cell size increase ) . Genetics and diet affect the relative contributions of these two mechanisms to the growth of adipose tissue in obesity . In this study , the size distributions of epididymal adipose cells from two mouse strains , obesity-resistant FVB/N and obesity-prone C57BL/6 , were measured after 2 , 4 , and 12 weeks under regular and high-fat feeding conditions . The total cell number in the epididymal fat pad was estimated from the fat pad mass and the normalized cell-size distribution . The cell number and volume-weighted mean cell size increase as a function of fat pad mass . To address adipose tissue growth precisely , we developed a mathematical model describing the evolution of the adipose cell-size distributions as a function of the increasing fat pad mass , instead of the increasing chronological time . Our model describes the recruitment of new adipose cells and their subsequent development in different strains , and with different diet regimens , with common mechanisms , but with diet- and genetics-dependent model parameters . Compared to the FVB/N strain , the C57BL/6 strain has greater recruitment of small adipose cells . Hyperplasia is enhanced by high-fat diet in a strain-dependent way , suggesting a synergistic interaction between genetics and diet . Moreover , high-fat feeding increases the rate of adipose cell size growth , independent of strain , reflecting the increase in calories requiring storage . Additionally , high-fat diet leads to a dramatic spreading of the size distribution of adipose cells in both strains; this implies an increase in size fluctuations of adipose cells through lipid turnover .
Obesity is an enlargement of adipose tissue to store excess energy intake . Hyperplasia ( cell number increase ) and hypertrophy ( cell size increase ) are two possible growth mechanisms . Adipose tissue obesity phenotypes are influenced by diet and genetics , as well as by their interaction [1]–[4] . Starting from Johnson and Hirsch's studies [5] , there is an extensive literature on adipose tissue growth in normal and abnormal development , characterizing the state of the tissue in terms of the mean cell size and cell number . Hyperplastic growth appears only at early stages in adipose tissue development [6] , [7] . Hypertrophy occurs prior to hyperplasia to meet the need for additional fat storage capacity in the progression of obesity [8] . However , it has proven difficult to understand how diet and genetics specifically affect hyperplasia and/or hypertrophy of adipose cells , because of limited longitudinal data about adipose tissue growth . Beyond these studies , it has recently become possible to measure cell-size distributions precisely . This detailed information , compared with the mean cell size and total cell number , can be used to compute many size-related quantities that permit a finer characterization of the adipose tissue growth process . Cumulants of the cell-size distribution can be used to compute other physiological quantities such as the volume-weighted mean cell size . The cell-size distribution can be used to estimate total cell number within a fat pad from its mass . Furthermore , it is believed that some specific metabolic properties , e . g . , insulin resistance [9] and adipokine secretion [10] , depend on the precise cell-size distribution rather than the mean cell size . Indeed , several studies have addressed the change of the size distribution of adipose cells under various conditions in chick embryo development [11] , lean and obese Zucker rats [12] , [13] , partially lipectomized Wistar rats [14] , rabbit biopsy [15] , and human adipose tissue [16] , [17] . These studies focused only on the static differences between cell-size distributions under different conditions . However , cross-sectional static cell-size distributions for a range of snapshots of animal development can be used to deduce the dynamics of adipose tissue growth , if we can appropriately analyze the snapshots with the help of mathematical modeling . Given present technical limitations , this may be the best available approach to a microscopic and longitudinal understanding of in vivo adipose tissue growth , although a recent experiment has obtained microscopic observations of lipid accumulation in lipid droplets of adipose cells [18] . To address genetic and dietary effects on the dynamic process of adipose tissue growth , we obtained cell-size distributions of epididymal fat of obesity-resistant FVB/N ( hereafter FVB ) and obesity-prone C57BL/6 ( C57 ) mouse strains under standard chow and high-fat diets . The C57 mouse is the best characterized model of diet-induced obesity [19] , and the FVB mouse is a preferable model for generating transgenic mice [20] . These two commonly-used inbred mouse strains are genetically quite distant [21] , [22] , and they have distinct metabolic phenotypes: Compared with FVB mice , C57 mice have low circulating triglyceride levels [21] and increased triglyceride clearance [23] , [24]; FVB mice are characterized by relatively higher hepatic insulin resistance , counter-regulatory response to hypoglycemia , and reduced glucose-stimulated insulin secretion [25]; FVB mice are also known to be spontaneously hyperactive [26] and relatively lean since they appear to be less responsive to high-fat diet than C57 mice [27] . However , the development of diet-induced obesity in these two strains has not been formally compared . In this study , we developed a mathematical model predicting the change of the cell-size distribution as a function of the epididymal fat pad mass to analyze quantitatively the dynamic characteristics that depend on genetics and/or diet . The model of adipose tissue growth describes how many new cells are formed , how each cell grows depending on its size , and how lipid turnover leads to size fluctuations that cause a spreading in the cell-size distribution . As the epididymal fat pad mass increases , the cell-size distribution changes in a systematic manner depending on both genetics and diet . Comparing experimental results with the theoretical growth model , we found that hypertrophy is strongly correlated with diet . Hyperplasia , on the other hand , is dependent on genetics . Diet-induced changes in hyperplasia are strain-dependent , suggesting an interaction between diet and genetics .
At the beginning of the experiment ( 5 weeks of age ) , C57 mice were significantly lighter than FVB mice ( Fig . 1A ) due to a difference in lean mass , although total fat mass was not different ( Fig . 1B ) . When mice were maintained on regular chow diet , the difference in body weight disappeared by the age of 11 weeks ( week 6 of experiment , Fig . 1A ) . Under regular diet conditions , FVB and C57 mice maintained comparable fat mass throughout the whole course of the experiment ( Fig . 1B ) . High-fat diet caused significant increase in body weight and fat mass in both strains; however , changes in body weight and fat mass were more dramatic in C57 mice . The C57 mice had twice as much fat after 12 weeks of high-fat feeding ( Fig . 1B ) . The overall difference in total fat mass between FVB and C57 mice correlated with proportional differences in the amounts of epididymal ( intra-abdominal ) , inguinal ( subcutaneous ) , and brown fat ( Table 1 ) . Caloric intake and activity were comparable in FVB HF and C57 HF mice; however , FVB HF mice had higher resting and total oxygen consumption , and higher rectal temperature , suggesting that increased energy expenditure rather than reduced caloric intake was the reason for relative resistance to high-fat diet-induced obesity in the FVB mice . Interestingly , during the first 2 weeks of high-fat feeding , FVB and C57 mice showed comparable increase in total fat mass ( Fig . 1B ) . C57 HF mice continued to increase fat mass rapidly until week 10 of the experiment , whereas FVB HF mice slowed down accumulation of fat around week 3 . In C57 mice , high-fat feeding caused a gradual increase of both epididymal and inguinal fat pads; in contrast , in FVB mice , epididymal fat mass increased only slightly after 4 weeks on high-fat feeding , while inguinal fat pad continued to increase in size throughout the course of experiment ( Fig . S1 ) . High-fat feeding caused significant increase in blood glucose and insulin levels in both FVB and C57 mice ( Table 1 ) . Insulin levels and glucose intolerance were higher in C57 HF mice than in FVB HF mice , suggesting more severe insulin resistance ( Fig . 2A ) . Consistent with previous reports [23] , [24] , C57 REG mice showed reduced serum triglyceride levels , compared with FVB REG mice with no difference in FFA ( Table 1 ) . This was not due to higher fat utilization , since respiratory exchange ratio ( Table 1 ) and the rates of fatty acid oxidation measured in vivo ( Fig . 2B ) and in isolated skeletal muscle ( Fig . 2C ) were comparable in FVB REG and C57 REG mice . More likely , lower serum triglycerides in C57 REG mice were caused by much more efficient clearance of circulating triglycerides as suggested by triglyceride clearance test ( Fig . 2D ) . High-fat feeding reduced circulating triglyceride levels in both FVB and C57 mice and improved triglyceride clearance in the latter strain ( Table 1 and Fig . 2D ) . Both strains showed comparable reduction of respiratory exchange ratio , suggesting comparable increase of fatty acid utilization under high-fat diet condition ( Table 1 ) . Taken together , these data suggest that the ability to efficiently clear triglyceride from circulation may contribute to the high capacity of fat accumulation in C57 mice . To test the underlying mechanism of different rate of fat accumulation in epididymal fat of FVB and C57 mice , we measured mass and cell-size distribution in tissue samples of epididymal fat collected at 0 , 2 , 4 , and 12 weeks of controlled feeding ( Fig . 3 ) . Since histological analysis does not allow accurate determination of adipocyte cell size , which will be discussed later , we measured cell size distribution using a Coulter Counter and estimated volume-weighted mean cell size and total cell number of epididymal fat pad from these measurements , which had similar values in other mouse study [5] . Strong correlations were observed between fat pad mass and volume-weighted mean cell size , and between fat pad mass and total cell number , regardless of strain and diet difference ( Fig . 4 ) . The first correlation gave a scaling relation , , between fat pad mass , , and volume-weighted mean cell size , ( Figs . 4A and 4B ) . In addition , an exponential relation was found between fat pad mass , , and total cell number , : where the initial fat pad mass , , was obtained from control mice; and the initial cell number , , and the rate of increase of cell number in fat pad mass , , were estimated from data ( Figs . 4C and 4D; Table 2 ) . The initial cell number , , in C57 mice was larger than the initial cell number in FVB mice ( Table 2; Figs . S2C and S2D ) . As the fat pad mass increases , the total cell number increases . The rate of increase of cell number , , was larger under regular diet than under high-fat diet , a tendency more evident in C57 mice ( Figs . 4C and 4D; Table 2 ) , suggesting a genetic difference . The ratios of between the results of regular and high-fat diets are 1 . 42 and 3 . 22 for FVB and C57 mice , respectively ( Table 2 ) . This may indicate an interaction between genetics and diet on the increase of cell number . Note that we also observed similar results with body weight and fat mass , since the three quantities ( epididymal fat pad mass , fat mass , and body weight ) are correlated with each other . However , the results with epididymal fat pad mass were the best fits: The mean square deviation between data and fit in Figs . 4A–D was 9 . 73 , 7 . 94 , 3 . 56×105 , and 5 . 54×106 , respectively; the result with body weight was 11 . 82 , 15 . 35 , 3 . 65×105 , and 4 . 82×106; the result with fat mass was 8 . 58 , 8 . 61 , 4 . 28×105 , and 6 . 11×106 . These two strong correlations , between fat pad mass and hypertrophy , and between fat pad mass and hyperplasia , suggest that the increase in adipose tissue can be described as a systematic growth process with respect to fat pad mass increase . We arranged cell-size distributions sorted with respect to epididymal fat pad mass ( Fig . 5 ) . Remarkably , the adipose tissue growth model in Eq . ( 1 ) describes the evolving pattern of cell-size distributions with respect to the fat pad mass increase . The model fitted experimental cell-size distributions quantitatively , despite the fact that all distributions are cross-sectional data obtained from individual animals . The different parameter values in the model , which fit each individual cell-size distribution from both strains and both diet regimens , gave quantitative differences in the epididymal adipose tissue growth process between strains and between diets ( Table 2 ) . First , the maximal size-dependent growth rate , , and the rate of cell-size fluctuations due to lipid turn over , , demonstrated a diet-induced difference , and a smaller strain-induced difference . Size-dependent growth and size fluctuations , which resulted in hypertrophy , appear to be regulated mainly by diet . Specifically , the size-dependent growth moved the large cell mode of the cell-size distributions in Fig . 5 to larger sizes , and the lipid turnover fluctuations increased the spread of the distribution around the large cell mode . It is important to note that the results must be carefully interpreted because every rate is a rate per unit fat pad mass increase , not per unit time increase . Second , the geometrical parameters ( , and ) , which determine the shape of the size-dependent growth rates , had essentially the same values regardless of the diet and strain difference , except for . Therefore , the lower critical size , which gives the size initializing cell size-dependent growth , and two scale parameters could be fixed at sl = 37 µm , ηl = 12 µm , and ηu = 63 µm , respectively . On the other hand , the upper critical size limiting cell size-dependent growth of big cells depended on diet; under the high-fat feeding , this cutoff size shifted to a larger size ( Fig . 6 ) . Under high-fat diet , the changes of parameters ( , , and ) enlarge the lipid-storage capacity of fat tissues through both hyperplasia and hypertrophy . Lower serum triglycerides in the high-fat diet condition ( Table 2 and Fig . 2D ) may be correlated with the increasing lipid storage in enlarged fat cells because no significant difference in fatty acid oxidation was found as suggested by no difference in respiratory exchange ratio ( Table 1 ) .
Our central finding is that hyperplasia and hypertrophy of adipose cells in the epididymal fat pad is a function of the fat pad mass , even though it may take individual animals different time periods to reach a given fat pad mass . Therefore , adipose tissue growth , represented as changes of the cell-size distribution , can be systematically modeled as a growth process with respect to fat pad mass increase; this may reflect a correlation between fat pad mass and the secretion of adipokines and other signaling molecules controlling adipose tissue growth . Accordingly , it should be noted that the rates in our model are not the usual rates per unit time increase but the rates per unit mass increase . Thus , several rates ( , , and ) in the model had larger values for animals on a chow diet than for those on a high-fat diet . However , if these rates were converted to the usual rates per unit time increase , they had larger values for the high-fat diet , because it takes less time for a unit increase in the fat pad mass from larger , and more numerous , cells on a high-fat diet than for an increase of the same magnitude from smaller , and fewer , cells on a chow diet ( Fig . S1 ) . It has been suggested that when obesity progresses , hypertrophy of adipose cells occurs early , and then triggers hyperplasia [8] . Our study showed that new cell recruitment increases exponentially as fat pad mass increases . Hypertrophy of adipose cells is the main contributor to fat pad mass increase , whereas hyperplasia does not contribute much to this increase because it occurs in small cells that have a much smaller volume of fat stored . Therefore , our model naturally embodies the concept that hyperplasia is affected by the hypertrophic growth of cells . On the other hand , it has been reported that hyperplasia of adipose cells occurs only at early development stages; hence , no new cell recruitment would be expected at late stages even under obesogenic conditions [6] , [7] . It may be the case that the age of the animals in our study ( 6 weeks old ) allows the occurrence of hyperplasia . The model developed here may give microscopic insights into the size-dependent growth of adipose cells that cannot be addressed by static cross-sectional studies . For example , we found the following specific properties of size-dependent cell growth: the lower critical size , , initializing lipid accumulation , did not depend on diet in the two mouse strains , whereas the upper critical size , , limiting cell growth from reaching an extraordinary size , was elevated on a high-fat diet . This size-dependence of cell growth is a testable hypothesis . Next , the cell-size fluctuation parameter , , was different between regular and high-fat diets; it is larger under high-fat diet , when it is transformed to units appropriate for per unit time change instead of unit fat pad mass change . Thus , the random process for fat cells to release and take up fat occurs more actively under a high-fat diet than under a regular diet . It may be of interest to see if these results can be generalized to other strains and organisms . Compared with the studies observing a single peak in cell-size distributions of fat cells [11] , [15]–[17] , [28] , we have observed bimodal cell-size distributions as reported by others [12] , [13] , [29]–[32] . Most studies [12] , [13] , [31] , [32] observing the bimodality used the Coulter Counter technology which has several advantages to assess the entire distribution of cell sizes [31]: First , the analyzed cells can be proven to be authentic adipose cells based on morphology and flotation; second , the volume of each cell is assessed regardless of shape and free of the artifacts of off-center sectioning as is the rule rather than the exception using histological approaches; finally , sufficient numbers of particles can be counted and sized to provide statistically significant complex curves . In contrast , microscopic methods for histology may not observe small cells due to the influence of microscope magnification [30] , small sample number , and sampling bias . However , when the Coulter Counter is used , non-adipocyte contamination may contribute to the cell-size distribution especially at small sizes , although our minimal cell diameter , 22 µm , is above the possible contamination ranges , 10 to 20 µm , mentioned by Mersmann and MacNeil [31] . To be certain , we again analyzed the modified data using only cell-size distributions above 35 µm diameter with the model , and reached the same conclusions ( data not shown ) . The nadir in the cell-size distribution ( Fig . 3 ) may separate two cell populations . DeMartinis and Francendese defined the small cells , with diameter smaller than 35 µm , as “very small fat cells” [29] . Based on our model , these cells have negligible size-dependent growth , because their size is smaller than the lower critical size , sl = 37 µm . Therefore , the size-dependent growth mechanism can naturally explain the origin of bimodality in the cell-size distribution of fat cells . Cells with size only above can grow with the size-dependent manner , but cells with size below can randomly grow with the size-fluctuation through lipid turnover . This separation causes the cell accumulation below the size , , which gives the lower peak in cell-size distributions . This cell population may serve as a potential reservoir for mature adipose cells . Their maturation process may be interpreted as follows: The fat cells reaching the critical size , , by random size fluctuations , then , can grow with a size-dependent growth mechanism . As mentioned above , the size fluctuation occurs more actively under a high-fat diet; therefore , the reservoir can accelerate the maturation process under the stimulating condition . In the tissue growth model , we included the recruitment of new cells and the growth of existing cells , but not the death of old cells , because the model was consistent with the data without the apoptosis of adipose cells . This result is also consistent with a study observing that epididymal fat tissue of C57BL/6 mice does not show significant cell death before 12 weeks under a high-fat diet [33] . However , extended high-fat diet finally induces apoptosis of fat cells [33] . Furthermore , one recent study has reported that human fat cells turn over on a ten-year time scale [7] . Our model , therefore , still needs enhancement to be more generally applied in various conditions . Cell death should be considered and the diet dependence of the model parameters should be formally incorporated . In this study , we focused on one fat depot , epididymal fat , with several reasons: 1 ) the weight of epididymal fat pads can be accessed more accurately than the weight of inguinal fat pads due to the ease of dissection; 2 ) the morphology of adipose cells in epididymal fat is more homogeneous than in inguinal fat which contains a lot of brown adipose tissue-like cells , particularly in mice resistant to diet-induced obesity; and 3 ) the difference between the genotypes was more evident in growth of epididymal fat , which reaches a plateau at 4 weeks in FVB mice , but continues to grow in C57 mice throughout the course of experiment , in contrast to the inguinal fat shows sustained growth in both strains ( Fig . S1 ) . Although we have not measured the cell-size distribution of other fat depots , we measured the mass change of inguinal and brown fat depots , which shows similar pattern with epididymal fat depot ( Fig . S1 ) . Thus , it is of interest to apply the model to other fat depots that have functional differences [34] , [35] , and to other species such as human , which is left for future study . We expect the model can be applied to such diverse data sets simply by adjusting the model parameters , because the model contains general tissue growth mechanisms for the recruitment of new cells and their subsequent development . Our data suggest that at least three factors may explain why C57 mice gain more fat than FVB mice do under high-fat diet: First , FVB mice have increased metabolic rate and increased rectal temperature , most likely due to the increased sympathetic tone . Although we did not detect significant differences in activity between the strains in our study , more comprehensive behavior measurements suggested that FVB mice are spontaneously hyperactive , compared with C57 mice [26] . They also have increased heart rate [36] and respond with hyperglycemia to a variety of treatments [37] . In addition , they are more responsive to stress associated with restraint and fasting [38] . All these data taken together suggest that the activity of sympathetic nervous system increased more in FVB mice than the activity in C57 mice . Second , compared with the FVB mice , C57 mice clear circulating triglycerides more efficiently [23] , [24] , which at least in part could be attributed to higher serum lipase activity [23] and higher capacity to store triglycerides in the liver [23] , [24] and the adipose tissue ( shown in this study ) . Although molecular mechanisms of triglyceride clearance are not fully understood , adipose tissue clearly contributes to clear triglycerides because the ability to clear circulating triglycerides is impaired in lipoatrophic mice [23] . In particular , it has been reported that high-fat diet enhanced triglyceride clearance [39] , which may be related to the induction of lipoprotein lipase activity in the adipose tissue [40] . Finally , compared with FVB mice , C57 mice showed greater recruitment of small adipose cells , particularly under high-fat diet . It has been suggested that new adipose cells can arise from progenitor cells which reside within the adult white fat depots [41] , [42] , and from other sources such as bone marrow-derived circulating progenitor cells [43] . Recruitment of both types of progenitors has been shown to be stimulated by high-fat diet [41] , [43] . It is possible that greater recruitment of smaller fat cells in C57 mice might be caused by a higher pool of precursor cells or thier higher intrinsic capacity for adipocyte differentiation . However , in vitro mesenchymal stem cell , isolated from the outer ear of FVB and C57 mice , differentiate into adipose cells equally well [44] . The attempt to compare differentiation of bone marrow stromal cell from FVB and C57 mice into adipose cells was not conclusive due to the very low yield and poor proliferative capacity of the cells isolated from C57 mice [45]; however , bone marrow does not appear to be the major source of new fat cells at least in mice [43] . Our model suggests the difference between genotypes in the recruitment of small adipose cells might be fat pad autonomous , but the molecular mechanism underlying this difference is unclear . Fat pad is a complex organ containing a variety of different cell types , including mature adipose , preadipose and vascular cells , nerves , macrophages , and fibroblasts . The number of adipocyte precursors and their proliferation in response to external signals varies between fat depots [35] . Further studies would be needed to determine how genotype specific interaction between different cell type and secreted factors may affect the rate of adipocyte recruitment to the specific fat depots . In summary , we have derived a mathematical model describing the growth of adipose tissue with cell-number and cell-size increases as a function of epididymal fat pad mass . Based on this dynamic model , we examined the effects of genetics and diet on adipose tissue growth . Comparing the cell-size distributions from two strains and two diets , we concluded that cell size change depends on diet , and cell number change depends on genetics and diet , as well as their interaction .
All procedures were approved by the Animal Care and Use Committee of the National Institute of Diabetes and Digestive and Kidney Diseases . Male FVB and C57 mice were obtained from The Jackson Laboratory ( Bar Harbor , ME ) . Mice were reared four per cage on a 12-h light/dark cycle ( lights on 06:00–18:00 ) . At the age of 5 weeks , mice of each strain were split into 2 groups . Half of the mice were fed regular chow NIH-07 diet ( hereafter REG; Zeigler Brothers , Inc . , Gardners , PA ) , containing 4 . 08 kcal/g ( 11% calories from fat , 62% from carbohydrates and 26% from protein ) . The other half was fed high-fat diet , F3282 ( hereafter HF; Bio-Serv , Frenchtown , NJ ) , containing 5 . 45 kcal/g ( 59% fat , 26% carbohydrate , and 15% protein ) . Water and diets were provided ad libitum . Five independent experiments were conducted , each using 4 groups of mice: FVB REG , FVB HF , C57 REG and C57 HF . In three experiments , mice were maintained on controlled diets for 12 weeks and used for body composition analysis , physiological characterization , and cell size distribution . Two additional sets of mice were euthanized after 2 weeks and 4 weeks of high-fat and control feeding for cell-size distribution only . Body composition , food intake , metabolic rate , glucose tolerance , triglyceride clearance , and fatty acid oxidation in isolated soleus muscle were measured as described previously [46] . Whole body fatty acid oxidation was measured as described in Gautam et al . [47] . Blood for biochemical assays was obtained from the tail vein in the non-fasted state . Glucose levels were measured using Glucometer Elite ( Bayer , Elkhart , IN ) . Serum insulin was assayed using radioimmunoassay ( Linco Research , St . Charles , MO ) . Serum triglycerides , cholesterol ( Thermo DMA , Louisville , CO ) and free fatty acid ( FFA ) ( Roche Applied Science , Indianapolis , IN ) were measured according to the manufacturers procedures . Cell-size distribution in epididymal fat was measured after 2 , 4 , and 12 weeks of high-fat and control feeding using Beckman Coulter Multisizer III as previously described [9] . Briefly , 20–30 mg of fat tissue were sampled from the midsection , by dissection and then removing the sample for fixation from the center of the cut epididymal fat . Tissue samples were immediately fixed in osmium tetroxide [48] , incubated in a water bath at 37°C for 48 h , and then adipose cell size was determined by a Beckman Coulter Multisizer III with a 400 µm aperture . The range of cell sizes that can effectively be measured using this aperture is 20–240 µm . The instrument was set to count 6 , 000 particles , and the fixed-cell suspension was diluted so that coincident counting was <10% . After collection of pulse sizes , the data were expressed as particle diameters and displayed as histograms of counts against diameter using linear bins and a linear scale for the x-axis ( Fig . 3 ) . Cell-size distribution was measured in four samples from each group , except for the C57 mice after 4-week high-fat diet exposure , which had only three available samples . A sample was taken from each fat pad and processed separately . Each sample was then counted at least twice . The curves from the two samples are then averaged , but only after examining the reproducibility between the two samples . The cell-size distribution includes all the information related to cell sizes in a tissue and its changes give a statistical view of the detailed growth process of each cell . To examine adipose tissue growth in terms of underlying microscopic processes , we consider a mathematical model quantifying the processes that change the cell-size distribution . The model can predict how many new cells are formed and how cells with different sizes grow as fat pad mass increases . The cell-number density of a certain size ( diameter ) at a given fat pad mass is the specific quantity to be considered . We consider how this cell-size distribution changes with an incremental change in fat pad mass . The evolution of the cell-size distribution with fat pad mass can be modeled by a partial differential equation , ( 1 ) This equation comprises three general components of the adipose tissue growth process . First , we assume that new cell recruitment occurs only at the minimal cell size observed , which is mathematically expressed as the delta function . The recruitment rate with respect to fat pad mass is given by the exponential function , ( 2 ) where is the initial total cell number at a given initial fat pad mass , and is the rate of increase in cell number per unit change in fat pad mass . The change of total cell number is the recruitment rate of new cells if cell death is negligible; we found no need to include apoptosis at any cell size to fit these experimental data . Therefore , this recruitment rate can be directly obtained from the experimental result using the relation between total cell number and fat pad mass by differentiating the function with respect to . Second , there is cell-size dependent cell growth . After maturation of adipose cells to a specific size , they may be able to accumulate fat , causing hypertrophy . In addition to this limiting growth rate of small adipose cells , there may be also an upper growth limit because large adipose cells cannot grow indefinitely by this growth process , though they may attain larger sizes by size fluctuations caused by lipid turnover . The rise and fall of cell-growth rate depending on cell size can be described with the general functional form multiplying two sigmoidal functions , ( 3 ) where represents the maximal growth rate; and are the lower and upper critical sizes , respectively , which give the half-maximal growth rate; and give their scale ( Fig . 6 ) . Finally , the last term in Eq . ( 1 ) represents cell-size fluctuations with the constant rate , , which reflect lipid turnover randomly occurring in adipose cells . This lipid turnover is the only growth mechanism for large cells above the upper critical size . Generally , the size-dependent growth of cells moves their size distribution to larger sizes , while size fluctuations spread the size distribution . The volume-weighted mean cell size was used as a measure of hypertrophy , ( 4 ) where is the cell size of the bin and is the relative frequency corresponding to the size at a given fat pad mass . This measure is meaningful from the functional point of view that the lipid storage capacity is proportional to cell volume . Note that the volume-weighted mean cell size gives an average cell size weighted by the lipid-storage capacity of large cells in the upper peak of the bimodal cell-size distribution ( Fig . 3 ) . In contrast , the usual number-weighted mean cell size , , may give a smaller average cell size due to the considerable contribution of small cells in the lower peak of the bimodal cell-size distribution . Clearly , the volume-weighted mean cell size is a better index of lipid-storage capacity . Total cell number is a direct measure reflecting hyperplasia . In this study , epididymal fat pad mass as well as cell-size distributions were measured . Therefore , the total cell number in the epididymal fat pad could be estimated from the relation between fat mass and volume . Here we used pure trioleine density ρ = 0 . 915 g/ml as the density of adipose cells since the density is similar to the actual fat pad density [49] . The total fat volume is where is the total cell number and is the average cell volume , which could be calculated from the cell-size distribution . Therefore the total cell number , or hyperplasia index , is ( 5 ) To optimize model parameters so that they can closely describe the evolution of cell-size distribution in experiment , we used the minimization of a “cost function” which quantifies the deviation between the model and experimental results . To define the cost function , the normalized cell-size distribution at a given fat mass was compared with simulation data with a parameter set : ( 6 ) where is the total number of cell size bins and is the total number of given fat mass . The scale of the cost function was calculated from the intrinsic fluctuation of experimental data , which can be defined as the squared deviation between the measured cell-size distribution and a smooth fitting function : ( 7 ) This intrinsic fluctuation is numerically about 10 percent of the squared deviation between experimental and model data . As the fitting function in Fig . 3 , we used a sum of two exponentials and a Gaussian , ( 8 ) a form that has been used to fit adipose cell-size distributions [9] . These parameter fits were performed using the nonlinear curve-fitting routine in MATLAB R2007a ( Natick , MA , USA ) . For the optimization process , we specifically used the parallel tempering Monte-Carlo method to find the global minimum of the cost function [50] . We used 10 uniformly spaced values ( 0 . 1 to 1 ) for the tempering parameter and ran ten chains in parallel with the updating probability . At every 20 steps , a pair of adjacent simulations on ten tempering parameters were randomly chosen and their parameter states were exchanged with probability . After equilibrium , 20 , 000 iterations were used with the fixed tempering parameter to estimate the optimal parameter values and their standard errors . We also used this method to estimate the initial total cell number , , and its rate of increase , , from the relation between fat mass and total cell number ( Figs . 4C and 4D ) . In the minimization between the fitting function and experimental data , we used a constraint that the initial cell number is equal regardless of diet conditions , i . e . , regular and high-fat diets . The average fat mass of four control mice before regular and high-fat diets was used as the initial fat mass , , which is 0 . 34 g and 0 . 29 g for FVB and C57 mice , respectively . We estimated the uncertainties , and , by propagating the 10 percent statistical fluctuations observed in the experimental data . We solved the following discrete version of our model , given as a continuous partial differential equation in Eq . ( 1 ) : ( 9 ) with mass interval δm = 0 . 1 mg and size interval δs = 0 . 73 µm . | Obesity is an enlargement of adipose tissue to store excess energy intake . Hyperplasia ( cell number increase ) and hypertrophy ( cell size increase ) are two possible growth mechanisms . The in vivo dynamic change of fat tissue cannot be monitored in real time due to current technical limitations . However , we can measure cell-size distributions of fat cells in individual animals . Our fundamental goal is to extract dynamic features of tissue remodeling from snapshots of cell-size distributions . We develop a mathematical model that interpolates between the cell-size distribution measurements and predicts the continuous change of the cell-size distribution with respect to fat pad mass increase . Our adipose tissue growth model includes three essential components: new cell recruitment , size-dependent cell growth , and cell-size fluctuations . In particular , we compared the adipose tissue growth of obesity-prone and obesity-resistant mice under a standard or a high-fat diet to examine the genetic and diet effect on adipose tissue growth . By applying our model to these different conditions , we found that the size increase of fat cells is dependent on diet . On the other hand , the diet-induced number increase of fat cells is dependent on strain , suggesting a synergy between genetics and diet . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/cell",
"growth",
"and",
"division",
"genetics",
"and",
"genomics/complex",
"traits",
"biophysics/theory",
"and",
"simulation",
"developmental",
"biology/cell",
"differentiation",
"diabetes",
"and",
"endocrinology/obesity",
"nutrition/obesity",
"physiology/integ... | 2009 | Hypertrophy and/or Hyperplasia: Dynamics of Adipose Tissue Growth |
The RIG-I like receptor pathway is stimulated during RNA virus infection by interaction between cytosolic RIG-I and viral RNA structures that contain short hairpin dsRNA and 5′ triphosphate ( 5′ppp ) terminal structure . In the present study , an RNA agonist of RIG-I was synthesized in vitro and shown to stimulate RIG-I-dependent antiviral responses at concentrations in the picomolar range . In human lung epithelial A549 cells , 5′pppRNA specifically stimulated multiple parameters of the innate antiviral response , including IRF3 , IRF7 and STAT1 activation , and induction of inflammatory and interferon stimulated genes - hallmarks of a fully functional antiviral response . Evaluation of the magnitude and duration of gene expression by transcriptional profiling identified a robust , sustained and diversified antiviral and inflammatory response characterized by enhanced pathogen recognition and interferon ( IFN ) signaling . Bioinformatics analysis further identified a transcriptional signature uniquely induced by 5′pppRNA , and not by IFNα-2b , that included a constellation of IRF7 and NF-kB target genes capable of mobilizing multiple arms of the innate and adaptive immune response . Treatment of primary PBMCs or lung epithelial A549 cells with 5′pppRNA provided significant protection against a spectrum of RNA and DNA viruses . In C57Bl/6 mice , intravenous administration of 5′pppRNA protected animals from a lethal challenge with H1N1 Influenza , reduced virus titers in mouse lungs and protected animals from virus-induced pneumonia . Strikingly , the RIG-I-specific transcriptional response afforded partial protection from influenza challenge , even in the absence of type I interferon signaling . This systems approach provides transcriptional , biochemical , and in vivo analysis of the antiviral efficacy of 5′pppRNA and highlights the therapeutic potential associated with the use of RIG-I agonists as broad spectrum antiviral agents .
The innate immune system has evolved numerous molecular sensors and signaling pathways to detect , contain and clear viral infections [1]–[4] . Viruses are sensed by a subset of pattern recognition receptors ( PRRs ) that recognize evolutionarily conserved structures known as pathogen-associated molecular patterns ( PAMPs ) . Classically , viral nucleic acids are the predominant PAMPs detected by these receptors during infection . These sensing steps contribute to the activation of signaling cascades that culminate in the early production of antiviral effector molecules , cytokines and chemokines responsible for the inhibition of viral replication and the induction of adaptive immune responses [2] , [5]–[7] . In addition to the nucleic acid sensing by a subset of endosome-associated Toll-like receptors ( TLR ) , viral RNA structures within the cytoplasm are recognized by members of the retinoic acid-inducible gene-I ( RIG-I ) -like receptors ( RLRs ) family , including the three DExD/H box RNA helicases RIG-I , Mda5 and LGP-2 [2] , [3] , [5] , [8]–[12] . RIG-I is a cytosolic multidomain protein that detects viral RNA through its helicase domain [13] , [14] . In addition to its RNA sensing domain , RIG-I also possesses an effector caspase activation and recruitment domain ( CARD ) that interacts with the mitochondrial adaptor MAVS , also known as VISA , IPS-1 , Cardif [15] , [16] . Viral RNA binding alters RIG-I conformation from an auto-inhibitory state to an open conformation exposing the CARD domain , resulting in the generation of an activated state characterized by ATP hydrolysis and ATP-driven translocation on RNA [17]–[19] . Activation of RIG-I also allows ubiquitination and/or binding to polyubiquitin . In recent studies , polyubiquitin binding has been shown to induce formation of RIG-I tetramers that activate downstream signaling by inducing the formation of prion-like fibrils composed of the MAVS adaptor [20] . MAVS then triggers the activation of IRF3 and NF-κB transcription proteins through the IKK-related serine kinases TBK1 and IKKε [10] , [21]–[23] , leading to the primary activation of the antiviral program , involving production of type I interferons ( IFNβ and IFNα ) , as well as pro-inflammatory cytokines and antiviral factors [5] , [24] , [25] . A secondary response involving the induction of IFN stimulated genes ( ISGs ) is induced by the binding of IFN to its cognate receptor ( IFNα/βR ) , which triggers the JAK-STAT pathway to amplify the antiviral immune response [6] , [26]–[29] . The nature of the ligand recognized by RIG-I has been the subject of intense study given that these PAMPs are the initial triggers of the antiviral immune response . In vitro synthesized RNA carrying an exposed 5′ terminal triphosphate ( 5′ppp ) moiety was first identified as RIG-I agonists [30]–[32] . The 5′ppp moiety is present at the end of viral and self RNA molecules generated by RNA polymerization; however , in eukaryotic cells , RNA processing in the nucleus cleaves the 5′ppp end and the RNA is capped prior to release into the cytoplasm . This mechanism distinguishes viral ‘non-self’ 5′pppRNA from cellular ‘self’ RNA , and renders it recognizable to the innate RIG-I sensor [30] , [31] , [33] . Further characterization of the RNA structure demonstrated that blunt base pairing at the 5′ end of the RNA , with a minimum double strand ( ds ) length of 20 nucleotides was also important for RIG-I signaling [17] , [33] , [34] . Furthermore , short dsRNA ( <300 bp ) triggered RIG-I , whereas long dsRNA ( >2000 bp ) such as poly I:C and lacking 5′ppp failed to trigger RIG-I , but was recognized by Mda5 [35] . Natural RNA extracted from virally infected cells , specifically the viral RNA genome or viral replicative intermediates , were also shown to activate RIG-I [30] , [31] , [36]–[38] . Interestingly , the highly conserved 5′ and 3′ untranslated regions ( UTRs ) of negative single strand RNA virus genomes display high base pair complementarity and the panhandle structure theoretically formed by the viral genome meets the requirements for RIG-I recognition [17] . The elucidation of the crystal structure of RIG-I highlighted the molecular interactions between RIG-I and 5′ppp dsRNA [18] , [39] , providing a structural basis for the conformational changes involved in exposing the CARD domain for effective downstream signaling [18] . Given the level of molecular understanding of the RIG-I ligand and subsequent signaling leading to induction of antiviral immune response , we sought to investigate the range of the protective innate immune response triggered by RIG-I agonists against viral infections . A short in vitro-synthesized 5′pppRNA derived from the 5′ and 3′ UTRs of the VSV genome activated the RIG-I signaling pathway and triggered a robust antiviral response that interfered with infection by several pathogenic viruses , including Dengue , HCV , H1N1 Influenza A/PR/8/34 and HIV-1 . Furthermore , intravenous delivery of the RNA agonist stimulated an antiviral state in vivo that protected mice from lethal influenza virus challenge . This report highlights the therapeutic potential of naturally derived RIG-I agonists as potent stimulators of the innate antiviral response , with the capacity to mobilize genes essential for the generation of efficient immunity against multiple infections .
A 5′ triphosphate containing RNA derived from the 5′ and 3′ UTRs of the negative-strand RNA virus Vesicular Stomatitis Virus ( VSV ) [17] was generated by in vitro transcription using T7 polymerase , an enzymatic reaction that synthesizes RNA molecules with a 5′ppp terminus [17] . Predicted panhandle secondary structure of the 5′pppRNA is depicted in Fig . 1A; gel analysis and nuclease sensitivity confirmed the generation of a single RNA product of the expected size ( 67 nucleotides ) . Transfection of increasing amounts of 5′pppRNA resulted in Ser396 phosphorylation of IRF3 at 8 h – a hallmark of immediate early activation of the antiviral response ( Fig . 1B , lane 2 to 6 ) . Induction of apoptosis was detected following treatment with higher concentrations of 5′pppRNA; the pro-apoptotic BH3-only protein NOXA – a direct transcriptional target of IRF3 [40] – as well as cleavage products of caspase 3 and PARP were up-regulated in a dose dependent manner ( Fig . 1B , lane 2–6 ) . Optimal induction of antiviral signaling with limited cytotoxicity was achieved at a concentration of 10 ng/ml ( ∼500 pM ) ( Fig . 1B; lane 4 ) . Importantly , the stimulation of immune signaling and apoptosis was dependent on the 5′ppp moiety; a homologous RNA without a 5′ppp terminus abrogated stimulation over a range of RNA concentrations ( Fig . 1B , lane 8–12 ) . To characterize the antiviral response triggered by 5′pppRNA , the kinetics of downstream RIG-I signaling were measured at different times ( 0–48 h ) after stimulation of A549 cells ( Fig . 1C ) . IRF3 homodimerization ( 1st panel ) and IRF3 phosphorylation at Ser396 ( 2nd panel ) were detected as early as 2 h post treatment with 5′pppRNA , and sustained until 24 h . Using a newly characterized anti-IRF7 antibody , induction of endogenous IRF7 was detected with kinetics that was delayed compared to IRF3 activation ( 4th vs . 3rd panel ) . IκBα phosphorylation was likewise detected as early as 2 h post-treatment and was sustained in A549 cells ( 6th panel ) . Altogether , IRF3 , IRF7 and NF-κB are required for optimal induction of the IFNβ promoter [27] . JAK-STAT signaling was detected at 4 h with Tyr701 phosphorylation of STAT1 ( 9th panel ) , as well as later accumulation at 24 h ( 10th panel ) . IFIT1 and RIG-I itself , were up-regulated 4 h post-treatment ( 11th and 12th panel ) whereas a second group of ISGs ( STAT1 and IRF7; 4th and 10th panel ) was induced between 6 h and 8 h after agonist treatment . IFNβ was detectable in cell culture supernatant as early as 6 h after 5′pppRNA treatment with a substantial release ( 4000 pg/ml ) that peaked at 12–24 h ( Fig . 1D , top panel ) . IFNα release was detected later at 12 h , and remained high thereafter ( 400 pg/ml ) ( Figure 1D ) . Thus 5′pppRNA triggers a full antiviral response as demonstrated by the activation of transcription factors IRF3 , IRF7 and NF-κB , release of interferons , JAK/STAT pathway activation and induction of ISGs . To address whether 5′pppRNA exclusively activated the RIG-I sensor , wild type mouse embryonic fibroblasts ( wt MEF ) and RIG-I−/− MEF were co-transfected with 5′pppRNA and type 1 IFN reporter constructs to measure promoter activity . 5′pppRNA activated the IFNβ promoter 60-fold and the IFNα promoter 450-fold in wt MEF; promoter activity was dependent on RIG-I since these promoters were not stimulated in RIG-I−/− MEF . As positive control , a constitutively active , CARD domain-containing RIG-I mutant [41] was used to bypass the requirement for RIG-I ( Figure 2A ) . Furthermore , induction of the IFN response was exclusively dependent on intact RIG-I signaling , since IFNβ promoter activity was not decreased in Mda5−/− , TLR3−/− or TLR7−/− MEFs ( Figure 2B ) . In A549 cells treated with 5′pppRNA , knocking down RIG-I abolished IRF3 and STAT1 phosphorylation , as well as IFIT1 and RIG-I upregulation compared to control siRNA-treated cells ( Fig . 2C; lane 6 vs . 4 ) . Of note , generation of the knock down by transient transfection of short RNA did not activate immune signaling ( Fig . 2C; lane 3 vs . 1 ) . Hence , this 5′pppRNA signals specifically via RIG-I . To evaluate the breadth of the host intrinsic response resulting from RNA agonist stimulation of RIG-I , modulation of the transcriptome of A549 cells stimulated with 5′pppRNA from 1 to 48 h was analysed by gene array using the Illumina platform . Figure 3A shows a waterfall plot of differentially expressed genes ( DEG; selected based on fold change ≥±2 , p-value ≤0 . 001 ) after 5′pppRNA stimulation . The number of genes up-regulated by 5′pppRNA administration steadily increased with time , while the majority of down-regulation occurred at 24–48 h ( Fig . 3A ) . The heatmap presents DEG with emphasis on the most highly deregulated genes over time ( Fig . 3B ) . Canonical pathway analysis using Ingenuity Pathway Analysis software identified IFN signaling , activation of IRFs by cytosolic PRRs , TNFR2 signaling and antigen presentation as the main up-regulated functional categories , while functions related to cell metabolism and cell cycle were down-regulated by RIG-I agonist treatment ( Fig . 3C ) . Subsequent kinetic analysis revealed that RIG-I agonist induced distinct temporal patterns of gene expression ( Fig . 3D , S1 and S2A ) . For example , some genes were highly expressed early at 6–12 h , including IFNB1 and the IFNλ family ( IL29 , IL-28A , IL28B ) , but the expression of these genes was not sustained throughout the time course and decreased at 24–48 h ( Fig . 3D; lest panel ) . A second subset of genes associated with the antiviral response were induced early at 6–8 h , but expression was sustained and markedly augmented at 24–48 h , as exemplified by IFI family members , IRF7 , and other ISGs ( Fig . 3D; middle panel ) . A third subset of genes was induced primarily at later time points , as part of the secondary response to 5′pppRNA treatment , and included HLA and CCL3 families ( Fig . 3D; right panel ) . Representative genes from different subsets were validated by quantitative real-time RT-PCR ( Fig . S1 ) . Overall , 5′pppRNA induced a robust bi-phasic transcriptional response , characterized by strong activation of antiviral and inflammatory gene signatures; the kinetics of the transcriptional profile mirrored the biochemical activation events detected in Figures 1 . In order to gain systems-wide insight into the RIG-I transcriptome , a functional clustering of 5′pppRNA-induced DEGs was performed . This functional clustering identified a variety of transcriptional sub-networks and biological processes regulated by RIG-I ( Fig . 4 ) . As expected , at 6 h ( Fig . 4A ) , induction of antiviral and inflammatory response programs downstream of IRF , NF-κB , STAT signaling were identified ( Fig . 4A and S2B ) ; expression of several cytokine and chemokine genes were also up-regulated . Concomitantly , genes related to Fos and TGF-β signaling , as well as hypoxic signaling via HIF-1α were down-regulated . At 24 h ( Fig . 4B ) , genes associated with pathogen recognition receptor signaling , the ubiquitin pathway , inflammation and apoptosis were also induced by RIG-I activation; interestingly , profiling of down-regulated genes identified functional clusters involved in cell cycle regulation , MYC signaling , and the heat shock response . Although the contribution of type I IFN to the antiviral response stimulated by 5′pppRNA is unquestionable , other factors may also augment the antiviral state established by 5′pppRNA . To define gene uniquely induced by 5′pppRNA , gene expression profiles of A549 cells stimulated with 5′pppRNA or with IFNα-2b for 6 and 24 h were compared . The heatmap in Figure 5A displays the expression profile of genes differentially up- and down-regulated ( fold change ≥2; p-value ≤0 . 001 ) by 5′pppRNA and IFN ( black and blue genes ) or exclusively by 5′pppRNA ( red and green genes ) . Interestingly , 5′pppRNA induced a significantly broader gene expression program compared to IFNα-2b , especially at 24 h . To determine whether differences in gene expression observed between 5′pppRNA and IFNα-2b were due to sub-optimal stimulation by IFNα-2b , higher concentrations of IFNα-2b within the range reported for in vitro applications were tested . Treatment with 5′pppRNA at 10 ng/ml was equivalent to treatment with IFNα-2b at 100 IU/ml in terms of IFNα levels released into cell culture supernatant . Increasing the concentration of IFNα-2b to 1000 IU/ml corresponded to levels of IFNα that were 8-times greater than physiological secretion following 10 ng/ml 5′pppRNA treatment ( 2500 pg/ml vs 480 pg/ml; Fig . 5B ) . Regardless of the amount of IFNα-2b used to activate cells , gene expression levels remained relatively unchanged ( Fig . 5A; IFNα-2b 100 IU/ml vs 1000 IU/ml ) , indicating that IFNα-2b treatment was saturating . Remarkably , the spectrum of genes differentially expressed by IFNα-2b treatment were virtually all contained within the transcriptome induced by 5′pppRNA treatment - 57 at 6 h and 134 at 24 h , out of 139 genes ( Fig . 5C ) - demonstrating that 5′pppRNA induced a complete IFN response by 24 h . Some of the genes characteristically activated as part of the IFN signature and maximally induced by 5′pppRNA are MX1 , IFIT1 , ISG15 ( Fig . 5A , black portion ) . This comparison also highlighted the fact that a surprisingly large number of genes are uniquely regulated by 5′pppRNA - 38 genes at 6 h and 730 genes at 24 h ( Fig . 5C ) . Most notably , IFNB1 and all three members of the IFNλ family , as well as the cytokines CCL5 , CXCL10 , IL-6 , and CCL3 , were highly induced by 5′pppRNA but not IFNα-2b treatment ( Fig . 5A; red genes ) . While IFN signaling was highly induced by both treatments , IFNα-2b strongly activated antigen presentation machinery , and 5′pppRNA preferentially stimulated dendritic cell maturation and crosstalk , linking innate and adaptive immunity as well as induction of a wider range of signaling pathways ( Fig . 5D ) . Significantly , 5′pppRNA preferentially stimulated a more extensive induction of IRF7 and NF-κB signaling nodes compared to IFNα-2b treatment ( Fig . S2C ) . Thus , 5′pppRNA treatment , besides inducing a complete IFN response , additionally stimulated the transcription of a large and unique set of inflammatory and antiviral genes . To determine if the RIG-I agonist was capable of inducing a functional antiviral response , A549 cells were treated with 5′pppRNA , and 24 h later , challenged with VSV , Dengue ( DENV ) , or Vaccinia viruses . All viruses established infection in untreated cells as assessed by flow cytometry ( 60% , 20% and 80% , respectively ) but in 5′pppRNA-treated cells , VSV and DENV infectivity was reduced to <0 . 5% , while infection with vaccinia was reduced to 10% ( Fig . 6A ) . Release of infectious VSV and DENV virus was completely blocked by 5′pppRNA treatment ( 1 . 7×109 and 4 . 3×106 PFU/mL in untreated cells , respectively vs . undetectable in treated cells; Fig . S3 ) . Similarly , in primary human CD14+ monocytes , DENV infection decreased from 53 . 7% to 2 . 6% in the presence of 5′pppRNA; in the CD14− fraction , DENV infectivity was lower ( 3% ) , but was likewise inhibited by RNA agonist treatment ( 0 . 4%; Fig . 6B ) . To demonstrate the requirement for intracellular delivery of 5′pppRNA , primary CD14+ cells from three patients were treated with 5′pppRNA alone , transfection reagent alone or the combination; DENV infectivity was reduced from ∼30% to ∼0 . 5% only upon transfection of the RNA agonist ( Fig . 6C ) . To evaluate the antiviral effect of 5′pppRNA against HIV infection , activated CD4+ T cells were pre-treated with supernatant isolated from 5′pppRNA-treated monocytes and then infected with HIV-GFP . In the absence of treatment , 24% of the activated CD4+ T cells were actively infected by HIV as determined by GFP expression by flow cytometry , whereas infection of CD4+ T cells treated with 5′pppRNA supernatants was reduced to 11% ( Fig . 6D ) . 5′pppRNA also had an antiviral effect against HCV in hepatocellular carcinoma cell line Huh7; HCV NS3 expression was inhibited by 5′pppRNA treatment ( Fig . 6E; lane 4 vs . 2 and 6 ) . The antiviral effect was fully dependent on RIG-I , as demonstrated in Huh7 . 5 cells ( that express a mutated inactive RIG-I ) by the absence of IFIT1 up-regulation following 5′pppRNA treatment ( Fig . 6E; lane 9 ) and NS3 expression comparable to untreated HCV-infected cells ( Fig . 6E; lane 10 vs . 8 and 12 ) . Thus , 5′pppRNA is a broad-spectrum antiviral agent able to trigger an efficient innate immune response in different cell types and prevent infection by RNA and DNA viruses . To further explore the inhibitory potential of 5′pppRNA , A549 cells were pre-treated with 5′pppRNA for 24 h and then infected with H1N1 A/PR/8/34 Influenza virus at increasing MOI ( 0 . 02 , 0 . 2 , 2 ) ; influenza replication was monitored by NS1 protein expression ( Fig . 7A ) and plaque assay ( Fig . 7B ) . Viral replication was blocked by 5′pppRNA pre-treatment , even at the highest MOI , as demonstrated by complete loss of NS1 expression and 40-fold decrease in viral titer at MOI 2 . In A549 cells pre-treated with decreasing concentrations of 5′pppRNA ( 10–0 . 1 ng/ml ) prior to influenza virus challenge ( 0 . 2 MOI ) , 5′pppRNA significantly blocked influenza replication at a concentration of 1 ng/ml , as demonstrated by a 3-fold reduction in NS1 protein expression ( Fig . 7C; lane 7 ) and a 7-fold reduction in virus titer by plaque assay ( Fig . 7D ) . To demonstrate that the antiviral activity of 5′pppRNA against influenza relies on RIG-I signaling , A549 cells were knocked down for RIG-I and infected with influenza; in the knockdown , ISGs were not induced ( Fig . 7E , lanes 3 vs . 6 ) and 5′pppRNA treatment failed to inhibit NS1 expression ( Fig . 7E; lanes 5 vs . 6 ) , indicating that the antiviral effect of 5′pppRNA is exclusively dependent on RIG-I . Next , to determine whether the RIG-I ‘unique’ gene expression profile characterized in Figure 5 could compensate for the IFN response , A549 cells were knocked down for the IFNα/βR; the knock down was efficient , as demonstrated by the absence of IFIT1 and RIG-I induction following IFNα-2b stimulation ( Fig . 7F; lane 6 ) . Interestingly , induction of ISGs was only partially reduced following 5′pppRNA treatment ( 2 . 2-fold reduction of IFIT1 vs . siCtrl; Fig . 7F; lane 5 vs . 2 ) ; this IFN-independent activation of innate signaling was sufficient to reduce viral NS1 expression by 2 . 4-fold ( Fig . 7E; lane 9 vs . 8 ) . Thus , in lung epithelial A549 cells , 5′pppRNA treatment can efficiently inhibit influenza H1N1 replication in a RIG-I-dependent manner and stimulate an antiviral and inflammatory response independently of IFN signaling to limit influenza infection in vitro . To determine the potential of 5′pppRNA in vivo , C57Bl/6 mice were inoculated intravenously with 5′pppRNA ( 25 µg ) in complex with the in vivo-jetPEI transfection reagent . 5′pppRNA stimulated a potent immune response in vivo characterized by IFNα and IFNβ secretion in the serum and lungs ( Fig . S4A ) as well as antiviral gene up-regulation ( Fig . S4B ) . The response was potent and rapid with serum IFNβ levels increased ∼20-fold compared to basal levels , as early as 6 h post administration ( Fig . S4A; top left panel ) . The immune activation observed in vivo correlated with an early and transient recruitment of neutrophils to the lungs along with a more sustained increase in macrophages and dendritic cells populations ( Fig . S4C ) . Next , to determine the antiviral potential of 5′pppRNA in vivo , mice were treated with 5′pppRNA 24 h before ( day −1 ) , and on the day of infection ( day 0 ) with a lethal inoculum of H1N1 A/PR/8/34 Influenza . Whereas all untreated , infected mice succumbed to infection by day 11 , all 5′pppRNA-treated mice fully recovered ( 100% survival ) ( Fig . 8A ) . Overall , weight loss was similar between the two groups ( Fig . 8B ) , although a noticeable delay of 2–3 days in the onset of weight loss was observed in 5′pppRNA-treated animals; treated mice then fully recovered within 12–14 days ( Fig . 8B ) . Influenza replication in the lungs was monitored by plaque assay over the course of infection with virus titers in the lungs of untreated mice reaching a maximum at day 3 post-infection ( Fig . 8C ) . A decrease in virus titer was noted by day 9 post-infection , possibly correlating with the onset of adaptive immunity , and all animals succumbed to influenza infection by day 11 ( Fig . 8A ) . Interestingly , 5′pppRNA treatment inhibited influenza virus replication in the lungs early after infection , within the first 24–48 h ( Fig . 8C; Day 1 ) ; by day 3 , virus titers in the lung had increased , although influenza titers were still ∼10-fold lower compared to titers in untreated mice ( Fig . 8C; Day 3 ) . By day 9 , the 5′pppRNA-treated animals had controlled the infection , as demonstrated by the decrease in viral titers . Continuous administration of 5′pppRNA at 24 h intervals post-infection had an additive therapeutic effect that further delayed viral replication ( Fig . 8D; 3 versus 2 doses of 5′pppRNA ) , indicating that antiviral immunity may be sustained with repetitive administration of 5′pppRNA . Furthermore , therapeutic administration of 5′pppRNA also controlled influenza viral replication; although prophylactic treatment was most effective at blocking influenza dissemination in the lung , administration of the RNA agonist on day 1 and day 2 after influenza infection also reduced viral lung titers by ∼10-fold ( Fig . 8E ) . The antiviral response triggered by 5′pppRNA in vivo was dependent on an intact RIG-I signaling; serum IFNβ release was abolished in MAVS−/− mice , whereas the absence of TLR3 did not affect 5′pppRNA-induced IFNβ release ( Fig . 8F ) . In agreement , MAVS−/− mice treated with 5′pppRNA did not control influenza lung titers ( 5-fold increase vs . wt mice ) and the titer was comparable to untreated wt mice ( Fig . 8G ) . To determine whether 5′pppRNA treatment was sufficient to protect against influenza in the absence of IFN signaling , IFNα/βR−/− mice were treated or not with 5′pppRNA and challenged with influenza H1N1 virus . While untreated IFNα/βR−/− animals succumbed to infection , 40% of the animals that received 5′pppRNA treatment survived , suggesting that an IFN-independent effect of 5′pppRNA functioned in the absence of the IFN response . Thus , intravenous administration of 5′pppRNA stimulated a potent and rapid immune response in vivo that delayed influenza H1N1 virus replication in the lungs of infected animals and rescued mice from a lethal inoculation of influenza H1N1 . To further evaluate the effect of RNA agonist administration on influenza-mediated pathology , histological sections of lungs from untreated and treated mice were prepared and analysed . 5′pppRNA treatment alone was characterized by a modest and rare leukocyte-to-endothelium attachment; mixed leukocyte populations ( mononuclear/polymorphonuclear ) infiltrated the perivascular space at 24 h after injection ( data not shown ) but the infiltration resolved and was limited to endothelial cell attachment at 3 and 8 days after intravenous administration ( Fig . 9A ) . Influenza virus infection induced severe and extensive inflammation and oedema in the perivascular space and the bronchial lumen at day 3 post-infection . In animals receiving the RNA agonist , influenza triggered a mild and infrequent inflammation that did not extend to the bronchial lumen at day 3 post-infection . Epithelial degeneration and loss of tissue integrity were more severe in the lungs of untreated , infected animals and correlated with epithelial hyperplasia observed at later times , compared to the lungs of animals treated with 5′pppRNA . Inflammation and epithelial damage progressed in untreated mice by day 8 ( Fig . 9E ) , and correlated with increased virus titer in the lungs ( Fig . 8C ) ; inflammation and epithelial damage was consistently less apparent in agonist-treated mice . Strikingly , the surface area of the lungs affected by pneumonia was significantly reduced in 5′pppRNA-treated mice compared to non-treated mice – on day 3 , 16% vs 35%; day 8 , 41% vs 73% ( Fig . 9C; bottom panel ) . Overall , influenza-mediated pneumonia was less severe in animals administered 5′pppRNA before influenza challenge , demonstrating that 5′pppRNA possesses an antiviral effect in vivo that limits influenza replication in the lung , limits lung damage and prevents influenza-mediated pneumonia and mortality .
RIG-I agonists are attractive potential antiviral agents , as triggering the innate cytosolic RIG-I pathway mimics the earliest steps of immune recognition and response to viral pathogens . In the present study , a short 5′pppRNA agonist of RIG-I derived from the 5′ and 3′ UTRs of the VSV genome stimulated an antiviral response that protected human lung epithelial A549 cells or human PBMCs from challenge with several viruses , including DENV , Influenza , HIV , VSV , HCV and Vaccinia virus . Intravenous administration of the 5′pppRNA agonist in mice stimulated an antiviral state in vivo that protected animals from lethal influenza virus challenge and controlled influenza virus-mediated pneumonia . Analysis of the dynamics of the host transcriptome following 5′pppRNA stimulation was characterized by antiviral and inflammation related gene expression patterns with transcriptional nodes of genes regulated by IRF , NF-κB , and STAT families . Virtually all of the genes activated by IFNα-2b were encompassed within the 5′pppRNA transcriptome; bioinformatics analysis also identified distinct gene patterns and functional processes that were uniquely induced or inhibited by 5′pppRNA . Because of its potency both in vitro and in vivo , 5′pppRNA represents a specific and powerful trigger of innate immunity and a novel approach to antiviral therapy . For the first time , an RNA-based agonist of RIG-I was shown to block the replication of multiple viruses; this broad-spectrum antiviral activity of 5′pppRNA is attributable in part to a potent stimulation of the inflammatory and antiviral response driven by the early induction of IRF , NF-κB , STAT , chemokines and pro-inflammatory cytokine genes . In parallel , we also observed an inhibition of genes involved in TGF-β signaling . Because of the immunosuppressive nature of the TGF-β pathway [42] , inhibition of this transcriptional node could further potentiate immune activation in response to 5′pppRNA agonist . The emergence of apoptosis and ubiquitin signaling nodes at later times ( 24 h ) suggests a role for cell death and ubiquitin-based signal modification in the antiviral response . 5′pppRNA stimulation of RIG-I triggered a complete IFN response . At 24 h post treatment , 5′pppRNA induced the expression of 97% of the genes stimulated by IFNα-2b treatment and the magnitude of ISG induction by 5′pppRNA was enhanced compared to IFNα-2b profile . Among the ISGs , the tripartite motif containing ( TRIM ) proteins , the IFITM proteins , MX1 and viperin exemplify the range of ISGs induced by RIG-I and all have been implicated as inhibitors of HIV-1 , Influenza , VSV , West Nile , Dengue , and HCV [43]–[53] . In a recent study , high throughput screening of antiviral effectors identified a panel of broadly acting antiviral molecules , with the combined expression of multiple ISGs providing additive inhibitory effects against HCV replication [29] . Of the 28 validated antiviral ISGs identified , 19 were induced by 5′pppRNA in A549 , including IRF1 , RIG-I , Mda5 , IFITM3 . The transcriptome analysis also identified a distinct subset of 968 genes specifically induced by 5′pppRNA - and not IFNα-2b - that additively or synergistically enhanced the antiviral response stimulated by 5′pppRNA treatment . Bioinformatics analysis identified a unique functional signature with up-regulated genes involved in inducing a wider range of signaling pathways and bridging innate and adaptive immune responses . The importance of genes uniquely induced by 5′pppRNA is highlighted by the antiviral response that limited influenza infection in vitro and in vivo , even in the absence of functional type I IFN signaling . We speculate that the extended range of genes induced by 5′pppRNA , compared to IFNα-2b , reflect the activation of multiple signaling pathways downstream of RIG-I/MAVS [10] , versus the more-limited transactivation potential of the IFN-regulated JAK-STAT axis [26] . Type III IFNs ( IL29 , IL28A , IL28B ) were among the most highly stimulated genes uniquely up-regulated in response to 5′pppRNA . Recent studies have demonstrated that both type I and type III IFNs activate similar components of the JAK-STAT pathways , although type III IFNs were shown to prolong the activation of JAK-STAT signaling and induce a delayed and stronger induction of ISGs , compared to type I IFNs [54] . IFNλs have been increasingly implicated in antiviral therapies: 1 ) IFNλ administration in mice stimulated expression of Mx1 and protected IFNαR−/− mice from lethal influenza challenge [55]; 2 ) IL29 blocked HIV-1 replication by inhibiting virus integration and post-transcriptional events [56]; and 3 ) the combination of IL29 and IFNα or IL29 and IFNγ enhanced the induction of antiviral genes and effectively inhibited HCV and VSV replication [57] . In addition , polymorphism in or near IFNλ3 gene correlated with spontaneous or treatment-induced clearance of Hepatitis C infection and IFNλ therapy is now actively investigated for the treatment of HCV [58] . Altogether these results indicate that specific induction of IFNλ by 5′pppRNA may contribute to immunity at the site of viral infection . Of note , negative regulators of the innate immune response were also detected after 5′pppRNA stimulation . In addition to the induction of SOCS1 , USP18 , and IFIT1 by both 5′pppRNA and IFNα-2b treatment , unique negative regulators activated exclusively by 5′pppRNA were also identified: SOCS3 contributes to the inhibition of the JAK/STAT signaling [59] , and hence limits the amplification of the IFN response; A20 and IκBα inhibit the activation of the NFκB signaling complex [60] , [61] , which would prevent excessive inflammation . This observation suggests that targeting an upstream viral sensor may provide activation as well as negative feed-back regulation to terminate the immune response and prevent uncontrolled inflammation; as such , this approach may offer an advantage over IFN therapy in terms of limiting potential toxic side-effects . Activation of the RIG-I signaling pathway using 5′pppRNA also induced an integrated set of genes and pathways that can efficiently bridge the innate and adaptive immune responses and utilize multiple arms of both systems . 5′pppRNA mobilized genes that enhance trafficking of immune cells such as neutrophils , monocytes , naïve and memory T cells and B cells , including CCL17 , CCL20 , CXCL10 , CCL3 , CCL5 and many others . We also observed the induction of genes important for the activation of the effector arm of the adaptive immune system such as IL-6 , which has been shown to enhance CD8+ T cell survival and killing potential [62] . These cytokines and chemokines certainly play a role in initiating the innate and adaptive immune cell response , which is critical for the generation of efficient immunity against multiple viral infections in vivo . Intravenous administration of the RIG-I agonist stimulated a potent immune response in vivo that reached the lungs and prevented mortality associated with virus challenge . Histopathology analysis revealed diminished influenza-mediated lung damage and recruitment of inflammatory cells in infected lungs following 5′pppRNA treatment . The rapid control of virus replication , as demonstrated by significantly reduced virus titers in the lungs within 3 days of virus inoculation , may have prevented excessive immune cell recruitment early after infection . Influenza infection generates a complex pathogenesis mediated in part by viral- and immune-mediated damage [63]; therefore , the activation and recruitment of limited numbers of specific immune cell types , such as neutrophils , alveolar macrophages and dendritic cells may generate a beneficial antiviral microenvironment that additionally favor the initiation of adaptive immune response , which would eventually contribute to the in vivo efficacy of the agonist . Other groups have recently reported that 5′pppRNA induces a protective RIG-I mediated antiviral response that inhibits the replication of influenza virus [64]–[66] . The inhibition of influenza infection by 5′pppRNA was dependent on the 5′ppp moiety and the secondary IFN response was crucial for mounting an effective antiviral response [64] , [65] . Recently , a short RNA molecule with dual functionality was developed - a siRNA against influenza NP gene and an agonist of the RIG-I pathway [66] . This 5′pppRNA inhibited influenza infection in vitro and in vivo , but the contribution of RIG-I activation to the inhibition of influenza was not demonstrated . The defective interfering RNA produced during Sendai virus life cycle is the best characterized natural RIG-I ligand and is known to induce strong inflammatory response [67] . Interestingly , this molecule has adjuvant potential and could stimulate an antibody-dependent response directed to influenza antigens . Compared to these earlier studies , we adopted a systems approach to provide biochemical , transcriptional and biological mechanistic explanations for the antiviral efficacy observed in vitro and in vivo . Activating natural host defense to prevent establishment and dissemination of viral infection is a valuable alternative strategy to antiviral drugs that specifically target viral processes . Interferon therapy has been used in the clinic for over two decades and has proven effective in the treatment of certain viral infections , mainly Hepatitis B and Hepatitis C [58] , as well as malignancies and autoimmune diseases [68] , [69] . However , IFN therapy is also associated with significant side effects that limit its use [70] . PolyI:C , another dsRNA immune modulator , is also being tested in vitro and in vivo and has demonstrated efficacy against respiratory infections [71]–[73] . Along with antiviral drugs , vaccination is the primary approach to reduce morbidity and mortality associated with viral infection . Increasing the immunogenicity of vaccines with molecular adjuvants eliciting cytokines , co-stimulatory molecules , or immunomodulatory factors enhance the vaccine-elicited immune responses . 5′pppRNA has the advantage of mimicking viral recognition to trigger an immune response analogous to natural viral infection . Furthermore , the response stimulated by 5′pppRNA was reminiscent of the integrated and multipotent response elicited early following immunization by the most protective vaccine identified so far , the yellow fever YD17 vaccine [74] . Therefore , an immune modulator such as a RIG-I agonist may not only function as an antiviral therapeutic , but may also serve as a vaccine adjuvant to increase the magnitude of the antiviral immune response elicited by vaccine epitopes . Thus , the present study not only demonstrated the prophylactic and therapeutic antiviral potential of 5′pppRNA , but also opens the door to further investigation of the potential of RIG-I agonists as vaccine adjuvants .
The sequence of the 5′pppRNA was derived from the 5′ and 3′ UTRs of the VSV genome as previously described [17] . In vitro transcribed RNA was prepared using the Ambion MEGAscript T7 High Yield Transcription Kit according to the manufacturer′s instruction ( Invitrogen , NY , USA ) . The template consisted of two complementary viral sequences containing T7 promoter that were annealed at 95°C for 5 minutes and cooled down gradually over night ( 5′-GAC GAA GAC AAA CAA ACC ATT ATT ATC ATT AAA ATT TTA TTT TTT ATC TGG TTT TGT GGT CTT CGT CTA TAG TGA GTC GTA TTA ATT TC-3′ ) . The in vitro transcription reactions proceeded for 16 hours . 5′pppRNA was purified and isolated using the Qiagen miRNA Mini Kit ( MD , USA ) . Homologous RNA without 5′ppp moiety was purchased from IDT ( Integrated DNA Technologies Inc , Iowa , USA ) ; dephosphorylation of the 5′pppRNA using CIAP ( Invitrogen , NY , USA ) generated identical results ( data not shown ) . Secondary structure was predicted using the RNAfold WebServer ( University of Vienna , Vienna , Austria ) . RNA was analysed on a denaturing 17% polyacrylamide , 7 M urea gel following digestion with 50 ng/ul of RNase A ( Ambion , CA , USA ) or 100 mU/ul of DNase I ( Ambion , CA , USA ) for 30 min . A549 were grown in F12K ( Invitrogen , NY , USA ) supplemented with 10% FBS and antibiotics . MEFs were grown in DMEM supplemented with 10% FBS , non-essential amino acids , and L-Glutamine ( Wisent , Quebec , Canada ) . WT and RIG-I −/− MEFS were kind gifts from Dr . Shizuo Akira ( Osaka University , Osaka , Japan ) [75] . WT , Mda5−/− , TLR3−/− , TLR7−/− MEFS were kind gifts from Dr . Michael Diamond ( Washington University , St Louis , USA ) [76] , [77] . Lipofectamine RNAiMax ( Invitrogen , NY , USA ) was used for transfections in A549 according to manufacturer's instructions . For luciferase assays , transfections were performed in wt and RIG-I−/−; wt and Mda5−/− , TLR3−/− , TLR7−/− MEFs using Lipofectamine 2000 ( Invitrogen , New York , USA ) or jetPRIME ( PolyPlus , France ) , respectively . Plasmids encoding GFP-ΔRIG-I , IRF-7 , pRLTK , IFNα4/pGL3 and IFNβ/pGL3 were previously described [78] . MEFs were co-transfected with 200 ng pRLTK reporter ( Renilla luciferase for internal control ) , 200 ng of reporter gene constructs , together with 5′pppRNA ( 500 ng/ml ) or 100 ng of a plasmid encoding a constitutively active form of RIG-I ( ΔRIG-I ) [41] . IRF7 plasmid ( 100 ng ) was added for transactivation of IFNα4 promoter . At 24 h after transfection , reporter gene activity was measured by Dual-Luciferase Reporter Assay , according to manufacturer's instructions ( Promega , Wisconsin , USA ) . Relative luciferase activity was measured as fold induction ( relative to the basal level of reporter gene ) . For siRNA knock down , A549 cells were transfected with 50 nM ( 30 pmol ) of human RIG-I ( sc-61480 ) , IFN-α/βR α ( sc-35637 ) and β ( sc-40091 ) chain , or control siRNA ( sc-37007 ) ( Santa Cruz Biotechnologies , Dallas , USA ) using Lipofectamine RNAi Max ( Invitrogen , NY , USA ) according to the manufacturer's guidelines . Treatment with 5′pppRNA was performed 48 hrs later . Whole cell extracts were separated in 8% acrylamide gel by SDS-PAGE and were transferred to a nitrocellulose membrane ( BioRad , Mississauga , Canada ) at 4°C for 1 h at 100 V in a buffer containing 30 mM Tris , 200 mM glycine and 20% methanol . Membranes were blocked for 1 h at room temperature in 5% dried milk ( wt/vol ) in PBS and 0 . 1% Tween-20 ( vol/vol ) and then were probed with primary antibodies: anti-pIRF3 at Ser396 ( EMD Millipore , Massachusetts , USA ) , anti-IRF3 ( IBL , Japan ) , anti-RIG-I ( EMD Millipore , Massachusetts , USA ) , anti-ISG56 ( Thermo Fischer Scientific , Massachusetts , USA ) , anti-pSTAT1 at Tyr701 ( Cell Signaling Technology , Inc , Massachusetts , USA ) , anti-STAT1 ( Santa Cruz Biotechnology ) , anti-NS1 ( Santa Cruz Biotechnology ) , anti-pIkBα at Ser32 ( Cell Signaling Technology , Inc , Massachusetts , USA ) , anti-IκBα ( Cell Signaling Technology , Inc , Massachusetts , USA ) , anti-NOXA ( EMD Millipore , Massachusetts , USA ) , anti-cleaved Caspase 3 ( Cell Signaling Technology , Inc , Massachusetts , USA ) , anti-PARP ( Cell Signaling Technology , Inc , Massachusetts , USA ) , anti-β-Actin ( EMD Millipore , Massachusetts , USA ) . Antibody signals were detected by chemiluminescence using secondary antibodies conjugated to horseradish peroxidise and an ECL detection kit ( Amersham Biosciences , Inc , NJ , USA ) Whole cell extracts were prepared in NP-40 lysis buffer ( 50 mM Tris , pH 7 . 4 , 150 mM NaCl , 30 mM NaF , 5 mM EDTA , 10% glycerol , 1 . 0 mM Na3VO4 , 40 mM β-glycerophosphate , 0 . 1 mM phenylmethylsulfonyl fluoride , 5 µg/ml of each leupeptin , pepstatin , and aproptinin , and 1% Nonidet P-40 ) . WCE was then subjected to electrophoresis on 7 . 5% native acrylamide gel , which was pre-run for 30 min at 4°C . The electrophoresis buffers were composed of an upper chamber buffer ( 25 mM Tris , pH 8 . 4 , 192 mM glycine , and 1% sodium deoxycholate ) and a lower chamber buffer ( 25 mM Tris , pH 8 . 4 , 192 mM glcine ) . Gels were soaked in SDS running buffer ( 25 mM Tris , pH 8 . 4 , 192 mM glycine , 0 . 1% SDS ) for 30 min at 25°C and were then transferred to nitrocellulose membrane ( Amersham Biosciences ) . Membranes were blocked in PBS containing 5% milk ( wt/vol ) and 0 . 05% Tween-20 ( vol/vol ) for 1 h at 25°C and blotted with an antibody against IRF3 ( IBL , Japan ) . Antibody signals were detected by chemiluminescence using secondary antibodies conjugated to horseradish peroxidise and an ECL detection kit ( Amersham Biosciences , Inc , NJ , USA ) The release of human IFNα ( multiple subunits ) and IFNβ in culture supernatants of A549 , and murine IFNα and IFNβ in serum or lung homogenate ( 20% w/v ) from mice in response to 5′pppRNA were measured by ELISA according to manufacturer's instructions ( PBL Biomedical Laboratories , Piscataway , NJ ) . PBMCs were isolated from freshly collected blood using a Lymphocyte Separation Medium ( Cellgro ) as per manufacturer's instructions . After isolation , total PBMCs were frozen in heat-inactivated FBS with 10% DMSO . On experimental days , PBMCs were thawed , washed and placed at 37°C for 1 hr in RPMI with 10% FBS supplemented with Benzonaze™ nuclease ( Novagen ) to prevent cell clumping . The optimal concentration of 5′pppRNA to efficiently activate PBMC with minimal cytotoxicity was 100 ng/mL ( data not shown ) . PBMCs were isolated from the blood of patients in a study both approved by IRB and by the VGTI Florida Institutional Biosafety Committee ( 2011-6-JH-1 ) . Written informed consent approved by the Vaccine and Gene Therapy Institute Florida Inc . ethics review board ( FWA#161 ) was provided and signed by study participants . Research was conformed to ethical guidelines established by the ethics committee of the OHSU VGTI and Martin Health System . VSV-GFP , which harbors the methionine 51 deletion in the matrix protein-coding sequence [79] , was kindly provided by J . Bell ( Ottawa Health Research Institute , CA ) . Virus stock was grown in Vero cells , concentrated from cell-free supernatants by centrifugation , and titrated by standard plaque assay as described previously [80] . The recombinant vaccinia-GFP virus ( VVΔE3L-REV ) , a revertant strain of the E3L deletion mutant , was kindly provided by Jingxin Cao ( National Microbiology Laboratory , Public Health Agency of Canada , Winnipeg ) [81] , [82] . Dengue virus serotype 2 ( DENV-2 ) strain New Guinea C was grown in C6/36 insect cells for 7 days . Briefly , cells were infected at a MOI of 0 . 5 , and 7 days after infection , cell supernatants were collected , clarified and stored at −80°C . Titers of DENV stocks were determined by serial dilution on Vero cells , with intracellular immunofluorescent staining of DENV E protein at 24 h post-infection and denoted as infectious units per ml . Titers of Dengue virions were determined by standard plaque assay in Vero cells; plaques were fixed , stained and counted 5 days later . In infection experiments , both PBMCs and A549 cells were infected in a small volume of medium without FBS for 1 hour at 37°C and then incubated with complete medium for 24 h prior to analysis . HIV-GFP virus is an NL4-3 based virus designed to co-express Nef and eGFP from a single bicistronic RNA . HIV-GFP particles were produced by transient transfection of pBR43IeG-nef+ plasmid into 293T cells as described previously [83] , [84] . Briefly , 293T cells were transfected with 22 . 5 µg of pBR43IeG-nef+ plasmid by polyethylenimine precipitation . Media was replaced 14–16 h post-transfection , and viral supernatants were harvested 48 hrs later , cleared by low-speed centrifugation and filtered through a 0 . 45 µm low binding protein filter . High-titer viral stocks were prepared by concentrating viral supernatants 100-fold through filtration columns ( Amicon ) , then aliquoted and stored at −80°C . Viral titers were determined by p24 level ( ELISA ) and TCID50; briefly , 10-fold serial dilutions of concentrated viral supernatants were used to infect PBMCs from two donors pre-activated for 3 days with 10 µg/ml of PHA . Half of the media was replaced on day 4 , and 7 days after infection , supernatants were harvested and processed for p24 by ELISA . The Reed–Muench method was used to calculate the TCID50 . For HIV infection , CD14+ monocytes were negatively selected using the EasySep Human Monocytes Enrichment Kit ( Stem Cell ) as per manufacturer's instructions . Isolated cells were transfected with 5′pppRNA ( 100 ng/ml ) using Lyovec ( Invitrogen ) according to the manufacturer's protocol . Supernatants were harvested 24 h after stimulation and briefly centrifuged to remove cell debris . CD4+ T cells were isolated using EasySep™ Human CD4+ T cells Enrichment Kit ( Stem Cell ) according to the manufacturer's guidelines . Purified CD14+ monocytes and CD4+ T cells were allowed to recover 1 h in RPMI containing 10% FBS at 37°C with 5% CO2 before experiments . For HIV infection , anti-CD3 Ab ( 0 . 5 µg/ml ) were immobilized for 2 hours in 24-well plate and CD4+ T cells were then added along with anti-CD28 Ab ( 1 µg/ml ) to allow activation of T cells for 2 days . After activation , cells were incubated for 4 hours with supernatant of 5′pppRNA-stimulated monocytes and infected with HIV-GFP at an MOI of 0 . 1 . Supernatant from the monocytes was left for another 4 h before adding complete medium . HCV RNA was synthesized using the Ambion MEGAscript T7 High Yield Transcription Kit using linearized pJFH1 DNA ( a generous gift Takaji Wakita; National Institute of Infectious Diseases , Shinjuku-ku ) as template . Huh7 cells were electroporated with 10 mg of HCV RNA and at 5 days post-transfection , virus containing supernatant was collected , filtered ( 0 . 45 µm ) and stored at −80°C ( HCVcc ) . Huh7 or Huh7 . 5 cells were pre-treated with 5′pppRNA ( 10 ng/mL ) for 24 h . Supernatants containing soluble factors induced following 5′pppRNA treatment was removed and kept aside during infection . Cells were washed once with PBS and infected with 0 . 5 ml of undiluted HCVcc for 4 h at 37°C; then , supernatant was added back . At 48 h post-infection , WCEs were prepared; expression of HCV NS3 protein was detected by Western blot ( Abcam , Toronto , Ca ) Influenza H1N1 strain A/Puerto Rico/8/34 was kindly provided by Veronika von Messling ( Duke-NUS , Singapore ) . Viral stock was amplified in Madin-Darby canine kidney ( MDCK ) cells and virus titer was determined by standard plaque assay [85] . Cells were infected in 1 ml medium without FBS for 1 hour at 37°C . Inoculum was aspirated and cells were incubated with complete medium for 24 hours , prior to analysis . For viral infections , supernatants containing soluble factors induced following 5′pppRNA treatment was removed and kept aside during infection . Cells were washed once with PBS and infected in a small volume of medium without FBS for 1 h at 37°C; then supernatant was added back for the indicated period of time . The percentage of cells infected with VSV , Vaccinia and HIV was determined based on GFP expression . The percentage of cells infected with Dengue was determined by standard intra-cellular staining . Cells were stained with a mouse IgG2a mAb specific for DENV-E-protein ( clone 4G2 ) followed by staining with a secondary anti-mouse antibody coupled to PE ( Jackson Immuno Research ) . PBMCs infected with DENV were first stained with anti-human CD14 Alexa Fluor 700 Ab ( BD Biosciences ) . Cells were analyzed on a LSRII flow cytometer ( Becton Dickinson ) . Compensation calculations and cell population analysis were done using FACS Diva . C57Bl/6 mice ( 8 weeks-old ) were obtained from Charles River Laboratories . MAVS−/− and WT ( mixed 129/SvEv-C57Bl/6 background ) were obtained from Z . Chen ( The Howard Hughes Medical Institute , US ) . TLR3−/− mice were obtained from Taconic . IFNα/βR−/− mice were bred on a C57Bl/6 background . For intra-cellular delivery , 25 ug of 5′pppRNA was complexed with in vivo-jetPEI ( PolyPlus , France ) at an N/P ratio of 8 as per manufacturer's instructions and administered intravenously via tail vein injection . Unless otherwise indicated , 5′pppRNA was administered on the day prior to infection ( Day −1 ) and on the day of infection ( Day 0 ) . Mice under 4% isoflurane anesthesia were infected intra-nasally with 500 PFU of Influenza A/PR/8/34 ( Day 0 ) . For viral titers , lungs were homogenized ( 20% wt/vol ) in DMEM and titers were determined by standard plaque assay as previously described [85] . Briefly , confluent Madin-Darby Canine Kidney Cells ( MDCK ) were incubated with 250 µL of serial Log10 dilutions for 30 minutes , the sample was aspirated , and cells overlaid with 3 ml of 1 . 6% agarose in DMEM . Plaques were fixed and counted 48 h later . All animal experimentations were performed according to the guidelines of the Canadian Council on Animal Care and approved by the McGill University Animal Care Committee . The IFNα/βR−/− animal experimentations were approved by the INRS Institutional Animal Care and Use Committee . All five lobes of the lungs were collected and fixed in 10% neutral-buffered formalin for 24 h . The organs were paraffin-embedded and 4 µm sections were cut and stained with hematoxyline and eosin staining ( H&E ) . The slides were analysed by a board-certified independent veterinary pathologist . The kinetics and the comparison to IFNα-2b were performed as two separate experiments . A549 cells were stimulated with either 5′pppRNA ( 10 ng/ml ) or IFNα-2b ( 100 IU/ml or 1000 IU/ml ) for designated times . IFNα-2b ( Intron A ) was purchased from Schering Plough ( Kenilworth , NJ ) . Cells were collected and lysed for RNA extraction ( Qiagen , Valencia , CA , USA ) . Reverse transcription reactions were performed to obtain cDNAs which were hybridized to the Illumina Human HT-12 version 4 Expression BeadChip according to the manufacturer's instruction , and quantified using an Illumina iScan System . The data were collected with Illumina GenomeStudio software . First , arrays displaying unusually low median intensity , low variability , or low correlation relative to the bulk of the arrays were discarded from the rest of the analysis . Quantile normalization , followed by a log2 transformation using the Bioconductor package LIMMA was applied to process microarrays . To account for variability between batches , the data were adjusted using the ComBat procedure ( http://dx . doi . org/10 . 1093/biostatistics/kxj037 ) . Missing values were imputed with the R package ( http://cran . r-project . org/web/packages/impute/index . html ) . In order to identify differentially expressed genes between treated and controls ( untreated ) samples , the LIMMA package [86] from Bioconductor was used . For data mining and functional analyses , genes that satisfied a p-value ( <0 . 001 ) with ≥2 fold change ( up or down ) were selected . Probes that do not map to annotated RefSeq genes and control probes were removed . The expected proportions of false positives ( FDR ) were estimated from the unadjusted p-value using the Benjamini and Hochberg method [87] . All network analysis was done with Ingenuity Pathway Analysis ( IPA: Ingenuity systems , Redwood City , CA ) . The differentially expressed genes selected based on above criteria were mapped to the ingenuity pathway knowledge base with different colors ( red: up-regulated; green: down-regulated ) . The significance of the association between the dataset and the canonical pathway was measured in two ways: ( 1 ) A ratio of the number of genes from the dataset that map to the pathway divided by the total number of genes that map to the canonical pathway was displayed; ( 2 ) by over-representation analysis Fisher's exact test was used to calculate a p-value determining the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone . The pathways were ranked with −log p-values . Microarray data have been deposited in the NCBI Gene Expression Omnibus . Total RNA was isolated from cells using RNeasy Kit ( Qiagen , Valencia , CA , USA ) . Spleen and lungs were homogenized in RLT buffer and RNA isolated as per manufacturer's instruction . 1 ug of RNA was reverse transcribed using High-Capacity cDNA Reverse Transcription Kits from Applied Biosystems according to manufacturer's instructions . Parallel reactions without reverse transcriptase were included as negative controls . Relative amount of an intracellular RNA of interest was quantified by real-time PCR on a 7500 fast real-time PCR system and expressed as a fold change using SYBR Green ( Roche ) according to the manufacture's protocol . All data presented are relative quantification with efficiency correction based on the relative expression of target genes versus GAPDH as reference gene . Primers sets used for these studies are presented in Table S1 . | Development of safe and effective drugs that inhibit virus replication remains a challenge . Activation of natural host defense using interferon ( IFN ) therapy has proven an effective treatment of certain viral infections . As a distinct variation on this concept , we analyzed the capacity of small RNA molecules that mimic viral components to trigger the host antiviral response and to inhibit the replication of several pathogenic human viruses . Using gene expression profiling , we identified robust antiviral and inflammatory gene signatures after treatment with a 5′-triphosphate containing RNA ( 5′pppRNA ) , including an integrated set of genes that is not regulated by IFN treatment . Delivery of 5′pppRNA into lung epithelial cells in vitro stimulated a strong antiviral immune response that inhibited the multiplication of several viruses . In a murine model of influenza infection , inoculation of the agonist protected animals from a lethal challenge of H1N1 influenza and inhibited virus replication in mouse lungs during the first 24–48 h after infection . This report highlights the therapeutic potential of naturally derived RIG-I agonists as potent stimulators of the innate antiviral response , with the capacity to block the replication of diverse human pathogenic viruses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome",
"expression",
"analysis",
"medicine",
"hiv",
"prevention",
"respiratory",
"infections",
"immune",
"activation",
"influenza",
"immunity",
"to",
"infections",
"immunology",
"dengue",
"pulmonology",
"immune",
"defense",
"infectious",
"disease",
"control",
"immunomo... | 2013 | Systems Analysis of a RIG-I Agonist Inducing Broad Spectrum Inhibition of Virus Infectivity |
Genome-wide association studies ( GWAS ) have demonstrated the ability to identify the strongest causal common variants in complex human diseases . However , to date , the massive data generated from GWAS have not been maximally explored to identify true associations that fail to meet the stringent level of association required to achieve genome-wide significance . Genetics of gene expression ( GGE ) studies have shown promise towards identifying DNA variations associated with disease and providing a path to functionally characterize findings from GWAS . Here , we present the first empiric study to systematically characterize the set of single nucleotide polymorphisms associated with expression ( eSNPs ) in liver , subcutaneous fat , and omental fat tissues , demonstrating these eSNPs are significantly more enriched for SNPs that associate with type 2 diabetes ( T2D ) in three large-scale GWAS than a matched set of randomly selected SNPs . This enrichment for T2D association increases as we restrict to eSNPs that correspond to genes comprising gene networks constructed from adipose gene expression data isolated from a mouse population segregating a T2D phenotype . Finally , by restricting to eSNPs corresponding to genes comprising an adipose subnetwork strongly predicted as causal for T2D , we dramatically increased the enrichment for SNPs associated with T2D and were able to identify a functionally related set of diabetes susceptibility genes . We identified and validated malic enzyme 1 ( Me1 ) as a key regulator of this T2D subnetwork in mouse and provided support for the association of this gene to T2D in humans . This integration of eSNPs and networks provides a novel approach to identify disease susceptibility networks rather than the single SNPs or genes traditionally identified through GWAS , thereby extracting additional value from the wealth of data currently being generated by GWAS .
Genome-wide association studies ( GWAS ) have revolutionized our ability to identify the causal determinants for common human diseases over the past several years , delivering an unprecedented number of DNA loci associated with a diversity of common human diseases like age-related macular degeneration [1] , [2] , Crohn's disease [3] , type 1 diabetes [3] , [4] , coronary artery disease [3] , [5] , HIV-1 infection [6] , and type 2 diabetes ( T2D ) [3] , [7]–[10] . One interesting characteristic of single nucleotide polymorphisms ( SNPs ) identified as associated with disease in these studies is that the great majority do not affect the coding sequence of genes , most often falling in introns or intergenic regions [11] . As a result , GWAS do not necessarily lead directly to the gene or genes in a given locus associated with disease , they do not typically inform the broader context in which the disease genes operate , and even in cases where the susceptibility gene is identified , GWAS do not usually indicate whether you would knock the gene down or activate it in order to treat the corresponding disease . Therefore , GWAS on their own provide limited insights into the mechanisms driving disease [12]–[14] . In addition , the amount of genetic variation explained by GWAS for a given disease is most often significantly less than the heritabilities estimated for the disease . For example , a number of studies estimate the genetic heritability for T2D to be as high as 40% , but the 18 DNA loci identified for T2D to date account for only ∼3% of the variation in T2D [10] . This raises the question of whether there are many more common DNA variants with smaller effects that are not being identified in the GWAS due to lack of power , whether there are many more rare variants with stronger effects that explain the missing variation , or some combination of the two [11] , [15] . In fact , in the span of just a few short years in which large-scale GWAS have been carried out , the realization that tractable drug targets and clinically useful biomarkers of disease are not immediately falling out of the data , has for some reduced the enthusiasm for the GWAS approach , intensifying the debate over whether GWAS are the best strategy to elucidate the causes of disease [16]–[18] . Some have attempted to look for enrichments in pre-defined sets of pathways defined by GO , KEGG or other pathway sources and found common variants involved in T2D risk are likely to occur in or near genes in multiple pathways [19] . One clear and immediate task to provide further insights into GWAS is to develop an understanding of the genetics of gene expression ( GGE ) to facilitate a systems-based understanding of disease . Recently , detailed GGE studies have provided a way to address several of these GWAS limitations [13] , [14] , [20]–[22] . By mapping the genetic architecture of gene expression in human populations , GGE studies can provide functional support for candidate genes within a given locus . This has been demonstrated a number of times , but most recently in identifying SORT1 , PSRC1 , and CELSR2 as candidate susceptibility genes for heart disease and plasma lipid levels [14] , and ORMDL3 as an asthma susceptibility gene [20] , [23] . More generally , GGE studies provide the necessary information to infer causal relationships among genes and between genes and clinical traits , leading to whole gene networks that provide a broader context within which to elucidate the biological function of any given gene with respect to diseases of interest [12]–[14] , [24] , [25] . One way GGE studies can impact interpretation of GWAS is by providing a way to reduce the dimensionality of the DNA variation space , limiting focus to those DNA variants that have been associated with expression traits and testing whether such SNPs are associated with disease [12] . The set of SNPs associated with expression ( eSNPs ) in disease-relevant tissues can be considered a functionally relevant subset of all SNPs across the human genome , given they associate with a biologically relevant event ( gene expression ) . However , the extent to which eSNPs inform on disease biology has not been comprehensively characterized for any disease . In this paper , we systematically examined whether eSNPs are more likely to associate with T2D compared to SNPs that a priori have no association to biologically relevant events . We assembled a comprehensive set of eSNPs identified in two GGE study cohorts representing three tissues [12]: liver , subcutaneous fat and omental fat tissues . Given the metabolic relevance of these tissues and the large-scale GWAS undertaken for T2D [26] , we tested whether this set of eSNPs was more likely to associate with T2D than randomly selected SNPs . We further constructed a co-expression network from subcutaneous adipose tissue isolated from a mouse population segregating T2D traits and asked whether eSNPs associated with genes comprising these networks and sub-networks were enriched for association with T2D ( Figure 1 ) . By comparing the relative enrichments for association to T2D at these increasing levels of granularity , we sought to identify disease-associated subnetworks whose member genes might play important roles in T2D pathogenesis .
We identified eSNPs from two GGE studies: 1 ) a liver tissue cohort comprised of 427 individuals [12] , and 2 ) a cohort comprised of ∼900 individuals from which liver , subcutaneous and omental adipose tissues were collected from each individual . The number of eSNPs from each tissue and the corresponding cohort sample sizes are summarized in Table S1 . To test whether the eSNPs were enriched for association to T2D , we assembled GWAS results from three previously published T2D studies: 1 ) the Wellcome Trust Case Control Cohort ( WTCCC ) [3] , 2 ) the Diabetes Genetics Initiative ( DGI ) [7] , and 3 ) the Diabetes Genetics Replication And Meta-analysis ( DIAGRAM ) Consortium [10] , which combines the results from WTCCC , DGI , and Finland–United States Investigation of NIDDM Genetics ( FUSION ) [8] . To assess whether these distributions were enriched for SNPs associated with T2D , we empirically estimated the null distribution by randomly sampling 100 , 000 sets of SNPs from a set of SNPs genotyped in each study ( chosen from the full set of SNPs in each study ) such that the SNP set size , the location distribution of the SNPs with respect to protein coding genes , and the minor allele frequency ( MAF ) distribution were similar to that of the eSNP set . The distribution of T2D eSNP association p values from the GWAS ( referred to here as PT2D ) differed significantly from the null distribution in that the eSNP PT2D values were skewed towards the significance end of the PT2D spectrum . For example , in the DGI study , 6 . 2% of the eSNPs ( 241 out of 3 , 888 total ) had PT2D<0 . 05 , compared to a mean of 5 . 2% ( 202 out of 3 , 888; 95% confidence interval ( CI ) : 4 . 6% to 5 . 8% ) over the 100 , 000 randomly generated matched sets ( Z = 3 . 16; p = 8 . 00×10−4 , Table 1 , Figure 2 ) , representing a 1 . 19 fold enrichment for SNPs in the eSNP set over the random sets . In addition to testing for enrichments of eSNPs with PT2D<0 . 05 , we compared the overall average PT2D of the eSNP set to randomly selected SNP sets matched to the eSNP set with respect to location and MAF . The results were similar to the enrichment observed for eSNPs with PT2D<0 . 05 ( Figure S1 ) . Because different SNP panels were used in the different GGE and GWA studies , many of the eSNPs were not genotyped in any of the T2D GWAS . Therefore , we recomputed the PT2D distributions based on all SNPs in strong linkage disequilibrium ( LD ) with the eSNPs . A SNP was considered in strong LD with an eSNP if the correlation between the two SNPs was >0 . 89 . These SNPs were considered to be representative of our eSNPs and were included in the analysis set ( referred to here as the expanded eSNP set ) in order to extract the most information from the GWAS data . We again tested whether this expanded eSNP set was enriched for SNPs associated with T2D by empirically estimating the null distribution . For example , in the DGI study , 1 , 516 SNPs in the expanded eSNP set of 24 , 220 SNPs ( 6 . 3% ) had PT2D<0 . 05 , compared to an average of 1 , 279 SNPs ( 5 . 3%; [95% CI: 4 . 9% to 5 . 7%] ) in the random sets ( Z = 5 . 05; p = 2 . 19×10−7 ) , representing a 1 . 19-fold enrichment for SNPs in the expanded set over the random sets . Similar enrichments were observed in the DGI and WTCCC studies ( Table 1 , Figure 2 ) . While the eSNP PT2D enrichments in liver , omental and subcutaneous tissue were statistically significant , the enrichment was modest ( 1 . 19 fold enrichment for the expanded eSNP set ) . One explanation for this could be that these enrichments were calculated using an eSNP set spanning three distinct tissues without considering how the expression traits relate to networks associated with disease . Therefore , even though the eSNPs considered herein were derived from metabolically active tissues , we considered the possibility that restricting attention to eSNPs corresponding to expression traits in T2D-relevant tissues that are most variable in populations segregating T2D traits may enhance the enrichment for eSNPs associated with T2D . Towards this end , we tested whether eSNPs corresponding to genes comprising an adipose tissue gene network constructed from an F2 intercross between C57BL/6 ob/ob and BTBR ob/ob mice ( referred to here as the B6×BTBR cross ) were enriched for association with T2D . The B6×BTBR cross has been previously established as a model population for T2D [27] . While the C57BL/6 ob/ob strain becomes obese and develops moderate hyperglycemia , it is compensated by hyperinsulinemia , preventing beta-cell failure and the development of a T2D phenotype . In contrast , the BTBR ob/ob strain develops obesity , accompanied by severe hyperglycemia and insulin resistance , ultimately resulting in beta-cell failure and a severe T2D phenotype . Therefore , the gene networks in T2D-relevant tissues in the B6×BTBR mice have the potential to provide insight into pathways and regulatory networks in obesity-induced diabetes [28] , [29] . In this setting , we define a gene network as a graphical model comprised of nodes and edges , where the nodes represent gene expression traits or clinical traits , and the edges represent significant , weighted correlations between the corresponding two nodes ( expression traits ) [30] . Because gene expression , DNA variations and T2D traits were all scored in B6×BTBR cross , there is the potential to identify tissue-specific subnetworks that are causally associated with T2D traits , given DNA variations can be treated as a perturbation on the gene expression and clinical traits , thereby enabling the edges in the network to be directed [12] , [24] , [31]–[33] . Of the 39 , 600 genes represented on the microarray used in this study , the upper 25 percent of the most differentially expressed genes were used as input to construct the coexpression network [30] . We then restricted our eSNP set to those omental adipose eSNPs corresponding to genes in the adipose network that mapped to human orthologs ( referred to as the adipose eSNP set ) and found the expanded eSNP set significantly more enriched for T2D associated SNPs compared to randomly selected eSNPs in the DIAGRAM and DGI studies . In the DGI study , of the 3 , 342 expanded eSNPs from the adipose set considered , 303 ( 9 . 07% ) were associated with T2D at the 0 . 05 significance level , compared to a mean of 6 . 2% [95% CI: 5 . 31% to 6 . 95%] in random expanded eSNP sets ( Z = 7 . 02; p = 1 . 10×10−12 ) . In the DIAGRAM study , 9 . 2% were associated with T2D at the 0 . 05 significance level , compared to a mean of 7 . 3% [95% CI: 6 . 93% to 7 . 65%] in random expanded eSNP sets ( Z = 10 . 40; p<10−16 ) . However , the adipose eSNP set was not significantly more enriched with small PT2D in the WTCCC study ( p = 0 . 35; Figure 2; Table 1 ) . The lack of significance in the WTCCC cohort was of interest , and given DIAGRAM contains both the DGI and WTCCC cohorts , the intermediate enrichment of DIAGRAM with respect to WTCCC and DGI reflects the strong significance in DGI and lack of significance in WTCCC . It is of particular note that one critical difference between the DGI and WTCCC studies was the matching of DGI cases and controls for BMI , whereas no such matching was done in the WTCCC study . As the adipose network was derived from a mouse cross whose parental strains are both on an ob/ob background , the BMI matching in DGI may confer more biological similarities to the cross design and hence better overlap . In addition , while the BMI matching in DGI may enhance power to identify beta-cell loci , rather than loci whose effect on T2D risk was mediated by obesity [34] , the BMI matching would not fully account for waist circumference , where those individuals with increased waist circumference compared to individuals with a similar BMI are at increased risk of T2D , where omental adipose tissue is thought to play a role [35] . The genes comprising the adipose and islet co-expression networks are not expected to uniformly affect T2D traits [12] , [13] . Figure 3A depicts the most highly connected expression traits in the adipose network as a topological overlap map [30] . The adipose network is composed of distinct subnetworks or modules that emerge among the highly interconnected expression traits [36] . Such co-expression subnetworks often contain genes of related biological function [37] . For example , the purple subnetwork in the adipose network was found to be the subnetwork most significantly associated with T2D traits . The genes comprising this subnetwork were enriched for the Panther biological process lipid , fatty acid and steroid metabolism ( p = 4 . 49×10−8 , Table 2 ) . The first principal component of the gene expression traits making up this subnetwork explained 45 . 6% of the expression variation of the subnetwork and was strongly positively correlated with several T2D clinical traits measured in the B6×BTBR mice: number of islets ( R = 0 . 52 , p<1×10−70 ) , plasma insulin levels ( R = 0 . 70 , p<1×10−70 ) , and plasma glucose levels ( R = −0 . 57 , p<3 . 9×10−41 ) . We next applied a previously described method for inferring causal relationships between the expression traits and T2D traits with respect to genetic loci controlling for the islet count phenotype and plasma glucose and insulin levels [24] . We have previously shown that subnetworks under the control of genetic loci that are also associated with disease traits can be enriched for genes predicted to cause disease trait variation [38] . The purple subnetwork was supported as the most strongly causal subnetwork for the T2D traits in the adipose coexpression network in the B6xBTBR cross ( referred to here as the T2D adipose causal subnetwork ) , with 36% ( Fisher Exact Test p = 5 . 26×10−68 ) , 27% ( Fisher Exact Test p = 1 . 40×10−50 ) and 29% ( Fisher Exact Test p = 1 . 55×10−44 ) of the genes in this subnetwork supported as causal for plasma insulin levels , plasma glucose levels , and number of islets , respectively ( Figure 3C , Table 2 , Table S2 ) . Therefore , while there are many subnetworks identified in the adipose network , they are not all associated with T2D traits , and in the context of the B6xBTBR cross there is a single subnetwork in adipose that is very significantly enriched for genes causally associated with T2D . Given the strong causal relationship inferred between the T2D adipose causal subnetwork and T2D in the BTBRxB6 cross , we tested whether omental adipose eSNPs corresponding to genes in this subnetwork were more significantly enriched for association to human T2D compared to the adipose network filtered eSNP sets . Astonishingly , of the 101 SNPs in the expanded eSNP set that were associated with the expression of genes in the T2D adipose causal subnetwork , 37 ( 37% ) corresponded to PT2D<0 . 05 in the DGI study , compared to an average of 9 . 0% SNPs ( [95% CI: 3 . 44% to 14 . 63%] ) in the matched random SNP sets sampled from the adipose network expanded eSNPs ( p<10−16 ) , further supporting that this subnetwork is an important network for human T2D , and further supporting that the this causal subnetwork may reflect important molecular states associated with increased omental fat mass and the link of this increased fat mass to T2D . Similar enrichments were observed in the DIAGRAM study , although as in the case of the adipose network , the WTCCC enrichment was not significant ( Table 1 , Figure 2 ) . In addition to the dramatic enrichments observed in restricting attention to those human omental eSNPs corresponding to genes in the B6xBTBR T2D adipose causal subnetwork , the eSNPs generated from the other tissues and from all tissues combined were also enriched for lower PT2D values . The increasing enrichment trend was consistently observed from all tissue-GWAS combinations ( Figure S2 ) . While the enrichment magnitude and significance levels were somewhat tissue dependent , there were no profound differences among liver , omental fat and subcutaneous fat tissue eSNPs , possibly reflecting that all three tissues are metabolically active and important in obesity and diabetes . When comparing eSNPs identified in independent tissues from the same cohort , 72% of the cis eSNPs identified in liver , 79% of those found in omental adipose , and 81% from subcutaneous adipose were also found in the other two tissues: 2 , 189 , 2 , 286 and 1 , 999 tissue specific eSNPs were identified in liver , omental adipose , and subcutaneous adipose , respectively . This is consistent with previous findings on tissue-specific effects [13] , [39] . Because there is reduced numbers of eSNPs represented in the tissue specific sets , there is reduced power overall to detect enrichments . We note that in cases where the WTCCC enrichments were not significant in restricting attention to omental eSNPs , the enrichments were significant when focusing on eSNPs over all tissues combined . By pooling eSNPs from liver and adipose tissue , our main aim was to increase power to detect enrichments by increasing the number of eSNPs . While pooling of eSNPs from the three tissues was a first step in our analysis , restricting attention to the most disease relevant tissue in this case resulted in the most dramatic enrichment , highlighting the importance of the mouse cross in identifying the most causal subnetworks for disease in the most disease relevant tissue corresponding to the disease relevant tissues we had available in the human cohort . While the enrichment of eSNPs associated with genes in the T2D adipose causal subnetwork was encouraging ( 37% of the eSNPs in this subnetwork associated with T2D at the nominal 0 . 05 significance level ) , the effect sizes were all small , providing for very little power to prioritize the list for direct experimental validation based on the human association data alone . Given the number of putative causal genes represented in this module , we could not carry out experimental validation on all of them . Therefore , we integrated the mouse and human data to prioritize the list of T2D susceptibility genes for validation . To identify susceptibility genes for validation , we identified genes in the T2D adipose causal subnetwork that harbored DNA variations in mouse and human that associated with its expression levels and that were supported as causal for T2D [24] . This is a natural filter to apply , given DNA variations that directly affect the activity of a gene in multiple species , and that in turn are supported as causing variations in disease traits [24] , strongly implicate such genes as affecting disease susceptibility [14] . Specifically , for validation purposes , we focused on genes meeting the following three criteria: 1 ) adipose expression levels in the B6xBTBR cross were associated with genotypes for markers proximal to the gene of interest ( i . e . , genes that gave rise to cis eQTL ) ; 2 ) supported as causal for T2D traits using a previously described statistical procedure to infer causal relationships between expression and clinical traits [24]; and 3 ) gave rise to an adipose cis-eSNP in humans that associated with T2D in human GWAS . Application of this filter identifies those expression traits in the B6xBTBR cross and human GGE cohorts that are perturbed by cis DNA variation , and that in turn associate with T2D traits , directly supporting the genes as causal for T2D in the B6xBTBR cross and the human population . Of the 159 expression traits in the T2D adipose causal subnetwork , 117 gave rise to cis or trans expression QTL ( eQTL ) in a distinct region on chromosome 9 ( from 65Mb to 95Mb of the chromosome ) . However , only 8 of these genes were identified with strong adipose cis-eQTLs ( i . e . , the structural genes were located within the chromosome 9 linkage region ) . Further , 5 of these 8 genes ( Anxa2 , Bcl2l10 , C730029A08Rik , Me1 , Paqr9 ) were supported by the mouse data as causal for T2D traits ( Figure 3B ) . Among these , only human ME1 adipose expression was associated with at least one cis-eSNP that was also nominally associated with T2D in the DIAGRAM study ( PT2D = 0 . 002 ) ( Figure S3 ) . Therefore , while Me1 was supported as causal in the B6xBTBR cross , it was one of hundreds of genes supported as causal for T2D traits , but then the only gene of those hundreds whose expression in humans associated with a SNP that also associated with human T2D . The role of Me1 in obesity [40]–[42] , energy homeostasis [43] and diabetes [44] has been well documented in the literature . Encoding a cytosolic NADP ( + ) -dependent enzyme involved in the formation of pyruvate from malate , it produces NADPH to supply reducing equivalents for lipogenesis , thus siphoning the reducing equivalents originally derived from glycolysis as NADH to NADPH for fatty acid synthesis [45] . Me1 is co-regulated together with fatty acid synthetic enzymes by Chrebp and Srebp-1c and is therefore described as a lipogenic enzyme . Further , we recently provided direct experimental support for the involvement of Me1 in obesity-related phenotypic characteristics and in gene networks associated with obesity using a Me1 knockout ( Me1−/− ) mouse model [31] . Here , we extended the validation experiment to T2D related traits . As shown in Table 3 , the Me1−/− mice fed a high fat diet ( HFD ) demonstrate significantly lower insulin levels compared to the controls ( p = 1 . 23×10−9 ) , thus validating our prediction . In addition , the Me1−/− mice showed lower serum glucose levels ( p = 3 . 30×10−6 ) and an improved glucose tolerance at week 23 ( Figure 3D ) , with a 29 . 5% decrease in the area under the oral glucose tolerance test ( OGTT ) curve ( AUC ) relative to wild-type mice ( p = 7 . 30×10−8 ) . All of these lines of evidence support a diabetes-resistant phenotype in Me1−/− mice . Furthermore , the Me1−/− mice also possessed significantly improved lipid profiles including lower total cholesterol ( p = 2 . 19×10−3 ) and triglyceride ( p = 1 . 40×10−7 ) levels . Consistent with the lower body fat reported earlier [31] , the serum leptin levels were also significantly lower in the Me1−/− mice than in the controls . Therefore , the Me1−/− mice appeared to be resistant to both diabetes and obesity development . In order to explore the mechanisms underlying the observed phenotypic changes in the context of the subnetworks identified by the eSNP filtering method , we constructed a single gene perturbation gene expression signature for Me1 , comprised of 2 , 958 genes , by identifying adipose genes differentially expressed between wild type and Me1−/− male mice . The molecular perturbation signature can serve as an important molecular validation that a putative causal gene underlying a linkage region associated with disease is in fact one of the genes in the linkage region explaining the linkage signal [46] . We found that the Me1 perturbation signature was significantly enriched for many metabolic pathways , including insulin receptor signaling pathway ( p = 2 . 27×10−5 ) , fatty acid ( p = 5 . 49×10−6 ) , amine ( p = 8 . 67×10−8 ) , lipid ( p = 5 . 47×10−7 ) , and monocarboxylic acid metabolic processes ( p = 4 . 73×10−7; similar to the purple mouse subnetwork depicted in Figure 3A ) . The Me1 perturbation signature was significantly enriched for expression traits in the T2D adipose causal subnetwork: 32 genes overlapped this network whereas only 13 would have been expected by chance , a greater than 2-fold enrichment ( Fisher Exact Test p = 2 . 95×10−7; Figure 3E , Table 2 ) . This serves as an important molecular validation of the eSNP filtering method and confirms the causal nature of a gene identified through this approach , Me1 .
GGE studies provide the necessary information to infer causal relationships among genes and between genes and clinical traits , leading to the construction of gene networks that underlie diseases of interest [12]–[14] , [24] , [25] . Three fundamental advances presented herein significantly extend this earlier work: 1 ) to our knowledge , we have demonstrated for the first time that SNPs that associate with human gene expression traits in metabolically relevant tissues are enriched for associating with T2D in multiple T2D studies; 2 ) the enrichment of eSNPs associating with T2D over randomly selected SNPs dramatically increased as we restricted attention to eSNPs corresponding first to genes comprising the co-expression network from adipose tissue isolated from a mouse population segregating T2D traits , and then to genes comprising a specific adipose subnetwork strongly supported as causal for T2D-associated traits; and 3 ) we demonstrated directly that causal gene networks provide a path to functionally informing on genetic loci found in GWAS to associate with disease . The inability of GWAS studies to directly elucidate the causal genes and their function with respect to disease is now widely accepted as a problem in search of a solution; we provide one possible solution . Our results taken together support the idea that common forms of disease like T2D are emergent properties of networks that respond to wide-spread variation ( genetic and environmental ) , as opposed to the result of single hits to single genes . The eSNPs for genes in the T2D adipose causal subnetwork that were enriched for associating with T2D were too subtly associated with the disease to be identified in a classic GWAS , due to lack of power . However , the associations were detectable by reducing the number of SNPs tested in a GWAS , given the focus was on those SNPs that associate with the expression of genes in a subnetwork supported as causal for the disease of interest . The causal reasoning we have used to identify causal relationships between genes and disease traits refers to a statistical inference procedure in which statistical associations between changes in DNA , changes in expression , and changes in complex phenotypes like disease are examined for patterns of statistical dependency among these variables that support directionality among them , where the directionality then provides the source of causal information . This stands in contrast to the classic use of causality in molecular biology or biochemistry , where causality between two proteins implies that one protein has been determined experimentally to physically interact with or to induce processes that directly affect another protein , and that this in turn leads to a phenotypic change of interest . Therefore , experimental validation in this setting is critical . Towards that end , ME1 was identified as a putative driver of a gene subnetwork containing key regulators of lipogenesis and was then validated in vivo as a gene capable of modulating multiple T2D traits . The genes whose adipose expression levels change in response to knocking out Me1 were enriched for genes that 1 ) fell in this subnetwork , and 2 ) were supported as causal for T2D in this mouse T2D population . As we have previously detailed , this provides direct experimental support for the gene as a causal regulator of the subnetwork [12] , [24] , [46] . The T2D adipose causal subnetwork contains several co-expressed genes encoding key lipogenic enzymes , such as fatty acid synthase ( Fasn ) , ATP citrate lyase ( Acly ) , stearoyl-Coenzyme A desaturase 2 ( Scd2 ) , lanosterol synthase ( Lss ) , farnesyl diphosphate synthetase ( Fdps ) , and phospholipase A2 , group V ( Pla2g5 ) . The abnormal liporegulation found in obesity has previously been implicated in the pathogenesis of diabetes [47] , [48] , especially around the deleterious effects of the elevated levels of triglycerides in peripheral tissues , referred to as “lipotoxicity” . Excess circulating fatty acids present during obesity can accumulate in skeletal muscle tissues , contributing to insulin resistance [49] , [50] , [51] . Another organ negatively impacted by lipotoxicity is the pancreatic islets , where elevated fatty acid levels have been shown to contribute to β-cell apoptosis , a process thought to involve the de novo formation of ceramide and increased nitric oxide ( NO ) production , resulting in impaired glucose-stimulated insulin secretion [52] , [53] , [54] . Due to the key role played by Me1 in fatty acid synthesis , we hypothesized that a genetic knockout of malic enzyme in mice fed a high-fat diet would severely perturb this pathway . This would in turn lead to a decrease in circulating free fatty acids and triglycerides , a diminished ectopic triglyceride deposition , and consequently an improved insulin sensitivity profile . Indeed , both male and female Me1−/− mice exhibited dramatically improved responses to an OGTT ( Figure 3D ) , as well as significantly lower plasma triglyceride levels ( Table 3; see Text S1 for further discussion of Me1 and diabetes ) . It is important to note that while an adipose subnetwork strongly supported as causal for diabetes in an experimental mouse population demonstrated increased T2D eSNP enrichment when compared to the adipose network as a whole , only moderate enrichments were observed for all eSNPs and adipose-specific eSNPs . One possible explanation could be the limited coverage of eSNPs . For instance , certain GWAS SNPs may not affect gene expression , rather , they may alter post-transcriptional mechanisms such as mRNA splicing , or protein function . In other words , eSNP selection based on the GGE might have missed classes of important functional GWAS SNPs , and thus caused a loss of power . Additionally , our GGE cohorts may not have been appropriately powered to pick up all relevant eSNPs for T2D . The eSNPs used in this study are primarily from liver and adipose tissues . Although these are relevant tissues for T2D , other key tissues such as islet , muscle , and even brain were not available for eSNP discovery and hence a significant percentage of tissue-specific eSNPs were missing from our analysis . This emphasizes the importance of tissue selection for the success of this type of approach . Since many aspects of disease pathology are confined to certain tissues , the ability for eSNPs to inform on the biology relies on having a tissue-appropriate set of eSNPs . Related to this is our characterization of human gene expression traits in non-T2D individuals , which may have caused us to miss many relevant T2D eSNPs . Our first GGE cohort was a population-based random sample , while the second was an obese cohort , hence neither represents an appropriately powered T2D-specific cohort . Finally , the sample sizes of the GGE cohorts were not powered well enough to pick out the types of modest effects found in large GWAS studies . In our analysis , we pooled the eSNPs from the two cohorts in the three tissues as a starting point , mainly to improve power to observe pathway-specific signals . Many of these caveats associated with limited coverage of eSNPs are being addressed via increased funding for very large GGE studies . Therefore , we think the results realized here provide the beginning lines of evidence that eSNPs may in fact generally be enriched with disease associating SNPs . The set of eSNPs used in our analyses were identified at a false discovery rate ( FDR ) <10% . The motivation for selecting what could be considered a high FDR threshold was to increase the number of eSNPs to enhance the power to detect patterns of enrichment , as opposed to limiting attention to only the highest confidence single genes associated with disease . We also consider the possibility that the effective FDR decreases as we apply the filtering process of restricting attention to eSNPs whose associated genes are present in co-expression networks and subnetworks supported as causal for diabetes traits . We therefore suspect that this filtering process enhances the enrichment for T2D association primarily by restricting eSNPs to disease susceptibility gene networks , although a reduced effective FDR may also play a role . Indeed , while we have singled out a single gene , Me1 , as playing a causal role in this network , the true value of the currently described eSNP filtering approach is in its ability to identify disease susceptibility networks rather than single SNPs or genes traditionally identified through GWAS . In fact , the knockout gene expression signature for Me1 was significantly enriched for genes in the T2D adipose causal gene network , providing direct experimental evidence of the high degree of interconnectivity within this network , where perturbing one gene supported as causal for disease affects many other genes supported as causal in this network , as we have previously shown for other disease causal networks [12] , [46] . We have shown for the first time that SNPs that are associated with transcript abundance are more likely to associate with a complex trait as well . This type of approach provides a way to reduce the dimensionality of the DNA variation space and can help us reconsider how to map complex disease using gene expression traits . This approach can also help prioritize GWAS findings , for instance , by including the eSNPs corresponding to genes in causal disease networks in testing for epistasis or for consideration in future genetic association studies . GWAS will continue to deliver high-confidence correlations between DNA changes at a given locus and disease-associated traits of interest . Our understanding of the individual genes at these loci that alter disease susceptibility and the broader context in which they operate can be enhanced by leveraging studies that seek to map the genetics of gene expression . Generating large-scale molecular profiling data sets in both human and experimental segregating populations potentially provides additional power to elucidate not only the genetic basis of disease , but also the impact the genetic basis of disease has on molecular networks that in turn drive physiological states associated with disease . Diabetes pathogenesis involves many pathways operating in different tissues and distinct physiological processes ( blunted insulin signaling and failure of beta cells to compensate by producing more insulin ) . Therefore , the integration of large-scale molecular profiling , genotypic , clinical , and other biologically relevant data will be critical if we hope to understand more fully how genetic and environmental perturbations lead to complex traits like disease . Integration of a diversity of data in this setting will be key , since no single data dimension will provide the complete answer .
For the liver-specific GGE cohort , more than 39 , 000 transcripts were profiled and 782 , 476 unique SNPs were genotyped in more than 400 human liver samples [12] . In this cohort , the genetics of gene expression analysis resulted in the detection of 3 , 309 unique eSNPs at an FDR<10% [14] . The eSNP processing and analysis were carried out as previous described [14] . All expression data have been deposited in the Gene Expression Omnibus database under accession number ( GSE9588 ) [14] . The second multi-tissue GGE cohort was comprised of patients who underwent RXY gastric bypass surgery . Liver , subcutaneous adipose and omental adipose tissues were collected from each patient at the time of surgery at Massachusetts General Hospital . Genomic DNA was extracted from liver tissue for each patient , and total RNA was extracted from liver , subcutaneous adipose and omental adipose tissues . Each RNA sample was profiled on a custom 44K Agilent array . RNA processing methods are detailed in Text S1 . Each DNA sample was genotyped on the Illumina 650Y BeadChip array . We successfully genotyped 950 samples . Identity by state ( IBS ) analysis was performed to identify related individuals within this cohort . Eighteen parent-offspring , 6 sibling and 8 second degree relatives were identified , and 4 of these were related as trios . Twenty-eight individuals were removed to eliminate IBS in the dataset , leaving 922 samples for use in the analysis . Demographic information including age , race , gender , height , type of surgery and year of surgery were collected for each patient ( Text S1 ) . We required that the minor allele frequency for a SNP be >5% in order to be considered in the analyses . Cis and trans acting expression quantitative trait loci ( eQTLs ) were identified using a method similar to that previously described [14] . The cis eQTL for a given expression trait were defined as those with corresponding SNPs located within 1 megabase ( Mb ) of the transcription start or stop of the associated structural gene . All other associations were considered trans . SNP associations were identified using the Kruskal-Wallis test . The association p values were adjusted to control for testing of multiple SNPs and expression traits using an empirically determined FDR constrained to be <10% . For cis eQTL , we only tested for associations to SNPs that were within 1 Mb of the annotated start or stop site of the corresponding structural gene . The empirical FDR permutations were restricted to SNPs within the cis regions . In the case of trans eQTL , all SNPs were tested for association to each of the expression traits . Where SNP associations were identified to the same trait in high LD with each other , the SNP with the most significant p value was reported . When comparing eSNPs identified in independent tissues from the same cohort , 72% of the cis eSNPs identified in liver , 79% of those found in omental adipose and 80 . 5% from subcutaneous adipose were also found in the other two tissues . Of the eSNPs detected , 2 , 189 , 2 , 286 and 1 , 999 were specific eSNPs to liver , omental adipose and subcutaneous adipose , respectively . When compared to the set of liver eSNPs from the first cohort there was a 66% overlap in eSNPs indentified between the two studies . The set of eSNPs used in the paper is the combined set of eSNPs from the four sources , comprising 18 , 785 unique eSNPs in total . 554 F2 mice were generated in a cross between two inbred strains , both containing the ob mutation at the leptin locus: C57BL/6 ob/ob and BTBR ob/ob ( referred to as the B6×BTBR cross ) [27] . All F2 animals were maintained on a chow diet for ten weeks and were clinically characterized with respect to obesity- and diabetes-related traits at four , six , eight and ten week time points . Further details regarding the plasma glucose and insulin measurements , as well as islet isolation procedures , can be found in Keller et al . [29] . At the time of necropsy , gonadal white adipose tissue was collected from 497 mice . RNA was prepared using the same methods as described previously [29] and hybridized to an Agilent custom murine gene expression microarray . Of the 39 , 600 transcribed sequences represented on the microarray , the top 25 percent rank ordered by degree of differential expression in the adipose tissue were included in the reconstruction process [29] . These gene expression traits were used to construct weighted , co-expression subnetworks comprised of the most highly connected nodes from each tissue and sex using previously described methods [30] ( Text S1 ) . QTL were detected for each of the expression and metabolic traits using a forward stepwise regression procedure [55] , [56] . QTL with pleiotropic effects on expression and metabolic traits were identified using a multivariate likelihood test [24] , [57] . The QTL , expression , and metabolic trait data were then integrated to assess whether each expression trait was supported as having a causal relationship with each of the metabolic traits , with respect to QTL detected with pleiotropic effects on the expression and metabolic traits [24] . A naturally occurring mouse mutant deficient in Me1 enzymatic activity was first reported by Lee et al . in 1980 [58] . The detailed methods for breeding , genotyping , and characterization of the Me1−/− mice have been described previously [31] , [59] . Littermate male Me1−/− and wild-type mice were challenged with a high fat diet ( HFD ) starting at 7–8 weeks of age for 19 weeks . An oral glucose tolerance test ( OGTT ) was performed at week 23–24 of age and terminal blood serum was collected at week 26–27 of age . For females , HFD was initiated at week 8–10 and continued for 19 weeks . OGTT was performed at week 26–28 of age and terminal serum samples were collected at week 27–28 of age . Mice were euthanized at the end of the HFD period . For OGTT , glucose was administered at 2g/kg of mouse mass via oral gavage , mice were fasted 18 hours Prior , and glucose levels were measured using a OneTouch Ultra glucometer ( LifeScan , Inc , Milpitas , CA ) at 0 , 30 , 60 , 90 , and 120 min . Serum was collected from blood using Becton Dickson ( Franklin Lakes , NJ ) Microtainer tubes with SST . Insulin and leptin were measured using Millipore's ( Billerica , MA ) Multiplexed Biomarker Immunoassay for Luminex xMap using a Bio-Rad's ( Hercules , CA ) Bio-Plex machine . The other serum parameters were measured using a colorimetric assay . Triglycerides were measured at OD 510 nm using reagents from Roche Diagnostics ( Indianapolis , IN ) . Cholesterol was measured using reagents from Stanbio ( Boerne , TX ) at OD 510 nm as well . The gonadal white adipose tissues were collected from 10 male Me1−/− mice and 10 male littermate wild-type ( wt ) control mice . The detailed methods have been described previously [31] , [59] . The adipose tissues were homogenized and total RNA extracted using Trizol reagent ( Invitrogen , CA ) . Three micrograms of total RNA was reverse transcribed and labeled with either Cy3 or Cy5 fluorochrome . Labeled complementary RNA ( cRNA ) from each animal was hybridized against a pool of labeled cRNAs constructed from equal aliquots of RNA from the control animals using Agilent arrays consisting of 39 , 556 non-control probes that represent 37 , 687 genes . Arrays were quantified on the basis of spot intensity relative to background , adjusted for experimental variation between arrays using average intensity over multiple channels , and fitted to a previously described error model to determine significance [60] , [61] . Gene expression measures are reported as the ratio of the mean log10 intensity ( mlratio ) . A Student's t-test was used to identify genes with significant differences between Me1−/− animals and the corresponding wt control mice . These genes were defined as “signature” genes , representing the perturbed gene expression signature as a result of single gene modification . The significance level was set to p<0 . 05 . The false discovery rate , calculated using Q-value [62] , at this significance level was 20% . | Genome-wide association studies ( GWAS ) seek to identify loci in which changes in DNA are correlated with disease . However , GWAS do not necessarily lead directly to genes associated with disease , and they do not typically inform the broader context in which disease genes operate , thereby providing limited insights into the mechanisms driving disease . One critical task to providing further insights into GWAS is developing an understanding of the genetics of gene expression ( GGE ) . We present the first empiric study demonstrating that SNPs in human cohorts that associate with gene expression in liver and adipose tissues are enriched for associating with Type 2 Diabetes ( T2D ) in humans . By filtering “eSNPs” based on causal gene networks defined in an experimental cross population segregating T2D traits , we demonstrate a dramatically increased enrichment of T2D SNPs that enhance our ability to assess T2D risk . We demonstrate the utility of this approach by identifying malic enzyme 1 ( ME1 ) as a novel T2D susceptibility gene in humans and then functionally validating the causal connection between ME1 and T2D in a mouse knockout model for Me1 . This approach provides a path to identifying disease susceptibility networks rather than single SNPs or genes traditionally identified through GWAS . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/disease",
"models",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/complex",
"traits",
"genetics",
"and",
"genomics/gene",
"expression"
] | 2010 | Liver and Adipose Expression Associated SNPs Are Enriched for Association to Type 2 Diabetes |
The influence of mono-ubiquitylation of histone H2B ( H2Bub ) on transcription via nucleosome reassembly has been widely documented . Recently , it has also been shown that H2Bub promotes recovery from replication stress; however , the underling molecular mechanism remains unclear . Here , we show that H2B ubiquitylation coordinates activation of the intra-S replication checkpoint and chromatin re-assembly , in order to limit fork progression and DNA damage in the presence of replication stress . In particular , we show that the absence of H2Bub affects replication dynamics ( enhanced fork progression and reduced origin firing ) , leading to γH2A accumulation and increased hydroxyurea sensitivity . Further genetic analysis indicates a role for H2Bub in transducing Rad53 phosphorylation . Concomitantly , we found that a change in replication dynamics is not due to a change in dNTP level , but is mediated by reduced Rad53 activation and destabilization of the RecQ helicase Sgs1 at the fork . Furthermore , we demonstrate that H2Bub facilitates the dissociation of the histone chaperone Asf1 from Rad53 , and nucleosome reassembly behind the fork is compromised in cells lacking H2Bub . Taken together , these results indicate that the regulation of H2B ubiquitylation is a key event in the maintenance of genome stability , through coordination of intra-S checkpoint activation , chromatin assembly and replication fork progression .
Recent evidence suggests that histone modifications can affect DNA replication , under both normal or stressed conditions , through effects on nucleosome dynamics and protein recruitment [1]–[3] . One such modification is acetylation of nascent histone H3 at lysine 56 ( H3K56Ac ) , which is regulated by the Asf1 histone chaperone and the Rtt109 acetyltransferase during the cell cycle [4] , [5] . Regulation of this modification is important for DNA replication , as failure to deacetylate H3K56Ac results in impaired S phase progression [6] , sensitivity to replication stress [7] , and spontaneous DNA damage [6] . H3K56Ac appears to facilitate nucleosome reassembly on daughter strands during S phase [8] . Acetylation of the N terminal lysines of histone H3 by Gcn5 also contributes to nucleosome assembly during DNA replication [9] . These findings suggest that replication-coupled nucleosome assembly may impact on both fork progression and the stability of stalled forks [1] , [3] . A second histone , H2B , is mono-ubiquitylated at lysine 123 ( K123 , K120 in human ) by the E2 enzyme Rad6 and the E3 enzyme Bre1 in Saccharomyces cerevisiae [10]–[13] . Mono-ubiquitylation of H2B ( H2Bub ) is best characterized in terms of its effects on transcriptional regulation in budding yeast [14] , [15] , which are mediated through downstream methylation of lysines 4 and 79 of H3 [16]–[19] . In addition , H2Bub has been demonstrated to affect transcription independently of its regulation of H3 methylation [20] , [21] . H2Bub enhances passage of RNA Polymerase II during transcription elongation by mediating nucleosome reassembly in both yeast and human [20] , [22] , [23] . Furthermore , H2Bub may also affect transcription and DNA repair through influencing chromatin structure [24] , [25] . It has been suggested that H2Bub mediates homologous recombination repair at DNA double-strand break ( DSB ) sites through relaxing chromatin structure in human cells [26] , [27] . H2Bub has also been shown to maintain replication fork stability by promoting replication-associated nucleosome formation in budding yeast , independently of its role in regulating H3K4 and K79 methylation [28] . During S phase , replication fork progression can be impaired by low dNTP pools or by DNA damage . Under these conditions , a sensor-response system activates the DNA replication ( intra-S phase ) checkpoint , which prevents fork collapse while controlling origin firing [29] , [30] . The mechanism by which the intra-S checkpoint is activated is still not yet fully understood . It is hypothesized that decoupling between polymerase and helicase leads to single strand DNA accumulation and activation of the kinases Mec1/ATR and their downstream effector , Rad53 [30] , [31] . Once a stalled fork has been stabilized by activation of the intra-S checkpoint , the damaged fork can resume DNA synthesis . The RecQ helicase Sgs1 is recruited to the stalled fork , where it facilitates its re-initiation through a mechanism involving the recombination repair pathway [32] . Sgs1 also facilitates the phosphorylation of Rad53 ( possibly through direct physical interaction ) , and this process is redundant with the DNA damage checkpoint proteins Rad24 and Esc2 [33] . Activation of the intra-S phase checkpoint affects DNA synthesis by altering both the rate of replication fork progression and the rate of DNA replication initiation events [34] . For instance , a recent report suggests that Mec1 promotes chromatin accessibility at or ahead of replication forks via a mechanism independent of its checkpoint role [35] . The authors argue that a chromatin regulatory process may serve as a means of restricting fork progression , in order to control and stabilize fork progression under replication stress . However , the mechanisms through which chromatin structure regulates replication progression are still poorly understood . In the current study , we used BrdU IP-chip to examine genome-wide DNA synthesis incorporation in wild type and H2Bub-deficient cells in the presence of hydroxyurea ( HU ) . We demonstrate that newly-synthesized DNA in cells lacking H2Bub displays a broader distribution and enrichment at origin-distal regions; these findings suggest faster replication fork progression in the mutant . Surprisingly , this phenomenon is independent of DNA damage-induced dNTPs , and is accompanied by delayed Rad53 activation and defective chromatin assembly . All of these effects contribute to replication fork instability and reduced cell viability under replication stress . Our data indicate that H2Bub is one of the limiting factors that regulate replication fork progression , and maintain fork stability in the presence of HU-induced stress .
The presence of 200 mM HU increased lethality in mutant cells lacking H2Bub ( htb-K123R mutant ) ( Fig . 1A ) , confirming the previously-hypothesized role of H2Bub in maintaining fork stability [28] . High doses of HU , an inhibitor of the ribonucleotide reductase , leads to a strong decrease in dNTP pools , that in turn leads to a decrease in replication speed and intra-S checkpoint activation [36] , [37] . In order to investigate the mechanisms by which H2Bub sustains cell viability during replication stress , we examined genome-wide origin firing and replication fork progression under HU in wild type and in htb-K123R cells . This was achieved by performing BrdU immunoprecipitation followed by hybridization on a high density oligonucleotide array . Wild-type ( WT ) or H2Bub-deficient mutant ( htb-K123R ) cells were pre-synchronized in G1 with α-factor ( Fig . 1B ) and then released into fresh media containing HU and BrdU . Under such conditions , BrdU is incorporated at active origins ( such as ARS305 and ARS607 ) , and BrdU track length correlates with the replication fork progression . Positions of ARS elements were identified by Mcm2 occupancy [38]; therefore , this assay can be used to monitor origin usage and replication fork progression on a genomic scale [37] . An unexpected finding was that the BrdU track lengths at most origins were significantly longer in htb-K123R cells ( average 12 . 73 kb ) than in WT cells ( average 8 . 17 kb; Fig . 1C–F and Fig . S1 ) , indicating extended progression of replication forks . This finding was corroborated by the observation that DNA content in the mutant was greater than in WT , as evidenced by FACS ( Fig . 1B ) . In addition , despite the semi-quantitative aspect of this technique , BrdU incorporation peaks were clearly reduced at the majority of firing origins in the mutant . This may be indicative of a decrease in origin firing . Taken together , these results suggest that replication fork stalling is reduced in the htb-K123R mutant during HU-induced stress , and this may lead to fork destabilization . It was previously reported that yeast cells with persistently-enlarged dNTP pools are prone to DNA damage [39] , and exhibit enhanced fork progression [37] , [40] under replication stress . Thus , it is possible that the enhanced HU sensitivity and increased fork progression in htb-K123R cells may be a consequence of enlarged dNTP pools; this in turn may be a direct consequence of ( i ) increased transcription of ribonucleotide reductase ( RNR ) genes or ( ii ) spontaneous DNA damage , or otherwise via an indirect mechanism that stimulates ribonucleotide production . To test this hypothesis , we directly examined the size of dNTP pools in htb-K123R cells . We observed that the dNTP concentration in htb-K123R cells is ∼40% greater than that of WT cells ( shown for four biological replicates in Fig . S2 ) . The cellular concentration of dNTP pools is regulated by the Rad53-Dun1 pathway during both normal and perturbed cell cycles , through multiple mechanisms [41] , [42] . Deletion of DUN1 stabilizes the RNR inhibitor Sml1 and decreases the size of the cellular dNTP pool , while deletion of SML1 increases pool size [43] . To further investigate whether the effect of H2Bub on fork stalling during replication stress is dependent on the concentration of dNTP pools , we deleted the DUN1 gene from WT and htb-K123R cells . Notably , the dNTP pools of both dun1Δ and dun1Δ htb-K123R were ∼50% the size of those in WT cells ( Fig . 2A ) . Since deletion of DUN1 suppressed the increase of dNTP in H2Bub mutants , we can conclude that the increase of dNTP level in H2Bub mutants is mediated by Dun1 . We next performed BrdU-IP chip experiments to determine BrdU track length in these cells . As expected , the BrdU track length of the dun1Δ cells was significantly shorter ( 6 . 57 kb; Fig . 2B & C ) than that in WT cells , probably due to the reduced concentration of dNTP [37] , [39] . Surprisingly , fork progression in dun1Δ htb-K123R was significantly faster than in WT cells ( 9 . 86 kb vs . 8 . 81 kb; Fig . 2B & C ) , despite the reduced size of the dNTP pool ( Fig . 2A ) ; this finding indicates that the increase in fork progression and instability in this mutant does not arise solely from the increase in the dNTP pool . In addition , we observed that deletion of DUN1 , but not SML1 , increased the sensitivity of htb-K123R cells to chronic HU exposure ( Fig . 2D ) . Therefore , we conclude that H2Bub has a role in controlling fork progression and cell survival in response to replication stress , which is independent of the Dun1-mediated regulation of ribonucleotide production . Origin firing and fork progression have been reported to be strongly co-regulated by cells in order to ensure normal completion of replication . In particular , an increase in replication speed leads to a decrease in origin firing [37] . Our BrdU immunoprecipitation and chip hybridization data are consistent with this reported tendency ( Fig . 1 ) . However , since this technique is only partially quantitative , we decided to confirm this observation using two-dimensional ( 2D ) gel analysis ( Fig . 3A ) . Replication intermediates migrate differently depending on their molecular weight and sterical conformation . In particular , the bubble arc reflects origin firing . Interestingly , both WT and H2Bub mutant exhibit similar replication kinetics at two early origins ( ARS305 and ARS607 ) ; replication intermediates appear one hour after alpha factor release in agreement with origin firing , and start to decrease after 2 hours , reflecting fork progression outside of the restriction fragment . However , we observed a strong reduction of replication intermediates in the H2Bub mutant compared to WT , likely due to a decrease in origin efficiency . To further delineate the role of H2Bub in origin firing , we measured incorporation of BrdU into chromatin , using BrdU-IP combined with quantitative-PCR . This experiment was performed at 20°C to slow down DNA replication . We found that replication efficiency at ARS305 and ARS607 was much lower in mutant than in WT cells . We did not observe DNA synthesis at a telomeric region ( TEL VI ) in either WT or the mutant , owing to the late onset of DNA replication at telomere . These data strongly suggest that origin firing is inefficient in cells lacking H2Bub ( Fig . S3 ) . We subsequently hypothesized that the change in replication dynamics in cells lacking H2Bub may affect the integrity of the fork , as previously observed [37] . In particular , γH2A accumulation in the H2Bub mutant confirmed the accumulation of damage in the absence of H2B ubiquitylation ( Fig . 3B ) . This accumulation may explain the hypersensitivity to hydroxyurea that we and others [28] have observed ( Fig . 3C ) . Stalled forks are detected by the intra-S phase checkpoint [30] . We reasoned that the instability of replication forks in htb-K123R cells may result from a defect in the activation of the intra-S phase checkpoint [34] . As such , we examined whether H2Bub interacts with factors that stabilize the replication fork during replication stress , by systematically examining the genetic interactions between htb-K123R and mutations in key components of this complex signaling system . Initially , we examined a hypomorphic allele of pol2-11 , which encodes a mutant form of Polε that causes defects in the intra-S phase checkpoint [44] . The htb-K123R and pol2-11 double mutant exhibited synthetic growth defects at the permissive temperature ( 23°C ) ( Fig . 4A , top left panel ) . This interaction was confirmed to be specific , because double mutants of htb-K123R and pol1-17 , pol3-14 , or pri2-1 ( replication defective mutants of DNA polymerase α and δ , and RNA primase , respectively [45] ) , exhibited subtle additive growth defects , or sensitivity to 50 mM HU at both the permissive ( 23°C; Fig . S4A ) and non-permissive ( 30°C; Fig . S4B ) temperature for growth . We next examined the effect of HU on strains containing htb-K123R and mec1-100 , an intra-S phase checkpoint defective allele of Mec1/ATR [46] , or deletion of MRC1 or SGS1 ( Fig . 4A ) . Single mutants of htb-K123R and mec1-100 grew in the presence of 10 and 25 mM HU , but the double mutant was highly sensitive to these concentrations of HU ( Fig . 4A , top right panel ) . Interestingly , similar phenotypes were observed upon combining htb-K123R with deletions of the genes encoding the checkpoint mediator protein Mrc1 or the RecQ helicase Sgs1 ( Fig . 4A , bottom left panel ) , suggesting that H2Bub stabilizes the replication fork independently of these proteins . We then examined whether H2Bub interacts with the kinase checkpoint effector , Rad53 . The rad53-11 mutant is checkpoint defective , with undetectable Rad53 activity [47] . Intriguingly , the hypersensitivity of rad53-11 to HU was partially reversed by htb-K123R ( Fig . 4A , bottom-right panel ) . Deletion of the H2Bub-specific E3 ligase Bre1 had similar effects to htb-K123R when combined with the mec1-100 , sgs1Δ , or rad53-11 mutations ( Fig . 4B , S4C ) , suggesting that the genetic interactions between H2Bub and components required for the re-initiation of stalled forks are at the level of chromatin structure , and are linked to its chromatin modifying activities . We next examined the effect of H2Bub deficiency on the viability of mec1-100 and rad53-11 cells in the presence of HU . The absence of H2Bub exacerbated the lethality observed in mec1-100 cells in S phase , while enhancing the viability of rad53-11 cells under the same conditions ( Fig . 4C ) . Overall , our genetic analyses suggest that H2Bub and the Mec1-dependent S-phase checkpoint function in parallel to preserve fork stability under replication stress . However , our finding that htb-K123R alleviates the rad53-11 growth defect under HU suggest that that H2Bub may function upstream of Rad53 and participate in the replication stress response . Comparing our data with earlier works [37] , [48] , [49] revealed several lines of evidence which suggest that H2Bub and the RecQ helicase Sgs1 have overlapping functions in maintaining fork stability under HU . First , both htb-K123R and sgs1Δ cells exhibit increased fork progression in HU ( Fig . 1F; [37] ) . Second , the absence of either Sgs1 or H2Bub reduces the stability of stalled replication forks under HU ( Fig . 1A and 3C; [48] ) . Third , the combination of htb-K123R or sgs1Δ with mec1-100 causes fork collapse and failure to recover from acute exposure to HU ( Fig . 4C; [49] ) . To better delineate the role of H2Bub in the replication stress response , we decided to investigate the interaction between H2Bub and the Sgs1 helicase further . We used ChIP to measure the recruitment of Sgs1 to the ARS305 and ARS607 early origins in WT and htb-K123R cells ( Fig . 5A ) . While Sgs1 was initially recruited efficiently to ARS305 and ARS607 in both strains , it failed to accumulate at ARS in the mutant , suggesting the association of Sgs1 with the replication fork was unstable in the absence of H2Bub ( Fig . 5A ) . Furthermore , HU-induced phosphorylation of Rad53 was unaffected in sgs1Δ , but delayed in htb-K123R cells ( Fig 5B ) . Rad53 activation is facilitated by the retention of Sgs1 at stalled forks [33] , and the current results suggest that H2Bub may be required for such retention , and thus Rad53 phosphorylation . Interestingly , a recent study demonstrated that RPA-coated single-stranded DNA replication intermediates ( ssDNA ) are reduced at initiated origins in htb-K123R cells under HU [28] . RPA is postulated to interact with Sgs1 at replication forks [50] . Thus , the reduced Sgs1 occupancy at replication forks and delayed Rad53 phosphorylation in the htb-K123R mutant may be explained by the decreased amount of ssDNA at replication forks . In addition , Rad53 phosphorylation was significantly impaired in a sgs1Δ htb-K123R double mutant ( Fig . 5B ) . This is also indicative of a Sgs1-independent role for H2Bub in Rad53 activation . Collectively , these results point to a functional role for H2B in replication and the checkpoint response , and are consistent with the observed epistatic interaction between H2Bub and Rad53 ( Fig . 4A–C ) . To further elucidate the interaction between Sgs1 and H2Bub , we used BrdU IP-chip to monitor fork progression in sgs1Δ and sgs1Δ htb-K123R mutant cells in the presence of HU . Consistent with a previous report [37] , the average BrdU track length in sgs1Δ cells was significantly increased as compared to WT cells ( 11 . 35 kb vs . 8 . 17 kb , respectively; Fig . 6A and 6B ) , similar to the increase observed in htb-K123R cells ( Fig . 1F ) . Remarkably , track lengths in the sgs1Δ htb-K123R double mutant ( 20 . 36 kb ) were even greater , being almost 2 . 5-fold longer than those in WT ( 8 . 17 kb; Fig . 6A and 6B ) . Flow cytometry was used to confirm that the double mutant contained greater amounts of DNA in the presence of 200 mM HU ( Fig . 6C ) . Increased fork progression in the absence of Sgs1 is believed to be a consequence of dNTP accumulation [37] . We confirmed that the dNTP concentration was increased in sgs1Δ ( ∼2 . 3 fold as compared to WT; Fig . 6D ) , but no additional increase was observed in the sgs1Δ htb-K123R double mutant , suggesting that elevated dNTP production does not underlie the defect in fork stalling in the double mutant . We have demonstrated that the replication fork becomes unstable and vulnerable to replication stress in H2Bub-deficient cells ( Fig . 1A & 3B ) , which could be due to continuous DNA synthesis under conditions of dNTP depletion . Thus , we reasoned that the rapidly moving replication fork may become highly unstable in the absence of both H2Bub and Sgs1 , a hypothesis supported by the observation that the double mutant was more sensitive to acute treatment with HU than either single mutant ( Fig . 6E ) . Taken together , our results so far suggest that H2Bub is involved in stalling the replication fork and maintaining its stability in response to HU-induced S phase block; furthermore , this function is performed in cooperation with Rad53 kinase activity and in parallel with Mec1 and Sgs1 during S phase . H2Bub has been shown to be required for nucleosome reassembly during RNA Polymerase II elongation [20] , [23] and DNA replication [28] . We therefore hypothesized that defective fork stalling in htb-K123R cells under replication stress may be a consequence of incomplete nucleosome assembly . We first confirmed that histone assembly on newly-synthesized DNA is defective in htb-K123R under HU . In WT cells , histone H3 was associated with the early firing origins ARS305 and ARS607 at all times post-G1 release into HU . H3 occupancy at these early origins was reduced upon entry into S phase in htb-K123R cells , but occupancy at the late origin ARS501 was unaffected ( Fig . 7A ) . These data suggest that in the absence of H2Bub , histone assembly is less efficient at firing origins . Mec1 was recently reported to increase chromatin accessibility at or ahead of replication forks , and promote fork progression in HU [35] . Thus , the mechanism promoting nucleosome assembly during DNA replication may inhibit fork progression under replication stress . We reasoned that if this were the case , deletion of genes encoding proteins involved in replication-coupled histone assembly ( such as the histone chaperones CAF-1 and Asf1 ) should sensitize htb-K123R cells to replication stress . Asf1 has dual roles; it associates with the RCF complex and MCM helicase and facilitates nucleosome disassembly during replication [51] , [52] , and it assists acetylation of H3 lysine 56 ( H3K56ac ) by presenting newly-synthesized H3/H4 dimers to the Rtt109 acetyltransferase [4] , [53] . Acetylation increases the affinity of H3 for CAF-1 [53] and promotes efficient chromatin assembly onto nascent DNA [8] . As a control , we deleted Hir1; this protein is implicated in replication-independent H3/H4 deposition [54] . We found that deletion of ASF1 or RTT109 greatly increased the HU sensitivity of htb-K123R cells , but deletion of CAC1 ( the largest subunit of CAF-1 ) or HIR1 had no such effect ( Fig . 7B ) . This suggests that H2Bub and Asf1-Rtt109 function synergistically to promote cell survival during replication stress . Moreover , deletion of Asf1 increased the sensitivity of htb-K123R cells to acute HU treatment ( Fig . 7C ) , suggesting that the stability of the replication fork was decreased further . Rad53 and Asf1 form a dynamic complex that dissociates in response to Rad53 phosphorylation under replication stress . Rad53 acts as a regulator of Asf1 availability and indirectly controls its chromatin assembly activity [55] , [56] . Our results suggest that H2Bub may affect Rad53 phosphorylation ( Fig . 5B ) . We hypothesized that H2Bub may contribute to nucleosome assembly by influencing the dynamic association between Asf1 and Rad53 , in addition to possessing a direct role in nucleosome assembly . To test this hypothesis , we tagged the genomic ASF1 gene with a triple HA tag , thereby enabling immunoprecipitation of Asf1 with an anti-HA antibody ( Fig . 7D , lanes 2 , 4 , 6 , and 8 ) . Rad53 co-precipitated with HA-tagged Asf1 efficiently in WT lysates ( Fig . 7D , lane 2 ) but not with un-tagged Asf1 ( Fig . 7D , lanes 1 , 3 , 5 , and 7 ) . Consistent with previously published results [55] , [56] , Rad53 association with Asf1 was reduced in the presence of HU in WT cells ( Fig . 7D , lane 6 ) . However , the association of Rad53 with Asf1 remained stable in htb-K123R cells in the presence of HU ( Fig . 7D , compare lanes 6 and 8 ) . These results suggest that H2Bub coordinates nucleosome assembly in response to replication stress by directly contributing to nucleosome formation , and by indirectly regulating the availability of Asf1 , which in turn deposits histones behind the advancing replication fork ( Fig . 7E ) .
How does the Bre1-H2Bub pathway modulate the cellular response to HU-induced replication block ? Interestingly , H2B ubiquitylation has been proposed to promote unwinding of the DNA chromatin complex ahead of the replication fork , and thereby stimulate fork progression in HU [28] . However , our results support an alternative role for H2Bub in restricting replication fork progression under conditions of HU stress . We found that histone occupancy around early origins in htb-K123R cells is reduced upon S phase entry in the presence of HU ( Fig . 7A ) . In addition , we also showed that the removal of both Asf1-Rtt109 and H2Bub synthetically increases the sensitivity to replication stress ( Fig . 7B and C ) . Moreover , we provide evidence that H2Bub controls the availability of Asf1 during replication stress ( Fig . 7D ) , which potentially contributes to histone deposition behind the advancing replication fork [55] , [56] . Thus , we propose that enhanced chromatin assembly on nascent DNA during replication stress may facilitate replication fork stalling in response to nucleotide depletion imposed by HU ( Fig . 8 ) , akin to the brakes on a train ( the replisome ) . Mec1 slows S phase progression by delaying late origin firing through intra-S phase checkpoint activation [30] , [34] , but it also promotes sustained replication fork progression at early origins [35] . The chromatin state seems to facilitate tight regulation of fork progression at early origins during replication stress . We cannot exclude the possibility that continuous DNA synthesis in the htb-K123R mutant may reflect the movement of DNA polymerase through inappropriately-assembled chromatin . In this scenario , chromatin structure at or ahead of the fork would be altered in the absence of H2Bub due to defective chromatin assembly during the previous round of replication . Mec1 and Bre1-H2Bub may have antagonistic effects on chromatin dynamics ahead of the replication fork; thus in the absence of H2Bub , forks may be inclined to move faster because of Mec1-induced chromatin accessibility [35] . Although the scenario outlined above is formally possible , our molecular and genetic analyses favor a second model in which nucleosome formation on nascent DNA serves as a negative feedback mechanism to regulate the progression of the replication fork under stress . Thus , we suggest that Mec1-mediated signaling and the Bre1-H2Bub pathway synergistically interact to ensure that replisomes travel in a controlled manner , thereby maintaining fork stability under replication stress . Checkpoint kinases Mec1 and Rad53 are essential for the maintenance of cell viability when replication is perturbed [30] , [31] . Our genetic analyses reveal the unexpected finding that H2Bub maintains fork stability in parallel with the Mec1-mediated intra-S checkpoint , but its effect is epistatic to that of a second checkpoint kinase , Rad53 . Our results support the hypothesis that Rad53 stabilizes replication forks independently of Mec1 [57] , [58] . Furthermore , our findings suggest a possible mechanism for the role of H2Bub in Rad53 activation ( Fig . 8 upper panel ) . We report that the stable association of Sgs1 with the replication fork is not only replication-dependent [48] , but also H2Bub-dependent ( Fig . 5A ) . It was previously demonstrated that Sgs1 helps recruit Rad53 to stalled forks via an interaction with RPA [50] . Intriguingly , it has been postulated that fork collapse followed by origin firing in yeast cells lacking H2Bub results in reduced levels of single-stranded DNA ( ssDNA ) and RPA during a G1 to HU shift [28] , consistent with our observation of reduced replication intermediates and increased DNA damage in htb-K123R cells ( Fig . 3 ) . However , it is also possible that the failure to accumulate RPA in htb-K123R cells may be caused by an increase in the rate of nascent DNA synthesis , thereby reducing the accumulation of ssDNA at stalled forks; this in turn reduces Sgs1 retention and delays Rad53 phosphorylation ( Fig . 8 , bottom panel ) . The reduced activity of Rad53 may have a negative feedback effect , thereby further compromising fork stability . The absence of both H2Bub and Sgs1 therefore further disrupts Rad53 activation and fork integrity . In support of our model that chromatin assembly serves as a negative feedback signal to regulate the progression of replication forks , several reports in budding yeast have established that chromatin assembly at replication forks is necessary to stabilize replication forks and prevent their collapse [52] , [59] , [60] . A recent report in mammals demonstrated that replication fork speed is dependent on the supply of new histones and efficient nucleosome assembly during an unperturbed cell cycle [61] . Human Asf1 has been shown to associate with the MCM complex through histone H3/H4 dimers [51] . In addition , Asf1 extracted from human cells exposed to HU exhibits an enhanced ability to assemble chromatin [62] . Thus , there may be two pools of Asf1 in cells . One is coupled to replication forks , while the other is sequestered by Rad53 . Replication stress triggers the release of the sequestered pool of Asf1 ( which occurs at least in part through H2Bub ) to promote chromatin formation ( Fig . 7E ) and restrict fork progression ( Fig . 1C–F ) under replication stress . However , defects in nucleosome assembly mediated by CAF-1 trigger DNA damage checkpoint activation and delay fork progression in human cells during an unperturbed cell cycle [63]–[65] . Our genetic analysis shows that Cac1 , unlike H2Bub and Asf1 , is not required by yeast cells to maintain growth under HU stress; hence chromatin assembly regulated by H2Bub and Asf1 under replication stress ( Fig . 7B ) likely occurs through pathways distinct from those mediated by CAF-1 [66] . In summary , we have provided evidence that H2Bub coordinates chromatin assembly and Rad53 activation during HU stress in parallel with other mechanisms that maintain fork stalling and stability during replication stress , including the intra-S phase checkpoint and the Sgs1 helicase . Our data indicate that H2Bub maintains genomic stability by creating an environment that integrates chromatin formation and checkpoint kinase activation , thereby maintaining stable replication and facilitating recovery from replication stress in concert with other components that mediate faithful DNA replication .
Yeast strains and plasmids used in this study are shown in supplementary tables S1 and S2 . All yeast cells were cultured in yeast extract peptone supplemented with 2% dextrose at 30°C . All analyses were performed during the log phase of growth . Cells were arrested in G1 by the addition of α-factor to a final concentration of 100 ng/ml ( bar1Δ strain ) for at least 3 hours ( the exact time differed depending on the strain ) . Cells were released from G1 arrest by washing with sterilized H2O three times , before being re-suspended in fresh media containing hydroxyurea ( HU; Sigma ) . For phenotypic screening , mid-log ( 0 . 4–0 . 8 ) phase cultures were collected and counted . Ten-fold serial dilutions were spotted onto YPD plates containing different doses of HU . Plates were subsequently incubated at 30°C for several days . Two different strain backgrounds were used in this study . With the exception of the strains used in the genetic analysis shown in Fig . 4 , all strains were in the YS131 background . The YS131 parental strain is derived from W303 , but both genomic copies of HTA1-HTB1 and HTA2-HTB2 are deleted , and cell viability is maintained by a plasmid-derived HTA1-HTB1 or HTA1-htb1-K123R . Earlier studies established that deletion of HTA2-HTB2 has negligible effects on mitotic growth and stress responses , and that the HTA1-HTB1 gene pair can compensate for the absence of the HTA2-HTB2 [67] , [68] . Therefore , we predict that hta2-htb2Δ would not affect the htb-K123R mutation . For the genetic analysis with the checkpoint mutants , we were conscious of an earlier report that the rad53 mutant is sensitive to histone dosage [69] . To prevent unexpected pleiotropic effects , we introduced genomic htb-K123R mutations [HTA1-htb1-K123R::NAT+ HTA2-htb2-K123R::HIS+] [70] into mec1-100 and rad53-11 for genetic analysis . We also compared the HU sensitivity of the htb-K123R mutants in both strain backgrounds to ensure that they give rise to the same replication defects , as shown in Fig . S5 . For gene disruptions , the indicated gene was deleted by high efficiency transformation , using a PCR product in which the target was replaced with the KanMX gene ( deletion library from SGD ) . The mutant alleles , pol1-17 [45] , pri2-1 [45] , pol2-11 [71] and pol3-14 [45] , were introduced into strain CFK1204 or CFK1231 through the gene replacement technique of Scherer and Davis [72] , thereby generating ts mutants . The plasmid used for gene replacement consisted of a 9-kb pol1 ( Ts ) , 3 . 3-kb pri2 ( Ts ) , 13-kb pol2 ( Ts ) , or 4-kb pol3 ( Ts ) fragment cloned into the XhoI site of YlP , HpaI site of YlPA16 , AgeI site of pRS306 , or KpnI site of pMJ14 . S . cerevisiae strains were designed in order to allow BrdU incorporation ( TK repeats ) ( Tables S1 and S2 ) . S . cerevisiae oligonucleotide microarrays were obtained from Affymetrix . BrdU-IP chip analysis was carried out as previously described [73] , [74] . Briefly , cells were synchronized with α-factor and then released into fresh YPD containing 0 . 2M HU and 200 µg/ml BrdU for 90 minutes . The collected cells were arrested in ice-cold buffer containing 0 . 1% Na-azide , and genomic DNA was extracted from 2×109 cells as described in the “QIAGEN Genomic DNA Handbook” . DNA was sheared to 300 bp by sonication , denatured , and mixed with 4 µg anti-BrdU monoclonal antibody ( MBL M1-11-3 ) as previously described [75] , [76] . Antibody-bound and unbound fractions were subsequently purified , and then amplified using the WGA2 GenomePlex Complete Genome Amplification Kit . A total of 2 µg of amplified DNA was digested with DNaseI to a mean size of 100 bp; the fragments were subsequently end-labeled with biotin-N11-ddATP [77] , and hybridized to the DNA chip . For DNA content analysis , approximately 1×107 cells were collected at each time point , and resuspended in 1 ml 70% ethanol ( ice-cold ) , before being stored at −80°C for at least one night ( samples were stored up to a maximum of 3 days ) . The cells were then washed twice with 1 ml 50 mM Tris-HCl ( pH 8 . 0 ) followed by RNAase A digestion ( 1 mg ml−1 of RNAase A in 50 mM Tris-Cl , pH 8 . 0 ) and proteinase K digestion ( 16 units ml−1 in 30 mM Tris-Cl , pH 8 . 0 ) . Finally , cells were stained with SYBR GREEN I buffer ( in 50 mM Tris-Cl , pH 8 . 0 ) at 4°C overnight . The cell size and DNA contents of 50 , 000 cells were examined on a FACSCanto II ( BD ) . Total genomic DNA was extracted according to the protocol of the QIAGEN Genomic DNA Handbook , using genomic-tip 100/G columns . 2D gel electrophoresis was carried out as originally described by Brewer and Fangman [78] . The DNA samples were digested with HindIII or SacI/ApaL1 , for ARS305 and ARS607 detection respectively , and then blotted onto a Nylon Gene Screen Plus membrane ( NEN ) . Membranes were probed with the BamHI-NcoI 3 . 0 kb fragment which spans ARS305 and was purified from plasmid A6C-110 ( kindly provided by C . Newlon , uMDNJ , Newark , NJ ) , or probed with a 3 . 0 kb PCR product that spans ARS607 . Signals were detected using a PhosphorImager Typhoon FLA 7000 ( GE Healthcare ) . To determine viability in response to acute doses of HU , cells were grown in culture media until they reached log phase . The cells were then arrested in G1 for 3 hours by the addition of α-factor , before being released into rich media containing 200 mM HU . Aliquots were removed from each culture at the indicated time point , plated onto YPD plates , and allowed to grow at 30°C for 2–3 days . Viability was estimated based on colony forming unit ( CFU ) counts , and was adjusted to that of wild-type at each time point . Yeast cell lysates were prepared using the TCA method [79] . Briefly , equivalent numbers of cells ( 1 . 5×108 ) were collected , resuspended in 200 µl TCA buffer ( 1 . 85 M NaOH and 7 . 4% β-mercaptoethanol ) , and placed on ice for 10 minutes . Following the addition of 200 µl of 20% TCA , the lysates were incubated on ice for 10 minutes . Pellets were subsequently collected , washed with 1 ml acetone , dried , and dissolved in 200 µl 0 . 1 M NaOH . The concentration of each sample was determined , and equal amounts were separated by SDS-PAGE , before being transferred to PVDF membranes for immunoblotting . The following antibodies were used: anti-GAPDH ( Sigma ) , anti-Flag ( Sigma ) and anti-phospho-Rad53 ( produced and characterized by A . Pellicioli and the IFOM antibody facility , and kindly provided by Dr . Foiani [80] ) . Secondary antibodies conjugated to horseradish peroxidase were detected using enhanced chemiluminescence ( Amersham Biosciences ) . The dNTP pools were analyzed as described by a recent study [81] . At a density from 0 . 4 to 0 . 8×107 cells/ml , ∼3 . 7×108 cells were collected onto a 0 . 8 µm nitrocellulose filter ( Millipore AB , Solna , Sweden ) . The filters were immersed in 700 ml of ice-cold extraction solution ( 12% w/v trichloroacetic acid , 15 mM MgCl2 ) in Eppendorf tubes . The following steps were carried out at 4°C . The tubes were vortexed for 30 s , incubated for 15 min , and vortexed again for 30 s . The filters were removed , and the solutions were centrifuged at 20 , 000× g for 1 min . After centrifugation , 700 ml of supernatant was added to 800 ml of ice-cold Freon–trioctylamine mixture [10 ml of 99% Freon ( 1 , 1 , 2-trichlorotrifluoroethane; Aldrich , Sigma-Aldrich Sweden AB , Stockholm , Sweden ) ] , and 2 . 8 ml of>99% trioctylamine ( Fluka , Sigma-Aldrich Sweden AB , Stockholm , Sweden ) . The samples were vortexed and centrifuged for 1 min at 20 , 000× g . The aqueous phase was collected and added to 700 ml of an ice-cold Freon–trioctylamine mixture . Aliquots ( 475 and 47 . 5 ml ) of the resulting aqueous phase were collected . The 475 ml aliquots were pH adjusted with 1M NH4HCO3 ( pH 8 . 9 ) , loaded onto boronate columns [Affi-Gel 601 ( Bio-Rad ) ] , and eluted with 50 mM NH4HCO3 , pH 8 . 9 , 15 mM MgCl2 to separate dNTPs and NTPs . The eluates with purified dNTPs were adjusted to pH 3 . 4 with 6M HCl , and separated on a Partisphere SAX-5 HPLC column ( 125 mm×4 . 6 mm , 5 µm , Hichrom , UK ) using the Hitachi HPLC EZChrom system . Nucleotides were isocratically eluted using 0 . 36M ammonium phosphate buffer ( pH 3 . 4 , 2 . 5% v/v acetonitrile ) . The 47 . 5 ml aliquots were adjusted to pH 3 . 4 and used to quantify NTPs by HPLC in the same way as dNTPs . The nucleotides were quantified by measuring the peak heights and comparing them to a standard curve . Yeast strains were grown to an OD600 of 0 . 4–0 . 8 , and fixed with 1% formaldehyde at room temperature ( RT ) for 20 min . Fixation was stopped by the addition of glycine to a final concentration of 125 mM for 5 min , and the cells were then collected and washed twice with ice-cold TBS ( 100 mM Tris at pH 7 . 5 , 0 . 9% NaCl ) . Cell pellets were stored at −80°C or resuspended immediately in 500 µl of FA lysis buffer ( 50 mM HEPES , pH 7 . 5 , 140 mM NaCl , 1 mM EDTA , 1% sodium deoxycholate , 0 . 1% SDS ) supplemented with fresh protease inhibitor cocktail ( Sigma ) , and lysed by vortexing with glass beads for 30 min at 4°C . Cell lysates were sonicated in a cooling water bath four times for 10 min each using a SONICATOR 3000 ( MISONIX ) , with each cycle consisting of 30 sec sonication on and 30 sec off . The average size of the resulting DNA fragments was between 200 and 500 base pairs . Following centrifugation at 13 . 5K for 30 min at 4°C , the solubilized chromatin was collected and adjusted to 500 µl with FA lysis buffer . Twenty microliters were removed for use as input chromatin . For immunoprecipitation , 10 OD equivalents of solubilized chromatin were incubated overnight at 4°C , together with 20 µl of protein G dynabeads ( Invitrogen ) that had been pre-bound with anti-H3 or anti-Myc ( Sgs1-13Myc ) . Immunoprecipitates were collected by a step-wise washing protocol , consisting of 1 . 5 ml FA-lysis buffer , 1 . 5 ml WASH I ( FA lysis buffer+0 . 5 M NaCl ) , 1 . 5 ml WASH II ( 10 mM Tris-Cl , pH 7 . 5 , 1 mM EDTA , 0 . 25 M LiCl , 0 . 5% NP-40 , 0 . 5% sodium deoxycholate ) , and 1 . 5 ml TE ( pH 8 . 0 ) for 5 min each at room temperature . The immuno-complexes were eluted by adding 0 . 25 ml Elution buffer ( 50 mM Tris-Cl , pH 7 . 5 , 10 mM EDTA , 1% SDS ) , and incubated first at 65°C for 20 minutes , and then at room temperature for 10 minutes with vortexing . DNA was purified using Qiaquick PCR purification spin-columns ( Qiagen ) , and used as template for quantitative-PCR . All the primers used is listed in Table S3 . The primers used in the histone H3 ChIP experiment were designed to amplify DNA fragments present at nucleosomes , as depicted in Figure S6 . For immunoprecipitations [56] , log phase WT or htb-K123R cells untreated ( − ) or treated ( + ) with 0 . 2M HU were collected , resuspended in buffer containing 50 mM Tris7 . 5 , 150 mM NaCl , 5 mM EDTA , 0 . 5% Triton X-100 , and proteinase inhibitors , and broken open by bead beating . A total of 5 mg of protein extract was diluted in 1 ml of the same buffer , and incubated with pre-bound anti-HA-protein G beads at 4°C for 2 . 5 hours , and then rotated at 4°C overnight . Beads were then washed with 1 ml buffer four times . SDS-loading dye was added , and samples were boiled and resolved on SDS-PAGE . Results are expressed as the mean ± SEM from the number of experiments indicated in the figure legends . Student's t-test was used to analyze statistical significance . | Eukaryotic DNA is organized into nucleosomes , which are the fundamental repeating units of chromatin . Coordination of chromatin structure is required for efficient and accurate DNA replication . Aberrant DNA replication results in mutations and chromosome rearrangements that may be associated with human disorders . Therefore , cellular surveillance mechanisms have evolved to counteract potential threats to DNA replication . These mechanisms include checkpoints and specialized enzymatic activities that prevent the replication and segregation of defective DNA molecules . We employed a genome-wide approach to investigate how chromatin structure affects DNA replication under stress . We report that coordination of chromatin assembly and checkpoint activity by a histone modification , H2B ubiquitylation ( H2Bub ) , is critical for the cell response to HU-induced replication stress . In cells with a mutation that abolishes H2Bub , replication progression is enhanced , and the forks are more susceptible to damage by environmental insults . The replication proteins on replicating DNA are akin to a train on the tracks , and movement of this train is carefully controlled . Our data indicate that H2Bub helps organize DNA in the nuclei during DNA replication; this process plays a similar role to the brakes on a train , serving to slow down replication , and maintaining stable progression of replication under environmental stress . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"genomics",
"dna",
"replication",
"cell",
"biology",
"chromosome",
"biology",
"genome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"epigenetics",
"computational",
"biology",
"chromatin"
] | 2014 | H2B Mono-ubiquitylation Facilitates Fork Stalling and Recovery during Replication Stress by Coordinating Rad53 Activation and Chromatin Assembly |
Seed dormancy is an important economic trait for agricultural production . Abscisic acid ( ABA ) and Gibberellins ( GA ) are the primary factors that regulate the transition from dormancy to germination , and they regulate this process antagonistically . The detailed regulatory mechanism involving crosstalk between ABA and GA , which underlies seed dormancy , requires further elucidation . Here , we report that ABI4 positively regulates primary seed dormancy , while negatively regulating cotyledon greening , by mediating the biogenesis of ABA and GA . Seeds of the Arabidopsis abi4 mutant that were subjected to short-term storage ( one or two weeks ) germinated significantly more quickly than Wild-Type ( WT ) , and abi4 cotyledons greened markedly more quickly than WT , while the rates of germination and greening were comparable when the seeds were subjected to longer-term storage ( six months ) . The ABA content of dry abi4 seeds was remarkably lower than that of WT , but the amounts were comparable after stratification . Consistently , the GA level of abi4 seeds was increased compared to WT . Further analysis showed that abi4 was resistant to treatment with paclobutrazol ( PAC ) , a GA biosynthesis inhibitor , during germination , while OE-ABI4 was sensitive to PAC , and exogenous GA rescued the delayed germination phenotype of OE-ABI4 . Analysis by qRT-PCR showed that the expression of genes involved in ABA and GA metabolism in dry and germinating seeds corresponded to hormonal measurements . Moreover , chromatin immunoprecipitation qPCR ( ChIP-qPCR ) and transient expression analysis showed that ABI4 repressed CYP707A1 and CYP707A2 expression by directly binding to those promoters , and the ABI4 binding elements are essential for this repression . Accordingly , further genetic analysis showed that abi4 recovered the delayed germination phenotype of cyp707a1 and cyp707a2 and further , rescued the non-germinating phenotype of ga1-t . Taken together , this study suggests that ABI4 is a key factor that regulates primary seed dormancy by mediating the balance between ABA and GA biogenesis .
Seed dormancy prevents or delays the germination of maturated seeds even when conditions are favorable for germination [1]–[3] . Seed dormancy is an important trait for diverse , important crop species including rapeseed , wheat , corn and rice , because seed dormancy inhibits pre-harvest spouting or vivipary [4] . Vivipary usually causes great economic loss to cereal production , including losses in seed quantity and quality , especially in humid regions worldwide [5] , [6] . On the other hand , deep seed dormancy is problematic , especially in the horticultural and forest industries , and chemical treatments may be required to promote germination [7] . Thus , the optimal level of seed dormancy is a valuable trait for agricultural production . Therefore , it is essential to understand the precise molecular mechanisms that control seed dormancy and germination . Diverse endogenous and environmental factors including phytohormones , nutrients , temperature and light affect seed dormancy through different pathways [3] , [8] . Extensive studies have shown that abscisic acid ( ABA ) and gibberellin acid ( GA ) are the primary endogenous factors that regulate the transition from dormancy to germination , and they regulate this process antagonistically [1] , [2] , [9]–[11] . ABA is essential for the induction and maintenance of seed dormancy , while GA is required for the release of dormancy and for the initiation of seed germination [9] , [10] . In line with this conclusion , some ABA-deficient mutants such as nced6 , nced3 , nced5 , nced9 , aba2 and aao3 are better able to germinate than WT seeds [9] , [12] . In support of these observations , the overexpression of the ABA biosynthesis gene ABA2 enhances ABA accumulation and maintains deep seed dormancy [13] . Further , overexpression of other ABA biosynthesis genes , NCED6 and NCED9 , even inhibits precocious germination of developing seeds due to increased ABA biogenesis [14] . By contrast , some ABA metabolic pathway mutants , such as cyp707a1 , cyp707a2 and cyp707a3 , accumulate higher ABA levels than WT and subsequently exhibit hyperdormancy in seeds [15]–[17] . In addition to ABA content , ABA signaling also positively regulates seed dormancy [2] , [3] . Although ABI1 and ABI2 are negative regulators in the ABA signaling pathway , the abi1-1 and abi2-1 mutants show the reduced dormancy levels [18] . This phenotype results from dominant-negative mutations and therefore , these two PP2Cs ( Protein Phosphatases type 2C ) are unable to bind to ABA receptors [19] , [20] . In addition , the abi3 mutant also shows reduced seed dormancy levels [21] . Furthermore , the allelic mutant abi3-3 even rescues the non-germinating phenotype of ga1 in the absence of exogenous GA treatment , indicating that ABI3 is a negative regulator of GA biosynthesis [22] . Although a previous study has concluded that abi5 does not reduce seed dormancy [23] , other studies have shown that this gene negatively regulates seed germination [24] , [25] . In contrast to ABA , GA negatively regulates seed dormancy [2] , [3] . Mutants severely defective in GA biosynthesis such as ga1 show deep seed dormancy and fail to germinate in the absence of exogenous GA [26] . On the other hand , mutants defective in GA 2-oxidases ( GA2ox ) , which deactivate bioactive GA , exhibit reduced seed dormancy and germinate normally , even in the dark [27] . Mutations in two negative regulators in the GA signal transduction pathway , rgl2 ( RGA-LIKE2 ) and spy ( SPINDLY ) , rescue the non-germination phenotype of ga1-3 in the absence of exogenous GA [28] , [29] . Combined with the conclusion that ABA and GA regulate seed dormancy antagonistically [3] , the ability to synthesize GA is enhanced in the aba2 mutant , indicating that ABA is involved in the suppression of GA biogenesis in both developing and imbibed seeds [9] . These pioneering studies demonstrated that ABA and GA biogenesis and signaling play key roles in the control of seed dormancy and germination . However , the detailed molecular mechanism by which the crosstalk between ABA and GA at the hormone biogenesis level regulates seed dormancy is largely unknown . ABI4 encodes an AP2/ERF transcription factor , which is an enhancer in the ABA signal transduction pathway that functions especially during seed development and germination [23] , [30] , [31] . Furthermore , ABI4 is also involved in other aspects of plant development including lipid mobilization from the embryo [32] , glucose responses [33] , [34] , salt responses [35] and the mitochondrial and chloroplast-nucleus retrograde signaling pathways [36]–[38] . Most recently , ABI4 was found to regulate ABA and cytokinin inhibition of lateral root development by reducing polar auxin transport [39] , as well as ABA- and jasmonate-dependent signaling pathway crosstalk [40] and the nitrogen deficiency stress response [41] . Except in young seedlings , the ABI4 transcript level is relatively low through most stages of vegetative growth but high in both developing and imbibed seeds [31] . The abundance of ABI4 protein is partially regulated by the 26S-proteasomal pathway [42] . These excellent studies demonstrate that ABI4 is a versatile factor , which functions in diverse signaling pathways and is tightly regulated at the post-transcriptional level . However , the role of ABI4 in crosstalk between ABA and GA has not yet been elucidated . As described above , many mutants in which the ABA signal is attenuated , such as abi1-1 , abi2-1 and abi3 , exhibit the reduced seed dormancy phenotype [18] , [21] . Although this protein is a positive regulator of the ABA signaling pathway , however , previous studies have concluded that ABI4 has no effect on seed dormancy [23] , and this opinion has been accepted in the field [3] , [43] . Recently , two studies have demonstrated that a mutation in a double-repeat AP2 domain transcription factor , CHOTTO1 , results in reduced primary seed dormancy , and , interestingly , ABI4 likely acts upstream of CHOTTO1 in the genetic pathway [10] , [44] . On the other hand , the ABI4 transcript level is relatively low at almost all growth stages except during seed maturity and germination [31] . These studies inspired us to reconfirm the effects of ABI4 on primary seed dormancy as well as postgerminative growth . Here , we show that abi4 mutant seeds indeed exhibited reduced primary seed dormancy and increased cotyledon greening . The differences in germination rates and cotyledon greening between abi4 and WT decreased moderately after stratification . After-ripening treatment caused the rates of germination and cotyledon greening to be comparable between abi4 and WT . In line with these results , the ABA content in abi4 dry seeds was significantly lower than that in WT , but the ABA levels were comparative after stratification treatment . Consistently , the GA level in abi4 seeds was upregulated compared to WT . Further analysis showed that abi4 was resistant to exogenous paclobutrazol ( PAC ) , a GA biosynthesis inhibitor , while OE-ABI4 was sensitive to PAC during germination , and exogenous GA rescued the delayed germination phenotype of OE-ABI4 . The qRT-PCR assay also showed that the transcript levels of some GA biosynthesis and ABA inactivation genes were upregulated in germinating abi4 seeds , while some GA inactivation and ABA biosynthesis genes were downregulated . ChIP-qPCR and transient expression assays showed that ABI4 indeed inhibits CYP707A1 and CYP707A2 transcription by directly binding to these promoters , and the CCAC cis-elements are essential for this repression . Further genetic analysis showed that abi4 restored the delayed germination phenotype of cyp707a1 and cyp707a2 , and importantly , mutation in ABI4 also rescued the non-germinating phenotype of ga1-t even in the absence of exogenous GA treatment , reconfirming that ABI4 is a negative regulator of GA biogenesis and a positive regulator of ABA biosynthesis during seed germination . Taken together , this study demonstrates that ABI4 plays pivotal and complex roles in fine-tuning the ABA/GA balance to control primary seed dormancy .
Occasionally we found that the abi4-1 seeds germinated more quickly than WT seeds when the siliques fell onto the surface of the soil . Thus , we decided to investigate the effect of the ABI4 gene on seed dormancy . The abi4-1 ( hereafter referred to as abi4 ) mutant was obtained from ABRC ( Arabidopsis Biological Resource Center at the Ohio State University; stock number CS8104 ) . This mutant contains a point mutation in the open reading frame in which the 469th base G is deleted , resulting in a frame shift at codon 157 and producing a protein containing the predicted DNA binding and dimerization domains but lacking the presumed activation domain [31] . To investigate the effect of abi4 on seed dormancy , the germination of the abi4 mutant and WT seeds was scored . Using seeds subjected to one week of dry storage , the germination rate of abi4 seeds was clearly shown to be significantly higher than that of WT without 4°C stratification treatment ( Figure 1A ) . At 1 . 5 days after sowing , the endosperms of most abi4 seeds were ruptured , and radicles emerged from some seeds , while the testas of WT seeds had not even ruptured at this time ( Figure 1A ) . The germination rate was nearly 80% for abi4 seeds at day 2; however , the germination rate for WT seeds was less than 40% at this time point ( Figure 1A ) . Consistent with the reduced dormancy level , the abi4 mutant also exhibited markedly faster cotyledon greening than the WT ( Figure S1A ) . In addition , it is noteworthy that at 4 . 5 days after sowing , the germination rates of abi4 and WT reached 100% ( Figure 1A ) . Taken together , the results of this time-course experiment show that the abi4 mutant indeed exhibits reduced seed dormancy . Previous studies have shown that stratification treatment reduces primary seed dormancy and thus promotes seed germination [10] , [45] . Therefore , we also investigated the effect of stratification on primary seed dormancy in abi4 . When seeds subjected to one week of dry storage were stratified for 3 days , the differences in the rates of germination and cotyledon greening between abi4 and WT were moderately reduced ( Figures 1B , S1B ) , compared with the larger difference shown in Figures 1A and S1A . However , the percentage of germination and cotyledon greening of abi4 was still higher than that of WT , and the growth rate of the radicle of abi4 was significantly faster than that of WT ( Figures 1B , S1B ) . Next , when we examined seeds subjected to two weeks of dry-storage , the similar trends were detected ( Figures 1C , S1C ) , and the differences between abi4 and WT decreased moderately with stratification treatment ( Figures 1D , S1D ) . Indeed , the abi4 seeds subjected to short-term storage germinated more quickly than WT , especially without stratification treatment ( Figure 1A , 1B , 1C and 1D ) . Subsequently , the effect of after-ripening on primary seed dormancy was investigated . The faster germination phenotype of abi4 seeds was abolished when the seeds were fully after-ripened , either with or without stratification treatment ( 6-month dry-storage; Figure 1E , 1F ) . Consistent with these results , we also did not detect differences in cotyledon greening rates between abi4 and WT when we employed fully after-ripened seeds ( Figure S1E , S1F ) . Altogether , these results suggest that abi4 reduces primary seed dormancy . A reduced primary seed dormancy level usually results in preharvest sprouting or vivipary in cereals , especially if moist conditions are encountered [46] . Therefore , we tested whether the abi4 mutant exhibits vivipary in developing seeds using a protocol employed in a previous study [14] . The results show that abi4 seeds in developing siliques indeed germinated more quickly than WT both on 1/2 MS medium and on soil ( Figure 1G ) . In particular , at 8 days after sowing , only a few seeds in WT siliques germinated , while young abi4 seedlings were already established ( Figure 1Ga , 1Gb ) . At 14 days after sowing , most of the abi4 and WT siliques produced seeds that germinated , and the cotyledons greened ( Figure 1Gc , 1Gd ) , indicating that the seed vigor in these developing siliques was normal . Therefore , we reasoned that the difference in germination rate between abi4 and WT siliques resulted from different seed dormancy levels . To further confirm that the reduced primary dormancy level phenotype of abi4 resulted from a mutation in the ABI4 locus , we obtained another T-DNA insertion mutant in this locus from ABRC ( the stock name is SALK_080095 , hereafter referred to as abi4-t ) . A previous study showed that this line is a knockout mutant [47] . Similar to abi4 , the decreased seed dormancy level , early germination phenotype of abi4-t was also observed when we analyzed seeds subjected to one week of dry storage ( Figure S2A ) , and the percentage of cotyledon greening of abi4-t was also significantly higher than that of WT seeds ( Figure S2B ) . On the other hand , the faster germination phenotype of abi4-t seeds ( compared with WT ) was abolished when fully after-ripened seeds were employed ( data not shown ) . The similar phenotype of the two allele mutants further proved that a mutation in the ABI4 locus is indeed responsible for the reduced primary seed dormancy phenotype . To demonstrate the reproductivity of our experiment , we employed several mutants with seed dormancy phenotype in ABA pathway as controls . Previous study demonstrated that snrk2 . 2/snrk2 . 3 obviously reduced seed dormancy level compared to WT [48] . This phenotype is resulted from the impaired ABA signaling in this double mutant [48] . In our growth condition , the reduced seed dormancy phenotype of snrk2 . 2/snrk2 . 3 was repeated perfectly ( Figure S3A ) . Notably , the decreased seed dormancy phenotype of snrk2 . 2/snrk2 . 3 was stronger than that of abi4 ( Figure S3A ) . Furthermore , the reduced seed dormancy phenotype of abi1-1 and abi2-1 compared to Ler seeds was also detected in the same condition ( Figure S3B ) , which was consistent with our current knowledge [18] . These results demonstrated that the present experimental condition is eligible and reliable . As described above , abi4 seeds subjected to short-term storage exhibited significantly reduced dormancy compared with WT seeds , but the difference in germination rate between abi4 and WT was decreased moderately or even abolished after stratification or longer period of storage ( Figure 1 ) . Since ABA positively regulates seed dormancy [1] , and stratification and after-ripening treatment reduce the ABA content [10] , we next examined endogenous ABA levels in dry and stratified abi4 seeds using a liquid chromatography–tandem mass spectrometry system . We chose seeds subjected to two weeks of dry storage for this experiment . As expected , the result showed that the ABA content in abi4 was significantly lower than that in WT ( Figure 2 ) when no stratified seeds were analyzed , suggesting that ABI4 positively regulates ABA biogenesis . On the other hand , stratification treatment impairs ABA biosynthesis [49] . Accordingly , after a 3-day stratification treatment , both WT and abi4 mutant seeds contained lower ABA levels , and importantly , the ABA levels were comparable between WT and abi4 ( Figure 2 ) . The trend of ABA level in dry or stratified abi4 seeds is similar to that of CHOTTO1 , a positive regulator of primary seed dormancy that may acts downstream of ABI4 in a genetic pathway [10] , [44] . The hormonal measurements described above revealed that the decreased ABA level in abi4 seeds is at least partially responsible for the reduced primary seed dormancy phenotype of this mutant ( Figure 2 ) . To further investigate the precise mechanism by which ABI4 regulates primary seed dormancy , ABI4-overexpressing plants were generated . The coding region of ABI4 was introduced into the vector pCanG-HA-GFP under the control of the CaMV ( Cauliflower mosaic virus ) 35S promoter and transformed into WT Arabidopsis . Several independent T3 homozygous lines were identified through qRT-PCR and western blot assays , and two of them were shown ( Figure 3A , 3B ) . Because ABI4 directly promotes ABI5 transcription [50] , we examined the ABI5 expression levels in those transgenic lines . The qRT-PCR assay showed that ABI5 transcription was indeed upregulated in these ABI4 overexpressing lines ( Figure 3C ) . Thus , we reasoned that these two overexpressing lines are functional and they were employed in further analysis . Because both the abi4 and abi4-t mutants showed the reduced primary seed dormancy phenotype ( Figures 1 , S2 ) , we first tested the seed dormancy level of OE-ABI4 seeds subjected to two-week dry-storage on normal 1/2 MS medium . The results showed that the two independent lines germinated slowly than WT ( Figure 3D ) , and accordingly , the cotyledon greening rates of these two lines were also moderately lower than that of WT ( Figure S4A ) . These results indicate that the seed dormancy level in OE-ABI4 seeds was higher than that of WT , which is in contrast to the phenotype of the both of abi4 mutants . Our results show that ABI4 positively regulates ABA biogenesis ( Figure 2 ) , and a previous study demonstrated that ABA is involved in the suppression of GA biosynthesis in imbibed seeds [9] . Thus , we speculated that the GA level in abi4 was higher than that in WT . To confirm this speculation , we analyzed the responsiveness of abi4 mutant and OE-ABI4 seeds to GA and PAC treatment . Our results showed that OE-ABI4 seeds were sensitive to PAC during germination ( Figure 3E ) and cotyledon greening ( Figure S4B ) , while abi4 was resistant ( Figure 3E , S4B ) . However , the rates of germination and cotyledon greening among abi4 , WT and OE-ABI4 were comparable when we used medium supplemented with exogenous GA ( Figure 3F , S4C ) . The increased resistance of abi4 to the GA biosynthesis inhibitor suggests that this mutant contains higher levels of active GA or possesses stronger GA signaling than the WT [51] . Combined with the fact that exogenous GA can rescue the delayed germination and cotyledon greening phenotypes of OE-ABI4 , we proposed that ABI4 attenuates GA biosynthesis to positively regulate seed dormancy . The responsiveness analysis of abi4 and OE-ABI4 seeds to GA and PAC treatments suggested that ABI4 negatively regulates GA biogenesis ( Figure 3E , 3F ) . Furthermore , because that ABA is involved in the suppression of GA biosynthesis during seed germination [9] , and the ABA measurements between abi4 and WT seeds also supported this speculation ( Figure 2 ) . Then , we examined the endogenous GA content in abi4 and WT seeds . The result showed that the active GA4 level in abi4 dry seeds was significantly higher than that in WT ( Figure 3G ) , suggesting that ABI4 indeed regulates GA biosynthesis negatively . Combined with the ABA quantification result , the endogenous hormone measurements demonstrated that the decreased ABA level and the increased GA content in abi4 seeds are responsible for the reduced primary seed dormancy of abi4 . To further confirm that ABI4 functions as an attenuator of GA biogenesis during seed germination , we analyzed the effect of the ABI4 mutation on the expression of GA biosynthesis genes and GA inactivation genes in dry and imbibed seeds . The results of qRT-PCR analysis showed that the transcript levels of GA biosynthesis genes , including GA3 , GA3ox1 , GA20ox1 , GA20ox3 , KAO1 , KAO2 and GA20ox2 , were upregulated to varying degrees in abi4 seeds after imbibition ( Figure 4A ) . The expression levels of GA3 and KAO2 in dry abi4 seeds were 2-fold higher than that in WT , and this trend was maintained throughout the imbibition treatment process ( Figure 4A ) . Higher levels of GA3ox1 mRNA were detected in the abi4 mutant after 6 hours of imbibition ( Figure 4A ) . The transcripts of KAO1 , GA20ox1 , GA20ox2 and GA20ox3 were higher in abi4 than in WT during the entire imbibition process , although the differences were not significant ( Figure 4A ) . By contrast , the transcript level of GA2ox8 , a key GA inactivation gene , was lower in the abi4 mutant than in the WT ( Figure 4B ) . The increased expression of GA biosynthesis genes and the decreased expression of GA inactivation genes in imbibed seeds are accordance with the GA measurements in abi4 mutant seeds which contains higher levels of active GA than the WT ( Figure 3G ) . Consistent with this , the RGL3 gene , which encodes a DELLA transcription regulator that represses testa rupture during seed germination [52] , was downregulated in both dry and imbibed abi4 seeds ( Figure 4C ) . Since ABA and GA regulate seed germination antagonistically [2] , the expression levels of ABA biosynthesis and inactivation genes in dry and imbibed seeds were also analyzed . Analysis by qRT-PCR showed that the mRNA levels of ABA biosynthesis genes , including NCED2 and NCED3 , were downregulated in abi4 ( Figure 4D ) , while the ABA inactivation genes such as CYP707A1 , CYP707A2 and CYP707A3 were upregulated ( Figure 4E ) . The higher transcription levels of these three inactivation genes in abi4 were maintained throughout the entire imbibition process . Notably , the expression level of CYP707A2 in abi4 was almost 4-fold higher than that in WT at 6 hours after imbibition ( Figure 4E ) . The high transcription levels of ABA inactivation genes , and the low level of ABA synthesis in abi4 , explain the results of ABA measurement ( Figure 2 ) . Together , the transcript levels of GA biosynthesis and ABA inactivation genes were upregulated in germinating abi4 seeds , while GA inactivation and ABA biosynthesis genes were downregulated . These results are consistent with the notion that ABI4 negatively regulates GA biosynthesis while positively regulating ABA biogenesis ( Figures 2 , 3 ) . Previous studies have demonstrated that ABI4 is a versatile transcription factor that binds to the CACCG motif to promote the expression of some genes; this factor also binds to the CCAC element to directly inhibit the transcription of some genes [38] , [50] . To investigate whether ABI4 directly regulates the expression some GA and ABA metabolism genes , we first examined the promoters of the genes described in Figure 4 because the expression levels of these genes were altered in abi4 during germination . CYP707A1 , CYP707A2 and CYP707A3 were most interesting because 6 , 5 and 7 CCAC motifs were detected in these three promoters , respectively ( Figure 5A , 5B and 5C ) . This inspired us to examine whether ABI4 directly binds to these promoters in vivo . We then conducted a ChIP ( chromatin immunoprecipitation ) -qPCR assay with the ABI4 transgenic lines to examine whether ABI4 binds to these promoters directly . Because ABI4 binds directly to the promoter of ABI5 [50] , a DNA element of the ABI5 promoter was used as positive control . Two independent OE-ABI4 transgenic lines ( OE1 and OE2 ) were subjected to ChIP-qPCR analysis , which produced similar results . We determined that the promoters of CYP707A1 and CYP707A2 were enriched in the chromatin immunoprecipitated DNA using the anti-GFP antibody ( Figure 5D ) , especially the P2 and P3 regions in CYP707A1 and the P5 region in CYP707A2 . This result indicates that ABI4 directly binds to the promoters of CYP707A1 and CYP707A2 in vivo . However , we did not detected significant enrichment of all of the elements tested from promoter of CYP707A3 ( Figure 5D ) , although this promoter possesses 7 CCAC motifs ( Figure 5C ) . These results indicate that ABI4 may repress CYP707A1 and CYP707A2 expression by directly binding to the promoters of these genes . Physiological and molecular evidence support the notion that the biogenesis of ABA and GA during seed germination is affected by ABI4 , and ABI4 positively regulates primary seed dormancy . To further confirm this conclusion , we subsequently dissected the genetic relationship between ABI4 and hormone metabolism genes . GA1 encodes ent- ent-copalyl diphosphate synthase synthase , a key enzyme that catalyzes a relatively early biochemical reaction in the biosynthesis of GA [53] , [54] . The ga1 loss-of-function alleles cause GA deficiency and abolish seed germination in the absence of exogenous GA [26] , [54] . OE-ABI4 seeds were sensitive to PAC during germination , while abi4 seeds were resistant ( Figure 3E ) , and further , the GA biogenesis was attenuated in abi4 seeds compared to WT ( Figure 3G ) , indicating that GA biosynthesis is indeed negatively regulated by ABI4 . Therefore , we examined whether abi4 could rescue the non-germination phenotype conferred by ga1-t . First , we created a double mutant between the abi4 and ga1-t homozygous mutants through genetic crossing . Subsequently , seed germination was analyzed in the abi4 , ga1-t and abi4/ga1-t double mutants using seeds subjected to two weeks of dry storage . The results showed that the abi4/ga1-t double mutants germinated normally , and the cotyledons greened normally , even in the absence of exogenous GA , while ga1-t did not germinate under these condition ( Figures 6A , S7A ) . As expected , the exogenous application of GA restored the germination of ga1-t . Furthermore , the abi4/ga1-t double mutant also germinated , and the cotyledons greened slightly more quickly than those of ga1-t in the presence of exogenous GA ( Figures 6B , S7B ) . These results demonstrate that ABI4 indeed negatively regulates GA biogenesis from the view of genetics . On the other hand , the cyp707a1 and cyp707a2 mutants accumulate higher levels of ABA than the WT and thus exhibit the delayed germination phenotype [15] . Since the ABA level was downregulated in the abi4 mutant ( Figure 2 ) and ABI4 directly inhibits CYP707A1 and CYP707A2 expression by binding to those promoters ( Figure 5D to 5H ) , we tested whether abi4 could rescue the germination defect phenotype of cyp707a1 and cyp707a2 . Therefore , abi4/cyp707a1 and abi4/cyp707a2 double mutants were created between abi4 and the homozygous mutant cyp707a1 ( SALK_069127 ) and cyp707a2 ( SALK_083966C ) , respectively . Our results showed that the seeds of these double mutants showed higher germination frequencies than the corresponding cyp single mutants , cyp707a1 and cyp707a2 , but lower than abi4 when the seeds were subjected to two weeks of dry storage ( Figure 6C , 6D , right panel ) . Given that ABI4 positively regulates ABA biogenesis , we speculated that the reason responsible for the recovery of abi4/cyp707a1 and abi4/cyp707a2 regarding the delayed germination phenotype of cyp707a2 and cyp707a2 is that the ABA biogenesis is impaired in these double mutants . For this end , we further detected the ABA content in the cyp707a2 single mutant and abi4/cyp707a2 double mutant respectively . Indeed , our results revealed that the ABA level in cyp707a2 is significantly increased compared to WT ( Figure S8 ) , which is consistent with the previous study [15] . Importantly , we detected that the ABA content in abi4/cyp707a2 is decreased compared to cyp707a2 single mutant ( Figure S8 ) . These results indicate that a mutation in the ABI4 locus recovers the reduced germination potential of cyp707a1 and cyp707a2 through attenuating the ABA biogenesis . Together , these genetic analyses between CYP707A1 , CYP707A2 , GA1 and ABI4 further confirmed the notion that ABI4 indeed positively regulates ABA biosynthesis and negatively regulates GA biogenesis .
Pre-harvest sprouting of diverse cereal seeds usually occurs under humid conditions during harvest time and results in the germination of grains that are still on the mother plant . Sprouting , which results from the reduced dormancy level of crop seeds , lowers the value of crop seeds in terms of both quantity and quality [2] , [6] . Therefore , pre-harvest sprouting has attracted increasing amounts of attention from researchers , especially in agronomic regions; the precise molecular mechanism underlying seed dormancy and pre-harvest sprouting is worth exploring . In the present study , the abi4 seeds obviously germinated significantly more quickly than WT when the seeds were subjected to short-term storage; this mutant even exhibited the vivipary phenotype ( Figure 1 ) . On the other hand , it is noteworthy that the percentages of germination of abi4 and WT seeds were comparable at 4 . 5 days after sowing ( all reached nearly 100%; Figure 1A to 1D ) , which is in accordance with previous result [23] . In a previous study , the germination rate was scored at 5 days after sowing , and the abi4 mutant showed the same degree of dormancy as WT seeds ( both genotypes reached 100% germination ) [23] . Therefore , ABI4 was thought to have no effect on seed dormancy . Subsequent studies and reviews cited this conclusion [3] , [31] , [43] . Recently , two studies showed that CHOTTO1 regulates primary seed dormancy positively , and , more interestingly , ABI4 likely acts in the same genetic pathway as CHOTTO1 [10] , [44] . Both studies , along with our own occasionally observation that about abi4 germinated more quickly than WT when the siliques fell onto the soil , inspired us to reconfirm the effect of ABI4 on primary seed dormancy . We speculate that the reason for the previous conclusion ( that ABI4 has no effect on seed dormancy ) is that the germination rates were not scored using detailed time-course analysis [23] . Seed dormancy can be classified as primary or secondary seed dormancy [55] . Freshly harvested seeds , or dormant seeds subjected to short-term storage , are deemed to have primary dormancy , which is induced by ABA during seed maturation on the mother plant and is abolished by longer period of dry-storage treatment ( after-ripening ) [10] , [56] , [57] . By contrast , secondary dormancy can be induced in seeds with non-deep physiological dormancy after seed dispersal , and it is often associated with annual dormancy cycles in seed banks [56] . In the present study , abi4 seeds subjected to shorter period of dry-storage showed reduced seed dormancy levels and even the vivipary phenotype ( Figure 1A to 1D , 1G ) . By contrast , the germination frequencies of abi4 and WT were comparable when the seeds subjected to longer period of storage ( Figure 1E , 1F ) . On the other hand , further investigation revealed that OE-ABI4 seeds subjected to two weeks of storage germinated more slowly than WT seeds ( Figure 3D ) , and the cotyledon greening rates of different genotypes were consistent with the dormancy levels ( Figure S4A ) . Taken together , we conclude that ABI4 indeed positively regulates primary seed dormancy . After confirming the effect of ABI4 on primary seed dormancy , we dissected the molecular mechanism underlying this phenotype . The reduced primary seed dormancy of abi4 was moderately decreased by stratification and was even abolished by longer period of after-ripening treatment ( Figure 1B , 1D , 1E , 1F ) . Furthermore , stratification and after-ripening treatments reduce ABA content [10] , [49] . Therefore , we tested the ABA levels in dry and imbibed seeds . As expected , the ABA content in dry abi4 seeds was lower than that in WT and became comparable after stratification ( Figure 2 ) . This result is similar to previously reported results about CHOTTO1 , which also positively regulates primary seed dormancy [10] , [44] . In these studies , the ABA level was downregulated in the cho1 mutant , which was responsible for the reduced primary seed dormancy phenotype of cho1 [10] . Therefore , we conclude that the decreased ABA level in the abi4 mutant is at least partially responsible for the reduced primary seed dormancy phenotype , and further , ABI4 positively regulates ABA biogenesis . On the other hand , GA biosynthesis is enhanced in the ABA deficient aba2 mutant , indicating that ABA is involved in the suppression of GA biosynthesis in both developing and imbibed seeds [9] . Because the ABA content in abi4 seeds was markedly downregulated ( Figure 2 ) , we tested the responses of abi4 and OE-ABI4 to PAC and GA during seed germination . A previous report showed that the increased resistance to PAC suggests that the mutant contains higher levels of active GA or stronger GA signaling than the WT [51] . We found that OE-ABI4 was sensitive to PAC and abi4 was resistant , while exogenous GA rescued the delayed germination phenotype of OE-ABI4 ( Figures 3D to 3F , S4 ) , and further , the GA measurements result showed that abi4 seeds indeed contain higher levels of active GA4 than the WT ( Figure 3G ) . These results are consistent with the ABA measurements ( Figure 2 ) . Therefore , we propose that ABI4 attenuates GA biosynthesis and promotes ABA biosynthesis to precisely regulate seed germination . To further confirm the changes in ABA and GA content during seed germination , we also investigated the expression levels of ABA and GA biosynthetic and inactivation genes in dry and imbibed seeds . The results showed that the expression of most genes involved in ABA and GA metabolism was altered in dry and imbibed abi4 seeds ( Figure 4 ) , which is consistent with the results of ABA and GA quantification , and the analysis of the responsiveness of OE-ABI4 and abi4 to GA and PAC treatments ( Figures 2 , 3 , S4 ) . These results were similar to results obtained from the analysis of sorghum grains , i . e . , changes in the expression level of GA metabolism genes affects the seed dormancy and germination potential of sorghum grains [58] . In particular , the expression levels of CYP707A1 and CYP707A2 were remarkably decreased in the abi4 seeds ( Figure 4E ) . Furthermore , ChIP-qPCR analysis and the tobacco transient expression assays revealed that ABI4 inhibits both of the two ABA inactivation genes ( CYP707A1 and CYP707A2 ) expression by directly binds to the promoters ( Figure 5D ) . In addition , the CCAC motifs in these promoters are important and the inhibition effect of ABI4 on its transcription was depended on the CCAC cis-element ( Figure S6 ) . Further evidence confirming the regulation of ABA biogenesis by ABI4 was obtained by genetic analysis; the abi4 mutant rescued the delayed germination phenotype of cyp707a1 and cyp707a2 ( Figure 6C , 6D ) . Accordingly , our further experimental evidences demonstrated that ABI4 directly repress CYP707A1 and CYP707A2 expression to promote ABA biosynthesis ( Figure 5E to 5H ) , and the higher expression level of CYP707A1 and CYP707A2 in the absence of ABI4 result in reduced ABA content and , subsequently , the decreased seed dormancy level ( Figures 2 , 5 ) . Notably , except for cyp707a1 and cyp707a2 , abi4 also rescued the non-germination phenotype of ga1-t without exogenous GA treatment ( Figure 6A ) , suggesting that ABI4 is indeed involved in regulation of GA biogenesis . Mutation at early stage of GA synthesis gene does not totally abolish GA in plant , and the ga1-3 , an allele mutant of ga1-t , contains very low level of GA [59] . In abi4 and abi4/ga1-t double mutants , reduced ABA contents and activated downstream GA synthesis and down regulated GA metabolic gene transcription might increase GA/ABA ratio in seeds , thus promotes the germination of abi4/ga1-t double mutant ( Figures 4 , 6A ) . abi4 has the similar effects of spy , rgl2 and abi3 on the ga1 mutant [22] , [28] , [29]; these genes also are negative regulators of the GA biogenesis or signaling pathway . Taken together , we conclude that ABI4 regulates ABA biogenesis positively , and GA biosynthesis negatively , during seed germination . Previous elegant studies demonstrated that ABI4 is a key ABA signaling component per se [31] , and in this study , we further showed that ABI4 is also involved in ABA and GA biogenesis ( Figures 2 , 3 ) . High GA level could induce the transcription of α-amylase gene , whose product in turn hydrolyzes the seed coat which is essential for normal germination process . In opposite , ABA inhibits seed germination through suppressing the α-amylase gene expression [3] . Furthermore , previous study revealed that ABA is involved in the suppression of GA biogenesis [9] . Therefore , the decreased ABA level in abi4 seeds could further activates the GA biogenesis , and subsequently , the increased GA content further promotes the α-amylase gene transcription . Accordingly , the seed dormancy level of abi4 is decreased . Although the decreased ABA level and increased GA content in abi4 seeds are responsible for the reduced primary seed dormancy in this mutant ( Figures 2 , 3 , S4 ) , it is noteworthy that reduced seed dormancy was also detected when the short-term stored abi4 seeds were stratified ( Figure 1B , 1D ) , even the corresponding ABA levels were comparable between abi4 and WT after stratification treatment ( Figure 2 ) . These results suggest that ABA signaling plays an important role in the control of primary seed dormancy . Indeed , previous studies have demonstrated that ABI4 positively regulates ABA signaling during seed germination [31] , [60] , and our results are consistent with this conclusion ( Figure 1D ) . The other evidence about the key regulators in ABA signaling involved in seed dormancy control was from the analysis of the mutation in ABI3 locus . Similar to abi4 , abi3 also was found to show the decreased seed dormancy [18] . ABI3 , ABI4 and ABI5 were demonstrated to work in the same pathway in ABA signaling . Whether ABI5 is also involved in seed dormancy still need to be addressed in the future . Therefore , ABA signaling might also play a positive role during the control of seed dormancy . Taken together , the present study demonstrates that ABI4 positively regulates primary seed dormancy by mediating the biogenesis of ABA and GA . Further , this study also strongly suggests that ABI4 plays a pivotal role in these two signaling pathways . Further functional dissection of ABI4 during the biosynthesis and signaling of ABA and GA is necessary to obtain a deeper understanding of the crosstalk between these two hormones .
Arabidopsis ecotype Columbia-0 was used as the wild type in this study . The point mutant abi4-1 ( CS8104 ) and the T-DNA insertion mutants abi4-t ( SALK_080095 ) , cyp707a1 ( SALK_069127 ) and cyp707a2 ( SALK_083966C ) were obtained from the ABRC ( The Ohio State University , Columbus , OH , USA ) . It is noted that the T-DNA insertion mutant SALK_080095 was named as abi4-2 [47] . But the name of abi4-2 has been given much earlier to the other mutant harboring a point mutant in ABI4 gene [35] . Thus the T-DNA insertion line SALK_080095 was named as abi4-t in this work . The ga1-t mutant ( SALK_023192 ) in the Columbia-0 background was a gift from Dr . Xiangdong Fu ( Institute of Genetics and Developmental Biology , Chinese Academy of Sciences , Beijing ) . The abi1-1 , abi2-1 , snrk2 . 2/snrk2 . 3 mutants seeds were supplied by Dr . Zhizhong Gong ( College of Biological Sciences , China Agricultural University , Beijing ) . Arabidopsis seeds were surface-sterilized with 10% bleach and washed at least four times with sterile water . Sterile seeds were suspended in 0 . 2% agarose and sown on 1/2 MS medium plus 1% sucrose . The seeds were stratified on plates in the dark at 4°C for 0 or 3 days , depending on the experiment , and then transferred to a tissue culture room at 22°C under a 16-h-light/8-h-dark photoperiod . For ga1-t , the seeds were soaked in 100 µM GA solution for 3 days at 4°C , as the ga1-t mutant cannot germinate in the absence of exogenous GA . Normal 1/2 MS medium was supplemented with 1% sucrose and , unless otherwise noted , GA ( product number G7645 , Sigma-Aldrich Company ltd , USA ) or PAC ( product number 46046 , Sigma-Aldrich Company ltd , USA ) was added as needed . Transgenic plants carrying constitutively expressing ABI4 were generated . To produce 35S-ABI4 plants , the 987-bp CDS ( coding sequence ) fragment was amplified by PCR and then cloned into the vector pCanG-HA-GFP , in which ABI4 was expressed under the control of the CaMV 35S promoter . Transformation of Arabidopsis was performed by the vacuum infiltration method using the Agrobacterium tumefaciens strain EHA105 [61] . T2 seeds were germinated on MS plates containing 50 mg/mL kanamycin for vector pCanG-HA-GFP , and the resistant seedlings were transferred to soil to obtain homozygous T3 seeds . For more detailed phenotypic analysis , two independent T3 homozygous lines containing a single insertion were employed . To test germination rates , seeds were collected at the same time . Seeds subjected to various periods of dry storage were sown on normal 1/2 MS medium or 1/2 MS medium supplemented with various concentrations of GA or PAC . Radicle emergence was scored at the indicated time points , and at the same time , the percentages of cotyledon greening were also scored . For each germination test , approximately ≥45 seeds per genotype were used , and three experimental replications were performed . The average germination percentage ± SE ( standard error ) of triplicate experiments was calculated . For photography , a Leica MZ16 FA stereomicroscope was employed ( Leica Company , Germany ) . Photographs were taken using the same settings at the indicated time points . The vivipary assay was performed according to a previously described protocol [51] . Developing siliques at the long-green stage were collected from the same sites of plants with various genotypes , sterilized with 70% ethanol for 1 minute and 25% bleach for 10 minutes and plated on 1/2 MS medium or damp soil . Total RNA preparation ( from dry or imbibed seeds at various times ) , first-strand cDNA synthesis and qRT-PCR were performed as previously described [62] . DNase I-treated total RNA ( 2 µg ) was denatured and subjected to reverse transcription using Moloneymurine leukemia virus reverse transcriptase ( 200 units per reaction; Promega Corporation ) . Quantitative PCR was performed using the SsoFast EvaGreen Supermix ( Bio-Rad ) and a CFX96 Touch Real-Time PCR Detection System ( Bio-Rad ) . Gene expression was quantified at the logarithmic phase using the expression of the housekeeping 18S RNA as an internal control . Three biological replicates were performed for each experiment . Primer sequences for qRT-PCR are shown in Table S1 . To test the ABI4 protein levels in transgenic plants ( 35S-ABI4-GFP ) , western blotting was performed according to previously described protocols [62] , [63] . Approximately two-week-old seedlings grown on 1/2 MS medium were ground in liquid nitrogen and extracted with 4 M urea buffer . Crude extracts were separated by SDS-PAGE and transferred onto nitrocellulose membranes . The membranes were stained with 0 . 2% Ponceau S , with Rubisco serving as an internal control . The anti-GFP antibody was purchased from Santa Cruz Biotechnology , Inc . ChIP was performed as previously described [64] , with minor modifications . Transgenic seeds containing 35S-ABI4-GFP were grown on 1/2 MS medium for approximately 2 weeks . The seedlings were then harvested ( 1 . 5 g ) and crosslinked with 1% formaldehyde for 30 minutes under a vacuum; the crosslinking was stopped with 0 . 125 M glycine . The seedlings were ground in liquid nitrogen , and the nuclei were isolated . Immunoprecipitations were performed with the anti-GFP antibody and protein G beads . Immunoprecipitation in the absence of anti-GFP served as the control ( CK ) . DNA was precipitated by isopropanol , washed with 70% ethanol and dissolved in 10 µl water containing 20 µg/mL RNase . The qRT-PCR analysis was performed using specific primers corresponding to different promoter regions of CYP707A1 , CYP707A2 and CYP707A3 . TUB4 was used as an internal control . Since ABI4 directly binds to the promoter of ABI5 [50] , this promoter was employed as a positive control . Primers used in the ChIP-qPCR assay are shown in Table S1 . This transient expression assay was performed in N . benthamiana leaves as previously described [65] . The 2329 bp for native CYP707A1 promoter ( Pro-CYP707A1 ) and 2015 bp for native CYP707A2 ( Pro-CYP707A2 ) were amplified separately from genomic DNA . In addition , the several mutated CYP707A1 promoter fragments ( including Pro-CYP707A1 ( m1 ) , Pro-CYP707A1 ( m2 ) , Pro-CYP707A1 ( m1+m2 ) ) were generated by PCR amplification . All these five promoter fragments were cloned into pENTR using the pENTR Directional TOPO cloning kit ( Invitrogen ) . Then , these promoter versions were fused with the luciferase reporter gene LUC through the Gateway reactions into the plant binary vector pGWB35 [66]to generate the several reporters constructs . The effector construct was the pCanG-ABI4-GFP . For analysis of ABA content in dry or imbibed seeds , the seeds were ground in liquid nitrogen , and 150 mg of seed powder was homogenized and extracted for 24 h in methanol containing D6-ABA ( purchased from OIChemIm Co . Ltd . ) as an internal standard . Purification was performed with an Oasis Max solid phase extract cartridge ( 150 mg/6 cc; Waters ) and eluted with 5% formic acid in methanol . The elution was dried and reconstituted , and it was then injected into a liquid chromatography–tandem mass spectrometry system consisting of an Acquity ultra performance liquid chromatograph ( Acquity UPLC; Waters ) and a triple quadruple tandem mass spectrometer ( Quattro Premier XE; Waters ) . Three biological replications were performed . The endogenous gibberellins were determined by the method described [67] . Arabidopsis seeds ( 200 mg ) were frozen in liquid nitrogen , ground to fine powder , and extracted with 80% ( v/v ) methanol . GA isotope standards were added to plant samples before grinding . The crude extracts were purified by reversed-phase solid-phase extraction , ethyl ether extraction and derivatization . The resulting mixture was injected into capillary electrophoresis-mass spectrometry ( CE-MS ) for quantitative analysis . | Seed dormancy prevents or delays germination in maturated seeds . The optimal level of seed dormancy is a valuable trait for agricultural production and post-harvest management . High ABA and low GA content in seeds promote seed dormancy . However , the precise molecular mechanisms controlling seed dormancy and germination remain unclear . We found that ABI4 , the key transcription factor in the ABA signaling pathway , indeed controls primary seed dormancy . This result contradicts the previous conclusion that ABI4 is not involved in the control of seed dormancy . Several lines of evidence support our conclusion . For example , detailed physiological analysis of the germination of abi4 seeds that were harvested immediately and stored for various periods of time and subjected to various treatments allowed us to conclude that ABI4 negatively regulates primary seed dormancy . The molecular mechanism responsible for this control is as follows: ABI4 directly or indirectly regulates the key genes of the ABA and GA biogenesis pathways , which then regulates the ABA and GA contents in seeds . Importantly , further genetic interactions between CYP707A1 , CYP707A2 , GA1 , and ABI4 also support our conclusion . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"plant",
"science",
"plant",
"growth",
"and",
"development",
"plant",
"genetics",
"biology",
"plant",
"physiology"
] | 2013 | ABI4 Regulates Primary Seed Dormancy by Regulating the Biogenesis of Abscisic Acid and Gibberellins in Arabidopsis |
Inference about the causal structure that induces correlations between two traits can be achieved by combining genetic associations with a mediation-based approach , as is done in the causal inference test ( CIT ) . However , we show that measurement error in the phenotypes can lead to the CIT inferring the wrong causal direction , and that increasing sample sizes has the adverse effect of increasing confidence in the wrong answer . This problem is likely to be general to other mediation-based approaches . Here we introduce an extension to Mendelian randomisation , a method that uses genetic associations in an instrumentation framework , that enables inference of the causal direction between traits , with some advantages . First , it can be performed using only summary level data from genome-wide association studies; second , it is less susceptible to bias in the presence of measurement error or unmeasured confounding . We apply the method to infer the causal direction between DNA methylation and gene expression levels . Our results demonstrate that , in general , DNA methylation is more likely to be the causal factor , but this result is highly susceptible to bias induced by systematic differences in measurement error between the platforms , and by horizontal pleiotropy . We emphasise that , where possible , implementing MR and appropriate sensitivity analyses alongside other approaches such as CIT is important to triangulate reliable conclusions about causality .
We model a system whereby some exposure x has a causal influence βx on an outcome y such that y = α x + β x x + ϵ x In addition , the exposure is influenced by a SNP g with an effect of βg such that x = α g + β g g + ϵ g The α* terms represent intercepts , and henceforth can be ignored . The ϵ* terms denote random error , assumed independently and normally distributed with mean zero . Mediation-based analyses that test whether x causally relates to y rely on evaluating whether the influence of g on y can be accounted for by conditioning on x , such that c o v ( g , y - y ^ ) = 0 where y ^ = β ^ x x and assuming no intercept y - y ^ = ϵ x . MR analysis estimates the causal influence of x on y by using the instrument as a proxy for x , such that x ^ = β ^ g g y = β M R x ^ + ϵ M R where βMR ≠ 0 denotes the existence of causality , and βMR is an estimate of the causal effect . Measurement error of an exposure can be modeled as a transformation of the true value ( x ) that leads to the observed value , xo = f ( x ) . For example , following Pierce and VanderWeele [32] we can define f ( x ) = α m x + β m x x + ϵ m x where αmx and βmx influence the error in the measurement of x by altering its scale , and ϵmx represents the imprecision ( or noise ) in the measurement of x . Measurement imprecision can represent imprecise measurement due to limits on sensitivity of measuring equipment , or arise because of phenotypes being imprecisely defined . The same model of measurement error can be applied to the outcome variable y . In this study we assume there is no measurement error in the SNP . Common genetic variants are typically less susceptible to measurement error due to strict quality control procedures prior to genome wide association studies . Any non-differential measurement error that might be present ( either because the SNP is poorly typed or because the SNP is not in complete linkage disequilibrium with the causal variant ) will reduce power in MR but will not incur bias [3 , 13 , 32] . We also assume that measurement error in the exposure and the outcome are uncorrelated .
In the causal inference test ( CIT ) , the 4th condition ( see Methods ) employs mediation for causal inference , and can be expressed as c o v ( g , y - y ^ ) = 0 , where y ^ = α ^ x + β ^ x x o . When measurement error in scale and imprecision is introduced , such that yo is the measured value of y , it can be shown using basic covariance properties ( S1 Text ) that c o v ( g , y - y ^ ) = c o v ( g , y o ) - c o v ( g , y ^ o ) = β m y β g β x v a r ( g ) - D β m y β g β x v a r ( g ) where D = β m x 2 v a r ( x ) β m x 2 v a r ( x ) + v a r ( ϵ m x ) Thus an observational study will find c o v ( g , y o - y o ^ ) = 0 when the true model is causal only when D = 1 . Therefore , if there is any measurement error that incurs imprecision in x ( i . e . var ( ϵmx ) ≠ 0 ) then there will remain an association between g and yo|xo , which is in violation of the the 4th condition of the CIT . Note that scale transformation of x or y without any incurred imprecision is insufficient to lead to a violation of the test statistic assumptions , and henceforth mention of measurement error will relate to imprecision unless otherwise stated . We performed simulations to verify that this problem does arise using the CIT method . Fig 2 shows that when there is no measurement error in the exposure or outcome variables ( ρx , xo = ρy , yo = 1 ) the CIT is reliable in identifying the correct causal direction . However , as measurement error increases in the exposure variable , eventually the CIT is more likely to infer a robust causal association in the wrong direction . Also of concern here is that increasing sample size does not solve the issue , indeed it only strengthens the apparent evidence for the incorrect inference . If we do not know whether the SNP g has a primary influence on x or y then CIT can attempt to infer the causal direction . Though bi-directional MR can be used to orient causal directions [27] , this requires knowledge of a valid instrument for each trait , and we were motivated to develop the MR Steiger method that could operate on summary data to orient the direction of causality using the same conditions as the CIT , where the underlying biology of a single SNP is not fully understood . We go on to explore the scenarios in which the method is likely to return the correct or incorrect causal directions . We performed simulations to compare the power and type 1 error rates of MR and CIT in detecting a causal association between simulated variables under different levels of imprecision simulated in the exposure . Comparing the performance of methods with different sets of assumptions can be difficult , but a basic comparison is shown in Fig 3 . We observe that the CIT is more conservative under the null model of no association owing to the omnibus test statistic comprising several statistical tests . The FDR using a p-value threshold of 0 . 05 appears to be close to zero , whereas for the MR Steiger method the FDR is around 0 . 05 . Using the same p-value thresholds to declare significance in the non-null simulations , the general trend appears to be that the CIT power reduces as measurement error in the exposure increases more steeply than that of the MR Steiger method . For a particular association , it is of interest to identify the range of possible measurement error values for which the method will give results that agree or disagree with the empirically inferred causal direction ( Fig 4a , S2 Text ) . This metric can be used to evaluate the reliability of MR Steiger test . We show that in the presence of measurement imprecision , d = ρx , xo − ρx , yρy , yo ( S2 Text ) determines the range of parameters around which the MR Steiger test is liable to provide the wrong direction of causality ( i . e . if d > 0 then the MR Steiger test is likely to be correct about the causal direction ) . Fig 4b shows that when there is no measurement error in x , the MR Steiger test is unlikely to infer the wrong direction of causality even if there is measurement error in y . It also shows that in most cases where x is measured with error , especially when the causal effect between x and y is not very large , the sensitivity of the MR Steiger test to measurement error is relatively low . Unmeasured confounding between the exposure and outcome can also give rise to problems with the MR Steiger approach ( S3 Text ) . The relationship between unmeasured confounding and causal orientation is complex across the parameter space of possible confounding values ( S2 Fig ) . Based on the range of parameter values that we explored , when the magnitude of the observational variance explained between the exposure and the outcome is below 0 . 2 the MR Steiger method is unlikely to return the incorrect causal direction due to unmeasured confounding . We used simulations to explore the performance of the MR Steiger approach in comparison to CIT for different levels of measurement error . The performance was compared in terms of the rate at which evidence of a causal relationship is obtained for the correct direction of causality , and the rate at which evidence of a causal relationship is obtained where the reported direction of causality is incorrect . Simulations were performed for two models , one for a “causal model” where there was a causal relationship between x and y; and one for a “non-causal model” where x and y were not causally related , but had a confounded association induced by the SNP g influencing x and y independently . Fig 5a shows that , for the “causal model” , the MR analysis is indeed liable to infer the wrong direction of causality when d < 0 , and that this erroneous result is more likely to occur with increasing sample size . However , the CIT is in general more fallible to reporting a robust causal association for the wrong direction of causality . When d > 0 we find that in most cases the MR Steiger method has greater power to obtain evidence for causality than CIT , and always obtains the correct direction of causality . The CIT , unlike the MR Steiger test , is able to distinguish the “non-causal model” from the “causal model” ( Methods , Fig 5b ) , but it is evident that measurement error will often lead the CIT to identify the causal model as true , when in fact the underlying model is this non-causal model . We used the MR Steiger test to infer the direction of causality between DNA methylation and gene expression levels between 458 putative associations . We found that the causal direction commonly goes in both directions ( Fig 6a ) , but assuming no or equal measurement error , DNA methylation levels were the predominant causal factor ( p = 1 . 3 × 10−5 ) . The median reliability ( R ) of the 458 tests was 3 . 92 ( 5%-95% quantiles 1 . 08–37 . 11 ) . We then went on to predict the causal directions of the associations for varying levels of systematic measurement error for the different platforms . Fig 6a shows that the conclusions about the direction of causality between DNA methylation and gene expression are very sensitive to measurement error . We made a strong assumption that either methylation influenced gene expression or vice versa , but it is certainly possible that the SNP is solely or additionally influencing some other trait that confounds the association between gene expression and DNA methylation . We performed two sample MR [30] for each association in the direction of causality inferred by the Stieger test . We observed that the sign of the MR estimate was generally in the same direction as the Pearson correlation coefficient reported by Shakhbazov et al [39] ( Fig 6b ) . There was a moderate correlation between the absolute magnitudes of the causal correlation and the observational Pearson correlation ( r = 0 . 45 ) . Together these inferences suggest that even in estimating associations between ‘omic’ variables , which are considered to be low level phenotypes , it is important to use causal inference methods over observational associations to infer causal effect sizes . We also observed that for associations where methylation caused gene expression the causal effect was more likely to be negative than for the associations where gene expression caused methylation ( OR = 0 . 61 ( 95% CI 0 . 36–1 . 03 ) , Fig 6c ) , suggesting that reducing methylation levels at a controlling CpG typically leads to increased gene expression levels , consistent with expectation [40] .
Researchers are often confronted with the problem of making causal inferences using a statistical framework on observational data . In the epidemiological literature issues of measurement error in mediation analysis are relatively well explored [41] . Our analysis extends this to related methods such as CIT that are used in predominantly ’omic data . These methods are indeed susceptible to the same problem as standard mediation based analysis , and specifically we show that as measurement error in the ( true ) exposure variable increases , CIT is likely to have reduced statistical power , and liable to infer the wrong direction of causality . We also demonstrate that , though unintuitive , increasing sample size does not resolve the issue , rather it leads to more extreme p-values for the model that predicts the wrong direction of causality . Under many circumstances a practical solution to this problem is to use Mendelian randomisation instead of methods such as the CIT or similar that are based on mediation . Inferring the existence of causality using Mendelian randomisation is robust in the face of measurement error and , if the researcher has knowledge about the biology of the instrument being used in the analysis , can offer a direct solution to the issues that the CIT faces . This assumption is often reasonable , for example SNPs are commonly used as instruments when they are found in genes with known biological relevance for the trait of interest . But on many occasions , especially in the realm of ’omic data , this is not the case , and methods based on mediation have been valuable in order to be able to both ascertain if there is a causal association and to infer the direction of causality . Here we have described a simple extension to MR which can be used as an alternative to or in conjunction with mediation based methods . We show that this method is still liable to measurement error , but because it has different properties to the CIT it offers several main advantages . First , it uses a formal statistical framework to test for the reliability of the assumed direction of causality . Second , after testing in a comprehensive range of scenarios the MR based approach is less likely to infer the wrong direction of causality compared to CIT , while substantially improving power over CIT in the cases where d > 0 . We demonstrate this new method by evaluating the causal relationships of 458 known associations between DNA methylation and gene expression levels using summary level data . The inferred causal direction is heavily influenced by how much measurement error is present in the different assaying platforms . For example , if DNA methylation measures typically have lower or equal measurement error compared to gene expression measures then our analysis suggests that DNA methylation levels would be more often the causal factor in the association . Indeed , previous studies which have evaluated measurement error in these platforms do support this position [42 , 43] , though making strong conclusions for this analysis is difficult because measurement error is likely to be study specific . We also haven’t accounted for the influence of winner’s curse , which can inflate estimates of the variance explained by SNPs , with higher inflation expected amongst lower powered studies . Using p-values for genetic associations from replication studies will mitigate this problem . In our simulations we focused on the simple case of a single instrument in a single sample setting with a view to making a fair comparison between MR and the various mediation-based methods available . However , if there is only a single instrument it is difficult to separate between the two competing models of g instrumenting a trait which causes another trait , and g having pleiotropic effects on both traits independently [44] . Under certain conditions of measurement error the CIT test can distinguish these models . We also note that it is straightforward to extend the MR Steiger approach to multiple instruments , requiring only that the total variance explained by all instruments be calculated under the assumption that they are independent . Multiple instruments can indeed help to distinguish between the causal and pleiotropic models , for example by evaluating the proportionality of the SNP-exposure and SNP-outcome effects [16] . Additionally , if there is at least one instrument for each trait then bi-directional MR can offer solutions to inferring the causal direction [16 , 28 , 45] . We restricted the simulations to evaluating the causal inference between quantitative traits , but it is possible that the analysis could be extended to binary traits by using the genetic variance explained on the liability scale , taking into account the population prevalence [46] . However , our analysis goes beyond many previous explorations of measurement error by assessing the impacts of both imprecision ( noise ) and linear transformations of the true variable on causal inference . Our new method attempts to infer causal directions under the assumption that horizontal pleiotropy ( the influence of the instrument on the outcome through a mechanism other than the exposure ) is not present . Recent method developments in MR [24 , 25] have focused on accounting for the issues that horizontal pleiotropy can introduce when multiple instruments are available , but how they perform in the presence of measurement error remains to be explored . An important advantage that MR confers over most mediation based analysis is that it can be performed in two samples , which can considerably improve power and expand the scope of analysis . However , whether there is a substantive difference in two sample MR versus one sample MR in how measurement error has an effect is not yet fully understood . We have also assumed no measurement error in the genetic instrument , which is not unreasonable given the strict QC protocols that ensure high quality genotype data is available to most studies . We have restricted the scope to only exploring non-differential measurement error and avoided the complications incurred if measurement error in the exposure and outcome is correlated . We have also not addressed other issues pertaining to instrumental variables which are relevant to the question of instrument-exposure specification . One such problem is exposure misspecification , for example an instrument could associate with several closely related putative outcomes , with only one of them actually having a causal effect on the outcome . This problem has shown to be the case for SNPs influencing different lipid fractions , for example [47 , 48] . Mediation based network approaches , that go beyond analyses of two variables , are very well established [37] and have a number of extensions that make them valuable tools , including for example network construction . But because they are predicated on the basic underlying principles of mediation they are liable to suffer from the same issues of measurement error . Recent advances in MR methodology , for example applying MR to genetical genomics [49] , multivariate MR [48] and mediation through MR [50–52] may offer more robust alternatives for these more complicated problems . The overarching result from our simulations is that , regardless of the method used , inferring the causal direction using an instrument of unknown biology is highly sensitive to measurement error . With the presence of measurement error near ubiquitous in most observational data , and our ability to measure it limited , we argue that it needs to be central to any consideration of approaches which are used in attempt to strengthen causal inference , and any putative results should be accompanied with appropriate sensitivity analysis that assesses their robustness under varying levels of measurement error .
Here we describe how the CIT method [4] is implemented in the R package R/cit [18] . Assume an exposure x is instrumented by a SNP g , and the exposure x causes an outcome y , as described above . The following tests are then performed: The term in the 4th test can be rewritten as c o v ( g , y | x ) = c o v ( g , y - y ^ ) where y - y ^ = y - ( α ^ g + β ^ g x ) is the residual of y after adjusting for x , and x is assumed to mediate the association between the SNP and the outcome . The condition in the 4th test is formulated as an equivalence testing problem that is estimated using simulations , comparing the estimate from the data against empirically obtained estimates for simulated variables where the independence model is true ( full details are given in [4] ) . We note here that this approach is liable to fail , even when there is a true causal relationship , when confounders of the exposure and outcome are present , as these will induce collider bias . If all four tests reject the null hypothesis then it is inferred that x causes y . The CIT measures the strength of causality by generating an omnibus p-value , pCIT , which is simply the largest ( least extreme ) p-value of the four tests , the intuition being that causal inference is only as strong as the weakest link in the chain of tests . Now we describe how we used the CIT method in our simulations . The cit . cp function was used to obtain an omnibus p-value . To infer the direction of causality using the CIT method , an omnibus p-value generated by CIT for each of two tests—pCIT , x→y , was estimated for the direction of x causing y ( Model 1 ) , and for the direction of y causing x , pCIT , y→x ( Model 2 ) . The results from each of these methods can then be used in combination to infer the existance and direction of causality . For some significance threshold α there are four possible outcomes from these two tests , and their interpretations are as follows: For the purposes of compiling simulation results we use an arbitrary α = 0 . 05 value , though we stress that for real analyses it is not good practice to rely on p-values for making causal inference , nor is it reliable to depend on arbitrary significance thresholds [53] . Two stage least squares ( 2SLS ) is a commonly used technique for performing MR when the exposure , outcome and instrument data are all available in the same sample . A p-value for this test , pMR , was obtained using the systemfit function in the R package R/systemfit [54] . Note that the value of pMR is identical when using the same genetic variant to instrument the influence of the exposure x on the outcome y , or erroneously , instrumenting the outcome y on the exposure x . The method that we will now describe is designed to distinguish between two models , x → y or y → x . Unlike the CIT framework , this approach cannot infer if the true model is x ← g → y . We also assume all genetic effects are additive . To infer the direction of causality it is desirable to know which of the variables , x or y , is being directly influenced by the instrument g . This can be achieved by assessing which of the two variables has the biggest absolute correlation with g ( S2 Text ) , formalised by testing for a difference in the correlations ρgx and ρgy using Steiger’s Z-test for correlated correlations within a population [55] . It is calculated as Z = ( Z g x - Z g y ) N - 3 2 ( 1 - ρ x y ) h where Fisher’s z-transformation is used to obtain Z g * = 1 2 ln ( 1 + ρ g * 1 - ρ g * ) , h = 1 - ( f r m 2 ) 1 - r m 2 where f = 1 - ρ x y 2 ( 1 - r m 2 ) and r m 2 = 1 2 ( ρ g x 2 + ρ g y 2 ) . The Z value is interpreted such that Z { > 0 , x → y < 0 , y → x = 0 , x ⊥ ⊥ y and a p-value , pSteiger is generated from the Z value to indicate the probability of obtaining a difference between correlations ρgx and ρgy at least as large as the one observed , under the null hypothesis that both correlations are identical . The existence of causality and its direction is inferred based on combining information from the MR analysis and the Steiger test . The MR analysis indicates whether there is a potential causal relationship ( pMR ) , and the Steiger test indicates the direction ( sign ( Z ) ) of the causal relationship and the confidence of the direction ( pSteiger ) . For the purposes of compiling simulation results , these can be combined using an arbitrary α = 0 . 05 value: Note that the same correlation test approach can be applied to a two-sample MR setting . Two-sample MR refers to the case where the SNP-exposure association and SNP-outcome association are calculated in different samples ( e . g . from publicly available summary statistics [26 , 30] ) . Here the Steiger test of two independent correlations can be applied where . Z = Z g x - Z g y 1 / ( N 1 - 3 ) + 1 / ( N 2 - 3 ) An advantage of using the Steiger test in the two sample context is that it can compare correlations in independent samples where sample sizes are different . Steiger test statistics were calculated using the r . test function in the R package R/psych [56] . The Steiger test assumes that there is a causal relationship between the two variables , and that the SNP is a valid instrument for one of them . However it is liable to give incorrect causal directions under some other circumstances . First , some levels of horizontal pleiotropy , where the SNP influences the outcome through some pathway other than the exposure , could induce problems because this is a means by which the instrument is invalid . Second , some differential values of measurement error between the exposure and the outcome could lead to incorrect inference of the causal direction ( S2 Text ) . Third , some levels of unmeasured confounding between the exposure and the outcome could lead to inference of the wrong causal direction ( S3 Text ) . The Steiger test for inferring if x → y is based on evaluating ρgx > ρgy . However , ρgx ( or ρgy ) are underestimated if x ( or y ) are measured imprecisely . If , for example , x has lower measurement precision than y then we might empirically obtain ρg , xo < ρg , yo because ρg , xo could be underestimated more than ρg , yo . As we show in S2 Text it is possible to infer the bounds of measurement error on xo or yo given known genetic associations . The maximum measurement imprecision of xo is ρg , xo , because it is known that at least that much of the variance has been explained in xo by g . The minimum is 0 , denoting perfectly measured trait values ( the same logic applies to yo ) . It is possible to simulate what the inferred causal direction would be for all values within these bounds . To evaluate how reliable , R , the inference of the causal direction is to potential measurement error in x and y we need to predict the values of ρgy − ρgx for those values of measurement error . We offer two tools in which to do this . First , the user can provide values of measurement error for x and y and obtain a revised inference of the causal direction . Second , we integrate over the entire range of ρgy − ρgx values for possible measurement error values , assuming that any measurement error value is equally likely . Across all possible values of measurement error in x and y we find the volume that agrees with the inferred direction of causality and the volume that disagrees with the inferred direction of causality , and take the ratio of these two values . A ratio R = 1 indicates that the inferred causal direction is highly sensitive to measurement error , because equal weight of the measurement error parameter space supports each direction of causality . In general , the R value denotes that the inferred direction of causality is R times more likely to be the empirical result than the opposite direction ( S2 Text ) . Simulations were conducted by creating variables of sample size n for the exposure x , the measured values of the exposure xo , the outcome y , the measured values of the outcome yo and the instrument g . One of two models are simulated , the “causal model” where x causes y and g is an instrument for x; or the “non-causal model” where g influences a confounder u which in turn causes both x and y . Here x and y are correlated but not causally related . Each variable in the causal model was simulated such that: g ∼ B i n o m ( 2 , 0 . 5 ) x = α g + β g g + ϵ g x o = α m x + β m x x + ϵ m x y = α x + β x x + ϵ x y o = α m y + β m y y + ϵ m y where non-differential measurement error is represented by a noise ( measurement imprecision ) term ϵ m * ∼ N ( 0 , σ m * 2 ) , and measurement bias terms αm* and βm* for the exposure variable x and the outcome variable y . Note that following the first section of the Results we no longer include the bias terms for simplicity . We have formulated the non-causal model as: y = α g y + β g y g + ϵ g y All α values were set to 0 , and β values set to 1 . Normally distributed values of ϵ* were generated such that c o r ( g , x ) 2 = 0 . 1 c o r ( x , y ) 2 = { 0 . 2 , 0 . 4 , 0 . 6 , 0 . 8 } σ m x 2 = { 0 , 0 . 2 , 0 . 4 , 0 . 6 , 0 . 8 , 1 } σ m y 2 = { 0 , 0 . 2 , 0 . 4 , 0 . 6 , 0 . 8 , 1 } n = { 100 , 1000 , 10000 } giving a total of 432 combinations of parameters . Simulations using each of these sets of variables were performed 100 times , and the CIT and MR methods were applied to each in order to evaluate the causal association of the simulated variables . Similar patterns of results were obtained for different values of cor ( g , x ) . Two sample MR [30] was performed using summary statistics for genetic influences on gene expression and DNA methylation . To do this we obtained a list of 458 gene expression—DNA methylation associations as reported in Shakhbazov et al [39] . These were filtered to be located on the same chromosome , have robust correlations after correcting for multiple testing , and to share a SNP that had a robust cis-acting effect on both the DNA methylation probe and the gene expression probe . Because only summary statistics were available ( effect , standard error , effect allele , sample size , p-values ) for the instrumental SNP on the methylation and gene expression levels , the Steiger test of two independent correlations was used to infer the direction of causality for each of the associations . The Wald ratio test was then used to estimate the causal effect size for the estimated direction for each association . All analysis was performed using the R programming language [57] and code is made available at https://github . com/explodecomputer/causal-directions and implemented in the MR-Base ( http://wwww . mrbase . org ) platform [26] . | Understanding the causal relationships between pairs of traits is crucial for unravelling the causes of disease . To this end , results from genome-wide association studies are valuable because if a trait is known to be influenced by a genetic variant then this knowledge can be used to test the trait’s causal influences on other traits and diseases . Here we discuss scenarios where the nature of the genetic association with the causal trait can lead existing causal inference methods to give the wrong direction of causality . We introduce a new method that can be applied to summary level data and is potentially less susceptible to problems such as measurement error , and apply it to evaluate the causal relationships between DNA methylation levels and gene expression . While our results show that DNA methylation is more likely to be the causal factor , we point out that is it crucial to acknowledge that systematic differences in measurement error between the platforms could influence such conclusions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"infographics",
"genome-wide",
"association",
"studies",
"social",
"sciences",
"simulation",
"and",
"modeling",
"mathematics",
"statistics",
"(mathematics)",
"test",
"statistics",
"genome",
"analysis",
"epigenetics",
"dna",
"molecular",
"genetics",
"dna",
"methylation",
"... | 2017 | Orienting the causal relationship between imprecisely measured traits using GWAS summary data |
Developmental constraints have been postulated to limit the space of feasible phenotypes and thus shape animal evolution . These constraints have been suggested to be the strongest during either early or mid-embryogenesis , which corresponds to the early conservation model or the hourglass model , respectively . Conflicting results have been reported , but in recent studies of animal transcriptomes the hourglass model has been favored . Studies usually report descriptive statistics calculated for all genes over all developmental time points . This introduces dependencies between the sets of compared genes and may lead to biased results . Here we overcome this problem using an alternative modular analysis . We used the Iterative Signature Algorithm to identify distinct modules of genes co-expressed specifically in consecutive stages of zebrafish development . We then performed a detailed comparison of several gene properties between modules , allowing for a less biased and more powerful analysis . Notably , our analysis corroborated the hourglass pattern at the regulatory level , with sequences of regulatory regions being most conserved for genes expressed in mid-development but not at the level of gene sequence , age , or expression , in contrast to some previous studies . The early conservation model was supported with gene duplication and birth that were the most rare for genes expressed in early development . Finally , for all gene properties , we observed the least conservation for genes expressed in late development or adult , consistent with both models . Overall , with the modular approach , we showed that different levels of molecular evolution follow different patterns of developmental constraints . Thus both models are valid , but with respect to different genomic features .
Developmental constraints have been suggested to play an important role in shaping the evolution of embryonic development in animals . Briefly , the concept of developmental constraints assumes that the scope of developmental mechanisms limits the set of phenotypes that may evolve . Thus , morphological similarities between embryos of different species could reflect these underlying constraints [1] . Two main models of embryonic developmental constraints have been put forward . The early conservation model predicts that the highest developmental constraints occur at the beginning of embryogenesis . This corresponds to von Baer's third law [2] , postulating that embryos of different species progressively diverge from one another during ontogeny . However , in modern times , the highest morphological similarity between embryos of different species was observed in the phylotypic stage ( i . e . , mid-embryogenesis ) [3]–[5] . Consequently , Duboule [6] and Raff [7] proposed the so-called hourglass model , which has since become widely accepted ( see , e . g . , [8] , [9] ) . It predicts the highest developmental constraints during mid-embryogenesis . At the genomic level , the hourglass model was originally linked to the expression of Hox genes in animals [6] . More recently , the emphasis has shifted to the relation , if any , between developmental constraints and the evolution and function of the genome ( reviewed in [9] ) . Different studies have reported several characteristics supporting the hourglass model in animals on the genomic level . Hazkani-Covo et al . [10] reported the highest protein sequence similarity between mouse and human for genes expressed in mid-development . In two influential papers , Domazet-Lošo and Tautz [11] reported that the genes expressed in mid-development of zebrafish are older than genes expressed early or late , while Kalinka et al . [12] showed that genes expressed in mid-development of fruit flies have the highest expression conservation . Similarly , Irie and Kuratani [13] reported the highest expression conservation between zebrafish , frog , chicken and mouse , for genes expressed in mid-development . Very recently , the hourglass model was argued to hold also for plants embryogenesis with respect to gene age and sequence conservation [14] . However , some of these results do not hold out under detailed analyses ( see Box 1 and Text S1 ) . For example , applying a standard log-transformation [15] , [16] to microarray signal intensities used in [11] changes the reported pattern such that it no longer supports the hourglass model ( Figure 1 ) . Moreover , other studies have also found genetic patterns supporting an early conservation model [17] , [18] . In most of the studies of developmental constraints the authors compared descriptive statistics of all genes across all developmental time-points ( e . g . , median expression [17] , weighted mean age [11] , mean expression correlation [13] ) . Such an approach introduces dependencies between the sets of genes which are compared , and consequently can produce results biased by genes expressed at many time-points . For example , housekeeping genes contribute to the average gene expression at all time points , and hence dilute trends . To overcome this essential problem , we have used a modularization approach , which we applied to the recently published transcriptome data of zebrafish development [11] . We decomposed the genes into independent sets , i . e . , modules , that contained genes overexpressed solely in one of seven developmental stages: cleavage/blastula , gastrula , segmentation , pharyngula , larva , juvenile and adult . This decomposition allowed us to compare only sets of genes that have specific functions during embryonic development . For each of the seven modules , we studied five properties of its genes: 1 ) gene sequence conservation , 2 ) gene age , 3 ) gene expression conservation , 4 ) gene orthology relationships , and 5 ) regulatory elements conservation . Here , we show that different levels of molecular evolution follow different patterns of developmental constraints . First , the regulatory elements are most conserved for transcription factors expressed at mid-development , consistent with the hourglass model . Contrary to what has been reported previously [10] , [11] , [13] , we did not detect the hourglass pattern for gene sequence , age and expression . Second , constraints on gene duplication and on new gene introduction are the strongest in early development , supporting the early conservation model ( consistent with [17] ) . Finally , all gene properties displayed the least conservation in late development and adult , which is in agreement with both models of developmental constraints .
Our goal was to analyze the developmental constraints acting on different gene properties . To this end we identified and analyzed groups of genes co-expressed during distinct developmental stages . We applied the Iterative Signature Algorithm ( ISA ) [19] , [20] to the zebrafish expression data published by Domazet-Lošo and Tautz [11] , which measured the dynamics of the transcriptome during development with a resolution of 60 time points . The ISA is a modularization algorithm that finds genes with similar expression profiles and groups them into so-called transcription modules . In order to detect modules of genes with specific expression during the zebrafish development , we initialized the ISA with seven idealized expression profiles that corresponded to successive developmental stages ( see Text S1 and Figure S10 ) . We obtained seven modules , each containing genes overexpressed during one of the following developmental stages: cleavage/blastula , gastrula , segmentation , pharyngula , larva , juvenile and adult ( Figure 2 ) . Overall , the modules covered the entire development . The phylotypic stage in which the hourglass model predicts the highest evolutionary constraint corresponds to the segmentation and pharyngula modules . We will refer to these two modules as phylotypic modules . The cleavage/blastula and gastrula modules will be referred to as early modules , and larva , juvenile and adult modules as late modules . The adjacent modules partially overlapped in their gene content . In order to allow for unbiased cross-module comparisons , genes belonging to two modules were kept in the one with the highest ISA gene score ( see Methods ) ; this concerned 534 genes in total . The seven modules , i . e . , cleavage/blastula , gastrula , pharyngula , segmentation , larva , juvenile and adult , contained 444 , 820 , 487 , 414 , 415 , 290 and 207 genes , respectively ( see Table S3 for the lists of the genes ) . Overall , 3077 different genes were present in these modules , which implies a significant reduction of the number of genes being analyzed in comparison to the original data ( 14293 genes on the microarray ) . In particular , the ISA removed the bias related to the genes expressed uniformly across development ( like housekeeping genes ) . We verified the function of genes in modules detected by the ISA by comparing them to relevant known lists of genes . We found that the cleavage/blastula module was significantly enriched in maternal genes identified in [21] ( 36 genes vs . 19 expected by chance; hypergeometric test , ) , and the gastrula module was highly significantly enriched in post-midblastula transition ( post-MBT ) genes identified in [21] ( 78 genes vs . 25 expected by chance; hypergeometric test , ) . We confirmed the relevance of the segmentation and pharyngula modules by verifying that they were enriched in Hox genes ( 24 and 7 genes vs . 1 expected by chance , respectively; hypergeometric test , and , respectively ) , which is consistent with their role in mid-development [22] . We did not have any gold standard for genes expressed at the late stages of development . However , since the early and phylotypic modules were enriched in genes with relevant functions , we are confident that the same is true for the late modules . Moreover , gene ontology ( GO ) enrichment analysis confirmed that genes from the modules were enriched in functions relevant to the respective developmental stages . For example , the cleavage/blastula module was enriched in genes involved in protein phosphorylation and dephosphorylation processes , which is consistent with kinase-dependent control of cell cycle and regulation of mid-blastula transition ( MBT ) in vertebrates [23] , [24] . The pharyngula module was enriched in genes associated with cell differentiation , and anatomical structure development . Finally , the adult module was enriched in genes involved in responses to environment , although not significantly ( Table S2 ) . We checked whether the sequences of genes from different modules evolved under different selective pressure . To this end , we calculated the non-synonymous to synonymous substitution ratios ( ) for genes in the modules and asked if the ratio was significantly lower for any of them . With the early conservation model , we would expect the lowest values for genes from early modules . Whereas with the hourglass model , we would expect the lowest values for genes from the phylotypic modules . In the cleavage/blastula module the median was not different from the median for all genes ( equal to 0 . 15 ) . In the other four modules covering embryonic development the median was lower than the median for all genes ( Figure 3A ) , and the difference was significant for all but the segmentation module ( randomization test , for the gastrula , pharyngula and larva modules ) . In the juvenile module , the median was significantly higher than the median for all genes ( randomization test , ) . In the adult module , the median was also higher than the median for all genes , but the difference was not significant . When analyzing separately sites under purifying selection or evolving neutrally , we also find weaker purifying selection during post-embryonic stages ( see Text S1 and Figure S11 ) . These results were consistent with the study by Roux and Robinson-Rechavi [17] , who also reported equally low values during the entire zebrafish embryogenesis , and a small increase in mid-larva , juvenile and adult . In contrast , Hazkani-Covo et al . [10] reported an hourglass pattern for protein distance between mouse and human genes expressed during development . However , the trend was not significant . In [17] some evidence for early conservation was reported in mouse . Projecting the genes from zebrafish modules to mouse-human orthologs , we found equal conservation across development ( Figure S12 ) . Overall , data analyses support similar evolutionary constraints on sequences of genes expressed during whole embryogenesis of zebrafish , while for mouse more developmental data is needed to be conclusive . The differences in age of genes expressed during different stages of the development have been suggested to be a good indicator of evolutionary constraints [11] , [25] . Thus , we investigated the age of genes belonging to different modules . We dated each gene by its first appearance in the phylogeny and assigned it to one of the five age groups: 1 ) Fungi/Metazoa , 2 ) Bilateria , 3 ) Coelomata+Chordata , 4 ) Euteleostomi and 5 ) Clupeocephala+Danio rerio . Next , for each module we calculated the age distribution of its genes , i . e . , the number of genes belonging to each age group , and compared it with the age distribution of all genes . For all but the cleavage/blastula module we detected significant age variations which differed across modules ( Figure 3B; chi-square goodness of fit test , all ) . The oldest genes which belong to the Fungi/Metazoa class were overrepresented in the gastrula module ( 36 . 7% of genes in the module vs . 25 . 7% of all genes ) . The younger Bilateria genes were overrepresented in the phylotypic modules ( 45 . 5% and 52 . 1% of genes in the segmentation and pharyngula modules , respectively , vs . 34 . 4% of all genes ) . The youngest genes were overrepresented in the late modules ( e . g . , for Euteleostomi genes: 25 . 7% , 35 . 1% and 35 . 6% of such genes in larva , juvenile and adult modules , respectively , vs . 18% of all genes ) . In contrast , Domazet-Lošo and Tautz [11] reported that genes expressed in early and late development tend to be younger than genes expressed in mid-development , supporting the hourglass model . Yet , that result does not hold for log-transformed gene expression levels ( Box 1 ) , and is not recovered with measures of gene age other than the transcriptome age index ( see Text S1 and Figure S6 ) . With the modular approach we observed that the age of expressed genes decreased throughout ontogeny . This pattern suggests that the oldest evolutionary stages tend to express the oldest genes . Both gene duplication and gene loss can impact phenotypic evolution [26]–[30] . The outcome of these events can be summarized by the resulting gene family size . Consequently , constrained developmental stages should display less changes in gene family size than other stages . To test this hypothesis , for each zebrafish module we calculated the number of its genes that were in 1 ) one-to-one , 2 ) one-to-many , 3 ) many-to-many , and 4 ) no orthology relation to mouse genes ( i . e . , no ortholog detectable by the criteria used in Ensembl Compara [31] ) . We compared the observed distributions with the distribution of the ortholog relationships for all genes . We detected significant variations of the ortholog relationship for the cleavage/blastula module and for all three late modules ( chi-square goodness of fit test , all ) . Moreover , the pattern of variation itself differed across different modules . The number of one-to-one orthologs decreased throughout development ( Figure 3C ) . It was significantly higher than expected only in the cleavage/blastula module ( 54 . 6% of genes in the module vs . 45 . 4% of all genes ) . In contrast , the number of genes with no orthologous relationship increased throughout development ( Figure 3C ) . It was significantly higher than expected only in the juvenile and adult modules ( 38 . 2% and 38 . 4% of genes in the two modules , respectively , vs . 20 . 4% of all genes ) , consistent with the excess of “young” genes . A similar pattern was observed for many-to-many orthologs ( 10 . 4% and 7 . 8% of genes in the two modules , respectively , vs . 3 . 9% of all genes ) . Finally , the number of one-to-many orthologs was higher than expected only in the larva module ( 45 . 6% of genes in the module vs . 30 . 3% of all genes ) , and did not differ from expectation in all other modules . These results were consistent with [17] in which the genes retained in duplicates after the teleost-specific whole genome duplication were reported to have low expression early in the development . Here , we recovered an analogous pattern with the modular approach , showing that the genes expressed early in the development are retained in duplicates less often than genes expressed later . Note that our observation is not limited to whole genome duplication . In addition , we detected the highest number of novel genes amongst genes expressed late in the development . Changes in gene expression are one of the main sources of morphological variation [32]–[34] . The developmental constraints on gene expression might differ from those on the gene sequence [35]–[37] . Thus , for each module , we compared the mean expression profile of its genes with the mean expression profile of their one-to-one orthologs in mouse . We used two different data sets [13] , [38] with expression values of mouse genes during the development . The use of two data sets was necessary , because there does not exist a single experiment covering the entire mouse development . The incompatibility of the two microarrays impaired the statistical strength of the analysis . For this reasons the results reported here should be regarded rather as qualitative than quantitative . Since homology cannot be defined for individual developmental stages between zebrafish and mouse , we first mapped every time point to its broad metastage defined in Bgee database [39] ( Figure 4 ) . Next , we calculated the mean expression level in every metastage . This resulted in six expression values for each gene during the development of mouse and zebrafish: zygote , cleavage , blastula , neurula , organogenesis , and post-embryonic stage . Note that the mouse microarrays did not cover the gastrula stage at all . For each module we calculated the Pearson's correlation between the mean expression of its genes and their mouse orthologs across the six metastages . For the cleavage/blastula module no correlation was detected , probably due to the incompatibility of the two mouse microarrays . Nevertheless , there exists a plausible , biological interpretation of the differences in gene expression between the early stages of zebrafish and mouse development . Zebrafish and mouse form two different embryological structures during blastulation , a blastula and a blastocyst , respectively . The blastocyst is a mammalian innovation that consists of an embryoblast ( that develop into structures of the fetus ) and a trophoblast ( that form the extraembryonic tissue ) . In contrast , there is no extraembryonic tissue in zebrafish . Overall , the lack of correlation between gene expression for the early stages of mouse and zebrafish development could be explained by these structural differences . For other modules the correlation was positive ( Figure 3D ) , however due to the low number of data points in the analysis , no correlation values were significant ( all ) . These results stood in contrast with the report by Irie and Kuratani [13] who showed the highest conservation of gene expression in mid-development . However , a re-analysis of their data suggested that this observation was not significant ( see Text S1 and Figure S9 ) . Also , both their and our studies shared problems related to the use of two data sets from different sources to cover mouse development . This and the lack of a straightforward homology between ontogenies of different species made it difficult to conclude on the conservation of gene expression during vertebrate development . The cis-regulatory hypothesis asserts that most morphological evolution is due to changes in cis-regulatory sequences [40]–[42] . A reasonable prediction of this hypothesis is slower cis-element turnover in morphologically conserved developmental periods . We examined the presence of highly conserved non-coding elements ( HCNEs ) [43] and of transposon-free regions ( TFRs ) [44] in the proximity of genes from each module . In the analysis of HCNEs , we counted their number between zebrafish and mouse ( detected with 70% identity ) in regions of 500 base pairs upstream from the transcription start site . We found that only genes from the phylotypic modules were significantly enriched in HCNEs ( hypergeometric test , , and for segmentation and pharyngula modules , respectively ) . We tested the sensitivity of the results by changing the analyzed regions' length to 200 and 1000 base pairs upstream from the transcription start site , by looking for HCNEs in introns , and using HCNEs detected with identity of 90% . In all cases , we obtained similar results ( see Table S1 ) . In the analysis of TFRs , we counted the number of genes from each module that have been associated with TFRs in zebrafish . Importantly , these TFRs were reported to be conserved between vertebrates as distant as zebrafish and human . We found that only genes from the pharyngula module were significantly enriched in TFRs ( hypergeometric test , ) . The highly conserved non-coding elements and transposon-free regions are often associated with developmental regulatory genes , and with transcription factors ( TFs ) in particular [43]–[47] . In order to confirm this association , we calculated the fractions of genes with HCNEs or with TFRs in their proximity . We observed that for both features this fraction was higher for TFs than for all genes . Importantly , we observed that only the phylotypic modules were enriched in TFs ( Figure 3E ) . This partially explained the enrichment in HCNEs and TFRs for genes expressed in mid-development . In addition , HCNEs were more often present in the proximity of TFs from the pharyngula module than in the proximity of TFs in general ( Figure 3E; 8 . 8% of TFs from the pharyngula module had at least one HCNE in their proximity , and only 3 . 7% of all TFs had at least one HCNEs in their proximity ) . Also TFRs were more often present in the proximity of TFs from the phylotypic modules than in the proximity of TFs in general ( Figure 3E; 31% and 45% of TFs from the segmentation and pharyngula modules , respectively , had TFRs in their proximity , and only 26% of all TFs had TFRs in their proximity ) . Consequently , the enrichment in HCNEs and TFRs for genes expressed in the phylotypic stage seems to be related to the regulation of developmental processes . Interestingly , only few Hox genes from phylotypic modules were associated with HCNEs ( four Hox genes from segmentation module ) , and with TFRs ( six Hox genes from segmentation module , and one Hox gene from pharyngula module ) . In addition , we checked for genes that preserved their specific ancestral order in the genome across metazoans ( so called conserved ancestral microsyntenic pairs , [48] ) and are known to be involved in the regulation of development . We found that they were slightly overrepresented in the segmentation module , but only at the limit of statistical significance ( see Text S1 ) . Finally , we checked for core developmental genes in each module ( see [47] for the list of genes ) . These genes are known to be involved in the regulation of development , and to have highly conserved regulatory regions within different taxa , including , nematodes , insects and vertebrates [47] . We detected a significant enrichment in these genes only in the pharyngula module ( 20 core genes; hypergeometric test , ) , supporting the hourglass model .
Our goal was to study developmental constraints acting on various gene properties in vertebrates . Overall , we analyzed and compared five gene characteristics , namely the conservation of gene sequence , gene expression , and regulatory elements , as well as age and orthology relationships . To this end we identified distinct sets of genes with time-specific expression in zebrafish development , i . e . , genes over-expressed in one of the seven consecutive stages: cleavage/blastula , gastrula , segmentation , pharyngula , larva , juvenile and adult . We believe that the change in expression level is a reliable indicator of gene involvement in different stages , although genes might also play a role outside the stages of their highest expression . Moreover , the modules contained genes overexpressed in relation to other stages , regardless of the absolute values of their expression . Thus , lowly expressed genes were also considered by the modularization algorithm , as long as they displayed some variance in expression levels over developmental time . Several features do not show any significant pattern over embryonic development , often in contradiction to previous reports . There is notably no evidence for change in selective pressure acting on sequences of protein-coding genes ( i . e . , ) over development ( in contrast to [10] ) . Unfortunately , the available data does not allow a strong conclusion concerning the conservation of expression ( in contrast to [13] ) , despite the probable importance of this feature in the evolution of development . In this respect , the situation in vertebrates stands in contrast to the relatively clear results in flies [12] , where the evolution of expression has been shown to be most constrained in mid-development . Gene orthology relations support the early conservation model . We show that early stages are less prone to tolerate both gene duplication ( consistent with [17] ) and gene introduction . The deficit in duplication in early development could also be due to a lack of opportunities for neo- or sub-functionalization in the anatomically simpler stages , which is not exclusive with strong purifying selection . The interpretation of transcriptome age is less straightforward . Our observations suggest that the oldest evolutionary stages tend to express of the oldest genes . It is possible that early stages are evolutionarily oldest , and that this is why they are enriched in oldest genes . Consequently , it is the presence of young genes in a module that would mark relaxed developmental constraints during the corresponding stage . However , neither early nor phylotypic modules are enriched in young genes ( Euteleostomi and Clupeocephala+Danio rerio ) , which suggests similar developmental constraints in early and mid-ontogeny . In any case , we do not find any support for the hypothesis that the phylotypic stage would be characterized by the oldest transcriptome ( in contrast to [11] ) . While the modularization approach does not support several previous hypotheses of genomic traces of the phylotypic period , it allows us to distinguish a strong signal of conservation of gene regulation in mid-development . While this had not yet been reported in genomic studies , it is consistent with early descriptions of the phylotypic stage as characterized by Hox genes body patterning activity [6] . Of note , the patterns that we observe are robust to the removal of Hox genes ( data not shown ) , so they are more general than this original observation . We observed an excess of HCNEs only for genes expressed in the pharyngula module , and an excess of TFRs only for genes expressed in the phylotypic modules . The enrichment in HCNEs and TFRs has been related to developmental regulatory genes , and to transcription factors in particular [43] , [45]–[47] . Indeed , we observed that more TFs were expressed in mid-development than in other stages . Also , we showed that a significant proportion of TFs expressed in mid-development had conserved regulatory regions ( i . e . , HCNEs and TFRs ) , in contrast to TFs expressed early or late . Consequently , the enrichment in HCNEs and TFRs for genes expressed in mid-development can be explained by both a higher number of TFs and a higher number of HCNEs and TFRs for these TFs , than for genes expressed earlier or later . Moreover , the pharyngula module was associated with core developmental genes . Overall , these results suggest that mid-developmental processes have extremely high conservation of regulation . This conservation could translate into observed common traits of the phylum expressed at the phenotypic level during mid-development . In addition , core developmental genes are known to be present in different taxa ( e . g . , nematodes , insects and vertebrates ) , in each of which they have a conserved regulation that evolved in parallel [47] . This could explain why the phylotypic stage is observed not only in vertebrates [49] , but also in other phyla , e . g . , in arthropods [4] , [12] . Finally , for all of the features which we have considered there is at least some trend towards weaker evolutionary constraints in the latest stages: is higher in post-embryonic stages and there are less sites under purifying selection ( Figure S11 ) ; correlation of expression is lowest for maternal , larval and adult genes; young genes and genes with duplications in fishes or other vertebrates are overrepresented in late modules; and genes expressed in juveniles and adults have the less HCNEs and TFRs . Although not all of these trends are significant , no feature shows stronger conservation in late development or adult . Thus , while different aspects of gene evolution show constraints at different times of development , there appears to be a generally faster evolution of all aspects of larval , juvenile and adult genes . Whether this is due to lower constraints ( i . e . , less purifying selection ) or to stronger involvement in adaptation ( i . e . , more diversifying selection ) , remains an open question . In summary , we studied evidence for , or against , any particular pattern of developmental constraints by considering sets of genes with time-specific expression patterns . Comparing such independent sets of genes with a clear function during embryogenesis resulted in cleaner and more fine-grained characterization of evolutionary patterns than previously reported . Notably , we showed that different levels of molecular evolution follow different patterns of developmental constraints . The sequence of regulatory regions is most conserved for genes expressed in mid-development , consistent with the hourglass model . Gene duplication and new gene introduction is most constrained during early development , supporting the early conservation model . Whereas , all gene properties coherently show the least conservation for the latest stages , consistent with both the early conservation and the hourglass models .
Microarray data of zebrafish development were downloaded from NCBI's Gene Expression Omnibus [50] ( GSE24616 ) . This study was performed on the Agilent Zebrafish ( V2 ) Gene Expression Microarray . In total , expression profiles for 60 developmental stages ( from unfertilized egg to adults stages ) were measured . The last ten stages ( 55 days–1 year 6 months ) were measured separately for male and female . Two replicates were made per time point , resulting in microarrays in total . For each microarray , values of gProccessedSignal were log10 transformed and normalized as follows . Separately for each replicate , we equalized the expression signals between microarrays using the spike-ins reference , to account for different amounts of RNA present throughout development . To this aim , we first quantile normalized the expression signal of all spike-ins from all microarrays . Then , for each spike-in level we took the median value of expression signal before and after quantile normalization . This resulted in 10 pairs of expression signals ( original signal vs . normalized signal ) . With linear interpolation between these points , we obtained a piecewise linear curve that defined a mapping from original to normalized expression signals , which we used to equalize the expression signals from all microarrays . This was done by projecting each expression signal onto the piecewise linear curve and calculating the corresponding normalized value . Finally , we quantile normalized the data within replicates and computed the mean value for each gene within replicates . Expression values measured separately for males and females were averaged for each time point . Microarray data of mouse development were downloaded from Array Express ( E-MEXP-51 and E-MTAB-368 ) . The E-MEXP-51 study was performed on ( ) F1 mice using Affymetrix GeneChip Murine Genome U74Av2 . In total , expression profiles for 10 early developmental stages ( zygote , early 2-cell , mid 2-cell , late 2-cell , 4 cell , 8 cell , 16 cell , early blastocyst , mid-blastocyst , late blastocyst ) were measured . 2–4 replicates were made per time point . The data were normalized using gcRMA package . The E-MTAB-368 study was performed on C57BL/6 mice using Affymetrix GeneChip Mouse Genome 430 2 . 0 . In total , expression profiles for 8 mid and late developmental stages ( E7 . 5 , E8 . 5 , E9 . 5 , E10 . 5 , E12 . 5 , E14 . 5 , E16 . 5 , E18 . 5 ) were measured . 2–3 replicates were made per time point . The data were normalized using gcRMA package . Agilent probe sets were mapped to their corresponding zebrafish genes ( Ensembl release 63 [51] ) using BioMart [52] . Probe sets which did not map unambiguously to an Ensembl gene were excluded from the analysis . A total of 19049 probe sets corresponding to 14293 zebrafish genes were taken into account in our analysis . Affymetrix probe sets were mapped to their corresponding mouse genes ( Ensembl release 63 [51] ) using BioMart [52] . Probe sets which did not map unambiguously to an Ensembl gene were excluded from the analysis . For genes that were mapped by several probe sets we used the signal averaged across the probe sets . A total of 2883 mouse genes mapped by probe sets present on both mouse microarrays were taken into account in the gene expression analysis . The ISA identifies modules by an iterative procedure . A detailed description of the algorithm in the general case is given in [19] ( see also http://www2 . unil . ch/cbg/homepage/downloads/ISA_tutorial . pdf ) . In this specific study , the algorithm was initialized with seven candidate seeds , each consisting of one artificial expression profile corresponding to one of the zebrafish developmental stages ( see Text S1 for details ) . Next , these seeds were refined through iterations by adding or removing genes and developmental time points until the processes converge to stable sets , which are referred to as ( transcription ) modules . Each developmental time point and gene received a score indicating their membership ( if non-zero ) and contribution to a given module . The closest the score for a gene or developmental time point was to one , the stronger the association between the gene/developmental time point and the rest of the module . The ISA was run twice with the following sets of thresholds: 1 ) and , and 2 ) and , for genes and developmental time points , respectively . We obtained the pharyngula module only in the case of , and all other modules with both and . All the modules contained their corresponding idealized profile . For further analysis , we kept a single module per developmental stage . From the pair of modules , we chose the one in which the idealized profile had a higher gene score . Overall , segmentation , pharyngula and juvenile modules were obtained with , and cleavage/blastula , gastrula , larva , and adult modules were obtained with . Gene ontology ( GO ) association for all genes mapped by zebrafish probe sets were downloaded from Ensembl release 63 [51] , using BioMart [52] . GO enrichment was tested by Fisher's exact test , using the Bioconductor package topGO [53] version 2 . 2 . 0 . The reference set consisted of all Ensembl genes mapped by probe sets of the microarray used . The “elim” algorithm of topGO was used to eliminate the ( tree-like ) hierarchical dependency of the GO terms . To correct for multiple testing the Bonferroni correction was applied . For every module GO categories with corrected P-value lower than 0 . 01 were reported , if less then ten GO categories were significant we reported the top ten ( see Table S2 ) . Ensembl Perl API release 70 [54] was used to extract all Ensembl Compara gene trees ( and alignments ) with a Clupeocephala ( bony fishes ) root . Sequences with too many gaps , or undefined nucleotides , were removed from the tree and alignment by MaxAlign ( version 1 . 1 ) [55] . Only trees without duplication ( one-to-one orthologs ) and with at least six leaves were kept . This resulted in 6769 trees . The site model from codeml [56] ( PAML package release 4 . 6; models M1a and M2a in codeml ) was used to predict sites-specific selection in these trees . Finally , 916 trees were removed due to the lack of zebrafish genes , and 81 were removed due to lack of expression data on the zebrafish microarray . This resulted in 5772 trees . For every gene tree we calculated its mean value ( ) . For every module we calculated the median ratio of its genes , where was the number of genes belonging to one of the 5772 trees . Next , we generated 10000 sets of randomly chosen genes . For each set we calculated the median ratio . Thus , we constructed a sampling distribution of the median values for a set of genes . Then we calculated the probability that the median of the original module was sampled from the constructed distribution . This allowed us to assess if the observed median ratio was significantly different from the expected median value . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . To study the age of genes belonging to different modules we dated the genes by their first appearance in the phylogeny . This consisted of retrieving the age of the oldest node of their Gene tree in Ensembl release 63 [51] . Genes' age was described with one of the following categories: Fungi/Metazoa , Bilateria , Coelomata , Chordata , Euteleostomi , Clupeocephala , and Danio rerio . To fit the chi-square test requirements ( more than 5 elements in a group ) we merged the genes into five age categories: Fungi/Metazoa , Bilateria , Coelomata+Chordata , Euteleostomi , Clupeocephala+Danio rerio . Next , for every module we calculated the age distribution of its genes . We performed chi-square goodness of fit test to compare the observed and expected distributions of age classes in the modules . The expected distribution was estimated by classifying all zebrafish genes into one of the five age categories . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . Homology information of zebrafish and mouse genes was retrieved from Ensembl release 63 [51] , using BioMart [52] . A total of 17482 pairs of zebrafish-mouse orthologous genes had expression information in the zebrafish microarray data ( 14293 zebrafish genes and 11322 mouse genes ) . Among them there were 6441 one-to-one orthologous pairs , 5048 one-to-many orthologous pairs , and 2993 many-to-many orthologous pairs . 2901 zebrafish genes showed no orthology relationship with mouse genome . From further analysis we excluded 99 “apparent-one-to-one” gene pairs . For every module we calculated the number of genes that were in one-to-one , one-to-many , many-to-many and no orthology relation to mouse genes . Next , we performed chi-square goodness of fit test to compare the observed and expected distributions of orthology classes in the modules . The expected distribution was estimated by classifying all zebrafish genes into one of the four orthology categories . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . To study expression conservation between zebrafish genes assigned to the modules and their mouse one-to-one orthologs , we used gene expression data for 2883 orthologous gene pairs ( the limiting factor being the mapping to both mouse microarrays ) . For genes that were mapped by several probe sets we averaged their signal across the probe sets for both species . In order to compare gene expression between two species , we first calculated the mean expression for zebrafish genes present in the modules and their one-to-one mouse orthologs . Due to the incompatibility of two mouse microarray data used it was difficult to provide a meaningful comparison of expression for the two species . To calculate the correlation between expression profiles between zebrafish and mouse we reduced their expression profiles to six metastages: zygote , cleavage , blastula , neurula , organogenesis , and post-embryonic stage ( see [39] for detailed definition of metastage ) . For every module and every metastage we calculated the mean expression level for zebrafish genes and their mouse one-to-one orthologs , and next we calculated the Pearson's correlation coefficient between them . Location data for highly conserved non-coding elements ( HCNE ) between zebrafish and mouse ( 70% of identity ) was retrieved from Ancora [43] ( http://ancora . genereg . net/downloads/danRer7/vs_mouse ) . The file HCNE_danRer7_mm9_70pc_50col . bed . gz was downloaded and used in the analysis . For each of the 14293 Ensembl genes considered in our analysis , we calculated the number of HCNE in regions of 500 base pairs upstream from the transcription start site . Next , for every module we performed a hypergeometric test to assess if they were significantly enriched in genes with HCNE . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . In additional analyses , we calculated the number of HCNE in regions of 200 and 1000 base pairs upstream from the transcription start site , as well as in introns . Also , we repeated the analysis with HCNEs of 90% identity ( see Text S1 ) . Location data for transposon-free regions ( TFRs ) in zebrafish was retrieved from [44] ( http://www . biomedcentral . com/content/supplementary/1471-2164-8-470-S1 . txt ) . First , each TFR was associated with Ensembl ID [51] of its closest transcript from genome assembly Zv6 . Then for each Ensembl transcript ID we retrieved an Ensembl gene ID from genome assembly Zv7 . For every module we performed a hypergeometric test to assess if they were significantly enriched in genes with TFRs in their proximity . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . The set of transcription factors ( TFs ) was defined based on GO category annotation: GO: 0006355 , regulation of transcription , DNA-dependent . Among 14293 Ensembl genes , 957 were annotated as transcription factors . For every module we performed a hypergeometric test to assess if they were significantly enriched in TFs . Next , we performed a hypergeometric test to assess if the TFs present in the modules were enriched in HCNEs and TFRs . To correct for multiple testing we applied the Bonferroni correction . We used 0 . 01 as a significance level . | During development , vertebrate embryos pass through a “phylotypic” stage , during which their morphology is most similar between different species . This gave rise to the hourglass model , which predicts the highest developmental constraints during mid-embryogenesis . In the last decade , a large effort has been made to uncover the relation between developmental constraints and the evolution of genome . Several studies reported gene characteristics that change according to the hourglass model , e . g . sequence conservation , age , or expression . Here , we first show that some of the previous conclusions do not hold out under detailed analysis of the data . Then , we discuss the disadvantages of the standard evo-devo approach , i . e . comparing descriptive statistics of all genes across development . Results of such analysis are biased by genes expressed constantly during development ( housekeeping genes ) . To overcome this limitation , we use a modularization approach , which reduces the complexity of the data and assures independency between the sets of genes which are compared . We identified distinct sets of genes ( modules ) with time-specific expression in zebrafish development and analyzed their conservation of sequence , gene expression , and regulatory elements , as well as their age and orthology relationships . Interestingly , we found different patterns of developmental constraints for different gene properties . Only conserved regulatory regions follow an hourglass pattern . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genome",
"expression",
"analysis",
"microarrays",
"developmental",
"biology",
"genomics",
"genome",
"evolution",
"comparative",
"genomics",
"biology",
"computational",
"biology",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"developmental",
"biology"
] | 2013 | The Hourglass and the Early Conservation Models—Co-Existing Patterns of Developmental Constraints in Vertebrates |
In nature , closely related species may hybridize while still retaining their distinctive identities . Chromosomal regions that experience reduced recombination in hybrids , such as within inversions , have been hypothesized to contribute to the maintenance of species integrity . Here , we examine genomic sequences from closely related fruit fly taxa of the Drosophila pseudoobscura subgroup to reconstruct their evolutionary histories and past patterns of genic exchange . Partial genomic assemblies were generated from two subspecies of Drosophila pseudoobscura ( D . ps . ) and an outgroup species , D . miranda . These new assemblies were compared to available assemblies of D . ps . pseudoobscura and D . persimilis , two species with overlapping ranges in western North America . Within inverted regions , nucleotide divergence among each pair of the three species is comparable , whereas divergence between D . ps . pseudoobscura and D . persimilis in non-inverted regions is much lower and closer to levels of intraspecific variation . Using molecular markers flanking each of the major chromosomal inversions , we identify strong crossover suppression in F1 hybrids extending over 2 megabase pairs ( Mbp ) beyond the inversion breakpoints . These regions of crossover suppression also exhibit the high nucleotide divergence associated with inverted regions . Finally , by comparison to a geographically isolated subspecies , D . ps . bogotana , our results suggest that autosomal gene exchange between the North American species , D . ps . pseudoobscura and D . persimilis , occurred since the split of the subspecies , likely within the last 200 , 000 years . We conclude that chromosomal rearrangements have been vital to the ongoing persistence of these species despite recent hybridization . Our study serves as a proof-of-principle on how whole genome sequencing can be applied to formulate and test hypotheses about species formation in lesser-known non-model systems .
The Drosophila pseudoobscura species subgroup is comprised of two D . pseudoobscura subspecies ( D . ps . pseudoobscura and D . ps . bogotana ) , and two closely related species , D . persimilis and D . miranda . The D . pseudoobscura subspecies are geographically isolated ( D . ps . pseudoobscura ranges across the western half of North America and D . ps . bogotana is restricted to Colombia in South America ) , share chromosomal arrangements , and represent the earliest stages of species divergence [13] . D . persimilis and D . miranda are restricted to the west coast of North America , where they co-occur with D . ps . pseudoobscura . Both D . pseudoobscura subspecies differ from the close relative species D . persimilis by fixed ( or nearly fixed ) chromosomal inversion differences on three of their major chromosome arms , and F1 hybrid males from crosses between these species are sterile ( though females are fertile ) . In contrast , D . miranda is an outgroup species which cannot produce any fertile hybrids with D . pseudoobscura or D . persimilis [14] . The relative relationships of these species as ( ( ( D . ps . pseudoobscura-D . ps . bogotana ) D . persimilis ) D . miranda ) is well established by DNA sequences , chromosomal inversions , and reproductive isolation [15] , [16] . Overall , this system provides us with a pair of taxa that hybridize and have experienced introgression ( D . ps . pseudoobscura and D . persimilis [17]–[19] ) , and two taxa that have not experienced recent introgression from any close relatives ( ingroup , D . ps . bogotana , and outgroup , D . miranda ) . Nucleotide divergence between the hybridizing species D . ps . pseudoobscura and D . persimilis is high within and near the three chromosomal inversions [20] , [21] , which are linked to factors conferring hybrid sterility , mating discrimination , and other barriers to gene flow [22] , [23] . Based on these observations , we hypothesized that inversions facilitate the distinction of these species despite ongoing natural hybridization . However , it has been difficult to fully disentangle complications that result from ancestral polymorphisms shared between these species [21] , [24] and underlying assumptions found in many statistical tests for introgression [25]–[27] . Two recent studies also reached differing conclusions about whether gene exchange between these species occurred during the initial divergence process or later [20] , [21] . To better understand the genealogical history of this subgroup , we use published genome sequence assemblies of D . ps . pseudoobscura and D . persimilis [12] , [28] , along with three novel partial genomic sequences that we generated using 454/Roche technology ( Table S1 ) : one from D . miranda , one from a second strain of the North American subspecies , D . ps . pseudoobscura , and one from the South American subspecies , D . ps . bogotana . By providing controls for divergence in the absence of gene flow , these newly obtained genomic sequences allow for more robust analysis than previous studies . Our new results suggest that the chromosomal regions inverted between D . persimilis and D . ps . pseudoobscura arose in allopatry , and that D . pseudoobscura , D . persimilis , and D . miranda all diverged within a relatively short time frame . We also find compelling evidence for autosomal gene exchange between D . persimilis and D . ps . pseudoobscura in collinear regions since the split of D . ps . pseudoobscura and D . ps . bogotana , likely within the past 200 , 000 years . Overall , our analyses utilize genome sequence data in an existing framework to demonstrate the importance of chromosomal inversions in maintaining the persistence of hybridizing species and to consolidate previous tentative conclusions about divergence in this group . Further , this research serves as a model for how whole genome shotgun sequence data can be used with a reference genome sequence to address fundamental questions regarding evolutionary changes leading to the formation of species .
Figure 1 presents sliding window estimates of polymorphism within the North American subspecies , D . ps . pseudoobscura , and divergence between D . ps . pseudoobscura and each of D . persimilis and D . miranda across four of the five major chromosome arms . Each datapoint within the sliding window represents the fraction of bases differentiating two genome sequences along a 500 kilobase pairs ( kbp ) interval , iterated every 100 kbp . Very similar plots were generated for intergenic regions or introns alone ( not shown ) . We only scored positions for which aligned sequences were available for all four taxa ( Table S2 ) , hence eliminating the possibility that a particular region of high or low divergence would be represented in some estimates but not others . Nucleotide polymorphism estimated within D . ps . pseudoobscura was confirmed to be in the same range as that observed in polymorphism studies of focal genomic regions of this species [18] ( see Table S3 ) . Fixed inversions on chromosomes XL , XR , and 2 distinguish D . ps . pseudoobscura and D . persimilis , and their breakpoints are superimposed on Figure 1 . Corroborating previous work , nucleotide diversity within D . ps . pseudoobscura and divergence between D . ps . pseudoobscura and D . persimilis are low in regions near the centromere [21] , [29] . The latter observation was previously interpreted “as reflective of ancestral patterns of polymorphism rather than the process of divergence between these species” [21] . Consistent with this interpretation , we observe that diversity within D . ps . pseudoobscura and divergence between D . ps . pseudoobscura and D . persimilis were correlated on every chromosome arm ( r = 0 . 418–0 . 535 , P<0 . 01 for each ) [see also 24] . The species pair , D . ps . pseudoobscura and D . miranda , exhibit a different pattern . There was no consistent decline in divergence between these two species in regions near the centromeres . Furthermore , diversity within D . ps . pseudoobscura was not significantly correlated with divergence to D . miranda along any chromosome arm except chromosome 4 ( r = 0 . 330 , P = 0 . 018 ) , suggesting that D . miranda and D . ps . pseudoobscura are not sharing many polymorphisms . According to a model where D . miranda is the outgroup , we predict that the range ( maximum minus minimum ) of divergences across windows should be greater for the purportedly more divergent species pair , D . miranda - D . ps . pseudoobscura , than the pair of more recently diverged species , D . persimilis - D . ps . pseudoobscura . While this prediction was met for windows along the collinear chromosome 4 , we observed instead a greater range of divergences in the D . persimilis - D . ps . pseudoobscura pairing on the chromosome arms ( XL , XR , and 2 ) that harbor inversions distinguishing these species ( see also Table S4 , Figure S1 ) . These observations are inconsistent with a more recent divergence of this latter species pair , and are more consistent with the presence of multiple genealogical histories along the genome . Inversions prevent gene exchange because the products of recombination are not recovered . We confirmed that recombinant products are not recovered within 2 . 1 megabase pairs ( Mbp ) of fixed inversions along chromosome XL , XR , and 2 in heterozygotes ( D . ps . pseudoobscura - D . persimilis interspecies hybrids ) . We recovered 0 . 25%–0 . 55% recombinants at markers 2 . 8 Mbp outside of each inversion , indicating that complete recombination suppression extends greater than 2 . 1 Mbp , but not more than 2 . 8 Mbp outside inversions . Strong crossover suppression , resulting in less than one percent recombinants , is observed relative to one marker 3 . 35 Mbp outside of the XR chromosomal inversion . Crossing over is largely restored at 4 . 55 Mbp outside inversions , with a crossover rate greater than 5% observed from one marker on chromosome 2 ( see Table S5 ) . The lack of recombination and introgression should produce a distinct signature in nucleotide divergence within and near chromosomal inversions . We found that , along the three chromosome arms bearing inversions , nucleotide diversity within D . ps . pseudoobscura was comparable to D . persimilis nucleotide divergence when estimated on sequence greater than 2 . 5 Mbp outside the inverted regions . In contrast , divergence between D . persimilis and D . ps . pseudoobscura was comparable to divergence between D . miranda and D . pseudoobscura in regions inside and within 2 . 5 Mbp flanking the inversions . The consistency of this pattern across independent chromosomal arms suggests either that all three inversions arose at approximately the same time as the split from the ancestor of D . miranda , or that the ancestral populations of these species were already separated ( i . e . , allopatric ) when the inversions arose ( see Discussion ) . Recent gene flow is not expected between the South American D . ps . bogotana and either of the North American taxa D . ps . pseudoobscura or D . persimilis . Analyses of nucleotide sequence data suggests that the D . pseudoobscura subspecies diverged from a common ancestor 200 , 000 years ago [19] , [30] . As such , we can use the isolated subspecies as a “negative control” to test for recent introgression between North American D . ps . pseudoobscura and D . persimilis . Because of hybridization between the North American taxa , a very simple expectation is that D . persimilis ( Dper ) should be more similar in sequence to North American ( NA ) than South American ( SA ) subspecies of D . pseudoobscura ( Figure 2 ) . We limited the dataset to sites where we have 454/Roche sequence reads for both D . pseudoobscura subspecies , and tested this hypothesis using regions far from the inversion on chromosome 2 and all along collinear chromosome 4 . Aligned bases were categorized as [Dper = NA≠SA] or [Dper = SA≠NA] . No two bases were scored that were within 500 bp of each other , hence reducing artifacts from non-recombining haplotype blocks . We observed an excess of the first category ( 7073 vs . 6797 , Binomial Sign Test P = 0 . 0096 ) , indicating that divergence is lower between D . persimilis and North American D . ps . pseudoobscura than between D . persimilis and South American D . ps . bogotana . The above test does not account for possible faster divergence within the South American subspecies lineage , either through increased mutation rate or more frequent fixation of slightly deleterious alleles . Testing for differences in lineage rates , we did not observe greater divergence between the South American subspecies and the non-hybridizing species , D . miranda ( Dmir ) , than between the North American subspecies and D . miranda ( Dxy = 0 . 019 for both , P = 0 . 221 ) . Nonetheless , we can test for recent gene exchange more rigorously by specifically counting “shared-derived” base pair substitutions polarized with D . miranda . Counts of [Dmir = SA≠Dper = NA] were compared to counts of individual base pairs in which [Dmir = NA≠Dper = SA] , where the latter half of the inequality denotes potential shared-derived bases . Again , we observe a slight , borderline significant excess of the first category ( 219 vs . 185 , Binomial Sign test P = 0 . 05 ) , suggesting that D . persimilis and North American D . ps . pseudoobscura share more derived bases . Finally , introgression between species is not expected to be homogeneous outside inverted regions . The Alcohol dehydrogenase ( Adh ) region has been reported to have introgressed recently between these species using analyses independent of divergence from the South American species , D . ps . bogotana [19] . Further , it can be introgressed in the laboratory and made homozygous in a foreign genetic background with no deleterious effects [23] . We examined base pair counts of [Dmir = SA≠Dper = NA] vs . [Dmir = NA≠Dper = SA] for this region . In this region which bears Adh ( chromosome 4 “group1” , extending 4 Mbp starting at position 14 . 4 million in Figure 1 ) , we again observed a significant and dramatic excess of the first category ( 27 vs . 10 , P = 0 . 00382 ) . We applied the same analyses to test for recent gene exchange along X-linked regions from both XL and XR distant from the inversion breakpoints . We observed a nonsignificant difference in number of bases categorized as [Dper = NA≠SA] vs . [Dper = SA≠NA] on this chromosome ( 1200 vs . 1131 , Binomial Sign Test P = 0 . 079 ) . When we polarized the bases and compared ( Dmir = SA≠Dper = NA ) vs . ( Dmir = NA≠Dper = SA ) , we observed a nonsignificant difference opposite in direction to our expectation ( 46 vs . 62 ) . However , there was only 27% as much sequence to analyze more than 2 . 5 Mbp from inversions on the X-chromosome than on the autosomes .
If two species share extensive polymorphism through introgression or incomplete lineage sorting resulting from a recent split , we predict that nucleotide sequence diversity within species should be correlated with average pairwise nucleotide differences between species . Extensive polymorphism sharing was shown previously in the case of D . ps . pseudoobscura and D . persimilis [18] , [20] , [21] , [31] . In contrast , we find that nucleotide sequence differences between D . ps . pseudoobscura and D . miranda were uncorrelated with nucleotide sequence differences between two strains of D . ps . pseudoobscura . This finding suggests that our comparisons to D . miranda are not hindered by introgression or extensive shared ancestral polymorphisms [but see 32] . Previous DNA sequence-based studies observed that D . ps . pseudoobscura and D . persimilis share variation far outside the fixed inversions that distinguish these species [20] . Here , we note that divergence between D . ps . pseudoobscura and D . persimilis is higher and more comparable to differences between two strains of D . ps . pseudoobscura in regions distant from the inversions . In contrast , divergence between D . ps . pseudoobscura and D . persimilis is comparable to that between D . ps . pseudoobscura and the non-hybridizing outgroup , D . miranda , for regions inside and just outside the chromosomal inversions that separate them . These new results can be used to formulate a hypothesis for the evolutionary history of these species and reconcile previously contradictory inferences . Machado et al [20] suggested that D . ps . pseudoobscura and D . persimilis largely speciated in allopatry , close in time to the split of these species from D . miranda , and recent secondary contact between the first two resulted in the dissolution of differences outside the inverted regions . In contrast , Noor et al [21] noted that significant differences in divergence among the XL , XR , and 2-chromosome inverted regions suggest instead that D . ps . pseudoobscura and D . persimilis speciated under a sympatric “divergence-with-gene-flow” model . In other words , differences in divergence between inversions reveal when each inversion arose as both species evolved in sympatry . Our study recapitulates both sets of results and allows us to suggest a resolution . As in Noor et al [21] , we observe that the XL chromosome arm inversion was most different in sequence between D . ps . pseudoobscura and D . persimilis , followed by chromosome 2 and finally chromosome arm XR ( see Figure 1 ) . However , we also observe that divergence between D . pseudoobscura and D . persimilis within each inverted region was similar to the divergence between D . ps . pseudoobscura and D . miranda ( shown at single loci by [20] ) . Because we observe the same XL>2>XR ranking in D . ps . pseudoobscura divergence from D . miranda that was shown previously for divergence from D . persimilis [21] , we now interpret this variation among chromosomes as reflective of differences in mutational processes rather than differences in time since separation . Our new , combined observations suggest two possible interpretations . First , the three inversions independently may have arisen very close in time ( near the time of the split from D . miranda ) from the D . pseudoobscura-D . persimilis ancestor , and these three derived forms segregated exclusively in D . persimilis . Alternatively , and arguably more parsimoniously , the three species diverged close in time , D . persimilis acquired three new inversions sometime after the split from D . pseudoobscura , and secondary contact between D . persimilis and D . pseudoobscura homogenized the noninverted regions . Many recent studies have analyzed DNA sequence polymorphism and divergence to identify the statistical signature of recent introgression . However , these tests did not typically identify a time frame within which introgression occurred except as variance in the time of divergence [25] , [26] . Instead , most tests merely reject or fail to reject a model of divergence in total isolation . Here , we use a comparison between subspecies to infer the timing of introgression between D . persimilis and D . ps . pseudoobscura . One D . pseudoobscura subspecies co-occurs and hybridizes with D . persimilis while the other subspecies lives isolated on a different continent . Hence , we can attribute differences in divergence between D . persimilis and these D . pseudoobscura subspecies to hybridization that has occurred more recently than the split of the subspecies , estimated to have been 200 , 000 years ago [19] , [30] . We observe a slight but statistically significant difference in divergence across uninverted ( collinear ) autosomal regions between D . persimilis and the two D . pseudoobscura subspecies , suggesting recent introgression between the co-occurring taxa , but we fail to detect such evidence for introgression across comparable regions of the X-chromosome . Although we detected a statistically significant signature of introgression along autosomal loci , the signature was faint , suggesting that recent gene exchange has not been extensive . DNA sequence-based studies previously identified the statistical signature of historical introgression [18] , [20] , [31] , but these studies interpreted this gene exchange as ancient based on the lack of longer shared haplotypes [33] . Similarly , an allozyme-based meta-analysis failed to detect differences between D . ps . pseudoobscura populations co-occurring with D . persimilis compared to those elsewhere in North America [34] , suggesting a lack of extensive recent introgression . Given the high levels of gene exchange among populations within D . ps . pseudoobscura , the approach used by Kulathinal and Singh [24] does not have enough resolution to detect the low levels of gene flux we infer here . Again , our sparse genomic sequence data helps to refine these earlier results . A significant difference between sex-linked and autosomal loci in introgression has been a recurring theme in divergence population genetics [e . g . ] , [ 35] , [36]–[38] . However , in most systems , we lack knowledge of the karyotype ( e . g . , inversion differences ) or other factors which may make the sex chromosomes and particular autosomes inappropriate for comparison . In D . ps . pseudoobscura and D . persimilis , however , we observe evidence for introgression on the autosomes while not on the X-chromosome in regions outside the inversions , suggesting that these differences may be reflective of sex-linkage per se . This observation may be consistent with a higher density of factors conferring hybrid sterility or other barriers to gene flow on the X-chromosome than on the autosomes [e . g . , 39] . In this study , we used sparse whole-genome shotgun sequences from multiple taxa to infer the evolutionary history of a species group and to identify genomic features associated with their divergence . Our system was well-leveraged in that we initially began the investigation already having an assembled and annotated full-genome sequence for two of the focal species [12] , [28] as well as genetic mapping data localizing factors that reduced potential gene exchange [22] , [23] . Nonetheless , the cost of next-generation sequencing is dropping for both model and non-model systems , even between the execution of this study and its publication . Because of cost constraints , our study approached these questions using light resequencing ( effectively utilizing the power of millions of markers ) but producing extensive gaps and a majority of aligned positions being covered by single sequence traces . However , our approach serves as a proof-of-principle for future genomic studies on lesser developed systems . We attempted to reduce systematic biases by applying stringent filters , specific tests ( including averaging across 500 kbp windows ) and by employing the use of a well-assembled reference genome sequence . Future , more rigorous approaches enabled by less-expensive sequencing technologies will allow researchers greater power to infer historical evolutionary processes such as speciation and historical introgression in non-model systems .
In this comparative study , a total of five genomes representing four species of the obscura subgroup were sampled . Adult females from inbred lines of D . miranda ( from Mather , California; San Diego stock #14011-0101 . 08 ) and the subspecies , D . ps . bogotana ( from El Recreo , Colombia; San Diego stock #14011-0121 . 152 ) were each extracted and purified using the Gentra PureGene DNA isolation kit . For D . miranda , genomic DNA was nebulized and single stranded libraries generated before being sequenced at light coverage on a single Roche/454 Life Sciences GS-FLX run at Duke University's IGSP core sequencing facility , yielding approximately 100 Mbp of sequence ( see Table S1 ) . D . ps . bogotana genomic DNA was similarly sequenced in one half of one run at Duke University's IGSP core sequencing facility and one half of one run at 454 Life Sciences . These genome sequence traces were submitted to the NCBI Short Read Archive ( SRA ) as accession SRA008268 . Additionally , two previously sequenced and assembled genomes , D . ps . pseudoobscura ( Release 2 ) and D . persimilis ( Release 1 ) , were used for comparative analysis [12] , [28] . Finally , to estimate nucleotide diversity within D . ps . pseudoobscura , previously sequenced Roche/454 reads ( NCBI SRA accession SRA000268 ) from a second line ( from Flagstaff , Arizona; San Diego stock number 14011-0121 . 151; [24] ) were reassembled syntenically to D . ps . pseudoobscura . All Roche/454 reads were syntenically aligned against reference D . ps . pseudoobscura ( Release 2 ) linkage groups . Individual base calls were filtered to exclude nucleotides that are: within 3 base pairs of an alignment gap , harbor low quality scores ( below 10 ) , contain greater than 30% mismatches within a 7 base pair window , are in regions of high divergence ( divergence to D . persimilis is greater than 30% in a 7 base pair window ) . Alignments from the two previously sequenced reference genomes , Drosophila ps . pseudoobscura and D . persimilis were obtained via chain files from the UCSC Genome Browser ( genome . ucsc . edu ) . Site-specific annotation information such as intron and codon position was extracted from D . ps . pseudoobscura Release 2 . 3 annotations from FlyBase ( flybase . org ) . Chromosome arms ( including ordered contigs ) 2 , 4 , XL , and XR were used ( see [40] for contig details ) , representing roughly 80% of the total genome . We did not survey chromosome 3 because of complications from its inversion polymorphism within each of these species [41] . Chromosome arms XL , XR , and 2 differ by single inversions between D . pseudoobscura and D . persimilis , and the breakpoints of these inversions have been mapped [21] , [42] . Using microsatellite markers that flank the sides of each inversion , we surveyed the extent of recombination in F1 hybrids between D . ps . pseudoobscura and D . persimilis . The published genome lines of both species ( San Diego stock numbers #14011-0121 . 94 and #14011-0111 . 49 ) were used in this cross and recombinants were screened among 384 progeny of F1 females backcrossed to D . pseudoobscura . The following markers were used to assay recombination rate at varying distances from the inversions – chromosome 2 inversion: DPS2019 ( 2 . 77 Mbp from inversion on telomeric side ) , DPS2026 ( associated with inversion ) and DPS2031 ( 2 . 8 Mbp from inversion on centromeric side ) , XL inversion: DPSX_7446z ( 2 . 84 Mbp from inversion on centromeric side ) , DPSX046 ( associated with inversion ) , DPSX008 ( 0 . 4 Mbp from inversion on telomeric side ) , and DPSXL_3a_0 . 8 ( 2 . 8 Mbp from inversion on telomeric side ) , XR inversion: DPSXR_6_2 . 7 ( 3 . 35 Mbp from inversion on centromeric side ) , DPSX063 ( associated with inversion ) , DPSX037nA3 ( 1 . 4 Mbp from inversion on telomeric side ) , DPSX037N ( 2 . 1 Mbp from inversion on telomeric side ) , and DPSX058 ( 2 . 8 Mbp from inversion on telomeric side ) . Primer sequences are available upon request . | The transformation of populations into distinct species depends on whether hybridization , recombination , and subsequent gene introgression can be suppressed between diverging species . We use partial genome sequences to reconstruct this evolutionary process in the Drosophila pseudoobscura species subgroup , which includes the hybridizing species pair D . pseudoobscura pseudoobscura and D . persimilis . Recent models suggest that chromosomal inversions can facilitate the persistence of hybridizing species because of their effects on recombination , whereby inverted regions would exhibit higher nucleotide divergence than non-inverted regions . Indeed , D . pseudoobscura-D . persimilis nucleotide divergence outside these inverted regions is lower than within or near inversions , resembling D . ps . pseudoobscura levels of within-species nucleotide diversity . We also observe that recombination suppression in F1 hybrids extends greater than 2 Mbp outside the inversion breakpoints . Furthermore , when genomic sequence of D . persimilis is compared to two sister subspecies—the hybridizing subspecies , D . ps . pseudoobscura , and a non-hybridizing control subspecies , D . ps . bogotana—autosomal divergence is lower in the former , demonstrating recent gene exchange . These lines of evidence support a speciation model in which the two hybridizing species persist despite the presence of recent genic introgression in collinear regions of the genome because of the reduced recombinational effects of the inversions that distinguish them . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/population",
"genetics",
"evolutionary",
"biology/genomics",
"genetics",
"and",
"genomics/comparative",
"genomics",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics"
] | 2009 | The Genomics of Speciation in Drosophila: Diversity, Divergence, and Introgression Estimated Using Low-Coverage Genome Sequencing |
RNA processing events that take place on the transcribed pre-mRNA include capping , splicing , editing , 3′ processing , and polyadenylation . Most of these processes occur co-transcriptionally while the RNA polymerase II ( Pol II ) enzyme is engaged in transcriptional elongation . How Pol II elongation rates are influenced by splicing is not well understood . We generated a family of inducible gene constructs containing increasing numbers of introns and exons , which were stably integrated in human cells to serve as actively transcribing gene loci . By monitoring the association of the transcription and splicing machineries on these genes in vivo , we showed that only U1 snRNP localized to the intronless gene , consistent with a splicing-independent role for U1 snRNP in transcription . In contrast , all snRNPs accumulated on intron-containing genes , and increasing the number of introns increased the amount of spliceosome components recruited . This indicates that nascent RNA can assemble multiple spliceosomes simultaneously . Kinetic measurements of Pol II elongation in vivo , Pol II ChIP , as well as use of Spliceostatin and Meayamycin splicing inhibitors showed that polymerase elongation rates were uncoupled from ongoing splicing . This study shows that transcription elongation kinetics proceed independently of splicing at the model genes studied here . Surprisingly , retention of polyadenylated mRNA was detected at the transcription site after transcription termination . This suggests that the polymerase is released from chromatin prior to the completion of splicing , and the pre-mRNA is post-transcriptionally processed while still tethered to chromatin near the gene end .
The processes of transcription and RNA processing are co-transcriptionally integrated [1]–[7] . A variety of studies have demonstrated that pre-mRNA splicing is coupled to transcription and that intron removal can occur at the site of transcription while the polymerase is still engaged in active transcription . Other mRNA processing events such as 5′ capping and 3′-end formation are also coupled to transcription . A unique feature of the eukaryotic RNA Pol II is the presence of a long and highly conserved tail at the C-terminus of the large subunit ( Rbp1 ) of the polymerase , termed the CTD ( carboxy-terminal repeat domain ) . It is proposed that the CTD serves as a “jumping board” for protein factors that assemble on the nascent transcript as it emerges from the polymerase [8]–[10] . For instance , certain splicing factors from the SR protein family were found associated with the phosphorylated CTD [11]–[14] . SR proteins are most probably not associated with the transcriptional initiation complex but join in only after the elongation related CTD-phosphorylation has begun . This was demonstrated by showing that RNA Pol II poised at the Fos gene promoter is not associated with SR proteins when the gene is in an uninduced state , and only after transcription had begun were SR proteins recruited [15] . Chromatin immunoprecipitation ( ChIP ) experiments have also demonstrated the association of RNA processing factors with transcribing genes [16]–[19] . Co-transcriptional splicing implies that splicing , at least in part , should occur at the site of transcription . This is supported by kinetic studies showing that splicing is completed within a short ∼5–10 min time-frame from the time of transcription , and these times are not influenced by intron length [20]–[24] . Splicing factors are dynamically recruited to active gene loci [25]–[29] . The kinetics of transcription can influence alternative splice site selection . Polymerase slowdown and pausing affected the pattern of alternative splicing in vivo as seen with a mutant polymerase form that was intrinsically slow-elongating and favored the inclusion of an exon [30] . The current model is that slowing down of the polymerase or polymerase pausing leads to the selection of a weak splice site [1] , [31] . In order to examine the opposite relationship , namely whether the kinetics of the elongating RNA Pol II can be influenced by splicing events occurring co-transcriptionally on the associated pre-mRNA , we generated a series of comparable cell lines harboring inducible gene constructs that contain increasing numbers of introns and exons . These genes were stably integrated into a human cell line and formed tandem arrays containing multiple copies of the integrated gene , thereby creating a genomic locus that upon transcriptional activation is easily monitored using live-cell imaging . Such transcription sites have been used previously to analyze the kinetics of recruitment of a variety of nuclear factors to actively transcribing genes ( reviewed in [32] , [33] ) . Previous work has demonstrated that the real-time kinetics of RNA Pol II elongation on an active gene can be faster than measured by many biochemical measurements , and is coupled to stochastic pauses of the enzyme [34] , [35] . Recent studies have confirmed that Pol II elongation can proceed at rates of 3–4 kb/min [21] , [22] , [36] . In this study we used the above in vivo approach to examine Pol II elongation rates either on genes that do not contain an intron and therefore are not expected to recruit the spliceosome to the nascent transcript or on genes containing varying numbers of intron/exons and that undergo co-transcriptional splicing . We show that polymerase elongation kinetics are not modulated by splicing events taking place on the emerging pre-mRNA , that increased splicing leads to an increase in the splicing factors recruited to the mRNA , and that post-transcriptional splicing can proceed at the site of transcription without the presence of the polymerase .
In order to examine the elongation kinetics of RNA Pol II during splicing , we generated the following gene constructs . The first construct consisted of the human β-globin mini-gene with three exons and two introns and was therefore designated E3 ( Figure 1A ) [37] , [38] . Exon 3 was truncated and fused in-frame to a cyan fluorescent protein ( CFP ) coding region containing the peroxisomal targeting tripeptide Ser-Lys-Leu ( SKL ) in its C-terminus [39] , thereby generating cyan fluorescing peroxisomes throughout the cytoplasm , indicative of productive gene expression . Immediately downstream were a series of 18 MS2 sequence-repeats . When transcribed , these repeats form 18 stem-loops in the mRNA and provide high-affinity binding sites for the MS2 coat protein [40] , [41] . The detection of these mRNAs in vivo is accomplished by the binding of dimers of the YFP-MS2 coat-protein to each stem-loop , thereby forming fluorescently labeled mRNAs [42] . This occurs simultaneously during active transcription [34] , [41] . β-globin was chosen as a model gene since it has been extensively used for in vitro splicing assays [43] . In this manner we could use a system that is close to the endogenous state of a gene . The first intron of β-globin is short ( 131 nt ) , while most typical genes have a long first intron [44] . The length of the second β-globin intron is 992 nt , which is close to the median size of internal introns ( ∼1 , 400 nt ) [44] . Thereby , we could use repeats of the second intron to generate additional genes containing increasing numbers of introns . In order to compare between genes , we generated additional gene constructs ( Figure 1A ) : ( a ) E1 – an intronless gene in which exons 1 , 2 and their introns were deleted , thereby leaving part of exon 3 + CFP-SKL to serve as a single exon; and ( b ) E4 , in which intron 2 , flanked by the splice sites and part of exons 2 and 3 , was duplicated , thereby forming a gene construct with 4 exons and 3 introns . The E6 construct ( containing another two repeats of intron 2 ) will be described later on . All the constructs were under the inducible control of the Tet-On system . In the presence of the rtTA ( Tet-On ) transactivator and the addition of doxycycline ( dox ) to the medium , transcription was induced . Each of the constructs was stably integrated into human U2OS Tet-On stable cells . The different sized gene constructs produced the correctly spliced mRNAs as seen by RT-PCR on mRNA extracted from the three cell lines following transcriptional activation by dox: E3 = 2 , 279 nt ( unspliced 3 , 402 nt ) , E4 = 2 , 321 nt ( unspliced 4 , 436 nt ) ( Figure 1B ) . The E1 mRNA ( 1 , 941 nt ) did not undergo splicing . The genes were co-integrated with a plasmid containing 256 lac operator repeats ( lacO ) in order to enable the detection of the genomic locus of integration [29] , [45] . Using a fusion protein of the lac repressor protein ( LacI ) , e . g . RFP-LacI , the genomic locus was tagged , allowing the detection of the integration locus independent of the transcriptional status of the gene . Selection of positive stable clones was performed after induction with dox and identification of cytoplasmic CFP-peroxisomes in all cell clones ( Figures 1C and S1 ) , altogether indicating accurate transcription , splicing , export , and translation . Co-integration of the 256 lacO repeats was verified by the expression of RFP-lacI and the detection of single genomic integration sites ( Figure 1C ) . YFP-MS2-labeled mRNAs were detected at the transcription site , as well as in the nucleoplasm and the cytoplasm as expected ( Figure 1C ) . Several stable clones were kept for each gene , and we proceeded to work with three cell lines designated E1 , E3 , and E4 , in accordance with the integrated gene . Stable transfection results in the integration of multiple copies of the gene construct at one genomic locus that is termed a “tandem gene array , ” which is useful for live-cell imaging experiments of active transcription sites [32] . We confirmed that RNA Pol II in both active phosphorylated forms was recruited to the transcription sites using the H14 ( CTD phosphorylated on serine 5 ) , H5 ( CTD phosphorylated on serine 2 ) , and 8WG16 antibodies ( specific to the CTD repeats ) ( Figures 1D and S2A , B ) . RNA FISH with probes to different regions in the pre-mRNA and mRNA were used to detect the distribution of the mRNAs in the cells following dox induction . Probes to exon 1 , the CFP coding region ( exon 3 ) , or the MS2 region ( 3′UTR ) of the E3 gene , detected the active transcription sites as well as cytoplasmic mRNAs in induced cells , and some nuclear signal , suggesting robust nucleo-cytoplasmic transport of the mRNAs to the cytoplasm ( Figure 1E ) . No RNA FISH signal was seen in un-induced cells . Video S1 shows the activation of transcription , nucleo-cytoplasmatic transport of the tagged mRNA , and the appearance of the CFP peroxisomes in the cytoplasm . Altogether , we generated a series of cell lines in which the kinetics of transcription could be monitored on integrated genes in real-time . Probes to the first or second intron regions displayed signal only at the sites of transcription without additional signal in the nucleoplasm or cytoplasm ( Figure 2A , B ) , showing that the splicing of introns occurred at the site of transcription . In order for co-transcriptional splicing to occur , splicing factors must be recruited to the pre-mRNA either by diffusing from the nucleoplasm or possibly via the CTD of RNA Pol II . A previous proteomic study immunopurified more than 100 proteins associated with RNA Pol II in a non-transcribing extract [46] , among them U1 snRNP components and factors of the SR family . In order to assay the recruitment of splicing factors in vivo we examined whether different factors were enriched on the active E3 and E4 gene arrays , using immunofluorescence to endogenous proteins or expression of fluorescently-tagged versions of these proteins . snRNPs , splicing factors , and other RNA processing factors were detected ( Figures S2C–J and S3 ) . These data are summarized in Table 1 . In addition , using RNA FISH probes specific to the different snRNAs we could detect U1 , U2 , U4 , U5 , and U6 snRNAs at the site ( Figures 2C–E , S2D–E , and Table 1 ) . Not all molecules expected to be present at the transcription sites were detected , e . g . PSF , p54nrb , and 7SK-RNA ( Figure S3B , C , and F ) . We conclude that co-transcriptional splicing is occurring on the pre-mRNAs associated with the active transcription sites , and that the mRNA processing factors are recruited either via association with the transcribing polymerase or directly with the nascent mRNAs . We then tested for the recruitment of the spliceosomal U snRNPs to the intronless E1 gene . Using an antibody that recognizes the 2 , 2 , 7-trimethylguanosine ( m3G ) cap of snRNA ( anti-TMG ) we could detect TMG staining on the transcription site of E1 ( Figure S4A ) . RNA FISH showed that U1 snRNA was present on the site , as well as the U1A protein ( Figures 2F and S4B ) . However , all other snRNAs or U2AF65 were not found on the E1 sites ( Figure 2G–H , S4C–E and Table 1 ) . The recruitment of U1 snRNA to the E1 and E3 genes was measured and compared by correlating the intensity of the U1 and U2 signals on the transcription site to the same two signals within the nucleoplasm . In the nucleoplasm we expected to find co-localization of U1 and U2 snRNA signals , since spliced genes are predominant . In the case of the E3 and E4 genes , the U1/U2 correlation at the transcription site was the same as in the nucleoplasm ( Figure 2I ) . However , in the case of the E1 gene a different correlation was observed between U1 and U2 at the transcription site compared to the U1/U2 nucleoplasmic ratio ( Figure 2J ) . We then quantified this recruitment . The measured ratio of U1 and U2 snRNA signals on the different genes demonstrated a constant ratio between U1 and U2 on the intron-containing genes ( E3 , E4 , and E6 ) , while on the E1 gene the presence of U1 snRNA was dominant and U2 was not present ( Figure 2K ) . Interestingly in a recent study , the association of U1 components on a gene construct in which splicing did not take place due to mutations in the 5′- and 3′-splice sites was demonstrated [47] . We therefore conclude that U1 snRNPs are associated with the transcriptional machinery on all transcribing genes , no matter whether the pre-mRNA contains introns or not , probably through interactions with the CTD of the active polymerase [46] , [47] . After showing that pre-mRNA splicing is taking place at the site of transcription in the E3 and E4 genes , we examined whether polymerase elongation kinetics differed between the intronless and intron-containing genes . We used the previously described in vivo elongation assay [34] , [35] , [41] to measure the kinetics of RNA Pol II on the three genes . In this assay , active transcription sites are detected with the YFP-MS2 coat protein . This fluorescent signal at the transcription site consists of the steady state kinetics of mRNA synthesis as well as mRNA release from the site . Using a fluorescence recovery after photobleaching ( FRAP ) approach , the YFP-MS2 signal at the transcription site is photobleached and the recovery of the YFP-MS2 signal is monitored over time . Fluorescence recovery signifies the generation of new MS2 stem-loops in the nascent transcripts and their binding by fluorescent YFP-MS2 molecules entering from the surrounding nucleoplasm ( Figure 3A and 3B ) . This same approach was implemented on the E1 , E3 , and E4 genes . The cell lines were co-transfected with RFP-LacI and YFP-MS2 , and transcription was induced by dox for at least 6 h until transcription reached steady state . Active transcription sites at steady state did not show any significant difference in YFP-MS2 intensity for at least 20 min . Using a 3D-FRAP microscope system we could photobleach the transcription site ( YFP-MS2 signal ) and continue to detect the gene locus using the RFP-LacI DNA label . The advantage of using the 3D-FRAP microscope is that rapid 3 dimensional ( 3D ) -imaging over time ( 4D ) could be performed . Thereby , the signal of the whole nuclear volume could be collected with no effect on the measurements , and motion of the transcription site was easily tracked ( Figure 3A and Video S2 ) . This is in comparison to most confocal microscopes that collect the information only in 2 dimensions , and therefore measurements would be sensitive to any DNA motion in the z-axis . Transcription sites were followed for extended periods ( up to 30 min ) . The YFP-MS2 signal intensity on the transcription sites was quantified using a tracking algorithm . We measured recovery kinetics on the transcription sites of all the three gene-types in order to examine whether different elongation kinetics could be identified . We compared a polymerase passing through the MS2 region of an intronless E1 gene to a polymerase moving through the MS2 region of the intron-containing E3/E4 genes that undergo co-transcriptional splicing ( after the exons and introns are transcribed ) . The FRAP recovery curves measured on the E1 , E3 , and E4 genes all showed similar recovery kinetics ( Figure 3C ) , indicating that the polymerase transcribing the MS2 region was not affected by the presence of active splicing from the upstream sequences . We conclude that upstream splicing events do not necessarily affect the elongation kinetics on our model genes . However , it is possible that other differences such as in intron sequence or intron size ( e . g . the β-globin first intron is shorter than average ) might cause some effect on elongation . These experiments were performed on several microscope stations to verify reproducibility ( Figure S5 ) . Altogether , these measurements indicate that polymerase transcription rates are not modulated by the co-transcriptionally assembled spliceosomal machinery . We examined a similar gene construct in which additional introns and exons were very closely spaced . Thereby , we expected transcription to end before the completion of splicing of all introns , which would allow the further examination of elongation kinetics . The E6 gene was generated using the E4 construct and consisted of 6 exons and 5 introns by using duplications of intron 2 ( Figure 1A ) . In this manner , we generated a gene with 6 exons ( close to the 8 exon median size of genes [44] ) . The E6 mRNA ( 2 , 437 nt ) was correctly spliced ( unspliced 6 , 536 nt , Figure 1B ) , the CFP-SKL peroxisomal protein was expressed normally ( Figure S1 and Video S1 ) , and similar recruitment of splicing factors was observed ( unpublished data ) . First , we verified that there was an increase in accumulation of introns at the transcription site of the E6 gene compared to the other genes , as expected . 3D volumes of RNA-FISH labeled cells were collected , deconvolved , and analyzed . We calculated the fluorescence ratio of intron 2 in comparison to the last exon ( CFP exon ) , in all the cell lines . Higher levels of introns generated were detected on the E6 gene ( Figure 4A ) compared to the other genes . As expected , there was a gradual accumulation of intron 2 at the active transcription sites in dependence to the number of introns in each gene . We next asked whether there would be increased accumulation of the splicing machinery on genes containing more introns . We therefore quantified the relative amounts of the splicing machinery ( U snRNAs ) that accumulated on each transcript using 3D volumes of RNA-FISH labeled cells ( Figure 4B and 4C ) . Quantification of the fluorescence ratio of U5 or U6 snRNAs in comparison to the first exon present in each nascent transcript demonstrated a higher accumulation of snRNAs that was dependent on the number of introns in the gene . This indicated that the addition of exons resulted in an increase in the recruitment of the splicing machinery per transcript . These data also imply that the elongation machinery does not wait for each splicing event to complete before moving forward on the gene , but rather continues to elongate while more spliceosomal units accumulate on the transcript . Next , polymerase kinetics on the E6 gene were measured . Interestingly YFP-MS2 FRAP curves showed significantly slower recovery kinetics compared to the previous measurements on the E1 , E3 , and E4 genes ( Figure 4D ) . In order to verify that this was not a clone-related difference , and that the measurements were not affected by the number of gene units in the tandem array or from the DNA integration site , we compared two different E6 clones , E6-22 ( 40 gene copies ) to clone E6-3 ( 9 gene copies ) . The copy number of gene integration in each cell line was quantified by real-time PCR on genomic DNA . We found that both E6 clones recovered with identical and relatively slower kinetics ( Figure S6A and S6B ) . To identify the cause for the kinetic difference between the genes , we analyzed the measured FRAP recovery curves . Two previous studies have used this MS2 FRAP approach to measure Pol II transcription rates on specific genes [34] , [41] . The assay measures the recovery of fluorescence as the polymerase transcribes the MS2 region right after the exons and introns are transcribed ( Figure 3B ) . The recovery signal shows bi-phasic kinetics ( Figure S7A ) , which best fit to two exponentials . This implies that the measurements are detecting two processes that are occurring in parallel , which require the interpretation of the significance of each phase in the curves . The abovementioned studies have used extensive modeling to study such FRAP curves and demonstrated that the fast phase of fluorescence recovery is generated by the actively elongating polymerases . The slow phase of recovery was explained as either a population of polymerases in a paused state along the gene [34] or a termination delay [41] . The slow kinetics on the E6 gene could therefore be caused either by polymerase slowdown or by retention of the mRNA after termination . This was further tested . We first examined whether the kinetic effect was due to a change in polymerase activity . FRAP of the active polymerase transcribing on the E3 and E6 genes was analyzed using a transfected GFP-Pol II that was recruited to the active genes . Similar FRAP recovery kinetics were measured on both genes ( Figure 5A ) . This measurement differs from the above measurement of YFP-MS2 FRAP recovery . While the latter measures elongation through the MS2 region only , Pol II FRAP encompasses the whole polymerase transcriptional process on the gene , ranging from initiation through elongation and termination [34] . The Pol II FRAP measurements ( Figure 5A ) indicated that the polymerases on E3 and E6 genes were functioning with similar kinetics with regard to the general transcriptional flow . However , since only the very slow recovery phase is attributed to the elongating phase of the polymerase ( consists of only <30% of the signal ) [34] , it is difficult to resolve between the polymerase kinetics on E3 and E6 . Therefore , we obtained chromatin immunoprecipitation ( ChIP ) information to compare the distribution of Pol II along the genes , which might detect regions of paused polymerases . Using pull-down of the polymerase with an anti-Pol II antibody followed by real-time PCR on 7 different gene regions , we corroborated the polymerase FRAP findings and demonstrated that there was no statistical differences in the density of Pol II along the two genes ( Figure 5B ) . Additionally , from these data we can learn that the highest Pol II density is found at the promoter and the polyA site , which is in agreement with other studies [16] . Altogether , these results showed that the detected kinetic change in E6 was not directly connected to the polymerase . To examine whether the E6 kinetic delay was due to ongoing splicing , we used two splicing inhibitors , Spliceostatin A ( SSA ) [48] and Meayamycin [49] . SSA and Meayamycin are inhibitors of pre-mRNA splicing via SF3b binding , which binds to the intron branch-point and acts as a subcomplex in the U2 snRNP . SSA treatment caused the accumulation and retention of unspliced pre-mRNA in the nucleoplasm , which continued to accumulate in nuclear speckles , as detected by RNA-FISH with an intron probe ( Figure 5C ) , as previously described [48] . We now show this same effect also for Meayamycin ( Figure S8A ) . Interestingly , the dispersal of the intronless E1 mRNA was not affected , and it did not accumulate in speckles ( Figure 5D ) . Then we measured YFP-MS2 recoveries by FRAP on the E6 gene in untreated cells as described above and incubated the same transcribing cells with either SSA or with Meayamycin . Under both treatments the slow kinetics of E6 were transformed to the regular kinetics of E3/E4/E1 genes ( Figure 5E and S8B ) , whereas splicing inhibition did not cause any effect on the kinetics of the E3 gene ( Figure 5E ) . On the other hand , FRAP of GFP-Pol II in the presence of SSA on the E6 gene ( Figure S8C ) , and ChIP experiments in the presence of SSA ( Figure 5F ) did not show any change in kinetics . These results demonstrate that splicing inhibition did not modify Pol II distribution or kinetics along the genes . This emphasized that the changes in kinetics between E6 and E3 genes were due to active splicing events and were not connected to the elongation activity of the polymerase . The above experiments suggested that the slow recovery FRAP curves of the YFP-MS2 experiments were due to a delay of the nascent transcripts on the E6 sites caused by splicing but without connection to the polymerase itself . First , we verified that the full transcripts were 3′-end processed and were present at the site of transcription of all the genes prior to release . This was shown by the recruitment of GFP-PABP2 to the active transcription sites and by the presence of polyA tails as seen by RNA-FISH ( Figure 6A and 6B ) . RNA-FISH measurements of the ratio of polyA signal to the last exon ( CFP ) on the sites showed that a constant ratio was retained on all the transcription sites ( Figure 6C ) , meaning that there was no accumulation of mRNAs lacking polyA on the different genes and no change in polyadenylation levels . This suggested that there was an accumulation of transcripts that were not attached to Pol II , at the transcription sites . To characterize the delay time of the mRNAs on the E6 transcription sites , we next examined the kinetics of mRNA release from the gene . Actinomycin D ( ActD ) was added to actively transcribing cells containing YFP-MS2 tagged transcription sites . The “shut-down” of the transcription sites was monitored over time under the microscope . Indeed , slower clearing of the mRNAs from the gene was observed in E6 cells ( t1/2 = 17 min ) compared to E3 cells ( t1/2 = 7 min ) and E4 ( t1/2 = 8 . 9 min ) ( Figure 6D , Video S3 ) . This experiment correlates with the YFP-MS2 FRAP measurements ( Figure 4D ) showing slower mRNA kinetics on the E6 gene . Finally , to demonstrate that only the mRNAs disconnected from the polymerase were stalled on the E6 gene , we quantified the intensity ratio of the polymerase signal and the mRNA signal on the transcription sites using anti-Pol II immunofluorescence and RNA-FISH to the first exon . If indeed each mRNA is attached to one polymerase , we would expect a constant ratio for all cell lines . However , we found that the proportion of mRNA to Pol II on E6 transcription sites was three times higher than on E3 and E4 sites ( Figure 6E ) . When SSA was added and splicing was inhibited , the mRNA/Pol ratio on the E6 sites was reduced to the levels of E3 and E4 sites , meaning that under conditions of splicing inhibition , the unspliced E6 pre-mRNAs were being released without stalling similarly to the E3 and E4 mRNAs . These experiments implied that in a situation like the E6 gene which contains closely spaced introns , and in which the polymerase has finished transcribing yet the pre-mRNA is not fully spliced , the mRNA is retained at the transcription site until the completion of splicing . To better understand the given observed results we simulated the transcriptional process using a stochastic Monte-Carlo simulation ( Figure S9A and Video S4 ) . The simulation mimicked the transcriptional process based on the transcriptional models described in the previous studies [34] , [41] . Kinetic parameters were obtained by fitting the simulations to the experimental FRAP curves of E3 ( Figures 6F ) . See Material and Methods for explanation of parameter scanning . The parameters that gave the best fit were: an elongation speed of 3 . 6 kb/min , agreeing with previous measurements [21] , [22] , [36]; ∼42% of the polymerases had entered a paused state for an average time of 6 min [34]; and an average retention time ( after end of elongation ) of 50 s before release from the transcription site ( Figures S9 and S10 ) [41] . We found that the number of genes in the array and the initiation rate had no effect on the measured kinetic results , due to the normalization of the data to the steady-state pre-bleach signal . This observation was confirmed experimentally by measuring FRAP experiments with low amounts of DRB to abrogate elongation . Indeed , no kinetic effect on E3 or E6 YFP-MS2 FRAP recoveries was observed ( Figure S11 ) , while the addition of SSA even in the presence of DRB transformed E6 kinetics to E3 kinetics , signifying the effect of splicing on the retention of mRNA at the site of transcription . We then searched for a reliable fit to the E6 data by keeping the kinetic parameters of E3 and changing only the time-scale of transcript release . We found that the average time for a transcript to remain at the E6 site of transcription was on average ∼10 min ( Figure 6F ) . This is in agreement with the actinomycin D experiment that showed a 10 min difference in transcript clearance times between E3 and E6 ( Figure 6D ) . We also used the simulations to calculate the ratio of polymerases to nascent transcripts on the active genes , and compared to the experimental data . The simulation showed a 3-fold increase in this ratio on the E6 gene , similar to the result obtained with quantitative FISH ( Figure 6E ) . We then used this simulation to examine the intron/exon ratio on the different genes . In this experiment ( Figure 4A ) we found a gradual increase in intron signal at the transcription sites that correlated to the number of introns in each of the genes . The simulation output was compared to these RNA-FISH measurements . The simulation provided a ratio between the number of polymerases/transcripts present on the duplicated intron 2 region and the number of transcripts that had already passed through the CFP region . Simulations were run for the different genes with the abovementioned kinetic parameters assigned to E3 ( short retention time ) . Since this experiment follows the polymerase moving along the genes ( = elongation ) , we expected an increase in the intron/exon ratio from E1 to E6 genes . Indeed , the simulation showed the same increment in intron/transcript ratio as seen in the experimental results ( compared in Figure 6G ) . However , when the calculated long retention kinetics at the end point of the gene was included in the simulation of E6 kinetics , we obtained an output that showed a low intron/exon ratio , which did not agree with the experimental data ( Figure 6G ) . Therefore , the continuous flow of intron generation observed with the simulation , in contrast with the slow kinetics of the transcript , imply that the release time of the transcript from the gene site depends on splicing and not on the polymerase kinetics .
Gene expression profiles are influenced by the presence or lack of an intron . It has been known for many years that the addition of an intron into the cDNA of a gene sequence enhances its expression [50] . Indeed , the act of intron removal can have effects on the initiation of transcription , mRNA polyadenylation , mRNA export , and the cytoplasmic fate of the transcript such as translation , mRNA localization , and mRNA turnover [51] . For instance , several studies have demonstrated that the presence of an intron in the pre-mRNA and subsequent splicing can elevate the rates of mRNA export and thereby the accumulation of message in the cytoplasm [52]–[54] . Hence , the presence of an intron exerts an overall positive influence on gene expression . How does intron removal affect the levels of transcription and mRNA production ? Some level of regulation can be found on the DNA itself , for instance by intronic sequences that modulate nucleosome assembly in the vicinity of the promoter [55] . Splicing signals in the 5′ promoter-proximal intron can stimulate polymerase II initiation , processivity , and recruitment of transcription factors [56]–[59] , and complexes containing transcription and splicing factors have been identified [60] . Nonetheless , the increase of Pol II activity on intron-containing genes versus intronless genes was modest [56] . Recently , the depletion of the SR protein SC35 in cells was shown to attenuate the elongation of Pol II within the gene sequence [61] . In this study we wished to examine the effects of splicing on Pol II elongation in vivo , independent of promoter influences and transcriptional initiation events . Therefore , all the gene constructs were controlled by the same inducible promoter , and elongation kinetics were examined within the gene as far as possible from the initiation site . Elongation rates were measured immediately downstream of the region transcribing the exon-intron-exon elements , or in other words we could monitor the polymerase speed as it moved along the DNA after the splicing codes in the pre-mRNA had already been synthesized . This meant that polymerase movement towards the 3′-end of the gene was occurring simultaneously with splicing events accompanying the newly synthesized pre-mRNA . We measured intron accumulation on the active transcription sites that was dependent on intron numbers . The presence of various splicing machinery factors at the transcription sites was detected only when the gene loci were actively transcribing . Furthermore , RNA-FISH to intronic and exonic sequences demonstrated the presence of the introns only at the sites of transcription and not in the nucleoplasm , signifying that this is the location of the removal of these introns . However , when splicing was inhibited with Spliceostatin or Meayamycin , intron-containing pre-mRNA could be detected also throughout the nucleus , in particular within nuclear speckles , as previously described for SSA [48] . The intronless E1 mRNA was not affected by these treatments . Splicing factors associated with the actively transcribing genes belonged to the U snRNP family , including snRNAs ( U1 , U2 , U4 , U5 , U6 ) and protein components ( U1A , U1-70K , U2AF65 ) . Also SR proteins such as SC35 , SF2/ASF , and 9G8 were recruited to the active transcription sites upon activation . Other factors detected at the sites were the transcription/splicing-related protein TLS/FUS , splicing regulator PTB , the elongation regulator cyclin T1 , and polyA-binding protein 2 ( PABP2 ) . However , factors like PSF and p54nrb [62] previously demonstrated to be associated with the Pol II CTD [63] were not detected at the transcribing genes , nor were they pulled down with a purified polymerase [46] . Distinguishing between the two approaches might hint to transcription factor dynamics . While purified protein complexes can give an idea of the “presence” of a factor in a transcription complex , the imaging assays can indicate the longevity of “residence time” at the active gene . The fact that these factors were not identified accumulating on the transcribing gene arrays using antibodies to the endogenous proteins could indicate short-lived or low-affinity interactions with other factors that do associate with the actively transcribing gene . In this study we examined spliceosomal recruitment to comparable intron-containing genes as well as an intronless gene . As anticipated , spliceosomal factors were not found on the active transcription site of an intronless gene , except for components of the U1 snRNP . These data are in agreement with a recent report that quantified the recruitment of splicing factors to an intron-containing gene in which the splice sites were mutated and therefore did not undergo splicing [64] . In that study the recruitment of U1 snRNA and U1-70K were detected on the gene . In our analysis the intronless E1 gene does not contain any splice sites and therefore strengthens the notion that U1 snRNP is associated with the CTD of Pol II by default and presumably accompanies the polymerase during transcription in order to scan for the presence of a 5′ splice site . Only when such positive identification of splicing signals has occurred will the rest of the splicing machinery be recruited to the pre-mRNA . We quantified an increased accumulation of spliceosomal components on the pre-mRNA that was dependent on the number of introns within the gene sequence . These observations assist in understanding the dynamics of spliceosome assembly on multiple introns and would be consistent with the scenario of a step-wise model of spliceosome assembly on each intron [65] , as well as exon-tethering by the polymerase [66] . Our live-cell analysis suggests that the polymerase does not wait for splicing to take place but rather runs ahead such that more introns accumulate . Together with the exon-tethering model it would mean that exons are being held through the CTD and thereby recruiting the spliceosome for each splice site . The kinetic approach we used to measure Pol II transcription kinetics enables the quantification of elongation that is free of influences of initiation at the promoter [34] . The region in which elongation is measured is the MS2 portion of the mRNA in which the generation of each additional MS2 stem-loop results in an increase of YFP-MS2 signals on the transcription site . We inserted the MS2 region downstream of the exon-intron-exon modules and at the 3′-end of the mRNA , thereby monitoring the elongating polymerase when associated with the splicing machinery . When we compared the elongation kinetics of the polymerase moving through the MS2 region after transcribing an E3 or E4 mRNA , we found that the kinetics were similar and in fact did not differ than when transcribing an intronless E1 mRNA . A recent study examining transcription of a variety of endogenous genes along intronic and exonic regions using quantitative RT-PCR has shown that polymerase rates were similar on all genes and did not differ when transcribing intron or exon regions [21] . By generating a gene with many and closely spaced introns and exons ( E6 ) , we created a situation in which the transcription machinery should have reached the end of the gene before full splicing of the transcript was completed . Although the generation of the E6 and E4 genes required the duplication of intron 2 and the flanking splice sites , we did not find that this had an effect on the splicing outcome ( RT-PCR ) of E6 and E4 pre-mRNAs , nor on the expression levels of the CFP-peroxisome protein product . Also , there was no indication of aberrant localization of the E6 mRNA on the transcription site or in nuclear speckles , which could be indicative of aberrant mRNA production . The FRAP recovery kinetics of the E6 gene were significantly slower than the E3 and E4 genes . In order to examine whether the polymerase was the cause of these slow kinetics , or if the splicing machinery was responsible , we conducted a series of experiments . Using splicing inhibitors we saw that the impediment of the slow kinetics was released and returned to the levels of E3/E4/E1 genes , demonstrating that the slow-down was due to an active splicing environment . On the other hand , GFP-Pol II recovery kinetics and ChIP distribution patterns demonstrated that the polymerase was not paused or hindered during elongation . Use of actinomycin D that inhibits transcription by intercalating into the DNA and obstructing the path of the polymerase demonstrated that the clearance of the mRNAs from the E6 gene was slower . These transcripts were polyadenylated as seen by recruitment of PABP2 and the presence of polyA tails . The four different transcripts showed polyA accumulation on the transcription sites implying that the transcripts remained attached to the gene until 3′-end processing occurred . The quantification of the ratio of polymerase to nascent transcripts showed that there was accumulation of pre-mRNAs on the E6 gene , but not on the other genes . This accumulation was alleviated when splicing was inhibited . Together with parameter estimation by the simulation we demonstrate that unspliced E6 transcripts are retained at the site of transcription for up to 10 min before being released . The plausibility of a co-transcriptional splicing mechanism can be imagined just by the fact that human gene exons tend to be short compared to extremely larger intronic sequences [67] , and genes contain an average of 5–12 exons [68] . This means that if co-transcriptional splicing did not take place , then the splicing machinery would have to hold back until transcription has completed . This would not make much biological sense , especially when envisaging the transcriptional timeframe of very large genes such as the dystrophin gene , which lies in the range of 16 h [69] . In fact it has been proposed that the immediate association of U1 snRNP and splicing factors with the emerging pre-mRNA is a means to block the association with hnRNPs , which are splicing inhibitory proteins [46] . We demonstrated that the polymerase activity is uncoupled from splicing once transcription has terminated . In a scenario in which transcription has ended and splicing continues , the pre-mRNA and splicing machinery are retained at the site of transcription until the completion of splicing .
E4 generation: The E3 construct ( termed E3-minus in [37] ) was digested with BstXI . Re-circularization was prevented by Shrimp Alkaline Phosphatase ( Fermentas ) . An adaptor composed of the hybridized A1 & A2 sequences was ligated using T4 DNA ligase ( 5u/µl , Fermentas , Ontario , Canada ) . A1:TTGATATCACCGGTGGGCCCGGAAAAGAATTCGTCCATCACT A2:ATGGACGAATTCTTTTCCGGGCCCACCGGTGATATCAAAGTG The adaptor was EcoRV/BstXI digested , and a PmlI/BstXI intron2-exon3 insert that was separately removed from E3 was then ligated into the adaptor to generate E4 . E6 was constructed using the same strategy but on the E4 construct . E1 was generated by first digesting the E3 plasmid BstXI/SacII , followed by ligation with an adaptor composed of A9 & A10 . A9:AAGCTTTGATCAGCGGCCGCATCACT A10:GGTTCGAAACTAGTCGCCGGCGTA RNA extractions were carried out with the RNeasy mini kit ( Qiagen ) . cDNA was synthesized using the RevertAid™ First Strand cDNA Synthesis Kit and an oligo ( dT ) primer ( Fermentas ) . RT-PCR and sequencing of the correctly spliced mRNAs were performed with primers to the first exon and last exon regions . A3: AGCAACCTCAAACAGACACC A4: GGTCTTGTAGTTGCCGTCGT Other plasmids used in this study were: RFP-LacI , CFP-LacI , YFP-MS2 ( -NLS ) , GFP-Pol II [34] , ( NES- ) YFP-MS2 ( -NLS ) [36] , MS2-mCherry , GFP-PABP2 , GFP-PTB , GFP-TLS , and GFP-cyclin T1 . Human U2OS osteosarcoma Tet-On cells ( Clontech ) were maintained in low glucose Dulbecco's modified Eagle's medium ( DMEM , Biological Industries , Israel ) containing 10% fetal bovine serum ( FBS , HyClone , Logan , UT ) . For transient transfections , cells were transfected with 1–5 µg of plasmid DNA and 40 µg of sheared salmon sperm DNA ( Sigma ) using electroporation ( Gene Pulser Xcell , Bio-Rad , Hercules , CA ) or with FuGENE 6 transfection reagent ( Roche ) . Stable co-transfections were performed by electroporation of the E genes together with a 256 lacO repeat plasmid and a puromycin resistance plasmid using a 1∶10∶10 ratio ( puromycin res . , lacO , E gene ) . Cells were cultured for 1–2 d without selection and then selected for 2–3 wk with 100 µg/ml puromycin ( AG scientific , San Diego , CA ) . Single colonies were picked and positive clones were selected after the addition of doxycycline ( 1 µg/ml ) and identification of CFP-peroxisomes and by transient transfection of RFP-LacI and YFP-MS2 . For drug treatments cells were imaged before and right after the addition of 5 mg/ml actinomycin D ( Sigma ) . For DRB treatment , cells were treated with 20 µg/µl for 2 h before imaging . For splicing inhibition , cells were photobleached before and after incubation with 100 ng/ml of Spliceostation for 6 h , kindly provided by Minoru Yoshida ( RIKEN Advanced Science Institute , Japan ) , or Meayamycin ( 300 nM ) for 6 h before imaging , kindly provided by Kazunori Koide ( University of Pittsburgh , USA ) . gDNA was isolated using the PUREGENE™ DNA Purification System ( Gentra ) . For the generation of standards a 10-fold dilution was prepared . Primers used for the amplification of CFP region were GGATCACTCTCGGCATGGAC ( forward ) and TGCACATACCGGAGCCATTG ( reverse ) . Primers used for the human β-globin reference gene were CAGTGCAGGCTGCCTATCAGA ( forward ) and GAATCCAGATGCTCAAGGCCCTT ( reverse ) . The reaction ( 20 µl total volume ) contained 1× SYBR green mix ( Thermo scientific , Waltham , MA ) , 1 µl of gDNA dilution , and 0 . 5 mM of each primer and water . The reaction was performed using the Chromo4™ Real-Time Detector ( Biorad ) following the protocol: 15 min at 95°C for enzyme activation , followed by 40 cycles consisting of 15 s denaturation at 92°C , 10 s annealing at 64°C , and 20 s extension at 72°C . Fluorescent signal was measured at the end of each cycle . A final dissociation stage was performed to generate a melting curve for verification of amplification product specificity . For calculating the gene copy number , we first calculated the reaction efficiency for each set of primers by using triplicates of 5× dilutions of gDNA , using the provided software Opticon2 ( Biorad ) . We measured the average cycle number for each gDNA sample using both primer sets and calculated the concentration ( Qc ) of each reaction by: . Then we could calculate the gene copy number of the samples: We verified that the U2OS cell line has only two alleles of the β-globin gene by comparing to a non-cancerous cell line . ChIP and quantitative PCR were performed as previously described [18] . After an overnight induction with 1 µg/ml doxycycline , cells were harvested and lysed in SDS lysis buffer . Protein concentration was determined via Amido Black and 1 mg total protein was used per IP . For SSA treatment , cells were treated with 100 ng/ml SSA prior to lysis . Antibodies used for the ChIP were 5 µg/IP CTD4H8 against RNA polymerase II ( Santa Cruz Biotechnology , SC-47701 ) and 10 µg/IP non-immune mouse IgG ( Sigma , I5381 ) for the mock ChIPs . Uncrosslinking and Proteinase K treatment ( 100 µg/IP ) were performed for 6 h at 65°C . DNA was extracted using the Qiagen PCR purification kit and used as a template for quantitative PCR . Primer sequences used in the qPCR analysis can be obtained upon request . For each amplicon , ChIP signal relative to input was calculated based on the Ct values . The values obtained for control antibody were then subtracted from those of specific antibodies . Thus , the following equation was used: 2 ( Ct ( Input ) −Ct ( spIP ) ) −2 ( Ct ( Input ) −Ct ( cIP ) ) , where: spIP−IP with specific antibody and cIP−IP with a control antibody . Values obtained for amplicons within gene regions were further normalized to the promotor region . Cells grown on coverslips were fixed for 20 min in 4% PFA and then permeabilized in 0 . 5% Triton X-100 for 3 min . After blocking with 5% BSA , cells were immunostained for 1 h with: mouse anti-RNA Pol II antibody ( clones H14 , H5 , 8WG16 ( Covance ) ) , mouse anti-SF2/ASF ( Zymed ) , mouse anti-PSF ( B92 , Sigma ) , mouse anti-2 , 2 , 7-trimethylguanosine ( TMG ) ( Mercury ) , rabbit anti-U1A , rabbit anti-U2AF65 ( Abcam ) , mouse anti-SC35 ( Sigma ) , mouse anti-9G8 , and mouse anti-p54nrb ( BD Bioscience ) . After subsequent washes the cells were incubated for 1 h with a secondary Ab: Cy5-labeled goat anti-mouse IgG , Cy3-labeled goat anti-mouse IgG , FITC-labeled anti-mouse IgM ( Jackson ImmunoResearch , West Grove , PA ) , Alexa488-labeled goat anti-mouse IgG , Mouse anti-rabbit Alexa488 , and rabbit anti-mouse Alexa594 ( Molecular Probes ) . Coverslips were mounted in mounting medium . RNA FISH was performed on dox activated cells as previously described [70] using 10 ng ( probe per slide ) of 5′-Cy3 or Cy5 labeled fluorescent probes , as follows: CFP: ATATAGACGTTGTGGCTGATGTAGTTGTACTCCAGCTTGTGCCCCA GGATA Exon1: CATGAATTCTTTGCCAAAGTGATGGGCCAGCACACAGACCAGCA CGTTGC Intron1: TCTCGTGACAGAGAAGGGTGTAAAAGCTTCTAGCCTTTTCTCTTA CCTTA . Intron2: TAGCAAAAGGGCCTAGCTTGGACTCAGAATAATCCAGCCTTATC CCAACC MS2: CTAGGCAATTAGGTACCTTAGGATCTAATGAACCCGGGAATACT GCAGAC PolyA: an oligo ( dT ) probe The U snRNA set of probes containing 5-amino-allyl thymidines ( Eurogentec ) were labeled with Cy3 or Cy5 monoreactive dyes ( GE Healthcare ) . U1-1: CtGGGAAAACCACCtTCGTGATCAtGGTATCTCCCCtGCCAGGTAAGtAT U1-2: CGAACGCAGtCCCCCACtACCACAAATTAtGCAGTCGAGtTTCCCACAtT U2: AGtGGACGGAGCAAGCtCCTATTCCAtCTCCCTGCtCCAAAAATCCATtT U4-1: tAGCAATAAtCGCGCCTCGGAtAAACCTCATtGGCTACGATACtGCCACT U4-2: AGACtGTCAAAAATtGCCAATGCCGACtATATTGCAAGtCGTCACGGCGGtA U5: GGCAAGGCTCAAAAAATTGGGTTAAGACTCAGAGTTGTTCCTCTCCACGGTA U6: CGtGTCATCCTtGCGCAGGGGCCAtGCTAATCTTCtCTGTATCGTtCCAA 7SK snRNA: TtGGATGTGTCTGGAGTCTtGGAAGCTTGACTACtT In some cases , an immunostaining was performed after FISH using the standard protocol . Wide-field fluorescence images were obtained using the Cell∧R system based on an Olympus IX81 fully motorized inverted microscope ( 63× Plan-Apo , 1 . 42NA and 100× Plan-Apo , 1 . 4NA ) fitted with an Orca-AG CCD camera ( Hamamatsu ) , rapid wavelength switching , active focus control ( ZDC ) , a motorized stage ( Scan IM , Märzhäuser , Wetzlar-Steindorf , Germany ) , and driven by the Cell∧R software . The microscope is equipped with an on-scope incubator which includes temperature and CO2 control ( Life Imaging Services , Reinach , Switzerland ) . For time-lapse imaging of ActD treatment , several cell positions under non-treated conditions were chosen and recorded . Then 5 mg/ml of actinomycin D was added and cells were imaged every 5 min for 1 . 5 h . Images and videos were analyzed using self-written ImageJ software macros ( National Institutes of Health , Bethesda , MD ) . The co-localization intensity profile was generated with an ImageJ macro . The channel intensity correlation was generated by Matlab . The 4D image sequences were sum Z-projections . The transcription site was tracked and the mean intensity was measured at every time-point by finding the local maxima of the red ( RFP-LacI ) for each time point in a specific region of interest ( ROI ) . The background was taken from an ROI outside of the cell and subtracted from all other measurements . Photobleach correction was obtained using another area in the nucleus or in another nucleus . Images acquired before drug addition were used as the initial conditions for normalization of the data . The half-time of site inactivation was calculated for each transcription site using Matlab . Each transcription site measured intensity dataset was fitted to a bi-exponential equation f ( x ) = a*exp ( b*x ) +c*exp ( d*x ) . We then searched for the , which brings the equation to the value of . For each E cell type we collected at least 7 cell transcription site measurements , and the averages and standard error were calculated . FRAP experiments were performed using a 3D-FRAP system ( Photometrics ) built on an Olympus IX81 microscope ( 63× Plan-Apo , 1 . 4 NA ) equipped with an EM-CCD ( Quant-EM , Roper ) , 405 & 491nm lasers , Lambda DG-4 light source ( Sutter ) , XY&Z stages ( Prior ) , and driven by MetaMorph ( Molecular Devices ) . Experiments were performed at 37°C with 5% CO2 using a live-cell chamber system ( Tokai ) . Before and after the bleaching , cells were imaged in the YFP channel for the detection of YFP-MS2 ( mRNPs ) and in the CFP channel for the detection of CFP-lacI ( genomic locus ) . For each acquisition , 7 Z-slices were taken every 350 nm . The active transcription site was bleached using the 491 nm laser . Six pre-bleach images were acquired . Post-bleach images were acquired in a sequence of 3 time frequencies: 15 images every 3 s , 15 images every 6 s , and 26 ( or 45 for long experiments ) images every 30 s . Experiments were performed on two additional systems to verify their consistency: ( 1 ) a similar 3D FRAP system in the Darzacq laboratory under the same conditions; ( 2 ) Zeiss LSM 510 Meta laser scanning confocal microscope ( Plan-Apochromat 63× , 1 . 4 NA oil objective ( Jena , Germany ) ) . Cells were scanned at 488 nm for GFP-MS2 or 514 nm for YFP-MS2 , together with a 543 nm laser for the detection of RFP-lacI at the genomic locus . Post-bleach recovery images were acquired every 1 s . The experiments were analyzed using self-written ImageJ macros that sum the projected Z-stacks at every time-point , and track and measure the mean intensity of the transcription site at every time-point by finding the local maxima in a specific region of interest ( ROI ) . For each time-point , the background taken from a ROI outside of the cell was subtracted from all other measurements . T ( t ) and I ( t ) were measured for each time-point as the average intensity of the nucleus and the average intensity in the bleached ROI , respectively . The average of the pre-bleach images used as the initial conditions are referred to as Ti = nuclear intensity and Ii = intensity in ROI before bleaching . Ic ( t ) is the corrected intensity of the bleached ROI at time t [42] , [71]: Data from at least 10 experiments for each cell-line were collected and the averaged FRAP measurements were fitted to the simulation . The diffusion of the YFP-MS2 protein during the FRAP analysis was disregarded since the diffusion rate of free nucleoplasmic YFP-MS2 is very rapid , while the bound YFP-MS2 is associated with high affinity to the mRNA , and does not detach , diffuse , and bind again at the transcription site [34] , [41] , [42] . Quantitative RNA FISH was used to compare the ratio of FISH signals from two different probes in two different channels . Although each cell line had a different number of gene-copies integrated , the ratio per transcript is comparable . Z stacks ( 91 slices , 200 nm steps ) of the immunostained or RNA FISH samples were collected . Images were acquired for the quantification of the exons , introns , U snRNAs , or Pol II present on the transcription sites . Each experiment in which all E cell types were analyzed were acquired under the same conditions and on the same day . Images were deconvolved using a point spread function ( PSF ) based deconvolution algorithm Huygens Essential II ( Scientific Volume Imaging , Hilversum , The Netherlands ) . The sum of pixel values at the transcription site in each channel were measured using Imaris ( Bitplane , Saint Paul , MN ) , and the ratio between the channels is presented . Monte Carlo simulations ( Matlab ) of the transcriptional process were based on the mechanistic models of [34] , [41] . The simulation was used to simulate the FRAP experiments , and by fitting the simulation to the experimental data we could retrieve the kinetic parameters of the process . The simulation performs stochastic decisions by using random numbers obtained from the simulation and checks whether they are smaller than the kinetic parameters . For each gene ( E1 , E3 , E4 , and E6 ) , we generated a set of identical arrays where the length of each array corresponded to the length of the gene . RNA polymerases were implemented as counters that “slide” over these arrays . “Transcriptional initiation” occurred stochastically at a given “initiation rate , ” after which a constant speed of elongation was retained . Whenever a polymerase moves through the “MS2 region” of the gene , the polymerase/transcript accumulates a “fluorescent signal” that was maintained until the end of the gene . Polymerases were stochastically released at a given termination rate [41] . Polymerases could also randomly enter and exit a paused state [34] . The simulation reached steady state after long times , as expected . To simulate the FRAP experiments , we tracked the total “fluorescent” signal . After the system entered steady state , we “bleached” ( deleted ) the fluorescent molecules and measured the buildup of “fluorescent” signal during the recovery . The signal was normalized as in the FRAP experimental procedure . The kinetic parameters , namely , the elongation speed , the termination rate , and the rates of entering and exiting the paused state , were chosen as those that best fit the experimental FRAP data . We found that the number of gene arrays and the initiation rate had no effect on the measured kinetic results , due to the normalization to the pre-bleach signal . To find the above four kinetic parameters that best fit the experimental data we had to explore a four-dimensional parameter space . For each configuration , the mean square difference ( MSD ) between the simulated FRAP curve and the experimental curve was calculated ( Figure S10 ) . For the E3 gene , the kinetic parameters that gave the best fit ( lowest MSD ) were: 3 . 6 kb/min , ∼42% of the polymerases entering a paused state for an average of 6 min , and transcript retention of 50 s after the end of elongation . To obtain the kinetic parameters for the E6 gene , we used the best fit values of the E3 elongation speed and pausing transition rates , and changed only the termination rate . | The pre-mRNA emerging from RNA polymerase II during eukaryotic transcription undergoes a series of processing events . These include 5′-capping , intron excision and exon ligation during splicing , 3′-end processing , and polyadenylation . Processing events occur co-transcriptionally , meaning that a variety of enzymes assemble on the pre-mRNA while the polymerase is still engaged in transcription . The concept of co-transcriptional mRNA processing raises questions about the possible coupling between the transcribing polymerase and the processing machineries . Here we examine how the co-transcriptional assembly of the splicing machinery ( the spliceosome ) might affect the elongation kinetics of the RNA polymerase . Using live-cell microscopy , we followed the kinetics of transcription of genes containing increasing numbers of introns and measured the recruitment of transcription and splicing factors . Surprisingly , a sub-set of splicing factors was recruited to an intronless gene , implying that there is a polymerase-coupled scanning mechanism for intronic sequences . There was no difference in polymerase elongation rates on genes with or without introns , suggesting that the spliceosome does not modulate elongation kinetics . Experiments including inhibition of splicing or transcription , together with stochastic computational simulation , demonstrated that pre-mRNAs can be retained on the gene when polymerase termination precedes completion of splicing . Altogether we show that polymerase elongation kinetics are not affected by splicing events on the emerging pre-mRNA , that increased splicing leads to more splicing factors being recruited to the mRNA , and that post-transcriptional splicing can proceed at the site of transcription in the absence of the polymerase . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/gene",
"expression",
"cell",
"biology/nuclear",
"structure",
"and",
"function"
] | 2011 | The In Vivo Kinetics of RNA Polymerase II Elongation during Co-Transcriptional Splicing |
The QT interval ( QT ) is heritable and its prolongation is a risk factor for ventricular tachyarrhythmias and sudden death . Most genetic studies of QT have examined European ancestral populations; however , the increased genetic diversity in African Americans provides opportunities to narrow association signals and identify population-specific variants . We therefore evaluated 6 , 670 SNPs spanning eleven previously identified QT loci in 8 , 644 African American participants from two Population Architecture using Genomics and Epidemiology ( PAGE ) studies: the Atherosclerosis Risk in Communities study and Women's Health Initiative Clinical Trial . Of the fifteen known independent QT variants at the eleven previously identified loci , six were significantly associated with QT in African American populations ( P≤1 . 20×10−4 ) : ATP1B1 , PLN1 , KCNQ1 , NDRG4 , and two NOS1AP independent signals . We also identified three population-specific signals significantly associated with QT in African Americans ( P≤1 . 37×10−5 ) : one at NOS1AP and two at ATP1B1 . Linkage disequilibrium ( LD ) patterns in African Americans assisted in narrowing the region likely to contain the functional variants for several loci . For example , African American LD patterns showed that 0 SNPs were in LD with NOS1AP signal rs12143842 , compared with European LD patterns that indicated 87 SNPs , which spanned 114 . 2 Kb , were in LD with rs12143842 . Finally , bioinformatic-based characterization of the nine African American signals pointed to functional candidates located exclusively within non-coding regions , including predicted binding sites for transcription factors such as TBX5 , which has been implicated in cardiac structure and conductance . In this detailed evaluation of QT loci , we identified several African Americans SNPs that better define the association with QT and successfully narrowed intervals surrounding established loci . These results demonstrate that the same loci influence variation in QT across multiple populations , that novel signals exist in African Americans , and that the SNPs identified as strong candidates for functional evaluation implicate gene regulatory dysfunction in QT prolongation .
The QT interval ( QT ) , as measured by the resting 12-lead electrocardiogram ( ECG ) , reflects the duration of ventricular depolarization and repolarization , providing a non-invasive assessment of an average ventricular action potential . QT prolongation is an established risk factor for ventricular tachyarrhythmias [1] , coronary heart disease [2] , and sudden cardiovascular as well as all-cause death [3] . Although numerous factors influencing QT have been identified , including heart rate [4] , structural heart disease [5] , [6] , gender [7] , [8] , age [9] , and medication use [10]–[12] , a large portion of the variance in QT remains unexplained . Several lines of evidence support a genetic contribution to QT . Initial evidence was provided by studies of inherited cardiac arrhythmias including long- and short-QT syndromes , which identified rare and highly penetrant mutations in ion channel and ion channel associated genes associated with QT [13] . Family studies have also reported that ventricular repolarization ( as measured by QT ) is heritable [14]–[17] . In addition , recent genome-wide association ( GWA ) studies performed in populations of predominantly European descent have identified common SNPs in twelve loci , including NOS1AP , KCNH2 , and PLN [18]–[23] , that influence the distribution of QT . To date , the majority of published GWA studies examining QT have been performed in populations of European descent , although one study also examined an Indian Asian population [23] . It is therefore unclear whether previously identified QT loci are relevant in other racial groups such as African Americans or whether there are population-specific SNPs influencing QT . Furthermore , the increased genetic diversity in populations of African ancestry provides opportunities for the narrowing and fine-mapping of loci identified in European and Indian Asian populations [24] . Fine-mapping , which includes the dense genotyping of common and rare SNPs at already established loci , is a helpful next step in the identification of functional polymorphisms underlying the QT distribution . For example , dense genotyping can capture rarer SNPs that may be inadequately represented on frequently used genome-wide genotyping arrays , and which may have large effects , thereby potentially helping to explain a larger fraction of the QT heritability [25] , which is estimated to range from 35% to 52% [14]–[17] . In this study , we evaluated eleven QT loci previously identified in populations of European and Indian Asian descent in 8 , 644 African American participants from the Atherosclerosis Risk in Communities study ( ARIC ) and Women's Health Initiative clinical trial ( CT ) . In addition to testing the previously reported QT index SNPs at the known loci , we also searched for stronger markers of the index signal and investigated evidence for independent , novel SNPs influencing QT in African Americans . For loci associated with QT in African Americans , we also investigated whether patterns of linkage disequilibrium ( LD ) within African Americans could narrow the regions likely to harbor the biologically relevant variant . Finally , we queried bioinformatic databases and performed related in silico analyses to identify potential candidate polymorphisms for follow-up functional evaluation .
The Metabochip , a custom array containing approximately 195 , 000 SNPs , is designed to facilitate fine-mapping of loci associated with cardiovascular and metabolic traits , including QT , blood pressure , cholesterol , type 2 diabetes , and anthropometrics . To identify and fine-map signals associated with QT in African Americans , 6 , 670 SNPs from eleven previously identified QT loci represented on the Metabochip are examined ( Table S2; Methods ) . The number of SNPs at each locus with minor allele frequency ( MAF ) estimates ≥1% ranges from 51 to 1 , 371 , corresponding to regions spanning 67 Kb to 664 Kb in size ( median size: 275 Kb ) . On average , imputation quality for metabochip SNPs imputed in WHI SNP Health Association Resource ( SHARe ) participants ( n = 3 , 531 ) is high across the 11 loci , with the exception of RNF207 , KCNH2 , and KCNQ1 ( Table S2 ) . As described in the Methods , SNPs with imputation quality scores <0 . 95 are discarded . The six published GWA studies of QT ( five studies of European ancestral populations , one study of European and Indian Asian ancestral populations ) reported 25 index SNPs ( P≤5 . 0×10−8 , Table 1; D' estimates provided in Table S3 ) across eleven loci , which together represent fifteen independent signals ( at r2≥0 . 20 in European ancestry populations ) . The fifteen signals include three independent signals at NOS1AP and two independent signals at both PLN and KCNH2 . We first test the fifteen independent signals for association with QT in African Americans . Specifically , we evaluate the index SNPs for each of the fifteen independent signals as well as all SNPs in LD with the index SNPs ( r2≥0 . 20 using European ancestral LD patterns ) to determine whether any of the fifteen independent signals generalize to African Americans . The significance criterion , αa = 1 . 20×10−4 , is based on the number of tag SNPs in African Americans that capture ( r2≥0 . 80 , using African American LD patterns ) all SNPs that are correlated with the index SNPs ( r2≥0 . 20; determined using European ancestral LD patterns; see Methods ) . Six of the fifteen independent signals are significantly associated with QT in African Americans ( Table 1 ) : NOS1AP independent signals 1 and 2 , ATP1B1 , PLN independent signal 1 , KCNQ1 , and NDRG4 . Of those that are not significantly associated with QT ( P-value>1 . 20×10−4 ) , estimates for index SNPs representing the two KCNH2 independent signals as well as the RNF207 index SNP show a consistent magnitude and direction of effect when compared to published estimates ( Table S4 ) . The best marker in African Americans for the first NOS1AP independent signal ( rs12143842 ) and KCNQ1 ( rs12296050 ) is identical to the European index SNP . For the ATP1B1 and NDRG4 index SNPs as well as the second NOS1AP independent signal , the best marker in African Americans shows only a slightly more significant association than the index signal ( <1 order of magnitude change in the P-value; Table 1 ) . In contrast , at the first PLN independent signal , the three index SNPs are not significantly associated with QT in African Americans ( P-value range: 0 . 034–0 . 60 ) , although a substantially stronger marker of the index signal is detected ( rs56403768; P-value = 3 . 8×10−5 ) . In Europeans , rs56403768 is correlated with the three index SNPs ( LD r2 range: 0 . 68–0 . 88 ) . However , patterns of correlation are weaker in African Americans ( LD r2 range: 0 . 39–0 . 50 ) . We then evaluate evidence for additional independent signals at the eleven previously identified QT loci , focusing on SNPs that are uncorrelated with the index signals in European populations . Here , statistical significance is defined using an efficient Monte Carlo approach that accounts for LD between SNPs at the eleven previously identified QT loci ( αb = 1 . 37×10−5; see Methods ) . Eleven SNPs at two loci that are uncorrelated with the index SNPs ( r2≤0 . 20; evaluated using both African American and European ancestral LD patterns ) exceed our significance threshold . Conditional analysis confirms that these eleven SNPs represent three novel associations , one flanking the NOS1AP locus and two residing within or nearby ATP1B1 ( Table 2; Figures 1 and S1 ) . Of note , the novel NOS1AP SNP is monomorphic in populations of European ancestry . Together , the six best markers in African Americans and the three novel SNPs explain 1 . 6% of the variance in heart rate-corrected QT . Next , we examine whether African American LD patterns can assist with the narrowing of association signals at the six loci that generalized to African Americans . An example of variation in LD patterns by ancestral population is shown by the first NOS1AP independent signal . Among African Americans , 0 SNPs are correlated ( r2≥0 . 50 ) with rs12143842 , which is the best marker in African Americans and the index SNP reported by three prior QT GWA studies ( Figure 1; Table 3 ) [20]–[22] . This is in contrast to LD patterns estimated in Europeans for NOS1AP independent signal 1 , where 87 SNPs are correlated with the three index signals that characterize this independent signal ( r2≥0 . 50 ) , representing a region spanning 114 . 2 Kb . For the five remaining markers ( Figures 1 and S1 , S2 , S3 , S4 ) , African American populations exhibit lower levels of LD as compared to European populations ( mean narrowing = 48 . 2 Kb ) . Likewise , fine-mapping in African Americans helped to narrow the association signals for all six loci . Bioinformatic analysis of the six best markers in African Americans and the three novel and independent SNPs ( Table S5; Methods ) does not identify any correlated non-synonymous coding variants; instead , all signals harbor variants that occur solely within non-coding regions with the potential for influencing cis-regulation ( Table S6 ) . Several variants occur within candidate regulatory elements ( promoter regions and DNase hypersensitive sites in human cardiomyocytes ) , including three ( rs3864884 , rs1646010 , and rs27097 ) that are predicted to have allele-specific binding affinities for various transcription factors of relevance to cardiac function ( Table S7 ) . Though further functional characterization is outside the scope of this study , rs3864884 , rs1646010 , rs27097 , and rs37036 represent compelling candidates for follow-up evaluation .
In this study composed of approximately 8 , 600 African American participants , we evaluated fifteen independent signals across eleven loci that were previously associated with QT in populations of European and Indian Asian ancestry at genome-wide significant levels . For five independent signals – the two NOS1AP independent signals , ATP1B1 , NDRG4 , and KCNQ1 – the best markers in African Americans were either identical to or only slightly more significant than the index signal . These five markers are therefore not considered better signals than the index SNP . However , for the first PLN locus , the three previously identified index SNPs were not significantly associated with QT in African Americans . This result suggests that rs56403768 , the best marker in African Americans , is a better proxy of a biologically important PLN allele and may help improve localization of the true association . In addition to generalizing six previously characterized QT loci , we also identified three novel and independent signals for NOS1AP and ATP1B1 . Notably , rs79163067 , the novel NOS1AP signal , was monomorphic in European populations . When these three novel and independent variants were combined with the fifteen independent QT loci previously identified in populations of European and Indian Asian ancestry , our results suggest that to-date at least 18 independent variants influence QT . Finally , we showed that evaluating LD patterns in admixed populations such as African Americans assisted with the narrowing of intervals flanking the putative causal variants . This narrowing was particularly evident for the first NOS1AP locus that included index SNP rs12143842 , the most frequently reported SNP in the QT GWA study literature to-date . Rs12143842 also was the SNP that explained the majority of variance in heart rate-corrected QT as well as the best marker in African Americans for the first NOS1AP independent signal . Rs12143842 has yet to be evaluated in any functional studies , and although our bioinformatics characterization did not identify any compelling functional candidates , the SNP resides less than five Kb from the annotated NOS1AP promoter . Future studies evaluating the functional relevance of rs12143842 are clearly indicated . Of the nine QT SNPs we identified ( i . e . the six best markers in African Americans and the three novel SNPs ) , all resided in or were in LD with SNPs residing in candidate long-range regulatory elements in human cardiomyocytes , annotated promoter regions , and highly conserved non-coding elements , and as such , strongly implicate gene regulatory dysfunction in QT prolongation . One of the more striking predictions was with rs3864884 , where the major allele [C] is predicted to bind an entirely different set of transcription factors ( TFs ) than the minor allele [T] . Notably , only the major allele is predicted to bind Hairy-related TFs , which are involved in regulating cardiac morphogenesis [26] . The minor allele is predicted to bind TBX5 and AhR; the former has been linked to a number of cardiac phenotypes including Holt-Oram syndrome [27] and atrioventricular conduction [28] and the latter regulates cardiac size [29] , a known risk factor for QT-prolongation and cardiac sudden death [30] . Consistent with this prediction , the minor allele is associated with an increased QT in this population . We were unable to identify associations with RNF207 , SCN5A , KCNH2 , LITAF , LIG3 , and KCNJ2 based on our statistical significance threshold . RNF207 was reported in two previous QT GWA studies . However , because of design priorities , only 51 SNPs with MAF≥1% were available for analysis . Future fine-mapping efforts in African Americans or other admixed populations that include denser genotyping of this locus may therefore be useful . SCN5A , KCNH2 , and LITAF were all reported by two prior GWA studies of QT . SCN5A also has also been implicated in GWA studies of PR in populations of European [31] and African ancestry [28] as well as QRS interval duration in populations of European descent [32] . Our inability to detect signals at these three loci may simply reflect inadequate power , especially for KCNH2 , for which estimates of the index SNPs in African Americans were consistent with the published literature . KCNJ2 is a biologically plausible locus influencing QT , as it harbors mutations causing rare , familial forms of long QT syndrome [33] . Yet , the high P-values and the dense genotyping coverage of SNPs suggest that KCNJ2 does not influence QT in African American populations . The genetic architecture of African Americans and other admixed populations is on average characterized by lower correlation between SNPs when compared to European populations . Such populations therefore are valuable for the fine-mapping of previously identified loci , as fewer SNPs are expected to be correlated with the underlying functional variant , which is expected to be the same in populations of different ethnicity . We therefore anticipate that future fine-mapping efforts that include populations with different ethnic backgrounds will be useful for the further refinement of loci influencing QT as well as the identification of population-specific variants , as demonstrated by the current report . Several limitations of the present study warrant further consideration in order to inform future efforts for fine-mapping and functional characterization of QT loci . First , although the Metabochip includes dense genotyping of most QT loci , it is possible that the causative variants are not included on the Metabochip . Second , our functional characterization is based on in silico analyses and requires experimental validation . Third , the majority of study participants were female . It is unclear how a predominantly female population may have influenced the results presented herein , considering the well-known dependence of QT on gender [7] , [8] . Finally , our results , which are consistent with prior studies [22] , show that common SNPs only explain a very small fraction of the variance in QT , although heritability estimates suggest a substantial genetic component . These modest effect sizes corroborate the multifactorial etiology of QT and demonstrate that substantially greater efforts are required to explain the “missing heritability” . Future efforts with increased sample sizes that examine rare variants , gene-gene and gene-environment interactions , and structural variants poorly captured on existing arrays are clearly needed [34] . In conclusion , our findings provide compelling evidence that the same genes influence variation in QT across ancestral populations and that additional , independent signals exist in African Americans . Moreover , all SNPs identified as strong candidates for functional evaluation implicate gene regulatory dysfunction . Further characterization of these loci , including direct sequencing and large-scale genotyping in African Americans and other admixed populations , may provide more information on the genetic and molecular mechanisms underlying QT .
The Institutional Review Board at all participating institutions approved the study protocol . This study was conducted according to the principles expressed in the Declaration of Helsinki . The Population Architecture using Genomics and Epidemiology ( PAGE ) study is a National Human Genome Research Institute funded effort examining the epidemiologic architecture of common genetic variants that have been reproducibly associated with human diseases and traits [35] . The PAGE study consists of a coordinating center and four consortia with access to large , diverse population-based studies including three National Health and Nutrition Examination Surveys , the Multiethnic Cohort , the WHI , the ARIC study , the Coronary Artery Risk Disease in Young Adults study , the Cardiovascular Health Study , the Hispanic Community Health Study/Study of Latinos , the Strong Heart Study , and the Strong Heart Family Study . This PAGE Metabochip study included African American participants from the ARIC and WHI CT studies . Participants from the other PAGE studies were excluded from this effort due to the unavailability of ECG measures and/or genotype data . Genotypes of WHI CT participants were obtained in three phases: two sets of women were directly genotyped on the Metabochip platform by PAGE investigators during wave 1 ( n = 797 ) and wave 2 ( n = 1 , 128 ) and women ( n = 3 , 531 ) with Metabochip variants imputed from previous genome-wide SNP data provided by the WHI SHARe [36] . Participants meeting the following criteria were excluded from the study: QT unavailable , atrial fibrillation/atrial flutter on ECG , left or right bundle branch block on ECG , QRS duration >120 milliseconds , intraventricular conduction delay on ECG , pacemaker implant antedating ECG , ancestry outlier , excessive heterozygosity , low call rate , or second member of first degree relative pair . Further details on the ARIC and WHI CT studies are provided in Text S1 ( Participating Studies ) . For each study , certified technicians digitally recorded resting , supine ( or semi-recumbent ) , standard 12-lead ECGs at study baseline for each participant using Marquette MAC PC machines ( GE Healthcare , Milwaukee , WI , USA ) . The ARIC and WHI CT studies used comparable procedures for preparing participants , placing electrodes , recording , transmitting , processing , and controlling quality of the ECGs . QT was measured electronically using the Marquette 12SL algorithm . The Metabochip was a custom Illumina iSELECT array that contained approximately 195 , 000 SNPs and was designed to support large scale follow up of putative associations for cardiovascular and metabolic traits including QT , blood pressure , cholesterol , type 2 diabetes , and anthropometrics . Approximately 33% of the Metabochip SNPs were included as replication targets and 62% for fine-mapping . In total , 257 loci were selected for fine-mapping , with the surrounding regions totaling 45 . 5 Mb accounting for overlaps ( 14 . 2 Mb for the densest fine-mapping regions ) . Eleven QT loci identified in previous GWA studies in populations of European and Asian ancestry were represented on the Metabochip ( Table 1 ) . The only published QT locus that is not represented on the Metabochip is an intergenic region on 13q14 reported by Marroni et al [19] , but not replicated by other published GWA studies of populations with similar ancestral backgrounds . SNPs reported in the literature but not genotyped on the Metabochip ( NOS1AP , rs10494366; NDRG4 , rs7188697 , rs37062 ) were represented by proxies , defined as SNPs in high LD ( r2≥0 . 90 ) with the index SNP using HapMap YRI data . Samples were genotyped at the Human Genetics Center of the University of Texas-Houston and the Translational Genomics Research Institute for ARIC and WHI , respectively , following each genotyping center's standard procedures . HapMap YRI ( Yoruba in Ibadan , Nigeria ) samples were also genotyped independently by each study to facilitate cross-study quality control . Genotypes were called separately for each study , albeit with a common protocol and common personnel , with GenomeStudio using the GenCall 2 . 0 algorithm . Because the Metabochip includes SNPs with much lower MAFs than are usually called with GenCall , SNPs were recalled using the GenoSNP genotype-calling algorithm [37] . SNPs with call rates <95% , Hardy-Weinberg equilibrium P<10−6 , >1 Mendelian error ( in 30 YRI trios ) , >2 replication errors , or >3 . 3% discordant calls in YRI across genotyping centers or against the HapMap database were considered quality control failures . Samples with call rates <0 . 95 or an inbreeding coefficient F>0 . 15 were excluded from further analysis [38] . Prior to analyses , related participants were identified using PLINK [39] by estimating identical-by-descent statistics for all pairs . When apparent first-degree relative pairs were identified , the member from each pair with the lower call rate was excluded from further analysis . Principal components of ancestry were determined using the Eigensoft software [40] , [41] and apparent ancestral outliers were excluded from further analysis . Briefly , n = 1 , 962 WHI participants who were genotyped on both the Affymetrix 6 . 0 and Metabochip genotyping platforms were used to infer Metabochip genotypes to the n = 8 , 421 population of WHI participants genotyped on the Affymetrix 6 . 0 array [36] . Before phasing and imputation , Affymetrix 6 . 0 SNPs with genotype call rates <90% , Hardy-Weinberg P-values<10−6 , or MAF<0 . 01 were removed . Participants with call rates <95% , those who demonstrated excess heterozygosity , were part of a first-degree relative pair , or who were identified as an ancestry outlier were excluded . This yielded a set of 987 , 749 SNPs for the 1 , 962 reference participants . Mean concordance rates for the 23 , 703 SNPs in common was 99 . 7% . Haplotypes were reconstructed using MaCH and were used as a reference to impute Metabochip data into the 6 , 459 WHI participants with only Affymetrix 6 . 0 data . Liu et al . , ( 2012 ) demonstrated the ability to impute 99 . 9% ( 97 . 5% , 83 . 6% , 52 . 0% , 20 . 5% ) of SNPs with MAF≥0 . 05 ( 0 . 03–0 . 05 , 0 . 01–0 . 03 , 0 . 005–0 . 01 , and 0 . 001–0 . 005 ) with average dosage r2 = 94 . 7% ( 92 . 1% , 89 . 0% , 83 . 1% , and 79 . 7% ) , respectively . For this analysis , all imputed SNPs with r2<0 . 95 were excluded . To interpret fine-mapping results , LD in our African American PAGE Metabochip sample was calculated in 500 Kb sliding windows using PLINK . In addition , Metabochip LD and frequency information ( but not individual-level information ) was provided by the Malmö Diet and Cancer Study on 2 , 143 control participants from a Swedish population [42] to facilitate LD and MAF comparisons to populations of European ancestry . HapMap CEU LD data were used for previously published GWA studies in European populations , as not all European index variants were represented on the Metabochip . Regional association plots use positions from NCBI build 36 . Recombination rates were estimated from HapMap phase II data . Linear regression models were used to study the association between QT and 6 , 670 SNPs from 11 regions fine-mapped for QT assuming an additive genetic model and including age , sex , study center , ancestry principal components , and heart rate as covariates . Study-specific association results were combined using an inverse variance meta-analysis approach as implemented in METAL [43] . For each QT locus , it is expected that SNPs associated with QT in African Americans will be correlated with the index SNP reported in Europeans . Therefore , we first identified and tested SNPs that are correlated ( r2≥0 . 20 ) with the index signals in Europeans using LD statistics estimated in the Malmö Diet and Cancer Study . In order to determine the appropriate multiple testing threshold for declaration of whether the previously identified signals were significantly associated with QT in African Americans , i . e . generalizability , we then estimated the number of tag SNPs needed to capture all common alleles ( r2≥0 . 80 ) using African American LD patterns . The multiple testing threshold for declaring generalization was αa = 0 . 05/415 , where 415 = the total number of tags identified using African American LD patterns . To identify significant population-specific SNPs influencing QT that were not correlated with the index signal in Europeans ( i . e . r2<0 . 20 , which was estimated in the Malmö Diet and Cancer Study ) , we used an efficient Monte Carlo approach that accounts for LD between SNPs at the previously identified QT loci ( αb = 1 . 37×10−5 ) [44] . Conditional analyses were then performed to determine the number of independent signals the population-specific SNPs represent . Specifically , analyses were repeated for each locus including the SNP with the smallest P – value as a covariate . This approach was performed adjusting for successively less significant SNPs until no SNPs with P –values lower than αb = 1 . 37×10−5 were identified . To facilitate comparability with previous reports examining the proportion of variance in QT explained by common SNPs , heart rate-corrected QT [45] was regressed on the six best markers in African Americans and the three population-specific variants assuming an additive genetic model and including age , sex , study center , ancestry principal components as covariates . For each of the nine QT SNPs ( i . e . the six best markers in African Americans and the three novel SNPs ) , we identified all SNPs in LD ( r2≥0 . 5 ) using the genotypes from the African American population described in this study . We refer to these SNP sets as Trait Associated SNP ( TAS ) blocks . We assigned each TAS to one or more of the functional annotation datasets listed in Table S5 . These datasets are not mutually exclusive . For example , a TAS can reside in both a candidate regulatory element ( dataset #7 ) and a CTCF binding site ( dataset #10 ) . For TASs that occur within predicted transcription factor binding sites ( datasets #3 and #8 ) , we calculated transcription factor binding affinity for each TAS allele using PWM-scan [46] , as described previously [47] . For TASs that occur within 3′ untranslated regions , we used the TargetScanS algorithm to determine whether they disrupt likely microRNA target sites ( dataset #5 ) . To define candidate non-promoter regulatory elements of greatest relevance to QT ( dataset #7 ) , we restricted the analysis of DNase I hypersensitive sites ( open chromatin loci ) to only those present in human cardiomyocytes . | The QT interval ( QT ) provides a measure of a ventricular action potential , and its prolongation is associated with sudden death and ventricular arrhythmias . Genome-wide association studies performed in European populations have identified common genetic variants that influence QT . However , it is unclear whether these variants are relevant in other populations , including African Americans . The increased genetic diversity in African Americans also provides opportunities to narrow association signals and identify candidates for functional evaluation . We therefore used data from 8 , 644 African Americans to further characterize previously identified QT loci . Of the fifteen known independent QT variants at the eleven previously identified QT loci , six were associated with QT in African Americans . We also identified three variants that were independent from previously reported signals and narrowed intervals flanking association signals using patterns of linkage disequilibrium . Finally , bioinformatic-based characterization pointed to candidates located outside protein coding regions . Our results underscore the utility of genetic studies in African ancestral populations to identify novel variants and narrow intervals surrounding established loci . These results suggest that known QT loci are important in African Americans and that further characterization of these loci in other populations may provide additional insights into the genetic and molecular mechanisms underlying QT . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"population",
"genetics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics",
"human",
"genetics"
] | 2012 | Fine-Mapping and Initial Characterization of QT Interval Loci in African Americans |
Cattle and other ruminants produce large quantities of methane ( ~110 million metric tonnes per annum ) , which is a potent greenhouse gas affecting global climate change . Methane ( CH4 ) is a natural by-product of gastro-enteric microbial fermentation of feedstuffs in the rumen and contributes to 6% of total CH4 emissions from anthropogenic-related sources . The extent to which the host genome and rumen microbiome influence CH4 emission is not yet well known . This study confirms individual variation in CH4 production was influenced by individual host ( cow ) genotype , as well as the host’s rumen microbiome composition . Abundance of a small proportion of bacteria and archaea taxa were influenced to a limited extent by the host’s genotype and certain taxa were associated with CH4 emissions . However , the cumulative effect of all bacteria and archaea on CH4 production was 13% , the host genetics ( heritability ) was 21% and the two are largely independent . This study demonstrates variation in CH4 emission is likely not modulated through cow genetic effects on the rumen microbiome . Therefore , the rumen microbiome and cow genome could be targeted independently , by breeding low methane-emitting cows and in parallel , by investigating possible strategies that target changes in the rumen microbiome to reduce CH4 emissions in the cattle industry .
Methane ( CH4 ) is a potent greenhouse gas ( GHG ) with a climate change potential ~32 times greater than carbon dioxide ( CO2 ) [1] and an atmospheric half-life of 12 years , which is substantially shorter than CO2 ( > 100 years ) [2] . Therefore , reducing CH4 emissions from anthropogenic-related sources has been identified as a key area for mitigating climate change with immediate effects[2 , 3] . Livestock accounts for 14 . 5% of anthropogenic-related GHG emissions and enteric CH4 emissions from ruminants accounts for 5 . 8%[3] . Furthermore , CH4 emissions from livestock is predicted to markedly increase due to an expected doubling in the global milk and meat demand by 2050[4] . Ruminants , the most widespread livestock species , can digest a wide variety of high fiber feedstuffs due to the distinct microbiome in their rumen . Methane is a natural by-product of gastro-enteric fermentation of high fiber plant biomass by microbial enzymatic activity in the rumen [5] . Bacteria , protozoa , and fungi in the rumen produce CO2 and hydrogen ( H2 ) , which are converted to CH4 , primarily by archaea known as methanogens . Approximately 99% of CH4 emitted from cattle is released in the breath by eructation and respiration[6] . The emission of CH4 is also a crucial pathway for maintaining H2 balance and ruminal pH , as the optimal conditions for anaerobic fermentation by the rumen microbial community is limited to a narrow range of partial pressure of H2 and pH [7] . Hydrogenase-expressing bacteria convert metabolic hydrogen from anaerobic fermentation into H2 which is then converted to CH4 via methanogenesis [7] . Furthermore , emitted CH4 has a caloric value and represents a 2–12% net loss of a cow’s gross energy intake[8 , 9] . Consequently , cattle and other ruminants with increased efficiency to digest high fiber feedstuffs but reduced CH4 production could in principal benefit the global climate and concurrently improve the profitability and sustainability of cattle production . Mitigation to decrease CH4 production by cattle to date has been largely unsuccessful , as the available measures are temporary and not cumulative . Large international research approaches target the rumen microbial communities through feed additives ( chemical or biological ) , feed formulations , and anti-methanogen vaccines[10] . However , rumen microbial species rapid adaptation to changes in the substrate results in resistance to treatments and CH4 production returns to pre-treatment levels[11] . Conversely , rumen transplantation studies ( transfaunation ) show that the rumen bacterial community recovered to near pre-transfaunation composition after a short period of time[12] . This indicated the existence of a degree of host influence on rumen microbial composition[12] . Host genotype in cattle was reported to explain inter-animal differences in CH4 production[13 , 14] and the rumen microbial community influenced CH4 production[15] . However , empirical evidence linking the host’s genetic influence over the rumen microbial community and CH4 production is rather limited[15] . A promising strategy is genetic selection for low CH4 emitting cows , as it is sustainable , persistent , and cumulative over subsequent generations . Whether the host influences the rumen microbial community , and consequently CH4 production , or the two interact to affect CH4 production is currently unknown . If reduced CH4 production in cows is a consequence of poor symbiosis with rumen microbes and thus fiber digestibility , there is a risk selection for reduced CH4 production will act against the very symbiosis which has aided ruminants and rumen microbes’ coexistence . Thus , the extent to which the rumen microbiome is under the host genetic influence needs elucidation . If host genetics impose a strong influence on rumen microbial composition , traits influenced by rumen microbes could be improved by using rumen microbial composition as indicator traits in selection . However , should host genetics impose a strong influence on rumen microbial composition and selection for CH4 production proceed without cognizance of rumen microbial composition , there is a risk of unfavorable correlated responses in rumen microbial composition . We hypothesized that: 1 ) the relative composition of the microbiome in the rumen is heritable i . e . controlled by host genome and 2 ) variation in methane emission from rumen is influenced by both the cow genome and rumen microbial content .
Methane concentration in the exhalation-breath of 750 lactating Holstein dairy cows from farmer herds in Denmark was measured individually during automated machine milking for one week . Within-week methane measurements had a high repeatability coefficient of 0 . 70 ± 0 . 02 ( estimate ± SE ) . Estimated average daily methane emission was 395 . 8 ± 63 . 5 g/d ( mean ± SD ) , which was consistent with reports from the literature[16] . Considerable variation in estimated CH4 emission among cows was observed . The top 10% methane emitting cows ( 519 . 28 ± 28 . 5 g/d ) had a 41% mean difference from the low 10% emitting cows ( 303 . 8 ± 11 . 9 g/d ) ( S1 Fig ) . Results from linear mixed model with pedigree records indicated methane emission was moderately heritable , 0 . 19 ± 0 . 09 ( heritability coefficient , h2 ± S . E ) , which was consistent with previous findings in lactating Holstein cows in Denmark[13] . We identified 3 , 894 bacterial operational taxonomic units ( OTUs , ≥ 97% identity ) and 189 archaeal OTUs , which were present in a minimum of 50% of the cow samples ( 50% threshold maximizes the variation in a binary trait i . e . presence or absence ) . Taxonomic classification revealed generic bacterial and archaeal composition . The predominant bacterial phylum found was Bacteroidetes 72 . 2% ± 6 . 5 ( mean ± SD ) , followed by Firmicutes ( 18 . 3% ± 5 . 6 ) and Tenericutes ( 2 . 8% ± 1 . 0 ) . Absconditabacteria , Spirochaetes , Fibrobacteres , and Proteobacteria each comprised less than 2% , and another 20 phyla constituted 1% of all sequence reads . The archaeal community was dominated by two families , Methanobacteriaceae and Methanomassiliicoccaceae ( 35% ± 22 . 1 ) of the orders Methanobacteriales ( 64 . 2% ± 22 . 2; mean ± SD ) and the recently proposed order Methanomassiliicoccales and class Thermoplasmata[17] , respectively . The remaining archaeal community was comprised of 10 families , which were low in abundance , cumulatively accounting for less than 1% of all archaeal sequence reads . OTU abundance and OTU abundance collapsed at genus and family levels were used as microbial phenotypes . The heritability thereof was estimated using a linear mixed model with pedigree records ( known as ‘animal models’ ) , which partitions total variance into additive genetic and environmental variance[18] . We calculated 95% confidence intervals for OTU h2 estimates and found for 6% of bacterial and 12% of archaeal OTUs , the estimates were significantly higher than zero ( P < 0 . 05 ) , ranging from 16–44% ( Fig 1 ) and 18–33% ( Fig 2 ) , respectively . Due to the high number of independent tests , we calculated false discovery rate ( FDR ) corrected P—values for h2 estimates with a FDR threshold of 15% ( S1 Table ) . Heritability of bacterial and archaeal abundance was further estimated at the genus level . In total eight bacterial genera out of 144 showed significant h2 estimates ranging from 0 . 17 to 0 . 25 ( Table 1 ) . Only a single archaeal genus , Methanobrevibacter , had a h2 estimate significantly different from zero ( 0 . 22 ± 0 . 09 ) . However , Methanosphaera and Methanomicrococcus might also be under host additive genetic control with heritability estimates approaching significance thresholds ( Table 1 ) . Associations between relative bacterial and archaeal OTUs , genera abundance , and host CH4 emissions were tested , while simultaneously controlling for environmental factors and familial structures common in livestock due to relatedness among study samples [19 , 20] . The OTU or genera log-transformed abundance present in > 50% of cows were fit as an explanatory variable in a linear mixed model for CH4 production . Numerous significant OTUs were detected but failed to pass the threshold for multiple testing ( FDR ≤ 0 . 15 ) ( Supplementary Table 1 ) . This was a hypothesis-generating analysis and not directed at specific hypothesis testing therefore we reported the significance and FDR corrected values ( S1 Table ) . Seven genera in total were detected , which exceeded the significance threshold at FDR of 15% . The -log10 P-values are plotted in Fig 3 . Traditionally , dimension reduction techniques such as principal coordinates analysis ( PCoA ) are used to summarize community composition differences between individuals ( Beta diversity ) into clusters , which are further examined for associated biological or explanatory variables . Differences in bacterial and archaeal community structures were estimated for the entire sample population at OTU level using the Bray-Curtis[21] dissimilarity metric ( PCoA , Fig 4A and 4B ) . Briefly , the Bray-Curtis dissimilarity is the sum of minimum counts of shared species in two animals divided by the sum of counts of all species in each animal , where 0 indicates the same composition and 1 indicates no shared composition . Analysis revealed clustering of cows into ‘ruminotypes’ for both bacterial and archaeal community composition , which both associated significantly with high and low CH4 emitters at opposing polar regions ( Mann-Whitney test , P < 0 . 001 ) but failed to cluster distinctly from the intermediate CH4 emitters . Analysis of community structures using ANOVA revealed bacterial PCo1 was partly explained by non-genetic factors: parity ( i . e . lactation number ) ( 3 . 6% ) , sequencing batch ( 2% ) and lactation stage ( 1% ) . A genetic analysis controlling for these factors showed PCo1 was likely heritable ( 0 . 20 ± 0 . 10 ) and thus influenced by the host additive genetics . Bacterial PCo2 was partly explained by the herd of origin ( < 1% ) and parity ( < 1% ) and was not heritable ( 0 . 02 ± 0 . 05 ) . Similar findings were observed for archaea , with the variation in PCo1 partly explained by herd ( < 1% ) , parity ( 19 . 9% ) , sequencing batch ( 5% ) and lactation stage ( < 1% ) . The genetic analysis controlling for these factors exhibited moderate heritability ( 0 . 39 ± 0 . 05 ) . Archaeal PCo2 variation was partly explained by herd ( < 1% ) and parity ( < 1% ) , which were likely not heritable ( 0 . 05 ± 0 . 05 ) . The relative proportion of variation in CH4 emissions due to rumen microbial composition and host additive genetic components was estimated individually and jointly using linear mixed models . Likelihood ratio tests revealed that fitting either random effect of rumen microbial composition or individual cow’s polygenic component fitted the data significantly better than the null model i . e . including only fixed effects ( P < 0 . 001 ) . The model fitting both random effects ( microbial composition and polygenic component ) was significantly better ( P < 0 . 001 ) than models including only one random effect . The proportion of variance in CH4 production explained by the microbiome , here defined as microbiability ( m2 ) , was calculated in analogy to the heritability ( h2 ) [22 , 23] . The contrast between the two intra-class correlation coefficients h2 and m2 with their respective standard errors for all models are depicted in Fig 5 . The m2 of CH4 emission estimated individually was 0 . 15 ± 0 . 08 ( estimate ± S . E ) and the h2 estimated individually was 0 . 19 ± 0 . 09 . Simultaneous estimates of both effects indicated slightly lower microbiability ( 0 . 13 ± 0 . 08 ) , whereas h2 exhibited a corresponding increase ( 0 . 21 ± 0 . 09 ) as compared to the preceding models fitting only one of the random effects . The combined microbial abundance and additive genetic effects were responsible for ~ 34% of the total phenotypic variation in CH4 emissions .
The results of this study show that estimated CH4 emissions from a dairy cow were partially under the influence of host ( cow’s ) additive genetics , which explained 19% of the total variation . Of the rumen bacterial OTUs , a modest ~ 6% were associated with host additive genetics exhibiting significant heritability estimates ( 16–44% ) ( Fig 1 ) . Similarly , only ~ 12% of archaeal OTU abundance was influenced by host additive genetics , with heritability estimates ranging from 18–33% ( Fig 2 ) . However , bacterial and archaeal heritability estimates failed to pass the threshold for multiple testing . Our test was conservative as a large number of taxa were analyzed with many OTUs having little or no influence by the host genome . Studies with larger sample sizes would give more reliable estimates of the heritabilities , especially for lower heritable OTUs . The h2 estimates observed in this study were consistent with findings of intestinal microbiota in mice[24 , 25] and humans[26 , 27] and confirm that the majority of variation in rumen microbial abundance is due to factors other than host additive genetics [28] . Interestingly , the patterns of h2 with phylogeny differed between the bacteria and the archaea ( Fig 1 and Fig 2 ) . Heritable OTUs were distributed throughout the bacterial microbiome whereas archaea showed increased heritability within the Thermoplasmatales . This highlights the value of collating phylogeny with heritability estimates to focus research into possible mechanisms which predispose differential relative abundance of certain taxa across genetically related cows . The method employed to sample rumen contents is high-throughput and less invasive than surgical procedures , making it better suited to sampling large numbers of cows under commercial farm conditions . Large sample size is critical in genetic evaluations . However , it is important to note that the floral rumen scoop is inserted into an undefined portion of the rumen and likely samples the liquid phase . Recognizing that rumen microbial communities differ between liquid , solid and epimural phases[29] , studies testing the repeatability and representativeness of sampling are needed . We utilized linear mixed model analysis to test for associations between bacterial and archaeal OTUs , genera and families with estimated CH4 emissions , while concurrently accounting for effects such as parity , lactation stage , herd of origin and familial structure from the pedigree . Several bacterial genera associated with CH4 emission were detected . Out of these , four were found either to be affected by methane inhibitors or related to H2 production and other methanogenesis substrates . Three were moderately heritable ( 0 . 17–0 . 25 ) ( S1 Table ) . One of the identified bacteria , Sporobacter , with a mean relative abundance of 0 . 01% ( Ruminococcaceae , Clostridiales , Firmicutes ) , belongs to a group with only a single cultured representative , Sporobacter termitidis , isolated from the intestine of wood-feeding termites ( Nasutitemes lujae ) , also known for producing large amounts of CH4 . However , when this isolate was co-cultured with an archaea species , Methanospirillum hungatei , CH4 was not produced . S . termitidis was found to generate acetate and methylsulfides , but not H2 or CO2 , therefore interspecies H2 transfer did not occur and facilitate CH4 production[30] . The recent discovery and proposed archaeal order Methanomassiliicoccales species found to utilize methylsulfides and H2 in methanogenesis[31] , provides a possible mechanism for methylsulfide producers to contribute to CH4 production when H2 producers are present . Methanomassilicoccales was prevalent in our samples ( mean relative abundance 35% ) ; therefore , Sporobacter could potentially be contributing to CH4 production via a similar pathway . We also detected Sphaerochaeta with a mean relative abundance of 0 . 01% , associated with estimated CH4 production . Genomes from cultured Sphaerochaeta isolates revealed acetate , formate , ethanol , H2 , and CO2 were potential fermentation end products[32] , many of which are methanogenic archaea substrates[33] . Furthermore , seed extracts from Perilla frutescens ( Lamiaceae ) , a medicinal herb , decreased CH4 production in vitro from rumen samples of lactating dairy cows and decreased Sphaerochaeta abundance[34] . Interestingly , Caro-Quintaro et al . [32] reported up to 40% of the genes from Spaerochaeta species were exchanged with members of Clostridiales ( Firmicutes ) and this inter-order-species horizontal gene transfer was most extensive in mesophilic anaerobic bacteria , such as the conditions found in termite and ruminant guts[35] . Here 16S rRNA gene sequencing is used as a proxy for metabolic activity but cannot account for inter-order-species horizontal gene transfer . Therefore , full metagenome sequence may have an advantage over the 16S rRNA gene to describe rumen microbial contents . One bacterial genus detected in the present study , which is positively associated with estimated CH4 production , is classified in the yet uncultured BS11 gut group of the Bacteroidales ( mean relative abundance 1 . 4% ) . The relative abundance of the BS11 group reportedly decreased concomitantly with CH4 production by dietary methanogenic inhibitors , such as P . frutescens seed extract , mentioned previously[34] , monesin and essential oil supplementation in dairy cattle[36 , 37] , and bromochloromethane in Japanese goats[38] . Thus , supporting our finding of a positive association between BS11 and CH4 production . Solden et al . [39] employed metagenomics sequencing and shotgun proteomics approaches to phylogenetically and metabolically resolve the BS11 gut group . They resolved two genera within the group and both exhibited multiple pathways to ferment hemicellulose , a capability previously unknown for BS11 . The resulting fermentation end products included acetate , butyrate , propionate , CO2 , H2[39] the latter two being methanogenesis substrates . Genes encoding ‘fucose sensing’ pathways were found for only one of the proposed BS11 genera , offering a possible mechanism for interaction between genes in the BS11 group and the host[15] . However , further studies are needed to elucidate the links between CH4 inhibitors , host genes and CH4 production . Due to the absence of cultured rumen bacteria isolates , an understanding of the metabolic function in many bacterial genera remains in its infancy . However , from the isolates discussed above , results suggested CH4 emissions depend on abundance of bacterial taxa that produce substrates for methanogenesis , such as H2 . Remarkably , associations between archaeal relative abundance and estimated CH4 production were not detected in the present study , despite the knowledge that archaea are directly responsible for CH4 production . A meta-transcriptome study in sheep found archaeal transcription pathways and not simply abundance , contributed to inter-animal differences in CH4 production[40] . This study was congruent with conclusions reached in two recent reviews , which examined results from dairy cattle and other ruminant studies employing 16S rRNA[41] and ‘meta-omics’ approaches[42] , where bacteria abundance produced and utilized H2 or stabilized pH , which affected CH4 emissions and feed efficiency and archaeal activity matched substrate availability . The combined effects of the bacterial and archaeal community structure ( beta diversity ) on estimated CH4 emissions were investigated by conducting PCoA on the archaeal and bacterial communities , which revealed 2–3 clusters for archaea ( Fig 4A ) and two clusters for bacteria ( Fig 4B ) . Beta diversity is a non-parametric distance measure used in microbiology and ecology to assess the differences between environments or samples ( in this case cows ) as opposed to alpha diversity which takes into account the diversity within cows . Clusters of a similar nature were first reported in intestinal bacterial community types in humans[43 , 44] , chimpanzees[45] , mice[46] and pigs[47] , referred to as “enterotypes” , and found associated with specific host phenotypes . This concept was extended to sheep rumen bacterial communities and referred to as “ruminotypes”[48] . The ruminotypes observed herein followed a continuous gradient and did not form discrete clusters , which is consistent with the latest findings in microbiome stratification . [49] . Importantly , we found that animal and farm factors like herd of origin , parity and lactation stage , as well as technical factors , i . e . sequencing batch , contributed to the observed variation and stratification in ruminotypes . Similar findings were reported in rumen bacterial richness at different lactation stages and over different parities[50] , suggesting later parities ( higher parity cows are older ) decreased bacterial richness and increased production[51] . We detected a moderate heritable genetic component acting along PCo1 axis , with h2 of 20% for bacterial and 39% for archaea , when controlling for lactation stage and parity , demonstrating the first evidence of host additive genetic influence on rumen bacterial and archaeal community structure ( beta diversity ) . All the above-mentioned factors contribute to microbiome structure and associations with host phenotypes . An association was detected between the highest and lowest CH4 emitters and bacterial and archaeal ruminotypes along PCo1 , however , ruminotype cluster memberships were not exclusive to high and low emitters . This suggested ruminal bacterial and archaeal community structure provided a modest contribution to CH4 emission . Kittlemann et al . [48] surveyed microbial community composition in multiple sheep cohorts with low and high CH4 yield ( methane emission per kg dry matter intake , CH4/DMI ) . A ruminotype “S” associated with low CH4 yield and enriched with Sharpea azabuensis was reported . A follow up study in sheep also found low CH4 yielding sheep to be associated with ruminotype “S” , enriched with Sharpea spp . It was hypothesized a smaller rumen size and higher turnover rate promoted faster growing bacteria , such as Sharpea , which favor hetero-fermentative growth on soluble sugars , resulting in lower H2 production and subsequently decreased CH4 formation by hydrogenotrophic methanogens[52] . Smuts et al . [53] reported passage rate ( and consequently turnover rate ) in sheep was heritable , indicating a possible mechanism for host genetics to influence ruminotypes . Methane emission phenotypes differed between the sheep and the present study . Kittlemann et al . [48] assessed the amount of CH4 production per unit of DMI but not CH4 production directly . DMI measurements are not currently recorded on dairy cattle under commercial farms due to the high costs and therefore , CH4 emissions in the present study could not be corrected for feed intake . In light of the differences in phenotype definitions and similarities in ruminotypes between studies , it would be of interest in future work to obtain DMI records on cows and test if the ruminotypes observed show an increased relationship with CH4 yield . The heritability estimates for PCo1 and PCo2 indicates these measures could potentially be used as indicator traits in genetic selection should they be highly correlated to a trait of interest , however PCo1 and PCo2 ( beta diversity ) does not account for the total rumen microbial variation within and between individuals . The method employed to measure CH4 production in the present study is high throughput and non-invasive , making it practically viable for measuring large numbers of animals under commercial farm conditions . However , the cost trade off of this method is that it makes use of milk yield and body weight in the estimation of CH4 production . Validation of this method with the ‘gold standard method’ ( climate respiration chambers ) has yielded highly correlated ( r = 0 . 8–0 . 89 ) and concordant ( concordance correlation coefficient = 0 . 84 ) results in dairy cattle [54 , 55] . However , the effects of body weight and milk yield on estimation of CH4 cannot be discounted and further research into the relationships between these variables and the rumen microbiome would be of value . In this study , we quantified the combined effects of all rumen bacterial and archaeal OTUs simultaneously on estimated host CH4 emissions using a microbial relationship matrix among cows . This is a parametric approach similar to assessing both alpha and beta diversity , as total rumen microbial variation within and between individuals is taken into account simultaneously . We expressed the combined effects as the variance ratio due to microbial composition to the total variance in estimated CH4 emissions ( m2 , microbiability ) , an analogy to h2 . Estimated CH4 emissions had 15% m2 , indicating the combined rumen bacteria and archaea abundance of dairy cattle was associated with a considerable amount of variation in estimated CH4 emissions among animals . Ross et al . [56] first proposed the generation of metagenomic relationship matrices in dairy cattle and reported a CH4 emission prediction accuracy of 0 . 47 , explaining 22% of the total variation in CH4 production [57] . However , Ross et al . [57] did not have sufficient data to estimate h2 or microbiability ( m2 ) in CH4 production . A study with 207 pigs employing 16S rRNA sequencing of gut microbes , found eight of the 49 bacterial genera to be heritable and estimated m2 and h2 for feed intake ( m2 = 0 . 16 , h2 = 0 . 11 ) , daily gain ( m2 = 0 . 28 , h2 = 0 . 42 ) and feed conversion ratio ( m2 = 0 . 21 , h2 = 0 . 19 ) [23] . Only daily gain had higher h2 compared with m2 . These findings suggest agreement with holobiont theory , where variation in the genome and microbiome can cause variation in some complex traits , on which artificial , natural selection and genetic drift can act [58 , 59] . However , the aforementioned study did not have adequate numbers of animals to estimate m2 and h2 simultaneously to assess the relative interactions between additive genetics and the microbiome . Thus , it was unable to assess if host additive genetics co-influences the microbiome and variation in phenotypes . In contrast , we estimated m2 and h2 concurrently to examine the shared information between the two effects . Microbiability of estimated CH4 production decreased by two percentage points to 13% and h2 exhibited a corresponding increase from 19 to 21% . This result indicated host genetic effects do interact with the microbial community composition but are not the primary mechanism for host genetic effects on estimated CH4 emissions . A possible explanation for the negligible amount of shared influence between the two relationship matrices might be the small percentage of heritable bacterial and archaeal OTUs . This implies that the rumen bacterial and archaeal communities affected estimated host CH4 emissions independently and host genetics influenced a small portion of these bacteria and archaea . The combined host additive genetics and rumen microbial community composition explained ~ 34% of the total variance in estimated CH4 emissions in dairy cattle . Thus , breeding for low CH4 production can be expected to result in limited correlated genetic responses to shape the rumen microbiome and breeding can likely proceed without taking cognizance of the rumen microbiome for this trait . However , larger studies estimating genetic correlations between rumen microbiota and CH4 emissions and better functional annotation of rumen microbiota are needed to confirm this . Microbiability estimates can be used as a tool for quantifying the cumulative effects of microbial abundance on phenotypes , e . g . complex diseases and quantitative traits . However , further research is required to elucidate the biological mechanisms shaping microbiability . For example , animal factors known to affect CH4 production and rumen microbial populations , such as passage rates or individual differences in feed intake might influence microbiability estimates . Human intestinal microbiome studies find that numerous disease phenotypes are associated with microbial richness , species abundance , and microbial community structure[60 , 61] . Subsequent work using stool consistency and opaque markers as proxies for colonic transit time found all three metrics and disease phenotypes are partially confounded with colonic transit time[62 , 63] . Similarly , in sheep studies , low CH4 yielding sheep are associated with lower retention time and smaller rumens[64] , relationships with specific rumen microbial clusters[48] and different bacterial and archaeal species[52] . Therefore , studies are needed to determine if microbial differences among subjects associated with phenotypic differences are causative or are consequences of unknown extraneous factors . It is also necessary to clarify the mechanisms which allow rumen microbes to be passed on to successive generations , to assess the efficacy of perturbations of the rumen microbiome such as probiotics and rumen transplants aimed at desired changes to the rumen microbiome and associated changes in phenotypes[65] . Regardless of the underlying biology , quantifying the relative contribution of rumen microbes and additive genetics to complex phenotypes helps characterize whether the host genome and microbiome are acting jointly as a holobiont and highlights the merits of targeting microorganisms to achieve a specific change in a phenotype or selective breeding . Furthermore , providing additional information , such as relative abundance of rumen fungi and protozoa , or ‘meta-omics’ , including meta-transcriptomics or meta-proteomics data can be readily adopted and incorporated into this methodology , offering insights into economically important livestock and disease traits in humans .
Methane production by dairy cows is not only influenced by factors such as feed intake and composition among others , but also the cow’s individual genetic composition and rumen microbial composition . Each cow’s additive genetic effects influence a modest amount of variation in the abundance of a small percentage of rumen bacterial and archaeal taxa , and thereby contribute to variation in rumen microbiome composition and function . We detected associations between CH4 emissions and rumen bacteria abundance , which are known to produce methanogenesis substrates , suggesting bacteria driven CH4 production pathways . Although we detected a heritable component to ruminotypes , the association to CH4 production was weak . Concurrently , host additive genetic effects and rumen microbes contributed to inter-animal differences in CH4 production , however negligible interaction was observed between microbiability and heritability . Consequently , cow additive genetic effects on CH4 emissions were largely unmodulated by cow additive genetic effects on rumen bacteria and archaea abundance . Strategies to reduce CH4 emissions in ruminants can be optimized by a multifaceted approach , for instance , selective breeding to unlock host’s genetic potential and strategies which may effect desired changes in the rumen microbiota like rumen transplantation , and probiotics .
Methane emissions from 750 lactating Holstein cows in five commercial herds were recorded using a portable Fourier Transform Infrared unit ( FTIR; Gasmet DX-4000 , Gasmet Technologies , Helsinki , Finland ) [13 , 66] and one research herd using a permanently installed non-dispersive infrared ( NDIR; Guardian NG/Gascard Edinburgh Instruments Ltd . , Livingston , UK ) [67] . Briefly , the FTIR and NDIR equipment were installed within the feed bins of automated milking systems ( AMS ) in each commercial herd with the FTIR for seven consecutive days and the NDIR were permanently placed in the research herd . The FTIR and NDIR device inlets were installed in the AMS feed bins and methane ( CH4 ) and carbon dioxide ( CO2 ) gas concentrations ( ppm ) sampled continuously every 5 s and 1 s , respectively[66 , 67] . Cows were milked individually in the AMS and milked on average ( 18 . 2 ± 3 . 4 ) times during the seven-day period , for durations ranging from five minutes to 12 . 2 minutes . Mean CH4 and CO2 gas concentrations were corrected for environmental factors , including diurnal variation and day to day differences using a linear mixed model following Difford et al . [67] to approximate daily averages . Measurement stability was assessed by model repeatability and used as data quality control . All herds practiced indoor feeding strategies with ad libitum access to feed and water . A total mixed ration ( TMR ) was provided , consisting primarily of rolled barley , corn silage , grass clover silage , rapeseed meal , soybean meal and up to 3 kg of concentrate supplement given during milking . Although all commercial herds employed a standardized TMR recipe , ingredient-specific differences among farms were expected to contribute to differences in TMR dietary values over herds . Weekly mean values for milk yield and body weight were combined with weekly gas concentrations , as described in Lassen et al . [66] and applied to predict cow heat production[68] . During each week of CH4 and CO2 recording at different herds , milk samples were collected to estimate milk fat and protein percentages . Cow fat and protein corrected milk yield ( FPCM ) was estimated following the national recording scheme ( RYK , Skejby , Denmark ) [69] . Methane production ( L/day ) was estimated using the CH4 to CO2 ratio and predicted CO2 emission[70] from the conversion of cow heat production units to CO2 production , following Madsen et al . [71] and then converted to ( g/d ) using CH4 density at standard temperature and pressure . Holstein cow pedigree records were traced in the Danish national database ( NAV , Skejby , Denmark ) as far back as 1926 to construct a pedigree-based relationship matrix for the quantitative genetic analysis . Immediately following the CH4 recording period , rumen content samples were drawn from individual cows by oral insertion of the probe “Flora Rumen Scoop” [72] . Approximately 40 mL of the liquid fraction containing particulate matter was drawn from the rumen using this method . Trained technicians conducted the sampling to ensure correct probe insertion into the rumen following a previously established protocol [72] , recognizing that the location of the flora rumen scoop may differ somewhat from sampling to sampling . The entire “Flora Rumen Scoop” was rinsed vigorously between animal sampling to minimize cross-contamination . Samples were labeled , immediately placed on ice , and transferred to the laboratory within two hours for further processing . Each 40 mL sample was mixed vigorously , a subsample of 1 . 2 mL rumen fluid was collected , and transferred to a 1 . 5 mL vial , then snap frozen in liquid nitrogen , before storing at -80°C , until shipped on dry ice to a commercial sequencing company ( GATC Biotech , Constance , Germany ) for analysis . DNA extraction , sequencing library construction and sequencing were conducted by GATC Biotech ( Constance , Germany ) . Rumen samples were defrosted at 4°C overnight and vortexed until homogenous . A representative sample ( 500 μl ) containing rumen liquid and solids was used for DNA isolation using the Qiagen QIAamp stool kit ( Valencia , United States of America ) following the manufacturer’s instructions , modified for the larger sample size[73] . Two primer sets were used to create 16S rRNA libraries , one set for all bacteria and one set for all archaea . Universal bacterial 16S rRNA gene primers ( covering the V1-V3 variable regions ) 27F: 5’-AGAGTTTGATCCTGGCTCAG-3’ and 534R: 5’-ATTACCGCGGCTGCTGG-3’ were used to generate the bacterial amplicon libraries ( expected amplicon size 508 bp ) [74] . Universal archaeal 16S rRNA gene primers ( covering the V4-V6 variable regions ) S-D-Arch-0519-a-S-15 5’-CAGCMGCCGCGGTAA-3’ and S-D-Arch-1041-a-A-18 5’-GGCCATGCACCWCCTCTC-3’ were used to generate the archaeal amplicon libraries ( expected amplicon size 542 bp ) [75] . Following protocols standardized by GATC Biotech , PCR amplifications were conducted with GoTaq Green polymerase ( Promega , Madison , USA ) with 30 PCR cycles and a 60°C annealing temperature for the archaeal amplicon libraries and 25 PCR cycles with a 60°C annealing temperature for the bacterial amplicon libraries . The 16S rRNA amplicons were purified using the Axyprep Fragment Select bead purification system ( Axygen Biosciences , New York , USA ) , according to the manufacturer’s instructions . The size and purity of the PCR product was verified on a Fragment Analyzer using a High Sensitivity NGS Fragment Analysis Kit ( Advanced Analytical Technologies , Ankeny , USA ) . Multiplex indices and Illumina overhang adapters were added to both amplicon libraries in a second PCR amplification round ( six cycles ) , followed by Fragment Analyzer analysis to confirm the correct size of the amplicons ( Advanced Analytical Technologies , Ankeny , USA ) . Ninety-six libraries were pooled in equimolar concentrations and sequenced with an Illumina sequencing instrument using the 300 bp paired-end read mode , according to the manufacturer’s specifications . Approximately half the samples were run using the illumina MiSeq platform and half with the HiSeq platform . The 300 bp paired end protocol was adapted to HiSeq by GATC Biotech . The specific samples entered into sequencing batches within each sequencing platform were recorded for subsequent significance testing to examine possible differences between sequencing batches and sequencing platforms in statistical analyses . Bacterial and archaeal sequence reads underwent quality control , processing and were clustered into operational taxonomic units ( OTUs ) using the LotuS pipeline[76] with the following options: Sequence truncation length and minimum sequence length after barcode and primer removal was 230 bp . Minimum average sequence quality score was 27 , the maximum number of ambiguous bases was 0 , maximum homonucleotide run was set to 8 . Sequences were filtered away if any of the 50 bp segments in a sequence had average scores below 25 or if the expected number of errors exceeded 2 . 5 in the binomial error model . The low-quality sequence ends were trimmed by applying a sliding window quality filter with a width of 20 bp and a minimum average quality score within the window of 25 . Sequences were truncated if the probabilistic accumulated error exceeded 0 . 75 . The reads were de-replicated and sequences with a minimum of 10 replicates were retained for OTU clustering within the Lotus pipeline . Sequence pairs were merged with Flash[77] and clustered into OTUs based on sequence similarity ( 97% ) with UPARSE[78] and chimeric sequences removed with UCHIME reference-based chimera detection[79] . Representative sequences from each OTU were aligned with ClustalO[80] and a phylogenetic tree built with FastTree2[81] . Representative sequences , the OTU table , and phylogenetic trees were transferred to QIIME ( version 1 . 9 . 0 ) [82] , where further analyses were performed . Taxonomy was assigned to each OTU using the RDP classifier with a confidence level of 0 . 8[83] using greengenes ( gg_13_8_otus ) as the reference database . Unclassified OTUs and OTUs classified to non-target kingdoms were filtered from the OTU tables , i . e . only OTUs classified as k_Bacteria were maintained for the bacterial primer set and similarly OTUs classified as k_Archaea maintained for the archaeal primer set . Finally , samples with < 50 , 000 sequences were removed and OTUs containing < 10 sequences were filtered out of the OTU table . All handling of animals were conducted according to 'Metagenomics in Dairy Cows' protocol . The protocol and study were approved by The Animal Experiments Inspectorate , Danish Veterinary and Food Administration , Ministry of Environment and Food of Denmark ( Approval number 2016-15-0201-00959 ) . | Methane is a potent greenhouse gas and ruminant livestock contribute a substantial amount of total methane from human activities . Variation between cows’ methane production has been found partly due to their genetics ( heritable ) , making genetic selection a promising strategy for breeding low methane emitting cows . We hypothesized that the total methane production by a cow is affected by rumen microbes which are directly responsible for production of methane , as well as the cows’ own genetics and their interaction . We sampled the rumen contents of 750 dairy cows and found the relative abundance of some bacteria and archaea to be heritable and associated with methane production , but the majority of variation in relative abundance of rumen bacteria and archaea is due to non-genetic factors . We compared the amount of variation in methane production associated with host genetics as well as rumen bacteria and archaea and found the host genetics to explain 21% and rumen microbes 13% . Importantly , the two were largely independent of each other , so breeding for low methane emitting cows is unlikely to result in unfavorable changes in the rumen microbiome . However , further functional annotation of rumen microbiota is needed to confirm this . Strategies that target each source of variation can be conducted in parallel to optimize reduction in methane production from dairy cows . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusions",
"Materials",
"and",
"methods"
] | [
"taxonomy",
"livestock",
"ecology",
"and",
"environmental",
"sciences",
"archaeal",
"taxonomy",
"microbiome",
"chemical",
"compounds",
"community",
"structure",
"atmospheric",
"science",
"microbiology",
"data",
"management",
"archaea",
"bacteria",
"methane",
"microbial",
... | 2018 | Host genetics and the rumen microbiome jointly associate with methane emissions in dairy cows |
Many bacterial pathogens utilize a type III secretion system to deliver multiple effector proteins into host cells . Here we found that the type III effectors , NleE from enteropathogenic E . coli ( EPEC ) and OspZ from Shigella , blocked translocation of the p65 subunit of the transcription factor , NF-κB , to the host cell nucleus . NF-κB inhibition by NleE was associated with decreased IL-8 expression in EPEC-infected intestinal epithelial cells . Ectopically expressed NleE also blocked nuclear translocation of p65 and c-Rel , but not p50 or STAT1/2 . NleE homologues from other attaching and effacing pathogens as well OspZ from Shigella flexneri 6 and Shigella boydii , also inhibited NF-κB activation and p65 nuclear import; however , a truncated form of OspZ from S . flexneri 2a that carries a 36 amino acid deletion at the C-terminus had no inhibitory activity . We determined that the C-termini of NleE and full length OspZ were functionally interchangeable and identified a six amino acid motif , IDSY ( M/I ) K , that was important for both NleE- and OspZ-mediated inhibition of NF-κB activity . We also established that NleB , encoded directly upstream from NleE , suppressed NF-κB activation . Whereas NleE inhibited both TNFα and IL-1β stimulated p65 nuclear translocation and IκB degradation , NleB inhibited the TNFα pathway only . Neither NleE nor NleB inhibited AP-1 activation , suggesting that the modulatory activity of the effectors was specific for NF-κB signaling . Overall our data show that EPEC and Shigella have evolved similar T3SS-dependent means to manipulate host inflammatory pathways by interfering with the activation of selected host transcriptional regulators .
Many bacterial pathogens have the ability to “inject” virulence effector proteins into the host cell using a type III secretion system ( T3SS ) . The effector proteins perform a variety of functions that allow the pathogen to persist in the host and cause disease [1] . Enteropathogenic Escherichia coli ( EPEC ) and enterohemorrhagic E . coli ( EHEC ) deliver T3SS effector proteins to the intestinal epithelium that mediate attaching and effacing lesion ( A/E ) lesion formation . A/E lesions are characterized by intimate bacterial attachment , effacement of the brush border microvilli and actin pedestal formation [2] . T3SS effectors from other pathogens such as Salmonella and Shigella have various roles in invasion , intracellular survival and the inhibition of innate immune responses through targeting host inflammatory signaling pathways [1] . Many of the T3SS effectors belong to conserved protein families that are found in a range of bacterial pathogens of plants and animals . For example , the OspF family of T3SS effectors from Shigella , Salmonella and Pseudomonas exhibit phosphothreonine lyase activity and induce irreversible dephosphorylation of mitogen-activated protein kinases ( MAPKs ) in the host cell nucleus [3] , [4] , [5] . In Shigella , this leads to gene-specific repression of a subset of NF-κB regulated genes , including IL8 [3] . Given the remarkable specificity of their biochemical function , the discovery of the mechanism of action of T3SS effectors remains an important step towards understanding the pathogenesis of many bacterial infections . The activation of gene expression during inflammation is tightly regulated by transcription factors such as NF-κB . The NF-κB/Rel family comprises five members that share an N-terminal Rel homology domain that mediates DNA binding , dimerization and nuclear translocation [6] . The p65 , c-Rel and RelB NF-κB subunits have an additional C-terminal transactivation domain , which strongly activates transcription from NF-κB-binding sites in target genes . The p50 and p52 subunits lack the transactivation domain but still bind to NF-κB consensus sites and act as transcriptional repressors [6] . The most abundant form of NF-κB in mammalian tissues is a p65/p50 dimer that activates the expression of multiple cytokine genes in response to inflammatory signals . In resting cells , NF-κB subunits are held in an inactive form in the cytoplasm by binding IκB proteins . Activation of NF-κB signaling stimulates the phosphorylation and proteosomal degradation of IκB , whereupon the NF-κB dimer is transported into the nucleus through the nuclear pore complex [6] . The canonical NF-κB pathway is stimulated by a range cell surface receptors such as the TNF receptor , IL-1 receptor , Toll-like receptors and T-cell receptor . Although the upstream components of these pathways vary , they converge at the point of IκB kinase complex ( IKK ) -mediated phosphorylation of and subsequent degradation of IκB [7] . Enteropathogenic Escherichia coli ( EPEC ) and enterohemorrhagic E . coli ( EHEC ) utilize a type III secretion system ( T3SS ) to deliver effector proteins to the intestinal epithelium that induce actin pedestal formation [2] . Multiple additional effectors are transported into the host cell where their targets and effects on host cell biology remain largely uncharacterized [8] . NleE is a highly conserved 27 kDa T3SS effector protein of A/E pathogens encoded in an operon with the 38 kDa effector , NleB . The NleE homologue in the invasive pathogen , Shigella , is called OspZ [9] , [10] . While investigating the effect of EPEC infection on NF-κB activation , we observed that wild type EPEC prevented translocation of the p65 subunit of NF-κB to the host cell nucleus , whereas an nleE mutant was defective for this activity . Here we report that NleE inhibits p65 nuclear translocation , thereby reducing the IL-8 response during bacterial infection , and that OspZ shares this activity . In addition , we show that NleB suppresses NF-κB activation but appears to act in distinct manner to NleE .
Recent work has shown that EPEC and EHEC infection inhibits inflammatory cytokine production and NF-κB activation [11] , [12] , [13] . Previously , we found that translocated NleE localised to the host cell nucleus and we postulated that NleE had a role in subverting innate immune signaling [10] . Here we investigated the effect of NleE on NF-κB activation during EPEC infection . As actin accumulation beneath adherent EPEC depends on successful translocation of the T3SS effector , Tir [2] , we used the fluorescent actin staining ( FAS ) test as a general marker for the translocation of T3SS effectors . HeLa cells were infected with wild type EPEC E2348/69 , a T3SS ( escF ) mutant , an nleE deletion mutant of EPEC or an nleE mutant complemented with full length nleE . Cell monolayers were either infected for 4 h and left unstimulated or infected for 90 min and stimulated with tumour necrosis factor α ( TNFα ) or interleukin-1β ( IL-1β ) for 30 min . Nuclear translocation of the p65 NF-κB subunit was visualised by immunofluorescence microscopy of FAS-positive cells for EPEC E2348/69 , ΔnleE and ΔnleE ( pNleE ) ( Fig . 1A ) and cells with adherent bacteria for ΔescF . In unstimulated cells , there was no significant difference in p65 nuclear exclusion between wild-type infected cells and the escF mutant ( Fig . 1B ) . In contrast , in cells stimulated with TNFα or IL-1β , wild type EPEC E2348/69 inhibited p65 transport to the nucleus , whereas the escF mutant had little inhibitory effect on p65 nuclear translocation ( Fig . 1B ) . The nleE mutant also showed greatly reduced inhibition of p65 nuclear transport in response to TNFα or IL-1β compared to wild type EPEC E2348/69 which was restored upon complementation of the nleE mutant with a copy of full length nleE . Similar results were obtained in response to IL-1β in Caco-2 intestinal epithelial cells ( Fig . S1 ) . Caco-2 intestinal epithelial cells were then utilised to determine if the inhibition of p65 translocation resulted in the suppression of IL-8 production . Caco-2 cells were incubated with wild type EPEC E2348/69 , a T3SS ( espB ) mutant , the nleE mutant or the nleE mutant complemented with nleE . Cells were then stimulated with IL-1β [14] . Compared to infection with the espB mutant , wild type EPEC inhibited IL-8 production from Caco-2 cells . The nleE mutant showed a diminished capacity to inhibit IL-8 production , which was complemented to wild type levels upon reintroduction of full length nleE ( Fig . 2A ) . A similar trend was observed in CaCo-2 cells left unstimulated , although the differences were not as great as in IL-1β-stimulated cells ( Fig . 2B ) . Real time PCR analysis of IL8 from Caco-2 cells infected with derivatives of EPEC and stimulated with IL-1β , showed that levels of IL8 mRNA were suppressed by NleE expression ( Fig . 2C ) . To determine if NleE was sufficient for the inhibition of p65 nuclear translocation , we expressed GFP-NleE or GFP transiently in HeLa cells . Upon stimulation with TNFα , the NF-κB p65 subunit was excluded from the nucleus in the presence of ectopically expressed GFP-NleE but not GFP alone ( Fig . 3A ) . To ensure that this effect was not an artefact arising from the over expression of GFP-NleE , we examined a range of transfected cells exhibiting low , moderate and high GFP expression . Even in cells exhibiting low levels of GFP-NleE , p65 was excluded from the nucleus upon stimulation with TNFα ( Fig . 3A and data not shown ) . We also investigated the effect of NleE on nuclear localization of other NF-κB proteins , c-Rel and p50 . Similar to p65 , NleE blocked nuclear translocation of c-Rel in response to TNFα ( Fig . 3B ) , however p50 nuclear localization was unaffected by the presence of NleE ( Fig . 3C ) . A dual-luciferase reporter system measuring the activation of NF-κB-dependent transcription confirmed the absence of NF-κB p65 nuclear activity in GFP-NleE transfected cells stimulated with TNFα ( Fig . 3D ) . To determine if NleE-mediated inhibition of signaling affected other transcription factors , we tested the ability of NleE to inhibit nuclear translocation and activation of STAT1 and STAT2 . Following stimulation with interferon α , both STAT1 and STAT2 were translocated to the nucleus in the presence of GFP or GFP-NleE ( Fig . 4A and B ) . NleE also had no impact on STAT1/2 activation using a ISRE-Luc luciferase reporter ( Fig . 4C ) [15] . This suggests NleE acts on a subset of signaling pathways that includes NF-κB but not STAT1/2 . To determine if the function of NleE and OspZ was conserved across A/E pathogens and Shigella , GFP-NleE and GFP-OspZ fusions generated from enterohemorrhagic E . coli O157:H7 , Citrobacter rodentium , Shigella boydii and Shigella flexneri were expressed by transfection in HeLa cells . NleE from EHEC O157:H7 and C . rodentium , and full length OspZ from S . boydii and S . flexneri serogroup 6 inhibited NF-κB activation and p65 nuclear translocation ( Fig . 5A and B ) . In contrast , OspZ from S . flexneri serogroup 2a which carries a 36 amino acid truncation at the C-terminus ( Fig . S2 ) , had no impact on NF-κB activation and did not block p65 nuclear translocation . Further screening of three clinical isolates of S . flexneri 2a revealed that all strains encoded a truncated OspZ protein . Similar to the truncated form of OspZ from S . flexneri 2a , a GFP-NleE truncation lacking the C-terminal 36 amino acid residues , GFP-NleE1-188 was unable to prevent NF-κB activation . However , the C-terminal region was not sufficient for this antagonism , as GFP-NleE188-224 did not inhibit NF-κB activation ( Fig . 5C ) . Interestingly , in contrast to the other full length GFP-NleE/OspZ fusion proteins , GFP-OspZ from S . flexneri 6 and S . flexneri 2a was largely excluded from the nucleus ( Fig . 5B ) . Although the molecular basis of this is unknown , it may indicate that the mechanism of action of NleE/OspZ is in the cytoplasm of the cell since GFP-OspZSF6 inhibited NF-κB activation to the same degree as GFP-NleE ( Fig . 5A ) . To examine further the C-terminal region of NleE and OspZ , we performed a deletion analysis of NleE . Whereas truncated NleE1-214 , inhibited NF-κB activation to the same degree as full length NleE , truncated NleE1-208 was inactive ( Fig . 6A ) . This suggested that the region between amino acid residues 208 and 214 , with the motif IDSYMK was critical for NleE function . We then constructed a deleted form of NleE lacking amino acids I209DSYMK214 which was unable to inhibit NF-κB activation ( Fig . 6A ) as was a corresponding deleted form of OspZ from S . flexneri 6 lacking amino acids I209DSYIK214 ( Fig . 6B ) . In fact GFP-OspZΔIDSYMK appeared to have a pro-inflammatory effect even in unstimulated cells ( Fig . 6B ) . The C-terminal 50 amino acids of NleE and OspZ are strongly predicted by Jpred [16] to form an alpha-helical region , therefore it was possible that the IDSY ( M/I ) K deletion had disrupted protein secondary structure ( Fig . 6C ) . To account for this possibility and preserve the predicted alpha helix , we changed all six amino acids to alanine to generate GFP-NleE6A . NleE6A had the same alpha-helical prediction by Jpred as NleE ( Fig . 6C ) . Similar to GFP-NleEΔIDSYMK , GFP-NleE6A was unable to inhibit NF-κB activation ( Fig . 6D ) , and NleE6A delivered by the T3SS during infection was also unable to inhibit p65 nuclear translocation ( Fig . 1B ) . Further mutation of individual amino acids within the IDSYMK motif of NleE to alanine did not make a significant difference to NF-κB inhibition compared to the fulllength GFP-NleE fusion ( Fig . 6D ) . Expression of all GFP-NleE derivatives was tested by immunoblot using anti-GFP antibodies to ensure that differences in activity were not due to differences in the levels of GFP fusion proteins ( Fig . S3 ) . To determine if variations in amino acid sequence at the C-termini of NleE and full length OspZ had any functional significance ( Fig . 6C ) , we performed a domain swap by exchanging the last 40 amino acids of NleE with the last 46 amino acids of OspZ and vice versa . Both chimeric forms of NleE and OspZ ( NleE-OspZcterm and OspZ-NleEcterm ) were fully functional ( Fig . 6E ) and they retained their native localization ( data not shown ) . Therefore these regions were functionally interchangeable , suggesting that NleE and OspZ use the same molecular mechanism to inhibit NF-κB activation . To compare the activity of NleE with another T3SS effector , NleH , reported to interfere with NF- κB activation [17] , we generated GFP-NleH1 and GFP-NleH2 fusions from EPEC E2348/69 and tested these for their ability to inhibit NF-κB activity following stimulation with TNFα . In this system , NleE showed greater inhibition of NF-κB activation than either NleH1 or NleH2 ( Fig . 7A ) . To test for possible non-specific effects on NF-κB activation of effector over-expression by transfection , we also tested the effect of NleD and NleB on NF-κB activation following stimulation with TNFα . While NleD had no effect on luciferase induction in response to TNFα , NleB inhibited NF-κB activation as effectively as NleE ( Fig . 7A ) . NleB is encoded directly upstream from NleE and this organization is highly conserved among A/E pathogens [18] . Fluorescence microscopy of GFP-NleB transfected cells stimulated with TNFα confirmed that GFP-NleB inhibited p65 translocation ( Fig . 7B ) . NleE inhibited NF-κB activation in response to both TNFα and IL-1β so we tested the ability of NleB to inhibit IL-1β signaling . Whereas , GFP-NleE was effective against both TNFα and IL-1β stimulation , GFP-NleB had no effect on NF-κB activation stimulated by IL-1β ( Fig . 7C ) . This suggested that the two effectors act at different points in the NF-κB signaling cascade . To examine the effect of NleE and NleB on other signaling pathways , we used an AP-1 reporter to monitor JNK/MAPK signaling . Neither NleE nor NleB inhibited AP-1 activation by phorbol 12-myristate 13-acetate ( PMA ) ( Fig . 7D ) , and NleB also had no effect on STAT1/2 activation ( data not shown ) . This suggests that the effectors target only a subset of signaling pathways involving NF-κB . To ensure that NleE and NleB translocated by the LEE-encoded T3SS conferred the same phenotype as ectopic expression of the effectors by transfection , we infected HeLa cells with wild type EPEC E2348/69 and a double island mutant that lacked the genomic regions , PP4 and IE6 [19] . The double island mutant was used to eliminate genes encoding NleE and NleB in IE6 as well as NleB2 , a close homologue of NleB , encoded in PP4 [19] . The ΔPP4/IE6 island mutant was complemented with pNleE or pNleB to examine the contribution of each effector to the inhibition of p65 translocation . In unstimulated cells , there was no difference in p65 nuclear translocation between uninfected cells and those infected with wild type EPEC E2348/69 or the T3SS mutants ( escN and escF ) ( Fig . 8A and B and Fig . 1B ) . This suggested that , over a 4 h infection , bacterial products such as flagellin and lipopolysaccharide were not sufficient to stimulate signaling . In contrast , the ΔPP4/IE6 island mutant induced substantial p65 nuclear translocation ( Fig . 8B ) , which may indicate that the translocation and/or biochemical function of some effectors is proinflammatory . In infected cells , both NleB and NleE injected by the T3SS had the capacity to inhibit p65 nuclear translocation in response to TNFα but only NleE was effective in response to IL-1β ( Fig . 8A and B ) . Since EPEC infection has been reported to inhibit IκB degradation [12] , a critical event in the activation of NF-κB , the effect of NleE and NleB on IκB degradation was examined here in TNFα and IL-1β stimulated cells . Ectopically expressed GFP-NleE and GFP-OspZ inhibited IκB degradation in response to both stimuli whereas GFP-NleB inhibited IκB degradation in response to TNFα only ( Fig . 9A ) . GFP-NleE6A lacked the ability to inhibit IκB degradation as did GFP-NleEΔIDSYMK and GFP-OspZΔIDSYMK ( Fig . 9B ) . In addition , we tested whether NleB and NleE delivered by the T3SS had the same effect as ectopically expressed protein on IκB degradation stimulated by TNFα and IL-1β . Wild type EPEC E2348/69 , ΔnleE ( pNleE ) , ΔPP4/IE6 ( pNleE ) or ΔPP4/IE6 ( pNleB ) inhibited IκB degradation in HeLa cells stimulated with TNFα but IκB degradation was not inhibited in cells infected with ΔPP4/IE6 ( pNleB ) and stimulated with IL-1β . This suggests that NleE and NleB act in different ways to interfere with NF-κB signaling ( Fig . 9C ) .
The activation of NF-κB signaling is a critical host response to infection . In this study , we found that the T3SS effector NleE from EPEC prevented nuclear translocation of the p65 NF-κB subunit , leading to diminished IL8 expression and a compromised IL-8 response . The inhibition of p65 nuclear translocation occurred when NleE was expressed ectopically or when NleE was delivered through the T3SS by infection . We also observed that NleE inhibited nuclear translocation of c-Rel but not nuclear import of activated p50 , STAT1 or STAT2 . Both p65 and c-Rel are structurally similar and contain transcriptional activation domains that initiate gene expression [6] . In contrast , p50 lacks a transcriptional activation domain , so that p50/p50 homodimers act as transcriptional repressors . Thus , NleE appears to obstruct nuclear translocation of Rel family transcriptional activators while allowing nuclear import of a transcriptional repressor , resulting in the suppression of IL8 expression . The selectivity of NleE for p65 and c-Rel is not unprecedented as lack of nuclear translocation of p65 and c-Rel but not p50 was recently reported for oestrogen-induced inhibition of NF-κB activation , although the mechanism is unknown [20] . NleE is one of the conserved core type III effectors of A/E pathogens [19] . We observed that ectopically expressed NleE from EHEC O157:H7 and the murine pathogen , C . rodentium also inhibited NF-κB activation and p65 translocation . A close homologue of NleE , OspZ , is found in Shigella , however in S . flexneri 2a , OspZ is truncated to the length of NleE1-188 [10] . Both the truncated form of OspZ from S . flexneri 2a and a C-terminal 36 amino acid deletion mutant of NleE were inactive , suggesting that the C-terminus was critical for the immunosuppressive function of NleE . However , this region was not sufficient for inhibition of p65 nuclear translocation as a region encompassing the last 36 amino acids of NleE alone was unable to prevent NF-κB activation . A domain swap between NleE and OspZ of the last ∼40 amino acids showed that these regions were functionally interchangeable and we identified a 6-amino acid motif , IDSY ( M/I ) K , that was critical for both NleE and OspZ function . Although A/E pathogens stimulate an inflammatory response in vivo and proteins such as flagellin are recognised by TLR5 [21] , [22] , previous work has suggested that A/E pathogens modulate that inflammatory response by inhibiting p65 nuclear translocation as well as IκB degradation [11] , [12] , [23] . Here , we found that NleE inhibited nuclear translocation of p65 by preventing IκB degradation in response to TNFα and IL-1β . In contrast , we found that NleB inhibited IκB degradation in response to TNFα only . Since TNFα and IL-1β signaling converges at the point of IKK phosphorylation ( Fig . 9D ) [24] , NleE may act on IKK or IκB itself to prevent IκB degradation . TAK1 or other MAPK may also have involvement in IKK phosphorylation leading to JNK activation [24] , however , JNK signaling , represented here by the AP-1 reporter , was not affected by NleE and so we predict that NleE interferes with IKK or IκB function directly . Indeed while this work was under review , Nadler et al reported that NleE inhibits IKK phosphorylation [25] . The authors also proposed that NleB assists the inhibition of IκB degradation by NleE [25] . Here we hypothesize that NleB acts upstream of IKK in the TNFα pathway since NleB did not inhibit IκB degradation in response to IL-1β ( Fig . 9D ) . We therefore propose a model where NleE and NleB act at different points in the NF-κB signaling pathway and each plays a distinct role in the inhibition of p65 nuclear translocation . In the TNFα pathway , NleE and NleB have overlapping and somewhat redundant inhibitory roles as complementation of the ΔPP4/IE6 double island mutant with either NleE or NleB was sufficient to block p65 nuclear translocation . In the IL-1β pathway however , NleB was not able to compensate for the lack of NleE . Although we believe that NleB acts independently of NleE , these results do not exclude the possibility that in IL-1β stimulated cells , NleB acts in concert with NleE [25] . The fact that both NleE and NleB inhibit NF-κB activation raises the possibility that more effectors contribute to the suppression of innate signaling pathways . Although compromised compared to wild type EPEC , the nleE mutant showed significantly greater inhibition of IL-8 secretion than an T3SS mutant , which lacks the ability to translocate all T3SS effectors . While one of the additional effectors inhibiting p65 translocation is clearly NleB , a close homologue , NleB2 may also have anti-inflammatory activity and perhaps other effectors in the genomic islands , PP4 and IE6 . In addition , NleH1 and NleH2 were recently reported to interfere with the activation of NF-κB by binding ribosomal protein S3 ( RPS3 ) , a co-factor of nuclear NF-κB complexes , and sequestering it in the cytoplasm [17] . We also found that ectopically expressed NleH1 and NleH2 inhibited NF-κB activation , but not to the same degree as NleE and NleB . Together these anti-inflammatory effectors may balance the action of other effectors that through their biochemical activity stimulate inflammatory signaling , as suggested by the ΔPP4/IE6 double island mutant , which showed increased p65 nuclear translocation in uninfected cells compared to a T3SS mutant . Therefore , despite the fact that EPEC and Shigella infection ultimately induces gut inflammation , we propose that NleE/OspZ and NleB contribute to pathogenesis by inhibiting an initial host inflammatory response to allow the bacteria to persist in the early stages of infection . A multi-effector attack on NF-κB signaling occurs during Shigella infection , which modulates NF-κB activation through the effectors OspG and OspF [3] , [26] . Our studies suggest that Shigella strains carrying full length OspZ have evolved a further distinct mechanism to modulate NF-κB signaling . This makes the absence of a functional OspZ protein in S . flexneri 2a curious and may also explain previous findings that OspZ from S . flexneri 2a potentially enhanced inflammation by inducing polymorphonucleocyte migration across a polarized epithelium [10] . The truncation rendering OspZ inactive in S . flexneri was serotype specific however , as S . flexneri 6 encoded functional full length OspZ , similar to S . boydii . In this study , we have ascribed a function to the NleE/OspZ family of T3SS effectors shared by attaching and effacing pathogens and Shigella as well as the EPEC effector , NleB . Despite the remarkably different infection strategies of these two groups of pathogens , they appear to have a mutual need to inhibit the host inflammatory response during infection . NleE , NleB and OspZ are the latest T3SS effectors to target NF-κB activation and the expression of NF-κB-dependent genes . Neither NleE nor NleB inhibited STAT1/2 or AP-1 signaling , suggesting that the proteins target the NF-κB pathway specifically . The ongoing identification of T3SS effectors that act on this and other inflammatory pathways will continue to provide insight into the molecular mechanisms by which bacterial pathogens inhibit immune signaling and establish infection .
The bacterial strains and plasmids used in this study are listed in Table 1 . The construction of vectors and culturing of bacterial strains for infection is described in detail in the supplementary methods ( Protocol S1 ) . EPEC strains were used to infect HeLa cells for 4 h without exogenous stimulation or for 90 min after which the media was replaced with DMEM supplemented with 20 ng/ml TNFα or 10 ng/ml IL-1β ( eBioscience , San Diego , CA ) and the infection was continued for a further 30 min . For Caco-2 cells , EPEC infection continued for 4 h after which the cells were washed , treated with 100 µg/ml gentamicin for 2 h . For mRNA analysis monolayers were incubated for 3 h in media supplemented with 50 ug/ml gentamicin with or without 5 ng/ml IL-1β . For analysis of IL-8 secretion , monolayers were infected for 4 h and incubated for 24 h in media supplemented with 50 ug/ml gentamicin with or without 5 ng/ml IL-1β . IL-8 secretion into cell culture supernatants was measured by ELISA ( Peprotech EC ) . The expression of IL8 from total RNA was determined using the comparative quantification method included in Rotor-Gene 1 . 7 software ( Qiagen ) as described in the supplementary methods using gene specific primers ( Protocol S1 ) . Plasmids were transfected into HeLa cells for ectopic expression of GFP fusion proteins using Lipofectamine 2000 in accordance with the manufacturer's recommendation ( Invitrogen , Carlsbad CA , USA ) . Transfected HeLa cells were treated with 20 ng/ml TNFα , 10 ng/ml IL-1β or IFNα ( 500 U/ml; Calbiochem , La Jolla , CA , USA ) for 30 min at 37°C and 5% CO2 . Transfected or infected cells were fixed in 3 . 7% ( wt/vol ) formaldehyde ( Sigma ) in PBS for 10 min and permeablized with acetone-methanol ( 1∶1 , vol/vol ) at −20°C for 15 min . Following a 30 min blocking in PBS with 3% ( wt/vol ) bovine serum albumin ( Amresco , Ohio , USA ) samples were exposed to rabbit polyclonal anti-p65 ( SC-109 , Santa Cruz , Santa Cruz CA , USA ) , anti-c-Rel ( #4727 , Cell Signaling , Beverly MA , USA ) , anti-STAT1 ( SC-345 , Santa Cruz ) , anti-STAT2 ( SC-476 , Santa Cruz ) or mouse monoclonal anti-p50 ( 2E6 , Novus Biologicals , Littleton CO , USA ) . Antibodies were used at a 1∶100 , or 1∶50 for anti-c-Rel , ( vol/vol ) diluted in blocking solution for 1 h at 20°C . Alexa Fluor 488 or Alexa Fluor 568 ( Invitrogen ) conjugated anti-mouse or anti-rabbit immunoglobulin G were used at 1∶2000 . Coverslips were mounted onto microscope slides with Prolong Gold containing 4′ , 6-diamidino-2-phenylindole ( DAPI; Invitrogen ) . For the fluorescence actin staining ( FAS ) test , HeLa and Caco-2 cells were infected with bacterial strains , fixed and permeabilized as described above and cells were incubated with 0 . 5 mg/ml phalloidin conjugated to rhodamine for 30 min . Images were acquired using a confocal laser scanning microscope ( Leica LCS SP2 confocal imaging system ) with a 100x/1 . 4 NA HCX PL APO CS oil immersion objective . Nuclear exclusion of NF-κB , STAT1 and STAT2 was quantified from at least 3 independent experiments for both transfection and infection studies . To examine the activity of NF-κB , a dual luciferase reporter system was employed . HeLa cells were seeded into 24-well trays and co-transfected with derivatives of pEGFP-C2 ( 1 . 0 µg ) together with 0 . 2 µg of pNF-κB-Luc ( Clontech , Palo Alto CA , USA ) and 0 . 04 µg of pRL-TK ( Promega , Madison WI , USA ) . Approximately 24 h after transfection , cells were left untreated or stimulated with 20 ng/ml TNFα or 10 ng/ml IL-1β for 16 h . Firefly and Renilla luciferase levels were measured using the Dual-luciferase reporter assay system ( Promega ) in the Topcount NXT instrument . For each sample , the expression of firefly luciferase was normalized for Renilla luciferase measurements and NF-κB activity was expressed relative to unstimulated pEGFP-C2 transfected cells . To measure the induction of STAT1 and 2 , the IFN-α/β-responsive luciferase reporter plasmid p ( 9-27 ) 4th ( –39 ) Lucter ( ISRE-Luc ) [15] was used in combination with the Renilla luciferase plasmid pRL-TK . HeLa cells were transfected with both plasmids as described above and stimulated with IFNα ( 500 U/ml; Calbiochem ) for 30 min . Luciferase activity was measured as described above . To measure the induction of AP-1 , the cAMP response element ( CRE ) -dependent luciferase vector pAP-1-Luc was used in combination with the Renilla luciferase plasmid pRL-TK . HeLa cells were transfected with both plasmids as described above and stimulated with 25 ng/ml phorbol 12-myristate 13-acetate ( PMA ) for 30 min . Luciferase activity was measured as described above . To test the effect of ectopically expressed NleE , NleB and OspZ and on IκB degradation , HeLa cells were mock transfected or transfected with pEGFP-C2 or pEGFP-NleE , pEGFP-NleB , pEGFP-OspZ and derivatives and incubated for 16 h before being left untreated or treated with TNF-α or IL-1β for 10 , 20 or 30 min . Cell lysis was performed by incubating cells in cold lysis buffer ( 50 mM Tris-HCl pH 8 . 0 , 150 mM NaCl , 5 mM EDTA , 1% NP-40 ) on ice for 5 min before collecting lysate and incubating on ice for a further 10 min . Cell debri was pelleted and equal volumes of supernatant were collected for SDS-PAGE . Proteins transferred to nitrocellulose membranes were probed with mouse monoclonal anti-IκBα ( Cell signaling ) diluted 1∶1000 or rabbit polyclonal anti-p65 ( Santa Cruz ) diluted 1∶1000 . For infection studies , HeLa cells were infected with derivatives of EPEC E2348/69 for 90 min before stimulation with TNF-α or IL-1β for 30 min as described above . | Bacterial intestinal pathogens have evolved distinct ways of colonizing the gut and causing disease . Enteropathogenic E . coli ( EPEC ) and its close relative enterohemorrhagic E . coli O157:H7 ( EHEC ) are extracellular pathogens that cause a characteristic lesion on the intestinal mucosa known as an attaching and effacing lesion . In contrast , Shigella is an intracellular pathogen that invades the intestinal mucosa and spreads from cell to cell . Both pathogens utilize a bacterial type III secretion system that “injects” virulence effector proteins into the host cell upon contact . We have discovered that an effector shared by EPEC/EHEC and Shigella , known as NleE or OspZ , as well as another EPEC/EHEC effector , NleB , inhibit the host cell inflammatory response by preventing translocation of the immune regulator NF-κB to the cell nucleus . Thus , although EPEC/EHEC and Shigella have evolved different colonization strategies , they share a common virulence determinant that suppresses the inflammatory response of the host , and both pathogens mediate a multi-effector attack on NF-κB signaling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"biology/cell",
"signaling",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis",
"infectious",
"diseases/gastrointestinal",
"infections"
] | 2010 | The Type III Effectors NleE and NleB from Enteropathogenic E. coli and OspZ from Shigella Block Nuclear Translocation of NF-κB p65 |
Scrub typhus is a febrile infection caused by the obligate intracellular bacterium Orientia tsutsugamushi , which causes significant morbidity and mortality across the Asia-Pacific region . The control of this vector-borne disease is challenging due to humans being dead-end hosts , vertical maintenance of the pathogen in the vector itself , and a potentially large rodent reservoir of unclear significance , coupled with a lack of accurate diagnostic tests . Development of an effective vaccine is highly desirable . This however requires better characterization of the natural immune response of this neglected but important disease . Here we implement a novel IFN-γ ELISpot assay as a tool for studying O . tsutsugamushi induced cellular immune responses in an experimental scrub typhus rhesus macaque model and human populations . Whole cell antigen for O . tsutsugamushi ( OT-WCA ) was prepared by heat inactivation of Karp-strain bacteria . Rhesus macaques were infected intradermally with O . tsutsugamushi . Freshly isolated peripheral blood mononuclear cells ( PBMC ) from infected ( n = 10 ) and uninfected animals ( n = 5 ) were stimulated with OT-WCA , and IFN-γ secreting cells quantitated by ELISpot assay at five time points over 28 days . PBMC were then assayed from people in a scrub typhus-endemic region of Thailand ( n = 105 ) and responses compared to those from a partially exposed population in a non-endemic region ( n = 14 ) , and to a naïve population in UK ( n = 12 ) . Mean results at Day 0 prior to O . tsutsugamushi infection were 12 ( 95% CI 0–25 ) and 15 ( 2–27 ) spot-forming cells ( SFC ) /106 PBMC for infected and control macaques respectively . Strong O . tsutsugamushi-specific IFN-γ responses were seen post infection , with ELISpot responses 20-fold higher than baseline at Day 7 ( mean 235 , 95% CI 200–270 SFC/106 PBMC ) , 105-fold higher at Day 14 ( mean 1261 , 95% CI 1 , 097–1 , 425 SFC/106 PBMC ) , 125-fold higher at Day 21 ( mean 1 , 498 , 95% CI 1 , 496–1 , 500 SFC/106 PBMC ) and 118-fold higher at Day 28 ( mean 1 , 416 , 95% CI 1 , 306–1 , 527 SFC/106 PBMC ) . No significant change was found in the control group at any time point compared to baseline . Humans from a scrub typhus endemic region of Thailand had mean responses of 189 ( 95% CI 88–290 ) SFC/106 PBMC compared to mean responses of 40 ( 95% CI 9–71 ) SFC/106 PBMC in people from a non-endemic region and 3 ( 95% CI 0–7 ) SFC/106 PBMC in naïve controls . In summary , this highly sensitive assay will enable field immunogenicity studies and further characterization of the host response to O . tsutsugamushi , and provides a link between human and animal models to accelerate vaccine development .
Scrub typhus is a zoonotic illness caused by the intracellular bacterium Orientia tsutsugamushi which is endemic mainly across the Asia-Pacific region [1] . The pathogen is transmitted by the bite of larval Trombiculid mites known as chiggers [2 , 3] . Scrub typhus is a febrile illness with a wide spectrum of disease severity from mild febrile illness to potentially fatal illness influenced by O . tsutsugamushi strains and host immune status . The specific skin lesion , known as an eschar , has been reported in up to 68% of Thai patients with scrub typhus [4] . The disease is treatable by antibiotics such as doxycycline , tetracycline or chloramphenicol [5] , although emergence of antibiotic resistant strains has been reported in northern Thailand [6] . If untreated , the mortality is around 6% [7] . Awareness of and research into scrub typhus has been limited by its non-specific symptoms and difficulty of diagnosis , and a vaccine is highly desirable [8 , 9] . A greater understanding of the host immune response to O . tsutsugamushi is required for vaccine design . As an obligate intracellular pathogen , cellular immunity is likely to be necessary for host control of infection . Several studies have reported a role for type 1 cell mediated immunity and specifically IFN-γ production in response to O . tsutsugamushi for immune protection against infection . Increased levels of IFN-γ and other type 1 cytokines are seen in the blood of patients with scrub typhus compared to controls [10–13] . Adoptive transfer experiments of monocyte-depleted splenocytes [14] and antigen-specific IFN-γ T-cells in murine models [15] have supported an important role for cell mediated immunity in protection against death from scrub typhus . Replication of O . tsutsugamushi inside macrophages is impaired by extrinsic IFN-γ [16] . O . tsutsugamushi infected monocyte-derived dendritic cells induce production of IFN-γ from CD4+ T-cells [17] . Scrub typhus is universally fatal in CD8-deficient mice ( compared to 50% fatality in wild type mice ) [18] , and CD8+ T-cells play a vital protective role in control of O . tsutsugamushi growth [19] . In humans , CD8+ T-cell proliferation was seen during the convalescent phase of scrub typhus in patients [20] . These studies suggest the importance of developing a reliable method of monitoring O . tsutsugamushi-specific IFN-γ responses during scrub typhus . The enzyme-linked immunospot ( ELISpot ) assay is a very sensitive technique allowing quantification of antigen-specific T-cells at the single cell level from peripheral blood by detection of IFN-γ or other secreted cytokines [21 , 22] . The ex-vivo IFN-γ ELISpot assay is the most widely used technique for monitoring T-cell-based immune responses against intracellular pathogens such as HIV [23] , tuberculosis [24] and malaria [25] . There are several advantages of the ELISpot assay for use in clinical trials: it has high sensitivity , is relatively easy to perform , uses low number of cells in the assay , does not require expensive instrumentation , and has the potential for high throughput screening with numerous specific peptides , or to an entire pathogen proteome using overlapping peptides of varying lengths . The ELISpot is up to 200 times more sensitive for cytokine detection than ELISA [26 , 27] and significantly more sensitive than flow-cytometric based techniques [28] . Non-human primates ( NHP ) represent a model for investigating immunity to scrub typhus and provide valuable information to develop potential candidate vaccines for future testing in the clinical setting . Infection of cynomolgus macaques ( Macaca fascicularis ) with O . tsutsugamushi causes infection and illness which closely resemble the course of scrub typhus in humans [29 , 30] . Due to some limitations of using cynomolgus macaques , such as reagent and antibody availability , the rhesus macaque ( Macaca mulatta ) is likely to be more suitable for preclinical vaccine evaluation . In this study , we developed a novel ex-vivo IFN-γ ELISpot assay using whole cell antigen of O . tsutsugamushi ( OT-WCA ) as an antigen to determine magnitude and frequency of cellular responses in peripheral blood mononuclear cells ( PBMC ) of rhesus macaques . Our results indicate that our ex-vivo IFN-γ ELISpot assay can be used to determine immune responses against O . tsutsugamushi with high sensitivity and potentially high specificity for evaluation of vaccine candidate efficacy against O . tsutsugamushi in rhesus macaques and human clinical trials .
All animal research was performed strictly under approved IACUC protocol by the Institutional Animal Care and Use Committee and Biosafety Review Committee at the Armed Forces Research Institute of Medical Sciences ( AFRIMS ) Bangkok , Thailand , an AAALAC International-accredited facility . The IACUC protocol numbers are: PN12-01 ( approved 31st Jan 2012 ) , and PN13-12 ( approved 24th Jan 2014 ) . The animal research was conducted in compliance with Thai laws , the Animal Welfare Act , and all applicable U . S . Department of Agriculture , Office of Laboratory Animal Welfare and U . S . Department of Defense guidelines . All animal research adhered to the Guide for the Care and Use of Laboratory Animals , NRC Publication ( 8th Edition ) [31] . Animals were housed individually in standard squeeze-type stainless steel cages with a minimum floor space of 4 . 4 square feet equipped with standard enrichments and exposed to ambient environmental conditions inside an Animal Biosafety Level 3 ( ABSL-3 ) containment laboratory . Monkeys were fed daily with commercially prepared old-world primate extruded feed and supplemented with fresh fruit or vegetable four times per week . Fresh chlorinated water ( 5–10 ppm ) was provided ad libitum via automatic water valves . Cages were cleaned daily and sanitized biweekly . All procedures were performed under anesthesia using ketamine hydrochloride , and all efforts were made to minimize stress , improve housing conditions , and to provide enrichment opportunities . Animals were euthanized by ketamine hydrochloride injection followed by barbiturate in accordance with the Guidelines for the Euthanasia of Animals ( 2013 Edition of the American Veterinary Medical Association ) . The O . tsutsugamushi Karp strain ( New Guinea ) used in this study was provided by the Naval Medical Research Center ( NMRC ) , Silver Spring , MD , USA and is a well-characterized strain from a pre-existing collection of Orientia strains at the NMRC , previously used in related studies [30 , 32] . Fifteen ( 9 male and 6 female ) Indian-origin rhesus macaques ( M . mulatta housed in AFRIMS ( an AAALAC-accredited Program ) were used in this study . Environmental conditions were maintained in accordance with the Guide for the Care and Use of Laboratory Animals 8th edition ( 2011 ) [31] . The animals ranged from 3 to 5 year of age and weighed between 4 . 1 and 5 . 9 kg at the start of the study . The animals were evaluated to ensure that they were negative for SIV , SRV , STLV-1 and herpes B virus and had no O . tsutsugamushi exposure from the experimental history . Additionally , their antibody titers to O . tsutsugamushi were confirmed to be negative prior to the start of the experiment . Aliquots of inoculum containing defined concentrations of O . tsutsugamushi Karp strain at a dose of either 107 or 107 . 8 muLD50 ( cultured and prepared in yolk sacs of chicken eggs ) , were used for intradermal ( ID ) injections of a total of 10 macaques . The inoculum was prepared in collaboration with the Mahidol-Oxford Tropical Medicine Research Unit ( MORU ) and the Naval Medical Research Center ( NMRC ) , Silver Spring , Maryland , USA . The dosages were administered in infected groups on day 0 . The inoculum was suspended in 100 μl of Snyder’s buffer and was applied to the anterior medial aspect of the left thigh . A total of 5 macaques were used as control group . The control macaques received ID injection with homogenized un-infected inocula in Snyder’s buffer at the identical site on the anterior medial aspect of the left thigh . Blood samples for PBMC isolation were collected at 5 time points starting from Day 0 prior to ID inoculation and every week up to Day 28 . Bacteremia was determined using a previously describes qPCR assay for O . tsutsugamushi-specific 47 kDa gene [33] . DNA was extracted from 200 μl whole blood from each macaque using DNeasy Blood & Tissue Kit ( Qiagen , Valencia , CA , USA ) according to the manufacturer’s instructions . The O . tsutsugamushi-specific 47 kDa htra gene real time PCR assay was used as previously described using a CFX96 Real Time PCR Detection System ( Biorad , Hercules , CA “no template” negative controls were run with each reaction and plasmid DNA served for standard curves in serial dilution from 106 to 1 copies/μl of 47 kDa gene [34] . PBMC were separated from 12 ml heparinized blood samples by density centrifugation within 3 hours of blood draw . In brief , 3 ml of Ficoll-HyPaque singular ( Pharmacia , Peapack , NJ ) was preloaded in a 14 ml LeucoSep tube ( Greiner Bio-One ) by centrifugation for 1 min at 1 , 000 × g . The whole blood was added to the LeucoSep tube and centrifuged for 15 min at 800 × g at room temperature . The cell suspension was collected , and the cells were washed twice in complete RPMI medium [RPMI 1640 ( Sigma-Aldrich , St . Louis , MO ) containing 10% FBS ( Invitrogen Corp . , Carlsbad , CA ) , 2 mM L-glutamine , 50 U/ml gentamicin ( Quality Biological Inc . , Gaithersburg , MD ) , and 0 . 1 mM non-essential amino acids ( Sigma-Aldrich , St . Louis , MO ) ] for 5 min at 640 × g and 7 min at 470 × g , respectively . After final washing , the pellet was resuspended in complete RPMI medium before counting . The O . tsutsugamushi Karp strain was cultivated in L929 cells ( mouse fibroblast cell line ) cultured in RPMI1640 supplemented with 10% fetal bovine serum ( FBS ) and 2 mM L-glutamine . The stage of infection was determined by indirect immunofluorescence assay ( IFA ) ; when infected L929 cells approached 100% positive infection , the cells were harvested by centrifugation at 6 , 000 x g for 30 min at 4°C . The cells were pelleted and disrupted using glass beads ( 0 . 1 mm , Next Advance , Averill Park , NY ) , then homogenized with a bullet blender for 1 min . After centrifugation at 300 x g for 10 min to remove cell debris , the supernatant was collected and filtered through a 2 . 0 μm syringe filter . The supernatant was then centrifuged at 11 , 000 x g for 10 min to collect the bacterial pellet . After washing the pellet with PBS , the bacteria were resuspended in 50 μl PBS and heated at 80°C for 1 hour . The OT-WCA suspension was then aliquoted and stored at 4°C until used , with immunogenicity of the antigen confirmed up to five years of storage . The total protein concentration of OT-WCA protein was determined by BCA assay ( BCA1 kit , Sigma-Aldrich , St . Louis , MO ) . All processes were performed in a biosafety level 3 laboratory . The protein concentration of the stock solution for the OT-WCA used in this study was 1 . 41 mg/ml . Phytohemagglutinin ( PHA ) was used at a final concentration of 5 μg/ml and complete RPMI as described above was added to positive control wells and negative wells , respectively . The kinetics and magnitude of the cellular responses to whole O . tsutsugamushi were assessed by ex-vivo IFN-γ ELISpot following an 18 hour stimulation of PBMC with OT-WCA for each time point of the study . Fresh PBMC were used in all ELISpot assays , and separate ELISpot kits ( Mabtech , AB , Sweden ) for human ( 3420-2A ) and monkey ( 3421M-2A ) cells were used . Briefly , 96-well Multiscreen-I plates ( Millipore , UK ) were coated for 3 hours with 10 μg/ml GZ-4 anti-human IFN-γ ( Mabtech , AB , Sweden ) at room temperature . Fresh PBMC were added in duplicate wells at 2x105 PBMC in 50 μl per well and 50 μl of OT-WCA was added at the optimal concentration . For human studies , a T-cell epitope pool ( Mabtech , AB , Sweden ) at a final concentration of 1 μg/mL was used as control antigens . After 18 hours , secreted IFN-γ was detected by adding 1 μg/ml biotinylated mAb 7-B6-1-biotin for IFN-γ , which recognises an epitope completely conserved between human and macaques in the helical region of human IFN-γ , ( Mabtech , AB , Sweden ) for 3 hours and followed by 1 μg/ml streptavidin alkaline phosphatase ( Mabtech , AB , Sweden ) . The plates were developed using the AP Conjugate Substrate Kit ( Biorad , USA ) according to the manufacturer’s instructions . ELISpot plates were scanned using a CTL ELISpot reader ( Cellular Technology Limited , USA ) . Spots were then counted by Immunospot 3 . 1 software , using the manufacturer’s automated SmartCount™ settings . Results were expressed as IFN-γ spot-forming cells ( SFC ) per million PBMC . Background responses in unstimulated control wells were always less than 20 spots/106 PBMC , and were subtracted from those measured in OT-WCA stimulated wells . The recruitment of human subjects for immunological studies has been described previously in a study of melioidosis [35] . PBMC were isolated from subjects in Ubon Ratchathani , ( northeastern Thailand—an endemic area for scrub typhus ) , Bangkok , ( central Thailand—a non-endemic for scrub typhus ) and Oxford , UK ( non-endemic for scrub typhus ) participating in a study of melioidosis . Responses to O . tsutsugamushi were evaluated using the same ex vivo IFN-γ ELISpot assay , with PBMC from Ubon Ratchathani subjects known to known to be reactive to OT-WCA used in Oxford as positive controls . Human anti-Orientia antibodies ( IgM/IgG ) were detected using IFA for scrub typhus , based on pooled whole-cell antigens from three strains of O . tsutsugamushi ( Karp , Kato and Gilliam strains ) as previously described [36] . IFN-γ ELISpot responses were compared for people with IFA IgG titers of ≥ 1:400 compared to those with titres < 1:400 . Statistical analyses were performed using GraphPad Prism Software v . 6 . The results between the control group versus the infected group are expressed as means and were compared using the non-parametric Mann-Whitney U- test . Significant differences between timepoints within a group were determined with the non-parametric Wilcoxon t-test . The relationship between IFN-γ ELISpot responses and IFA IgG titers was evaluated using Spearman’s rank correlation test . Two-tailed P values < 0 . 05 were considered significant .
Frozen PBMC from rhesus macaques collected at Day 14 post inoculation ( pi ) with O . tsutsugamushi ( BRI-02 ) and from a control uninfected macaque ( BRI-06 ) were stimulated in duplicate with 50 μl of OT-WCA prepared at 3 different concentrations: 1 . 41 ( 1:50 dilution ) , 0 . 71 ( 1:100 ) and 0 . 14 ( 1:500 ) μg/well . PHA and ‘complete media’ were used as positive and negative controls respectively , ( Fig 1 and Table 1 ) . Strong IFN-γ ELISpot responses to OT-WCA were observed in infected macaques , whereas no responses to OT-WCA was observed in the uninfected macaque . Strong responses to PHA stimulation were found in positive control wells ( Fig 1 and Table 1 ) . Concentrations of OT-WCA above 1 . 41 μg/well were not tested because a blackout of the spot count was likely , and the optimized 0 . 14 μg/well concentration was selected for OT-WCA for the stimulation in further testing of experimental macaques . The O . tsutsugamushi-specific cellular immune responses were measured with ex-vivo IFN-γ ELISpot assay from freshly isolated PBMC of rhesus macaques at five time points after O . tsutsugamushi infection ( Day 0 , 7 , 14 , 21 and 28 ) to study the kinetics and magnitude of the responses over time ( Fig 2 ) . Freshly isolated PBMC from each macaque were tested in duplicate with OT-WCA at 0 . 14 μg/well . Each well contained 2x105 PBMC , and SFCs were quantitated by the CTL ELISpot reader—therefore responses were multiplied by 5 to provide SFC / million PBMC . Strong responses were found from PHA stimulated wells in all macaques , and background responses assessed by media only were always less than 20 SFC/106 PBMC ( Table 2 ) . Specific responses to O . tsutsugamushi were calculated by subtraction of corresponding media only wells from OT-WCA . Wells with very high responses resulted in blackout of the spot count and are represented as 300 spots ( 1 , 500 SFC/106 PBMC ) , corresponding to the highest spot count that can be measured by CTL ELISpot reader in our experiment . Overall , O . tsutsugamushi-specific IFN-γ responses were observed from all infected macaques . At Day 7 , the response to OT-WCA from infected macaques ( mean 235 SFC/106 , 95% CI 200–270 SFC/106 ) was 20-fold higher than baseline level ( Day 0; mean 12 SFC/106 and 95% CI 0–25 SFC/106 ) . At Day 14 the specific IFN-γ responses rose >100-fold ( mean 1 , 261 95% CI 1 , 097–1 , 425 SFC/106 PBMC ) , and were 125-fold higher at Day 21 ( mean 1 , 498 , 95% CI 1 , 496–1 , 500 SFC/106 PBMC ) and 118-fold higher at Day 28 ( mean 1 , 416 , 95% CI 1 , 306–1 , 527 SFC/106 PBMC ) . No change in specific IFN-γ response to O . tsutsugamushi was found for uninfected control macaques . An overview of the adjusted spots count is shown in Table 2 . We explored the relationship in the macaque model between ex-vivo IFN-γ ELISpot of cellular immune responses to O . tsutsugamushi and antibody responses to O . tsutsugamushi as measured by an IgG and IgM IFA assay . We saw a correlation between the magnitude of the cellular response and the reciprocal titers of the IgG-based IFAs in the non-human primate model ( r 0 . 79 = P < 0 . 001 by Spearman’s rank test ) ( Fig 3 , panel A ) . A significant correlation was not seen for each individual time point . This may be because there are only ten data points per timepoint and this is insufficient to make a correlation . In addition , the relationship in the first 28 days is limited by the IgG rising more slowly than the Elispot response , the latter plateauing by Day 21 ( Fig 3 , panel A ) , whilst IgG responses to infections are generally believed to peak later at around 4 to 6 weeks [37] . We also investigated the relationship of cellular immune responses to O . tsutsugamushi and bacterial load in blood ( expressed as AUC of bacteremia ) at Day 14 , which corresponds to the peak bacteremia phase in this model , and found an inverse correlation , with increased SFCs /106 PBMC relating to lower bacterial loads ( Fig 3 , panel B ) . In order to explore the feasibility of using this assay in human populations , responses to OT-WCA were measured in patients participating in a study of immune responses to a different disease ( melioidosis ) . Subjects living in the scrub typhus endemic region of Ubon Ratchathani , Northeast Thailand ( n = 105 ) , had a mean IFN-γ ELISpot response of 189 SFC / 106 PBMC ( 95% CI 88–290 ) , compared to 40 SFC / 106 PBMC ( 95% CI 9–71 ) in subjects living in a non scrub typhus endemic area ( Bangkok , Thailand , n = 14 ) some of whom may have grown up in or travelled to an endemic part of Thailand , and 3 SFC / 106 PBMC ( 95% CI 0–7 ) for subjects in Oxford , UK ( n = 12 ) who had never encountered scrub typhus . 17/105 subjects ( 16% ) in the endemic region had high responses greater than 200 SFC / 106 PBMC compared to none in the non-endemic and naïve regions ( Fig 4 , panel A ) . No differences between groups were seen for responses to a control panel of common T-cell epitopes for Epstein–Barr virus ( EBV ) , cytomegalovirus ( CMV ) , influenza etc ( “T-cell control panel” , Fig 4 , panel A ) . Provisional studies showed that ELISpot counts were greatly reduced in responders if cryopreserved PBMC rather than fresh PBMC were used , suggesting a requirement for fresh antigen presenting cells to optimally process whole bacteria . As for the macaque model , we saw a correlation between the magnitude of the cellular response and the reciprocal titers of the IgG-based IFAs in the non-human primate model ( Spearman’s R = 0 . 57 , P < 0 . 001 by Spearman’s rank test ) ( Fig 4 , panel B ) . When cellular immune responses to OT-WCA antigen measured by ex vivo IFN-γ ELISpot were compared between humans with either an IFA IgG titer of 1:400 or above ( n = 22 ) or less than 1:400 ( n = 84 ) , significantly higher cellular immune responses were found in people with the higher IFA ( Fig 4 , panel C ) .
We have established a highly sensitive method for measuring the magnitude and kinetics of the adaptive cellular immune response to scrub typhus in rhesus macaque monkeys . This study builds on previous work demonstrating production of IFN-γ in the host defense against O . tsutsugamushi . The OT-WCA ELISpot assay was developed using PBMC from rhesus macaques ( Macaca mulatta ) , the most commonly used NHP model for preclinical vaccine development . To evaluate antigen-specific cellular responses to Orientia after infection with O . tsutsugamushi strain Karp , freshly isolated PBMC from a total of ten O . tsutsugamushi infected rhesus macaques and five uninfected control macaques were tested with the novel ELISpot assay from Day 0 through Day 28 . O . tsutsugamushi-specific IFN-γ responses were observed post infection from all infected macaques compared to uninfected macaques , with a 20-fold ( mean 235 , 95% CI 200–270 SFC/106 PBMC ) increase at Day 7 , a 100-fold increase at Day 14 and maintenance of high level responses to the end of the study ( Day 28 ) . The maximum measurable response was limited by the SmartCount™ software , which does not read any higher once a blackout of the well is obtained . No significant increases in cellular responses were found in the uninfected control group at any time point . Preliminary human studies in subjects from an endemic area , a non-endemic area where some of the population has previous exposure , and from non-exposed subjects gives support to the specificity of this assay for studying human populations , although the role of cross-reactivity to other rickettsial group pathogens merits further exploration . The major reason for developing the IFN-γ ELISpot assay for O . tsutsugamushi is to allow immunogenicity monitoring in animal models , in scrub typhus exposed but healthy populations , in patient populations , and for future vaccine trials . IFN-γ was chosen as the read-out cytokine because of previous work demonstrating IFN-γ responses in scrub typhus patients [10–13] and mouse studies [14 , 15] . The inverse relationship seen in this study for macaques 14 days post infection , where higher IFN- γ responses are associated with lower bacterial loads also lends support for the importance of the IFN-γ response in control of the bacteria , although other factors such as sicker animals having lower immune responses may be relevant . The importance of IFN-γ in response to intracellular pathogens has been reaffirmed recently by transcriptomic studies [38–40] , and by a study demonstrating the link between BCG-specific T cells secreting IFN-γ and reduced risk of developing tuberculosis in South African infants [41] . However , other cytokine responses are important , for example IL-2 is involved in the development of memory responses and antibodies following vaccination against malaria [42] , Hepatitis B [43] and tick-borne encephalitis [44] . Further studies by flow cytometry of multifunctional T-cell responses following scrub typhus infection are underway for humans and macaques . For this initial characterization of the cellular response to Orientia , the OT-WCA was based on a single strain of high relevance in human disease ( Karp strain ) . The antigen used for the IFA assay is based on pooled whole-cell antigens from three strains of O . tsutsugamushi ( Karp , Kato and Gilliam ) as this is the standard assay in the field . The conventionally used IFA slides based on the three reference strains have served over many years to document humoral responses against single strain Orientia infections , and results have been validated via gene sequencing methods [45 , 46] . Ongoing work in the laboratory is exploring the immunogenicity of different strains and culture conditions . A potential bias to the ELISpot assay is the potential for persistent presence of live O . tsutsugamushi bacteria in the PBMC of infected monkeys , driving the antigen specific response . However , we would expect to see responses in the media only in wells of the Day 7 and 14 macaque PBMC if residual bacteria were contributing to the measurable response . The cell culture media used contains penicillin and streptomycin to limit antimicrobial contamination in the laboratory , which may have a partial efficacy against the bacteria , but these antibiotics were used uniformly in all samples so should not introduce bias . Previous vaccine development studies in cynomolgus macaques [29 , 30] have demonstrated the presence of cellular immunity to two recombinant proteins derived from Karp stain O . tsutsugamushi ( Kp r47b and Kp r56 ) using a 36-hour ELISpot assay . The magnitude of the IFN-γ responses to the whole bacteria reported in this manuscript was much higher than the levels to the individual proteins ( range 125–800 SFC/106 PBMC ) , alongside a lack of responses in unexposed animals . This is likely to be associated with the new and modified approach to purify O . tsutsugamushi to produce the OT-WCA suspension as described above . This OT-WCA ELISpot assay therefore allows better demonstration of cellular immunity to the whole bacteria , and can be performed as a positive control alongside exploration of immunity to specific vaccine candidate proteins such as Kp r47b and Kp r56 . The magnitude of cellular immune response seen to OT-WCA compared to reported responses to the 47 and 56 kDa proteins suggest that there are other immunodominant antigens in O . tsutsugamushi for discovery as vaccine candidates . Stimulation with this preparation of killed O . tsutsugamushi in scrub typhus naïve monkeys did not invoke measurable innate cellular responses by this assay , for example from macrophages and NK cells via pattern recognition receptor pathways . This may be due to inactivated whole cell antigen being suboptimal at this dose for induction of innate pathways , or involvement of other key cytokines in pattern recognition , such as TNF-α , IL-12 and IL-1β not measured by this assay . In addition , potential evasion of the host innate immune response by O . tsutsugamushi is of interest in understanding the pathogenesis of natural infection and will be the subject of ongoing studies . Scrub typhus diagnostics is a major difficulty for both management of patients and for epidemiological studies , and the lack of a clear-cut “gold standard” reference means that statistical modeling has been required to evaluate novel alternative diagnostic tests [36] . The OT-WCA IFN-γ ELISpot assay uses the same technology platform as the T-SPOT interferon-gamma release assay ( IGRA ) developed for diagnosis of tuberculosis [47] . Due to the laboratory processing requirements of ELISpot assays using fresh cells we do not believe there will be a role for using the OT-WCA IFN-γ ELISpot assay for real-time diagnosis and patient management of scrub typhus in endemic areas . However , given the unexpectedly high sensitivity and potential high specificity observed in this study , the OT-WCA IFN-γ ELISpot assay could potentially be used as a reference standard in research studies evaluating novel diagnostics . However evaluation regarding cross-reactivity with other rickettsial group bacteria and/or cross-protection in heterologous re-infection/infection in scrub typhus are needed . The optimization of OT-WCA preparation for antigen stimulation studies provides a way of exploring the host response to O . tsutsugamushi by flow cytometry , thus allowing detailed characterization of cell phenotype , infected cells , secretion of cytokines and pathways of response to the bacteria . Immune-phenotyping studies of infected cells in eschar biopsies from scrub typhus patients have demonstrated a cellular tropism of O . tsutsugamushi for antigen presenting cells in the skin [48] . Although the ELISpot assay is a highly sensitive method for measuring cellular responses [21 , 22] , this method is unable to identify the cell phenotype secreting the cytokine . Further studies including a small volume , whole blood stimulation assay are underway to define which cells contribute most to IFN-γ secretion , and to characterize other cytokines associated with O . tsutsugamushi infection . This study did not address the cross-reactivity of immunity to Karp strain compared to other bacterial strains , and cross-reactivity with other rickettsial group bacteria in the region . Further studies are addressing this complex issue . In summary , we have successfully developed for the first time a novel ex vivo IFN-γ ELISpot assay to whole O . tsutsugamushi antigen . This assay will allow field immunogenicity studies , pave the way for more detailed flow cytometry studies of response to O . tsutsugamushi antigen and provide a link between human and animal models to enhance vaccine development . | Scrub typhus is a disease caused by bacteria that invade cells in our immune system and blood vessels . It is transmitted by mites and is treatable with antibiotics . Unfortunately diagnosis is difficult and requires techniques that are not easily accessible everywhere . Currently , there is no scrub typhus vaccine available . In order to improve diagnostics and vaccine development in future , we need to better understand our immune response against these bacteria . In this study , we developed a test where these bacteria were killed and prepared by a new purification method to stimulate the immune cells in our blood -not antibodies . We evaluated this test in hospitalized patients with scrub typhus disease and also in non-human primates to study the responses over time . The test proved to be very accurate and useful to study natural immune responses , and we found differences in responses in areas where scrub typhus is common , compared to areas where it is not common . This test will allow us to investigate the immune response to scrub typhus more in-depth in the future , and will support the development of better diagnostic tests and vaccines against scrub typhus . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"typhus",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"vertebrates",
"animals",
"mammals",
"primates",
"animal",
"... | 2017 | Strong interferon-gamma mediated cellular immunity to scrub typhus demonstrated using a novel whole cell antigen ELISpot assay in rhesus macaques and humans |
Trypanosoma cruzi is the causal agent of Chagas Disease that is endemic in Latin American , afflicting more than ten million people approximately . This disease has two phases , acute and chronic . The acute phase is often asymptomatic , but with time it progresses to the chronic phase , affecting the heart and gastrointestinal tract and can be lethal . Chronic Chagas cardiomyopathy involves an inflammatory vasculopathy . Endothelial activation during Chagas disease entails the expression of cell adhesion molecules such as E-selectin , vascular cell adhesion molecule-1 ( VCAM-1 ) and intercellular cell adhesion molecule-1 ( ICAM-1 ) through a mechanism involving NF-κB activation . Currently , specific trypanocidal therapy remains on benznidazole , although new triazole derivatives are promising . A novel strategy is proposed that aims at some pathophysiological processes to facilitate current antiparasitic therapy , decreasing treatment length or doses and slowing disease progress . Simvastatin has anti-inflammatory actions , including improvement of endothelial function , by inducing a novel pro-resolving lipid , the 5-lypoxygenase derivative 15-epi-lipoxin A4 ( 15-epi-LXA4 ) , which belongs to aspirin-triggered lipoxins . Herein , we propose modifying endothelial activation with simvastatin or benznidazole and evaluate the pathways involved , including induction of 15-epi-LXA4 . The effect of 5 μM simvastatin or 20 μM benznidazole upon endothelial activation was assessed in EA . hy926 or HUVEC cells , by E-selectin , ICAM-1 and VCAM-1 expression . 15-epi-LXA4 production and the relationship of both drugs with the NFκB pathway , as measured by IKK-IKB phosphorylation and nuclear migration of p65 protein was also assayed . Both drugs were administered to cell cultures 16 hours before the infection with T . cruzi parasites . Indeed , 5 μM simvastatin as well as 20 μM benznidazole prevented the increase in E-selectin , ICAM-1 and VCAM-1 expression in T . cruzi-infected endothelial cells by decreasing the NF-κB pathway . In conclusion , Simvastatin and benznidazole prevent endothelial activation induced by T . cruzi infection , and the effect of simvastatin is mediated by the inhibition of the NFκB pathway by inducing 15-epi-LXA4 production .
Chagas disease ( CD ) afflicts more than ten million people in Latin-America , where it is endemic , and worldwide as a consequence of migration [1] . This disease is caused by Trypanosoma cruzi , a vector-borne flagellate protozoan that infects virtually any nucleated cell in its mammalian hosts [2] . CD evolves from an acute , frequently asymptomatic , unspecific phase towards a chronic , silent phase . In total , 30% of chronically infected patients develop clinical manifestations due to gastrointestinal or cardiac involvement . Finally , patients die because of cardiovascular complications such as heart failure , arrhythmias , or thromboembolism secondary to ventricular aneurysms . Indeed , chronic Chagas cardiomyopathy ( CCC ) is responsible for the high burden of disease and explains its high mortality [3] . CCC pathophysiology involves parasite permanence in myocardial tissue and persistence of immune system activation including generation of autoantibodies against cardiac cholinergic receptors and ultimately microvascular damage [4] . T . cruzi reportedly induces endothelial activation [5] as revealed by an increase in the expression of endothelial cell adhesion molecules ( ECAMs ) such as E-Selectin , vascular cell adhesion molecule-1 ( VCAM-1 ) and intercellular cell adhesion molecule-1 ( ICAM-1 ) [6] through a mechanism involving NF-κB activation [7] . Endothelial activation induces vasoconstriction , inflammatory cell recruitment favoring immune cell homing , and generation of a procoagulant environment that promotes local ischemia [8 , 9] . Current drug therapy is not one hundred percent curative , especially during the chronic phase , and has diverse adverse events that affect patient compliance and often require treatment suspension . Nonetheless , current advances in trypanocidal therapy have not generated drugs that exceed the effectiveness of current medications , although several triazole derivatives are promising [4] . Thus , a novel strategy is proposed that aims at some pathophysiological processes to facilitate current antiparasitic therapy , decreasing treatment length or doses and slowing disease progress . Previously , it was suggested that aspirin , a well-known and widely used medication , could perform this function [10] . Herein , we present evidence that statins , mainly simvastatin , can play a similar role . This drug decreases inflammatory infiltration in the hearts of T . cruzi-infected dogs [11] . This anti-inflammatory effect is part of the pleiotropic effects of statins , which has been related to a novel pro-resolving lipid that is an aspirin-triggered lipoxin , 15-epi-lipoxin A4 ( 15-epi-LXA4 ) [12] . The present report provides evidence that links the effects of simvastatin and benznidazole to T . cruzi-infection induced endothelial activation and the relationship between endothelial activation and 15-epi-LXA4 production . These two drugs decreased CAM expression and leukocyte adhesion in an in vitro infection model . Furthermore , the effect of benznidazole on endothelial activation is independent of the parasite , suggesting an independent anti-inflammatory action .
EA . hy926 cells ( ATCC CRL2922 ) are a human umbilical vein cell line established by fusing primary human umbilical vein cells with a thioguanine-resistant clone of A549 by exposure to polyethylene glycol ( PEG ) . Hybrid clones were selected in HAT medium and screened for factor VIII-related antigen . The cell line was cultured following reported conditions [13] . Cells were cultured on Iscove's Modified Dulbecco's Medium ( IMDM , Biological Industries , Israel ) supplemented with 10% v/v FBS , 100 U/mL penicillin , and 100 mg/mL streptomycin at 37°C and 5% CO2 . HL-60 cells ( ATCC CCL240 ) are a promyelocytic cell line that was derived by S . J . Collins et al [14] . Peripheral blood leukocytes were obtained by leukapheresis from a 36-year-old Caucasian female with acute promyelocytic leukemia . The cell line was cultured with Iscove's Modified Dulbecco's Medium plus 10% v/v FBS . HUVECs ( C-015-10C , Cascade Biologics , Life Technologies , USA ) are primary human umbilical vein endothelial cells that are pooled from multiple donors . Cells were cultured in medium 200 ( Cascade Biologics , USA ) that had been supplemented with low serum growth supplement ( LSGS , Cascade Biologics ) . T . cruzi trypomastigotes ( Dm28c clone [15] ) from our collection , were obtained from infected EA . hy926 cells . Cells were exposed to trypomastigotes ( Dm28c clone ) at a multiplicity of infection ( MOI ) of 5 . Trypomastigotes were allowed to infect cells for 24 hours , after which the supernatant was removed and fresh medium was added . Trypomastigotes were released from EA . hy926 cells after four days of infection . The parasites were harvested and collected for viability assays and further cell infections . Simvastatin , benznidazole , AA-861 ( 2- ( 12-hydroxydodeca-5 , 10-diynyl ) -3 , 5 , 6-trimethyl-p-benzoquinone ) were obtained from Sigma-Aldrich , USA . 15-epi 15-epi-lipoxin A4 ( cat#90415 ) , and 5 ( S ) , 6 ( R ) -Lipoxin A4 methylester ( cat#10033 ) were from Cayman Chemical , USA . All drugs were dissolved in DMSO and controls were incubated with DMSO vehicle alone . DMSO final concentration at cell cultures was 0 . 025% v/v . For most of the experiments , simvastatin and benznidazol concentration was 5 and 20 μM , respectively . It has been reported that 5 μM simvastatin is effective in decreasing inflammation and expression of ECAMs [16] [17] . 20 μM benznidazole correspond to the IC50 for its trypanocidal action [18] . The effect of the drug on all cells and parasite viability was evaluated through the tetrazolium salt ( MTT , Sigma-Aldrich ) reduction assay as described [19] . Drugs at concentrations ranging from 1 to 20 μM , dissolved in DMSO at a 0 . 025% v/v final concentration , were applied to 2 . 5x105 cells/mL or 106 parasite/mL culture medium . Cultures were incubated for 24 hours before adding MTT for 4 hours . The plates were incubated overnight with 10% SDS w/v in 0 . 01 M HCl at 37°C , and optical density ( OD ) was determined using a microplate reader ( Labsystems Multiskan MS , Finland ) at 570 nm . Under these conditions , the OD was directly proportional to viable cell number per well . All of the experiments were performed at least three times , and the data are shown as the means and their standard deviations from triplicate cultures . 5×105 EA . hy926 or HUVEC cells/well were seeded in 6-well plates and exposed to various simvastatin and/or benznidazole concentrations for 24 hours . Then , cells were infected with trypomastigotes at a MOI of 10 and incubated for 16 hours . Cells were detached with 1X EDTA-PBS at 0 . 5 mM . The harvested cells were washed with cold PBS and centrifuged at 800 x g for 5 minutes . Then , cells were washed with flow cytometry buffer three times . 100 μL of cell suspensions were incubated for 45 min at 4°C in the dark with mouse monoclonal anti- human E-selectin ( 5 μL undiluted ) , ICAM-1 ( 3 μL undiluted ) , and VCAM-1 ( 1 μL undiluted ) that were conjugated with PE , FITC and APC , respectively ( BioLegend , USA ) . Cell suspensions were analyzed by flow cytometry using a FACSAria-III flow cytometer ( BD Biosciences , USA ) . A homogeneous cell population was selected by size vs . granularity in log scale for these two conditions . 2x104 EA . hy926 cells/slide were seeded in Lab-Tek II Chamber SlideTM ( ThermoScientific , USA ) and allowed to adhere overnight . Then , cells were incubated with 5 μM simvastatin or 20 μM benznidazole for 24 hours prior to T . cruzi- infection ( Dm28c clone ) at a MOI of 10 . After 16 hours of infection , cells were fixed in 4% formaldehyde–0 . 1 M phosphate buffer ( pH 7 . 3 ) for 10 min . Cells were blocked with 3% bovine serum albumin for 1 hour . Then , cells were incubated with monoclonal antibodies against E-selectin ( 1:100 ) , VCAM-1 ( 1:250 ) and ICAM-1 ( 1:100 ) ( from Abcam , UK ) overnight at 4°C . The samples were washed with PBS and incubated with anti-rabbit IgG that had been conjugated with fluorescein ( 1:100 ) from Sigma-Aldrich for 1 h . Finally , nuclei were stained with DAPI for 5 minutes and mounted with Dako Fluorescence Mounting ( Dako , USA ) . The cells were photographed using a Nikon Eclipse 400 fluorescence microscope ( Nikon , Japan ) , and images were analyzed by mean intensity using ImageJ software ( ImageJ 1 . 47v ) . 5x105 endothelial cells were rinsed once with PBS and scraped . Proteins were extracted into 100 μL RIPA buffer plus a protease and phosphatase inhibitor cocktail ( Radio-Immunoprecipitation Assay; Millipore Corporation , USA ) at 4°C . Total protein was quantified with the Lowry assay . Samples in loading buffer ( 10% SDS , 50% glycerol , 0 . 5 M Tris , 0 . 1% bromophenol blue , and 1 M dithiothreitol , pH 6 . 8 ) were electrophoretically separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane . Nitrocellulose membranes were incubated with blocking solution ( 5% w/v nonfat dry milk , 0 . 1% Tween-20 in Tris-buffered saline ) at room temperature for 1 hour . Then , the membranes were incubated at 4°C overnight with 1 mL solution of primary antibodies against E-selectin ( 1:1000 ) and α-Tubulin ( 1:10 . 000 ) from Sigma-Aldrich , ICAM-1 ( 1:2000 ) and VCAM-1 ( 1:1000 ) from Abcam , 7 ml solution of p-IKK ( 1:1000 ) , IKK ( 1:1000 ) , p-IκB ( 1:1000 ) and IκB ( 1:1000 ) from Cell signaling ( Cell Signaling Technologies , USA ) . Bound antibodies were detected with horseradish peroxidase-conjugated secondary antibodies and visualized by Lumminata forte ( Millipore ) . The nitrocellulose was stripped between reprobes using Mild stripping ( Millipore ) for 10 minutes . Developed films were scanned , and band densitometry was analyzed using ImageJ software ( ImageJ 1 . 47v ) . 2x104 EA . hy926 cells/slide were seeded in Lab-Tek II Chamber Slides for 12 h . Cells were exposed to 5 μM simvastatin or 20 μM benznidazole for 24 hours prior to T . cruzi- infection ( Dm28c clone ) at a MOI of 10 for 16 hours . Then , the cells were washed and fixed in cold methanol ( 70% ) overnight . The fixed cells were then washed , and 1 mL PBS ( pH 7 . 4 ) was added . DNA was stained with DAPI ( NucBlue; Molecular Probes , USA ) following the manufacturer's instructions . The cells were photographed using a Nikon Eclipse 400 fluorescence microscope using 358 nm ( excitation ) and 461 nm ( emission ) wavelengths . In total , ten pictures were obtained per well , and each picture was counted using MATLAB software . 2x104 EA . hy926 or HUVEC cells/slide cultures were seeded in Lab-Tek II Chamber SlidesTM . Cells were treated with 1–20 μM simvastatin . After drug treatment , cells were fixed at room temperature for 10 min in 3 . 7% formaldehyde ( v/v ) in phosphate-buffered saline ( PBS ) for 10 minutes . Cytoskeleton was assessed using a F-Actin Visualization Biochem Kit ( Cytoskeleton , USA ) , following manufacturer’s instructions . Briefly , cells were washed with cytoskeletal buffer , followed by permeabilization for 5 min with permeabilization buffer and blocking with serum-containing buffer ( 3% FBS in PBS with 0 . 02% sodium azide ) . The cells were incubated with tetrarhodamine isothiocyanate ( TRITC ) -phalloidin for 30 minutes to stain cytoskeletal F-actin . DNA was stained with DAPI ( NucBlue; Molecular Probes , USA ) following the manufacturer's instructions . The cells were photographed using a Nikon Eclipse 400 fluorescence microscope . The leukocyte adhesion assay was performed with the Cytoselect Leukocyte-Endothelium Adhesion Assay kit ( CBA-210; Cell Biolabs , USA ) . Briefly , EA . hy926 cells were cultured in 96-well plates that had been previously coated with gelatin for 1 hour . Confluent monolayers were treated with 1–10 μM simvastatin or 1–20 μM benznidazole for 24 hours followed by T . cruzi infection for 16 hours . LeukoTracker-labeled leukocytes ( HL60 cells ) were added to the monolayer and incubated for 90 min . After thorough washing , cells were lysed , and fluorescence was measured at 480 nm excitation/520 nm emission . The percentage of adherent leukocytes was calculated: % adherence = adherent signal/total signal . All of the determinations were performed in triplicate using a fluorescence microplate reader ( Varioskan , Thermo Scientific ) . 8x105 EA . hy926 cells/well were cultured in 6-well plates ( IMDM 10% SFB ) . Then , the cells were treated with simvastatin 5 μM for 24 hours and infected at a MOI of 10 for 16 hours . 15-epi-LXA4 determination was performed by: 2x104 EA . hy926 cells were seeded in Lab-Tek II Chamber SlidesTM for 12 hours . Cells were fixed in 3 . 7% formaldehyde–0 . 1 M phosphate buffer ( pH 7 . 3 ) for 10 min . They were then washed with cytoskeletal buffer followed by permeabilization for 5 min with Triton-X100 0 . 5% . Cells were blocked with 3% bovine serum albumin for 1 hour . Then , cells were incubated with monoclonal antibodies against p65 ( 1:100 ) from cell signaling overnight at 4°C . The samples were washed with PBS and incubated with anti-rabbit IgG conjugated with fluorescein ( 1:100 ) from Sigma-Aldrich for 1 hour . Finally , nuclei were stained with DAPI for 5 minutes and mounted in Dako fluorescence mounting media . The cells were photographed and ten pictures per well were obtained; images were analyzed using ImageJ software ( ImageJ 1 . 47v ) . Statistical significance was established at p<0 . 05 . The results represent the mean ± SD of triplicates . Normal data distribution was assessed using D'Agostino-Pearsons and Shapiro-Wilk analysis . One- and two-way ANOVA analysis ( with Tukey’s or Bonferroni’s post-tests ) was performed when required . All of the statistical analyses were performed using GraphPad Prism ( 5 . 0 ) software . For the analysis of the effect of the combination of simvastatin and benznidazoles on ECAM expression , the combinatory index ( CI ) and isobolographic analysis was performed using CompuSyn software ( ComboSyn , Inc . Paramus , NJ ) in accordance with the Chou and Talalay’s principle [20] . The interaction between simvastatin and benznidazole , on E-Selectin , ICAM-1 and VCAM-1 was investigated by calculating the CI , where CI < 1 , CI = 1 and CI > 1 indicate synergism , additive and antagonism , respectively .
To determine the optimal time point of maximal expression of ECAMs in T . cruzi-infected endothelial cells , a kinetic pattern of expression was determined by flow cytometry . Fig 1 demonstrates the mean fluorescence intensity ( MFI ) and representative histograms for each ECAM analyzed as obtained after 48 and 72 hours of T . cruzi infection . E-Selectin , ICAM-1 and VCAM-1 surface expression on endothelial-like EA . hy926 ( Fig 1A and 1B ) and HUVEC ( Fig 1C and 1D ) cells increased in a time-dependent manner . However , the expression behavior was slightly different from each cell model . In EA . hy926 cells , after reaching their maximum expression at 16 hours , a steady state was attained with a slow trend to decreased expression without reaching control values ( Fig 1A ) . For HUVECs , there was a statistically significant increase at 16 hours for all three adhesion molecules ( two-way Anova with bonferroni post-test , p<0 . 05 ) . However , peak expression was reached at 48 hours and declined until 72 hours , and only a sustained ICAM-1 expression remained ( Fig 1C ) . Constitutive ICAM-1 expression could likely account for this observation [21] . Indeed , as demonstrated in Fig 1C and 1D , there is a slight deviation in the ICAM-1 fluorescence signal in both HUVEC and EA . hy926 uninfected cells . As a consequence of these results , peak ECAM expression was set at 16 hours for further assays . To evaluate the effect of simvastatin or benznidazole on ECAMs expression EA . hy926 and HUVEC cells were incubated with 5 μM simvastatin or 20 μM benznidazole for 24 hours and then were infected with T . cruzi trypomastigotes for 16 hours ( Fig 2 ) . The simvastatin and benznidazole concentrations used here provided the best effect on ECAM expression without cytotoxic effects on endothelial cells , either cytoskeletal alterations or on cell viability ( Table 1 , S2 and S3 Figs ) [22] ) . Nevertheless , even at concentrations as low as 1 μM , the decrease in ECAM expression was observed ( Fig 2A ) . Both drugs significantly prevented E-selectin , ICAM-1 , and VCAM-1 expression as evidenced by flow cytometry ( Fig 2A–2D ) and immunofluorescence ( Figs 2E and 3F ) ( Two-way ANOVA with Tukey post-test , p >0 . 01 ) . We expected that the effect of simvastatin on ECAMs would complement the trypanocidal action of benznidazole . Thus , studying their combination was necessary . To study the effect of the combination of simvastatin and benznidazole on E-selectin , ICAM-1 , and VCAM-1 expression , EA . hy926 and HUVEC cells were incubated with varying concentrations of both drugs , using the same procedure as for Fig 2 . For all three ECAM evaluated , at any combinatory point the CI values were >5 . Thus , accordingly to Chou and Talalay these values are indicative of antagonism [20] . A decrease in endothelial adhesion molecule expression , as a result of simvastatin or benznidazole administration , might be explained by intracellular sequestration of these molecules . If this is the case , total protein levels would be similar to those of uninfected cells . In Fig 3 , total E-selectin ( Fig 3A ) , ICAM-1 ( Fig 3B ) and VCAM-1 ( Fig 3C ) protein expression analyses are shown . When compared with infected , untreated cells , a slight decrease in total E-selectin protein expression was observed . This effect was more evident and significant with VCAM-1 and ICAM-1 ( One-way ANOVA with Tukey post-test , p<0 . 05 ) , where total protein content returned to values similar to uninfected controls . Thus , simvastatin or benznidazole administration prevented T . cruzi-triggered activation of intracellular mechanisms , which explains the increased ECAM expression on the surface of EA . hy926 cells . Conversely , the lack of response in adhesion molecule expression could obey eventually to a trypanocidal effect of simvastatin or benznidazole . Therefore , the burden of intracellular amastigotes in simvastatin or benznidazole-treated and subsequently T . cruzi-infected EA . hy926 cells was determined 72 hours after establishing the infection . To evaluate parasite load , DAPI-stained cells were photographed ( 20 . 12 ± 3 . 8 cells per photograph , 20 photographs per experimental condition; infected to uninfected cells ratio of 1:5 ) and the number of intracellular amastigotes was determined by manual count . As demonstrated in Fig 4 and S4 Fig , intracellular amastigote content is not influenced by drug pretreatment . Administration of simvastatin and benznidazole prior to T . cruzi infection prevented endothelial activation and the subsequent increase in adhesion molecule expression . Hence , it is important to clarify whether these drugs affect cell adherence . Consequently , an adhesion assay was performed using HL-60 leukocytes loaded with Leuko Tracker , which were co-incubated with T . cruzi-infected endothelial cells . The results of the analysis are demonstrated in Fig 5 . Indeed , when leukocyte adhesiveness was assessed , there was a significant increase in cell adhesion to the T . cruzi-infected endothelial cells ( One-way ANOVA and Tukey post-test , p<0 . 05 ) . However , upon simvastatin or benznidazole treatment the adhesiveness decreased and reached similar values compared to those observed in uninfected cells . This finding is important because it links the decreased adhesion molecule expression with the physiological consequence of efficiently reducing leukocyte adhesion . NF-κB pathway is involved is several inflammatory events , including sepsis , where endothelial activation is induced [23] . In addition , statins decrease endothelial inflammation as part of their pleiotropic effects . Thus , simvastatin could decrease ECAM expression by affecting NF-κB pathway . To evaluate the effect of simvastatin and benznidazole upon NF-κB pathway activation , we assessed total and phosphorylated IKK and IkB protein levels by Western blot analysis ( Fig 6 ) . After 60 minutes of incubating endothelial cells with T . cruzi trypomastigotes , the respective phosphorylated IKK form increased significantly compared with non-stimulated cells ( Fig 6A ) ( One-way ANOVA and Tukey post-test , p<0 . 05 ) . Similarly , p-IkB increased at the same time point ( Fig 6B ) . When the endothelial cells were incubated with simvastatin or benznidazole after previous T . cruzi incubation , these two drugs significantly decreased the response of the Ikk-IkB system ( Fig 6C and 6D ) ( One-way ANOVA and Tukey post-test , p<0 . 05 ) . Thus , NF-κB is modulated by simvastatin and benznidazole , which prevents the activation of these two essential NF-κB pathway proteins during a challenge with the parasite . This is demonstrated by decreased nuclear p65 localization when infected endothelial cells are previously incubated with simvastatin or benznidazole ( Fig 6E ) . Aside from nuclear p65 localization after T . cruzi challenge , cytosolic p65 levels were also increased compared with uninfected controls . Among the pleiotropic effects of statins , their anti-inflammatory actions include the induction of pro-resolutive , anti-inflammatory molecules such as 15-epi-LXA4 , which may mediate the decreased leukocyte adhesion [24] . Indeed , simvastatin induced 15-epi-LXA4 in T . cruzi-infected endothelial cells . In Fig 7A and Table 2 , 5 μM simvastatin was added after incubation of the EA-hy926 cells with T . cruzi trypomastigotes for 16 hours . 15-epi-LXA4 levels rose slightly with simvastatin alone ( Fig 7A ) ; while the T . cruzi infection progressed , there was a significant increase in 15-epi-LXA4 levels ( Fig 7A and Table 1 ) ( One-way ANOVA and Tukey post-test , p<0 . 001 ) . In this experiment , the model that had been used so far was modified by adding drug treatment after establishment of the infection . Notwithstanding , the most important fact is that the simvastatin did not induce the production of this eicosanoid in the absence of an activating factor of endothelial cells such as the parasite . In contrast , when 5-lipoxygenase activity is inhibited in the presence of a competitive inhibitor ( AA-861 ) [25] , ICAM expression was restored ( Fig 7B ) . Although AA-861 concentration appears high ( 50 μM ) , there are several reports that used this inhibitor at concentrations as high as 100 μM , without reporting off-target effects [26–29] . Thus , there is a connection between the increased 15-epi-LXA4 and the action of 5-lipooxygenase when simvastatin was administered . The addition of exogenous 15-epi-LXA4 before endothelial cell infection , as has been done in the other experiments with simvastatin , decreased IKK-IκB pathway activity in a dose-dependent fashion ( Fig 8A–8C ) . Furthermore , 100 nM 15-epi-LXA4 at the highest concentration used decreased nuclear p65 migration ( Fig 8D ) . Thus , NF-κB pathway activity is decreased in a similar manner as simvastatin . In fact , 15-epi-LXA4 similarly reduced endothelial adhesion molecule expression ( Fig 8E ) .
In Chagas cardiomyopathy , there is an inflammatory vasculopathy , which is demonstrated by sustained expression of the adhesion molecules E-selectin , VCAM-1 and ICAM-1 during T . cruzi infection . ICAM-1 is expressed at a low level constitutively on the cell surface . ECAM expression increased significantly only after inflammatory stimuli such as LPS , TNF-α [30 , 31] or intracellular infection with T . cruzi [32] , which tends to be sustained over time . Furthermore , according to the results reported here , both simvastatin and benznidazole prevented the increased expression of these molecules as T . cruzi infection was installed . Indeed , decreased expression of CAMs had functional consequences . Leukocyte adhesion was also decreased by both simvastatin and benznidazole . Albeit this functional decrease in cell adhesion is more evident with simvastatin , the most striking result is the action of benznidazole . The reduction in CAM expression and , consequently , cell adhesion was independent of the trypanocidal activity of benznidazole . Indeed , the anti-inflammatory effect of benznidazole had already been reported [33] . In experimental sepsis models , benznidazole attenuated NF-κB and the MAPK pathway activities , highlighting its immunomodulatory capacity [34 , 35] . Undoubtedly , these models did not attend the antiparasitic capacity of benznidazole . Neither did our experiments . Herein , we report its ability to prevent endothelial activation during T . cruzi infection for the first time . That is a critical factor to support the usefulness of this drug for treatment during the chronic phase of CD . In our experimental model , simvastatin was administered before producing infection with T . cruzi . Then , the culture medium was changed to fresh medium without drug and incubated the endothelial cells with the parasite for 16 hours . Therefore , it is unlikely that the drug presents trypanocidal activity because it was not present when the infection occurred . The idea of this model is to cause an environment where the infection cannot be spread , due to the action of simvastatin on endothelium activation . However , simvastatin and other 3-hydroxy-3-methylglutaryl-coenzyme A ( HMGCoA reductase ) inhibitors can inhibit parasite growth , especially in the replicative forms . In the parasite , this enzyme is essential for mevalonate and ergosterol synthesis , and is inhibited by statins , affecting parasite viability . In the work reported by Silva et al . [36] , simvastatin decreased epimastigote proliferation but at concentrations in the millimolar range . Apart from this observation , our results support the findings of the mentioned report since the reduction of endothelial activation is only part of the broad anti-inflammatory effect of simvastatin in that murine model of acute CD . Previously , we reported that benznidazole prevented endothelial damage in a murine model of chronic Chagas heart disease [6] . However , the effect of simvastatin on cellular models of T . cruzi-induced endothelial activation was not yet studied . It was interesting to find that cytosolic p65 levels were increased after T . cruzi challenge . This finding suggests that this protein could be upregulated during an inflammatory drive . However , the most outstanding result is that both simvastatin and benznidazole decreased NF-κB pathway activation through IKK-IκB pathway inactivation and decreasing nuclear migration of p65 . Nonetheless , a small fraction of p65 remained in the nuclei of benznidazole treated cells , and to a lesser extent , in cells that had been incubated with simvastatin . Thus , despite drug treatment , an inflammatory drive might persist due to this low-grade NF-κB activation . These findings are supported by previous reports in other experimental models [34 , 35 , 37] . Furthermore , according to our results , the effect of simvastatin on this signal transduction pathway and finally on adhesion molecule expression took place through 15-epi-lipoxin A4 generation . This eicosanoid decreased p65 migration towards endothelial cell nuclei similar to simvastatin . However , benznidazole did not induce 15-epi-lipoxin A4 production ( Fig 7A ) ; therefore , the mechanism of IKK-IκB pathway inhibition could be addressed through another yet unknown mechanism . Considering that simvastatin and benzidazole share similar effects on ECAM expression and NF-κB pathway , it is possible to think that their combined effect could be synergistic . Unexpectedly , the effect was rather antagonistic . The explanations for antagonism are diverse . It is possible that two drugs acting on the same target may behave as antagonistic [38 , 39] . Thus , our findings should not be surprising since both drugs interfere in the same signaling pathway , NFκB pathway . In any case , it is necessary to consider that both drugs act within a complex network of biological functions . Thus , it is not easy to conclude , based on an in vitro model , if this antagonism is important in a living model of CD . In any case , this interaction should not invalidate our findings for benznidazole . In this report , we confirm the involvement of benznidazole in the NFkB pathway , and through this pathway , its modulation of the expression of ECAMs in endothelial cells infected with T . cruzi , regardless of its trypanocidal capacity . The role of pro-resolving lipids in T . cruzi-induced inflammatory processes is becoming increasingly important to understand chagasic cardiomyopathy [40 , 41] . The role of LTB4 and PAF as produced by macrophages has already been reported to control parasitemia in in vivo CD models [40 , 42] . LTB4 , which is dependent on 5-LO activity , is involved in the decrease in inflammation , collagen deposition and lymphocyte migration to the myocardium [40] . In addition , 5-LO derivatives are increasingly associated with acute inflammatory process resolution [43–46] . Thus , these mediators could be more involved in the acute phase of Chagas disease . However , in a chronic model of Chagas cardiomyopathy , simvastatin decreased inflammation [11] . Most likely , simvastatin , by inducing 15-epi-LXA4 , prevented leukocyte migration into the myocardium by decreasing endothelial activation [44] , thus contributing to reduced myocardial damage . In any case , it is important to verify this hypothesis in future studies using in vivo CCC models . It is also important to consider the immunomodulatory and anti-inflammatory role of benznidazole in Chagas cardiomyopathy progression . In conclusion , simvastatin and benznidazole prevent endothelial activation , as demonstrated by decreased expression of the adhesion molecules E-selectin , ICAM-1 , and VCAM-1 , which involves decreasing NF-κB pathway activity , and at least for the case of simvastatin , by increased 15-epi-LXA4 production . | Chagas disease , caused by the protozoan Trypanosoma cruzi , affects more than 10 million people in Latin America . In the chronic phase , a lethal complication may develop: Chronic Chagasic Cardiomyopathy . In this condition the vascular lining , the endothelium , is involved and participates in disease progression . Benznidazole is the current treatment for Chagas disease . However , other useful drugs could be added to Chagas disease chemotherapy to improve associated processes such as endothelial dysfunction , thus decreasing length and adverse events of the conventional therapy . Simvastatin , a drug that decreases blood cholesterol , also has anti-inflammatory effects and improves endothelial function . Thus , we studied the effect of simvastatin and benznidazole on endothelial cells and their relation with the production 15-epi-lipoxin A4 , an anti-inflammatory molecule . We found that simvastatin and benznidazole decreased endothelial activation since they reduced the adhesion of inflammatory cells . Simvastatin and benznidazole inhibited NFκB pathway , which is pro-inflammatory and in the case of simvastatin , this effect was mediated by the production of 15-epi-lipoxin A4 . Thus , we provide the bases that support the future use of simvastatin in the treatment of cardiac Chagas disease . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Simvastatin and Benznidazole-Mediated Prevention of Trypanosoma cruzi-Induced Endothelial Activation: Role of 15-epi-lipoxin A4 in the Action of Simvastatin |
Over long timescales , neuronal dynamics can be robust to quite large perturbations , such as changes in white matter connectivity and grey matter structure through processes including learning , aging , development and certain disease processes . One possible explanation is that robust dynamics are facilitated by homeostatic mechanisms that can dynamically rebalance brain networks . In this study , we simulate a cortical brain network using the Wilson-Cowan neural mass model with conduction delays and noise , and use inhibitory synaptic plasticity ( ISP ) to dynamically achieve a spatially local balance between excitation and inhibition . Using MEG data from 55 subjects we find that ISP enables us to simultaneously achieve high correlation with multiple measures of functional connectivity , including amplitude envelope correlation and phase locking . Further , we find that ISP successfully achieves local E/I balance , and can consistently predict the functional connectivity computed from real MEG data , for a much wider range of model parameters than is possible with a model without ISP .
Healthy resting brain dynamics exhibit several characteristic spatiotemporal features , including structured functional connectivity between brain regions detectable by a variety of different measures . The mechanisms that give rise to this functional connectivity are still under investigation , and large-scale biophysical models offer a parsimonious way to mechanistically explore how brain structure and neuronal properties give rise to functional connectivity . There has been considerable interest in using coupled networks of oscillators to relate large-scale functional connectivity to network properties including connectivity strength , time delays , and graph structure . These oscillator models span very simple oscillators such as the Kuramoto model [1–7] , through to more sophisticated oscillators based on Hopf bifurcations [8 , 9] , and biophysical neural mass models such as the Wilson-Cowan model [10–15] . Functional connectivity in these models is typically measured using amplitude envelope correlations in MEG data , or in fMRI either by simulating the model on slow BOLD timescales [13 , 16] , convolving model predictions using a hemodynamic response function [17–19] , or incorporating a hemodynamic model ( e . g . , Balloon-Windkessel ) [20–23] . Achieving realistic brain activity in biophysical models typically requires extremely fine tuning of parameters [24] . This contrasts with real brains , whose dynamics can be robust to quite large changes in white matter connectivity and grey matter structure , whether through learning , aging , development or disease [25] . There is increasing evidence that a fine balance between excitation and inhibition underpins a wide range of dynamics found in the brain [23 , 26–32] . However , how is this balance maintained when the brain is perturbed by changes that disrupt this balance ? One possible explanation is that there are homeostatic mechanisms that continuously adjust brain networks to achieve the necessary balance . The balance between excitation and inhibition is often framed in terms of correlations in excitatory and inhibitory activity when analysing electrophysiological data , but it could equally be framed in terms of excitatory and inhibitory connection strengths at the connectome level . Inhibitory synaptic plasticity ( ISP ) , where the strength of local inhibitory connections changes depending on whether excitatory activity is above or below a target level of activity , is a parsimonious , biophysically plausible mechanism that directly modulates the balance between excitation and inhibition [33–36] . ISP can be readily implemented in simple biophysical models [27 , 37] , and integration of ISP into a whole-brain neural mass model has been previously demonstrated with neural dynamics occurring on fMRI timescales [13] . However , the effects of ISP on electrophysiological timescales and the robustness of ISP regarding conductance delays are unexplored . Models on electrophysiological timescales have complex fast dynamics that could affect the outcome of plasticity on long timescales . The aim of this study is to use a biophysical model to investigate whether ISP is successfully able to balance excitation and inhibition for models with ongoing oscillations on electrophysiological timescales , and to investigate the effect that ISP has on population-level dynamics and functional connectivity measured using MEG . For this purpose , we selected the well-established Wilson-Cowan neural mass model , which includes both excitatory and inhibitory populations and can be readily extended to incorporate ISP . By simulating neural activity on electrophysiological timescales and including conduction delays between brain regions , we investigate the effect of ISP on functional connectivity metrics typically used in MEG , using both amplitude envelope correlations and phase based measures of connectivity .
All participants gave written informed consent and ethical approval was granted by the University of Nottingham Medical School Research Ethics Committee . We simulate neural activity using a network of neural masses , using the Wilson-Cowan model [10] . We divide the cortex into regions based on a grey matter parcellation , and model the dynamics in each brain region by an excitatory and inhibitory population of neurons ( a unit ) , connected as shown in Fig 1A . Long-range white matter connections link excitatory populations across different brain regions with distance-dependent propagation delays . In accordance with previous work using this model [12 , 13 , 15] , inhibitory connections are purely local , although long-range inhibition can be readily included in the future in the same way as long-range excitation . Activity in the neural populations is governed by the equations [12 , 13]: τedEk ( t ) dt=−Ek ( t ) +S ( ceeEk ( t ) +ciek ( t ) Ik ( t ) +P+ξ ( t ) +C∑j=1nWjkEj ( t−τjk ) ) , ( 1 ) τidIk ( t ) dt=−Ik ( t ) +S ( ceiEk ( t ) +ξ ( t ) ) , ( 2 ) where Ek and Ik are the mean firing rates in the excitatory and inhibitory populations in brain region k , τe and τi are excitatory/inhibitory time constants , cab is the local connection strength from population a to population b , P is constant external excitatory input , and ξ ( t ) is a noisy input signal added into the dynamics . Long range white-matter connections Wjk from region j to region k incorporate a time delay τjk , and are multiplied by a global coupling scaling C . Indices k and j range over the number of brain regions in the parcellation , n . The nonlinear response function S is a sigmoid function given by S ( x ) =11+e−x−μσ , ( 3 ) where μ and σ represent the mean firing threshold and variation in threshold for neurons in each population . Note that the numerator of Eq ( 3 ) corresponds to the maximum firing rate of the neural populations , so all firing rates here are represented as a non-dimensional fraction of the maximum firing rate . Actual firing rates could be simulated by multiplying the sigmoid by a dimensional constant , and accordingly dividing the connection strengths by the same constant . This changes the units of the state variables , but does not affect their dynamics with regard to the time constants , delays , or any of our measures of functional connectivity . Without ISP , the local inhibitory coupling strength is constant and identical for all brain regions k , with ciek ( t ) =cie ( 0 ) . To implement ISP , we use the framework developed by Vogels et al . [27] in which ISP is modelled as a spike-timing-dependent process dependent on both pre- and postsynaptic activity . The local inhibitory connection strength changes depending on the activity of the corresponding local excitatory population , according to [13 , 27] τispdciek ( t ) dt=Ik ( t ) ( Ek ( t ) −ρ ) , ( 4 ) where τisp is the learning rate , and ρ is the target excitatory activity level , and the initial value ciek ( 0 ) =cie ( 0 ) . As we are focused on slow plasticity where plasticity is decoupled from fast neural activity [38 , 39] , the primary constraint on τisp is that it is large enough to ensure separation of timescales between Eqs ( 1 and 2 ) and Eq ( 4 ) . Once τisp is sufficiently large , further increasing it will change how long the system needs to be simulated until plasticity ceases , but will not change the dynamics of the system once that point is reached . The fact that we only include long-range E-E connections results in an initial imbalance in the network which ISP acts against . In the real brain , the network would already be close to balanced , and ISP would instead serve to oppose more subtle changes , such as those introduced by other plasticity mechanisms . The parameter values used in this study are listed Table 1 and are based on those used in previous work by Deco et al . [12] . The parameters have been rescaled such that the maximum value of the sigmoid response is 1 and μ = 1 for simplicity , but this does not alter the dynamics of the system . The inhibitory time constant is larger than the excitatory time constant , consistent with previous studies [40 , 41] and cellular measurements in rodent [42] . For these parameters , an isolated unit transitions from a stable steady state to oscillations as the value of P increases beyond 0 . 34 ( a Hopf bifurcation ) . The value of P at which the Hopf bifurcation occurs depends on the ratio τe/τi , while the frequency can be changed by rescaling both time constants in proportion . The parameters in Table 1 result in intrinsic oscillations at ~11 Hz , which is suitable for investigating functional connectivity in typical MEG frequency bands . In accordance with previous work [12] , we set P = 0 . 31 such that the individual units are just below their oscillatory threshold , and large oscillations in the network arise only when long-range coupling is included . We set the nominal ISP target level of activity to ρ = 0 . 15 , which is a relatively low level of activity that corresponds to an isolated unit being in an oscillatory state . In this study , we simulated a cortical brain network using the Desikan-Killiany parcellation [43] with 68 brain cortical regions covering the brain cortex bilaterally . Structural connectivity weights between these regions were estimated using diffusion MRI probabilistic tractography and data from the Human Connectome Project [44–46] . Briefly , fibre orientations were estimated from distortion-corrected data [47] using a model-based spherical deconvolution approach , as implemented in FSL [48 , 49] . Up to three fibre orientations were detected per white matter voxel and were used for probabilistic tractography , which was performed using FSL’s probtrackx2 ( https://fsl . fmrib . ox . ac . uk/fsl/fslwiki/FDT ) [50] . The white/grey matter boundary surface was used as a seed , since this reduces biases observed using whole-brain seeding [51 , 52] . Streamlines were seeded from N = 60 , 000 standard-space vertices [46] on the boundary surface ( 10 , 000 streamlines per seed ) . Anatomical constraints were imposed to reduce false positives . Specifically , we allowed streamlines to hit the white matter/grey matter boundary not more than twice , and also streamlines were allowed to enter subcortical volumes , propagate within them , but terminate upon exit , as suggested in [53] . The pial surface was further used as a termination mask to ensure estimated paths do not “jump” between neighbouring gyri . The number of streamlines reaching each vertex in the WM/GM boundary was recorded , and this was normalised by the total number of valid streamlines propagated , giving a dense NxN “connectivity” matrix . Using the cortical parcellation , this matrix was reduced to a 68 x 68 parcellated connectivity matrix , by computing for each pair of regions the mean connectivity between all pairs of vertices they were comprised of . Forty subjects were processed and their resulting connectivity matrices were averaged . The connection matrix was log-transformed to account for algorithmic bias in the tractography [54] , and normalised by dividing by the largest value , yielding the structural connectivity used in the simulations shown in Fig 1B and 1C . Propagation delays between brain regions were approximated using the barycentric distances between regions ( shown in Fig 1D ) , with a uniform conduction velocity that was varied as one of the model parameters . To generate model predictions comparable to MEG , we integrated the system with a sufficiently small time step ( Δt = 1 × 10−4 s ) using a 4th order Runge-Kutta scheme for 500 seconds . The noise signal ξ ( t ) was sampled from a Gaussian distribution ( zero mean , standard deviation 0 . 01 ) in each population at each time step , and was interpolated for the intermediate steps in the integration ( noting that this means our noise input is not white noise ) . The initial conditions for the system were randomized , and the first 15 seconds of the simulation were discarded to remove transient effects due to the initial conditions . The simulated output was downsampled to 300Hz to reduce storage requirements for comparison to typical MEG data . In accordance with previous studies , we include a small amount of noise to improve the robustness of the simulation [15] . Without noise , we find that the system is prone to becoming trapped in highly periodic oscillatory states that are not physiologically realistic . However , these states are only marginally stable , such that even a very small amount of noise is sufficient for the system to escape the attractor . Thus these states are also not biologically plausible , because the ubiquitous external perturbations and sources of noise in the brain , e . g . thermal noise , synaptic transmission failures , would prevent the system from remaining in such a state [55 , 56] . We select the noise amplitude such that it is much smaller than the nonlinear oscillations that drive functional connectivity in the model . Our results therefore primarily reflect the nonlinear interactions between populations rather than noise , and our results are qualitatively robust to moderate variation of the step size and noise amplitude ( as shown in Fig A2 in S1 Supplementary Material ) . To simulate ISP , we integrated the system for 1500 seconds . To improve computational tractability , we accelerate ISP in the initial portion of the simulation . For the first 500 seconds we set τisp = 2 . 5s , the next 500 seconds τisp = 10s , and the final 500 seconds τisp = 20s . This procedure strikes a balance between ensuring that ISP converges in a computationally feasible amount of time , while also ensuring that the timescale differences are sufficient to decouple the neural oscillations from ISP . At the end of this period , ISP was disabled , and the simulation run for a further 500 seconds . As we do not expect changes in synaptic strength to drive fast neuronal dynamics over short periods of time and are focused on the dynamics of the balanced networks that are the end result of ISP , we turned off ISP for the final run to ensure that our measures of activity and stationary functional connectivity are not affected by ongoing changes in synaptic strength [13] . Synchrony and variability in synchrony ( metastability ) are common global metrics of activity for coupled oscillator models . In particular , variability in synchrony is associated with realistic brain activity [1 , 57 , 58] . The synchrony order parameter R ( t ) is obtained by summing over the analytic phase of each oscillator in the network [1 , 4 , 57] R ( t ) eiθ ( t ) =1N∑k=1Neiϕk ( t ) ( 5 ) This order parameter is 1 if the oscillators are completely synchronized ( their phases are the same ) , and 0 if their phases are uniformly randomly distributed . For phase oscillators such as the Kuramoto model , the phase of the oscillator is well defined because it is simply the state variable of the model . In contrast , for the Wilson-Cowan model it is necessary to use a technique like the Hilbert transform to estimate phase . The analytic signal is derived from the original signal using Xa ( t ) =X ( t ) +iH ( t ) where Xa is the analytic signal , X is the original signal , and H is the Hilbert transform of X . Writing this in polar form gives Xa ( t ) =A ( t ) eiϕ ( t ) where A ( t ) is the amplitude envelope timecourse , and ϕ ( t ) is the phase timecourse . This phase is only readily interpretable for narrowband signals , which means the oscillatory activity needs to be filtered prior to estimating synchrony . We used a 4th order , two pass Butterworth filter implemented by FieldTrip [59] . We used eyes open resting state MEG data from 55 healthy controls ( ages 18–48 , mean age 26 . 5 , 35 males ) , acquired at the University of Nottingham as part of the UK MEG Partnership . Data were acquired with a 275-channel axial gradiometer CTF MEG system ( MISL , Conquitlam , Canada ) at a sampling rate of 1200 Hz , and downsampled to 600Hz with a 300 Hz low pass anti-aliasing filter . Synthetic third order gradiometer correction was applied to reduce external interference . Subjects were seated in the scanner and presented with a fixation target while 300 seconds of data were recorded . Head position was continuously tracked using three head position indicator ( HPI ) coils , placed at the nasion and left and right preauricular points . Structural MRI scans for each subject were acquired at 0 . 8mm isotropic resolution using a Philips 7T Achieva MRI scanner running a phase-sensitive inversion recovery sequence [60] . To coregister the MEG system geometry to the subject’s anatomical MRI , a Polhemus FASTRAK 3D digitiser system was used to record the position of the fiducial points and the subject’s head shape . The locations of the MEG sensors relative to individual brain anatomy were determined by registering the digitized head surface to the structural MRI . The structural MRI was registered to the MNI152 standard brain and all source space analysis was then performed in MNI space . For analysis , the data were imported into SPM12 format , and downsampled to 250 Hz using an anti-aliasing low-pass filter . High-pass filtering ( cut-off 0 . 1 Hz ) and a notch filter were subsequently applied to attenuate slow drifts and line noise . For outlier detection , resting state data were epoched into pseudo-trials of 2s length , and the signal standard deviation was estimated once per trial . Subsequently these estimates were subject to a robust fit using the bisquare distribution . Extreme trials were identified by a regression coefficient smaller than 0 . 05 and were discarded . The data were visually inspected for remaining artefacts . The data were bandpass filtered from 1-45Hz , and then beamformed onto an 8mm grid [61] . For each brain region , a single activity timecourse was computed as the first principal component of the activity in voxels belonging to that region [62] . To compensate for spatial leakage ( which can affect certain connectivity metrics ) , we use the symmetric multivariate orthogonalisation algorithm developed by Colclough et al . [62] . In summary , this multivariate algorithm removes all zero-lag correlations between all parcel timecourses simultaneously , by projecting them onto a new orthogonal basis . We also apply the orthogonalisation procedure to the model when computing connectivity metrics that require leakage correction in MEG data , to ensure that the underlying activity in the model is consistent with experimental data when processed through the same analysis pipeline . That is , the model should generate predictions comparable to experimental data when zero lag correlations are removed , even though there is no spatial leakage in the model , because genuine zero lag correlations have also been removed in the data . We analyse static functional connectivity in the data and in the model using three metrics–amplitude envelope correlation ( AEC ) , phase locking value ( PLV ) , and phase lag index ( PLI ) [63] . Of these , AEC and PLV are affected by spatial leakage and require signal orthogonalisation , while PLI is insensitive to spatial leakage and does not require orthogonalisation . We include PLI in addition to AEC and PLV to ensure the model does not rely on the orthogonalisation procedure to produce connectivity similar to experimental data . To compute the amplitude envelope correlation , analytic signal corresponding to the orthogonalised parcel timecourses is computed , and then the amplitude envelope component is extracted and downsampled to 1Hz [64–67] . The connectivity matrix is obtained by computing the Pearson correlation between every pair of the downsampled envelope timecourses . AEC requires orthogonalisation because additive mixing of the parcel timecourses due to spatial leakage introduces zero-lag correlations in the raw signals that consequently result in correlations in the envelope timecourses . Removing all zero-lag correlations implies that any remaining envelope correlations are due to mechanisms other than spatial leakage . The phase locking value [68] is calculated from the phase component of the analytic signal , again computed using the orthogonalised parcel timecourses [63] . PLV is a pairwise measure of synchronization between two signals , given by PLVij=|〈ei ( ϕi ( t ) −ϕj ( t ) 〉| If the phase difference is constant over time , then the PLV will approach 1 , whereas if the phases of the two signals are random relative to each other , then the PLV will be 0 . PLV requires orthogonalisation because additive mixing of the parcel timecourses introduces a constant phase relationship between them , artificially increasing the PLV . The phase lag index [62 , 69] aims to quantify asymmetry in the distribution of the phase difference between two signals . It is calculated from the phase component of the analytic signal using the raw ( non-orthogonalized ) parcel timecourses: PLIij=|〈sign ( sin ( ϕi ( t ) −ϕj ( t ) ) 〉| PLI does not require signal orthogonalization because while spatial leakage decreases the phase difference between the signals , it does not make the distribution of phase differences less symmetric . The overall analysis pipeline from simulated or measured data to connectivity estimates is shown in Fig 2 . As excitatory pyramidal cells contribute most strongly to EEG/MEG signals , we associate activity in the excitatory populations of the model with signals in experimental data [70–72] , although future work may wish to model contributions from both excitatory and inhibitory populations by modelling postsynaptic currents in more detail .
To assess the influence of ISP in the model behaviour , we quantified similarity between connectivity matrices by computing the correlation coefficient for the upper triangular part of each ( symmetric ) connectivity matrix for different scenarios . As shown in Fig 4A , without ISP the model exhibits high similarity with real group average AEC and PLV over a narrow range of coupling strengths , and with intermediate delays ( 5-15ms ) . The functional connectivity estimated using PLI is acceptable , but not optimal in this same regime–instead , the best fit with experimental data is obtained at somewhat shorter delays , and slightly stronger coupling strengths . With the inclusion of ISP , the model can reproduce alpha band functional connectivity over a wide range of global coupling strengths , as shown in Fig 4B . We find that the range of delays over which realistic functional connectivity is exhibited is relatively unchanged by the addition of ISP ( when collapsed over coupling strength ) , while sensitivity to long-range coupling is greatly reduced , as expected . Notably , in the model with ISP , the optimal parameter regime for matching experimental data is the same for all three functional connectivity metrics . This demonstrates that it is possible to select a single operating point that matches functional connectivity in both amplitude and phase connectivity measures , and regardless of whether spatial leakage correction is applied or not . The extent to which the model can predict real group average functional connectivity is limited by the fact that we use an average structural connectivity matrix , rather than averaging functional connectivity over individual connectomes ( notwithstanding the uncertainties in estimation of the connectomes themselves ) . However , the magnitude of the similarity between the model and data shown in Fig 4A and 4B is difficult to interpret without a point of comparison . The correlation between the matrices shown in Fig 5 is 0 . 48 for AEC , 0 . 43 for PLV , and 0 . 28 for PLI–does this indicate that the model is performing well ? Given that we are comparing functional connectivity estimated from a single connectome to the group average , a suitable point of comparison is the typical level of similarity observed between real individual’s functional connectivity profile and the real group average . To estimate this , we computed the correlation between each real subject’s functional connectivity and the real group average connectivity computed with that subject left out . The similarity in connectivity between real individuals and the real group average was 0 . 60 ± 0 . 17 for AEC , 0 . 29 ± 0 . 09 for PLV , and 0 . 22 ± 0 . 06 for PLI . We can then test whether the model’s performance is significantly different by computing the difference between model similarity and mean individual similarity , and then dividing by the standard deviation of the individual similarity . This converts the model similarity to a Z-score based on individual variability , shown in Fig 4C . The null hypothesis that the model’s similarity comes from the same distribution is then rejected if |Z| > 1 . 96 . For the optimal parameter regime , |Z| < 1 . 96 in all three connectivity measures , which indicates that our observed model performance is not significantly different to the typical similarity observed in real individuals . The functional connectivity profiles for a representative set of parameters in the optimal parameter regime ( the red dot in Fig 4 ) are shown in Fig 5 . In general , the model predicts stronger functional connectivity than is seen in the group average data , although we note that real individual functional connectivity can also be considerably stronger than the group average . To verify that this is not a consequence of the low level of noise included in the simulation , but is an intrinsic property of the nonlinear dynamics of the system , we repeated the simulation shown in Fig 5 for a range of different noise amplitudes ( as shown in Fig A2 in S1 Supplementary Material ) . We find that increasing the noise by a factor of 10 provides qualitatively similar results , including the magnitude of the connectivity metrics . This robustness partly reflects the fact that the response time of the system limits the effect of the noise , acting as a low-pass filter that averages out the noise over short periods of time . The time-averaged synchrony ( mean value of R ( t ) as defined in Eq 5 ) and metastability ( standard deviation of R ( t ) ) are shown in Fig 6 . Compared to previous work using coupled Kuramoto phase oscillators [1] , the Wilson-Cowan model is more sensitive to network coupling strength because the amplitude of oscillations in neural activity can vary . Networks of Kuramoto oscillators exhibit highly synchronized oscillations when strongly coupled , but this is not the case for the Wilson-Cowan model , as shown in Fig 6A and 6B . Without ISP , the model exhibits high synchrony over a narrow range of couplings , but a wide range of delays . As coupling becomes stronger , synchrony drops markedly . As coupling strength is increased further , the global synchrony rises again . This occurs because strong coupling causes the most strongly connected neural populations to enter a high-activity stable steady state . In this regime , it is difficult to interpret the global synchrony because different brain regions exhibit qualitatively different dynamics . When coupling increases further ( for our parameters , when global coupling is stronger than 0 . 21 ) , all the units converge to a high activity steady state with noise-driven fluctuations in activity and low synchrony . However , when excitation and inhibition are locally balanced , the global network dynamics become much more similar to those seen with simple phase oscillators . We recover the same regime of high synchrony occurring at progressively longer delays as coupling strength increases , as well as the existence of a metastable regime with high variability in synchrony at the interface between the high and low synchrony regimes of the model ( Fig 6C and 6D ) . While synchrony and metastability are useful to analyse the model dynamics , they are not commonly used to study MEG data because they are extremely coarse measures that aggregate activity over the entire brain at each time point . As these measures are susceptible to spatial leakage , comparison to data must include a leakage correction step . We show synchrony and metastability in the model including orthogonalisation in Fig A3 in S1 Supplementary Material . In summary , in the optimal parameter regime of the model , the predicted synchrony and metastability are both larger than in data by a factor of 2–3 , although the effect of measurement noise in the data is not accounted for in the model . We hypothesised that ISP would serve as a mechanism to balance excitation and inhibition in the network . This balance is inherently disrupted because we only include long-range excitatory connections . Balancing excitation and inhibition therefore requires that inhibition is increased more for brain regions that have stronger long-range couplings . To quantify whether such a balance has been achieved , we can examine the correlation between the node strength ( row-sum of the structural connectivity matrix , which corresponds to the sum of long-range excitatory inputs ) for each brain region , and the local inhibitory synaptic strength after ISP . Note that without ISP , the inhibitory synaptic strengths are independent of node strength and they are therefore uncorrelated . As shown in Fig 7A , for most delay/coupling parameter combinations , ISP successfully balances excitation and inhibition , with a strong correlation between long range excitation and local inhibition . The relationship is particularly clear in the low synchrony regime , either with weak coupling or large delays . In the high synchrony regime , ISP is somewhat less successful at balancing excitation and inhibition , and the E/I balance shown in Fig 7A becomes a snapshot of ongoing changes in inhibition ( as shown in Fig A5 in S1 Supplementary Material ) . Fig 7B shows the clear relationship between excitation and inhibition in the metastable regime , where realistic functional connectivity is achieved . To investigate the sensitivity of our findings to the choice of target activity level ρ , we repeated our simulations for ρ = 0 . 10 and ρ = 0 . 30 corresponding to a low-activity target and a high-activity target , respectively . For an isolated unit , at the Hopf bifurcation the mean excitatory activity level is 0 . 12 , which means that a target of ρ = 0 . 10 is low enough that for low coupling values , the units can be in a low-activity , noise-driven steady state rather than undergoing nonlinear oscillations . In contrast , all tested parameters provide nonlinear oscillations for ρ = 0 . 15 and ρ = 0 . 30 . As shown in Fig 8 , with ρ = 0 . 10 we did not find realistic functional connectivity for any of the tested delays and couplings . The low ISP target gives rise to two distinct regimes ( as shown in Fig A4 in S1 Supplementary Material ) –a stochastic regime at low coupling strength , and a synchronous oscillatory regime at high coupling strength . In the stochastic regime , realistic functional connectivity can be produced when the model is close to instability , so that some of the eigenmodes of the system are weakly damped and thus their corresponding connectivity profiles become visible against the noise . In contrast to previous work [21] , the stochastic regime does not give rise to realistic functional connectivity in the present study , suggesting that ISP has moved the system too far from instability . At higher coupling strengths , the oscillatory regime gives rise to stronger functional connectivity , but the orthogonalisation procedure eliminates most of the AEC because the raw signals are unrealistically highly correlated . As a result , neither regime gives rise to realistic functional connectivity . On the other hand , with ρ = 0 . 30 our findings are qualitatively similar , with a parameter regime that gives rise to realistic functional connectivity at increasingly long delays as coupling strength increases . The magnitude of similarity between model and data is also comparable . However , the optimal parameter regime is shifted to shorter delays for the same coupling strengths . Notably , with ρ = 0 . 30 it is possible to obtain realistic functional connectivity even with no delays in the network , which is not possible with ρ = 0 . 15 . Finally , we assessed the importance of locally balancing excitation and inhibition within every brain region , by testing the effect of coarsely balancing excitation and inhibition at the global level without introducing heterogeneity in the model parameters . We approximated this global balance by using the results of the ISP simulations–for each global coupling and delay , we computed the spatial average of cie after ISP , and then used this average within every brain region . In effect , this is an offline optimization . Fig 9A shows the resulting similarity in functional connectivity measures , and Fig 9B shows the homogeneous value of cie used for each delay and coupling value tested . As with the fine local balancing of excitation and inhibition in Fig 4B , sensitivity to global coupling is greatly reduced compared to Fig 4A where there is no regulation of inhibition . For PLI and PLV , the coarsely balanced simulations reach a maximum correlation between model and data that is similar to the locally balanced simulations , while for AEC performance of the coarsely balanced simulations is somewhat lower . We also examined the synchrony and metastability for these simulations , shown in Fig 9C and 9D . Comparing these to Fig 6C and 6D , we find qualitatively similar regimes of high synchrony , low , synchrony , metastability , and realistic functional connectivity as in the locally balanced ISP case . For some points in parameter space , there are ongoing changes in inhibitory synaptic strength , indicating that the homeostatic mechanism is unable to converge . If ISP fails to converge , this indicates there are still sizable discrepancies between the target activity level and the actual activity level , which drives ongoing plasticity . We can quantify this by computing the standard deviation of cie , after providing enough time for the system to converge first , if indeed convergence is possible . Convergence of ISP is examined in detail in the supplementary material , as shown in Figs A5 and A6 in S1 Supplementary Material . In summary , convergence tends to be robust for parameters that give rise to realistic neural activity , but fails in the high-synchrony regime , and in the low-synchrony regime depending on the ISP target .
There is considerable freedom in how to quantify similarity between model and data functional connectivity . A natural choice is to correlate the unique parts of the connectivity matrices , but even this is a nontrivial decision when working with MEG/EEG because the connectivity profiles are frequency band specific . Because the relatively simple model we used here does not predict band-specific connectivity , in this study we avoid the issue of combining connectivity across frequency bands by focusing only on the alpha band . We used the amount of individual variability in functional connectivity to provide an interpretation of the similarity in functional connectivity between model and data . The anatomical connectivity matrix we used in this study was an average across multiple subjects computed from the HCP dataset , and comes from different subjects to those used for our estimates of functional connectivity . The functional connectivity from the model is analogous to an individual estimate of functional connectivity , and for an ideal biophysical model we would expect that the simulated functional connectivity should not be significantly less correlated with the group average connectivity than any real individual . We found that in all connectivity measures , the model performance was not significantly different to individual variability . Interestingly , the model’s similarity for AEC is lower than the mean individual similarity regardless of model parameters , whereas in the optimal parameter regime , the model outperforms individual similarity in data for PLV and PLI . This likely reflects the relative dominance of different sources of variability in the data—AEC is known to be a more reproducible measure than PLV and PLI both within and across subjects [63 , 77] , which suggests that there is relatively more noise in PLV and PLI than AEC . This results in a reduction in similarity between single-scan estimates of individual functional connectivity and the group average for those metrics . Having achieved model performance consistent with individual variability where the metric is the correlation between functional connectivity matrices , future work may focus on increasing similarity in other metrics , rather than further improving correlation . For example , the Kolmogorov-Smirnov distance between the distributions of values in the connectivity matrices would quantify the difference in magnitude of connectivity , which is at present quite large ( as shown in Fig 5 , particularly for PLV and PLI ) . In this study , we applied the same analysis pipeline to the model predictions and to the data , including spatial leakage correction by signal orthogonalisation . Including an orthogonalisation step to compensate for source leakage is an integral part of most MEG analysis pipelines , to eliminate beamformer-induced connectivity [62 , 63 , 67 , 77–79] . This procedure may also remove some genuine neuronal connectivity , but how much genuine neuronal connectivity is removed cannot be determined based on the experimental data alone . In the model , by design all functional connectivity can only be neuronal in origin . However , it is possible to produce functional connectivity in the model that would not survive orthogonalisation–for example , the highly-synchronized regime has strong zero lag correlations that are greatly affected by orthogonalisation . By including orthogonalisation in our analysis , we ensure that the model predictions arise through mechanisms that are compatible with the experimental data . Future work could also investigate using a forward model to simulate MEG sensor timecourses , whether to compare activity to data in sensor space where spatial leakage correction is not necessary , or to then use a beamformer step to introduce realistic spatial leakage into the model signals . We also investigated the effect of changing the target activity level , and found that with the higher ISP target , the model could exhibit high similarity to experimental data even with zero delay . This is an interesting finding because some studies do not include propagation delays [8 , 14 , 80] and still find realistic functional connectivity , while others find that delays are required to obtain realistic model predictions [1 , 12 , 81] . Our model is able to exhibit both kinds of behaviour , depending on the target level of activity . In this study , we have leveraged the fast timescale of MEG to also examine phase-based metrics of functional connectivity . There are significant propagation delays in brain networks , and although the delays may not be critical when modelling fMRI , they are important in MEG and EEG because they are on a similar timescale to oscillations in neural activity . Phase metrics are expected to be sensitive to propagation delays between brain regions , and intuitively this suggests that the optimal parameter regime in the model should have a nonzero delay . However , for the high ISP target simulation at zero delay , the model is still fairly well correlated with data even in PLV and PLI . There are contrasting approaches in previous studies regarding the selection of intrinsic oscillatory frequency for uncoupled brain regions . A key factor is whether frequency suppression occurs in the model when brain regions are coupled [1 , 82] . For models that display frequency suppression such as the Kuramoto model , individual brain regions typically have an intrinsic frequency in the gamma band [1] , while models that do not use oscillators with intrinsic frequencies in bands of interest e . g . alpha [8] or otherwise examine functional connectivity by incorporating a hemodynamic response function that reduces dependence on oscillatory frequency [12] . The former approach is motivated by both the hypothesis that isolated neural masses resonate in the gamma band , and by the fact that frequency suppression would otherwise shift oscillations to even lower frequencies ( e . g . intrinsic alpha oscillations becoming network delta oscillations ) . The latter approach advocates that due to other connectivity such as thalamocortical interactions , sufficiently large brain regions may oscillate intrinsically at much lower frequencies . In this study , we used a model with local oscillations in the alpha band , because for the Wilson-Cowan model we do not see large frequency suppression at the network level for parameters that produce realistic functional connectivity . This alpha band timescale is much faster than fMRI , although there are even faster dynamics that can be investigated with MEG . A related open question is how to produce different band-specific patterns of functional connectivity using a single anatomical connectivity matrix . One promising direction may be to introduce multiple timescales of dynamics into the local model–for example , by introducing additional populations with different intrinsic oscillatory frequencies [41] , having the local effective time constants depend on network properties [14] , or by using a conduction-based neural mass model [83 , 84] that incorporates multiple timescales through the inclusion of multiple receptor types , each with a different time constant . We have examined inhibitory synaptic plasticity due to its biological plausibility as a mechanism for balancing excitation and inhibition , but it is not the only possibility . Because the fundamental imbalance in excitation and inhibition in this study originates with the fact that we only include long-range E-E connections , an obvious question is whether E/I balance could be achieved simply by including long-range E-I connections as well . The clear proportionality between long-range excitation and local inhibition suggests that this may be possible to some extent , and including E-I connections based on the same anatomical connectivity matrix will naturally mean that the increase in input to the inhibitory populations is proportionate to long-range excitation . However , balancing excitation and inhibition would likely require different global coupling strengths C for the E-E and E-I connections , and this in turn would necessitate some sort of optimization or homeostatic mechanism to tune them . Thus it is unlikely that including long range E-I connections would eliminate the need for a homeostatic mechanism . Similarly , balancing excitation and inhibition through long range E-I connections requires that these connections be proportionate to long-range excitation , but achieving this balance in connection strengths in the real brain would likely require its own regulatory mechanism . Finally , there are also other mechanisms that could regulate excitatory activity–for example , intrinsic plasticity that directly modulates the excitability of the excitatory populations [85 , 86] . These may offer alternate routes to balancing excitation and inhibition at the network level . One of the proposed roles of ISP is to oppose perturbations to the network , improving the robustness of dynamics to changes such as lesions . Previous studies have modelled lesions by disconnecting regions from the network and examined the effect this has on network [3 , 6 , 73 , 87] . In some cases , removing a region can dramatically change oscillatory activity in the network . Does balancing excitation and inhibition restore functional connectivity ? There is evidence that offline balancing of inhibition is able to restore functional connectivity as measured by fMRI [88] . It is an open question whether the same is true for more biophysically plausible online homeostatic mechanisms , although the results presented here suggest that the outcome may be comparable . Aside from frequency-specific functional connectivity , there is also emerging evidence that changes in oscillatory frequency are linked to connectivity and functional hierarchy , with areas higher in the hierarchy exhibiting lower frequency oscillations [89–91] . In a previous study , a similar model to the one used here was able to produce this behavior through an upscaling of excitatory inputs depending on the region’s position in the hierarchy [14] . This upscaling affects different local connections to ISP , and could therefore be implemented in conjunction with ISP . Intuitively , ISP should act to oppose the upscaling of excitation , but whether this will eliminate the frequency suppression effect is unclear . More generally , this points to a broader question of the interaction and potential interference between homeostatic mechanisms and their effect on desirable network inhomogeneities . In the present study , we approximated the delays in the network using Euclidean distance and a uniform conduction velocity , which matches previous recent work [1 , 65 , 73 , 92] . The effect of delays in brain networks and the sensitivity of dynamics to a distribution of delays is still an open question [93–97] . Specifically , the effect of tract length and myelination , both of which affect the propagation delay between brain regions , is yet to be explored . The framework developed in this study is suitable for investigating alternate methods for calculating delays , and we plan to investigate the effect of accounting for tract length and myelination in future work . In this study , we focused on reproducing static functional connectivity in the model . Having reached levels of performance within the range of individual variability in data , future work on resting state connectivity may focus on reproducing more features in the data rather than pursuing higher correlations . In particular , the nonstationary nature of real brain dynamics is well established [78 , 98 , 99] , and there have been a range of recent developments in methods to quantify fast transient dynamics with regard to functional connectivity in electrophysiological data [21 , 100–102] . What network properties are responsible for these dynamics is still an open question , although there a wide range of possible mechanisms that produce transient dynamics in models [1 , 9 , 55 , 103–105] . The next step in this direction is to examine transient states in our model in more detail , to characterize properties such as the statistics of state transitions , and the transient patterns of functional connectivity and frequency content . | Recently there has been much interest in investigating the role of synaptic plasticity in supporting healthy brain activity . In particular , the balance between excitation and inhibition in the brain is believed to play a critical role in brain dynamics , and it is likely that this balance is regulated by homeostatic mechanisms . Biophysical models of the brain have previously been used to predict functional connectivity , but are typically extremely sensitive to changes in parameter values and require extremely fine tuning to achieve realistic dynamics . In this study , we investigated whether including a homeostatic plasticity mechanism would improve the robustness of simulated neural dynamics . We focused on functional connectivity in MEG data , which can resolve fast oscillations in neural activity , unlike fMRI . We found that including a simple plasticity rule to balance excitation and inhibition resulted in more realistic model predictions , and reduced sensitivity to changes in model parameters . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"neural",
"networks",
"nervous",
"system",
"homeostatic",
"mechanisms",
"neuroscience",
"magnetic",
"resonance",
"imaging",
"simulation",
"and",
"modeling... | 2018 | A biophysical model of dynamic balancing of excitation and inhibition in fast oscillatory large-scale networks |
Studies of the 1918 H1N1 influenza pandemic , the H5N1 avian influenza outbreak , and the 2009 H1N1 pandemic illustrate that sex and pregnancy contribute to severe outcome from infection , suggesting a role for sex steroids . To test the hypothesis that the sexes respond differently to influenza , the pathogenesis of influenza A virus infection was investigated in adult male and female C57BL/6 mice . Influenza infection reduced reproductive function in females and resulted in greater body mass loss , hypothermia , and mortality in females than males . Whereas lung virus titers were similar between the sexes , females had higher induction of proinflammatory cytokines and chemokines , including TNF-α , IFN-γ , IL-6 , and CCL2 , in their lungs than males . Removal of the gonads in both sexes eliminated the sex difference in influenza pathogenesis . Manipulation of testosterone or dihydrotestosterone concentrations in males did not significantly impact virus pathogenesis . Conversely , females administered high doses of estradiol had a ≥10-fold lower induction of TNF-α and CCL2 in the lungs and increased rates of survival as compared with females that had either low or no estradiol . The protective effects of estradiol on proinflammatory cytokines and chemokines , morbidity , and mortality were primarily mediated by signaling through estrogen receptor α ( ERα ) . In summary , females suffer a worse outcome from influenza A virus infection than males , which can be reversed by administration of high doses of estradiol to females and reflects differences in the induction of proinflammatory responses and not in virus load .
Males and females differ in their responses to infection with many viral pathogens , including human immunodeficiency virus ( HIV ) , herpes simplex viruses , and hantaviruses [1] . Although societal and behavioral factors can influence exposure to viruses and access to vaccines and treatments for infection [2] , genetic and physiological differences between the sexes can cause differential immune responses to viruses [3] . Because females tend to mount higher innate [4] , [5] , cell-mediated [5] , [6] , [7] , and humoral [8] immune responses than males , viral loads are often reduced among females [1] . Heightened immunity in females also can lead to the development of immunopathology following viral infection [5] . Elevated immunity in females represents a balance between immune responses conferring protection and causing pathology . Growing evidence links sex differences in immune function with circulating sex steroid hormones [9] , [10] . Receptors for sex steroids are expressed in a variety of lymphoid cells [11] , [12] . Androgens , including dihydrotestosterone ( DHT ) and testosterone ( T ) , suppress the activity of immune cells [9] , [13] , [14] . Estradiol ( E2 ) can have divergent effects , with low doses enhancing proinflammatory cytokine production ( e . g . , IL-1 , IL-6 , and TNF-α ) and T helper cell type 1 ( Th1 ) responses and high or sustained concentrations reducing production of proinflammatory cytokines and augmenting Th2 responses and humoral immunity [15] . Elevated E2 also attenuates production of CXC chemokine ligand ( CXCL ) -8 , CXCL10 , chemokine ( C-C motif ) ligand 2 ( CCL2 ) , and CCL20 and recruitment of leukocytes and monocytes into several tissues , including the lungs [16] , [17] , [18] , [19] . The anti-inflammatory effects of high E2 are mediated by signaling through estrogen receptors ( ERs ) , which inhibits activation of NF-κB-mediated inflammatory responses [20] . Observational data for influenza reveal that the outcome of pandemic influenza as well as avian H5N1 is generally worse for young adult females [21] . In the United States , during the 1957 H2N2 pandemic , mortality was higher among females than males 1–44 years of age [22] . Worldwide as of 2008 , females were 1 . 6 times less likely to survive H5N1 infection than males [23] . During the 2009 H1N1 pandemic , a significant majority of patients hospitalized with severe 2009 H1N1 disease were young adult females ( 15–49 years of age ) [21] , [24] , [25] , [26] , [27] , [28] . Pregnancy and other risk factors ( e . g . , asthma and chronic obstructive pulmonary disorder ) contribute to the severity of disease in females [21] . The mechanisms mediating how the sexes differ in response to influenza virus infection as well as the effects of sex steroids on influenza pathogenesis remain largely undefined . We hypothesize that biological differences in immune responses may explain variation between the sexes during influenza virus infection . Several studies reveal that excessive proinflammatory responses ( i . e . , the cytokine storm ) contribute significantly to morbidity and mortality from influenza virus infection [29] , [30] , [31] , [32] , [33] . Our data reveal that females experience greater morbidity and mortality than males , which can be reversed by administration of exogenous E2 or an ERα agonist to females . Our data further indicate that sex differences and the effects of E2 on influenza pathogenesis reflect differences in the production of proinflammatory cytokines and chemokines as opposed to differences in virus load .
To examine whether the sexes respond differently to influenza A virus infection , adult male and female C57BL/6 mice were inoculated with 102 TCID50 of influenza A/Puerto Rico/8/1934 ( PR8; H1N1 ) and monitored daily for changes in morbidity and mortality for 21 days . Females showed a greater percent reduction in body mass ( Fig . 1A ) and body temperature ( Fig . 1B ) than males , with these differences being most pronounced 7–13 days post-inoculation ( p . i . ) ( MANOVA sex x day P<0 . 0001 in each case ) . Survival following influenza infection was significantly reduced in females compared with males ( log rank P<0 . 001 ) , in which no females survived infection with PR8 , whereas 47% of the males survived through 21 days p . i . ( Fig . 1C; Χ2 P<0 . 05 ) . The average day of death was 2 days earlier for females ( 9 . 4±0 . 6 days ) than males ( 11 . 5±0 . 7 days ) ( t-test P<0 . 05 ) . Titers of infectious virus peaked 3–5 days after infection , but sex differences were not observed ( Fig . 1D ) , suggesting that changes in virus load alone were not responsible for the observed sex differences in morbidity and mortality . Highly pathogenic influenza viruses cause severe disease by initiating profound proinflammatory cytokine and chemokine responses [29] , [33] . Inflammatory cytokine responses increased in both sexes , in a time-dependent manner as documented previously [30] , [34] , [35] , [36] . Interleukin-1β , IL-12p70 , IL-10 , and TGF-β were induced within 24 h p . i . ; IFN-β , IL-6 , TNF-α , CCL2 , and CCL3 were induced within 72 h p . i . ; and IFN-γ and IL-10 were induced 7 days p . i . in both sexes ( Table S1; 2-way ANOVAs , main effect of day P<0 . 05 ) . Females showed a greater induction of CCL2 , TNF-α , IFN-γ , and IL-6 than males ( Fig . 1E-H , 2-way ANOVAs sex x day P<0 . 05 ) . Inflammatory immune responses induced by fatal infection with pathogens ( e . g . , HIV ) affect the brain to reduce reproductive function , appetite , and thermoregulation [37] , [38] , which in mice can result in greater suppression of reproductive activity in females than males [39] . To evaluate the effects of PR8 infection on reproductive physiology , T concentrations in males and E2 concentrations in females were evaluated in plasma samples collected from separate mice at several time-points during the first week p . i . Infection of males reduced circulating T concentrations on days 3–7 p . i . as compared with uninfected males ( Fig . 2A , 1-way ANOVA P<0 . 05 ) . In females , infection with PR8 appeared to cause persistently low E2 concentrations ( Fig . 2B ) ; single time-point sampling of cyclical hormones , however , is difficult to accurately evaluate in females [40] . To better characterize the hormonal milieu of females during influenza virus infection , we monitored the estrous cycles of female mice before and after infection with PR8 . The average duration of the estrous cycle increased significantly following influenza virus infection ( Fig . 2C; paired t-test P<0 . 05 ) and this increase in estrous cycle length was attributed to an increase in the duration of diestrus ( Fig . 2D; paired t-test P<0 . 05 ) . As diestrus is the stage of the estrus cycle that corresponds with the follicular phase , when both estrogens and progesterone are at their nadir [40] , these data suggest that influenza virus infection suppresses ovarian function in females and results in persistently low circulating E2 . To establish whether gonadal secretions modulate sex differences in influenza pathogenesis , we compared morbidity and mortality following PR8 infection in gonadally-intact ( sham ) and gonadectomized ( gdx ) male and female mice . Consistent with previous data ( Fig . 1 ) , hypothermia ( MANOVA sex x day P<0 . 0001 ) and mortality ( log rank P<0 . 001 ) following influenza virus infection were greater among gonadally-intact ( sham ) females than gonadally-intact ( sham ) males ( Fig . 3A and B ) . Gonadectomy of males resulted in more pronounced hypothermia ( MANOVA treatment x day P<0 . 003 ) and death ( log rank P = 0 . 04 ) as compared with gonadally-intact males ( Fig . 3A and B ) . Mortality ( log rank P = 0 . 046 ) , but not morbidity , was lower among gdx than gonadally-intact females during influenza virus infection . Among gdx animals , removal of the gonads in male and female mice eliminated the dimorphism in hypothermia and survival ( Fig . 3A and B ) . In summary , sex differences in influenza pathogenesis are partially mediated by the presence of gonadal secretions . To establish whether androgens in males and estrogens in females affect responses to influenza , we examined the effects of removal and replacement of sex steroids on influenza pathogenesis . Among males , gdx animals that received exogenous T had greater concentrations of circulating T ( 14 . 74±0 . 65 ng/ml ) than gdx males ( 1 . 00±0 . 01 ng/ml ) ( t-test P<0 . 05 ) . Treatment with either T or DHT did not significantly reverse the effects of gdx on either hypothermia ( Fig . 4A ) or mortality ( Fig . 4C ) . Titers of PR8 also did not differ among hormonally-manipulated and gonadally-intact males ( Fig . 4D ) . Manipulation of androgens in males affected concentrations of CCL3 , IFN-γ , and IL-10 , but not in a discernable pattern associated with morbidity and mortality ( Table S2; 2-way ANOVAs treatment x day P<0 . 05 ) . Among females , those that received exogenous E2 had significantly higher serum concentrations of E2 ( 978±39 pg/ml ) than gdx females ( 1±26 pg/ml ) ( t-test P<0 . 05 ) . Administration of exogenous E2 mitigated hypothermia ( Fig . 5A; MANOVA treatment x day P<0 . 005 ) and mortality ( Fig . 5B; log rank P<0 . 001 ) following PR8 infection as compared with gonadally-intact and gdx female mice . Females that received E2 were more likely to survive PR8 infection and those that died had a later day of death ( 12 . 8±1 . 2 days ) than did sham ( 8 . 6±0 . 4 days ) or gdx ( 10 . 7±1 . 1 days ) female mice ( 1-way ANOVA P<0 . 05 ) . Administration of E2 did not affect virus replication kinetics ( Fig . 5C ) , but diminished the rise in TNF-α and CCL2 in the lungs that was apparent among gonadally-intact and gdx female mice ( Fig . 5D and E; 2-way ANOVA treatment x sex P<0 . 001 ) . Although hormone manipulation in females altered other cytokines , including IFN-γ , IL-10 , IL-12 ( p70 ) , and CCL3 ( Table S3; 2-way ANOVAs treatment x sex P<0 . 05 ) , the patterns were not associated with changes in morbidity or mortality . In summary , females with low ( sham ) or no ( gdx ) circulating E2 suffer a worse outcome from infection and have higher proinflammatory responses than females with high E2 . The anti-inflammatory effects of high E2 are mediated by signaling through two nuclear receptors , ERα and ERβ [41] , which antagonizes nuclear factor kappa B ( NF-κB ) activity [20] . To determine which ER was mediating the effects of E2 on influenza pathogenesis , gdx females were administered E2 , vehicle , or vehicle containing agonists specific to ERα ( Propylpyrazole-triol; PPT ) or ERβ ( diarylpropionitrile; DPN ) . Treatment with the ERα agonist , but not vehicle or the ERβ agonist , reduced hypothermia ( Fig . 6A; MANOVA treatment x day P<0 . 0001 ) and increased rates of survival ( Fig . 6B; log rank P<0 . 01 ) to levels that were similar to females treated with E2 . Titers of PR8 in the lungs peaked for all females at Day 3 p . i . , but were not affected by ER manipulation ( data not shown ) . Treatment with the ERα agonist reduced TNF-α and CCL2 ( Fig . 6C and D; 1-way ANOVA P<0 . 001 ) in the lungs to levels that were similar to those of females treated with E2 . Treatment with the ERβ agonist reduced TNF-α ( Fig . 6C; 1-way ANOVA P<0 . 001 ) , but not CCL2 ( Fig . 6D ) as compared with vehicle-treated females . Although administration of the ERα and ERβ agonists altered IL-6 and IL-10 , these patterns were not correlated with morbidity and mortality ( Table S4; 1-way ANOVAs P<0 . 05 ) . The effects of E2 on proinflammatory responses to infection , in particular CCL2 responses , and disease outcome are primarily mediated by signaling through ERα .
Although epidemiological data suggest that females experience more severe disease and suffer a worse outcome from influenza virus infection than males [21] , whether these differences reflect sex or gender is difficult to assess as both factors can affect exposure and vulnerability to influenza A viruses [21] . Using a small animal model , our data and data from others [42] illustrate that there are distinct biological differences in how males and females respond to influenza . Disease associated with highly pathogenic influenza viruses and the clinical manifestations that ensue in humans can be mediated by the proinflammatory response ( e . g . , TNF-α , IL-6 , CCL2 , CCL3 , and CXCL10 ) initiated by the host in response to infection [29] , [30] , [31] , [32] , [33] . Studies of patients infected with avian influenza viruses further reveal that higher proinflammatory responses are correlated with mortality during infection [33] . Elevated production of CCL3 and CCL2 and expression of CCR2 recruit monocytes and neutrophils into the lungs and regulate inflammation and influenza A virus replication [43] . The data from the present study illustrate that inflammatory immune responses , including induction of CCL2 , IFN-γ , IL-6 , and TNF-α , are elevated in the lungs of females compared with males . Infectious virus titers , however , do not differ between the sexes and are not altered by hormones . Similarly , infection of adult BALB/c mice with a mouse-adapted H3N1 influenza A virus results in greater lung hyperresponsiveness to methacholine challenge and production of CCL2 , but not virus titers , in females compared with males [42] . These data support and extend the hypothesis that host-mediated immunopathology rather than virus replication underlies influenza pathogenesis . Sex differences in disease outcome are likely mediated by multiple factors , including sex steroids , glucocorticoids , and the direct activity of sex chromosomal genes [44] . In the present study , removal of gonadal secretions in both males and females reduced the sex difference in morbidity and mortality , illustrating that the sex difference in influenza pathogenesis is reversible and that activational sex steroids in adulthood affect the outcome of infection . Furthermore , sex differences in response to influenza A virus infection are not observed among pre-pubertal mice [42] . Within males , however , manipulation of androgens did not significantly affect influenza pathogenesis , suggesting that some androgenic effects may be organized early during sexual differentiation . The extent to which sex differences in immunity are hard-wired early during development must be considered [44] . Our data reveal that estrogens are one mechanism mediating influenza pathogenesis in females . Infection with influenza virus disrupted reproductive function in gonadally-intact females , resulting in a prolonged state of diestrus , which is the stage of the reproductive cycle when E2 and progesterone concentrations are at their lowest [40] . Gonadectomized females ( i . e . , females with no circulating E2 ) and gonadally-intact females ( i . e . , females with low circulating E2 as a result of infection ) produced higher inflammatory responses and suffered a worse outcome from infection than gdx females administered exogenous E2 . These data support the hypothesis that low concentrations of E2 in females promote excessive inflammatory responses that contribute to disease pathogenesis [15] . In the present study , exogenous administration of E2 reduced the induction of pulmonary inflammatory responses and protected females against influenza . High doses of estrogens also are protective in animal models of multiple sclerosis ( MS ) , in which supraphysiological doses of estrogens reduce inflammatory responses and progression of this autoimmune disease [45] , [46] . In contrast , low cyclical levels of estrogens in gonadally-intact females have little effect on the outcome of MS . High E2 has potent anti-inflammatory actions , including repression of proinflammatory gene transcription and cytokine production [11] , [47] , which is partially mediated by inhibition of NF-κB transcriptional activity [48] . The anti-inflammatory effects of estrogens have been observed in several models for diseases , including autoimmunity , atherosclerosis , arthritis , inflammatory bowel disease , and asthma [15]; our data reveal that influenza is another disease of public health importance that is influenced by estrogens . The proinflammatory effects of low or no E2 and the anti-inflammatory effects of high E2 in females are mediated by signaling through the ER , which regulates the activity of NF-κB [20] . Administration of an ERα , but not an ERβ , agonist protected females against influenza infection . ERα has been identified in several immune cells , including DCs , macrophages , and T cells , whereas ERβ is expressed in epithelial cells , macrophages , and B cells [11] , [41] . The differential effects of ERα and ERβ agonists in vivo provide insight into the cell types that may be responsible for the exacerbated inflammatory responses observed in influenza-infected females with low or no E2 . Our data suggest a number of important avenues of research that require further investigation . Mechanisms in addition to low circulating E2 likely influence sex differences in influenza pathogenesis and we are actively investigating the effect of other sex-specific factors on viral disease . In addition to the mouse-adapted PR8 ( H1N1 ) , sex differences are reported in response to mouse-adapted H3N1 [42] and H3N2 ( Lorenzo et al . unpublished data ) ; mouse models utilizing mouse-adapted influenza A viruses , however , may not completely reflect virus pathogenesis in humans . Because clinical isolates of influenza A viruses cause limited pathology in mice [49] , [50] , [51] , examination of sex differences and the effects of sex steroids in response to non-adapted strains of influenza in mice would need to be limited to highly pathogenic viruses such as the 1918 virus strain and avian H5N1 viruses . Alternatively , sex differences in response to infection with clinical isolates of influenza A viruses could be evaluated in alternative animal models , such as ferrets [49] . The use of mouse-adapted strains of influenza to demonstrate sex differences and effects of sex steroids on influenza pathogenesis in mice reveal significant differences and suggest that these differences should be considered in evaluations of epidemiological and clinical human data . The observation that elevated E2 reduces , rather than elevates , the severity of influenza A virus infection does not explain why pregnancy is associated with worse outcome after infection . Elevation of E2 concentrations in non-pregnant females does not completely recapitulate pregnancy as several other hormones , including progesterone , estriol , and glucocorticoids also dramatically change during pregnancy and can impact immune function [52] , [53] . While sex steroid-modulation of influenza pathogenesis likely contributes to the increased severity of disease during pregnancy , the data from the current study suggest that E2 is not the mechanism mediating severe outcome of infection during pregnancy . There is clinical relevance to uncovering the mechanisms mediating how sex and sex steroids affect responses to influenza viruses as this may result in preventative measures and treatments that are optimized for each sex . Most epidemiological and clinical studies of influenza in humans do not partition or analyze data by sex and a majority of animal studies of influenza either use only females or do not report the sex of their animals [21] . The data from the present study provide evidence that the pathogenesis of influenza virus infection differs between the sexes and is influenced by the effects of sex hormones on inflammatory immune responses .
Adult ( 6–8 weeks old ) male ( total n = 308 ) and female ( total n = 356 ) C57BL/6 mice were purchased from NCI Frederick , housed 5/microisolater cage with food and water available ad libitum , and handled using Biosafety Level ( BSL ) -2 practices . All experiments were performed in compliance with the standards outlined in the National Research Council's Guide to the Care and Use of Laboratory Animals . The animal protocol ( MO09H26 ) was reviewed and approved by the Johns Hopkins University Animal Care and Use Committee . All efforts were made to minimize animal suffering . Male and female mice were anesthetized with 2 . 5% isoflurane ( Baxter Healthcare Corporation , Deerfield , IL ) mixed with oxygen and bilaterally gonadectomized as previously described [54] , [55] , [56] . All animals were given two weeks to recover prior to infection . Vaginal cell samples were collected at 1600–1700 h , smeared onto clean glass slides , fixed , stained with Diff-Quick Staining kit ( Andwin Scientific , Addison , IL ) , and diagnosed for stage of estrus based on the cellular profile of each sample: proestrus ( 80–100% intact , healthy epithelial cells ) , estrus ( 100% cornified epithelial cells ) , diestrus I ( ∼50% cornified epithelial cells and 50% leukocytes ) , and diestrus II ( 80–100% leukocytes ) [57] , [58] . Only females that exhibited at least 3 regular estrous cycles ( 16/20 ) prior to infection were included . Hormone and placebo capsules were made with 10 mm of silastic tubing ( inner diameter = 0 . 04 in; outer diameter = 0 . 085 in; VWR International , Bridgeport , NJ ) and sealed with silastic medical adhesive ( Dow Corning , Midland , MI ) . Hormone capsules were left empty ( placebo ) or filled with 5 mm of testosterone ( T ) , dihydrotestosterone ( DHT ) , or 17β-estradiol ( E2 ) purchased from Sigma Aldrich ( St . Louis , MO ) and sealed with 2 . 5 mm of adhesive at either end [56] . Capsules were incubated in 0 . 9% saline at 37°C overnight prior to implantation . Propylpyrazole-triol ( PPT ) and diarylpropionitrile ( DPN ) were purchased from Tocris Bioscience ( Ellisville , MO ) , suspended in Miglyol 812N oil ( kindly provided by Sasol , Hamburg , Germany ) , and administered at 10 mg/kg and 8 mg/kg , respectively [59] . Animals received daily subcutaneous injections of either vehicle ( 90% Miglyol , 10% EtOH ) or vehicle containing PPT or DPN . For experiments in which morbidity and mortality were monitored for up to 21 days post-inoculation , animals were not euthanized as death was an approved endpoint in our IACUC protocol . Body mass and rectal temperature were measured daily between 0800 and 1000 h . Animals were weighed to the nearest hundredth of a gram and rectal temperature was monitored with a Thermalert TH-5 monitor ( 25°C–45°C ) and RET-3 rectal microprobe for mice ( Physitemp Instruments , Inc . , NJ ) , which stably evaluates body temperature to the nearest 0 . 1°C in 3–5 seconds . For time course experiments , animals were randomly assigned to be euthanized at 0 , 1 , 3 , 5 , or 7 days p . i . , at which time they were anesthetized with isoflurane and terminally bled from the retro-orbital sinus into heparinized tubes and plasma was stored at −80°C and used to measure hormones . Whole lungs were collected , snap-frozen , and stored at −80°C until homogenized in Dulbecco's Modified Eagles Medium ( DMEM ) supplemented with 1% Penicillin/Streptomycin and 1% L-glutamine ( Invitrogen , Carlsbad , CA ) . Homogenates were centrifuged and the supernatants were stored at −80°C and used to measure virus titers and cytokine concentrations . Mouse-adapted influenza A virus , A/Puerto Rico/8/34 ( PR8 ) was provided by Dr . Maryna Eichelberger at the Food and Drug Administration . Mice were anesthetized with Ketamine/Xylazine ( 80 mg/kg and 6 mg/kg , respectively ) and intranasally inoculated with 30 µl of vehicle ( DMEM ) or 102 50% tissue culture infective dose ( TCID50 ) units of PR8 diluted in DMEM ( which corresponds to 1 . 24 50% mouse lethal dose ( MLD50 ) for males and 0 . 78 MLD50 for females , based on the Reed and Meunch method ) between 0800 and 1000 h . The challenge dose was selected based on previous experiments that quantified the lethal dose that killed 50% of the animals ( LD50 ) . Virus quantification was performed using the TCID50 method measuring cytopathic effects ( CPE ) of influenza A virus on a monolayer of Madin Darby Canine Kidney ( MDCK ) cells [60] . Hormones were concentrated from plasma by ether extraction and hormone quantification was performed using testosterone and estradiol enzyme immunosorbant assays ( EIA ) purchased from Cayman Chemicals ( Ann Arbor , MI ) . Supernatants from lung homogenates were used to measure CCL3 , IFN-β , IL-1β , and TGF-β by ELISAs ( R&D Systems , BD Biosciences , PBL Biomedical Laboratories ) and CCL2 , IL-12 ( p70 ) , TNF-α , INF-γ , IL-10 , and IL-6 with the mouse inflammation cytometric bead array ( BD Biosciences ) . Kaplan Meier survival curves were compared using log rank analyses . The proportion of animals that survived influenza A virus infection was compared among experimental groups using χ2 analyses . Morbidity data were analyzed with multivariate ANOVAs ( MANOVAs ) with one within-subjects variable ( days ) and one between-subjects variable ( sex or treatment ) and significant interactions were further analyzed using planned comparisons . Virus titers and protein concentrations were analyzed with 2-way ANOVAs with day p . i . and sex/treatment as the independent variables and significant interactions were further analyzed using the Bonferroni method for pairwise multiple comparisons . Hormone concentrations were analyzed by t-tests or 1-way ANOVAs followed by Bonferroni post hoc tests . Changes in estrous cycle length were evaluated using paired t-tests . Mean differences were considered statistically significant if P<0 . 05 ( SYSTAT 13 , Systat Software , Chicago , IL ) . | Sex and pregnancy affect the outcome of infection with seasonal , avian , and pandemic influenza viruses among young adults . Males and females are biologically different , yet the implications of these differences on influenza A virus pathogenesis are not well characterized . Generally , females mount more robust immune responses to viral challenge than males , which can result in more efficient virus clearance at the cost of developing immune-mediated pathology . In this study , we tested the hypothesis that sex and sex steroid hormones differentially impact the outcome of influenza A virus infection in mice . Our data illustrate that influenza A virus dysregulates reproductive function as well as cytokine and chemokine production in females , rendering them significantly more susceptible to weight loss , hypothermia , and death than males . Administration of a high dose of estradiol or an estrogen receptor α agonist to females suppresses the excessive induction of cytokines and chemokines and increases survival following infection . The protective effects of estradiol on influenza pathogenesis reflect changes in the induction of proinflammatory responses and not in virus load . Uncovering the mechanisms mediating how sex and sex steroid hormones affect influenza pathogenesis may result in preventative measures and treatments that are optimized for both sexes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"immunopathology",
"virology",
"immunology",
"biology",
"microbiology",
"immune",
"response"
] | 2011 | Elevated 17β-Estradiol Protects Females from Influenza A Virus Pathogenesis by Suppressing Inflammatory Responses |
Bacterial chemotaxis is one of the best studied signal transduction pathways . CheW is a scaffold protein that mediates the association of the chemoreceptors and the CheA kinase in a ternary signaling complex . The effects of replacing conserved Arg62 of CheW with other residues suggested that the scaffold protein plays a more complex role than simply binding its partner proteins . Although R62A CheW had essentially the same affinity for chemoreceptors and CheA , cells expressing the mutant protein are impaired in chemotaxis . Using a combination of molecular dynamics simulations ( MD ) , NMR spectroscopy , and circular dichroism ( CD ) , we addressed the role of Arg62 . Here we show that Arg62 forms a salt bridge with another highly conserved residue , Glu38 . Although this interaction is unimportant for overall protein stability , it is essential to maintain the correct alignment of the chemoreceptor and kinase binding sites of CheW . Computational and experimental data suggest that the role of the salt bridge in maintaining the alignment of the two partner binding sites is fundamental to the function of the signaling complex but not to its assembly . We conclude that a key feature of CheW is to maintain the specific geometry between the two interaction sites required for its function as a scaffold .
The Escherichia coli chemotaxis pathway employs dedicated chemoreceptors that are anchored in the membrane and detect signals from both outside and inside the cell [1] . Chemoreceptors relay this information to the CheA histidine kinase , which then communicates the information to its cognate response regulator CheY . In a phosphorylated form , the CheY protein binds to flagellar motors to cause a change in the direction of its rotation , thus converting the initial signal detected by chemoreceptors into a behavioral response – a change in the swimming direction . This pathway also employs the receptor-modifying enzymes CheB and CheR as well as the CheZ phosphatase , which acts on CheY [2] . The key features of this remarkable system include high sensitivity , wide dynamic range , signal integration , memory , and precise adaptation [3]–[7] , all of which are consequences of a highly ordered arrangement of chemoreceptor and kinase proteins at the cell pole [4] , [8] , [9] . The geometry of a hexagonal array with a lattice spacing of 12 nm is conserved over long evolutionary distances [9] , indicating the importance of precise interactions among members of the complex . In addition to the chemoreceptors and the CheA kinase , this complex also contains the CheW protein , which is interchangeably referred to as a docking , scaffold , coupling , or adaptor protein [10]–[12] . The crystal structure of CheW [10] , [13] , [14] reveals a fold composed of two five-stranded β-barrel subdomains connected by a hydrophobic core . Within the chemotaxis signaling complex , the CheW fold is present not only as a stand-alone adaptor but also as a homologous domain within the CheA kinase [15] . Furthermore , the two subdomains of CheW are topologically similar to the SH3 domain [15] , which is widely distributed among scaffold proteins in eukaryotic signal transduction systems [16] . Thus , elucidating the structure/function relationships of CheW will have a broader impact in understanding the role of scaffold proteins in signal transduction system in all organisms . CheW is required for proper activation of the kinase by the chemoreceptor [17] and is essential for the formation of the chemotaxis complex [18] . Overexpression of CheW disrupts formation of chemoreceptor trimers by blocking trimer contacts [12] , [19] , [20] , thereby impairing chemotaxis [21] . The binding sites for CheA and the chemoreceptor on CheW have been mapped using various experimental approaches [12] , [19] , [22]–[24] . The overall results were consistent with CheW being a scaffold protein . However , the replacement of Arg62 ( throughout the text , numbers are for E . coli CheW ) with His , which moderately affected in vitro binding affinity of CheW for both its binding partners , completely abolishes chemotaxis . This finding indicates that CheW plays a role in addition to holding CheA and the chemoreceptors together [24] . Our view on the role of scaffold proteins in signal transduction is rapidly changing . They can no longer be viewed as nothing more than molecular “glue” . It is clear that their dynamics must play a central role in their communication with partner proteins in signaling proteins [25] . Although X-ray and NMR structures are excellent starting points , they do not describe the dynamics properties of proteins . These properties can be studied only by tracking the time-dependent positions of all atoms in the system through molecular dynamics simulations , a methodology that has improved dramatically in recent years [26] . In this study , we sought to gain a deeper insight into the structure/function relationship of CheW by using a combination of sequence analysis , NMR spectroscopy , circular dichroism ( CD ) , and molecular dynamics ( MD ) . This approach revealed the existence of an evolutionarily conserved salt bridge on the surface of CheW that is responsible for maintaining the stability of a specific geometry within the signaling complex that is essential for its function .
NMR data were collected at 30°C with a Varian Inova 600 MHz spectrometer equipped with a four-channel ( 1H , 13C , 15N , and 2H ) cryoprobe and Z-axis pulsed field gradients . NMR data were analyzed with the nmrPipe package and ANSIG3 . 3 [27] , [28] . The wild-type CheW backbone chemical-shift assignments were obtained from previous publication ( BMRB accession No . 15322 ) [13] . Of the 154 published assignments , 123 were transferred to our wild-type CheW 15N-HSQC spectrum . The remaining assignments ( 20% ) were not transferred because of the overlap or the weak intensity of these resonances under these experimental conditions . All NMR samples were analyzed in 30 mM Tris-HCl ( pH 7 . 3 ) , 30 mM NaCl , 0 . 2% sodium azide in 90% H2O and 10% D2O . The concentration of the NMR sample was 1 mM for the WT CheW and 1 . 5 mM for the R62A mutant . The longitudinal relaxation time T1 ( or inverse rates R1 ) , transverse relaxation time ( or inverse rates R2 ) , and the 1H-15N NOE factor of backbone amide 15N nuclei were measured using inverse-detected two-dimensional ( 2D ) experiments [29]–[33] . Measured delay times for R1 relaxation rate were 11 , 55 , 110 , 220 , 330 , 440 , 660 , 880 , and 1210 ms . Measured delay times for R2 relaxation rates were 16 . 5 , 33 , 49 . 5 , 66 , 82 . 6 , 99 . 1 , 115 . 6 , 132 . 1 , and 148 . 6 ms . A recycle delay of 1 . 5 s was used for both R1 and R2 measurements . R1 and R2 were extracted by fitting the peak intensities with a single exponential-decay function . The 1H-15N NOE factor was taken as the ratio of the peak intensities with and without proton saturation during 3 s of the 8 s recycle delay period [33] , [34] . Further analysis of the dynamics data was performed by using the MODELFREE program [30] , [32] , [35] , [36] to provide information on the internal and overall motions . The 15N R1 , R2 and 1H-15N NOE values were fitted to a single isotropic rotational diffusion model described by the overall correlation time τm . The model contains a contribution from fast internal motions described by the order parameter S2 and the correlation time τe and from additional exchange broadening ( Rex ) on the time scale of µs to ms . During the calculation , τm was fixed at 11 . 0 ns for wild-type and 11 . 6 ns for the mutant , and internal motional parameters were optimized [29]–[32] , [35]–[38] . For more accurate characterization of the chemical exchange contribution ( Rex ) to the transverse relaxation rate constant , a series of modified Carr-Purcell-Meiboom-Gill ( CPMG ) relaxation-dispersion experiments were performed [39]–[41] . The total CPMG period was kept constant at 80 . 0 ms while the delay τcp was varied for a total of 9 values ranging from 1 . 0 ms to 20 . 0 ms . The ΔRex term , with a base value at the fastest spin-echo rate or the shortest τcp = 1 ms , can be extracted by the following equation: ( 1 ) where I is the peak intensity at τcp and I0 is the peak intensity with τcp = 1 ms . The value of Rex is determined by the difference in chemical shift between two exchange sites ( Φex ) and the reduced lifetime of the exchange sites ( τex ) : ( 2 ) in which ; and pi and ωi are the population and Larmor frequency for the nuclear spin at site i , respectively , and τex is the reduced lifetime of the exchanging sites [38] . CD spectra were collected with an Aviv CD Spectrometer Model 202 . Wild-type and the R62A mutant variant of CheW were diluted to 7 µM in a 1 cm path-length quartz cuvette . Each sample was then titrated with 11 . 0 ml of the 7 µM protein and 9 . 5 M Urea ( Amresco , Ultra Pure Grade ) . The urea-induced denaturation experiments were controlled by a Microlab 500 series dual syringe auto-titrator , and the 220 nm CD signal for each data point was collected at 25°C . Four measurements were collected for the wild-type and two for the R62A variant ( Table S1 ) . The data were averaged and normalized . Assuming that CheW wild-type and R62A mutant undergo a two-state unfolding mechanism , the fraction unfolded curve vs . [urea] for each variant was fitted to a six parameters equation [42]: ( 3 ) where y is the CD signal , yf and yu are intercepts , mf and mu are the slopes of the pre and post-transition baselines , m is a measure of dependence of ΔG on urea concentration , and is an estimate of the conformational stability of the protein in 0 M of urea . We used non-linear least square fit to calculate each of the parameters , Table 1 . We collected 3738 CheW protein sequences available from draft and complete genomes , using the August 2012 release of the MIST database [43] . Using HMM provided by the authors [44] and HMMER 2 . 3 . 2 [45] , we assigned the CheW sequences to classes [44] . We selected only sequences from flagellar systems ( 2553 sequences ) , and to avoid contamination by proteins with unassigned or unknown domains in addition to the CheW domain , we used a length filter . Sequences shorter than the Pfam [46] model for CheW ( PF01584 ) or 100 amino acids larger than the model were discarded . Only 368 sequences were discarded in this process . The 2185 CheW sequences selected were separated according to their chemotaxis classes in individual files . Each file was subjected to multiple sequence alignment using algorithm L-INS-I from the package MAFFT [47] . To avoid redundancy , sequences with more than 98% identity were removed from the dataset . The final dataset contained 1429 sequences . Fig . S1 shows the distribution of these sequences in chemotaxis classes . The sets for classes F1 and F7 were used to calculate the identities presented in Table S2 . The atomic coordinates of E . coli CheW were obtained from the NMR structure deposited on PDB ( PDB code: 2HO9 ) [13] . There are 20 frames in the PDB file , and the frame with the lowest alpha carbon RMSD relative to the average of all frames was selected . Standard protocols for solvation and neutralization were used to build the 64×91×70 Å simulation cell with a total of 36193 atoms . After 1000 steps of energy minimization , the frame at 40 ns of equilibration at 298K and NPT ensemble was selected as the starting point for production simulations of the wild-type protein and to build the in silico mutants R62A and E38A . To ensure that all three simulation systems ( wild-type , mutant R62A , and E38A ) were similar , only 200 steps of energy minimization were applied to each of the two simulations cells with mutant proteins . All simulations were performed with NAMD2 [48] using CHARMM22 [49] force fields for proteins and the TIP3P model for water [50] in the NPT ensemble . Temperature and pressure were held constant at 298 K and 1 atm using a Nose-Hoover Langevin piston [51] with a period of 100 fs and a decay time of 50 fs . The integration time-stepping was set to 2 fs under a multiple time stepping scheme [52] , with bonded and non-bonded interactions calculated at every step , and long range electrostatics interactions calculated at every other step . For the description of the long range forces , van der Waals forces had a cutoff of 12 Å , and the switching function started at 10 Å to ensure smoothness . Electrostatic interactions were calculated using particle mesh Ewald ( PME ) with a grid-point density of over 1/Å . For the wild-type and both mutant proteins , ten 90 ns-long , independent simulations were produced . In each simulation , atom velocities were reinitialized , guaranteeing independence between runs . The same simulation settings described in the equilibration section were used . The computation was performed using 512 nodes in the Newton Cluster at The University of Tennessee-Knoxville , with a performance of ∼33 ns/day . To calculate the frame-average RMSD per residue , we executed the following procedure: ( 1 ) from each of the ten simulations with the wild-type structure , the frames in which Arg62 and Glu38 formed a salt-bridge in geometry A were selected . ( 2 ) For each one of the ten sets of frames , the RMSD per residue was calculated against the initial frame , which is common to all simulations . ( 3 ) The RMSDs per residue were independently averaged over the number of frames in each set . The RMSDs were calculated using the VMD tcl command “rmsd , ” and all atoms were taken into consideration . The same procedure was executed for the simulations with the mutant R62A structure . However , to produce the ten sets of frames , the same number of frames selected from the wild-type simulations ( 64% ) was randomly selected from each of the ten independent simulations . Statistical significance was calculated using two-tailed t-tests for each residue independently .
Protein residues in proteins that are conserved over long evolutionary distance usually play the most critical roles in their structure . The signal transduction pathway for chemotaxis originated early in the evolution of bacteria and diversified into many distinct classes , in which the repertoire of interacting proteins can be quite different [44] . For example , in the F1 ( F stands for systems that control flagellar motors and a number represents a clade on the chemotaxis phylogenetic tree , see [44] for details ) class exemplified by Bacillus subtilis and Thermotoga maritima , the CheW protein interacts with chemoreceptors that are structurally different from those in the F7 class exemplified by Escherichia coli [54] . Furthermore , within a genomic dataset , protein sequences in each class are unequal in numbers and in phylogenetic relatedness , which further complicates analysis . In order to identify residues that are critical to the function of the CheW protein , we assigned the CheW sequences collected from the MIST database [43] to chemotaxis classes and found that F1 and F7 , the most abundant classes , are comparable in size ( Fig . S1 ) . Therefore , we performed detailed sequence analysis only on the CheW-F1 and CheW-F7 subsets . Earlier analysis of CheW sequences indicated that it is a relatively poorly conserved protein [55] . Thus , it was not surprising to discover that , among the five most conserved positions in each class , only two Gly residues are absolutely conserved in both classes ( Table S2 ) . Conservation of a Gly residue usually indicates that it performs unique structural role , either by allowing sharp turns and bends or its location in a spatially constrained environment [56] , [57] . Indeed , Gly63 is located at a critical turn on the CheW tertiary structure , and Gly57 is present in a beta sheet bend ( Fig . 1 ) . An unexpected finding , however , was a nearly absolute conservation of two charged residues ( Arg62 and Glu38 in the E . coli protein ) in the F7 class ( Table S2 ) . We therefore focused our investigation on the properties of CheW-F7 , which includes the E . coli CheW protein . Arg62 and Glu38 are in close proximity in the tertiary structure ( Fig . 1B ) . Both Arg62 and Glu38 ( along with some other residues ) have been implicated as functionally important in previous experimental studies with the E . coli protein . Mutations targeting Glu38 reduce the binding constant between CheW and the Tar chemoreceptor , making it a likely candidate for being located in the receptor-binding site [23] . Substitutions in residues in close proximity to Arg62 decrease the binding affinity between CheW and CheA; however , substitutions at Arg62 itself do not appreciably affect binding affinities for either CheW or CheA although they impair chemotaxis [23] . Thus , defining the role of this conserved residue remains a challenge , despite the fact that it has been approached by different experimental techniques [12] , [22] , [23] . Their physical proximity and their opposite charges suggest that Arg62 and Glu38 residues interact . Furthermore , the highest level of evolutionary conservation of both residues suggests this interaction is critical to protein function . A salt bridge between Arg and Glu can be inferred from an in-silico model if the pair of residues meet the following criteria: i ) the centroids of the side-chain charged groups are within 4 . 0 Å of each other; and ii ) at least one pair of carbonyl oxygen and side-chain nitrogen atoms are within 4 . 0 Å of each other . When the ion pair only meets the latter criterion , it is inferred to be forming a N-O bridge [58] . By these criteria , only 4 out of 20 in the ensemble of NMR models resolved for CheW from E . coli [13] identify a salt bridge between Arg62 and Glu38 . To explore these static models further , we performed ten independent MD simulations of 90 ns each; with a total of 450 thousand frames after an equilibration period of 30 ns ( see Materials and Methods ) . In our simulations , 84% of the frames met both criteria , and 11% met only the latter . In only 5% of the frames were neither of these criteria met . The temporal evolution of the distance between centroids of the side-chain charged groups is shown in Fig . S2 . In subsequent analysis , we found two distinct geometries for the salt bridge: ( A ) atoms NH1 and NH2 are within 4 . 0 Å of two distinct oxygen atoms in the Glu side-chain , and ( B ) both atoms NE and NH2 are within 4 . 0 Å from a different oxygen atom in the Glu side-chain ( Fig . 2 ) . In 64% of all frames the residues were in geometry A , and in 20% they were in geometry B . To confirm the existence of the salt bridge experimentally , we first attempted to measure the pKa for the wild-type CheW and for a mutant targeting Glu38 using pH titration . Unfortunately , wild-type CheW precipitated at pH less than pH 6 . 0 , which prohibited the use of this method , which requires a larger excursion of pH titration for Glu in proteins [59] . However , because “self-interactions” between residues in a protein molecule are known to contribute to protein dynamics , we attempted to examine the role of Glu38 and Arg62 residues by using NMR . First , we compared the 15N-HSQC spectrum of E38A and R62A mutants to that of the wild-type protein ( Table S3 ) . The results showed that the E38A mutation caused a global structural perturbation , suggesting that it introduces severe structural changes ( Fig . S3 ) . Consequently , we did not pursue further studies with this mutant . On the other hand , the R62A mutation caused only local structural perturbations ( Fig . 3 ) while disrupting the interaction with Glu38 . Residues of the R62A mutant that showed significant chemical shift changes are mainly located in β4–β5 , the C-terminus of the β-strand containing Glu38 ( β3 ) , and residues in close proximity to these limited regions ( Table S3 ) . To investigate the significance of the interaction between Glu38 and Arg62 in more depth , we measured the relaxation parameters of the backbone 15N nuclei in both the wild-type and the R62A CheW proteins . The average longitudinal relaxation rate R1 was 1 . 299 s−1 for wild-type and 1 . 295 s−1 for the mutant ( Fig . 4a ) . The longitudinal relaxation is caused by fluctuations at the NMR transition frequencies and reflects the NMR excited state lifetime , which is not altered by the substitution in the position Arg62 . The average transverse relaxation rate R2 was 14 . 62 s−1 for wild-type and 15 . 38 s−1 for R62A mutant ( Fig . 4b ) . The transverse relaxation reflects any events that cause dephasing of spins in the xy plane , such as rotational diffusion or chemical exchange . Although the absolute values are characteristic of each individual molecule , the substantial difference in the average transverse relaxation rate between the wild-type protein and the R62A mutant protein suggests that the substitution at residue Arg62 causes a subtle difference in rotational diffusion . The average R2/R1 value is 11 . 36 for wild-type and 12 . 02 for the R62A mutant and reflects the rotational correlation time of each protein . The slight increase in R2 in the R62A mutant , despite the almost identical in R1 values , implies that there might be motions on the microsecond-to-millisecond time scale induced by conformational exchange in the mutant protein that leads to line broadening . The order parameter S2 obtained from the isotropic model shows that the majority of the backbone amides are rigid , although the loops and the turns connecting the β-sheets show some dynamic behavior , and the N- and C-termini are highly flexible in both wild-type and R62A CheW ( Fig . 4d ) . This conclusion is in agreement with a previous report [13] . Overall , the wild-type and mutant proteins had the same backbone dynamics . In the model-free analysis , the phenomenological transverse relaxation rate constant , Rex , was found to make a significant contribution to achieving adequate fit of the 15N relaxation data during the calculations of the order parameter . This finding is in line with the previous results that suggest conformational motions in CheW that occur on the microsecond-to-millisecond time scale . For accurate characterization of the Rex term , a series of Carr-Purcell-Meiboom-Gill ( CPMG ) 39–41 relaxation dispersion experiments were performed on both 15N labeled wild-type and R62A CheW [31] , [37] , [38] . The Rex terms from R1/R2/NOE fitting were similar to the results from the individual CPMG measurements . The differences between Rex measured at τcp values ranging from 20 ms to 1 ms for wild-type and R62A CheW are shown in Fig . 4e , and the differences between these two constructs are shown in Fig . 4f . In both proteins , the majority of the backbone 15N spins showed no significant differences in their relaxation rate constants . However , some residues located in β4 and β5 , in loops , and in the two helical regions , showed relatively larger Rex values in the mutant protein . Furthermore , the R62A change increased the Rex value in some residues in loop1 , α1 , α2 , β4 and β10 , indicating that there are increased conformational-exchange motions on the microsecond-to-millisecond time scale in this region ( Table S4 ) . This finding suggests that the R62A substitution , which disrupts the interaction between Glu38 and Arg62 , decreases the stability of the second subdomain of the CheW structure . In addition , we detected no difference in the folding stability of the wild-type and R62A proteins in denaturation curves measured by circular dichroism as function of urea concentration ( Fig . 5 ) . The conformational stability of both variants was the same ( Table 1 ) . This result suggests that the Arg62 – Glu38 salt bridge does not influence overall protein stability and is not important in protein folding , in agreement with the results from the MD simulations and the NMR spectroscopy data . However , the difference in the slope of the pre-transition region of the unfolded fraction curves , mf , suggests that the Arg62-Glu38 bridge has a stabilizing effect when the protein is close to its native conformation ( Fig . 5 ) . To understand the contribution of the salt bridge to local backbone stability in more detail , we analyzed the distance between alpha carbons of all relevant residues in 10 MD simulations . We found that the salt-bridge in geometry A maintains the distance between the alpha carbons of residues Glu38 and Arg62 at 12 . 3±0 . 3 Å . All other conformations assumed by the ion pair , including the salt-bridge in geometry B and the N-O bridge , slightly shift the distance between the alpha carbons and increase their relative motion range ( Fig . 6 ) . The independent distribution of each conformation is shown on Fig . S4 . This result indicates that the maintenance of the correct distance between the backbone atoms of Glu38 and Arg62 is compromised if the residues do not form a salt-bridge in geometry A . To further validate this finding , we carried out ten independent simulations of 90 ns each for the in silico E38A and R62A proteins . Both mutations intrinsically forbid salt-bridge formation . In both mutant proteins , the distance between the backbone atoms , and specifically alpha carbons , was not restricted , as it was in the wild-type protein with the salt-bridge in geometry A ( Fig . 6 ) . Overall , shifts in distances between alpha carbons ( Fig . 6 and Fig . S4 ) were significant ( >1 . 5 Å ) . For example , a helical displacement of only 2 Å initiates the signaling cascade in the transmembrane chemoreceptors [60]–[62] . Thus , we conclude that the formation of the salt-bridge between residues Glu38 and Arg62 in a specific geometry maintains the positions of their corresponding backbone atoms in a stable relationship . Arg62 is located close to a proposed CheA-binding site , and Glu38 is located within a proposed chemoreceptor-binding site . Consistent changes in alpha carbon fluctuations calculated for each variant show an increase in the motion of the chemoreceptor-binding site relative to the kinase-binding site . Local changes in backbone positions relative to these sites were seen in all frames in which the interaction between Arg62 and Glu38 was not maintained in geometry A . As revealed by the analyses of the order parameter derived from the molecular dynamics simulations , and in agreement with the values calculated in NMR studies , this local change in backbone mobility is not a result of changes in overall protein dynamics in the pico-to-nanosecond time scale ( Fig . S5 ) . To examine the consequences of salt bridge disruption on the chemoreceptor- and kinase-binding sites , we analyzed the difference in frame-averaged root mean square deviation ( RMSD ) per residue between frames collected from the simulations of the R62A protein in comparison with those of the wild-type protein , with the salt bridge in geometry A . ( The Ala substitution at Glu38 disrupts the interaction of CheW with the receptor [23] , so we did not perform the same analysis for the E38A protein ) . To measure the fluctuation of the chemoreceptor-binding region relative to the CheA-binding region , we aligned the frames using only the backbone atoms of residues Ile55 to Val68 , which is a proposed CheA-binding site [12] , [23] . The frames with the salt bridge in geometry A were selected separately from each simulation , and the final frame-averaged RMSD per residue value is an average of the values independently calculated for each simulation . Because only 64% of the frames from the wild-type simulation had a salt bridge in geometry A , we randomly selected 64% of the frames from all ten R62A simulations . Overall , the R62A mutant protein was more dynamic than the wild-type ( higher frame-averaged RMSD per residue ) ( Fig . 7A ) . However , considering the fluctuation of the results from simulation to simulation , only a few residues were significantly more dynamic in the R62A protein than in the wild-type protein ( p-value<0 . 00002 ) ( Table S5 ) . More than half of these residues were found in the chemoreceptor-binding region ( Fig . 7B ) , a result which further supports our hypothesis . Taken together , these results suggest that the most important consequence of disrupting the salt bridge between Glu38 and Arg62 is an increase in fluctuation of the relative positions between the kinase and receptor binding sites on the CheW surface .
The results presented here provide a compelling explanation for the strong evolutionary pressure on residues Arg62 and Glu38 of the chemoreceptor scaffold protein CheW . These residues are invariant in all currently available CheW sequences from the most populated chemotaxis class F7 , which contains chemotaxis systems of representatives of diverse bacterial phyla [44] . Glu38 was previously suggested to participate in the interaction with chemoreceptors [23] , however , earlier studies failed to propose a specific role for Arg62 despite the fact that this residue was recognized as conserved and shown to be critical for chemotaxis [24] . Using MD simulations , we demonstrated that Arg62 and Glu38 can form a stable salt-bridge with a specific geometry . This result could not be obtained using any other experimental method , such as pH titration , given the dynamic properties of the CheW protein , which precipitates in pH values lower than 6 . 0 . Simulations with the R62A mutant show that disruption of the salt bridge does not compromise the overall structure or dynamics of the protein . However , it results in a detectable loss in maintaining a stable relationship between the chemoreceptor- and the kinase-binding regions . NMR experiments showed that the R62A substitution only perturbs the CheW structure locally , in agreement with the MD results . In contrast , the chemical shifts of several important residues in the E38A mutant protein indicate that this position is important for the structural integrity of at least one ( C-terminal ) subdomain of CheW . Furthermore , the NMR relaxation-dispersion experiment suggests that there are local motions on microsecond-to-millisecond time scale in the R62A mutant . The increase in the rotational correlation time for the R62A mutant protein suggests that this substitution may lead to an overall subtle expansion of the molecule . Taken together , we propose that motions in pico-to-nanoseconds time scale explicitly caused by the disruption of the Glu38-Arg62 as observed in the MD simulations are likely to cause subtle changes in the protein structure without changing the overall protein stability . These changes are likely to allow new vibration modes in the microsecond-to-millisecond time scale with larger excursions than encountered in the wild-type protein , as suggested by the increase in the rotational correlation time . In addition , the analysis of urea-induced unfolding curves showed that disruption of the salt bridge does not affect the conformational stability of the protein . However , the difference in the slope of the pre-transition part of the curve of the wild-type and the R62A variant proteins suggests that the interaction between Arg62 and Glu38 provides a stabilizing role in the near native-protein conformations . Finally , there was a good agreement between the NMR and MD measurements of the order parameter for both the R62A and the wild-type proteins , which provided experimental validation of computer simulations . Our results show that the major structural difference in the R62A mutant is the destabilization of the relative position of the chemoreceptor-binding site relative to the kinase-binding site . On a larger scale , this translates into a relaxation of the precise orientation of the chemoreceptor relative to the kinase . Thus , the Glu38-Arg62 bridge is stabilizing , i . e . , it constrains flexibility and motion as do most of salt bridges [63] . We conclude that , although this salt bridge is not required for assembly of the signaling complex , it “tightens” the elements of the complex together thus enabling signal transduction . The stabilization provided by the Glu38-Arg62 salt-bridge appears to be required for chemotaxis . The retention of this feature in the entire chemotaxis class F7 attests to its importance . It is present in the CheW proteins in such important pathogens as Bordetella bronchiseptica ( BB2547 , CheW locus tag ) , Clostridium difficile ( CD0536 ) , Salmonella enterica ( t0957 ) , Pseudomonas aeruginosa ( PA0177 ) , Vibrio cholerae ( VCA1094 ) , Yersinia pestis ( YPO1667 ) and many others . However , whether or not the same mechanism is utilized in CheW proteins from other chemotaxis classes is unclear . For example , analysis of the crystal structure of the CheW protein ( TTE0700 , locus tag ) from of Thermoanaerobacter tengcongensis ( class F1 ) reveals the same Glu38-Arg62 salt-bridge ( residues Glu33 and Arg57 in the TTE0700 sequence ) [14] . However , neither of the two CheW proteins from Thermotoga maritima [64] ( both from class F1 ) have the conserved glutamate at the same position . Analysis of the crystal structures [65] , [66] and the NMR model [10] of T . maritima CheW2 ( TM0701 , locus tag ) suggests a similar mechanism through a salt-bridge formed by another Glu-Arg pair ( Glu31 and Arg58 in the TM0701 sequence ) . Thus , CheW proteins from other chemotaxis classes might utilize different amino acid positions for the same strategy - stabilizing the relative position between the chemoreceptor- and the kinase-binding sites . Recent advances in determining crystal structures of interacting CheW and CheA proteins and electron cryotomography of signaling arrays have provided static models of the entire chemotaxis signaling complex [65] , [67] . However , protein interactions are dynamic , not static . In recent years , a number of studies have aimed at improving our understanding of protein-protein interactions and their roles in biological processes by revealing their evolution and dynamic properties . ( For a review , see [68] ) . Our study builds on these recent developments . We still do not know the molecular mechanisms of signal transduction in the signaling complex , and studies of protein dynamics will provide a complement to other avenues of current research in chemotaxis . CheW appears to be a rapidly evolving and highly dynamic protein . These two features usually correlate: Properties that make proteins conformationally dynamic also facilitate rapid evolution [69] . CheW proteins from different prokaryotes share very little sequence similarity [55] , in striking contrast to their interacting partners - chemoreceptors and CheA [44] , [54] . Likewise , low sequence similarity and high diversification are observed in SH3 domains from eukaryotic scaffold proteins [16] , [70] that are topologically similar to CheW [15] . Therefore , the high conformational dynamics coupled with stabilization mechanisms of the type discussed here for CheW may be important universal properties of scaffold proteins that participate in assembling arrays of proteins involved in signal transduction . | Signal transduction is a universal biological process and a common target of drug design . The chemotaxis machinery in Escherichia coli is a model signal transduction system , and the CheW protein is one of its core components . CheW is thought to work as a scaffold protein that mediates the formation of the signaling complex with the CheA histidine kinase and the chemoreceptors . A mutation targeting a highly conserved residue , Arg62 , impairs chemotaxis while maintaining normal binding affinity for both partners , suggesting that CheW might play a more complex role than previously proposed . Using a series of molecular dynamics simulations , we found that the residue Arg62 can form a stable salt bridge with another highly conserved residue , Glu38 . We determined that this bridge does not contribute to the overall stability of the protein . However , the bridge stabilizes the local backbone structure of CheW and stabilizes the relative position of the binding sites for the chemoreceptor and kinase . The geometry of these interactions appears to be vital for the function of the signaling complex . We validated and complemented our computational findings using NMR spectroscopy and circular dichroism analysis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | Conformational Coupling between Receptor and Kinase Binding Sites through a Conserved Salt Bridge in a Signaling Complex Scaffold Protein |
SH3 domains are peptide recognition modules that mediate the assembly of diverse biological complexes . We scanned billions of phage-displayed peptides to map the binding specificities of the SH3 domain family in the budding yeast , Saccharomyces cerevisiae . Although most of the SH3 domains fall into the canonical classes I and II , each domain utilizes distinct features of its cognate ligands to achieve binding selectivity . Furthermore , we uncovered several SH3 domains with specificity profiles that clearly deviate from the two canonical classes . In conjunction with phage display , we used yeast two-hybrid and peptide array screening to independently identify SH3 domain binding partners . The results from the three complementary techniques were integrated using a Bayesian algorithm to generate a high-confidence yeast SH3 domain interaction map . The interaction map was enriched for proteins involved in endocytosis , revealing a set of SH3-mediated interactions that underlie formation of protein complexes essential to this biological pathway . We used the SH3 domain interaction network to predict the dynamic localization of several previously uncharacterized endocytic proteins , and our analysis suggests a novel role for the SH3 domains of Lsb3p and Lsb4p as hubs that recruit and assemble several endocytic complexes .
Families of peptide recognition modules ( PRMs ) , such as PDZ ( PSD-95/Discs-large/ZO-1 ) , SH2 ( Src homology 2 ) , and SH3 ( Src homology 3 ) domains bind peptide motifs within proteins to mediate protein–protein interactions required for the assembly of stable or transient biological complexes [1] . Thus , PRMs function to dynamically orchestrate biological pathways [1] . PRM family members can be identified directly from whole-genome sequences; therefore , it is possible to explore the recognition specificity of entire families using a variety of different experimental approaches [2] , [3] . Here , we explore the potential for mapping SH3 domain protein interaction networks by a Bayesian integration of results from three complementary experimental screening approaches: phage display , peptide array , and yeast two-hybrid analysis . In general , PRMs engage in protein–protein interactions by recognizing a core motif common to a domain family as well as additional ligand features that are more specific for each family member as is the case for PDZ domains [3] . Initial studies determined that SH3 domains bind to proline-rich sequences containing a core PXXP motif ( where X is any amino acid ) flanked by a positively charged residue [4] , [5] . Class I domains bind to ligands conforming to the consensus +XXPXXP ( where + is either arginine or lysine ) , and do so in an orientation opposite to that of class II domains , which recognize PXXPX+ motifs [6] , [7] . More recently , a number of alternative SH3 domain binding motifs have been identified , highlighting a wider breadth of SH3 specificities [8]–[11] . A general genome-wide analysis of PRMs would involve defining all the domains from their primary sequence , profiling their ligand-binding specificities in detail , predicting natural ligands for each domain , and mapping large-scale protein–protein interaction networks for each domain family . Here , we present the first high-resolution analysis of the yeast SH3 domain family . First , we used large-scale phage-displayed peptide libraries and extensive sequencing to generate high-resolution binding profiles , which we show accurately represent binding specificity across multiple SH3 domain ligand positions . Second , we used the resulting specificity profiles to identify putative interactions in the yeast proteome , which were subsequently confirmed using oriented synthetic peptide arrays . Third , we conducted large-scale yeast two-hybrid screens to generate a physical interaction network mediated by the set of yeast SH3 domains . Finally , the datasets were integrated using Bayesian networks to generate a global SH3 domain interaction map in yeast . Applying the integrated probabilistic network revealed an intricate array of SH3-mediated interactions amongst proteins that make up the endocytic machinery . Investigation and comparison of the dynamics of protein localization within this network showed that the modular network predictions of the spatiotemporal dynamics of several novel endocytotic components were correct . In particular , our analysis predicts that the SH3 domains from Lsb3p and Lsb4p interact with multiple endocytic proteins and therefore might act as hubs to cluster these proteins at sites of endocytosis .
We used peptide phage display to conduct a large-scale analysis of yeast SH3 domain specificity . We cloned DNA fragments encoding all 27 unique yeast SH3 domains using boundaries taken as the union of the domain lengths identified by three domain detection tools , BLAST [12] , PFAM [13] , and SMART [14] , with an additional ten amino acids included on either side of the overlapping domain region to facilitate cloning ( Table S1 ) . We expressed the domains in bacteria as proteins fused to the C-terminus of glutathione S-transferase ( GST ) and purified 24 out of 27 fusion proteins in a stable , soluble form . For two of the three recalcitrant domains , the C-terminal domain of Bem1p ( Bem1-2 ) and the Bud14p domain , we extended the sequence boundaries by examining the conservation of the domain regions across diverse fungal species . Based on this analysis , the domain boundaries for these two SH3 domains were extended , enabling the isolation of stable GST fusion proteins ( Table S2 and Text S1 ) . The third recalcitrant domain , the N-terminal domain of Sla1p ( Sla1-1 ) , could only be purified in tandem with Sla1-2 , and we denoted the dual domain protein as Sla1-1/2 . The GST-SH3 domain fusion proteins were used as targets in binding selections with a combination of random and biased peptide–phage libraries . We were successful in obtaining ligands for all SH3 domains except Bud14 and Cdc25 , and we isolated a total of 1 , 871 unique peptides . These results extend results from our previous study [2] and represent nearly an 8-fold increase in binding data . The set of aligned ligands for each domain was used to compile a position weight matrix ( PWM ) , which captures the frequency of amino acid preferences at each ligand position . Some ligand sets contained two distinct groups of ligands , and for these , two separate PWMs were compiled ( see below ) . From each PWM , a sequence logo [15] was generated to graphically represent the specificity at each amino acid position in each ligand set . To compare the binding specificities for the yeast SH3 domain family on a global scale , we clustered all domains in an unrooted tree based on their specificities ( Figure 1 and Figure S1 ) . We generated a set of 10 , 000 random peptides from the yeast proteome and used these to score each phage-derived PWM . The match of each PWM with each peptide was calculated using an information score yielding a 10 , 000-dimensional profile vector for each PWM . This profile vector describes the binding specificity in a cellular context by sampling the sequences that the domains would be exposed to in the cell . The similarity between PWMs was computed as the Pearson correlation between these vectors . PWMs were then clustered according to this similarity measure using a complete linkage algorithm . Hence , the tree represents natural yeast SH3 domain specificity as it clusters binding profiles based on endogenous protein ligands . Overall , our results are consistent with previous findings [2]; in addition , this higher resolution analysis reveals that each domain exhibits specificity across multiple ligand positions , including the core motif and flanking positions . Our specificity map reveals that the majority of yeast SH3 domains have specificities that can be defined as class I or II , with eight and 12 representatives , respectively . Notably , the SH3 domains from Cyk3p and Rvs167p , and a protein fragment containing the two N-terminal domains from Sla1p ( Sla1-1/2 ) exhibit dual specificity for both ligand classes ( Figure 1 ) . Furthermore , the specificity map uncovered many specificity profiles that do not cluster with either of the canonical classes . For instance , the SH3 domains of Bem1p , Hof1p , Myo3p , and Myo5p comprise a distinct cluster , which we denote as class III , and are characterized by their preference for poly-proline ligands , without the requirement for flanking charged residues . The SH3 domains of the paralogs Pin3p and Lsb1p exhibit dual specificity , recognizing class II ligands and a ligand set ( +XXXPXP ) that resembles class I ligands , but with different residue spacings , thus was left unclassified . The SH3 domain of Pex13p also exhibits dual specificity for class II ligands and for a second motif characterized by a positively charged residue located between proline residues , which does not fit any defined class . The specificity profiles for the paralogs Boi1p and Boi2p ( PXXXPX+ ) resemble class II , but with proline spacings that differ from the canonical binding motif , and have also been left unclassified . Finally , as observed previously [2] , the SH3 domain of Fus1p exhibits a unique specificity profile that does not include prolines . To compare the intrinsic specificities of yeast SH3 domains , we quantified the specificities using a specificity potential ( SP ) score , which was applied previously to the PDZ domain family [3] . The SP value summarizes the specificity observed in each column of a PWM as a numerical value ranging from zero ( least specific ) to one ( most specific; Table S3 ) . We had sufficient peptide data ( n>10 ) to calculate reliable SP scores for 26 distinct specificity profiles . By summing the SP score across all PWM columns , we calculated a total SP ( SPt ) score for each SH3 domain specificity profile . Most yeast SH3 domains exhibit similar intrinsic specificities with SPt values ranging from four to six ( Figure 2A ) . Furthermore , domains that recognize more than one class of ligands do so with approximately the same level of specificity for each class . This analysis reveals that the Cyk3p SH3 domain [16]–[18] has an unusually high SPt value for class II ligands , which stems from its strong preference for an Asp-Tyr motif downstream of the Arg residue of the canonical class II motif ( Figure 1 ) . To assess the specificity contribution from different elements in the binding profiles , we quantified separately the SP scores for the positions within or outside the core motif for the various specificity profiles ( Figure 2B ) . The core positions for classes I and II only contribute roughly half of the SPt value , with the other half being contributed by other positions that define distinct specificity niches . Analogously , residues outside the core positions contribute approximately the same level of specificity for the unique sets of ligands recognized by Lsb1/Pin3 and Boi1/Boi2 ( Figure 2B ) . For class III domains , we found that recognition of proline accounts for approximately 60% of the SPt . Taken together , these results highlight the importance of residues outside the core positions for mediating specificity in SH3 domain–ligand interactions . Phage display generally selects high-affinity ligands through an iterative panning process , and high-resolution PWMs have been used to predict preferences in selectivity that reflect binding affinities for PDZ domain–ligand interactions [3] , [19] . To assess the accuracy of our phage-derived data for SH3 domains , we examined the SH3 domain of Sho1p and determined the correlation of PWM scores derived from phage display to differences in Gibbs free energy ( ΔΔG ) derived from previous in vitro binding assays with synthetic peptides [20] ( Table S4 ) . We observed an excellent correlation between the two datasets ( r2 = 0 . 97; p = 7 . 8×10−5; Figure 3A ) , and a significant correlation was also observed for similar datasets for the SH3 domain of Abp1p [21] ( r2 = 0 . 73; p = 2 . 1×10−4; Figure S2 and Table S5 ) . For the SH3 domain of Sho1p , the correlation between binding affinity and PWM score match is proportional to the number of peptides used to generate the PWM , and good correlation is observed for datasets containing >30 peptides ( r2>0 . 8 ) . Notably , 22 of our SH3 domain specificity profiles are derived from 30 or more ligands , suggesting that the majority of our phage-derived data can predict accurately the relative in vitro affinities of peptide ligands for SH3 domains . The use of synthesized peptide arrays offers an alternative and independent approach to query PRM–ligand interactions . In an ideal scenario , unique peptides representing the entire proteome of the organism would be spotted onto an array and assayed individually for interactions with a PRM of interest . However , in practice , a filtering step is required to generate an array of manageable size . In a strategy dubbed WISE ( whole interactome scanning experiment ) , natural ligands for PRMs are identified by computationally scanning the proteome for sequence patterns similar to known ligands , and these are tested for interactions using synthetic peptide array ( SPOT ) technology . Proteome scanning can use regular expressions ( REs ) , which describe discrete text patterns , or PWMs , which describe probabilistic positional frequency-based patterns . Although both methods rely heavily on the quality of the information they are based upon , REs run a higher risk of missing candidate ligands ( higher false-negative rate ) , whereas PWMs often fail to catch strict position-specific rules . Following identification of putative natural ligands by either filtering method , the peptides are tested for interactions by SPOT . We used the WISE approach to generate a yeast SH3 domain interaction network independently by creating a set of 15 REs based on SH3 domain specificity profiles identified in this study and previously [2] , and searching for matches in the yeast proteome ( Text S1 ) . The stringency of the REs was set very low in order to maximize the number of putative ligands tested on the array . Although this approach potentially identifies a number of false positives , the goal is to capture as many interactions as possible , thus minimizing the number of false negatives . This analysis identified 2 , 953 peptides within 1 , 693 proteins ( almost one-third of all yeast ORFs; Table S6 ) . This defined set of peptides was synthesized on cellulose membranes according to a modified SPOT synthesis approach [22] . Subsequently , peptide arrays were screened for binding individually with 26 SH3 domains . In total , we identified 295 peptides that showed a positive signal with at least one SH3 domain ( Table S7 ) . Peptides identified by either PWMs or REs address the ability of a domain to bind to a ligand outside of its protein and cellular context , but the peptides are identified by independent computational approaches with different strengths and weaknesses . To address this , we used the PWMs to define a set of peptides of similar size to the one defined by the REs . Interestingly , this analysis revealed only an approximately 30% overlap between the peptide sets defined by REs and PWMs . To examine the PWM-defined peptides experimentally , we tested in the SPOT assay the ten peptides with the highest PWM score for each SH3 domain . Of the 230 PWM high-scoring peptides , 113 were not included in the original WISE interactome , and approximately 55 of these gave a significant SPOT signal with at least one SH3 domain ( Table S8 ) . The 55 peptides predicted by PWM but missed by RE that yielded a significant SPOT signal can be regarded as false-negative interactions for the RE approach; therefore , the false-negative hit rate for the RE-defined peptides appears to be approximately 20% . Notably , of the 230 PWM high-scoring peptides , 69 did not generate a SPOT signal , which suggests that the PWM false-positive rate is on the order of approximately 30% . The SPOT approach is semiquantitative , so we also examined the correlation between interaction signals and dissociation constants , but we found that , as reported previously [22] , it was much poorer than that observed with the phage-display score ( unpublished data ) . Thus , although SPOT assays can be used to validate PWM-predicted interactions , further development of the method is required to obtain highly accurate quantitative signals . Taken together , our SPOT analysis of the yeast SH3 interactome yields a weighted graph of more than 5 , 000 edges , which served as an additional source of semiquantitative information to be integrated into a map of yeast SH3 domain interactions . To complement the phage display and SPOT experiments , we performed large-scale yeast two-hybrid screens . We screened 22 yeast SH3 domain baits against a novel yeast activation domain ORFeome library [23] , which tests for interactions with full-length proteins , using an array-based approach as described previously [24] and repeating each screen twice ( Table S9 ) . In addition , 26 SH3 domain baits were screened against a randomly fragmented genomic library ( gDNA ) , which tests for interactions with protein fragments [25] ( Table S9 ) . In total , we identified 801 unique interactions , consisting of 241 and 587 interactions from the ORFeome or gDNA library screens , respectively ( Table S10 ) . Only 26 interactions were identified in both screens ( 10 . 7% or 4 . 4% of the interactions identified by the ORFeome or gDNA screens , respectively ) . Using the ORFeome screen , we identified an average of 11 . 0 interactions per SH3 domain , whereas we detected an average of 22 . 6 interactions per SH3 domain in the gDNA screen . One major reason for the difference in these numbers is that we sequenced approximately 200 positive single colonies from each gDNA library screen in an attempt to saturate the system . Furthermore , although we repeated the ORFeome screening twice , this is not expected to achieve complete saturation according to a recent analysis [23] . In total , we sequenced 3 , 965 yeast two-hybrid–positive colonies , and some interactions were captured multiple times ( 593 interactions were captured at least twice ) by each screening technique ( Table S10 ) . To assess the potential of identifying biologically relevant interactions , we examined the number of literature-validated interactions that were identified by each approach . To do so , we curated a comprehensive “gold-standard” set of 42 SH3 domain interactions from the literature ( see below ) . Within this gold-standard set , only five were identified by the ORFeome screen , whereas 28 were identified by the gDNA screen . Thus , with yeast SH3 domains , gDNA two-hybrid screening has a 2 . 5-fold lower false-negative rate than ORFeome analysis ( Figure S3 ) , which may reflect both that our screening of the gDNA library was more extensive and that it contains gene fragments corresponding to protein domains , which often behave better in the two-hybrid system [26] . Taken together , these results highlight the complementary nature of ORFeome and gDNA screening methods to experimentally identify protein interactions for PRMs . As yeast two-hybrid and phage display potentially query different regions in interaction space , we sought to determine the overlap between the two methods . The phage-derived PWMs were used to search the yeast proteome for matching peptide ligands based on a PWM-scoring algorithm . For each SH3 domain , the yeast proteins were ranked according to their associated PWM score . Subsequently , the fraction of yeast two-hybrid hits containing predicted ligands with a rank higher than a defined threshold ( x = 1 , 2 , …N , where N is the size of the yeast proteome ) was determined . We find that approximately 10% of two-hybrid positives rank among the top ten hits predicted by the PWM of the associated SH3 domain ( Figure 4 , dashed line ) . The fraction of yeast two-hybrid hits with peptide sequences ranked among the top ten PWM-predicted ligands is increased to more than 25% when considering interactions that are captured at least six times , suggesting that these interactions have a higher likelihood of representing bona fide SH3 domain ligands ( Figure 4 , solid line ) . The high fraction of yeast two-hybrid positives with high-scoring PWM matches , compared to those predicted for random interactions , suggests that the detailed binding specificity uncovered by phage-derived PWMs was recapitulated using the yeast two-hybrid system . Each experimental method has different strengths and biases , and the integration of data from independent techniques increases the accuracy of the resulting dataset substantially [27] . We generated a yeast SH3 domain protein–protein interaction network and used a statistical approach based on Bayesian networks [27] to assign each interaction a probability score . This score is based on the confidence level of the experimental data that defined the interaction benchmarked by the gold-standard set ( see Materials and Methods and Table S11 ) . A Bayesian networks formalism was chosen for the machine learning because it has been shown previously to perform well at integrating heterogeneous biological data [27] , [28] . The gold-standard set represents a list of manually curated interactions known to be mediated by a specific SH3 domain , compiled through an exhaustive literature search . Each interaction in the gold-standard set is supported by multiple experiments reported in one or more focused studies , which show the direct binding of the SH3 domain to its target , and its functional relevance . Each technique utilized in our analysis encompasses a quantitative measure: first , the phage-derived PWMs accurately represent relative binding affinities; second , interactions identified by SPOT peptide arrays can be binned and ranked based on intensity ( see Materials and Methods ) ; and third , interactions captured multiple times by yeast two-hybrid can be assigned a higher score than those captured only once . Furthermore , the different methods have complementary features . Whereas the phage display and SPOT peptide array signals correlate with and predict binding affinity , the yeast two-hybrid system identifies putative in vivo interactors of SH3 domains . We therefore integrated these datasets into a Bayesian model to identify highly likely SH3 domain–ligand interactions . All interactions in the gold-standard set were mapped specifically to an SH3 domain and , where applicable , to the peptide sequence within the interacting partner ( see Materials and Methods and Table S12 ) . We generated a negative set using random protein pairs under the constraint of never sharing or being in “adjacent” cellular compartments ( see Materials and Methods ) . To determine the sensitivity of each technique individually , we plotted their respective receiver-operating characteristic ( ROC ) curve , a standard assessment of accuracy , and examined the area under the curve ( AUC; Figure 5 ) . The phage-derived PWMs were found to exhibit the highest AUC ( 0 . 91; Figure 5A and Figure S4 ) , with the SPOT peptide array and yeast two-hybrid exhibiting a lesser value ( 0 . 85 in both cases; Figure 5B and 5C , respectively ) . Remarkably , the Bayesian network , which integrates the data from all three techniques , results in an AUC of 0 . 94 ( Figure 5D; p = 1 . 2×10−10; Figure S5 ) , suggesting that our probabilistic interaction network captures the vast majority of literature-validated interactions . The entire set of yeast SH3 domain–ligand interactions predicted by our Bayesian model is represented as a network diagram and summarized in table format ( Figure S6 and Table S12 ) . To assess how profile specificity translates into specificity at the level of the network , we computed for each SH3 domain , the fraction of its interactors in the Bayesian network that are targeted by at least one other domain ( Figure S7 ) . Our results show that many proteins ( 61% ) are targeted by only one SH3 domain . The other proteins ( 39% ) are predicted to bind to more than one SH3 domain . Furthermore , important differences between SH3 domains can be observed , some of them having very unique specificity ( e . g . , Fus1p SH3 domain ) , whereas others share most of their interactors with other domains . The latter is especially true for SH3 domains from paralogous proteins such as Boi1p/Boi2p , Lsb1p/Pin3p , Lsb3p/Lsb4p , and Myo3p/Myo5p . To study specificity further , we also distinguished the different predicted binding sites on each protein ( binding sites are predicted by the best PWM hit in the protein sequence ) , since a protein can be targeted by multiple SH3 domains but at different binding sites . Interestingly , the fraction of binding sites targeted by more than one SH3 domain is lower than the fraction of proteins targeted by more than one SH3 domain ( 29% against 39% ) , revealing that some proteins have multiple unique binding sites recognized by individual SH3 domains ( Figure S7 , grey bars ) . However , cases of possible competition are not completely removed by distinguishing the different binding sites . To assess the contribution of SH3 domains from the same protein and highly similar SH3 domains , we merged Bzz1-1 and Bzz1-2 , Sla1-1/2 and Sla1-3 , and the four pairs of close paralogs ( Boi1p/Boi2p , Lsb1p/Pin3p , Lsb3p/Lsb4p , and Myo3p/Myo5p ) , treating each of them as a single protein since they have highly similar specificity profiles . In this case , we found that 33% of all interactors are targeted by more than one SH3-containing protein in our network ( Figure S8 ) . As previously , we distinguished the different binding sites in each protein target and found that 23% of binding sites are targeted by more than one SH3-containing protein ( Figure S8 ) . Thus , the majority of interactions are expected to be insulated from competition effects , due to sequence differences among binding sites alone , though some competition among domains is likely . As one approach to assessing the biological relevance of interactions identified by the Bayesian model , we examined biological process annotation associated with the putative SH3 domain ligands , defined by Gene Ontology ( GO ) . We found a significant number of overrepresented biological processes known to be associated with yeast SH3 domain biology such as establishment of cell polarity and endocytosis ( p = 3×10−7 and p = 9×10−8 , respectively ) . Moreover , from a recently published set of approximately 60 known and putative endocytosis proteins , 29 were found to be connected with at least one SH3 domain in our interaction network [29] ( Figure S6 ) . In addition , by searching for highly interconnected nodes in the Bayesian interaction network , we identified a core of 31 proteins that engage in at least six interactions with each other ( k-core = 6; Figure S9 ) . Consistent with the GO term enrichment analysis described above , 14 of the proteins that emerge from the k-core analysis ( e . g . , Las17p , Myo3p , and Vrp1p ) have well-defined roles in endocytosis with a GO enrichment p-value of 5×10−8 [29] , [30] . Hence , we decided to focus on the SH3-mediated interactions underlying endocytosis in more detail . Endocytosis is a complex cellular process in which a dynamic array of protein interactions are sequentially coordinated to drive endocytic site initiation , membrane invagination , and vesicle scission [31] . Live-cell imaging analyses uncovered a detailed spatiotemporal map for the dynamic recruitment of numerous proteins to endocytic sites in budding yeast [31] , [32] . These studies proposed the existence of four dynamic protein modules that cooperate to drive vesicle formation: ( 1 ) the endocytic coat module , ( 2 ) the Wiskott-Aldrich syndrome protein ( WASP ) -myosin ( WASP/Myo ) module , ( 3 ) the scission ( or amphiphysin ) module , and ( 4 ) the actin module . Proteins in the endocytic modules arrive sequentially at sites of endocytosis with precisely defined dynamics and their assembly drives the steps of endocytic internalization . The first step in the endocytic pathway is the recruitment to the plasma membrane of the coat module proteins , which include clathrin , Sla1p , Pan1p , End3p , and Sla2p . The assembly of the coat module occurs prior to and independent of actin assembly . However , the subsequent movement of the coat proteins into the cell , and subsequent coat disassembly , are dependent upon actin polymerization . One to two minutes following coat module assembly , Las17p ( the yeast ortholog of WASP ) is recruited , which activates the Arp2/3 complex to promote actin assembly . The SH3 protein Sla1p is thought to inhibit the actin polymerizing function of Las17p . This inhibition appears to be relieved by the recruitment of members of the WASP/Myo complex , including Vrp1p and the SH3 proteins Bbc1p , Myo3p/Myo5p , and Bzz1p . Actin polymerization triggered by the WASP-myosin complex leads to recruitment of the actin module proteins , which include actin , Cap1p , Cap2p , Sac6p , Abp1p ( SH3 protein ) , and the Arp2/3 complex , leading to further actin polymerization . As the vesicle begins its movement into the cell , the scission module , consisting of Rvs161p and the SH3 protein Rvs167p , is recruited . Although the exact scission mechanism is unclear , the scission module promotes the release of the nascent endocytic vesicle [29] , [31] . In contrast to components of the coat module , proteins of the WASP/Myo module remain immotile at the plasma membrane as actin is being polymerized , and disassemble as the nascent vesicle is internalized [29] , [31] . Spatiotemporal characterization of protein dynamics by live-cell imaging has provided a detailed view of endocytosis , but our understanding of this pathway is far from complete . It has been established that numerous proteins arrive at sites of endocytosis in a choreographed manner , but it is not known how the sequential recruitment , assembly , and functions of endocytic proteins are achieved . Our Bayesian interaction network contains 29 of the 60 or so known yeast endocytosis proteins , including ten that contain SH3 domains . To gain insights into the roles of SH3-mediated interactions in endocytosis , we screened for putative ligands for these ten SH3 domains using our Bayesian scoring algorithm ( Table S13 ) . The interacting proteins were grouped with the respective protein modules described above , and the putative SH3-mediated interactions at each stage of endocytosis were determined ( Figure 6 and Table S13 ) . This analysis uncovered a vast array of putative SH3 domain–mediated interactions , with 53 connections among 19 known or putative endocytic proteins , and suggested that interactions are likely to become more prevalent as additional proteins are recruited to the endocytic site ( Figure 6 ) . Furthermore , the interaction network suggests that the majority of SH3 domain–mediated interactions are established 35 to 15 s prior to vesicle internalization ( Figure 6C to 6E ) . This timing suggests that SH3 domains play a particularly important role at the stages encompassing assembly of the WASP/Myo module , actin polymerization , membrane invagination , and vesicle scission . The network allows us to map potential interactions onto the temporal order of protein recruitment at the site of endocytosis , and these interactions likely mediate assembly of protein modules and coordinate activities between the modules ( Figure 6 ) . We therefore examined in greater detail the relationships between SH3 domain–mediated interactions and protein dynamics during endocytosis . For each protein , we summed Bayesian probability scores ( or interaction scores ) based on interactions with proteins from within its corresponding module compared to interactions with proteins from external modules ( Table S13 ) . This analysis revealed that proteins have the highest total interaction score for interactions occurring within the same module . This was the case for 11 of 13 endocytic proteins for which a suitable Bayesian probability score and dynamic data were available ( Table S13 ) . For instance , the network identified a large number of interactions between members of the WASP/Myo module ( Bbc1p , Bzz1p , Las17p , Vrp1p , Myo3p , and Myo5p ) , which arrive following the coat module , 35 to 25 s prior to vesicle internalization ( Figure 6C and 6D ) . Summing their Bayesian probability scores across all modules revealed that each of these proteins has the highest combined interaction score for interactions within the WASP/Myo module . This finding provides support for the conclusion that the SH3 domain–mediated interactions are required for the assembly of this module , and that interactions between these proteins are established upon their temporal recruitment to the endocytic site . Subsequent to the formation of an SH3 domain–mediated network within the WASP/Myo module , the network analysis points to the formation of an SH3 domain–mediated network within the actin module ( e . g . , Abp1p , Ark1p , Prk1p , and Sjl2p [33] ) , at 25 to 10 s prior to vesicle internalization ( Figure 6D and 6E ) . Interestingly , proteins from the actin module also appear likely to engage in many interactions with members of both the WASP/Myo and actin modules , suggesting extensive cross-talk between these two modules ( Figure 6 ) . However , the interaction scores for proteins within the same module were higher than those for proteins in different modules , underscoring the predictive potential of using interaction scores to place endocytic components into their respective modules ( Table S13 ) . Our network analysis , which incorporates both SH3 domain–mediated interactions and dynamics of endocytic proteins , suggests that members from the same endocytic module engage in tighter SH3 domain–mediated interactions and have similar spatiotemporal dynamics . This raises the possibility of predicting the dynamics of putative endocytic proteins based on their SH3 domain interaction profile . Thus , an uncharacterized endocytic protein is predicted to be part of the module within which it registered the highest interaction scores . For example , if an uncharacterized protein is implicated as a member of the WASP/Myo module because it has high scores with SH3 domains within the WASP/Myo module , then we predict that its dynamics will follow a similar pattern to those of other proteins in that module . Analogously , an uncharacterized SH3 domain protein would be predicted to be part of the module containing its best-predicted binding partners . To test our hypothesis , we quantitatively examined the protein dynamics of five uncharacterized endocytosis proteins ( Scd5p , Aim21p , Scp1p , Bsp1p , and Lsb3p ) , each of which had a high SH3 interaction score with at least one of the established endocytic modules ( Table S13 ) . We predicted that Scd5p , a protein first identified as a suppressor of defects in cells depleted of clathrin heavy chain ( Chc1p ) [34] , arrives with and is part of the late coat module ( with Sla1p , and Sla2p , etc . , but not with the early coat protein , clathrin ) and/or WASP/Myo module . We also predicted that Aim21p , a fungal-specific protein , is a component of the WASP/Myo module , and that Scp1p , a conserved member of the Calponin/transgelin family of actin-associated proteins [35] , is part of the actin module . Two closely related SH3 domain proteins , Lsb3p and its paralog Lsb4p , had high interaction scores across several modules , most significantly with early ( e . g . , coat and WASP/Myo ) and late ( actin ) modules ( Figure 6 and Table S13 ) . Notably , we observed that the score for a particular module did not exceed the median interaction score across all other modules by more than 2-fold . This unique pattern of interactions suggests that Lsb3p and Lsb4p may play a role to cluster and to coordinate the activities of several module components at the site of endocytosis . In addition , Bsp1p , an adapter that links the yeast synaptojanins , Inp52p and Inp53p , to the cortical actin cytoskeleton and participates in actin contractile ring function [36] , showed a similar interaction profile , and therefore , we speculated that it too might be a cross-module protein together with Lsb3p and Lsb4p . To test our predictions in vivo , each protein was C-terminally tagged with GFP and expressed from its endogenous locus in yeast cells . The dynamics of each protein were analyzed individually and in tandem with Abp1p-RFP . The dynamic localization analysis validated our approach for assigning proteins to endocytic modules based on their interaction scores ( Figure 7 and Table S13 ) . In agreement with earlier observations [37] , we found that Scd5p-GFP patches had a lifetime of 22±6 s ( Figure 7A ) . Simultaneous , two-color analysis with Abp1p-RFP , a marker for actin polymerization , revealed that Scd5p-GFP arrives prior to actin polymerization ( Figure 7B ) . However , Scd5p-GFP patches were immotile throughout their lifetime , like proteins of the WASP/Myo module ( Figure 7C ) . These dynamics establish Scd5p as a component of the WASP/Myo module with similar dynamics to Bzz1p , suggesting that it might participate in late coat formation and/or coordinate this module with the WASP/Myo module . Moreover , Scd5p was recently reported to have a role in phospho-regulation of the endocytic coat complexes and its spatial dynamics may have a role in this essential function [37] . For Aim21p-GFP , we observed that it is located in immotile patches with a lifetime of 10±1 s , similar to the patch dynamics reported for Bbc1p [29] ( Figure 7A and 7B ) . Furthermore , Aim21p-GFP arrives when actin begins to polymerize , as revealed by two-color analysis with Abp1p-RFP ( Figure 7C ) . Thus , as predicted , Aim21p localizes as a component of the WASP/Myo module ( Figure 6D and Table S13 ) . Scp1p is expected to be part of the actin module as it is predicted to bind the SH3 domain of Abp1p . Indeed , Scp1p-GFP formed patches with a lifetime of 15±2 s ( Figure 7A ) and colocalized with Abp1p ( Figure 7B and 7C ) [30] . These patch dynamics are indicative of proteins in the actin module . Scp1p patches had shorter lifetimes than Abp1p . However , Scp1p-GFP exhibited weak fluorescence intensity , which likely accounted for this lifetime decrease . Two-color analysis revealed strong colocalization between Scp1p and Abp1p with the fluorescence intensity of the patches peaking together ( Figure 7C; unpublished data ) . As mentioned above , Lsb3p and Lsb4p scored highly across all modules , predicting a long lifetime at the patch . As expected , Lsb3p-GFP patches were long-lived with a lifetime of 36±9 s ( Figure 7A ) . Lsb3p-GFP patches arrived at the cell cortex as an immotile patch , but showed an initial slow movement into the cell to a depth of about 200 nm . The initial slow movement was then followed by a fast , more randomly directed movement further into the cell ( Figure 7B ) . Two-color simultaneous imaging with Abp1p-RFP revealed that , like Sla1p , Lsb3p-GFP arrived early at endocytic sites but persisted late with the actin module proteins ( Figure 7C ) [30] . These dynamics are consistent with the prediction that the Lsb3p and Lsb4p SH3 domains interact with Sla1p and several actin module proteins ( Figure 6 ) . Thus , Lsb3p and Lsb4p appear to provide continuity in the context of a continuously evolving endocytic protein composition . Finally , we analyzed the dynamics of Bsp1p , which our model suggested interacts with proteins in all modules , similar to Lsb3p and Lsb4p . However , in contrast to the Lsb proteins , Bsp1p-GFP patches were short-lived with a lifetime of 13±2 s ( Figure 7A ) . Bsp1p-GFP colocalized well with Abp1p-RFP and displayed an Abp1p-like motility behavior ( Figure 7B and 7C ) [30] , suggesting that Bsp1p functions within the actin module . Two-color analysis revealed that Bsp1p consistently arrived approximately 1 to 2 s after Abp1p-RFP , in a manner similar to descriptions for Cof1p , Ark1p , or Prk1p [38] , [39] . Moreover , unlike other patch proteins , Bsp1p-GFP had an additional stable localization to the bud neck as described previously [40] . Bsp1p is not well characterized , and further studies are necessary to understand the nature of the discrepancy between its predicted interactions with proteins of multiple modules and its appearance only late in the pathway during the burst of actin assembly . Our SH3 domain network contains only approximately half of the 60 proteins implicated in endocytosis and , as underscored by the Bsp1p example , a number of SH3-independent interactions must control endocytosis protein localization . To emphasis this point , we also analyzed the dynamics of yeast twinfilin ( Twf1p ) , a highly conserved actin monomer-sequestering protein involved in regulation of the cortical actin cytoskeleton [41] , which was not predicted to bind to any SH3 domain . Similar to Scp1p and Bsp1p , Twf1p localized to the patch with a lifetime of 15±2 s ( Figure 7A ) . The patches were initially immotile at the cell surface but subsequently underwent a highly motile phase , after which the patch moved long range into the center of the cell ( Figure 7B ) , in a manner characteristic of proteins comprising the actin module . In summary , SH3 domain interactions are powerful predictors of spatiotemporal localization of yeast SH3 domain proteins . The putative SH3 domain–mediated interaction networks allowed us to accurately predict the dynamics of several previously uncharacterized proteins in the endocytic pathway and provided a detailed spatiotemporal map of the endocytic pathway ( Figure 8 ) .
We generated a specificity map for the SH3 domain family of budding yeast based on 1 , 871 unique peptide ligands isolated against 25 of the 27 domains . This map reveals that SH3 domains have a high level of intrinsic specificity and different domains recognize distinct sets of ligands . Notably , specificity was observed for ligand positions outside of the core positions , suggesting that SH3 domains utilize multiple features of their peptide ligands to achieve binding selectivity . A major challenge in functional proteomics is the development of accurate protein interaction networks . We have integrated the data from three independent screening techniques ( phage display , peptide arrays , and yeast two-hybrid ) into a Bayesian model to generate a yeast SH3 domain interaction map . Each technique has a semiquantitative measure that was captured by the probabilistic model . Our interaction map captures a significant proportion of literature-validated interactions and therefore serves as an accurate reference for additional in-depth studies of yeast SH3 domain biology . Proper interpretation and use of our interaction map requires consideration of additional factors such as cellular concentration , localization , and competition from other SH3 domains to identify physiologically relevant interactions . Applying our model to proteins involved in endocytosis revealed that there is a significant connection between the time at which a protein arrives at the endocytic patch and its predicted SH3 domain interactions . This correlation was used to accurately predict the spatiodynamics of several uncharacterized endocytic proteins . We found that Scd5p and Aim21p are both components of the WASP/Myo module , which drives vesicle internalization by nucleating actin filament assembly and generating myosin motor-based forces on the actin filaments [29] . Future studies will reveal how these proteins contribute to the function of the WASP/Myo module , but the presence of Scd5p in the WASP/Myo module may be important for its role in phospho-regulation of the endocytic machinery [37] . We also found that Scp1p , Bsp1p , and Twf1p are components of the actin module . Both Scp1p and Twf1p are known actin-binding proteins and may play a role in modulating actin dynamics at endocytic sites [35] , [41] . The novel dynamics observed for Lsb3p indicate that it is present across all modules . Based on conserved SH3 predictions and sequence homology , we propose that Lsb4p has similar dynamics . Their numerous predicted interactions , and their dynamics and association with multiple endocytic modules , suggest that Lsb3p and Lsb4p may play an important role in coordinating the activities of the various endocytic modules . The SH3 interaction predictions did not agree with the dynamics of Sla1p and Bsp1p . Sla1p appears in the coat module , whereas its predicted interactions are more consistent with it being a component of the WASP/Myo module . This may be explained by its established essential role in regulating the WASP/Myo module [29] . Perhaps Sla1p integrates its cargo adaptor role [42] with its roles in actin assembly to prevent premature actin nucleation , and perhaps its departure from the cell surface with the coat proteins separates it from the WASP/Myo proteins , further relieving its inhibition of actin polymerization . Unlike Sla1p , Bsp1p is less well studied and lacks any obvious homology with other proteins . Bsp1p has been linked to the actin module protein Sjl2p , which regulates phosphatidylinositol 4 , 5-bisphosphate levels [36] . Furthermore , Bsp1p plays a role in actomyosin ring function [40] , but it is unclear how this relates to its role at endocytic sites [40] . The delayed recruitment of Bsp1p to the actin module also suggests a role in endocytic site disassembly alongside Sjl2p , Ark1p , Prk1p , and Cof1p [38] , [39] . Given the conserved nature of endocytosis from yeast to human [31] , it will be of great interest to examine SH3 domain interaction networks in more complex organisms . We emphasize the feasibility of the approach presented here , given the recent discovery that orthologous protein modules generally have very similar specificity profiles [3] . Recent studies of PDZ domains have shown that PRMs are more specific than previously appreciated [3] , [43] , and we show that the same holds true for SH3 domains . The intrinsic specificity observed at the level of the protein domain itself suggests that there is significant selective pressure driving the domain into a specificity niche not utilized by other domains . As postulated previously [20] , an interplay between positive specificity selection acting on the protein interaction module and negative selection acting upon its cognate ligands would ensure high specificity without the requirement of a high-affinity interaction . It appears that such specific interactions have evolved and are necessary to create the dynamic and intricate signaling pathways required for cellular functions .
For cloning , the SH3 domain boundaries were defined as the union of the domain regions identified by BLAST [12] , PFAM [13] , and SMART [14] , plus an additional ten amino acids ( where applicable ) on either side as described previously [3] . DNA fragments encoding the identified domains were amplified from S . cerevisiae genomic DNA by the polymerase chain reaction ( PCR ) and cloned into a vector designed for the expression and purification of SH3 domains fused to the C-terminus of glutathione S-transferase , as described [44] . All plasmid constructs were verified by DNA sequencing . Phage-displayed peptide libraries ( >1010 unique members ) fused to the N-terminus of the gene-8 major coat protein of M13 filamentous phage were used to select peptide ligands for the collection of purified GST-SH3 fusion proteins . All domains were first screened using a random decapeptide library ( X10 , where X is any amino acid ) . Domains that failed to select peptides with the decapeptide library were subsequently screened using a biased peptide library ( X6-PXXP-X6 , where P is proline ) . Three SH3 domain proteins ( Cyk3 , Lsb4 , and Sla1-1/2 ) were also tested using a biased library containing a fixed positive charge ( X7-R/K-X7 , where R and K are arginine or lysine , respectively ) . Phage display selections were carried out as described [44] . Individual binding clones were tested for positive interactions with cognate yeast SH3 domains by phage ELISA as described [44] . The sequencing of approximately 3 , 000 clones resulted in the isolation of 1 , 871 unique peptide sequences , which were manually aligned by an expert ( RT ) . The phage library used to select peptides for each domain is indicated in Tables S1 and S2 . As some peptide files contain peptides selected from different libraries ( Cyk3-class II , Lsb4 , and Sla1-1/2-class II ) , the library from which each peptide was isolated is also indicated in each sequence file . For each SH3 domain , the set of peptide ligands was used to create a binding profile statistical model as a PWM . The specificity potential ( SP ) for a given column ( position ) of a PWM was calculated as is done for the letter height in a sequence logo [15] , except normalized to range from 0 to 1 instead of 0 to 4 . 32 ( log 20 ) . A SP value of one means the given PDZ domain is completely specific for a single amino acid at that position , and a value of zero means that there is no preferred amino acid at that position . We have also included a p-value to assess the statistical significance of these scores . The p-values were computed by statistical sampling: for each PWM , we generated 107 sequences of N randomly chosen amino acids , with N equal to the number of different peptides used to build the PWM . For each sequence , we computed the SP score , and from the distribution of SP scores , we computed the p-value of the SP scores for each column in the initial PWM ( Table S3 ) . Peptide arrays were semi-automatically prepared on cellulose- ( 3-amino-2-hydroxy-propyl ) -ether membranes [22] ( CAPE membranes ) using a SPOT robot ( Intavis ) and the standard SPOT synthesis protocol [45] . Array design was generated using the in-house software LISA . To exclude false-positive spots in the incubation experiment , all cysteine residues were replaced by serine . The CAPE membranes were used because of the better signal to noise ratio in the incubation experiments . The peptide arrays were incubated with the GST-SH3 fusion proteins , as described [22] . Analysis and quantification of membrane-bound GST-SH3 fusion proteins was carried out using a chemiluminescence substrate and a Lumi-Imager ( Roche Diagnostics ) . Analysis and quantification of SPOT signal intensities ( SI ) were executed with the software Genespotter ( MicroDiscovery ) following previously described rules [46] . DNA fragments encoding SH3 domains were amplified by PCR from a S . cerevisiae genomic DNA library , using sequence specific primers fused to common sequences used for homologous recombination cloning . Specifically , the forward primer was composed of the bait-specific primer and a 23-nucleotide common sequence ( CGACCCCGGGAATTCAGATCTAC ) , which is homologous to the upstream sequence of the SpeI site on pPC97 [23] . The reverse primer was composed of the bait-specific primer and a 23-nucleotide common sequence ( CGGGGACAAGGCAAGCTAAACTA ) , which is homologous to the 5′ of the KanMX6 cassette . The KanMX6 cassette was amplified by PCR with the forward primer ( TTTAGCTTGCCTTGTCCC ) and the reverse primer ( ATAGATCTCTGCAGGTCGACGGATCCCCGGGAATTGCCATTTTTCGACACTGGATGGC ) , using a KanMX6 cassette carrying plasmid , p2076 , as template . Along with the PCR-amplified KanMX6 cassette and the SpeI-cut pPC97 , bait coding sequence PCR product was transformed into Y8930 ( MATα trp1-901 leu2-3 , 112 ura3-52 his3-200 gal4Δ gal80Δ LYS2::GAL1-HIS3 GAL2-ADE2 met2::GAL7-lacZ cyhR ) , which was generated from PJ69-4α [47] . G418 positive yeast transformants were selected on SD-Leu+G418 medium , and yeast DNA was purified and transformed into Escherichia coli DH5-α . Constructed plasmid was purified from kanamycin-positive clone , verified by DNA sequencing , and transformed into Y8930 for Y2H screening . The whole library was assembled in an 1 , 536-spot array format on agar plates with each clone represented twice . ORFeome Y2H screening was performed as described [24] with some modifications . The optimal concentration of 3-amino-1 , 2 , 4-triazole ( 3-AT ) was tested for each bait before performing the screen . SD-Leu-Trp-His+3-AT selective medium was used for screening . Plates were incubated at 30°C for 5 to 10 d before scoring positive colonies . All pOBD plasmids were taken from Tong et al . [2] , and the pBDC plasmids were cloned by homologous recombination as described above ( Table S9 ) and verified by sequencing . All bait plasmids were transformed into Y8930 . A genomic DNA library [25] was transformed into Y8800 ( same genotype as Y8930 except opposite mating type ) . Screening was performed by mating methods as described previously [48] on SD-Leu-Trp-His+3AT plates . Up to 192 positive single colonies were picked from each screen . The identity of each positive colony was determined by colony-PCR and sequencing . To compile a comprehensive list of yeast SH3 domain–ligand interactions supported by one or more experiments ( referred to as the gold-standard set ) , we used a combination of automatic text mining and database searches to retrieve abstracts from the literature . The DOMINO database , specialized in domain–peptide interactions , already contained 22 entries for yeast SH3 domains , curated according to the MIMix standards from 14 papers [49] , [50] . A text-mining approach looking for co-occurrence in the abstract of names of yeast proteins together with “SH3” and a list of nouns and verbs indicating interactions yielded only two papers containing relevant information after manual inspection . An additional 19 papers were captured by manual searching PubMed ( http://www . ncbi . nlm . nih . gov/sites/entrez ) and the Saccharomyces Genome Database ( SGD; http://www . yeastgenome . org/ ) [51] , [52] . These 21 new papers , which were not already present in DOMINO , were read , and the information supporting interactions mediated by SH3 domains was captured in a MIMix format . A total of 56 new interactions were added to the DOMINO database by this curation effort . Since some interactions are supported by more than one report , this amounts to a total of 41 nonredundant interactions mediated by SH3 domains and supported by at least one experiment . One paper reported a yeast two-hybrid interaction between Las17p and a protein fragment encoding both SH3 domains of Bzz1p . As the SH3 domains were not tested individually for interaction with Las17p , we counted the interaction twice to account for the two SH3 domains , thus resulting in 42 total SH3-mediated interactions [53] . The Bzz1 domains were tested individually by GST pull-down and Western blot analysis , and both domains interact with Las17p ( A . Soulard and B . Winsor , personal communication ) . The curated gold-standard list is contained in Table S11 . The peptides from phage display were converted into position weight matrices ( PWMs ) by calculating the probability of occurrence for each amino acid at each position . Despite the large number of peptide sequences , we still substantially undersampled sequence space , and hence added pseudocounts . We scaled the number of pseudocounts added by the entropy of each position [54] . Each matrix was used to scan the yeast proteome to identify the best matches . We used the MOTIPS analysis pipeline to identify possible binders for each domain . Only the proteome-scanning module of the pipeline was utilized , which performs a highly optimized search in the yeast proteome for optimal matches to a given PWM . It works in an analogous fashion to earlier proteome scanners ( e . g . , the Scansite server ) [55] . We employed the Bayesian Network algorithm as implemented in the WEKA 3 . 4 . 13 Java libraries [56] . All pre- and post-processing of the data was carried out using custom code written in Perl and Java . Bayesian networks can efficiently integrate different types of data and accurately estimate the probability of interactions based on different experiments [27] . The different data sources were first preprocessed as follows: the Y2H hits were put in one of two bins , depending on whether the associated clone was found once or more than once and given scores of one or two , respectively . The resulting discrete data were then fed directly into the learning algorithm . In the preprocessing step , the SPOT peptide binding data was discretized into four bins . The discretized data were then used as one feature of the learning algorithm . To ensure a reliable set of gold-standard true-positive interactions for efficient machine learning , we used the curated list of 42 bona fide domain–peptide interactions for the yeast SH3 domains deposited in the DOMINO database , as described above [49] . We generated the true-negative set by using the “random with constraints” logic . Specifically , we included only pairs of proteins where protein A is annotated to localize to the cell membrane and where protein B is annotated to localize to the nucleus . Proteins with overlapping annotations were excluded as well . Although the first member of each gold-standard negative set was chosen to be one of the proteins containing SH3 domains , its interacting partner was under no such constraint . Since the proportion of real interactions is very low in the space of possible interactions , one can use random domain–ligand pairs to get a set likely to contain only negatives . However , we improved upon this set by filtering out only those pairs that do not occur in known interaction databases and are annotated to occur in nonadjacent cellular compartments . Specifically , we included only pairs of proteins where protein A is annotated to localize to the cell membrane and where protein B is annotated to localize to the nucleus . Proteins with overlapping annotations were excluded as well . Performance of each data source was evaluated using the AUC ( area under the curve ) in the ROC curve . This corresponds to an evaluation of how well each data source corresponds to the gold-standard data . Finally , the performance of the Bayesian data integration was assessed using the AUC in a ROC curve analysis with 10-fold cross-validation . Ten-fold cross-validation corresponds to splitting the gold standard into a training ( 9/10 ) and a testing ( 1/10 ) set ten times in a rotating fashion and evaluating its accuracy for each split . Using the gold-standard set , we classified the discretized input data into the “True” ( interacting ) and “False” ( noninteracting ) labels as well as a probability score of the interaction . We report all interactions assigned a probability score of >0 . 6 ( Table S12 ) . The networks were created using Cytoscape 2 . 6 [57] . On the basis of affinity data for the Sho1p and Abp1p SH3 domains , we estimate that this cutoff represents a dissociation constant of approximately Kd = 1 . 5 µM . Yeast strains were grown at 25°C in standard rich medium ( YPD ) or synthetic medium ( SD ) supplemented with appropriate amino acids . GFP tags were integrated chromosomally to generate C-terminal fusions of each protein , as described [58] . All strains expressing fluorescent fusion proteins had growth properties similar to the corresponding untagged strains . For microscopy , cells were grown in SD medium without tryptophan ( to minimize autofluorescence ) at 25°C until early log phase . Cells were attached to coverslips coated with concanavalin A , which were sealed to slides with vacuum grease ( Dow Corning ) . Imaging was done at room temperature using an Olympus IX81 or IX71 microscope equipped with 100× NA 1 . 4 objectives , and Orca II cameras ( Hamamatsu ) . Simultaneous two-color imaging was done using an image splitter ( Optical Insight ) to separate the red and green emission signals to two sides of the camera sensor using a 565-nm dichroic mirror , and 530/30-nm and 630/50-nm emission filters . To excite GFP or RFP , we used a 488-nm Argon ion laser ( Melles Griot ) or a mercury lamp filtered through a 575/20-nm filter , respectively . The excitation beams from these two light sources were combined using a beam splitter . After each experiment , images of immobilized microbeads that fluoresce at both green and red wavelengths were captured . These images were used to align the cell images . Image analysis was done as described [30] . Tracking of patches was done from single-color GFP movies to achieve the best signal-to-noise ratio . ImageJ ( http://rsbweb . nih . gov/ij/ ) was used for general manipulation of images and movies . | Significant diversity exists in protein structure and function , yet certain structural domains are used repeatedly across species to execute similar functions . The SH3 domain is one such common structural domain . It is found in signaling proteins and mediates protein–protein interactions by binding to short peptide sequences generally composed of proline . To investigate both the generality and selectivity of peptide binding by SH3 domains , we examined peptide specificity for almost all SH3 domains encoded within the proteome of the budding yeast , Saccharomyces cerevisiae , using a range of experimental methods . We found that although most of the intrinsic binding specificity for SH3 domains can be summarized by the two previously described canonical binding modes , each individual SH3 domain that we studied utilizes unique features of its cognate ligand to achieve binding selectivity . Moreover , some domains exhibit binding specificities that are distinct from the two canonical classes . We integrated peptide-SH3 domain binding data from three complementary screening techniques using a Bayesian statistical model to generate a protein–protein interaction network for the budding yeast SH3 domain family . This network was highly enriched in endocytosis proteins and their interactions . By examining these interactions in detail , we show that our SH3 domain network can be used to predict the temporal localization of several previously uncharacterized proteins to dynamic complexes that orchestrate the process of endocytosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"computational",
"biology/sequence",
"motif",
"analysis",
"biotechnology/protein",
"chemistry",
"and",
"proteomics",
"cell",
"biology",
"computational",
"biology/genomics",
"computational",
"biology",
"computational",
"biology/systems",
"biology"
] | 2009 | Bayesian Modeling of the Yeast SH3 Domain Interactome Predicts Spatiotemporal Dynamics of Endocytosis Proteins |
The HMT3522 progression series of human breast cells have been used to discover how tissue architecture , microenvironment and signaling molecules affect breast cell growth and behaviors . However , much remains to be elucidated about malignant and phenotypic reversion behaviors of the HMT3522-T4-2 cells of this series . We employed a “pan-cell-state” strategy , and analyzed jointly microarray profiles obtained from different state-specific cell populations from this progression and reversion model of the breast cells using a tree-lineage multi-network inference algorithm , Treegl . We found that different breast cell states contain distinct gene networks . The network specific to non-malignant HMT3522-S1 cells is dominated by genes involved in normal processes , whereas the T4-2-specific network is enriched with cancer-related genes . The networks specific to various conditions of the reverted T4-2 cells are enriched with pathways suggestive of compensatory effects , consistent with clinical data showing patient resistance to anticancer drugs . We validated the findings using an external dataset , and showed that aberrant expression values of certain hubs in the identified networks are associated with poor clinical outcomes . Thus , analysis of various reversion conditions ( including non-reverted ) of HMT3522 cells using Treegl can be a good model system to study drug effects on breast cancer .
A major challenge in systems biology is to uncover dynamic changes in cellular pathways that either respond to the changing microenvironment of cells , or drive cellular transformation during various biological processes such as cell cycle , differentiation , and development . These changes may involve rewiring of transcriptional regulatory circuitry or signal transduction pathways that control cellular behaviors . Such information is of particular importance for seeking a deep mechanistic understanding of cellular responses to drug treatments in various diseases , offering a more holistic view of both microscopic and macroscopic changes in the cellular functional machinery than has been available from traditional analyses which usually focus only on finding differential markers or close-up analysis of changes in a handful of molecules constituting parts of some selected pathways of interest . Network-based differential analysis naturally requires the availability of multiple networks each in principle corresponding to a specific biological condition in question , that are then topologically rewired across conditions [1] . However , most existing computational techniques for reconstructing molecular networks based on high-throughput data cannot capture such dynamic aspects of the network topology; instead , they represent the networks as an invariant graph . For example , it is common to infer a single invariant gene network using microarray data obtained from samples collected over time or multiple conditions . More sophisticated methods such as a trace-back algorithm [1] and DREM [2] , [3] do emphasize uncovering the dynamic changes of a network over time using time series data , but limitations in these algorithms allow only certain kinds of dynamic behaviors , such as “active path” [1] or bifurcating sequence of transcriptional activations [2] . Moreover , such methods are heuristic in nature and do not offer statistical guarantees on the asymptotic correctness of the inferred “transient” components in the network , making the results difficult to withstand the harsh standard on stability and robustness when sample quality and size become less ideal , as we face in the analysis to be conducted in this paper . Indeed , a number of in-depth investigations of disease models have suggested that over the course of cellular transformation in response to microenvironmental changes due to disease progression or drug-induced reversion , there may exist multiple underlying “themes” that determine each molecule's function and relationship with other molecules [4] , [5] . As a result , molecular networks at each cellular stage are context-dependent and can undergo systematic rewiring ( Figure 1 ) . For example , strong evidence of alterations of various pathways have been reported in the HMT3522 progression series of breast cells when malignant T4-2 cells were phenotypically reverted by various drugs , albeit only manifested by a small number of well-known signaling molecules as discussed below [6]–[8] . In this paper , we conduct an in-depth study of the structural changes in the gene regulatory networks underlying each cell state in both the non-reverted and the reverted HMT3522 progression series of breast cells . The HMT3522 cells have been shown to be an excellent model system for studying the roles of tissue architecture , microenvironment and signaling molecules involved in the nonmalignant and malignant growth and behaviors of breast cells , including the potential of various factors to cause phenotypic reversion of malignant cells to nonmalignant states . These cells originated from a nonmalignant human breast epithelial sample , HMT3522 [9] , [10] . HMT3522-S1_LBNL ( S1 ) cells are from early passages which are nonmalignant and dependent on exogenous epidermal growth factor ( EGF ) to grow . HMT3522-T4-2_LBNL ( T4-2 ) cells were generated from S1 cells by a multi-step process: 238 passages in medium without EGF followed by transplantation into a mouse which generated a tumor , and T4-2 cells were isolated from the serial passage of this tumor; thus T4-2 cells are malignant and tumorigenic [10] . Interestingly , when cultured in three-dimensional ( 3D ) laminin-rich extracellular matrices ( lrECM ) , S1 cells form polarized acinus structures with a central lumen which resemble the terminal milk-secreting alveolar units in normal breasts [6] , [11] , whereas T4-2 cells form disorganized structures under the same conditions . Signaling molecules such as EGFR , β1-integrin , PI3K , and MAPK are overexpressed in T4-2 cells relative to their levels in S1 . Crosstalk between these molecules plays pivotal roles in defining malignant behaviors of T4-2 cells , and downmodulation of them causes phenotypic reversion of T4-2 cells into growth-arrested , normal-looking cells ( also called T4R cells later ) which form structures resembling S1 acini but often without the lumen [8] . Other molecules , such as TACE or Rap1 , have also been shown to be important for reversion of T4-2 cells [12] , [13] . NFkappaB was identified as one of the transcriptional regulators involved in disorganization of T4-2 cells [14] . Despite significant efforts to delineate key signaling events responsible for phenotypic reversion of these malignant breast cells , many questions remain . For example , are T4-2 cells reverted by inhibitors of different molecules intrinsically the same ? What is involved in the phenotypic reversion of T4-2 cells at the systems level other than a few genes directly related to the signaling molecules mentioned above ? One classical approach to address these questions is to identify genes differentially expressed between different cell states . While this can lead to some information about marginal effects of the genes in a particular stage of cancer progression or reversion , it cannot yield insight into the underlying regulatory mechanisms that govern interaction of genes with one another to carry out complex cellular processes . Instead , we propose a network-based differential analysis , by reverse engineering gene regulatory networks of various conditions of the breast cells to depict a fuller picture of regulatory mechanisms of the cells . Many methods , as reviewed in [15] , [16] , have been proposed for reconstructing gene networks using gene expression microarray data . Most of them [17]–[19] , however , rely on the statistical assumption that the samples in question were independent and identically distributed ( i . i . d ) , and thus they either lead to estimation of a single network by pooling data from all the samples together , or lead to estimation of a network for each cell state independently . Since the breast cells in this study came from non-reverted HMT3522 cells as well as various conditions of the reverted cells , the regulatory mechanisms in different cell states can be significantly different; therefore , pooling data from different cell states together to estimate one single network does not reveal networks in their full depth . On the other hand , reconstructing a network specific to each cell state independently of the other ones can be statistically inaccurate due to a small sample size for each cell state . Recently , time-varying network detection methods have been proposed that allow information sharing across time and can thus recover a sequence of networks even with small sample sizes [20]–[23] . For example , Song et al . proposed a time-varying dynamic Bayesian network method to estimate a chain of evolving networks over time [22] . However , these methods estimate networks that evolve as a chain of networks over time , not as a series of networks shared by the tree-shaped phenotypic relationships as shown in Figure 1 . Due to the unique challenges we encountered to reconstruct networks that rewire over the tree-shaped phenotypic relationships , we recently proposed Treegl [24] , a network reconstruction algorithm that can effectively and jointly recover rewiring regulatory networks present in multiple related cell states . Our approach can not only recover a distinct network for each cell state and reveal sharp differences among networks for different cell states , but also capture and leverage similarities of the networks in the cell states nearby in the phenotypic tree , thereby leading to more accurate estimation of gene interactions in small sample size scenarios . This new angle of estimating networks can reveal information that has not been mined in traditional analysis . In this paper , we conduct an extensive network analysis of non-reverted HMT3522 cells ( normal S1 and malignant T4-2 cells ) as well as three different conditions of reverted T4-2 cells using gene expression microarray data obtained from these cells . It is notable that the same set of the gene expression data was first described and used in our previous work published in [24] , however , our focus then was to report the novel methodology behind the Treegl algorithm , but not a thorough biological analysis of the HMT3522 series of cells from which the gene expression data was generated . In this current work , we focus more on the biological findings discovered by a more comprehensive network analysis of the data using Treegl and other bioinformatics tools , and aim to provide better biological insights and understandings of the various breast cell states in the HMT3522 series . In particular , we estimated the network specific to each cell state using Treegl . Our results showed that while the S1-specific network contains predominantly nonmalignant pathways , the T4-specific network contains various cancer-related pathways , both findings consistent with biological evidence [4] . Furthermore , we found that the networks specific to various conditions of the T4-2 reverted cells are enriched with pathways suggestive of compensatory effects . In the T4-2 cells reverted by inhibition of either EGFR or β1-integrin , signaling pathways downstream of EGFR or β1-integrin , mainly via the PI3K-AKT-mTOR axis , are upregulated . Similarly , in the T4-2 cells reverted by either PI3K or MAPKK , we observed upregulation of the pathways both upstream and downstream of PI3K . These results are supported by clinical evidence showing patient resistance to the same anti-breast cancer drugs as we used in the study . Moreover , the compensatory signaling is also observed in the differential network of the T4-2 cells reverted by MMPIs , which involves genes participating in protein catabolic processes . Together , our findings suggest a common resistance mechanism employed by breast cancer cells to antagonize drug effects . Finally , in order to identify potential novel drug targets , we also investigated hubs ( i . e . , genes with high degrees , see details in Materials and Methods ) in the differential networks of the breast cells , and characterized specifically three hubs ( NEBL , HBEGF , and PAPD7 ) whose aberrant expression values are linked with the worst survival outcomes in the breast cancer patients to provide insight into their functional significance on the growth and development of breast cancer cells . Our data suggest that Treegl when applied to an effective disease model system , such as the HMT3522 cells , may potentially become an effective tool for elucidating disease mechanism and discovering novel drug targets , and thus help make personalized medicine possible .
However , our goal is not estimate a single network , but rather a collection of networks , one for each cell state . One simple solution is to estimate the network for each state independently of the others using the graphical lasso . However , this approach can result in poor quality of the networks due to the small sample size per cell state . To overcome this challenge , our recently proposed algorithm , Treegl [24] , utilizes the following strategy . Similar to the graphical lasso [26] , [27] , Treegl estimates the neighborhood of each gene independently of those of other genes using regularized regression . However , unlike previous methods learning only a single network , Treegl simultaneously estimates neighborhoods of a gene in multiple networks each corresponding to a unique state in the phenotypic tree of the breast cells . It is unique in that Treegl makes use of a total variation regularizer based on the progression and reversal relationships between pairs of cell states in question to bias the amount of topological differences between networks underlying the related states , and to allow information regarding probabilistic independencies between genes to propagate across all states either directly or indirectly related by phenotypes . Such a strategy can lead to highly statistically confident estimation of a gene Markov network [28] , even under small sample size scenarios . In the Materials and Methods section , we will offer details of a novel statistical regularization technique that makes this possible . From the theoretical standpoint , Treegl is an instance of the general varying coefficient varying structure ( VCVS ) formalism analyzed in [29] . The VCVS model encodes changing structures of gene networks in different cell states as a function of regression coefficients in regularized regression problems . Estimating these regression coefficients , and thus the associated network structures then reduces to solving a convex optimization problem jointly over all cell states . The global optimal solution for such a problem can be found using standard convex solvers . Moreover , the VCVS formalism allows one to theoretically examine and prove the statistical conditions under which changes in structures can be correctly estimated even in the high dimensional setting when . This distinguishes our approach from other methods [17] , [30] that are highly non-convex and therefore rely on local search heuristics that only find local optima . These existing methods also do not offer sound statistical machinery for addressing difficult conditions such as nonstationarity ( e . g . , time-evolving ) and high-dimensionality under small sample size as we encountered in our study . Having a theoretical framework allows us to trade-off model expressivity and learning complexity in a principled manner . For example , if we allow a complex and arbitrary network model ( i . e . , a dense network ) , then there would be no guarantees on the quality of the recovered network structure in small-sample size scenarios . Instead , by enforcing a restricted model ( i . e . , a sparse network ) , its likelihood function is by definition convex and an optimal solution may be found . Thus , the quality of the resulting solution can be theoretically characterized , and it can be determined under which conditions the correct underlying parameters ( network structure in this case ) are discovered . Fortunately , sparsity is also biologically justifiable . For example , it is common to find a transcription factor regulating a limited number of genes under specific conditions [31] . We first evaluate Treegl's performance on simulated microarray data . In order to find out how effectively Treegl can detect change points of multiple networks while sharing information among related cell states at the same time , we design the simulated networks as illustrated in Figure S1 . In particular , for each experiment , an artificial collection of 70 networks related by a tree-shaped lineage are generated , in which a sequence of 10 identical networks is connected to a network of different topology via a change/branching point ( see details in Materials and Methods ) . Then , a small number of samples are generated from each of the networks . It is important to note that Treegl does not know a priori which of the networks are identical and which are different and thus has to discover this based on the samples . In order to evaluate how well Treegl can recover the underlying network structures for the samples in the simulation data , we compare Treegl with the static method estimating a single network and the method estimating each network independently by plotting the precision-recall curves which show the recall for different values of precision based on the network estimated by the three methods . As illustrated in Figures 3 & S2 , Treegl performs favorably to the other two methods . It should also be noted that compared to the static method which produces only one network , Treegl can produce different networks related by the tree lineage . The independent method also produces different networks but it performs poorly compared to Treegl . In order to reverse engineer gene networks of the breast cells , we first used a phenotypic tree to represent the relationships of the cells ( Figure 1 ) . Due to the small sample size of the microarray data and imbalance of the sample abundance for different cell states ( see Materials and Methods for details ) — both of the problems pose significant challenges to network reconstruction — we used what is known of the interrelatedness of signaling pathways affecting phenotypic reversion to pool data derived from various samples in order to increase the power of the network inference . In particular , since EGFR and β1-integrin are cross-modulated in the HMT3522 cells [8] , we assumed that the gene networks in the T4-2 cells reverted by inhibiting either of the molecules share reasonable similarity , and hence we grouped data from these reverted cells together to form the EGFR/ITGB1-T4R group . Likewise , we grouped together data from T4-2 cells reverted by either a PI3K inhibitor , a MAPK inhibitor , or a dominant-negative Rap1 to form the PI3K/MAPKK-T4R group , because PI3K and MAPK are also cross-modulated in the breast cells and Rap1 signals through PI3K . A tree diagram illustrating the relationships of the cells are shown in Figure 1 . Based on these relationships , we reverse engineered gene networks for the HMT3522 cells using Treegl . It is important to point out that it would be nearly statistically impossible to reconstruct a cell-state-specific gene network using existing methodology based on three microarray samples per group as in the dataset we used here . Note that we will also refer to different groups of the cells in the phenotypic tree in Figure 1 as different cell conditions or states . The reconstructed networks for non-reverted and various conditions of the reverted HMT3522 cells are illustrated in Figure S3 . They share many topological similarities as well as differences . About 60% of the network edges are common to all cell conditions represented in the phenotypic tree diagram , consistent with underlying biological similarities shared between them . In the following , we concentrate on only the edges specific to each cell state , which we call the differential network for each cell state . In order to validate our network reconstruction results , we used an external microarray dataset . Since previous evidence suggests that there is a high correlation between the degrees and essentiality of genes in yeast networks , we hypothesize that i ) if a gene is indeed involved in the networks of the breast cells , its abnormal expression would have higher impact on the survival of breast cancer patients than those genes which are not in the networks; and ii ) if a gene is a hub ( with a high degree ) in the differential networks of breast cells , its abnormal expression would have higher impact on the survival of breast cancer patients than those with low degrees . We therefore investigated the effect of the aberrant expression values of the hubs in the differential networks of the HMT3522 cells on the survival of human breast cancer patients . The external dataset we used is a gene expression microarray dataset obtained from 295 primary human breast tumors [39] , employed previously to identify gene expression signatures which may be predictive of patient clinical outcomes . The same dataset was also used previously [12] to demonstrate the impact of abnormal expression of TACE , TGFA , and AREG , which were shown to play important roles in the HMT3522 series of the breast cancer cells grown in the 3D culture , on the survival of the same cohort of the breast cancer patients . In order to define hubs , we examined the distribution of genes with varying degrees in the differential networks . Figure S7 shows that while a majority of the genes have degrees of 1–3 , much fewer genes have a degree greater than 5 , and therefore , we designated hubs to be genes with degree greater than 5 in the differential networks . Note that the same criterion was also used to define hubs in [40] . Indeed , we found that 18% of the genes in the networks of the five breast cell states affect patient survival significantly , whereas that only 6% of the genes which are not in the breast cell networks but are present in the external dataset affect patient survival significantly . Our results also showed that 22% of the hubs in the differential networks of the breast cells affect patient survival significantly . GO analysis revealed that these significant hubs are enriched with genes involved in regulation of cell migration , mobility , growth , and proliferation ( see Table S4 for a list of the hubs ) , and all of these biological activities are known to be essential for cancer cell development and progression . Similarly , 23% of the genes with degree >10 and also with degree >20 affect patient survival significantly . However , for genes with degrees equal to one to five in the differential networks , the percentage of them affecting patient survival drops to 17% . Together , these results indicate that i ) genes in the breast cell networks indeed have higher tendency of influencing patient survival significantly than those not in the networks , and also that ii ) hubs in the differential networks are more likely to affect patient survival significantly than those with low degrees , suggesting the structures of the reconstructed networks are valid . In order to identify potential novel drug targets , we examined three hubs , NEBL , HBEGF , and PAPD7 , whose extreme ( i . e . , either lower or higher ) expression values are correlated with the lowest 15-year patient survival rates ( 35% , 30% and 34% , respectively ) and also with low 10-year survival rates ( 60% , 42% and 56% , respectively ) in the examined dataset ( Figure 6 ) . Previous evidence has shown that abnormal expression of TACE , TGFA , and AREG are associated with 62% , 61% and 54% of the 10-year survival rates respectively , and associated with 57% , 50% and 54% of the 15-year survival rates respectively , in the same cohort of the patients [12] . Our results , therefore , suggest that NEBL , HBEGF , and PAPD7 , similar to TACE , TGFA , and AREG , also play important roles in breast cells . Since little is known about NEBL , HBEGF , and PAPD7 , we examined their neighbors in the corresponding differential networks ( which we call neighborhood analysis ) to shed light on their functions in breast cancer . Figure 7A shows the NEBL subnetwork in the S1 differential network . NEBL encodes a member of the nebulin family of proteins , which bind actin and are components of focal adhesion complex . Our data showed that decreased expression of NEBL is associated with 36% of 15-year survival rate for breast cancer , suggesting a protective role of this protein when overexpressed . Genes interacting with NEBL in the NEBL subnetwork are mainly involved in energy production by oxidation of organic compounds , actin and cytoskeletal protein binding , regulation of growth , and anatomical structure morphogenesis , all of which are consistent with the biological evidence suggesting involvement of nebulin in migratory cells [41] . Figure 7B shows the HBEGF subnetwork in the T4-2 differential network . HBEGF encodes a heparin-binding EGF-like growth factor , which is an EGFR ligand [42] . We found that higher expression of HBEGF is correlated with 34% of 15-year survival rate ( Figure 6B ) , and that the neighbors of HBEGF in the HBEGF subnetwork are involved in diverse biological processes and functions , such as glycolysis , apoptosis , participating in laminin-5 complex , notch signaling pathway , and angiogenesis , all of which are in line with previous findings showing the high expression level of HBEGF is positively related to the aggressiveness of the breast tumors [43] and that HBEGF plays key roles in tumorigenicity and invasiveness of ovarian cancer [44] . Finally , we examined the PAPD7 subnetwork in the MMP-T4R differential network ( Figure 7C ) . PAPD7 encodes DNA polymerase sigma . Our data indicated that overexpression of PAPD7 is associated with 30% of 15-year survival rate and 8 . 5 years of median survival time ( Figure 6C ) , both of which are the worst patient outcomes correlated with all the hubs in the differential networks , suggesting that PAPD7 plays significant roles in breast cancer cells . Previous evidence showed that a homolog of PAPD7 in Saccharomyces cerevisiae Trf4 plays a key role in RNA quality control by degrading aberrant or unwanted RNAs in the nucleus [45] . Interestingly , our functional analysis revealed that genes interacting with PAPD7 in the MMP-T4R differential network are significantly enriched with those involved in RNA degradation and metabolic process , as well as regulation of Ras and small GTPase mediated signal transduction , and phosphatidylinositol signaling system ( Figure 7C ) , which implicates that similar to Trf4 , PAPD7 also participates in crucial functions such as RNA quality control in human cells . Taken together , these findings suggest that the three hubs , NEBL , HBEGF , and PAPD7 , in the differential networks play important roles in growth and development of breast cancer cells , and may thus become potential novel therapeutic targets . More important , these results also suggest that our reconstructed networks can not only reveal genes which have high impact on patient survival in specific cell conditions , but also can provide insight into their functions by neighborhood analysis , and thus facilitate personalized drug target discovery and identification , and help make personalized breast cancer therapy possible .
The problem of estimating rewiring networks simultaneously from multiple cell states in the phenotypic tree , as solved by Treegl , is fundamentally different from either estimating a single “average” network from the samples pooled from all states and subsequently “trace-out” active subnetworks corresponding to each state [1] , or estimating multiple networks independently . The latter strategies are common practices in the system biology community , which either directly or indirectly assume the network in question is static , and samples of the nodal states in the phenotypic tree are i . i . d . across ( when pooled ) or within cell states . In reality , such an assumption is biologically invalid as well as statistically unsubstantiated . The Treegl algorithm elegantly couples all the inference problems pertained to each network in the tree of multiple conditions , and achieves a globally optimal and statistically well behaving solution based on a principled VCVS model and a convex optimization formulation . In our analysis of the HMT3522 breast cancer cell lines , we reverse engineered 5 different gene networks specific to each cell state represented in the phenotypic tree . The S1 differential network contains genes predominantly involved in normal cellular activities , while the T4-2 differential network is enriched with pathways playing active roles in cancers . Interestingly , compensatory signaling appears to be a recurring theme of the T4-2 cells phenotypically reverted by different agents . In the T4-2 cells reverted by inhibition of either EGFR or β1-integrin ( i . e . , the EGFR/ITGB1-T4R group ) , despite the absence of the ErbB pathway , signaling events downstream of EGFR or β1-integrin , mainly via the PI3K-AKT-mTOR axis , seem to be upregulated . These results are supported by clinical evidence showing that some breast cancer patients exhibit drug resistance after being treated with EGFR inhibitors . Similarly , in the PI3K/MAPKK-T4R cells , their differential network is enriched with genes closely connected to PI3K , suggesting they are upmodulated to make up for the loss of PI3K signaling , also agreeing with clinical findings showing patient resistance to PI3K inhibitors . Likewise , the compensatory effect is observed in the differential network of the T4-2 cells reverted by MMPIs , which involves genes participating in protein catabolic processes presumably to make up for the loss of the MMP function . The effect of MMPIs for treating breast cancer patients was disappointing in clinical trials , but no conclusive evidence for ineffectiveness has been put forward [38] . Our results suggest that the failure of treating breast cancer patients by MMPIs involves upmodulation of the catabolic processes in the treated patients due to compensatory effect . Together , these results suggest despite phenotypic similarities , T4-2 cells reverted by various drugs are intrinsically different from one another; similar compensatory mechanisms , however , appear to be utilized by the T4-2 cells to antagonize effects of the different drugs . In order to compare our network-based approach with traditional statistical test-based approach , we also analyzed the gene expression data using ANOVA , and identified 1432 genes significantly differentially expressed ( FDR p-value<0 . 05 ) across different cell states; then we used pairwise t-tests to further identify significant differences between cell states . We found that due to small sample size problems , these traditional approaches are too stringent to reveal interesting signals . For example , we examined the genes differentially expressed between the T4-2 cells reverted by MMP inhibitors ( MMP-T4R ) and other cell states , in particularly between MMP-T4R and S1 , as well as between MMP-T4R and T4 . Our results show that there are 473 genes significantly differentially expressed in MMP-T4R , comparing to S1 , and the only two GO functional groups significantly enriched ( FDR p-value<0 . 05 ) among these genes are “mitotic cell cycle” and “sterol biosynthesis process . ” Comparing to T4 , there are 375 genes differentially expressed in MMP-T4R , and there are no GO groups significantly enriched among these genes . Moreover , we examined genes in the differential network of MMP-T4R which are involved in some of the significantly enriched GO groups , e . g . , “proteasome complex” and “cellular catabolic process” , both of which suggest compensatory signaling in the MMP-T4R cells . We found that among 12 genes in the differential network of MMP-T4R ( “PSME3 , PSMA4 , PSMB8 , PSMD10 , PSMA3 , PSMB9 , PSME2 , PSMD7 , PSMA6 , PSMC2 , PSMA2 , PSMD6” ) which are involved in “proteasome complex” , only two of them ( PSMA3 , PSMB9 ) significantly differ between MMP-T4R and S1 as identified by ANOVA , and two ( PSMC2 , PSMB9 ) significantly differ between MMP-T4R and T4 . Likewise , among 33 genes in the differential network of MMP-T4R which are involved in “cellular catabolic process” , only 5 genes ( “PSMB9 , ANAPC5 , USP18 , IDH1 , PSMA3” ) significantly differ between MMP-T4R and S1 as identified by ANOVA , and 3 genes ( “PSMB9 , USP18 , IDH1” ) differ between MMP-T4R and T4 . Furthermore , we looked into the 22 hubs in the differential networks which significantly affect patient survival , and found that only 8 ( 36% ) of them are differentially expressed across 5 cell states as identified by ANOVA and a majority ( 64% ) of them are not differentially expressed . Similarly , among 99 hubs in the differential networks , 43% are differentially expressed , while 57% are not . These results suggest that under small-sample-size scenarios , traditional statistical tests are too stringent to capture interesting signals , while our network-based differential analysis can leverage on similarities among different samples while revealing key differences which set them apart . In order to identify potential novel drug targets , we also investigated hubs in the breast cells whose aberrant expression values are significantly associated with survival outcomes of breast cancer patients . We found that genes in the networks of the breast cells have 2 times higher tendency than those not in the networks to affect patient survival in the cohort we studied . Also , hubs in the breast networks appear more likely to influence patient survival than genes with low degrees . Indeed , the proportion of the hubs with high degrees which are significant survival genes ( 22% for hubs with degree >5 , and 23% for hubs with degree >10 and also for those with degree >20 ) is not much higher than that ( 17% ) of the genes with low degrees . The reasons for this can be explained as follows . When previous evidence suggests that in yeast networks , a gene with a higher degree is more likely to be an essential gene [46] , [47] , an essential gene is defined as “the cell is unviable when the gene is knocked off” [47] . However , it is difficult to know/determine which genes are essential in humans . Nevertheless , in light of the definition of ‘essentiality’ in yeast , we think it is plausible to believe that the actual percentage of the hubs ( with degree >5 ) in the differential networks of the breast cells , which can affect patients significantly , is 22%+x% , rather than 22% , and the reasons why we cannot see the phenotypic effect of the x% of the hubs on patient survival may include: i ) these hubs are so essential to humans that any abnormality would lead to death , even before breast tumors were formed or diagnosed; and/or ii ) there are some redundant genes which can make up for the loss/gain of functions of these essential hubs . Despite the fact that our results suggest that the genes in the breast cell networks are more likely to affect patient survival than those which are not , and also that hubs in the differential networks tend to affect patient survival more than genes with low degrees , our data show that the distributions of the patient survival rates ( 5-year , 10-year or 15-years ) associated with these different groups of genes are not significantly different , suggesting that the patient survival rates are not only affected by degrees of genes in the breast cell networks , but also affected by the functionalities of the genes . We have also characterized the three hubs in the cell-state-specific differential networks whose aberrant expression values are linked with the worst survival outcomes in the breast cancer patients: NEBL in S1 cells , HBEGF in T4-2 cells , and PAPD7 in the MMP-T4R group of the reverted cells . Our results are not only in line with existing information known about these genes , but also provide insight into their functional significance on the growth and development of breast cancer cells . These hubs are promising to serve as potential drug targets for personalized breast cancer therapy . The major challenge of this work is the small sample size of the microarray data we have used for the network inference . The data was from 15 microarrays in total , and the T4-2 cells reverted by different agents had to be pooled together in order to increase the power of the network inference . Even though the sample grouping strategy is biologically justifiable ( see details in the Results section ) , our abilities to find differences between T4-2 cells reverted by different agents are limited due to mixed samples in the EGFR/ITGB1-T4R and PI3K/MAPKK-T4R groups of the reversion cells . For example , it is difficult to dissect which specific pathways are abnormally regulated ( compared to S1 cells ) in which reversion cell state: T4-2 reverted by EGFR inhibitors or by ITGB1 inhibitors . Likewise , it is also difficult to reveal differences in the T4-2 cells reverted by different agents in the PI3K/MAPKK-T4R group . Moreover , mixed samples can reduce power to detect interesting signals in the data . Despite suggesting compensatory events in the reversion cells , the enriched pathways in the EGFR/ITGB1-T4R and the PI3K/MAPKK-T4R cells are not significant ( unadjusted p-values<0 . 05 , but FDR p-values>0 . 1 ) . However , since our data agree well with clinical evidence , they may facilitate clinicians to identify specific molecules which lead to resistance in the drug-treated breast cancer patients . In order to overcome the limitations of the mixed samples , we also focus on finding similarities of the different T4-2 reversion cells . Our results show that we were able to discover a significant amount of information that agrees with the facts and evidence previously known in the literature . Moreover , we were also able to delineate a mechanistic framework at the systems level that can facilitate further elucidation of the mechanisms underlying different states of the breast cells in the progression and reversion model . Experimental validations are nevertheless needed to further verify our findings . In summary , this work demonstrates our recently developed Treegl algorithm can not only provide a holistic view ( i . e . , the so-called “pan-cell-state” view that echoes the emerging “pan-cancer” or “pan-disease” approach nowadays to biomedical analysis ) of the progression and reversion model of the breast cells worthy of further exploration , but also allows us to gain a deeper and systems-level understanding about the behaviors of nonmalignant and malignant breast cells , which may help novel drug target discovery and make personalized breast cancer therapy possible .
HMT3522 S1 and T4-2 cells were grown in 3D lrECM as previously described [6] , [48] . The T4-2 cells were reverted using each of the following reverting agents as described previously: an EGFR inhibitor Tyrphostin AG 1478 and a human EGFR-blocking monoclonal antibody mAb225 [7] , a β1-integrin inhibitor AIIB2 [6] , a MAPK inhibitor PD98059 [7] , a PI3K inhibitor LY294002 [8] , dominant-negative Rap1 [13]; an MMP inhibitor GM6001 [49] , and a broad-range inhibitor of MMPs and ADAMs , TNF protease inhibitor–2 ( TAPI-2 ) [12] . S1 , T4-2 and reverted T4-2 cells were isolated from 3D cultures with PBS/EDTA as previously described [50] . Total cellular RNA was extracted using RNeasy Mini Kit with on column DNase digestion ( Qiagen ) . RNA was quantified by measuring optical density at A260 and quality was verified by agarose gel electrophoresis . Purified total cellular RNA was biotin labeled and hybridized to the Affymetrix GeneChip human genome HG-U133A arrays as previously described [51] . Gene expression microarray data was obtained from 15 total RNA samples prepared from the HMT3522 breast cells grown in 3D lrECM and treated with various reverting agents or vehicle controls as mentioned above . Unfortunately , T4-2 cells reverted by some agents have only one sample per each reversion cell state . Even though our method , Treegl , is designed for small sample size scenarios , having only one sample per state is not enough for network inference — as it is known that it takes at least two samples to measure even a simple quantity like correlation . Thus , in order to increase the power of the network inference , we grouped the arrays into the following five categories with each having 3 samples: ( i ) S1 cells ( 3 arrays ) ; ( ii ) T4-2 cells ( 3 arrays ) ; ( iii ) the EGFR/ITGB1-T4R group , which contains two arrays of the T4-2 cells reverted by the EGFR inhibitor Tyrphostin AG 1478 and the human EGFR-blocking monoclonal antibody mAb225 , respectively , and one array of the T4-2 cells reverted by a β1-integrin inhibitor AIIB2; vi ) the PI3K/MAPKK-T4R group , which contains one array of the T4-2 cells reverted by a MAPK inhibitor PD98059 , one array of the T4-2 cells reverted by a PI3K inhibitor LY294002 , and one array of the T4-2 cells reverted by dominant-negative Rap1; and ( v ) the MMP-T4R group , which contains two arrays of the T4-2 cells reverted by an MMP inhibitor GM6001 , and one array of the T4-2 cells reverted by a broad-range inhibitor of MMPs and ADAMs , TAPI-2 . The biological justification on this grouping strategy is provided in the Results section . In order to identify networks specific to each state of the breast cells , we utilized a phenotypic tree model to represent the relationships of different states of the breast cells ( Figure 1 ) . In particular , since the HMT3522 series were originated from S1 cells , we positioned S1 cells as the root of the phenotypic tree . Then we made S1 cells the parent of T4-2 cells , since T4-2 cells were derived from S1 cells . Finally , we made T4-2 cells the parent of the three conditions of the T4-2 cells reverted by various agents ( the EGFR/ITGB1-T4R group , the PI3K/MAPKK-T4R group , and the MMP-T4R group ) . Raw gene expression data was preprocessed using the following procedure . The data from the perfect match ( PM ) probes on the Affymetrix arrays was first log2-transformed , and normalized using the CyclicLoess normalization method to minimize unwanted noise in the data [52] . We did not use the difference between the values from the PM probes and those from the mismatch ( MM ) probes ( i . e . , PM – MM ) to represent values of the probes for each gene , because it has been shown that the MM values can pick up both non-specific and specific signal of the probes , and thus PM-MM values may attenuate real signal values from the PM probes [53] . The normalized PM values were then summarized into gene expression values using the median polish technique [54] . For some transcripts , multiple probes on an array target the same transcript; the values of the probes were combined by taking the median of the values to represent the expression level of the corresponding transcript . There are 12 , 977 unique genes on the arrays . The complete microarray dataset is available at the Gene Expression Omnibus ( GEO ) database ( http://www . ncbi . nlm . nih . gov/geo - GSE42125 ) . To reduce biological noise in the data , we removed genes whose expression values showed low variability across different groups of the breast cells . In particular , for each gene , we calculated its median expression values in five different groups of the breast cell states . If the fold change value of a gene between any of the two groups was larger than 1 . 3 , we included the gene for the downstream analysis . The reasons why we used the fold change of 1 . 3 as the threshold value to filter genes are as follows: i ) based on our previous experience with human lung disease studies [55] , we found that a fold change of 1 . 2–1 . 3 is enough to elicit significant biological changes in humans; and ii ) When using the threshold value of 1 . 3 , 5 , 440 genes passed the filter , which we consider is a reasonable number for the downstream network analysis by Treegl . Then we applied Treegl to reconstruct gene networks in the five breast cell states using expression values of the qualified genes . We now give a mathematical formulation of representing the gene networks in order to introduce our algorithm . Consider the problem of modeling different gene networks , each corresponding to a unique cell state . Each cell state has i . i . d . microarray replicates . All the arrays in the dataset have the same set of genes . In our case , we have 5 different conditions of the breast cells in the phenotypic tree: S1 , T4 , EGFR/ITGB1-T4R , PI3K/MAPKK-T4R , and MMP-T4R . As commonly done , we model each gene network as a weighted undirected graph , where the vertices represent genes and the edges represent interactions in the network . Let represent a network in cell state , where denotes the set of genes that is fixed for all cell states and denotes the set of edges specific to the network for cell state . Let where be the vector of expression values of genes on array in cell state . We assume , i . e . that the vector of expression values follows a multivariate Gaussian distribution . We are interested in reconstructing a set of networks that are related by the phenotypic tree as shown in Figure 1 . For each cell state , let be the parent of the cell state in the tree; alternatively , we can also view as a descendant of . In our case , , , . We generally let correspond to S1 , correspond to T4 , and correspond to the EGFR/ITGB1-T4R group , the PI3K/MAPKK-T4R group , and the MMP-T4R group , respectively . Thus , in our formulation , recovering the structures of the gene regulatory networks in different breast cell states corresponds to estimating the network structure for each cell state . Consider first estimating the edge set of a single network from the data . As described in the Results section , we model the gene network for each cell state as a Gaussian Markov network . Therefore the inverse of the covariance matrix , called the precision matrix , , completely encodes the structure of the Markov network . An edge exists in the Markov network if and only if the corresponding precision matrix element is non-zero . A Gaussian Markov network , encoded via the precision matrix , allows us to model more sophisticated dependencies than a correlation network , which is encoded by the covariance matrix . In particular , the precision matrix elements are related to the partial correlation between and ( denoted as , see below for details ) . Formally , partial correlation between a pair of random variables given a set of controlling variables is defined as follows . Let and denote the residuals from performing linear regression of with and with , respectively . The partial correlation is then defined as the correlation between and . Unlike correlation , which simply measures the association between a pair of random variables , partial correlation intuitively measures the association between a pair of variables with a set of controlling variables removed ( where here is all the other genes ) . The partial correlation , due to its close relationship with the elements of the precision matrix , makes the latter much more suitable than the covariance matrix for distinguishing between indirect and direct relationships as shown in Figure 2 . Since our goal is to learn the structure of the Markov network , we are only concerned with estimating which precision matrix elements are zero and which are not ( rather than the exact precision matrix values ) . Therefore it suffices to estimate the partial correlation coefficients , which are proportional to the precision matrix elements by the equation . An estimation algorithm can be constructed by exploiting the relationships between the partial correlation coefficients and a linear regression model [56] . Specifically , consider a linear regression model where gene is treated as the response variable and all the other genes are covariates . The regression coefficient of covariate is then proportional to the partial correlation . The above facts enable us to use regression-based methods to estimate the elements of the precision matrix ( up to a proportionality constant ) , and thus the underlying network structure . In particular , our method is based on an efficient neighborhood selection algorithm [26] based on -norm regularized regression that works well in practice and has strong theoretical guarantees . In this approach , the neighborhood of each gene ( the set of edges incident to ) is estimated independently of the neighborhoods of other genes . After estimating each neighborhood , the results are then combined to produce the estimated network . In every neighborhood estimation step , gene is treated as a response variable , and all the other genes are the covariates . An penalized linear regression ( also known as the lasso [57] ) is performed to give an estimate of the regression coefficients . Then by leveraging the relationship between the regression coefficients and the partial correlation , the estimated gene network is constructed by adding an edge to if either or is non-zero ( max-symmetrization ) . Obviously , networks for each cell state can be estimated independently by using the method described above . However , this can lead to very poor estimates of the edge sets , because in common laboratory settings only a few replicates of gene expression data can be obtained . To overcome this limitation , we estimated the networks by assuming that the networks share similarities due to their relationships as suggested by the phenotypic tree , but also have some sharp differences . For example , for S1 and T4-2 cells , we assume they have considerable differences as the former is nonmalignant while the latter is tumorigenic; however , since T4-2 were derived from S1 , we also assume that these cells share substantial similarity . This is the motivation behind Treegl , the algorithm that we first presented in [24] . Treegl is unique in that it makes use of a total variation regularizer , which allows information to be shared across different cell states , and thus encourages the resulting networks to be similar while allowing differences in the networks to be revealed . More specifically , Treegl adopts the idea of neighborhood selection and additionally penalizes the differences between the neighborhoods of adjacent states in the breast cell phenotypic tree . This makes Treegl more effective in small-sample-size settings than existing approaches since it can estimate a collection of networks more robustly by leveraging the similarities among them . In summary , Treegl proposes the following optimization problem for jointly recovering the neighborhoods of genes for all the cell states in the phenotypic tree of the breast cells: In the equation above , the first term corresponds to the residual sum of squares as in normal linear regression . indicates the vector of the expression values of all genes except , and similarly , . is defined as . The second term ( corresponding to ) is a penalty on the edge weights ( similar to [26] ) , where denotes the norm of vector , which is the sum of the absolute values of the components of . This penalty promotes sparsity in the edge weights by enforcing most of the edge weights to be zero . The assumption of sparsity is biologically justifiable . For example , it is common to find a transcription factor regulating a limited number of genes under specific conditions [31] . The details of the regularization can be seen in [57] . The third term ( also called the total variation penalty ) associated with enforces sparsity of differences between S1 and T4-2 as well as between T4-2 and each of the T4R groups ( as illustrated in the tree structure in Figure 1 ) , but not between T4Rs and S1 . This encourages many ( but not all ) of the elements of to be identical to those of . The fourth term ( also associated with ) additionally penalizes the differences between each of the T4R groups and S1 , while allowing for sharp differences to be revealed between the two groups . Note that if the fourth term was not used , the T4R networks would be biased to be more similar to the T4-2 network than the S1 network . This would be undesirable , since it is unknown a priori whether each of the T4R states are more similar to T4-2 or S1 cells . and are regularization parameters that control the amount of penalization ( see below for details on how we selected these parameters ) . Because the minimization problem is convex , we solved it using the CVX solver [58] , as we described in [24] . In this work , we focus on genes linked by positive edges , because interaction of these genes is easier to interpret . For example , suppose genes X and Y are linked by positive edges , and genes Y and Z are also linked by positive edges . Intuitively , this suggests that genes X and Y are regulated in the same direction , that is , when gene X is up ( or down ) -regulated , gene Y is also up ( or down ) -regulated . Same is true for genes Y and Z which are also regulated in the same direction . As a result , we can also decide that genes X and Z are regulated in the same direction . On the other hand , interpreting interaction of genes linked by negative edges is more complicated . For example , suppose genes A and B are linked by a negative edge and genes B and C also linked by a negative edge . This intuitively means that A and B are regulated in the opposite direction , and that B and C are also regulated in the opposite direction; it is , however , unclear what is the relationship of A and C , which may be regulated in either the same or the opposite direction . Due to the reasons stated above , we chose to limit the scope of this work by focusing only on the positive edges to simplify our interpretation of the results . Choosing the regularization parameters and is a challenging problem in high dimensional statistics . Kolar and Xing proposed to use the Bayesian information criterion ( BIC ) score to select these parameters [59] . This approach can be useful in low dimensional settings; however , it does not perform well in high dimensional settings [60] . In this work , since we have a good knowledge of the biological properties of the S1 and T4-2 cells in the HTM3522 system , we employed a knowledge-based approach to tune and , namely , we tuned these parameters based on our prior knowledge about S1 and T4-2 cells , which turned out to be highly effective in the high dimensional , small sample size setting as we encountered in this work . Specifically , we first varied and in the set {4 , 4 . 5 , 5 , 5 . 5 , 6 , 6 . 5 , 7} and the set {0 . 5 , 1 , 1 . 5 , 2 , 2 . 5} , respectively , and generated cell-state-specific networks for each possible pair of and . These sets of and were chosen because the networks can be generated with reasonable sparsity . Then we examined the biological pathways significantly enriched in the differential network of the S1 and T4-2 cells , and found that when and , almost all of the enriched pathways in the T4-2 network make the best biological sense in that they are either well described in previous studies or are known pathways active in cancers . Since we used S1 and T4-2 cells to help tune the regularization parameters , we present and discuss mainly biological findings we made from the networks of the T4-2 reversion cells to avoid circular reasoning . We describe below how the networks in our simulation experiments were generated . Consider the following artificial collection of 70 networks , related by a tree: Treegl does not know a priori which networks are identical and which are not . The number ( ) of samples are then generated for each network under the Gaussian Graphical Model assumption . We vary the values of in the simulation experiments , and the results presented in Figures 3 are based on the values indicated in the figure . In each scenario , the number of edges is twice as much as the number of nodes . To evaluate Treegl , we conduct a total of 10 simulation experiments , and plot the precision-recall curves showing the recall for different values of precision based on the networks reconstructed by Treegl . The error bars in the curves indicate the first and third quartiles of the results . Details on how we generated the precision-recall curves and selected the regularization parameters can be found in [24] . To identify pathways significantly enriched in the gene networks of the 5 breast cell states estimated by Treegl , we performed pathway analysis on the list of the genes involved in each network using the Category Bioconductor package with minor modification ( http://www . bioconductor . org ) . The Category package uses hypergeometric tests to assess overrepresentation of the KEGG pathways among genes of interest . A list of 12 , 977 unique genes on the Affymetrix GeneChip Human Genome U133A was used as the reference gene list for the pathway analysis . A pathway is considered to be significant if p<0 . 1 with the FDR controlling procedure of Benjamini & Hochberg [61] . To find out genes significantly associated with certain diseases in the differential networks of the breast cell states , we performed pathway analysis as described above . For each differential network , pathways related to diseases and significantly enriched in the network were singled out; genes in the network that are involved in the enriched disease-related pathways were reported as the genes significantly associated with the diseases in the network . To identify functional groups of genes significantly enriched in the gene networks of the breast cells estimated by Treegl , we performed GO analysis on the list of the genes involved in each network using the GOstat program [62] . The GOstat program finds the enriched functional groups using Fisher's exact tests . The GOstat program was also used to identify functional groups of genes enriched among the neighborhoods ( or the subnetworks ) of the hubs significantly affecting patient survival . A functional group is considered to be significant if p<0 . 05 with the FDR controlling procedure of Benjamini & Hochberg . A list of 12 , 977 unique genes on the Affymetrix GeneChip Human Genome U133A was used as the reference gene list for the GOstat program . We define hubs as genes with positive degree greater than 5 in the differential networks of the breast cell states . Survival analysis was performed using microarray expression values of the hubs extracted from a gene expression microarray data set obtained from 295 primary human breast tumors [39] . For each hub , its expression values across all patients were divided into three groups: lower quartile , interquartile , and upper quartile groups . Kaplan–Meier curves were used to estimate the association of expression values of the hubs in the three groups with patient survival . The log-rank test was used to calculate p-values of the survival curves . A hub was considered as significant if the p value of its associated survival curve <0 . 05 after controlling for multiple testing using the Bonferroni procedure . | The HMT3522 isogenic human breast cancer progression series has been used to study the effect of various drugs on the reversion of the breast cancer cells . Despite significant efforts to delineate key signaling events responsible for phenotypic reversion of the malignant HMT3522-T4-2 ( T4-2 ) breast cells in this series , many questions remain . For example , what is involved in the phenotypic reversion of T4-2 cells at the systems level ? In order to answer this question , we analyzed gene expression microarray data obtained from these cells using our recently developed tree-evolving network inference algorithm Treegl . We reconstructed cell-state-specific gene networks using Treegl . Our functional analysis results show that we can not only unravel cell-state specific information characteristic of non-malignant HMT3522-S1 ( S1 ) and malignant T4-2 cells in the series , but can also provide insight into the T4-2 cells reverted by various agents . We found that the networks specific to various conditions of the T4-2 reverted cells are all suggestive of compensatory signaling effects , which , however , are mediated by different signaling pathways to antagonize different drug effects in the reverted cells . Our results demonstrate that the HMT3522 system when analyzed with Treegl may potentially become an effective tool for novel drug-target discovery and identification . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computer",
"applications",
"algorithms",
"systems",
"biology",
"computer",
"and",
"information",
"sciences",
"computer",
"modeling",
"mathematics",
"cell",
"biology",
"applied",
"mathematics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",... | 2014 | Network Analysis of Breast Cancer Progression and Reversal Using a Tree-Evolving Network Algorithm |
Genetic association studies , in particular the genome-wide association study ( GWAS ) design , have provided a wealth of novel insights into the aetiology of a wide range of human diseases and traits , in particular cardiovascular diseases and lipid biomarkers . The next challenge consists of understanding the molecular basis of these associations . The integration of multiple association datasets , including gene expression datasets , can contribute to this goal . We have developed a novel statistical methodology to assess whether two association signals are consistent with a shared causal variant . An application is the integration of disease scans with expression quantitative trait locus ( eQTL ) studies , but any pair of GWAS datasets can be integrated in this framework . We demonstrate the value of the approach by re-analysing a gene expression dataset in 966 liver samples with a published meta-analysis of lipid traits including >100 , 000 individuals of European ancestry . Combining all lipid biomarkers , our re-analysis supported 26 out of 38 reported colocalisation results with eQTLs and identified 14 new colocalisation results , hence highlighting the value of a formal statistical test . In three cases of reported eQTL-lipid pairs ( SYPL2 , IFT172 , TBKBP1 ) for which our analysis suggests that the eQTL pattern is not consistent with the lipid association , we identify alternative colocalisation results with SORT1 , GCKR , and KPNB1 , indicating that these genes are more likely to be causal in these genomic intervals . A key feature of the method is the ability to derive the output statistics from single SNP summary statistics , hence making it possible to perform systematic meta-analysis type comparisons across multiple GWAS datasets ( implemented online at http://coloc . cs . ucl . ac . uk/coloc/ ) . Our methodology provides information about candidate causal genes in associated intervals and has direct implications for the understanding of complex diseases as well as the design of drugs to target disease pathways .
In the last decade , hundreds of genomic loci affecting complex diseases and disease relevant intermediate phenotypes have been found and robustly replicated using genome-wide association studies ( GWAS , [1] ) . At the same time , gene expression measurements derived from microarray [2] or RNA sequencing [3] studies have been used extensively as an outcome trait for the GWAS design . Such studies are usually referred to as expression quantitative trait locus ( eQTL ) analysis . While GWAS datasets have provided a steady flow of positive and replicable results , the interpretation of these findings , and in particular the identification of underlying molecular mechanisms , has proven to be challenging . Integrating molecular level data and other disease relevant intermediate phenotypes with GWAS results is the natural step forward in order to understand the biological relevance of these results . This strategy has been explored before and allowed the identification of the genes and regulatory variations that are important for several diseases ( reviewed in [4] ) . In this context , a natural question to ask is whether two independent association signals at the same locus , typically generated by two GWAS studies , are consistent with a shared causal variant . If the answer is positive , we refer to this situation as colocalised traits , and the probability that both traits share a causal mechanism is greatly increased . A typical example involves an eQTL study and a disease association result , which points to the causal gene and the tissue in which the effect is mediated [5]–[7] . In fact , looking for overlaps between complex trait-associated variants and eQTL variants has been successfully used as evidence of a common causal molecular mechanism ( e . g . , [5] , [8] ) . The same questions can also be considered between pairs of eQTLs [9] , [10] , or pairs of diseases [11] . However , identifying the traits that share a common association signal is not a trivial statistical task . Visual comparison of overlaps of association signals with an expression dataset is a step in this direction ( using for example Sanger tool Genevar http://www . sanger . ac . uk/resources/software/genevar/ ) , but the abundance of eQTLs in the human genome and across different tissues makes an accidental overlap between these signals very likely [2] . Therefore visual comparison is not enough to make inferences about causality and formal statistical tests must be used to address this question . Nica et al . [5] proposed a methodology to rank the SNPs with an influence on two traits based on the residual association conditional on the most associated SNP . By comparing the GWAS SNP score with all other SNPs in the associated region , this method accounts for the local LD structure . However , this is not a formal test of a null hypothesis for , or against , colocalisation at the locus of interest . A formal test of colocalisation has been developed in a regression framework . This is based on testing a null hypothesis of proportionality of regression coefficients for two traits across any set of SNPs , an assumption which should hold whenever they share causal variant ( s ) [12] , [13] . No assumption is made about the number of causal variants , although the method does assume that in the case of multiple causal variants , all are shared . Both the ranking method and proportionality testing share the drawback of having to specify a subset of SNPs to base the test on , and Wallace [14] shows that this step can generate significant biases . The main sources of bias are overestimation of effect sizes at selected SNPs ( termed “Winner's curse” ) , and the fact that , owing to random fluctuations , the causal variant may not always be the most strongly associated one . These factors lead to rejection of colocalisation in situations where the causal SNP is in fact shared . Although this can be overcome in the case of proportionality testing by averaging over the uncertainty associated with the best SNP models [14] , perhaps the greatest limitation is the requirement for individual level genotype data , which are rarely available for large scale eQTL datasets . The success of GWAS meta-analyses has shown that there is considerable benefit in being able to derive association tests on the basis of summary statistics . With these advantages in mind , He et al . [7] developed a statistical test to match the pattern of gene expression with a GWAS dataset . This approach , coded in the software Sherlock , can accommodate p-values as input . However , their hypothesis of interest differs from the question of colocalisation , with the focus of the method being on genome-wide convergence of signals , assuming an abundance of trans eQTLs . In particular , SNPs that are not associated with gene expression do not contribute to the test statistic . Such variants can provide strong evidence against colocalisation if they are strongly associated with the GWAS outcome . These limitations motivate the development of novel methodologies to test for colocalisation between pairs of traits . Here , we derive a novel Bayesian statistical test for colocalisation that addresses many of the shortcomings of existing tools . Our analysis focuses on a single genomic region at a time , with a major focus on interpreting the pattern of LD at that locus . Our underlying model is closely related to the approach developed by Flutre et al . [10] , which considers the different but related problem of maximising the power to discover eQTLs in expression datasets of multiple tissues . A key feature of our approach is that it only requires single SNP p-values and their minor allele frequencies ( MAFs ) , or estimated allelic effect and standard error , combined with closed form analytical results that enable quick comparisons , even at the genome-wide scale . Our Bayesian procedure provides intuitive posterior probabilities that can be easily interpreted . A main application of our method is the systematic comparison between a new GWAS dataset and a large catalogue of association studies in order to identify novel shared mechanisms . We demonstrate the value of the method by re-analysing a large scale meta-analysis of blood lipids [15] in combination with a gene expression study in 966 liver samples [16] .
We consider a situation where two traits have been measured in two distinct datasets of unrelated individuals . We assume that samples are drawn from the same ethnic group , i . e . allele frequencies and pattern of linkage disequilibrium ( LD ) are identical in both populations . For each of the two samples , we consider for each variant a linear trend model between the outcome phenotypes Y and the genotypes X ( or a log-odds generalised linear model if one of the two outcome phenotypes Y is binary ) :We are interested in a situation where single variant association p-values and MAFs , or estimated regression coefficients and their estimated precisions , are available for both datasets at Q variants , typically SNPs but also indels . We make two additional assumptions and discuss later in this paper how these can be relaxed . Firstly , that the causal variant is included in the set of Q variants , either directly typed or well imputed [17]–[19] . Secondly , that at most one association is present for each trait in the genomic region of interest . We are interested in exploring whether the data support a shared causal variant for both traits . While the method is fully applicable to a case-control outcome , we consider two quantitative traits in this initial description . SNP causality in a region of Q variants can be summarised for each trait using a vector of length Q of ( 0 , 1 ) values , where 1 means that the variant is causally associated with the trait of interest and at most one entry is non-zero . A schematic illustration of this framework is provided in Figure 1 in a region that contains 8 SNPs . Each possible pair of vectors ( for traits 1 and 2 , which we refer to as “configuration” ) can be assigned to one of five hypotheses: In this framework , the colocalisation problem can be re-formulated as assessing the support for all configurations ( i . e . pairs of binary vectors ) in hypothesis . Our method is Bayesian in the sense that it integrates over all possible configurations . This process requires the definition of prior probabilities , which are defined at the SNP level ( Methods ) . A probability of the data can be computed for each configuration , and these probabilities can be summed over all configurations and combined with the prior to assess the support for each hypotheses . The result of this procedure is five posterior probabilities ( PP0 , PP1 , PP2 , PP3 and PP4 ) . A large posterior probability for hypothesis 3 , PP3 , indicates support for two independent causal SNPs associated with each trait . In contrast , if PP4 is large , the data support a single variant affecting both traits . An illustration of the method is shown in Figure 2 for negative ( Figure 2A–B , FRK gene and LDL , PP3 >90% ) and positive ( Figure 2C–D , SDC1 gene and total cholesterol , PP4 >80% ) colocalisation results . While the method uses Approximate Bayes Factor computations ( ABF , [20] , and Methods ) , no iterative computation scheme ( such as Markov Chain Monte Carlo ) is required . Therefore , computations are quick and do not require any specific computing infrastructure . Precisely , the computation time behaves as , where Q is the number of variants in the genomic region and d the number distinct associations ( typically d = 2 , assuming two traits and at most one causal variant per trait ) . Importantly , the use of ABF enable the computation of posterior probabilities from single variant association p-values and MAFs , although the estimated single SNP regression coefficients and their variances or standard errors are preferred for imputed data . Given the well-understood requirements for large sample size for GWAS data , we used simulations to investigate the power of our approach . We generated pairs of eQTL/biomarker datasets assuming a shared causal variant . We varied two parameters: the sample size of the biomarker dataset and the proportion of the biomarker variance explained by the shared genetic variant . We set the proportion of the eQTL variance explained by the shared variant to 10% and we used the original sample size of the liver eQTL dataset described herein [16] . Text S1 contains a description of the simulation procedure . Results are shown in Figure 3 . We find that given a sample size of 2 , 000 individuals for the biomarker dataset , the causal variant needs to explain close to 2% of the variance of the biomarker to provide reliable evidence in favour of a colocalised signal ( lower percentile for PP4 >80% ) . Until recently the assumption that , for a given GWAS signal , the causal variant in that interval had been genotyped was unrealistic . However , the application of imputation techniques [17]–[19] can provide genotype information about the majority of common genetic variants . Therefore , in situations where a common variant drives the GWAS signal , it is now plausible that , in imputed datasets , genotype information about this variant is available . Nevertheless , limited imputation quality can invalidate this hypothesis . This prompted us to investigate the implication of not including the causal variant in the genotype panel . To address this question , we used Illumina MetaboChip data and imputed the genotyped regions using the Minimac software ( [19] and Methods ) . We then selected only the subset of variants present in the Illumina 660K genotyping array . We simulated data under the assumption of a shared causal variant , with 4 , 000 individuals in the biomarker dataset . We then computed the PP4 statistic with and without restricting the SNP set to the Illumina 660K Chip SNPs ( Figure 4 ) . We also considered two different scenarios , with the causal SNP included/not included in the Illumina 660W panel ( Figures S1 and S2 for more exhaustive simulations ) . Our results show that when the causal variant is directly genotyped by the low density array , the use of imputed data is not essential ( Figure 4A ) . However , in cases where the causal variant is not typed or imputed in the low density panel , the variance of PP4 is much higher ( Figure 4B ) . In this situation , the resulting PP4 statistic tends to decrease even though considerable variability is observed . Inspection of simulation results in Figure 5 ( bottom row for tagging SNP , leftmost graph for shared causal variant ) shows that while PP4 tends to be lower than for its counterpart with complete genotype data ( top row , leftmost graph ) , PP3 remains low . This indicates that more probability is given to PP0 , PP1 and PP2 , which can be interpreted as a loss of power rather than misleading inference in favour of distinct variants for both traits . Statistical power may also be affected by the mode of inheritance of the causal variant . To address this , we simulated cases under a recessive pattern of inheritance . Our results show that if the true model is recessive , but the eQTL signal is nonetheless analysed using the trend test , then we will often also successfully detect a colocalised signal ( Figure S9 ) . We compared the behaviour of our proposed test with that of proportional colocalisation testing [12] , [14] in the specific case of a biomarker dataset with 10 , 000 samples ( Figure 5 , and also Figures S3 and S4 ) . Broadly , in the case of either a single common causal variant or two distinct causal variants , our proposed method could infer the simulated hypotheses correctly ( PP4 or PP3 >0 . 9 ) with good confidence , and PP3 >0 . 9 slightly more often than the proportional testing p-value <0 . 05 . A key advantage in our Bayesian approach is the ability to distinguish evidence for colocalisation ( i . e . high PP4 ) from a lack of power ( i . e . high PP0 , PP1 or PP2 ) . In both of these cases ( high PP4 or high PP0/PP1/PP2 ) , the use of the proportional approach leads to failure to reject the null even though the interpretation of these situations should differ . It has been proposed that gene expression may be subject to both global regulatory variation which acts across multiple tissues and secondary tissue specific regulators [21] . Neither approach covers this case explicitly in its construction , but it is instructive to examine their expected behaviour . The proportional approach tends to reject a null of colocalisation , suggesting that a single distinct causal variant can be sufficient to violate the null hypothesis of proportional regression coefficients . In contrast , the Bayesian approach tends to favour the shared variant in the cases covered by our simulations ( median PP4 > median PP3 ) , and either hypotheses H3 or H4 can potentially have strong support ( PP4 >0 . 9 in close to 50% of simulations , and PP3 >0 . 9 in around 25% of simulations ) . Of course , the ultimate goal should be to extend these tests to cover multiple causal variants , but in the meantime , it can be useful to know that a high PP4 in our proposed Bayesian analysis indicates strong support for “at least one causal variant” and that rejection of the null of proportionality of regression coefficients indicates that the two traits do not share all causal variants , not that they cannot share one . We have so far assumed that each trait is associated with at most one causal variant per locus . However , it is not unusual to observe two or more independent associations at a locus for a trait of interest [22] . In the presence of multiple independent associations , the assumption of a single variant per trait prompts the algorithm to consider only the strongest of these distinct association signals . Hence , the presence of additional associations that explain a smaller fraction of the variance of the trait , for example additional and independently associated rare variants , have a negligible impact on our computations . To illustrate this situation , we simulated datasets with two causal variants: one colocalised eQTL/biomarker signal plus a secondary independent “eQTL only” signal ( Figure S8 ) . These simulations confirm that the PP4 statistic is only affected in the presence of two independent associations that explain a similar proportion of the variance of the trait ( Figure S8 ) . The natural and statistically exact modification of our approach would compute , for each trait , Bayes factors for sets of SNPs rather than single SNPs ( up to N SNPs jointly to accommodate for N distinct associations per trait ) . However , this approach has two drawbacks . Firstly , the interpretation of the resulting posterior probabilities is more challenging in situations where some but not all of the variants are shared across both traits . More importantly , the typical approach consists of publishing single variant summary statistics , which would prevent the use of standard summary statistics , a key feature of our approach . Owing to the focus of our algorithm on the strongest association signal , an alternative approach to deal with multiple associations consists of using a stepwise regression strategy , which would then reveal the secondary association signals . Our colocalisation test can then be run on using the conditional p-values . We find this approach to be the most practical and illustrate below an application for a locus that contains several independent eQTL associations ( Figure 6 ) . In situations where only single SNP summary statistics are available , the approximate conditional meta-analysis framework proposed by Visscher et al . [23] can be used to obtain conditional p-values . Teslovich et al . [15] reported common variants associated with plasma concentrations of low-density lipoprotein cholesterol ( LDL ) , high-density lipoprotein cholesterol ( HDL ) and triglyceride ( TG ) levels in more than 100 , 000 individuals of European ancestry . They then reported the correlations between the lead SNPs at the loci they found and the expression levels of transcripts in liver . For the lipid dataset we have access only to summary statistics . The liver expression dataset used in this analysis is the same as the one used in [15] . In Teslovich et al . , regions are defined within 500 kilobases of the lead SNPs , and the threshold for significance is . At this threshold , they found 38 SNP-to-gene eQTLs in liver ( Supplementary Table 8 of [15] ) . Table S1 shows our results for these 38 previously reported colocalisations . A complete list of all our identified colocalisations ( independently of previous reports ) is provided in Tables S2 , S3 , S4 , S5 ( broken down by lipid traits ) . Using the coloc web server for this analysis with a PP4 >75 , it took 1 minute to complete chromosome 1 and approximately 7 minutes to analyse the entire imputed genome-wide data on a laptop . The majority of our results are consistent with the findings of Teslovich et al . , with 26 out of 38 loci having PP4 . To assess the role of the prior , we varied the critical parameter , which codes for the prior probability that a variant is associated with both traits . Here we report the results using the . The complete list of results is provided in Table S1 . Table 1 lists the previously reported lipid-eQTL for which we find strong support against the colocalisation hypothesis ( PP3 >75% ) . The LocusZoom association plots for each of these loci can be found in Figure S5 . In addition to the loci listed in Table 1 , we found strong evidence of distinct signals between HLA-DQ/HLA-DR and TC ( Table S1 ) but these results must be interpreted with caution owing to the extensive polymorphism in the major histocompatibility complex region . For only one locus ( CEP250 ) , we did not find a significant eQTL signal , pointing to potential differences in bioinformatics processing and/or imputation strategy . In such a situation , both PP3 and PP4 are low and PP0 , PP1 and PP2 concentrate most of the posterior distribution . Three loci ( TMEM50A , ANGPTL3 , PERLD1/PGAP3 ) do not have enough evidence to strongly support either colocalisation or absence of colocalisation ( Table S1 ) and these should remain marked as doubtful . One of these genes , ANGPTL3 is noteworthy . Examining this locus ( Figure S6 ) , it is clear that the pattern of association p-values is consistent between LDL and ANGPTL3 expression . However , the extent of LD is strong , with 98 strongly associated variants . In such a situation , there is uncertainty as to whether the data support a shared causal variant for both traits , or two distincts variants for eQTL/LDL . Because the data are consistent with both scenarios , the choice of prior becomes determinant . Accordingly , PP4 drops from 91% to 49% if one uses instead of . Table 2 lists the 14 colocalised loci ( 15 genes ) that were not reported by Teslovich et al . ( or in Global Lipids Genetics Consortium [24] for the gene NYNRIN ) , but for which our method finds strong support for colocalisation ( PP4 >75% ) . Figure S7 shows the LocusZoom plots for these colocalisation results . Eleven of these 15 genes are strong candidates for involvement in lipid metabolism and/or have been previously suggested as candidate genes: SDC1 , TGOLN2 , INHBB , UBXN2B , VLDLR , VIM , CYP26A1 , OGFOD1 , HP , HPR , PPARA . See Text S2 for a brief overview of the function of these genes . Four others genes have a less obvious link: CMTM6 , C6orf106 , CUX2 , ENSG00000259359 . Three previously reported genes ( SYPL2 , IFT172 , TBKBP1 ) which , based on our re-analysis , do not colocalise with the lipid traits , have a nearby gene with a high probability of colocalisation ( respectively , SORT1 , GCKR , KPNB1 ) . This suggests that these genes are more likely candidates in this region . To explore the possibility that secondary signals may colocalise , we applied the stepwise regression strategy described above to deal with several independent associations at a single locus . We performed colocalisation test using eQTL results conditional on the top eQTL associated variant . Two of the loci ( SYPL2/LDL or TC , APOC4 and TG ) showed evidence of colocalisation with expression after conditional analysis ( Table 1 ) . An example of this stepwise procedure for the gene SYPL2 and LDL is provided in Figure 6 . We find that the top liver eQTL signal is clearly discordant with LDL association ( Table 1 and Figure 6 ) . However , conditioning on the top eQTL signal reveals a second independent association for SYPL2 expression in liver . This secondary SYPL2 eQTL colocalises with the LDL association ( PP4 >90% , Figure 6 ) . We developed a web site designed for integration of GWAS results using only p-values and the sample size of the datasets ( http://coloc . cs . ucl . ac . uk/coloc/ ) . The website was developed using RWUI [25] . Results include a list of potentially causal genes with the associated PP4 with their respective plots and ABF , and can be viewed either interactively or returned by email . Researchers can request a genome-wide scan of results from a genetic association analysis , and obtain a list of genes with a high probability of mediating the GWAS signals in a particular tissue . The tool also allows visualisation of the signals within a genetic region of interest . The database and browser currently include the possibility of investigating colocalisation with liver [15] and brain [26] , [27] expression data , however the resource will soon be extended to include expression in different tissues . This method , as well as alternative approaches for colocalisation testing [12] , [14] , are also available with additional input options in an R package , coloc , from the Comprehensive R Archive Network ( http://cran . r-project . org/web/packages/coloc ) .
We have developed a novel Bayesian statistical procedure to assess whether two association signals are colocalised . Our method is best suited for associations detected by GWAS , which are likely to reflect common , imputable , variations with small effects , or a rare variants with large effect sizes . Our aim differs from a typical fine-mapping exercise in the sense that we are not interested in knowing which variant is likely to be causal but only whether a shared causal variant is plausible . The strength of this approach lies in its speed and analytical forms , combined with the fact that it can use single variant p-values when only these are available . Our results show that to provide an accurate answer to the colocalisation problem , high-density genotyping and/or accurate use of imputation techniques are key . The quality of the imputation is another important parameter . Indeed , while the variance of the regression coefficient can be estimated solely on the basis of the minor allele frequency for typed SNPs and sample size ( and the case control ratio in the case of a binary outcome ) [17] , [28] , this ignores the uncertainty due to imputation . Filtering out poorly imputed SNPs partially addresses this problem , with the drawback that it may exclude the causal variant ( s ) . Hence , providing estimates of the variance of the MLE , together with the effect estimates , will result in greater accuracy . This additional option is available on the coloc package in R ( http://cran . r-project . org/web/packages/coloc ) . We currently assume that each genetic variant is equally likely a priori to affect gene expression or trait . A straightforward addition to our methodology would consider location specific priors for each variant , which would depend for example on the distance to the gene of interest , or the presence of functional elements in this chromosome region [29] . Our computation of the BF also assumes that , under , the effect sizes of the shared variant on both traits are independent . This could be modified if , for example , one compares eQTLs across different tissue types , or the same trait in two different studies . [30] has proposed a framework to deal with correlated effect sizes , and these ideas could potentially be incorporated in our colocalisation test . Another related issue is the choice of prior probabilities for the various configurations . For the eQTL analysis , we used a prior probability for a cis-eQTL . A more stringent threshold may be better suited for trans-eQTLs where the variants are further away from the gene under genetic control . We also used a prior probability of for the lipid associations . Although our knowledge about this is still lacking , this estimate has been suggested in the literature in the context of GWAS [20] , [31] , [32] . We assigned a prior probability of for , which encodes the probability that a variant affects both traits . It has been shown that SNPs associated with complex traits are more likely to be eQTLs compared to other SNPs chosen at random from GWAS platforms [33] , and a higher weighting for these SNPs has been proposed when performing Bayesian association analyses [34] , [35] . Also , eQTLs have been shown to be enriched for disease-associated SNPs when a disease-relevant tissue is used [9] , [36] . Our sensitivity analysis for the parameter showed broadly consistent results ( Table S1 ) . In cases where GWAS data are available for both traits , [10] show that it is possible to estimate these parameters from the data using a hierarchical model . This addition is a possible extension of our approach . The interpretation of the posterior probabilities requires caution . For example , a low PP4 may not indicate evidence against colocalisation in situations where PP3 is also low . It may simply be the result of limited power , which is evidenced by high values of PP0 , PP1 and/or PP2 . Moreover , a high PP4 is a measure of correlation , not causality . To illustrate this , one can consider the relatively common situation where a single variant appears to affect the expression of several genes in a chromosome region ( as observed , for example , in the region surrounding the SORT1 gene ) . Several eQTLs will be colocalised , both between them and with the biomarker of interest . In this situation one would typically expect that a single gene is causally involved in the biomarker pathway but the colocalisation test with the biomarker will generate high PP4 values for all genes in the interval . We show that we can use conditional p-values to deal with multiple independent associations with the same trait at one locus . While we found this solution generally effective , Wallace [14] points out that this top SNP selection for the conditional analysis can create biases , although the bias is small in the case of large samples and/or strong effects . For difficult loci with multiple associations for both traits and available genotype data , it may be more appropriate to estimate Bayes factors for sets rather than single variants in order to obtain an exact answer . This extension would avoid the issue of SNP selection for the conditional analysis . Importantly , GWAS signals can be explained by eQTLs only when the causal variant affects the phenotype by altering the amount of mRNA produced , but not when the phenotype is affected by changing the type of protein produced , although the former seems to be the most common [33] . Furthermore , since many diseases manifest their phenotype in certain tissues exclusively [2] , [21] , [37] , [38] , colocalisation results will be dependent on the expression dataset used . In addition to identifying the causal genes , the identification of tissue specificity for the molecular effects underlying GWAS signals is a key outcome of our method . We anticipate that building a reference set of eQTL studies in multiple tissues will provide a useful check for every new GWAS dataset , pointing directly to potential candidate genes/tissue types where these effects are mediated . While this report focuses on finding shared signals between a biomarker dataset and a liver expression dataset , we plan to utilise summary results of multiple GWAS and eQTL studies , for a variety of cell types and traits . In fact , our method can utilise summary results from any association studies . Disease/disease , ( cis or trans ) eQTL/disease or disease/biomarkers comparisons are all of biological interest and use the same statistical framework . We expect that the fact that the test can be based on single SNP summary statistics will be key to overcome data sharing concerns , hence enabling a large scale implementation of this tool . The increasing availability of RNA-Seq eQTL studies will further increase the opportunity to detect isoform specific eQTLs and their relevance to disease studies . Owing to the increasing availability of GWAS datasets , the systematic application of this approach will potentially provide clues into the molecular mechanisms underlying GWAS signals and the aetiology of the disorders .
This paper re-analyses previously published datasets . All samples and patient data were handled in accordance with the policies and procedures of the participating organisations . We used in our analysis gene expression and genotype data from 966 human liver samples . The samples were collected post-mortem or during surgical resection from unrelated European-American subjects from two different non-overlapping studies , which have been described in [16] . The cohorts were both genotyped using Illumina 650Y BeadChip array , and 39 , 000 expression probes were profiled using Agilent human gene expression arrays . All of the expression data has been normalised as one unit even though they were part of different studies , since high concordance between data generated using the same array platforms has been previously reported . Probe sequences were searched against the human reference genome GRCh37 from 1000 Genomes using BLASTN . Multiple probes mapping to one gene were kept in order to examine possible splicing . The probes were kept and annotated to a specific gene if they were entirely included in genes defined by Ensembl ID or by HGNC symbol using the package biomaRt in R [39] . After mapping and annotating the probes , we were left with 40 , 548 mapped probes covering 24 , 927 genes . Quality control filters were applied both before and after imputation . Before imputation , individuals with more than 10% missing genotypes were removed , and SNPs showing a missing rate greater than 10% , a deviation for HWE at a p-value less than 0 . 001 were dropped . After imputation , monomorphic SNPs were excluded from analyses . To speed up the imputation process , the genome was broken into small chunks that were phased and imputed separately and then re-assembled . This was achieved using the ChunkChromosome tool ( http://genome . sph . umich . edu/wiki/ChunkxChromosome ) , and specifying chunks of 1000 SNPs , with an overlap window of 200 SNPs on each side , which improves accuracy near the edges during the phasing step . Each chunk was phased using the program MACH1 with the number of states set to 300 and the number of rounds of MCMC set to 20 for all chunks . Phased haplotypes were used as a basis for imputation of untyped SNPs using the software Minimac with 1000 Genomes European ancestry reference haplotypes ( phase1 version 3 , March 2012 ) to impute SNPs not genotyped on the Illumina array . Variants with a MAF less than 0 . 001 were also excluded post-imputation . The data was then collated in probability format that can be used by the R Package snpStats [39] . eQTL p-values , effect sizes , and standard errors were obtained by fitting a linear trend test regression between the expression of each gene and all variants 200 kilobases upstream and downstream from each probe . After filtering out the variants with MAF <0 . 001 , monomorphic SNPs , multi-allelic SNPs ( as reported in 1000 Genomes or in the Ensembl database ) and variants not sufficiently well imputed ( Rsq <0 . 3 , as defined by minimac http://genome . sph . umich . edu/wiki/minimac ) between both datasets , we applied our colocalisation procedure . We conducted conditional analysis on SNPs with p-values for the expression associations , and repeated the colocalisation test using expression data conditioned on the most significant SNP . The aim of this analysis is to explore whether additional signals for expression other than the main one are shared with the biomarker signal . The biomarker p-values from the meta-analyses ( with genomic control correction ) were obtained from a publicly available repository ( http://www . sph . umich . edu/csg/abecasis/public/lipids2010/ ) . The regional association plots for the eQTL and Biomarker datasets were created using LocusZoom [40] ( http://csg . sph . umich . edu/locuszoom/ ) . We call a “configuration” one possible combination of pairs of binary vectors indicating whether the variant is associated with the selected trait . We can group the configurations into five sets , , , , , , containing assignments of all SNPs Q to the functional role corresponding to the five hypothesis , , , , . We can compute the posterior probabilities given the data for each of these 5 hypothesis by summing over the relevant configurations: ( 1 ) where P ( S ) is the prior probability of a configuration , is the probability of the observed data D given a configuration S , and the sum is over all configurations S which are consistent with a given hypothesis , where h = ( 1 , 2 , 3 , 4 ) . Thus , the probability of the data given a configuration is weighted by the prior probability of that configuration . Next , to avoid computing the proportionality constant in Equation 1 , we can reformulate the posterior probability for each hypothesis by writing this quantity as a ratio . For example , the posterior probability under hypothesis 4 , dividing each of these terms by the baseline , is: ( 2 ) The ratios in the numerator and denominator of equation 2 are: ( 3 ) The first ratio inside the sum in this equation is a Bayes Factor ( BF ) for each configuration , and the second ratio is the prior odds of a configuration compared with the baseline configuration . The BF can be computed for each variant from the p-value , or estimated regression coefficient and variance of , using Wakefield's method . By summing over all configurations in we are effectively comparing the support in the data for one alternative hypothesis versus the null hypothesis . An in-depth description of the method making use of the current assumptions can be found in Text S1 . A Bayes Factor for each SNP and each trait 1 and 2 was computed using the Approximate Bayes Factor ( ABF , [20] ) . Wakefield's method yields a Bayes factor that measures relative support for a model in which the SNP is associated with the trait compared to the null model of no association . The equation used is the following: ( 4 ) where is the usual Z statistic and the shrinkage factor r is the ratio of the variance of the prior and total variance ( ) . Assuming a normal distribution , the p-value of each SNP can be converted to standard one-tailed Z-score by using inverse normal cumulative distribution function . So for a SNP , all that it is needed are the p-values from a standard regression output , and , the standard deviation of the normal prior N ( 0 , W ) on . The variance of the effect estimate , V , can be approximated using the MAF and sample size . However for imputed data it is preferable to use the variance outputted in standard regression analysis directly in the ABF equation . For the expression dataset used here , the variance and effect estimates from the regression analysis were used for computation of ABFs ( see Text S1 for more details ) . Prior probabilities are assigned at the SNP level and correspond to mutually exclusive events . We assigned a prior of for and , the probability that a SNP is associated with either of the two traits . Since all SNPs are assumed to have the same prior probability of association , this prior can be interpreted as an estimate for the proportion of SNPs that we expect to be associated with the trait in question . We also assigned a prior probability of for , the probability that one SNP is associated with both traits . This probability can be better understood when it is re-expressed as the conditional probability of a SNP being associated with trait 2 , given that it is associated with trait 1 . So assigning a probability of means that 1 in 100 SNPs that are associated with trait 1 is also associated with the other . As a sensitivity analysis , we ran the comparison with Teslovich et al . using two other prior probabilities for , which means 1 in 50 SNPs that are associated with one trait is also associated with the other; and which means 1 in 10 SNPs . To compute the ABF , we also needed to specify the standard deviation for the prior , and we set this to 0 . 20 for binary traits and 0 . 15 for quantitative traits ( more details in Text S2 ) . | Genome-wide association studies ( GWAS ) have found a large number of genetic regions ( “loci” ) affecting clinical end-points and phenotypes , many outside coding intervals . One approach to understanding the biological basis of these associations has been to explore whether GWAS signals from intermediate cellular phenotypes , in particular gene expression , are located in the same loci ( “colocalise” ) and are potentially mediating the disease signals . However , it is not clear how to assess whether the same variants are responsible for the two GWAS signals or whether it is distinct causal variants close to each other . In this paper , we describe a statistical method that can use simply single variant summary statistics to test for colocalisation of GWAS signals . We describe one application of our method to a meta-analysis of blood lipids and liver expression , although any two datasets resulting from association studies can be used . Our method is able to detect the subset of GWAS signals explained by regulatory effects and identify candidate genes affected by the same GWAS variants . As summary GWAS data are increasingly available , applications of colocalisation methods to integrate the findings will be essential for functional follow-up , and will also be particularly useful to identify tissue specific signals in eQTL datasets . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computer",
"applications",
"computer",
"and",
"information",
"sciences",
"web-based",
"applications",
"medicine",
"and",
"health",
"sciences",
"clinical",
"medicine",
"genetics",
"biology",
"and",
"life",
"sciences",
"population",
"biology",
"cardiology"
] | 2014 | Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics |
New systems genetics approaches are needed to rapidly identify host genes and genetic networks that regulate complex disease outcomes . Using genetically diverse animals from incipient lines of the Collaborative Cross mouse panel , we demonstrate a greatly expanded range of phenotypes relative to classical mouse models of SARS-CoV infection including lung pathology , weight loss and viral titer . Genetic mapping revealed several loci contributing to differential disease responses , including an 8 . 5Mb locus associated with vascular cuffing on chromosome 3 that contained 23 genes and 13 noncoding RNAs . Integrating phenotypic and genetic data narrowed this region to a single gene , Trim55 , an E3 ubiquitin ligase with a role in muscle fiber maintenance . Lung pathology and transcriptomic data from mice genetically deficient in Trim55 were used to validate its role in SARS-CoV-induced vascular cuffing and inflammation . These data establish the Collaborative Cross platform as a powerful genetic resource for uncovering genetic contributions of complex traits in microbial disease severity , inflammation and virus replication in models of outbred populations .
Severe Acute Respiratory Coronavirus ( SARS-CoV ) emerged in humans in Southeast Asia in 2002 and 2003 after evolving from related coronaviruses circulating in bats [1 , 2] . SARS-CoV caused an atypical pneumonia that was fatal in 10% of all patients and 50% of elderly patients [3 , 4] . Patients infected with SARS-CoV experienced fever , difficulty breathing and low blood oxygen saturation levels [5 , 6] . Severe cases developed diffuse alveolar damage ( DAD ) and acute respiratory distress syndrome ( ARDS ) and disease severity was positively associated with increased age [7] . Host genetic background is also thought to influence disease severity but this understanding is complicated by inconsistent sample collection , varying treatment regimens and the limited scope of the SARS epidemic in humans [3 , 8 , 9] . Existing animal models of SARS-CoV infection have revealed that this lethal pulmonary infection causes a denuding bronchiolitis and severe pneumonia which oftentimes progresses to acute respiratory failure [10 , 11 , 12] . More recently , a second emerging coronavirus designated Middle East Respiratory Coronavirus ( MERS-CoV ) emerged from bat and camel populations [13 , 14 , 15] , and has caused ~38% mortality . Given the complex interplay between environmental , viral and host genetic variation in driving viral disease severity , as well as the difficulty of studying those factors in episodic outbreaks of pathogens such as SARS-CoV , MERS-CoV and other highly virulent zoonotic pathogens that cross the species barrier at regular intervals , novel approaches are needed to understand and identify those factors contributing to these diseases . Host genetics play a critical role in regulating microbial disease severity , evidenced by the identification of highly penetrant host susceptibility alleles within CCR5 , FUT2 , IL-28B in controlling HIV , norovirus and HCV infection and disease severity , respectively [16 , 17 , 18] . However , most microbial infections cause complex disease phenotypes that are regulated by the interactions of oligogenic traits with reduced penetrance , making them extremely difficult to identify and validate in human populations during outbreaks . Mannose binding lectin ( MBL ) polymorphisms were alternatively associated with successful recovery from SARS-CoV infection and a poor outcome of infection [19 , 20] , reflecting the complexity of performing candidate gene or genome wide association studies with limited human samples . The generation of a mouse adapted strain of SARS-CoV , MA15 , allowed for development of a small animal model that replicated both human lung disease and the age-dependency of SARS-CoV pathogenesis [10] . MA15 infection of inbred mice deficient in various immune genes has greatly contributed to our understanding of the host response to SARS-CoV infection [21 , 22] . However , such studies have focused on extreme abrogation of rationally selected candidate genes and have not evaluated the role of undescribed polymorphisms in genes in a model mimicking the genetic diversity seen in the human population . As a complement to human genome wide association studies , here we apply a new approach designed to dissect the identity and contributions of monogenic and oligogenic variants on multiple traits in complex disease outcomes following acute virus infection in a mouse model of human populations . The Collaborative Cross ( CC ) , a novel eight-way recombinant inbred ( RI ) mouse strain panel , has recently become available to the scientific community [23 , 24 , 25] . The power of the CC for genetic mapping is enhanced by availability of complete genome sequences of the founder strains and rich bioinformatics resources [26 , 27 , 28] . The eight founder strains used to generate the CC ( A/J , C57BL/6J , 129S1/SvImJ , NOD/ShiLtJ , NZO/HILtJ , CAST/EiJ , PWK/PhJ and WSB/EiJ ) are phenotypically diverse and capture single nucleotide polymorphisms ( SNPs ) and insertion/deletions ( In/Dels ) at approximately twice the frequency of common variants in human populations [24 , 29 , 30 , 31 , 32] . The derivation of CC strains from these multiple founders has proven to be useful for identifying polymorphisms that are responsible for a variety of traits [23] . The CC supports precise genetic mapping and , because the CC strains are genetically reproducible , it also serves as a robust validation platform and reference resource for integrative systems genetics applications . Here , we studied incipient lines of the CC ( the preCC ) to identify host genes that contributed to SARS-CoV MA15 infection and pathogenesis . We identified four novel quantitative trail loci ( QTLs ) contributing to SARS-CoV pathogenesis . Within the HrS1 QTL , a combination of approaches applied to the CC platform predicted a single gene candidate , Trim55 , as the principle regulator of vascular cuffing after infection . Vascular cuffing is a commonly reported phenotype observed in response to a variety of insults including chemical injury and infection ( [33 , 34 , 35]; high levels of vascular cuffing have been observed in models of severely pathogenic SARS-CoV infection [21 , 22] . Fluid vascular cuffing has been reported to decrease lug compliance suggesting an important physiologic consequence of this response [36] . Using knockout mice , we confirmed the role of Trim55 in immune cell infiltration , demonstrating the utility of the CC platform for identifying single gene candidates that likely regulate novel immune functions in trans-endothelial migration and perivascular cuffing following virus infection .
Mice from the eight founder strains as well as 147 eight to twenty week old female preCC mice were infected with 105 plaque forming units ( PFU ) of mouse adapted SARS-CoV , designated MA15 [10] , and weight loss was observed over the course of a four day infection . At day four post infection mice were euthanized and tissue collected for assessment of viral load in the lung as well as virus-induced inflammation and pathology . A wide range of susceptibilities to SARS-CoV infection was found among the founder strains of the CC and the overall heritability of weight changes following SARS-CoV infection determined to have a coefficient of genetic determination of 0 . 72 . NOD/ShiLtJ mice were resistant to infection and gained weight over the course of the experiment ( Figs 1A and S1A ) . A/J , C57BL/6J , 129S1/SvImJ and NZO/HILtJ mice experienced moderate and transient weight loss as previously described [21 , 22] while CAST/EiJ , PWK/PhJ and WSB/EiJ mice demonstrated extreme susceptibility to SARS-CoV infection including substantial weight loss and death ( Figs 1A , S1A and S1B ) . Subsequent dose response studies using the three highly susceptible wild-derived strains indicated an LD50 of between 100 and 500 PFU for CAST/EiJ mice , between 500 PFU and 1000 PFU for PWK/PhJ and between 103 and 105 PFU for WSB/EiJ mice ( S1 Table ) . PreCC mice infected with SARS-CoV ranged from over 30% weight loss by day four post infection to over 10% weight gain ( Fig 1A ) , exceeding the range of susceptibilities observed in the founder strains . Additionally , 26 preCC mice ( 18% of the preCC cohort ) succumbed to infection prior to the day four harvest point indicating extreme susceptibility to SARS-CoV infection . Viral load in the lung at day four post infection was determined for each surviving preCC mouse as well as for each of the founder strains . Viral lung titers showed a heritability of 0 . 60 as measured by the coefficient of genetic determination amongst the 7 surviving founder strains . Amongst the founder strains , PWK/PhJ mice had the lowest viral loads in the lungs , with 1 . 75x103 PFU per lung at day four post infection ( Figs 1B and S1B ) . PWK/PhJ mice also showed significant weight loss and a low LD50 indicating that viral load was unlikely to be responsible for pathogenesis in these mice . In contrast , C57BL/6J mice had the highest amount of virus at 6 . 35x106 PFU per lung . Lung titers in the preCC mice ranged from below the limit of detection ( 100 PFU/lung ) to over 108 PFU per lung , greatly exceeding the range of viral loads in the founder strains . Some preCC mice had viral loads in the lung below the 100 PFU limit of detection , despite having substantial weight loss . CAST/EiJ mice are extremely susceptible to SARS-CoV infection and do not survive until the day four post infection timepoint . Fig 1C shows the relationship between weight loss and lung titer at day four post infection . We found no correlation between viral load in the lung at day four post infection and weight loss ( r = -0 . 014 , p = 0 . 8938 ) when excluding those animals with viral loads below the limit of detection . When those animals were included in the analysis there is a significant , but not very explanatory correlation ( r = -0 . 347 , p = 0 . 00019 ) between the two phenotypes . Multiple aspects of lung pathology were assessed in surviving preCC animals including disease and immune infiltrates in the airways , vasculature , alveoli and parenchyma and signs of DAD ( S2 Table ) . A wide variety of lung pathologies were found across the preCC mice including denudation of airway epithelial cells , airway debris , eosinophilia , hyaline membrane formation and vascular cuffing ( Fig 2A–2F ) . Quantification of the overall pathology score along with select data ranges are shown in S2 Fig . Hyaline membrane formation and pulmonary edema with accompanying inflammation in the alveoli was a hallmark of SARS-CoV infection in human cases and is also evident in aged mouse models of disease [11] . In contrast to young founder strain animals , robust hyaline membrane formation was observed in 13% of preCC mice at day four post-infection , demonstrating that improved animal models are one likely outcome of infection studies in the CC . Phenotypic correlations of varying strengths were observed between aspects of lung pathology , inflammation , viral load at day four post infection , as well as weight loss across the course of the study ( Fig 3 ) . We genotyped 140 preCC animals at high density , including several that succumbed to infection prior to the scheduled day four harvest . As previously described [23 , 27] , we conducted QTL mapping using Bagpipe ( http://valdarlab . unc . edu/software . html ) and the underlying eight founder strain haplotypes present in the CC to identify host genome regions containing polymorphisms significantly associated with SARS-induced disease responses . We identified four QTLs shown in Fig 4 , HrS1-4 ( Host response to SARS ) that contributed to disease associated phenotypes at day four post infection . We identified a significant main effect QTL for vascular cuffing ( Chr 3: 18286790–26668414 ) , which explained 26% of the variation in vascular cuffing . We also identified two highly suggestive ( genome-wide p-values based on permutations of 0 . 1>p>0 . 05 ) QTL for viral titer ( Chr 16: 31583769–36719997 ) and eosinophil infiltration ( Chr 15: 72103120–75803414 ) , explaining 22% and 26% of the variation in these traits respectively . Finally , we also searched for modifier QTL , those QTL additively influencing a trait of interest , but whose presence was initially masked by our three other identified QTL . We identified a significant QTL further influencing vascular cuffing ( Chr 13: 52822984–54946286 ) , explaining an additional 21% of the variance in this phenotype . HrS4 was a moderate peak even without considering HrS1 status , suggesting that these interactions are additive . Table 1 details each of the SARS susceptibility QTLs including LOD and p-values . Analysis of other phenotypes did not lead to discovery of QTLs passing the p<0 . 01 significance threshold . The genetic architecture of the preCC , with up to eight distinct haplotypes at each locus , provides unique opportunity for narrowing QTL regions to candidate genes or SNPs . To narrow QTL regions we estimated the additive allele effects associated with each haplotype and correlated these to the allelic states at candidate causative polymorphisms . Allele effects [23] describe the estimated effect of each of the eight founder haplotypes on the phenotype ( e . g . a large positive allele effect for the PWK/PhJ haplotype suggests that having a PWK/PhJ allele will increase the phenotypic trait value of interest ) . In our analysis we focused on polymorphisms corresponding to the largest contrast between allele effects at the peak QTL locus . For HrS1 we identified two haplotypes , C57BL/6J and WSB/EiJ increasing vascular cuffing relative to the haplotypes of the other six founder strains . For each of HrS2-4 , we identified a single founder haplotype altering the phenotype relative to the seven other founder haplotypes ( HrS2: PWK/PhJ haplotype reduced viral titer; HrS3: A/J haplotype increasing eosinophillic infiltration; HrS4: CAST/EiJ haplotype reduced vascular cuffing ) . We then used high coverage whole genome sequence from the eight founder strains [37] to identify either private SNPs or small In/Dels in the case of a single causative haplotype , or regions of shared descent ( in the case of two causative haplotypes ) to narrow down the large QTL regions . HrS1 was initially an 8 . 38 Mb region which contained 26 genes and 9 non-coding RNAs ( ncRNAs ) . Identification of the sub-regions where C57BL/6J and WSB/EiJ share private , common ancestry reduced this region to 449 kb , which contained only one gene , one pseudogene and one miRNA of unknown function ( Trim55 , GM7488 and AC107456 . 1 , respectively ) . Allele effects for all four QTLs can be seen in S3 Fig . The HrS2 QTL on chromosome 16 was a 5 . 4 Mb region containing 92 genes and 30 ncRNAs . Across the eight founder strains , there were 95 , 936 SNPs or small In/Dels , and 33 , 288 of these were private to PWK/PhJ . Seven ncRNAs and 74 genes had private PWK/PhJ SNPs or In/Dels ( S3 Table ) . We further prioritized these variants based on whether the PWK/PhJ private polymorphisms were likely to cause major functional changes to the gene ( missense , nonsense , stop gained/lost , splice alterations or nonsense mediated decay ) . When we did so , we further reduced this list to 48 candidate genes including several mucins as well as genes involved in T cell activation and apoptosis . The HrS3 QTL on chromosome 15 was a 3 . 7 Mb region containing six ncRNAs and 63 genes . There were a total of 71 , 208 SNPs or small In/Dels in the region , 932 of which were private to A/J . No ncRNAs and only 25 genes contained a private A/J SNP or In/Del , and we further reduced these to one candidate gene with major functional changes ( S4 Table ) , Bai1 . Bai1 is a high priority candidate gene given the association between eosinophils and angiogenesis [38]; however we have not chosen to focus on Bai1 at this time because of the limited availability of tools for working on an A/J genetic background . Finally , HrS4 on chromosome 13 was a 2 . 12 Mb region containing three ncRNAs and 30 genes . There were a total of 461 , 46 SNPs or In/Dels in the region , 9 , 732 being private to CAST/EiJ . 29 of the genes and all three ncRNAs contained private CAST/EiJ polymorphisms ( S5 Table ) . When we further prioritized based on major functional changes , we reduced the region to only one ncRNA and nine genes including Cdhr2 , a member of the protocadherin family [39] . We focused our validation efforts on Trim55 , the single HrS1 candidate and a member of the TRIM protein superfamily which has not previously been associated with any infectious disease phenotype . Although many TRIM proteins function in innate immunity and inflammation , Trim55 ( also known as muscle-specific RING finger 2 or Murf2 ) has only been studied in the context of muscle development and cardiac function [40 , 41] . Trim55 is expressed in smooth muscle surrounding blood vessels [42] , an appropriate location to influence perivascular cuffing phenotypes . Knockout mice on a C57BL/6J background have previously been reported [43] and were kindly made available to our laboratory . Groups of age matched Trim55-/- and C57BL/6J control mice were infected with 105 PFU of MA15 for four days and monitored daily for weight loss and signs of disease . Trim55-/- and C57BL/6J animals had similar weight loss profiles as well as similar viral loads in the lung at day four post infection ( Fig 5A and 5B ) and no differences in mortality . Hematoxylin and eosin stained lung sections showed significantly reduced vascular cuffing in the lungs of Trim55-/- ( mean score of 0 . 69 ) compared to control animals ( mean score of 1 . 15 ) ( p < 0 . 05 by students t test , Fig 5C ) , confirming the role of Trim55 in contributing to SARS-CoV-induced vascular cuffing . Additional mice were infected for flow cytometric analysis of inflammatory cell populations in the lung after MA15 infection . While we observed a general trend towards increased numbers of T cells , B cells and macrophages in the lungs of C57BL/6J control mice compared to the Trim55-/- mice , only monocyte numbers were significantly different between the two groups . Total monocytes , as well as the subset of Ly6C positive monocytes , were present in significantly higher numbers in the lungs of infected control mice compared to Trim55-/- mice ( Fig 5D ) . RNA was isolated from the lungs of mock and infected control and Trim55-/- mice at two and days four post infection . 168 genes were identified as differentially expressed ( DE , log2 fold change >2 relative to mock ) between the two strains , predominantly at day two post infection ( GEO accession GSE64660 ) . We then used Ingenuity Pathway Analysis software to identify functionally enriched gene categories . This analysis identified the granulocyte and agranulocyte diapedesis gene ontology categories as among the most significantly enriched ( first and third respectively ) from genes with DE between Trim55-/- and B6 controls ( Fig 6A ) . Diapedesis , or extravasation , is the process by which inflammatory monocytes and leukocytes bind to endothelial cells and migrate from the blood stream into surrounding injured tissues . The transcriptional analysis indicates decreased expression of tight junction genes and increased chemokine expression in C57BL/6J mice compared to that observed in Trim55-/- mice . Relative expression of genes involved in granulocyte adhesion and diapedesis at days two and four post infection is shown in Fig 6B and 6C .
Emerging coronaviruses like SARS-CoV and MERS-CoV cause high morbidity and mortality in human populations . Because of limited access to clear human disease responses and samples from acute infections , as well as the limited number of overall infected individuals , it is extremely challenging to define the role of host genetic polymorphism in human disease . Coronavirus pathogenesis is heavily influenced by host genetics , as evidence by the extreme resistance of SJL mice , which encode a defective variant CEACAM1 receptor for mouse hepatitis virus entry and infection [44] . Furthermore , genetic monomorphisms in the cheetah have resulted in extreme hypersensitivity to feline infectious peritonitis coronavirus infection , underscoring the importance of abundant genetic variation in controlling lethal coronavirus infection [45 , 46] . In this study we examined numerous phenotypes following SARS-CoV infection and identified 4 QTL related to various aspects of SARS-CoV pathogenesis . These data support previous predictions that the CC platform can identify genetic variants contributing moderate effect sizes ( e . g . ~20% ) to complex immune response traits . Two of the four identified QTL , on chromosome 3 and 13 respectively , pertained to perivascular cuffing . Perivascular cuffing in the lung is frequently observed during microbial and non-microbial lung disease [34 , 47 , 48] and is associated in part with extravasation , the process by which inflammatory cells migrate from the blood to surrounding tissues [49 , 50] . Previous reports of perivascular cuffing include lymphocyte and granulocyte involvement with limited insights into the genetic underpinnings of this phenotype . In vivo models of SARS-CoV infection have shown that vascular cuffing increases in cases of severe disease [21 , 22 , 35] and vascular congestion was observed in human SARS-CoV patients [7] . A recent study of pneumococcal infection [51] identified several QTL governing disease susceptibility including one on chromosome 13 . The authors also found an association between perivascular inflammation and susceptibility to infection but did not extend their genetic analysis to that phenotype; there was no overlap between their chromosome 13 QTL and Hrs4 . Analysis of pulmonary inflammation following hyperoxia-induced lung injury [52] identified QTL on chromosomes 1 , 2 , 4 , 6 and 7 and informative SNPS helped to identify Chrm2 as the causative gene on chromosome 6 . In this study we identified QTL contributing to 26% and 21% of the total vascular cuffing phenotypic variance , respectively . The limited numbers of candidate genes under the larger effect size QTL allowed us to test and validate the role of Trim55 in SARS-CoV-induced perivascular cuffing phenotype . The CC was conceived to expand upon the genetic variation and mapping precision found within classical recombinant inbred ( RI ) panels which often suffer from inability to narrow the numbers of candidate genes due to the close genetic relationship of the founding lines . The classical BxD panel—derived from C57BL/6J and DBA2/J founder strains–was used previously to identify Klra8 , the resistance gene to mouse cytomegalovirus ( Cmv–1 ) infection [53] . Importantly , the validation experiments were conducted over a decade after the initial identification of the Cmv–1 susceptibility locus [54] as the wide initial QTL interval was not sufficient for identification of specific candidate genes . The Collaborative Cross provides a significant advantage in comparison to two-way crosses and other bi-allelic RI strain panels–as illustrated by our study , allele effects associated with founder haplotypes can provide a substantial reduction in the list of plausible candidate loci . Moreover , the inclusion of a diverse set of founder strains increases the likelihood of variants existing at loci that can influence any given trait . Indeed , in our study five of the eight founder strains contributed minor , causative alleles to the four QTL we identified . As the breeding of the CC lines preserved genetic variation across the genome , the CC lacks genetic blind spots and has multiple variant alleles at each locus . With a wide range of phenotypes [23] , the CC recapitulates aspects of the genetic diversity of the human population , making it a powerful system for use in causal genetic analyses . This study was part of an early pilot project to demonstrate the utility of the CC panel [23] . As such we did not have access to fully inbred animals and were limited to a single animal per genotype . However , the increased control of the experimental conditions in these studies and high frequency of minor alleles within the CC population ( each allele is present in roughly 12 . 5% of CC genomes [23] whereas minor allele frequencies in the human population are typically much lower ) allowed us to identify multiple host genome regions contributing to differential SARS-CoV infection . Studies utilizing the full CC panel will be able to use the full potential of a reproducible genetic background to obtain repeated assays and high-precision phenotyping , even our limited proof-of-concept study proved to be adequate to identify multiple host genome regions contributing to differential responses to SARS-CoV infection . Trim55 is part of the well-known superfamily of TRIM proteins , specifically in the C-II subfamily . This subfamily consists of Trim54 , Trim55 and Trim63 , and is defined by an N-terminal that contains a Ring Finger domain , B-box 2 domain and a coiled-coil domain [42] . The C-II Trim family genes are solely expressed by muscle cells and to date have only been studied in the frame of muscle cell development and cardiac function . Trim55 and Trim63 , also known as Murf1 , mediate muscle cell protein turnover through their E3 ubiquitin-ligase activities and function in muscle wasting phenotypes [40 , 43 , 55] . Trim55 specifically functions in myosin and myofibril maintenance and knockdown studies correlate Trim55 levels with modified post-translational microtubule modifications and defects in myofibril assembly , critical components in extravasation [55] . Blood vessels are comprised of vascular endothelial cells , connective tissue and smooth muscle cells , all of which must be crossed during inflammatory cell trafficking into the lung . During extravasation , inflammatory cells tumble and bind to adhesion molecules , slowing their motion and expanding surface-surface interactions with endothelial cells [56] . Tissue Necrosis Factor-alpha and thrombin expression levels increase following SARS-CoV infection [12 , 21] and these proteins have both been shown to increase endothelial permeability [57 , 58] . Here we observed a complicated picture of altered chemokine and tight junction gene expression in the absence of Trim55 ( Fig 6B and 6C ) . Increased expression of Ccl24 , CCR3 , IL4 and Pdgfc in C57BL/6J mice compared to that in Trim55-/- mice at day four post infection correlates with increased inflammatory cell recruitment and binding to extracellular matrix proteins . These expression changes are consistent with changes in altered recruitment of inflammatory cells to the lung following SARS-CoV infection . Higher expression of Claudin19 at day two post infection in Trim55 deficient mice likely contributes to decreased tight junction permeability and reduced ability for inflammatory cells in the bloodstream to cross the endothelial barrier . Additionally , one of the high priority candidate genes under the modifier QTL on chromosome 13 is Cdhr2 , a cadherin superfamily member that may also play a role in extravasation of inflammatory cells into the infected lung . Different specific VE-cadherin residues are known to regulate leukocyte extravasation and vascular permeability [59] , demonstrating the importance of cadherin family members in these processes . More recent work details the role of Cdhr2 in intestinal brush border assembly via adhesion links between adjacent microvilli [60] . Intravascular crawling and signaling through RhoA induces actin , microfilament and microtubule reorganizations and the production of endothelial cell docking structures , which surround the inflammatory cell and span tight junctions [56] . Although controversial , myofibril contractile structures may also contribute in to the assembly of these structures . In any event , inflammatory cell transmigration requires the formation of actin-myosin II contractile structures which are attached to tight junction membranes by VE-cadherins , resulting in increased endothelial tension , and programmed separation and expansion of the tight junctions which allow for leukocyte/monocyte passage into the surrounding tissues [61] . It seems likely that Trim55 , with its roles in myosin and myofibril maintenance and microtubule organization , contributes to the programmed formation of endothelial docking structures and regulation of inflammatory cell transmigration; key features associated with the formation of perivascular cuffs around vessels in the lung . Our data ( Figs 5 and 6 ) demonstrate that Trim55 contributes to vascular cuffing following SARS-CoV infection . While the mechanism is not yet fully understood , the data strongly suggest that Trim55 is important for extravasation of inflammatory cells , and thus overall SARS-CoV pathogenesis , by altering intercellular junctions and chemotactic signals . Increased studies of Trim55 and Cdhr2 function within the CC population , either via specific crosses of lines with high and low alleles at the HrS1 and HrS4 loci , or via CRISPR-Cas9 modification of these loci will allow further insight into the role that these two genes play during SARS-CoV pathogenesis and recovery , as well as increasing understanding of the more general process of extravasation . The Collaborative Cross was conceived of as a resource to drive insight into a variety of biomedically important diseases via the reassortment of genetic variants and expansion of phenotypic ranges [62] . Indeed , previous studies with various preCC subsets have demonstrated expanded phenotypes in preCC mice body weight and hematological parameters [23 , 63 , 64] , response to Aspergillus [65] and susceptibility to Influenza A infection [27 , 66] . More recently it has been shown that novel combinations of alleles have also resulted in new models for human disease such a spontaneous colitis [67] , and that F1 hybrids of CC mice were used to create an improved mouse model for Ebola virus disease [68] including hemorrhagic signs of disease previously not observed in a small animal model . Within our study of SARS-CoV infection within the preCC , we showed more extreme disease phenotypes than those seen within the eight founder strains of the CC . These disease phenotypes included virus titer , weight loss , pathology and lethality . Further , we saw the emergence of new phenotypes including ARDS and DAD not traditionally seen within young inbred strains [11] . Importantly , our results highlight another exciting aspect of the nature of CC genome: transgressive segregation , or the release of cryptic genetic variation [69 , 70] . As the three wild-derived CC founders all showed mortality early in the course of SARS-CoV infection , genetic variants within these three strains impacting later-stage SARS-CoV responses would normally not be seen . Only via the reassortment of these alleles into a variety of genetic backgrounds ( some resistant to clinical disease , some susceptible ) were we able to show that alleles from all three wild-derived founders impacted perivascular cuffing or viral titer levels independent of their effects on clinical disease or SARS-CoV mortality . Collectively , these data support the hypothesis that the CC population represents a robust platform for developing improved animal models that more readily replicate disease phenotypes seen in human populations . All told , our identification of multiple QTL related to SARS-CoV pathogenesis , identification of a novel function for Trim55 , and the development of new models of acute lung injury , further solidify the utility of the CC as a valuable community resource for research of infectious diseases and other biological systems driven by complex host response networks .
Mouse studies were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All mouse studies were performed at the University of North Carolina ( Animal Welfare Assurance #A3410-01 ) using protocols approved by the UNC Institutional Animal Care and Use Committee ( IACUC ) . Recombinant mouse-adapted SARS-CoV ( MA15 ) was propagated and titered on Vero E6 cells . For virus titration half of the right lung was used to assess plaque forming units ( PFU ) per lung using Vero E6 cells with a detection limit of 100 PFU [71] . All experiments were performed in a Class II biological safety cabinet in a certified biosafety level 3 laboratory containing redundant exhaust fans while wearing personnel protective equipment including Tyvek suits , hoods , and HEPA-filtered powered air-purifying respirators ( PAPRs ) . 8–12 week old female animals from the 8 founder strains ( A/J , C57BL/6J , 129S1/SvImJ , NOD/ShiLtJ , NZO/HILtJ , CAST/EiJ , PWK/PhJ , and WSB/EiJ ) were obtained from the Jackson labs ( jax . org ) , and bred at UNC Chapel Hill under specific pathogen free conditions . 8–20 week old female pre-CC mice were bred at Oak Ridge National Laboratories under specific pathogen free conditions , and transferred directly into a BSL–3 containment laboratory at UNC Chapel Hill . One preCC mouse per line was infected , for the founder strains at day four n = 2 ( A/J ) , n = 3 ( C57BL/6J , 128S1/SvImJ , NOD/ShiLtJ , CAST/EiJ , PWK/PhJ and WSB/EiJ ) and n = 5 ( NZO/HILtJ ) . Trim55-/- ( Murf2-/- ) mice on a C57BL/6 background were a kind gift from Christian Witt at the University of Mannheim . Validation experiments used 8–12 week old female mice . All experiments were approved by the UNC Chapel Hill Institutional Animal Care and Use Committee . Animals were maintained in SealSafe ventilated caging system in a BSL3 laboratory , equipped with redundant fans as previously described by our group . Animals were lightly anesthetized via inhalation with Isoflurane ( Piramal ) . Following anesthesia , animals were infected intranasally with 105 pfu of mouse adapted SARS-CoV ( MA15 ) in 50 μL of phosphate buffered saline ( PBS , Gibco ) , while mock infected animals received only 50 μL of PBS . Animals were weighed daily and at four days post infection , animals were euthanized via Isoflurane overdose and tissues were taken for various assays . No blinding was used in any animal experiments and animals were not randomized; group sample size was chosen based on availability of age-matched mice . Pearson’s correlation was used to determine any correlation between weight loss and log-transformed viral load in the lung . The left lung was removed and submerged in 10% buffered formalin ( Fisher ) without inflation for 1 week . Tissues were embedded in paraffin , and 5 μm sections were prepared by the UNC Lineberger Comprehensive Cancer Center histopathology core facility . To determine the extent of inflammation , sections were stained with hematoxylin and eosin ( H & E ) and scored in a blinded manner by for a variety of metrics relating to the extent and severity of immune cell infiltration and pathological damage on a 0–3 ( none , mild , moderate , severe ) scale . Significant differences in lung pathology were determined by a two-sample student’s t test . Images were captured using an Olympus BX41 microscope with an Olympus DP71 camera . The right lung of each mouse was used for flow cytometric staining of inflammatory cells . Mice were perfused with PBS through the right ventricle before harvest , lung tissue was dissected and digested in RPMI ( Gibco ) supplemented with DNAse and Collagenase ( Roche ) . Samples were strained using a 70 micron filter ( BD ) and any residual red blood cells were lysed using ACK lysis buffer . The resulting single cell suspension was stained with two antibody panels using the following stains ( 1 ) FITC anti-Ly-6C clone AL21 ( BD ) , PE anti-SigLecF clone E50-2440 ( BD ) , PETR anti-CD11c clone N418 ( MP ) , PerCP anti-B220 clone RA3-6B2 ( MP ) , PE-Cy7 anti-Gr–1 clone RB6-8C5 ( eBio ) , eF450 anti-CD11b clone M1/70 ( eBio ) , APC anti-LCA clone 30-F11 ( eBio ) , APC-eF780 anti-MHC class II clone M5/114 ( eBio ) or ( 2 ) FITC anti-CD94 clone 18d3 ( eBio ) , PE anti-CD3Ɛ clone 145-2C11 ( eBio ) , PETR anti-CD4 clone RM4-5 ( MP ) , PerCP anti-CD8 clone 53–6 . 7 ( BD ) , PE-Cy7 anti-CD49b clone DX5 ( eBio ) , eF450 anti-LCA clone 30-F11 ( eBio ) , AF647 anti-CD19 clone 6D5 ( Biolegend ) , APC-eF780 anti-B220 clone RA3-6B2 ( eBio ) . While this FACS analysis was solely performed on mice of a C57BL6/J background , these antibodies have all been shown recognize the relevant antigens in each of the CC founder lines . Samples were run on a Beckman Coulter CyAN , and data analyzed within the Summit software . Significant differences in lung inflammatory cell populations were determined by a two-sample student’s t test . Genotyping and haplotype reconstruction were done as described in [23] . Briefly , each pre-CC animal was genotyped using the Mouse Diversity Array [72] ( Affymetrix ) at 372 , 249 well performing SNPs which were polymorphic across the founder strains [31] . Once genotypes were determined , founder strain haplotype probabilities were computed for all genotyped loci using the HAPPY algorithm [73] . Genetic map positions were based on the integrated mouse genetic map using mouse genome build 37 [74] . Linkage mapping was done as described in [23] . Briefly , QTL mapping was conducted using the BAGPIPE package [75] to regress each phenotype on the computed haplotypes in the interval between adjacent genotype markers , producing a LOD score in each interval to evaluate significance . Genome-wide significance was determined by permutation test , with 250 permutations conducted per scan . Phenotype data for mapping either satisfied the assumptions of normality or were log transformed to fit normality ( titer data ) . For the likely regions of identified QTL peaks , SNP data for the eight founder strains from the Sanger Institute Mouse Genomes Project [37] were downloaded and analyzed as described in Ferris et al [27] . Bagpipe is freely available at http://valdarlab . unc . edu/software . html . At two and four days after infection , mice were euthanized and a lung portion placed in RNAlater ( Applied Biosystems/Ambion ) and then stored at −80° . The tissues were subsequently homogenized in TriZol ( Life Technologies ) , and RNA extracted as previously described [12] . RNA samples were spectroscopically verified for purity , and the quality of the intact RNA was assessed using an Agilent 2100 Bioanalyzer . cRNA probes were generated from each sample by the use of an Agilent one-color Quick-Amp labeling kit . Each cRNA sample was then hybridized to Agilent mouse whole-genome oligonucleotide microarrays ( 4 x 44 ) based on the manufacturer’s instructions . Slides were scanned with an Agilent DNA microarray scanner , and the output images were then analyzed using Agilent Feature Extractor software . Microarray data has been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus database and is accessible through GEO accession GSE64660 . Raw Agilent Microarray files were feature extracted Agilent feature extractor version 10 . 7 . 3 . 1 . Raw Microarray files were background corrected using the “norm-exp” method with an offset of 1 and quantile normalized using Agi4x44PreProcess [76] in the R statistical software environment . Replicate probes were mean summarized , and all probes were required to pass Agilent QC flags for 75% replicates of at least one infected time point ( 41 , 267 probes passed ) . This microarray analysis was performed only on animals with a C56BL6/J background; thus it was not necessary to correct for probes with SNPs caused by the genetic variation of the other founder lines . Differential expression was determined by comparing MA15 infected samples ( C57Bl6/J mice vs . Trim55-/- mice ) with mock and each other to fit a linear model for each probe using the R package Limma . Criteria for differential expression was an absolute log2 FC of 1 and a q value of < 0 . 05 calculated using a moderated t test with subsequent Benjamini-Hochberg correction . Differentially expressed ( DE ) genes were observed for both C57BL/6J and Trim55-/- infected mice compared to time matched mocks at two and day four post infection . DE analysis was also run on the Trim55-/- infected mice against the C57BL/6J infected mice which provided a direct observation of the transcription signatures in the Trim55-/- against the MA15 infected mouse background . To identify genes with similar patterns of variation at early and late times post infection , day two and day four gene signatures were intersected separately and then combined . There was no intersection of DE genes between the day two and four time points when the Trim55-/- infected mice were run against the C57BL/6J infected mice . Functional analysis of statistically significant gene expression changes was performed using the Ingenuity Pathways Knowledge Base ( IPA; Ingenuity Systems ) [77] . Functional enrichment scores were calculated in IPA using all probes that passed our QC filter as the background data set . | New emerging pathogens are a significant threat to human health with at least six highly pathogenic viruses , including four respiratory viruses , having spread from animal hosts into the human population within the past 15 years . With the emergence of new pathogens , new and better animal models are needed in order to better understand the disease these pathogens cause; to assist in the rapid development of therapeutics; and importantly to evaluate the role of natural host genetic variation in regulating disease outcome . We used incipient lines of the Collaborative Cross , a newly available recombinant inbred mouse panel , to identify polymorphic host genes that contribute to SARS-CoV pathogenesis . We discovered new animal models that better capture the range of disease found in human SARS patients and also found four novel susceptibility loci governing various aspects of SARS-induced pathogenesis . By integrating statistical , genetic and bioinformatic approaches we were able to narrow candidate genome regions to highly likely candidate genes . We narrowed one locus to a single candidate gene , Trim55 , and confirmed its role in the inflammatory response to SARS-CoV infection through the use of knockout mice . This work identifies a novel function for Trim55 and also demonstrates the utility of the CC as a platform for identifying the genetic contributions of complex traits . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Genome Wide Identification of SARS-CoV Susceptibility Loci Using the Collaborative Cross |
Discrete clusters of circadian clock neurons temporally organize daily behaviors such as sleep and wake . In Drosophila , a network of just 150 neurons drives two peaks of timed activity in the morning and evening . A subset of these neurons expresses the neuropeptide pigment dispersing factor ( PDF ) , which is important for promoting morning behavior as well as maintaining robust free-running rhythmicity in constant conditions . Yet , how PDF acts on downstream circuits to mediate rhythmic behavior is unknown . Using circuit-directed rescue of PDF receptor mutants , we show that PDF targeting of just ∼30 non-PDF evening circadian neurons is sufficient to drive morning behavior . This function is not accompanied by large changes in core molecular oscillators in light-dark , indicating that PDF RECEPTOR likely regulates the output of these cells under these conditions . We find that PDF also acts on this focused set of non-PDF neurons to regulate both evening activity phase and period length , consistent with modest resetting effects on core oscillators . PDF likely acts on more distributed pacemaker neuron targets , including the PDF neurons themselves , to regulate rhythmic strength . Here we reveal defining features of the circuit-diagram for PDF peptide function in circadian behavior , revealing the direct neuronal targets of PDF as well as its behavioral functions at those sites . These studies define a key direct output circuit sufficient for multiple PDF dependent behaviors .
Circadian clocks act in many organisms to promote daily rhythms of behavior and physiology . In Drosophila , clock function under conditions of light-dark entrainment ( 12-h light∶12-h dark; LD ) is evident as increases in locomotor activity in advance of lights-on ( morning anticipation ) and lights-off ( evening anticipation ) . These rhythms are driven by well-conserved transcriptional feedback loops in which the basic helix-loop-helix transcription factor heterodimer , CLOCK/CYCLE , activates components such as period ( per ) , timeless ( tim ) , and clockwork orange ( cwo ) that feedback and regulate CLOCK/CYCLE binding to its cognate DNA targets [1]–[4] . These feedback loops generate daily gene expression rhythms . Approximately 150 pacemaker neurons in the adult Drosophila brain are implicated in the regulation of circadian locomotor behavior . These neurons can be roughly divided into the PIGMENT DISPERSING FACTOR ( PDF ) -expressing small and large LNv ( sLNv , lLNv ) , a single non-PDF sLNv , the dorsal lateral neurons ( LNd ) , and three groups of dorsal neurons ( DN1 , DN2 , and DN3 ) [5] . Ablation of PDF+ neurons results in substantial reduction in morning anticipation [6] , [7] . A functional clock in the small subset of PDF+ neurons is sufficient to drive morning behavior , and these cells have thus been dubbed “morning” ( M ) cells [8] . The large LNv have been observed to promote arousal especially during the light period [9]–[11] . A subset of ∼30 circadian pacemaker neurons , including the non-PDF sLNv , LNd , and/or a small subset of DN1s and DN3s [7] , [8] , [12] , [13] , are essential for evening anticipatory behavior , and are thus dubbed “evening” ( E ) cells . Mammalian circadian clocks may also have a similar morning and evening organization [14] , [15] . Drosophila also maintains robust locomotor activity rhythms during constant-dark conditions ( DD ) , reflecting the endogenous function of its circadian clock . The PDF-expressing LNv play a critical role in sustaining free-running rhythms , as ablation of the PDF+ LNv leads to decreased DD rhythmicity [6] . Moreover , tissue-specific rescue experiments indicate that the circadian clock component PERIOD ( PER ) [7] and the circadian output ion channel NARROW ABDOMEN ( NA ) [16] are each required in the PDF+ LNv to promote robust , sustained DD rhythmicity . The function of PDF neurons is instructive , as selectively altering the period of these cells drives changes in period length in several non-PDF neurons and sets the circadian period of locomotor activity [17] . It is not known if the ability of PDF neurons to influence non-PDF pacemaker neurons reflects a direct cellular connection . The PDF neuropeptide is implicated as the principal transmitter of the LNv group , as flies lacking Pdf function exhibit phenotypes similar to ablation of the PDF+ LNv [6] . In LD , these phenotypes include reduced morning behavior and advanced evening behavior . During DD , null Pdf01 mutants exhibit progressive dampening of locomotor rhythmicity and a slightly shortened period . A receptor for Drosophila PDF has been identified ( PDFR , aka han , groom-of-pdf , CG13758 ) , and loss of this receptor leads to circadian phenotypes essentially identical to Pdf01 mutants [18]–[20] . The DD behavioral phenotypes of Pdf01 mutants are accompanied by alterations in the molecular clock . PER oscillations in the DN1 of Pdf01 mutants rapidly damp during DD , indicating a role for PDF in sustaining molecular rhythms [21] . In contrast , the LNd of Pdf01 mutants exhibit persistent rhythms , but with an advance in the phase of PER oscillations , consistent with the observed short behavioral period of these flies [22] . Additionally , desynchronized PER nuclear localization rhythms are observed in the sLNv of Pdf01 mutants , but only after many days of DD [22] . These data suggest that PDF may also reset or synchronize these molecular clocks . However , molecular alterations have not been observed in Pdf01 mutants in LD [23] , suggesting that PDF may be acting downstream of the molecular clock under these conditions . While the molecular consequences of manipulating PDF/PDF RECEPTOR ( PDFR ) function have been well described , it was not previously known which of these effects reflected the direct actions of PDF on the affected cells or whether they were mediated by cellular intermediates . In addition , it was not known which of these direct cellular targets was mediating the multiple effects of PDF on behavior , especially under LD conditions . Here we demonstrate that PDFR expression limited to the ∼30 non-PDF evening cells can not only alter the timing of evening behavior , but also drive the amplitude of morning behavior . Our data indicate that the effect of PDFR expression on morning behavior does not likely occur through the core clock , but instead through the regulation of neuronal output . We also demonstrate a role for PDFR in non-PDF cells to reset evening phase and regulate period length , consistent with core clock resetting . Finally , we find that PDFR likely functions within a more distributed group of pacemaker neurons , including the PDF+ LNv , to promote sustained DD rhythmicity . This study defines the major direct targets for PDF in vivo and their functions in circadian behavior .
To define the neuroanatomical targets of PDF action in circadian behavior , we performed tissue-specific rescue of a Pdfr mutant using the GAL4-UAS system . For these experiments , we utilized a strong loss-of-function mutant allele of Pdfr , han5304 . Like null Pdf01 mutants , Pdfr han5304 flies display strongly reduced morning anticipation and phase-advanced evening anticipation in LD , as well as a reduced morning peak at the onset of DD [6] , [18] . Previous studies had suggested that PDFR functions in circadian neurons largely based on partial rescue using a single perGAL4 driver [18] . perGAL4 drivers , in addition to demonstrating expression in all major circadian pacemaker groups , also drive widespread expression in nominally noncircadian brain areas , including the central complex , antennal lobe , and lateral horn [24] , raising questions as to the precise site of PDFR function . To address this issue , we utilized clockGAL4 [25] , which drives broad expression among all major circadian neuronal groups [16] but relatively limited noncircadian expression , including the pars intercerebralis ( PI ) and cells surrounding circadian neurons [16] . Using this driver , we find that PDFR expression in Pdfr mutants rescues morning anticipation and the proper timing of LD evening behavior ( Figures 1A–1C and S1; p<0 . 05 ) . Given the relatively limited noncircadian expression of clockGAL4 , these results suggest a major function for PDFR in circadian neurons . We next assessed PDFR function specifically in the pacemaker neuron subsets known to control morning and evening behavior . We performed rescue using a GAL4 driver containing the promoter and first intron of the cryptochrome gene ( cryGAL4-13 ) [26] . cryGAL4-13 drives expression in both PDF-expressing morning cells and ∼30 non-PDF evening cells ( LNv , LNd , small subset of DN1 and DN3 ) , while promoting little or no expression in other circadian pacemaker neurons ( e . g . , most DN1 , all DN2 , and most DN3 ) or outside the circadian system [7] , [13] , [16] . cryGAL4-13 driven expression of UAS-Pdfr restores the timing of evening behavior ( p<0 . 0001 ) and also promotes significant restoration of morning behavior during LD and the first day of DD ( DD1; Figures 1D , 2A–2C; Table 1; p<0 . 001 ) . We examined and quantified morning behavior during DD1 as the lights-on response in LD can mask some of the clock-driven morning behavior . We then further restricted UAS-Pdfr expression specifically to either the evening cells or morning cells . Expression was restricted to evening cells by blocking GAL4 induction selectively in PDF+ cells using GAL80 ( PdfGAL80; cryGAL4-13 ) , while morning cell-specific expression was driven using PdfGAL4 . Expressing UAS-Pdfr only in non-PDF evening cells rescues both the timing of evening behavior and the magnitude of morning anticipation ( Figures 1E and 2D; Table 1; p<0 . 0001 ) . In contrast , UAS-Pdfr expression restricted to morning cells does not have comparable effects on morning or evening behaviors ( Figures 1F and 2E; Table 1 ) , as previously reported [18] . These findings suggest that the PDF+ LNv can communicate directly to the non-PDF “evening” cells through PDFR to promote morning behavior . We next examined whether the behavioral contribution of evening cells to morning behavior might be driven by changes in the circadian clock . The etiology of circadian phenotypes in flies with disrupted PDF signaling has largely focused on the role of PDF in synchronizing and/or resetting circadian clocks . These studies have largely identified changes in molecular oscillations of core clock components , such as PER , which reflect core clock timing , under constant darkness conditions . It has been proposed that light can compensate for the loss of PDF/PDFR as no large changes in the core clock have been described in Pdf01 mutants in LD [23] . However , these experiments were performed with only two time points . Given our interest in determining the molecular basis of morning and evening behavioral phenotypes in LD , we performed PER immunolabeling in wild-type ( UAS-Pdfr/+ ) and Pdfr mutant ( Pdfr han5304; UAS-Pdfr/+ ) flies during LD using four time points . Previous studies have shown that PER expression restricted to PDF neurons is sufficient to rescue morning anticipation of per01 mutants , suggesting that PDF actions do not require clock function in non-PDF neurons for morning behavior [8] . We asked whether changes in the LNv clock could account for loss of morning behavior in Pdfr mutants . However , no significant differences in PER oscillations were observed in the PDF+ sLNv important for morning behavior ( Figure 3A–3C ) . While small changes were observed in some pacemaker neuron clusters ( lLNv , LNd , and DN3 ) , high amplitude oscillations were observed in all pacemaker neurons groups in Pdfr mutants , in contrast to the highly significant reduction in morning behavior ( Figure 2 ) . In addition , clock function in the lLNv is not required for morning behavior [7] , [8] , [27] , suggesting ( but not excluding ) the possibility that these small molecular changes do not underlie changes in morning behavior . The LNd and a subset of DN3 have been implicated in regulating evening behavior [7] , [13] , and subtle changes in the LNd and/or DN3 may be responsible for the ∼2-h phase advance in evening anticipation ( Figure 1B ) . A higher temporal resolution will be necessary to definitively demonstrate a molecular phase shift in these cell clusters . Nonetheless , relatively small phase changes are unlikely to explain large amplitude changes in morning behavior . In this case , the dramatic effects of PDFR on morning behavior largely reflect its function in circadian output . In addition to defects in morning and evening behaviors , Pdf01 and Pdfr mutants exhibit decreased rhythmic power and shortened period length in DD [6] , [18] , [19] . To determine whether the anatomical requirements for PDFR function in free-running rhythmicity match those for morning and evening behaviors , we assessed DD rhythms in PDFR rescue flies . Expression of PDFR using a broad circadian driver clockGAL4 promotes significant rescue of DD rhythmicity , as reflected by rhythmic power ( p<0 . 0001 ) and period length ( p<0 . 0001; Table 2 ) . Period length of clockGAL4 rescue flies is slightly short ( 23 . 3+/−0 . 1 h ) , yet comparable to clockGAL4 driven overexpression of PDFR in a wild-type background ( 23 . 4+/−0 . 1 h; Table 2 ) , likely due to a modest ( ∼30 min ) overexpression effect ( Table 2 ) . Nonetheless , clockGAL4 rescue of period is statistically significant and supports a role for PDFR in circadian neurons to promote normal DD period and rhythmicity . To assess PDFR DD function in specific circadian neuron subsets , we analyzed DD rescue using GAL4 drivers with limited circadian expression . Expression of PDFR in morning and evening cells using cryGAL4-13 rescues DD period length and partially restores DD rhythmicity ( Table 2; p<0 . 0001 ) , suggesting broader or stronger DN expression provided in clockGAL4 may be needed for robust rhythmicity . However , further restriction of PDFR expression to evening cells using PdfGAL80 fully blocks the cryGAL4-13 rescue of rhythmic power ( p = 0 . 6; Table 2 ) . While we cannot rule out residual GAL4 activity , these data are consistent with full GAL80 repression of GAL4 activity in the LNv . These data uncouple LD and DD rescue and suggest a role for PDFR within the PDF+ LNv to promote DD rhythms . Yet consistent with previous findings [18] , PDFR expression restricted to PDF+ neurons ( PdfGAL4 ) has no significant effect on either free-running rhythmicity or period length in DD , indicating that PDFR function in these cells is not sufficient for normal DD rhythms ( Table 2 ) . To confirm a role for PDFR in the PDF+ neurons , we expressed PdfGAL80 in the context of pan-neuronal elavGAL4-mediated rescue . elavGAL4 expression results in strong rescue of all LD and DD phenotypes ( Figure 1G , Table 2; p<0 . 0001 ) . As with cryGAL4-13 , blocking elavGAL4 driven PDFR selectively in the PDF+ LNv using PdfGAL80 results in a substantial reduction in rhythmic power ( Table 2; p<0 . 0001 ) . Taken together , these data suggest that PDF/PDFR communication within the LNv plays an important role in sustaining robust DD rhythmicity . In addition , our rescue also suggests that other cells also contribute to DD rhythmicity . Notably , period length is also significantly rescued with all GAL4 drivers tested except PdfGAL4 ( Table 2; p<0 . 0001 ) . Yet unlike rhythmic power , period length rescue is unchanged when PDFR expression is blocked in PDF+ neurons via PdfGAL80 . In fact , PDFR expression restricted only to evening cells ( PdfGAL80; cryGAL4-13 ) retains significant rescue of period length , as evident from group activity profiles ( Figure 4 ) and individual fly analyses ( Table 2 ) . This remains true even if only strongly rhythmic flies ( Power≥40 ) are considered ( unpublished data; p<0 . 0001 ) . Thus , direct PDF communication among PDF-expressing neurons , as well as with other target neurons , is important for sustaining DD rhythms . In contrast , the PDF+ LNv communicate directly to non-PDF evening cells to set period length , indicating functional and anatomical specialization of PDF signaling . Taken together , our functional neuroanatomical approaches highlight PDFR function in circadian pacemaker neurons . However , reports of the PDFR expression pattern are conflicting . Two initial reports utilized independently generated antisera to assess PDFR expression in the Drosophila brain . One reported expression limited mainly to circadian neurons [18] , while the other observed broad expression that included only few circadian neurons [20] . We previously reported pdfr expression using in situ hybridization and noted expression in potential dorsal neurons and the PI [19] . A more recent report indicates that the reported immunofluorescence patterns may not represent specific PDFR signal [28] , calling into question the true PDFR expression pattern . We have made several additional attempts to generate specific antisera to PDFR but have yet to identify reproducible and robust signals ( unpublished data ) . To examine Pdfr expression , we instead used a P-element exchange strategy to insert a P{GAL4} element ∼40 bp upstream of the presumptive Pdfr transcription start site ( PdfrGAL4; see Materials and Methods ) [29] . A targeted GAL4 insertion into a locus of interest has been a valuable approach to report endogenous gene expression patterns [30] . If the GAL4 insertion falls under the control of enhancers that normally drive Pdfr expression , we predict that PdfrGAL4 will reflect endogenous Pdfr expression . In this case , PdfrGAL4 if combined with UAS-Pdfr should be able to rescue Pdfr mutant phenotypes . The original insert used to generate PdfrGAL4 displayed a modest circadian rhythmicity phenotype [20] and a ∼50% reduction in transcript levels ( unpublished data ) . Consistent with these data , we find that PdfrGAL4/Pdfrhan5304 flies display poor DD rhythmicity ( Table 2 ) . Importantly , this reduced rhythmicity is strongly rescued by PdfrGAL4 driven expression of UAS-Pdfr ( Table 2; p<0 . 0001 ) , suggesting that PdfrGAL4 is a faithful reporter of Pdfr expression . We then examined the driven expression pattern for PdfrGAL4 . Upon crossing PdfrGAL4 to UAS-nuclear green fluorescent protein ( GFP ) ( nGFP ) , we observe broad GFP expression in the adult Drosophila brain , including circadian neuron regions , PI , optic lobe , and ellipsoid body ( Figure 5A ) , the latter possibly consistent with noncircadian functions of PDF in arousal and geotaxis [9]–[11] , [20] , [31] . Given our rescue data , we more closely examined expression within circadian pacemaker neurons . To directly assess circadian expression , we labeled PdfrGAL4 UAS-nGFP brains with PER antisera . We observe prominent GFP expression in the sLNv , all LNd , and several DN1 ( Figure 5B and 5C ) . The PI and DN expression is consistent with our published in situ expression pattern [19] . Weak expression is observed in the lLNv . We also consistently observe expression in two DN3s , and we sometimes observe expression in one of the two DN2s ( unpublished data ) . These expression data nicely complement our functional neuroanatomy data . Expression in the sLNv is consistent with a role for PDFR in these cells to sustain free-running rhythmicity ( see PdfGAL80 ) , as the sLNv are known to be especially important for DD rhythmicity [8] , [17] . Pdfr expression in the LNd , DN1 , and DN3 subset is consistent with our data demonstrating an important role of the non-PDF evening cells in morning and evening behavior and DD period length .
Here we define the direct targets of PDF using circuit-specific rescue and find that the direct action of PDF on just ∼30 neurons , the so-called evening pacemaker neurons , mediates PDF dependent effects on morning , evening , and free-running behaviors . We corroborate our functional rescue data with a novel GAL4 enhancer trap reporting endogenous pdfr expression . We also provide strong evidence that PDF , in addition to its well-described effects on the core clock mechanism , also likely affects the output of pacemaker neurons providing novel mechanistic insight into PDFR function . These studies define a major direct conduit for in vivo PDF signaling in circadian behavior . A number of reports have examined the molecular consequences of manipulating PDF neuron function . Altering the core clock , output , or projections of PDF neurons alters the molecular clock in non-PDF circadian neurons and evening behavior under short days or in constant darkness [17] , [19] , [22] , [23] , [32]–[36] . However , these studies leave open a number of key questions important for elucidating the PDF circuit diagram . Not surprisingly , functional changes in PDF neurons can be propagated widely through the nervous system , not only to the direct cellular targets of that group of neurons ( primary target neurons ) , but to the targets of those targets ( secondary ) , and so on ( tertiary ) . Thus , the direct and indirect effects of PDF could not be distinguished in these papers . In addition , these studies do not identify the behavioral functions of PDFR ( or PDF ) at these different cellular targets particularly on LD morning and evening behavior . Some of these studies also rely on analysis of mutant flies with significant developmental abnormalities [34] , [36] . Measurements of PDF activation in ex vivo brains have also been used to infer direct cellular targets [28] . Bath application of PDF to cultured brains up-regulated cAMP levels . However , these assays required ∼1 min to observe significant activation . Given the slow response time course relative to the faster rate of synaptic transmission , PDF effects on a primary target neuron could be propagated through circuitry to secondary neurons to increase cAMP , on a similar minute time course . Thus , one cannot exclude the possibility that some of the observed responses may be indirect . In addition , effects might even reflect direct responses to nonphysiological levels of PDF . Moreover , this study does not address the behavioral functions of PDF at those sites . By using the direct molecular target of PDF , the PDF receptor , to rescue Pdfr mutant phenotypes , we functionally define these direct neuronal targets in vivo . We demonstrate that the expression of PDFR in a highly restricted group of neurons ( ∼30 neurons ) is sufficient to rescue morning behavior , evening phase , and circadian period , thus defining a major direct output circuit for multiple PDF-dependent behaviors ( Figure 6 ) . How does PDF function at these neuronal targets ? Previous studies have identified molecular clock changes especially under constant darkness conditions indicating that PDF acts to reset molecular oscillators . Consistent with this model , we observed that PDFR expression in the E cells can rescue circadian period and evening activity phase . Moreover , we identified subtle molecular changes in E cells in LD that are consistent with a small phase shift in molecular oscillations . Thus , at least some PDF-dependent behaviors can be attributed to its function in resetting clocks . While there are PDF effects on core clocks , our data also suggest an additional output function particularly in regulating morning behavior ( Figure 6 ) . Both Pdf and Pdfr mutants have been shown to have strong effects on the amplitude of morning behavior . Our studies similarly demonstrate major changes in the amplitude of morning behavior despite robust oscillations in both the sLNv ( which are sufficient for morning behavior ) as well as other circadian cell groups including the E cells . The published data further support this model . The finding that PDF can acutely affect neuronal firing rate in other insects [37] strongly suggested that such clock-independent output functions were possible , if not likely , as core clock changes and their subsequent translation into neuronal firing changes would take place over a longer time frame . It has previously been shown that rescue of clocks exclusively in the PDF neurons in the arrhythmic per01 mutant rescues morning behavior [8] . If morning behavior works by PDF targeting of the E cells , then PDF must act on the output of the E cells in these flies , as there is no clock in the E cells . In addition , manipulation of sodium channel activity shifts the phase of PDF rhythms and morning behavior but these are not accompanied by shifts in molecular oscillators in the sLNv , LNd , or DN1 , consistent with an output function [33] . Taken together , we believe our data coupled to the published literature support the notion that PDF can affect the output of E cells in addition to its phase resetting effects . Previous data have suggested that PDF activates MAPK phosphorylation in the dorsal brain just prior to the increase in morning activity [38] , suggesting that PDF release may promote morning activity . Consistent with this hypothesis , recent data suggest a role for the PDF expressing large LNv in driving locomotor activity and arousal [9]–[11] , [39] . These effects may be mediated through the sLNv , which in turn project to the dorsal brain [10] , [11] . PDF release in the morning may also reset oscillators in the E cells ( Figure 6 ) . Our identification of a role for so-called “E” cells in M behavior , also fits well with prior data suggesting that E cells can control M behavior and highlights additional complexity of the M-E model . Manipulating the clock in E cells can shift morning behavioral phase under long photoperiods [40] , whereas rescue of the arrhythmic per0 mutant in non-PDF neurons can rescue morning behavior [7] . However , these results were interpreted to indicate that E cell clocks signal through M cell clocks to drive morning behavior . Indeed , these authors proposed that M cells signal through unknown circuits to drive morning behavior [7] . Here we demonstrate that the E cells themselves are direct targets of the M cells to drive morning behavior . Given our data , E cells may signal to other pacemaker neurons or even nonpacemaker neurons rather than to M cells to drive morning behavior . How then does one reconcile the apparent observation that clock function is sufficient in E cells to drive morning behavior with the observation that E cells are not necessary for M behavior [7] ? One possibility is that redundant pathways control morning behavior . Thus , PDF communication to E cells is sufficient , but may not be necessary , to drive morning behavior . Nonetheless , these data demonstrate that the function of M and E cells is more intertwined than previously thought , necessitating a revision of the simplest versions of the M-E model . As E cells constitute a focused yet heterogeneous group of cells [7] , [13] , [35] , [41] , [42] , it will be of interest to determine whether distinct subsets of them are responsible for E and M behavior . E cells consist of the non-PDF small LNv , two DN3s , the LNd , which can be further subdivided by their expression of Neuropeptide F ( NPF ) [43] , and a subset of DN1s , two of which persist from larval development and the remainder that express the transcription factor GLASS [21] . We have attempted to rescue Pdfr mutant phenotypes using PdfGAL4 and npfGAL4 combined , but we fail to observe significant rescue of any LD or DD phenotypes ( unpublished data ) , suggesting a role for PDFR in E cells other than the NPF-expressing LNd . The DD period is likely driven from some or all LNd , as DN1 rhythms of Pdf01 mutants rapidly damp in DD while LNd rhythm persist with a short period for several days in DD [22] , comparable to the period of DD locomotor rhythmicity in Pdf01 or Pdfr mutants . Moreover , since the residual DD rhythms in Pdf01 and Pdfr mutants occur in the evening , we propose that the LNd may contribute to the phase advanced LD evening behavior in these flies . Nonetheless , disco mutant flies that lack intact LNs but retain DNs also retain evening anticipation; this suggests redundant LN and DN pathways for evening behavior [44] . GLASS+ DN1s are missing in glass mutant flies and these flies display an intact evening peak but an altered morning peak , in that this peak is poorly entrained and variable in phase [45] . This suggests that the GLASS+ DN1 may be important for morning behavior . Additional functional cell-specific reagents will be necessary to assess the relative contribution of the PDF-sLNv , LNd , DN1 , and DN3 in PDF-dependent circadian behaviors . While our data suggest that the E cells are an important conduit for PDF action in the brain especially for circadian period , phase , and morning behavior , we also find that multiple targets are likely important for regulating rhythmic strength ( Figure 6 ) . In E cell only rescue , we do not observe significant rescue of rhythmic strength , indicating that other cells are relevant . Knockdown of pan-neuronal rescue in PDF neurons substantially reduces rhythmic strength ( Table 2 ) . On the other hand , PdfGAL4-mediated rescue does not rescue DD rhythmicity . Thus , PDFR function in PDF neurons is necessary but not sufficient for DD rhythmic strength . Based on our expression analyses of PdfrGAL4 ( Figure 5 ) and PDF responsiveness by PDF application [28] , these target cells are likely the PDF+ small LNv . We have observed a similar function for the LNv in regulating rhythmic strength in tissue-specific rescue of na mutants [16] . Desynchronized molecular rhythms in these cells may contribute to the reduction in rhythmic strength observed in Pdf01 mutants [22] . Importantly , PDF neurons are not the only targets of PDF relevant to sustaining DD rhythms . Expression in broader sets of neurons including E cells ( cryGAL4-13 ) , most circadian pacemaker neurons ( clockGAL4 ) , and all neurons ( elavGAL4 ) results in progressively increasing levels of rhythmicity ( Table 2 ) . In addition , PdfGAL80 knockdown of pan-neuronal rescue does not suppress rhythmicity to mutant levels , further highlighting the role of both PDF neurons and non-PDF neurons in DD rhythmicity . The rescue data and PdfrGAL4 pattern presented here are also largely consistent with a report on PDF-responsiveness in the adult Drosophila brain [28] . Shafer et al . [28] observe PDF responsiveness in each of the circadian neuron groups ( PDF+ sLNv , non-PDF sLNv , lLNv , LNd , DN1 , DN2 , DN3 ) , albeit only weak responsiveness in a subset of lLNv assayed . The LN responsiveness matches PdfrGAL4 quite well , as we observe PdfrGAL4/UAS-GFP expression in all sLNv , all LNd , and weakly in a subset of lLNv ( Figure 5B ) . Among the DN clusters , we observe PdfrGAL4/UAS-nGFP in approximately half of the DN1 ( Figure 5C ) , reproducibly in two DN3 , and occasionally in one of the two DN2 ( unpublished data ) . Whereas Shafer et al . report PDF responsiveness in most DN cells assayed , these experiments were performed using a cryGAL4-39/UAS-Epac-cyclicAMP reporter . cryGAL4-39 expression has been reported to include only a subset of DN1s and DN3s , and ( in some reports ) DN2s , comparable to the DN pattern we describe for PdfrGAL4 [21] , [28] , [42] . Moreover , as noted above , these PDF-response measurements could reflect some degree of indirect responsiveness . Despite the likely complexity of PDF function in circadian behavior , the data presented here define a major direct output pathway for PDF-dependent circadian behaviors . These studies highlight both the function in resetting core clocks as well as communicating timing information downstream of these core oscillators . It will be of interest to further refine the targets in the circadian system as well as define the molecular and cellular mechanisms by which PDF acts on those neural circuits to regulate circadian behavior .
For rescue experiments , either Pdfrhan5304 [18] , Pdfrhan5304;; UAS-Pdfr [20] , or Pdfrhan5304; elavGAL4 [46] virgin females were crossed to y w , GAL4/GAL80 , or UAS males . For overexpression experiments , UAS-Pdfr ( line 10 ) flies were crossed to either y w ( control ) or specific GAL4/GAL80 strains . For PdfrGAL4 rescue experiments , female progeny were used for behavioral assays . For all other behavior , male progeny were assayed . Locomotor activity levels were monitored using Trikinetics Activity Monitors for 5 d of LD followed by 7 d of DD at 25°C . For LD analyses ( Figure 1 ) , activity levels from each fly were normalized and averaged within genotypes over 4 d , as described previously [47] . For DD analyses ( Figures 2 and 4 ) , activity levels were normalized and averaged over the last 2 d of LD followed by 7 d of DD . To calculate time of evening anticipation in LD ( Table 1 ) , we determined the largest 2-h increase in normalized average activity for each fly over the last 7 h of the light phase . The time designation refers to the end point of the maximal activity increase , as averaged among individual flies in each genotype . To quantitatively analyze morning behavior , we examined the first day of DD , as the lights-on peak in LD can mask the increase in morning behavior . To calculate DD Day 1 Morning Index ( Table 1 ) , normalized activity levels were averaged over three consecutive 30-min time points . For each genotype , maximum average activity of the group was determined for any two consecutive 30-min time points over the 6 h surrounding CT 0 ( ZT 21- CT3 ) . Minimum average activity was then determined for all time points before and after the observed maximum activity , up to 7 h before or after CT 0 ( ZT 17- CT 7 ) . Morning index value was obtained by subtracting the average of these minimum values from the maximum activity value . For DD rhythmicity ( Table 2 ) , chi-squared periodogram analyses were performed using Clocklab ( Actimetrics ) . Rhythmic flies were defined as those in which the chi-squared power was ≥10 above the significance line . Period calculations also considered all flies with rhythmic power ≥10 , with the exception of one outlier removed as indicated . All p-values reported were calculated using Student's two-tailed t-tests . Male Pdfrhan5304;UAS-Pdfr/+ and UAS-Pdfr/+ flies were entrained for 3–5 d at 25°C and anesthetized with CO2 . The flies were dissected in 3 . 7% formaldehyde diluted in PBS at ZT1 , ZT7 , ZT12 , and ZT18 . After fixing for 30 min at room temperature , the brains were rinsed three times in PBS and incubated in PBT ( PBS with 0 . 1% Triton ) for 10 min at room temperature . The brains were then incubated with 5% goat serum diluted in PBT for 30 min at room temperature , followed by overnight incubation of 1∶4 , 000 rabbit anti-PER diluted in PBT containing 5% goat serum at 4°C . After several PBT rinses , the brains were incubated with 1∶500 goat-anti-rabbit AlexaFluor 594 ( Amersham ) in PBT overnight at 4°C . Final rinses in PBT and PBS were followed by mounting in 80% glycerol diluted in PBS . All slides were coded as to sample identity and remained so until the numerical analysis stage . PER-stained specimens were photographed with 60× oil lens on a Nikon Eclipse 800 laser scanning confocal microscope . For a given experiment the microscope , laser , and filter settings were held constant , and all specimens were photographed in the same microscopy session . PER immunostaining was quantified from digitally projected Z stacks using ImageJ ( NIH ) . PER-stained soma were outlined to obtain average pixel intensity . On each projection image an unstained area was quantified to be used for background subtraction . All background-subtracted intensity measurements within a condition ( time and genotype ) were averaged . To combine experiments , background subtracted measurements were scaled to ZT1 of Pdfrhan5304;UAS-Pdfr/+ in that experiment . Statistical analysis was conducted in SPSS and Excel using ANOVA . Targeted transposition was used to replace P{EY11181} , a P-element insertion approximately 40 bp upstream of the Pdfr transcription start site , with P{GawB} , a P-element containing GAL4 . To perform targeted transposition , P{EY11181} , P{GawB} CyO flies were crossed to P-element transposase [29] . Strains in which P{GawB} mobilized to the X chromosome were identified by eye color and then analyzed by genomic PCR , to determine whether the GAL4 element had inserted into the pdfr upstream region . One strain ( PdfrGAL4-19 ) was identified using this method , and the insertion position of the GAL4 element was confirmed using inverse PCR ( Model Systems Genomics , Duke University ) . For expression analyses , PdfrGAL4 flies were crossed to UAS-nuclearGFP ( UAS-nGFP ) . Female progeny were entrained , dissected , and labeled with anti-PER protein as previously described [16] . Images were obtained using laser scanning confocal microscopy ( Nikon C1 ) . | Animals depend on being awake at the right time of day to find food and mates and fend off predators . Circadian pacemaker neurons in the brain play a crucial role in timing of specific behaviors to the appropriate times of day . These neurons are further specialized to those primarily responsible for morning and evening behavior . We have used the fruit fly Drosophila as a simple model system to elucidate the neural circuits important for timed daily behavior . In flies , a small group of clock neurons devoted to morning behavior express a neuropeptide , PIGMENT DISPERSING FACTOR ( PDF ) . Until now it was unclear what the direct neural targets of this peptide are and how its actions at those targets mediate timed behavior . Here we find that the so-called morning clock neurons communicate directly to other clock neurons , those responsible for evening behavior . This communication sustains high amplitude morning activity and sets the phase of evening activity as well as the period of activity rhythms in constant conditions . These studies reveal the circuit diagram for PDF function in circadian behavior . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/neuronal",
"and",
"glial",
"cell",
"biology",
"genetics",
"and",
"genomics/animal",
"genetics",
"neuroscience",
"mental",
"health/sleep",
"disorders",
"cell",
"biology/gene",
"expression"
] | 2009 | The Neuropeptide PDF Acts Directly on Evening Pacemaker Neurons to Regulate Multiple Features of Circadian Behavior |
Vasopressin neurons generate distinctive phasic patterned spike activity in response to elevated extracellular osmotic pressure . These spikes are generated in the cell body and are conducted down the axon to the axonal terminals where they trigger Ca2+ entry and subsequent exocytosis of hormone-containing vesicles and secretion of vasopressin . This mechanism is highly non-linear , subject to both frequency facilitation and fatigue , such that the rate of secretion depends on both the rate and patterning of the spike activity . Here we used computational modelling to investigate this relationship and how it shapes the overall response of the neuronal population . We generated a concise single compartment model of the secretion mechanism , fitted to experimentally observed profiles of facilitation and fatigue , and based on representations of the hypothesised underlying mechanisms . These mechanisms include spike broadening , Ca2+ channel inactivation , a Ca2+ sensitive K+ current , and releasable and reserve pools of vesicles . We coupled the secretion model to an existing integrate-and-fire based spiking model in order to study the secretion response to increasing synaptic input , and compared phasic and non-phasic spiking models to assess the functional value of the phasic spiking pattern . The secretory response of individual phasic cells is very non-linear , but the response of a heterogeneous population of phasic cells shows a much more linear response to increasing input , matching the linear response we observe experimentally , though in this respect , phasic cells have no apparent advantage over non-phasic cells . Another challenge for the cells is maintaining this linear response during chronic stimulation , and we show that the activity-dependent fatigue mechanism has a potentially useful function in helping to maintain secretion despite depletion of stores . Without this mechanism , secretion in response to a steady stimulus declines as the stored content declines .
Models of neuronal networks generally assume that the output of the neurons is well characterised by their spiking activity . However , for all neurons , their output is not the spikes themselves , but the neurotransmitter release that is triggered by those spikes , most commonly at synapses . Generally , the coupling between spike activity and transmitter release is nonlinear , subject to both frequency facilitation of release and to activity-dependent depression [1] , and when , as is often the case , neurons fire spikes in complex patterns , these non-linearities mean that the spike activity itself can be a poor approximation of their true output . Detailed study at presynaptic terminals is difficult however , and the peptide secreting nerve terminals of the posterior pituitary have served as a more technically accessible model system [2] , displaying similar complex activity patterning , and for which the characteristics of stimulus-secretion coupling have been well studied . Here we have studied the vasopressin neurons that project to the posterior pituitary , and develop a quantitatively precise model of both their spike activity and stimulus-secretion coupling , presenting a novel approach to modelling activity dependent facilitation and depression , based on abstractions of the underlying mechanisms . The magnocellular neuroendocrine neurons of the hypothalamus synthesise and secrete the hormones oxytocin and vasopressin . These hormones can be readily measured in the bloodstream , have well-understood physiological roles , and are secreted subject to well-characterised reflex pathways . The rare ability to measure the electrical activity of identified neurons in physiological circumstances together with the resulting activity-dependent secretion [3] has meant that these neurons have become important model systems for the study of stimulus-secretion coupling in peptidergic neurons , and both the mechanisms leading to spike activity and those leading to secretion have been extensively studied by a wide range of experimental approaches . By its actions at the kidney to regulate water loss , vasopressin has an essential role in the homeostatic regulation of plasma osmotic pressure . It is synthesised by magnocellular neurons in the supraoptic and paraventricular nuclei ( SON and PVN ) of the hypothalamus; these neurons project axons to the posterior pituitary gland , from where vasopressin is secreted into the blood stream following activity-dependent exocytosis of vesicles that are abundantly stored in axonal swellings and terminals . This secretion is triggered by spikes that are generated in the neuronal cell bodies and propagated down the axons . The vasopressin neurons are activated by increases in plasma osmotic pressure as a consequence partly of osmotically induced change in the cell volume , and partly of increased afferent input arising from other osmosensitive neurons in the forebrain [4] , [5] . When activated , they display a distinctive “phasic” pattern of spike discharge , and it has been proposed that the adaptive advantage of this patterning is that it optimises the efficiency of stimulus-secretion coupling . In the rat brain , there are about 9000 magnocellular vasopressin neurons , each of which individually generates spikes and manufactures and secretes hormone . These neurons are asynchronous in their spiking activity , and are quite heterogeneous , both in their levels of activity and in the membrane properties that determine specific features of their activity patterns . However , this individual phenotypic heterogeneity does not imply functional heterogeneity of the vasopressin cells: the physiologically important vasopressin signal is the plasma vasopressin concentration , which reflects the total secretion from the whole population . Although individual vasopressin cells generate complex phasic patterns of electrical activity , the plasma vasopressin concentration increases remarkably linearly in response to increasing osmotic pressure above a “set point” [6] , [7] . We have previously modelled the distinctive phasic spiking activity [8] in order to understand how this behaviour affects the coding of information in the vasopressin cell population . Because the phasic activity of vasopressin cells is asynchronous , and because of the properties of the mechanism that generates phasic activity , the average spike rate of the population increases relatively linearly in response to increased afferent input despite the short term non-linearity of individual cell responses . However , at the nerve terminals , stimulus-secretion coupling is also highly non-linear , and how much hormone is secreted per spike depends on multiple features of the spike activity pattern . At the axonal terminals , vesicle exocytosis is triggered by spike-generated Ca2+ entry , and , in both vasopressin and oxytocin neurons , the spikes broaden as spike frequency increases , increasing the amount of Ca2+ entry [9] , [10] . This spike broadening is thought to be partly responsible for an increase in secretion per spike that occurs with increasing frequencies of stimulation – a process called frequency facilitation [11] . This facilitation of vasopressin secretion peaks at about 15Hz , whereas oxytocin secretion per spike continues to increase up to frequencies of at least 50 Hz [12] . This difference suggests that in vasopressin terminals there is a competing , activity-dependent attenuation of stimulus-secretion coupling , possibly Ca2+-dependent inactivation of Ca2+ channels [13] . On a longer timescale , an additional mechanism produces activity-dependent fatigue of stimulus-secretion coupling in vasopressin cells ( and to a much lesser extent in oxytocin cells ) [14] . When axon terminals are stimulated at 13 Hz , the amount of vasopressin secreted declines progressively after about 10 s . However , if stimulation is interrupted by a silent period of 10 s or longer , the secretory response recovers . It was thus proposed that the function of phasic spiking is to optimise the secretion response , by minimising the consequences of fatigue while maximising those of facilitation [15]–[17] . However , there is a circularity in this logic . The differences between oxytocin and vasopressin neurons show that phasic firing is only efficient in vasopressin neurons because the particular properties of the vasopressin terminals make it so . We recently argued that there are other functional advantages to the phasic discharge pattern , and that the secretion mechanism in vasopressin cells may thus have evolved to confer efficiency on a spike pattern that has evolved for other reasons . Our objective here was to develop a model of stimulus-secretion coupling in vasopressin cells that concisely reproduces facilitation and fatigue , and which quantitatively and qualitatively matches experimentally observed responses to stimulation . The broader aim is to test how these features relate to the ability of vasopressin neurons to respond to osmotic pressure . By coupling the secretion model to our spiking model we can simulate secretion response to varied input activity . One of the challenges is to understand how the highly heterogeneous and non-linear vasopressin neurons act together to produce a highly linear secretion response to increasing osmotic input . Is cell heterogeneity just unavoidable noise , or does it have a useful functional role ? Our combined model shows that cell heterogeneity , in the form of varied input intensity across the population , combined with phasic patterning , acts to linearise the population response , achieving a profile of secretion that matches that observed in vivo . Interestingly , the linearising effect of heterogeneity only works in combination with phasic spiking , and not on its own . Thus in vasopressin cells , the non-linear stimulus secretion properties of the nerve terminals , in conjunction with the pattern generating characteristics of the vasopressin cells , and in conjunction with heterogeneity in the vasopressin cell population , all combine to generate robust linear stimulus-secretion coupling across a wide dynamic range of input .
The spike generating mechanism is modelled as described previously [8] , using an integrate-and-fire based spiking model modified to include a set of activity-dependent effects on excitability [18] that shape spike patterning and produce emergent bistability . The modifications include a hyperpolarising afterpotential ( HAP ) , a fast depolarising afterpotential ( DAP ) , a slow afterhyperpolarisation ( AHP ) , and a slow DAP based on a Ca2+-inactivated K+ leak current . The positive feedback of the activity-dependent slow DAP triggers bursting , but it is also opposed by the slower action of activity-dependent dendritic dynorphin release , which inactivates the DAP . The opposing effects combined with the random perturbations of the synaptic input produce a bistability that generates the successive periods of bursting and silence . Setting the K+ leak ( slow DAP ) conductance ( spiking model parameter gL ) to zero removes the mechanism that underlies bistability , producing model cells with very similar interspike interval distributions to vasopressin neurons but which fire continuously rather than phasically . We use this here to compare otherwise identical phasic and non-phasic cells . Synaptic input is simulated using a Poisson random process to generate EPSPs and IPSPs , represented by small ( 2 mV ) exponentially decaying positive and negative perturbations to the membrane potential . The spike outputs match experimental observations very closely , and the parameters of the model can be fit to in vivo recorded data to produce indistinguishably close fits to multiple statistical measures of spike patterning . Importantly the model also reproduces the observed in vivo behaviour in response to an increasing rate of input activity; matching the shift from slow irregular spiking , to phasic spiking , and eventually to continuous spiking activity . Our aim was to concisely model the mechanisms of stimulus-secretion coupling in a way that gives a robust qualitative and quantitative fit to the characteristics of spike driven secretion as measured experimentally . As a major simplification , we model secretion from a single cell as from a single compartment , rather than from thousands of individual terminals . We assume that the summed effect of the stochastic , discrete , and relatively rare secretion events at individual terminals can be approximated by a continuous model . The large number of terminals ( ∼2000 ) is likely to make this a safe assumption . It may be interesting to make a more detailed model including individual terminals , but we do not have the detailed experimental data against which to fit such a model . We fitted the model to experimental data reported in various studies of stimulus-secretion coupling in vitro in which vasopressin secretion from the isolated pituitary gland was measured in response to different patterns of electrical stimulation . These observations led to the description of stimulus-secretion coupling as being influenced by frequency facilitation , and by fatigue . Vesicle exocytosis depends on spike-evoked Ca2+ entry , and the rate of secretion depends on both the ion channels that propagate spikes and allow Ca2+ to enter , and the exocytotic apparatus that turns the Ca2+ signal into secretion . Frequency facilitation is thought to mainly reflect a modulation of Ca2+ entry in response to spike activity , using a spike broadening based mechanism detailed below . Modulation of the exocytotic response we include in the model by using a depletable vesicle store ( see below ) . Two hypotheses have been proposed to explain fatigue; that it is due to depletion of a readily-releasable pool of vesicles , or due to suppression of the secretion mechanism by a Ca2+-activated K+ conductance . The idea of depletion has previously been used to model secretion fatigue in oxytocin and vasopressin neurons [19] , and in other types of hormone-secreting cells [20]–[22] . However , these models have been used to fit experimental data where cells have been depleted by large and prolonged depolarisation . In recent work [23] we used detailed quantitative data to predict the rate of vesicle release in response to spike stimulation during physiological conditions , showing that release events at individual axonal terminals are rare , requiring several hundred spikes to trigger a single release event . It seems unlikely therefore that fatigue on the timescale of seconds is due to depletion of vesicle stores . Thus , we have based our present model on the likely effects of a Ca2+-activated K+ mechanism analogous to the mechanism underlying the afterhyperpolarisation observed at vasopressin cell bodies [24] . Spikes generated in the cell body are conducted down the axon to invade the secretory terminals – though whether a spike will invade a given terminal depends on the excitability of that terminal at the time of arrival of the spike . Like spikes at the cell body , spikes at the secretory terminals involve voltage-activated Na+ and Ca2+ channels , and terminal excitability ( membrane potential ) is modulated by various K+ conductances . Spike broadening is thought to be due to the voltage sensitive suppression of a K+ conductance , as demonstrated by experiments using the K+ channel blocker tetra-ethyl ammonia [2] , [25] . The other effects on terminal excitability appear to depend on Ca2+ entry . Ca2+ acts at the secretory terminals on at least two different timescales . Vesicle exocytosis depends more on the rate of Ca2+ entry than on intracellular Ca2+ concentration , and it is likely that exocytosis occurs at sites that are close to clusters of voltage-gated Ca2+ channels . These submembrane sites experience high Ca2+ concentrations in response to spike activity , but only transiently , as the Ca2+ swiftly diffuses into the cytosol . This is represented in the model ( Figure 1 ) by making the rate of secretion proportional to a ‘fast’ Ca2+ variable e that represents the sub-membrane Ca2+ concentration; spike broadening affects secretion by producing a larger rise in e . However , as mentioned above , the extent of facilitation of stimulus-secretion coupling in vasopressin terminals is limited by a competing , activity-dependent attenuation of stimulus-secretion coupling; we model this ( Figure 2 ) as arising from Ca2+-dependent inactivation of Ca2+ channels [13] . Prolonged pulse stimulation of terminals results in an eventual increase in the rate of failure of spike propagation at the terminal [26] , [27] , and this is thought to be due to intracellular ( cytosolic ) Ca2+ accumulation [10] and the activation of a highly [Ca2+]i sensitive , slow activating Ca2+-activated K+ conductance [28] . Thus in the model , a slow Ca2+ component c ( reflecting cytosolic Ca2+ concentration ) acts to reduce the terminal spike response . Its slow rate of accumulation and decay reproduces the slow development of fatigue , and the recovery during silent periods . The model also includes a finite vesicle store so that we can assess the useful function of these mechanisms in maintaining long-term response . The store has two compartments: a large reserve store ( r ) , and a smaller releasable pool ( p ) representing those vesicles in the axonal terminals that are relatively available for release . The secretion model consists of a set of differential equations which take as input the spike events generated by the spiking model , or defined by a stimulation protocol . The variables representing spike broadening ( b ) , cytosolic Ca2+ concentration ( c ) and submembrane Ca2+ concentration ( e ) are all incremented with each spike . Broadening ( b ) is incremented by a fixed step kb , and decays exponentially with half-life λb , converted to time constant τb: ( 1 ) where s = 1 if a spike is fired at time t , and s = 0 otherwise . All half-life parameters are converted to time constants using the formula in [8] . The Ca2+ variables , c and e , are incremented at a rate governed by Ca2+ entry ( Caent ) , with similar exponential decay: ( 2 ) ( 3 ) Ca2+ entry depends on spike broadening ( b ) , and is subject to submembrane and cytosolic Ca2+-dependent inactivation: ( 4 ) where parameter bbase gives a basal level for b . Entry is inhibited by c and e using two inverted Hill equations [29] with threshold and coefficient parameters , cθ , eθ , cn and en: ( 5 ) ( 6 ) The releasable vesicle pool ( p ) is depleted with secretion , x , and refilled at a rate proportional to the reserve store ( r ) , unless already full ( p = pmax ) . The refill rate is scaled by parameter β: ( 7 ) The reserve store ( r ) is depleted exponentially as it refills p , with its maximum ( and initial ) value defined by parameter rmax: ( 8 ) The rate of secretion ( vesicle exocytosis , x ) , is the product of the cube of the fast Ca2+ variable ( e ) and the releasable pool ( p ) : ( 9 ) We use a cube of e because Ca2+ activation of exocytosis is thought to be cooperative , and proportional to at least the square of the Ca2+ concentration close to the binding sites [30]; we found that using the cube gives a better fit to the in vitro data . Parameter α scales secretion to units that can be compared with experimental data . The final output of the model , vasopressin plasma concentration ( v ) , increases with secretion ( x ) and decays with half-life λv: ( 10 ) The parameter values for the figures in this paper are given in Tables 1 and 2 . The secretion model variables were initialised to 0 , except c = 0 . 03 , p = pmax , and r = rmax . The differential equations were integrated using the first order Euler method . We can do this safely since the step size ( 1 ms ) inherited from the spiking model is much smaller than any of the time constants in the secretion model . Using the same fixed time step makes it simple to couple the secretion model to the integrate-and-fire based spike model . The modelling software was developed in C++ , using the open source wxWidgets graphical interface library . A typical run of the full model , simulating 2000s of activity for a population of 100 neurons takes ∼20s on an Intel i7-2600K quad core processor . To generate a heterogeneous model cell population , we applied random variation to the rates of synaptic input received by each cell using the formula: ( 11 ) where Ire is the spiking model's input rate parameter , Ipop is the population input rate , and Isyn is a randomly generated value using a lognormal distribution ( see Results ) with mean = 0 and standard deviation = 0 . 5 , representing the synaptic connection density at a single neuron .
When the vasopressin nerve terminals of the posterior pituitary are stimulated electrically with brief pulses that trigger spikes in the axons , the amount of vasopressin secretion depends on the stimulation frequency: the secretion per stimulus pulse ( or per spike ) increases to a maximum at ∼15 Hz , beyond which it gently declines again . Our hypothesis is that this rise and fall is due to two competing activity-dependent mechanisms . At the beginning of a train of spikes , successive spikes broaden , and Ca2+ entry per spike increases . We fitted the spike broadening parameters , kb , bbase , and λb using the experimental data on spike broadening in [9] and [2] . A relatively slow decay is used , based on [31] which suggests that facilitation continues to increase over several seconds . The experimental data show a broadening of up to ∼100% , but this includes the Na+ current proportion of the spike , and so assuming that the broadening involves only the Ca2+ current ( as Ca2+ removal suggests ) , the proportional increase in the Ca2+ proportion of the spike , which we model , is larger . Thus the parameters bbase and kb are fitted to give a value for the Ca2+ component ( b + bbase ) ranging from 0 . 5 to ∼2 . 5 , illustrated in the b trace of Figure 2 . The competing inhibitory component ( modelled as Ca2+-dependent inactivation of Ca2+ entry ) uses both the fast and slow Ca2+ variables ( e and c ) to reduce Ca2+ entry per spike . The parameters for e and its inhibitory Hill equation were fitted to match the experimental data from [12] shown in Figure 3A . This frequency profile shows the sum of the effects of spike broadening and the competing inhibition , closely matched by the model in Figure 3B . These data were also used to adjust parameter α to scale the model's secretion units . Fatigue of the secretion response occurs over a timescale of tens of seconds , and recovery occurs on a similar timescale after stimulation is ended . The cytosolic Ca2+ concentration changes on a similar timescale , rising during stimulation , and decaying with a half-life of ∼20s [2]; our hypothesis for the model is that fatigue is due to the accumulation of cytosolic Ca2+ , which activates a Ca2+-dependent K+ current that hyperpolarises the terminals , leading to an increased rate of spike failure and reduced Ca2+ entry . The model parameters were fitted using the experimental data in [14] . We analysed the model secretion data by summing the secretion rate ( x ) values over each of four 18-s steps , to give a measure analogous to the cumulative secretion that was measured experimentally . We set λc , the half-life for the cytosolic Ca2+ variable c , to 20s , and manually fit parameters kc , θc and cn to match the fatigue profile in [14] , as measured in response to stimulation with regular 13 Hz spikes ( Figure 3c ) . The weakest fit of the model is to the 54–72 s period . This data point comes from a much lower n value than the others and so is less precisely determined , although margins of error are not provided because of the complex analytical approach that was used [14] . We thus consider that , as far as we can judge , the model is consistent with all available data , given the likely margins of error in those data . The detailed secretion rate profile is shown in Figure 3D . The core experimental result is that phasic stimulation produces more vasopressin secretion for a given mean spike rate than regular stimulation [15]–[17] . In these experiments , isolated pituitary glands were stimulated phasically using stimulus patterns generated using recordings taken from vasopressin neurons in vivo . Comparing a single phasic burst and regular stimulation at the same mean frequency [17] shows a more rapid decline in the rate of secretion using the burst pattern , tested here using spike model generated data ( Figure 4 ) . The faster spiking at the head of the burst causes an increased initial rate of secretion , followed by a steep decline , due mostly to the variation in spike rate and facilitation effect , rather than increased fatigue . In [15] the secretion response to regular and burst stimulation was further tested using a range of mean frequencies , and producing the equivalent data with the model shows a close match to the in vitro results . In particular , for both experimental data ( Figure 5A ) and model data ( Figure 5B ) , secretion is maximal in response to phasic stimulation at a mean frequency of ∼6 Hz . Similar levels of secretion can be achieved with regular stimulation , but only with much higher frequencies . In response to stimulation frequencies within the physiological range of mean firing rates for vasopressin neurons ( up to ∼8 Hz ) , phasic stimulation always releases more vasopressin than regular stimulation . We also used the model to compare non-phasic spike patterning ( by setting gL = 0 in the spiking model ) , simulating a continuously spiking neuron . The non-phasic pattern is more efficient than regular stimulation , but still much less efficient than the phasic pattern , as observed experimentally [16] . In the spiking model , we studied the effects of increasing synaptic input rate , and showed that , compared to otherwise identical non-phasic model cells , phasic cells respond to increases in input with a much more linear increase in mean spike rate . To extend this analysis to incorporate the secretory response , we coupled the spiking model to the secretion model . As previously , to investigate the adaptive consequences of the phasic firing pattern , we compared a representative phasic model cell ( Figure 6A ) with a cell that was identical except in lacking the mechanisms that allow the generation of phasic firing , by setting spiking model parameter gL = 0 . We simulated the plasma vasopressin concentration resulting from stimulation of the model phasic cell for 3000s at a range of fixed mean input frequencies ( using 25-Hz steps from 50 Hz to 1000 Hz ) by assuming a half-life of 2 min for the evoked vasopressin secretion [23] . This is a much higher rate than observed for these neurons in vitro , but the in vitro preparations are largely deafferented making this comparison unreliable; in vivo intracellular recordings from vasopressin neurons show a much higher rate of input than in vitro recordings ( and a lower specific impedance ) [32] , but the input rates in vivo have not been quantified . The density of GABA synapses in the supraoptic nucleus is between 14 and 27×106/mm3 of tissue , comprising about half of all synapses [33] . The total number of synapses per supraoptic neuron has been estimated to be ∼600 [34]; this is likely to be an underestimate , as the dendrites of magnocellular neurons extend well beyond the boundaries of the nucleus , and because it was based on an estimated neuronal population of the supraoptic nucleus that is much higher than subsequent estimates . For 600 synapses/neuron , a mean input rate of 1000 Hz represents an average input rate at each synapse of ∼1 . 5 Hz . The resulting frequency-response profiles for the phasic and non-phasic cell models ( parameters in Table 1 ) are shown in Figure 6B–E . The non-phasic model cell respond to increasing input with a non-linear increase in spike output ( Figure 6B left ) which is roughly mirrored by secretion ( Figure 6C left ) . In the phasic model however , the secretion output ( Figure 6C right ) is much more non-linear than the spike response ( Figure 6B right ) . Thus any possible benefits of the phasic firing pattern in linearising the cell response appear to be dissipated , and the faster firing of non-phasic cells compensates for the less efficient secretion response . Indeed , the non-phasic model appears to have a better secretion output , on the basis of dynamic range ( quartiles in Figure 6C ) , although examining secretion against mean spike rate does show a much more linear response with phasic cells in the 0–4 Hz range ( Figure 6D ) . The phasic cell response at faster spike rates becomes less linear , as bursts lengthen and secretion per spike becomes less efficient ( Figure 6E ) . One of the distinctive features of the vasopressin cell population is that it is highly heterogeneous in its spiking activity [35] , [36] . During osmotic challenge , the proportion of phasically firing cells increases , but some cells still fire slowly and irregularly , while others fire more quickly in a non-phasic continuous firing pattern . Previously with the spiking model , we simulated a heterogeneous population of vasopressin cells by varying a subset of the parameters related to the phasic firing mechanism [8] . However , these variations are insufficient to reproduce the observed range of firing behaviours , and here we tested a simpler and more powerful method . Instead of varying the intrinsic properties of cells , we varied the density of synaptic input received by each cell to generate a heterogeneous population of 100 model cells , using lognormal randomly generated values . This produces a distribution of firing rates and patterns that matches closely the heterogeneity observed experimentally amongst vasopressin neurons recorded in vivo ( [36] , Figure 7A ) . A survey [36] of 83 recorded phasic cells showed a mean spike rate of 4 . 2 Hz ( range 0 . 9 Hz to 10 . 7 Hz ) with standard deviation 1 . 8 . For Ipop = 460 Hz , 79 of the model population of 100 cells were fully phasic ( defined as having an activity quotient [37] between 0 . 1 to 0 . 9 ) . These phasic cells had a mean spike rate of 4 . 2 Hz ( range 1 . 1 Hz to 10 . 4 Hz ) with standard deviation 2 . 0 , very close to the population described from in vivo experiments ( Figure 7 ) . The secretion output from these 100 cells was summed and scaled to match a full size cell population , for comparison with in vivo data . We compared secretion from four model populations: homogeneous and heterogeneous non-phasic cells , and homogeneous and heterogeneous phasic cells ( Figure 8 ) . The homogeneous populations comprised 100 identical cell models each with the same synaptic input density , differing only in their Poisson randomly timed synaptic input events . Both the homogeneous and heterogeneous non-phasic populations show a similar non-linear mean spiking and summed secretion response ( Figure 8A and C ) ) to a single non-phasic neuron ( Figure 6B and C left ) . Similarly , the homogeneous phasic population shows just a smoothed version of the single phasic cell response ( Figure 6B , C right ) , with a mostly linear increase in spike rate ( Figure 8B ) and a much more non-linear secretion response ( Figure 8D ) . However , the input rate variation of the heterogeneous population has a strong effect on the spiking response to increased input , but only for the phasic cells . For phasic cells , heterogeneity produces a further linearization of the already more linear spike response ( Figure 8B ) . The secretion response from non-phasic cells is modestly linearised by introducing heterogeneity – but the secretory response from phasic cells is markedly linearised and the dynamic range considerably increased . Thus , population input heterogeneity has a strong linearising effect , but only when combined with phasic firing . Comparing the heterogeneous populations of phasic and non-phasic neurons , there is little difference in the linearity of the relationships between synaptic input and secretion ( Figure 8E and F ) . Comparing the overall model output to the experimental data of [6] shows that the model , over the physiological range of secretion , matches the linearity of secretion in response to increased stimulation ( Figure 9 ) . The other challenge for the vasopressin neurons is to sustain their response during prolonged osmotic challenge , over hours and days . They must balance immediate response to osmotic challenge with preserving vasopressin stores for as long as possible , to maintain life . Depleted vasopressin stores combined with lack of access to water rapidly lead to fatal dehydration . Vesicle stores are represented in the model as a reserve store ( r ) and a releasable pool ( p ) . The normal vasopressin content of the rat pituitary is approximately 1 µg [23] . However , to study the effects of depletion in the model on a timescale where we can still observe the response to individual spike bursts , we set the initial content to a very low value ( rmax = 100000 , equivalent to a content of 0 . 1 µg ) and simulated stimulation in 72-s bursts separated by 30-s silences , with regular intraburst activity at 13 Hz ( Figure 10 ) . The content of p ( pmax = 5000 ) , which is much smaller than the reserve store , was set large enough to support secretion during a typical burst with only partial depletion ( Figure 10 ) but also small enough to be fully replenished during a 30-s silence . We compared the model with and without fatigue , removing the effect of c by setting cinhib = 1 . In the model , p is replenished at a rate proportional to r . While this rate is greater than x ( the rate of secretion ) p is maintained near its maximum , and the relation between spike rate and secretion is maintained . Only when r is sufficiently depleted ( ∼50% ) does the secretion response decline . Thus , it seems that fatigue helps to maintain a consistent response to stimulation in the face of progressive store depletion . We then tested this with a simulation setting smax = 1000000 ( 1 µg ) and using spike model generated activity instead of regular stimulation . The spike model input rate was set at 600 Hz , producing a mean spike rate of 4 . 7 Hz ( at the high end of the physiological range ) . Figure 11 shows the model output , with and without fatigue , for 24 h of simulated stimulation . With fatigue , secretion is maintained at a fairly constant mean level for about 6 h , before declining progressively . Without fatigue , secretion declines from the beginning . Thus , in the model , the presence of a fatigue mechanism enables a consistent response to stimulation to be maintained until the stores have reached about 50% depletion . In a heterogeneous population , individual cells will become depleted at different times . To study this , we used the same heterogeneous population as used in Figures 8 and 9 , and simulated stimulation for 24 h with a population input rate of 400 Hz ( Figure 12A ) . The slowest cells maintain their response for the whole 24 h , but most active cells become depleted , and their responses become proportional to their respective reserve store levels well within 24 h ( Figure 12B ) . However , although the decline in the overall secretion of the population ( Figure 12C ) begins at a time determined by the most active cells , heterogeneity in the input rates results in heterogeneity in depletion rates , which reduces the rate of the decline in the population signal . We know from experimental data [38] that , when tested over up to 4 days of stimulation , vasopressin secretion declines in proportion to pituitary content , but there is currently not enough data to make a more detailed quantitative comparison with the model .
The present secretion model implements mechanistic elements derived from previous experimental studies . The rate of vesicle exocytosis depends on Ca2+ entry , and the model relates Ca2+ entry at the axonal terminals to spike activity . Facilitation of stimulus-secretion coupling arises as a consequence of the spike broadening that has been observed in recordings from the terminals . We assumed that most of the increase in spike duration reflects a high voltage-gated Ca2+ conductance , since the effect is blocked when Ca2+ is removed [9] . This results in a frequency-dependent increase in Ca2+ entry per spike . In vasopressin cells facilitation of stimulus-secretion coupling peaks at ∼15 Hz . By contrast , oxytocin cells continue to show increasing facilitation at spike rates as high as 50 Hz , and this suggests that vasopressin cells have some other competing inhibitory mechanism . To fit the measured frequency-response profile [12] , we added a Ca2+-dependent inactivation of Ca2+ entry . We modelled fatigue of stimulus-secretion coupling as the effect of a Ca2+-activated K+ current that reduces the probability that axonal spikes invade the terminals . The close match between intracellular Ca2+ time course and the development and recovery from fatigue [14] , [17] , the existence of suitable Ca2+ activated K+ channels , and activity-dependent spike failure [27] , all make this mechanism plausible , and we chose model parameters to match experimental data quantitatively . The data of [17] suggests a more rapid development of fatigue than [14] . The model is capable of fitting both , but for the results here we use parameters that fit our own data [14] . In both cases , the time resolution is limited , and ideally we would repeat the experiments , using spiking model generated data . We also experimented with an extended mechanism incorporating ATP- and adenosine-based mechanisms [40] , that are thought to modulate the Ca2+ activated K+ channel and an N-type Ca2+ channel respectively , but the improvement in fit was not sufficient to justify the increased complexity . Although the model doesn't use a releasable store to simulate fatigue , it does use a store mechanism to simulate the depletion that occurs on longer timescales , with the model tested for durations up to 24 h . The secretion of vesicles from the posterior pituitary has been studied quantitatively at the ultrastructural level [41]; these studies led to the conclusion that exocytosis can occur from all parts of the axons of magnocellular neurons – from nerve terminals , from the large axonal swellings that are a conspicuous feature of the gland and which contain most of the vesicles , and even from undilated axons . However in each of these compartments , and under different physiological conditions including different states of depletion , the number of vesicles that is released by a given stimulus appears to depend only on the number of vesicles that are close to the plasma membrane [41] . This implies that in any secretory compartment , exocytosis is a stochastic process , and that the probability of release from a compartment depends on the content of a releasable pool in that compartment . On long timescales ( tested over 4 days ) the rate of secretion is proportional to the total gland content [38] , suggesting that replenishment of the releasable pool ( close to the plasma membrane ) from a deeper store is also activity-dependent and probabilistic . In normal conditions the system functions to maintain body fluid homeostasis , and secretion is linearly proportional to osmotic pressure . We assume that it is also important to maintain a consistent response , and therefore an advantage to make secretion independent of the store content . However , in conditions of prolonged challenge it will be more important to preserve vasopressin stores , so that osmotic pressure can be maintained at a life preserving level for as long as possible , and so it would be of use to gradually reduce the secretion rate , as rationing becomes more urgent . In modelling the store and the releasable pool , we assumed that secretion is proportional to the releasable pool , and that it is refilled at a rate proportional to the store content . We did not include any synthesis in the model , as the long axons mean that there is a long delay ( 24 h ) between an increase in demand and an increase in the arrival rate of newly synthesised vesicles at the terminals [42] . In the present model , the releasable pool represents the summed total of all the terminals of a single cell . It was set large enough that a single burst would release only a small proportion of the content , with a refill rate parameter fast enough to refill the pool between bursts while the store content is at maximum . Under these conditions , we get a model cell that , in response to a large and sustained increase in activity ( comparable to that which accompanies systemic dehydration ) can maintain a stable secretion rate for a long period ( hours ) . However , once the store reaches a critical level of depletion , the releasable pool also becomes depleted , and thereafter the secretion rate becomes proportional to store content . These simulations give a strong indication of the likely adaptive value of fatigue . From experimental data we know that fatigue of stimulus-secretion coupling is a marked characteristic of vasopressin secretion but not of oxytocin secretion – so it is not an inevitable feature of stimulus-secretion coupling in neuroendocrine neurons . The present model results suggest that in conditions of chronically maintained stimulation , fatigue is important for maintaining a constant level of output despite declining stores . This may be important physiologically for vasopressin secretion , where the absolute level of vasopressin determines the degree of antidiuresis , but less important for oxytocin in its primary role as a hormone that mediates milk let-down – a reflex governed not by the mean sustained level of oxytocin secretion but by the frequency at which transient pulses of secretion occur [43] . Combining the new secretion model with our previously published spiking model allows us to examine the relationships between synaptic input , spiking activity , and secretion . As we showed previously , because the phasic activity of vasopressin cells is asynchronous , and because of the properties of the mechanism that generates phasic activity , the average spike rate of the population increases relatively linearly in response to increased afferent input despite the short term non-linearity of individual cell responses . However , as we show here , the combined effects of fatigue and facilitation make the secretory response to increasing input one that is highly non-linear . This is an obvious consequence of facilitation , which increases secretion per spike as spike rate increases . However , fatigue also makes the response less linear , by capping the secretion rate for a proportion of the spike activity - a proportion that increases as burst durations increase . The effect of this overall nonlinearity is to reduce the dynamic range of the secretory response compared to that of the spiking response . However , to properly understanding the relationship between the underlying mechanisms and the function of the cells , we need to look not at individual cells , but at how they behave as a population . To do this we used populations of 100 model cells to look at the relation between synaptic input ( assumed to be representative of osmotic pressure ) and summed population secretion , scaled to compare against full size cell populations ( ∼9000 cells ) . Testing a homogeneous population of 100 identical model cells has the expected smoothing effect of an averaged population signal , but otherwise matches the features of the single cell response . However , vasopressin neurons are a highly heterogeneous population , varying both in their firing rates , and spike patterning . Previously , with the spiking model , we simulated heterogeneity by adding random variation to a subset of the parameters which had been used to fit varied recorded cells . This is sufficient to produce a range of phasic firing patterns , but not the highly varied activity that is observed in vivo , and this is likely to underestimate population heterogeneity since identified cells are biased towards the more distinctive phasic pattern . To produce more varied spike rates we used a simpler method , applying random variation to the rates of synaptic input received by each model cell , using a log normal distribution ( Figure 8A ) . Tests using a heterogeneous population using the subset method showed only minor variation from the homogeneous response profile ( not shown ) . However , tests with input based heterogeneity show a marked linearisation of the population response , along with a large increase in the dynamic range , producing a profile which is much closer to the highly linear secretion response observed in vivo ( Figure 9 ) . The effect only works using the phasic cell model , with non-phasic cells showing very little difference between homogeneous and heterogeneous populations . The vasopressin neurons of the supraoptic nucleus are probably as homogeneous a population of neurons as any that exists within the CNS: there are about 2000 of these in each nucleus , and each has just a single axon that projects to the posterior pituitary gland with few if any collateral branches that terminate in the CNS . Thus they share a strictly common function in that the role of each is to contribute to the pooled output of vasopressin that enters the systemic circulation . While these neurons have many features in common , there is also considerable variability in their intrinsic properties , and their spontaneous firing rates and patterns vary substantially from cell to cell – even between immediately adjacent cells recorded simultaneously . The most parsimonious explanation of this variability is that it is not “designed in” , but is the consequence of intrinsic variability in the developmental determination of gene expression and neuronal wiring patterns . But however this variability arose , the question arises of whether it has any adaptive significance: does this variability affect information coding , and does it do so in a way that yields any apparent advantages ? This question has been addressed previously for several other neuronal systems , leading to the conclusion that there can indeed be various advantages of neural heterogeneity in information coding [44]–[46] . However , these previous studies all drew their conclusions from models that assumed that information is coded in the spike activity of neurons . However the information that is transmitted from neuron to neuron is mediated not by spike activity itself , but by the secretion that is evoked by that activity – and in all neurons the coupling between activity and neurosecretion is complex and highly nonlinear . Here we show that in the case of vasopressin cells at least , taking account of stimulus-secretion coupling radically changes our perception of the importance of heterogeneity . Here , we have modelled closely the particular properties of stimulus-secretion coupling in vasopressin cells . These properties vary in different neuronal types , but it seems that frequency facilitation is a common if not universal feature of neurotransmitter secretion from synapses , and that activity-dependent depletion and other mechanisms functionally equivalent to fatigue are also common if not universal . We have shown here that the functional properties of a network critically depend upon these properties . The output of a neuron is not its spike activity , but what it secretes in response to that activity , and representing this accurately in network models may be critical to properly understanding how they process information . | Vasopressin is a hormone that is secreted from specialised brain cells into the bloodstream; it acts at the kidneys to control water excretion , and thereby help to maintain a stable ‘osmotic pressure’ . Specialised cells in the brain sense osmotic pressure , and generate electrical signals which the thousands of vasopressin neurons process and respond to by producing and secreting vasopressin . In response to these signals , vasopressin neurons generate complex “phasic” patterns of electrical activity , and this activity leads to vasopressin secretion in a complex way that depends on both the rate and pattern of this activity . We have now built a computational model that describes both how the vasopressin neurons generate electrical activity and also how that activity leads to secretion . The model , which gives a very close fit to experimental data , allows us to explore the adaptive advantages of particular features of the vasopressin neurons . This analysis reveals the importance of heterogeneity in the properties of vasopressin neurons , and shows how the vasopressin system is optimally designed to maintain a consistent hormonal output in conditions where its stores of releasable hormone are severely depleted . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"neurophysiology",
"homeostatic",
"mechanisms",
"computational",
"neuroscience",
"single",
"neuron",
"function",
"endocrine",
"system",
"biology",
"anatomy",
"and",
"physiology",
"neuroendocrinology",
"neuroscience",
"computational",
"biology",
"coding",
"mechanisms",
"endocr... | 2013 | Spike Triggered Hormone Secretion in Vasopressin Cells; a Model Investigation of Mechanism and Heterogeneous Population Function |
Complex traits such as obesity are manifestations of intricate interactions of multiple genetic factors . However , such relationships are difficult to identify . Thanks to the recent advance in high-throughput technology , a large amount of data has been collected for various complex traits , including obesity . These data often measure different biological aspects of the traits of interest , including genotypic variations at the DNA level and gene expression alterations at the RNA level . Integration of such heterogeneous data provides promising opportunities to understand the genetic components and possibly genetic architecture of complex traits . In this paper , we propose a machine learning based method , module-guided Random Forests ( mgRF ) , to integrate genotypic and gene expression data to investigate genetic factors and molecular mechanism underlying complex traits . mgRF is an augmented Random Forests method enhanced by a network analysis for identifying multiple correlated variables of different types . We applied mgRF to genetic markers and gene expression data from a cohort of F2 female mouse intercross . mgRF outperformed several existing methods in our extensive comparison . Our new approach has an improved performance when combining both genotypic and gene expression data compared to using either one of the two types of data alone . The resulting predictive variables identified by mgRF provide information of perturbed pathways that are related to body weight . More importantly , the results uncovered intricate interactions among genetic markers and genes that have been overlooked if only one type of data was examined . Our results shed light on genetic mechanisms of obesity and our approach provides a promising complementary framework to the “genetics of gene expression” analysis for integrating genotypic and gene expression information for analyzing complex traits .
Most complex traits such as obesity involve a diverse set of genes , intricate interplay among them and subtle interaction between genetic and environment factors . One of the first steps toward a systematic understanding of the genetic basis of a complex trait is the identification of causal genetic elements , e . g . genes , genetic markers and/or single nucleotide polymorphisms ( SNPs ) , whose variations are responsible for the traits . The objective of this challenging task is two-fold: effectively identifying a subset of genetic elements out of a large pool of candidates whose patterns are characteristic of a trait of interest , and accurately predicting the phenotype with a model that accommodate interactions among selected genetic elements . Despite recent advances in high-throughput technologies that have produced an enormous amount of biological data , heterogeneous data types , non-linear relationships among genes and complex phenotypes have made this task difficult . Although conventional linkage analyses and association studies as well as the latest genome-wide association studies ( GWAS ) have produced a fruitful collection of genomic susceptibility loci for a variety of complex traits and diseases [1] , [2] , they have mainly been able to detect genetic elements of marginal effects while failed to respect epistatic interactions [3] , [4]; as a result , they have a low power for predicting phenotypes [5] . As an intermediate between genotype and phenotype , gene expression has been proven to be a rich and valuable source of information complementary to genotype information for dissecting complex traits . On one extreme using gene expression data alone , classifiers or regressors have been built to predict disease types or stages with only a small number of disease-related genes [6]–[8] . By integrating information of genetics and gene expression , genetics of gene expression-based approaches [9]–[11] and network-based approaches [12]–[14] have been independently developed and applied to identify genes related to complex traits . Recently a few machine learning based methods have been proposed to integrate both genotype and gene expression data to not only identify relevant genes , but also predict phenotypes based on selected genes . Ruderfer et al . [15] adopted a SVM classifier to predict drug responses ( i . e . , sensitivity or resistance ) in yeast . They showed that using both data of transcripts and genetic markers can improve prediction accuracy compared with using either transcripts or genetic markers alone . Based on the elastic net regularized regression [16] , Chen et al . [17] developed Camelot to predict quantitative response ( i . e . growth yield ) of yeast to 94 drugs using genotype and gene expression data collected from drug-free conditions from yeast segregants . Compared with the work by Ruderfer et al . [15] , Camelot was able to make accurate quantitative prediction on various drug treatments as opposed to dichotomic classes of drug response . Camelot also emphasized greatly on causal inference by incorporating a priori knowledge and adopting post statistic tests to select only handful genetic makers and expression transcripts as phenotype predictors . For example , for predicting the hydrogen peroxide response , only a single gene , DHH1 , passed their pre-filtering criteria and was then used to construct the final prediction model . Although appropriate for downstream experimental validation as Camelot always make the most conservative choices , it remains unknown whether its filtering steps could indeed help improve prediction accuracy and whether it would otherwise prevent further novel discovery besides the factors known to have large marginal effects . Random Forests ( RF ) [18] , an ensemble of classification or regression trees , has recently been successfully applied in various biological studies [19]–[23] . RF has many desirable characteristics that make it well suited for integrating both genotypic and gene expression information . It is well adapted for variable selection for high-dimensional data with competing prediction accuracy compared to the state-of-the-art machine learning techniques . RF is able to accommodate categorical ( e . g . genotype ) and continuous ( e . g . gene expression ) data . It can be used when the number of variables substantially exceeds the number of observations ( e . g . thousands of SNP markers and probes of gene expression versus a few hundred samples of subject ) [19]–[23] . Moreover , RF supports possible interactions among variables [4] , which is critical for systems-biology studies where interplays between genetic ( e . g . epistatically interacting SNPs ) and gene expression ( e . g . coactivator/corepressor ) must be taken into consideration . While promising , however , conventional RF algorithms have several drawbacks that limit their success on large biological problems . Firstly , even though RF allows possible interactions among variables , it does not incorporate possible correlation among variables; even worse , with correlated variables , it suffers from biases introduced in measuring variable importance ( VI ) [24] , [25] , which can result in incorrect or misleading variable rankings . Secondly , RF's prediction accuracy may decline significantly when the proportion of truly informative variables among all variables is small [26] . In this paper , we develop a new method , called module-guided Random Forests ( mgRF ) , to integrate genotypic and gene expression information to understand and possibly dissect complex relationships among different genetic elements underlying complex traits . mgRF combines the method of conventional RF and a network-based analysis to remedy the two aforementioned drawbacks of conventional RF by exploiting structural relationships , extracted from the network analysis , among different types of variables . As a test and application , we applied mgRF to the data of genetic markers and gene expression from a cohort of F2 female mouse intercross to examine its performance and demonstrate its ability to identify genetic elements that contribute to mouse weight , many of which were missed by the conventional RF algorithm . mgRF outperformed the state-of-the-art methods that combine information from multiple biological sources with more accurate predictions . Furthermore , using mgRF we investigated the interactions among multiple genetic elements underlying mouse weight . Statistically significant interactions of SNP-to-SNP , gene-to-gene , and SNP-to-gene identified by mgRF revealed genetic elements and their significant association underlying mouse weight . The results demonstrated a great expectation of mgRF as a complementary framework to “genetics of gene expression” analysis for dissecting genetic mechanism of complex traits , such as obesity .
In mgRF our main objectives are to capture intrinsic structures of variable ( genetic element ) correlation and/or interaction and to incorporate such information in the RF framework to predict a complex phenotype . The major steps of mgRF algorithm , outlined in Figure 1 , consist of the identification of variable modules from a variable correlation network ( Figures 1A and 1B ) and an iterative RFs construction process ( Figure 1C ) . In the first step we construct a correlation network and identify modules in the network to group highly-correlated variables , which may be in different types , using a network clustering method such as HQCut [27]–[29] . In the second step of mgRF , we iteratively construct a series of RFs guided by the previously identified network modules . Instead of randomly sampling variables in each node of regression tree , we adopt a two-stage candidate variable sampling procedure , where we first select a subset of modules and then choose one representative variable for each of the selected modules ( right panels in Figure 1C ) to correct the bias of variable importance while incorporating the variable association information . Except the first RF construction , we use a modified weighted sampling to improve the prediction accuracy by prioritizing informative variables among a large pool of variables . A key element of mgRF is to correct possible bias of variable importance measure and improve the performance of RF for high-dimensional data . This is done in part by introducing a module importance ( MI ) to each network module identified . Initially all MI and variable importance ( VI ) are set to 0 so that in the first iteration the sampling of modules and variables is un-weighted . After each iteration of RF , new MIs and VIs are re-estimated ( not accumulated ) . The values of MIs and VIs can typically converge within a small number of iterations , where little change can be observed between the last and the second to the last estimations . The final output of mgRF is an ensemble of trees as a model for future analysis and its corrected VIs ( cVIs ) and MIs for variable and module ranking . Details and parameter selections of the mgRF algorithm are described in Materials and Methods . To investigate the benefits of integrating genotypic and gene expression data , we examined the performance of different models on the genotypic and gene expression data of a cohort of F2 mouse intercross in a three-way comparison: ( 1 ) using only data of genetic markers ( genotype-only ) , ( 2 ) using only data of gene expression ( expression-only ) , and ( 3 ) using both genotypic and expression data ( combined ) . We first compared mgRF with group lasso [30] , elastic net [16] , SVR-RFE [8] , and the conventional RF algorithm ( see Text S1 ) in terms of the weight regression error ( Root-Mean-Square Error or RMSE ) in all three types of data using 10 trials of 10-fold cross-validation . The average cross-validation RMSEs of the methods compared are shown in Figure 2 . mgRF achieved the smallest average error compared to all the other competing methods in all types of data . Since we assessed each model using the same training and test data in each fold of the cross-validation , we can compute the paired t-test of RMSEs to evaluate the significance of the results . As shown in Table S1 , the RMSEs of all the other methods are significantly larger ( p<2 . 52×10−13 ) than that of mgRF . Furthermore , the running time of mgRF is slightly less than the conventional RF with better prediction accuracy ( Table S2 ) . It is noteworthy to mention that SVR-RFE , RF and mgRF outperformed the linear models , group lasso and elastic net , suggesting the benefits of incorporating non-linearity between variables and the mouse weight response . In particular , we examined the RMSEs of mouse weight using mgRF . The boxplot of prediction errors on the three data types are shown in Figure 3A . The combined data have the smallest error ( RMSE = 3 . 80 and R2 = 0 . 604 ) , followed by the expression-only data ( RMSE = 4 . 13 , and R2 = 0 . 534 ) , while the genotypic-only data have the highest error rate ( RMSE = 5 . 62 and R2 = 0 . 137 ) . Thus using either the genotypic or gene expression data alone is less effective than using the combined data . Although the standard error of RMSEs of different trials ( quartile bar in Figure 3A ) is relatively large comparing to the difference of mean RMSE between with and without genotypic data , there is a substantial improvement in pairwise comparisons using the same training samples ( Figures 3B to 3D ) . The two-dimensional co-ordinates of point in each of these plots indicate the RMSEs of mgRF trained with the same set of samples but with different data types . In Figures 3B and 3C , most of the points appear under the reference diagonal line , which confirms that both expression-only and combined data achieved better performance in a single fold than the genotype-only data ( paired one-tail t-test p-value≤4 . 7446×10−28 and ≤1 . 7272×10−32 , respectively ) . This was probably because in general the linkage signal of genetic markers is weak ( LOD score <4 ) , while gene expression is more closely related to the phenotype than genotypes . Furthermore , mgRF using both genotypic and gene expression data outperforms using expression-only data in more than 90% of the trials ( Figure 3D , paired one-tail t-test p-value≤1 . 691×10−19 ) , showing that combining genotypic and gene expression data can indeed improve the prediction power and suggesting that information of gene expression plays a role in bridging the gap between genotype variations and complex traits . The mgRF method used corrected variable importance ( cVI ) and module importance ( MI ) to identify variables and groups of variables that influence the trait of body weight . MIs were computed for network modules identified by the HQcut algorithm [27] , [29] , [31] . To assess and illustrate mgRF's ability for correcting the bias of variable importance , we evaluated different regression models regarding their abilities for recovering the true variable importance associated with the known data-generating pattern in a simulation study ( see Text S2 ) . mgRF was able to accurately recover the known pattern of variables' importance and the VI measure of mgRF was more stable than the other methods in all simulations , as discussed in Text S2 . When applied to the mouse weight data from a cohort of 132 samples and compared with modules identified by topological overlap matrix based methods [12] , [32] , HQcut produced much smaller modules , allowing only highly-correlated variables to be clustered in a module ( Figure S1 ) . HQcut identified 146 , 1036 and 1187 network modules ( see module structures in Table S3 , S4 , S5 ) in the genotype-only , expression-only and combined data , respectively . As expected , SNPs in one module were generally in linkage disequilibrium . Genes in one module were co-expressed and potentially functionally related . There were SNPs and genes assigned to the same module in the combined data set due to the large correlation values among those gene expression and SNPs . The top-ranked genetic markers and genes in the combined data largely overlapped with those identified by genotype-only and expression-only data types indicating the stability of mgRF in terms of variable ranking . Here we reported the top-ranked modules of genetic markers and genes in Table 1 and 2 . Among these top-ranked SNPs ( Table 1 ) , rs3662726 ( Chromosome 5 , 123 Mb ) is near Gofm2 ( gonadal fat mass 2 ) QTL which has been reported to confer increased fat mass in female mouse [33] . We also examined the LOD scores of SNPs using the traditional QTL mapping . Several “hotspot” QTLs on Chromosomes 1 , 3 , 5 , 7 , 10 , 15 and 19 were partially overlapped with the top-ranked markers by mgRF ( Figure 4 ) . Table S6 lists all the top-ranked SNPs . Note that several markers with low LOD scores were assigned relative high cVIs , suggesting that a SNP with low marginal effect can be identified by mgRF because of their interaction with other SNPs , which may contribute to the variation of body weight . It is important to note that there was little overlap among the 100 top-ranked genes from group lasso , elastic net , SVR-RFE , the conventional RF algorithm , and mgRF ( Figure S2A ) . The lack of consensus indicated that these algorithms identified their own top-ranked genes based on different ( unspecified ) assumptions on the given data and target models to be learned . Introduction of such assumptions seemed to be inevitable because of the lack of sufficient knowledge of the problem at hand and different objectives that these methods were devised to achieve . Nevertheless , all these methods strived to select predictive variables ( genetic factors ) . On top of finding individual predictive genetic factors , mgRF was particularly designed to identify such predictive genetic factors whose association might contribute more significantly than individual variables at the module level because it propagated the contribution of individual variables to highly correlated neighbors , rather than fully focusing on individual genes . Table S7 lists the top-ranked genes related to mouse weight from mgRF . Among these top-ranked genes ( Table 2 ) , monoacylglycerol O-acyltrasferase 1 ( Mogat1 , cVI = 11 . 84 ) in module 189 has been previously identified to be located within Chromosome 1 obesity QTL interval near D1Mit215 . Within this QTL interval on Chromosome 1 , insulin-like growth factor binding protein 2 ( Igfbp2 , cVI = 10 . 94 ) has expression levels in liver negatively correlated with mesenteric fat pad weights [34] . Igfbp2 ( appeared in module 32 ) has also been shown to prevent diet-induced obesity and insulin resistance in mice on overexpression [35] . In particular , module 32 contained Cyp2c37 ( cVI = 9 . 65 ) , C7orf24 ( cVI = 7 . 43 ) and Gpld1 ( cVI = 6 . 54 ) , which were not among the top 100 ranked genes from any of the other methods compared , probably due to their correlation with Igfbp2 . Raet1d ( cVI = 9 . 38 ) in module 189 was also not identified by the competing methods ( Table S8 ) probably due to its correlation with Mogat1 . Remarkably , Cyp2c37 has been previously recognized as being associated with fat mass [10] and Gpld1 had been shown to be associated with the level of adiponectin , a hormone secreted from adipose tissue which is negatively correlated with obesity [36] . It is viable to hypothesize that other genes identified by mgRF , which were neglected by the other methods , may potentially contribute to mouse weight variation . To further assess the biological significance of the genes identified by mgRF , we conducted a Gene Ontology ( GO ) enrichment analysis ( see Materials and Methods ) on the top-ranked genes from the methods that were compared . mgRF identified more enriched biological processes than the other methods ( Table S9 ) . In particular , genes identified by mgRF were enriched with many obesity-related processes , such as regulation of lipid storage ( p = 7 . 17×10−06 ) , positive regulation of cholesterol storage ( p = 1 . 07×10−05 ) , regulation of growth ( p = 0 . 000575 ) , and cellular response to cholesterol ( p = 0 . 00396 ) . In contrast , the enriched biological processes provided by the other methods were less significant and less biologically meaningful ( Tables S5B to S5E ) . As an example , Figure 5 shows a sub-network of some top-ranked genes from mgRF to compare the variable importance measures in mgRF and the conventional RF . The size of nodes represents relative cVIs from mgRF in Figure 5A and represents relative VIs from conventional RF in Figure 5B . In the conventional RF algorithm , one variable , Igfbp2 , has a larger importance than others . As a result , the importance of Igfbp2 overshadows several other correlated variables such as Fmo3 , Cyp2c37 , and Raet1d , which may in fact be equally important as Igfbp2 . In mgRF , several genes with the highest cVI , such as Mogat1 , MGC137458 , and Igfbp2 , which are known to be critical to body weight , were also hub nodes in the network with many edges . It was consistent with our previous studies on the importance of hub genes in the co-expression network [37] . Although the corrected variable importance ( cVI ) from mgRF quantifies the contribution of a genetic factor to the prediction power , it does not indicate whether the contribution is from the genetic factor alone or from its interaction between or association with other factors . One advantage of the RF method is its ability to incorporate variable interactions , which mgRF inherited . We devised a systematic statistical test ( see Materials and Methods ) to assess the significance of gene-to-gene , SNP-to-SNP , and SNP-to-gene interactions revealed by mgRF . To examine the biological relevance of genes identified in gene-to-gene interactions in the mouse weight data , we first tested the functional enrichment among 160 unique genes from the top 100 most significant pairs of interactions ( Table S10 ) . Interestingly , these genes were enriched with metabolic processes such as isoprenoid metabolic process ( p = 0 . 00675 ) , drug metabolic process ( p = 0 . 00841 ) , and terpenoid metabolic process ( p = 0 . 00117 ) , indicating obesity-related interaction among the identified genes . In particular , the pair of Avpr1a and Igfbp2 is one of the most significant interactions ( p = 4 . 15×10−07 ) , both of which are also among the most predictive genes . However , only 11 ( ∼7% ) of the 160 unique genes from the significant gene-to-gene interactions were overlapped with the top 100 most predictive genes identified by cVI ( Figure S2B ) . This suggested that our interaction test could indeed identify genes that were less significant when examined individually . For example , macrophage receptor with collagenous structure ( Marco ) was observed to interact with many other phenotype-related genes ( e . g . Dhrs4 , Cyp2d22 , and Pdia5 ) , even though its own cVI was relatively low . Among the top-ranked SNP-to-SNP interactions , we found significant interactions between SNPs on Chromosome 5 ( 123 Mb ) and Chromosome 19 ( 51 Mb ) ( p = 3 . 34×10−11 ) , on Chromosome 2 ( 96 Mb ) and Chromosome 9 ( 61 Mb ) ( p = 8 . 97×10−7 ) , and on Chromosome 1 ( 41 Mb ) and Chromosome 15 ( 62 Mb ) ( p = 2 . 94×10−06 , Table S11 ) . The most significant interaction was between p45558 ( Chromosome 5 , 123 Mb ) and p44890 ( Chromosome 19 , 51 Mb ) . Cis-eQTLs analysis [12] indicated that Bmp2 , a key regulator of adipogenesis , was a candidate gene of p45558 , and Cyp2c40 , known to be presented in Fatty acid metabolism , was a candidate gene of p44890 . For SNP-to-gene interactions ( Table S12 ) , there was little overlapping between genes involved in SNP-to-gene interactions and genes involved in gene-to-gene interactions ( Figure S2B ) . Of particular interest was the interaction between SNP p45334 ( Chromosome 1 , 77 Mb ) and gene Ehhadh , where Mogat1 is one of the candidate eQTL genes of p45334 and Ehhadh is annotated in the fatty acid metabolism pathway . We compared the top 50 unique SNPs involved in SNP-to-gene interactions with top SNPs ranked by cVIs . More than 20 of them were among the top-50 most predictive SNPs . On the other hand , only two genes involved in top 100 SNP-to-gene interactions were among the top 100 most predictive genes . We hypothesized that the most predictive SNPs did not interact with the most predictive genes because the information contained in these SNPs was redundant to these predictive genes , which usually were the expression traits of the corresponding predictive SNPs . In turn , by combining less predictive gene and marker profiles introduced extra information into the system and indeed improved the prediction accuracy . Genes involved in such interactions might reveal additional perturbed pathways underlying the trait of body weight .
A systems biology approach is necessary to dissect complex traits , such as obesity , and understand relationships among various genetic factors . Combing heterogeneous data from multiple sources will become increasingly important to model a large quantity of data and interpret results . In this paper , we proposed a novel approach that integrates the method of Random Forests and a network analysis to incorporate genotypic and gene expression data for revealing genetic factors and their interaction or association that are characteristic of complex traits . To overcome the curse of dimensionality , mgRF enhanced conventional RF with the module structure of a correlation network and the weighted sampling procedure . As a result , it successfully identified a small subset of both predictive and biological meaningful genes and genetic markers out of thousands of candidates . Meanwhile , mgRF was able to model the complex associations and possibly interactions between heterogeneous variables , which lead to interesting findings that can shed some lights on solving the genetic multiplicity problem underlying complex traits . To rectify the bias in ranking correlated variables in conventional RF , simple but effective strategies such as grouping correlated variables prior to model fitting [38] , [39] can be applied , where cluster centroids obtained from a hierarchical clustering could be used as supergenes to fit classification/regression models . Compared with Tolosi and Lengaue's work [39] , the major differences and novelty of mgRF are three folds . ( 1 ) We maintained the original feature space in RF models so that the importance of individual variables can be estimated . ( 2 ) Instead of using a simple hierarchical clustering , we adopted the network modeling method HQCut , which is able to automatically and accurately determine the number of modules ( clusters ) in the network . ( 3 ) We further utilized the learnt MI and VI to guide a weighted sampling of variables . Compared with other regression models , such as elastic net and support vector regressor ( SVR ) , mgRF naturally handles different types of variables in that the splitting points of continuous ( or ordinal ) variables preserve the order information , which , however , is disregarded in categorical variables . On the contrary , elastic net and SVR treat categorical variables as continuous ones , consequently imposing ordered information , which is related to how the categories should be encoded . There is a popular variable importance measure , permutation VI [18] , which measures the increase of out-of-bag prediction error with the variable to be measured being permutated . However , the permutation VI suffers from several shortcomings for large problems . It requires an excessive computation time , since each variable needs to be permuted dozens of times to ensure statistical stability . In addition , when the baseline prediction error is large , there is little chance for permutation to make a prediction worse , which leads to an uniformly low VI [25] . More critically , it is still subject to the bias of correlated variables [40] , [41] . Even though this problem can be corrected [24] , [42] , [43] , the incurred computation time of additional permutation for a solution will make the excessive computation cost prohibitive for large application . Using a BxH F2 mouse intercross data set , we showed that the proposed algorithm was effective on not only reducing prediction error , but also identifying a subset of genetic markers and genes that are associated with the trait of body weight . By integrating genotypic and gene expression data , mgRF achieved a lower prediction error compared to using either type of data alone . These results support the idea that gene expression plays an intermediate bridging role between genotypic variations and a phenotype . Genotypic data alone are insufficient for accurately predicting the body weight due to their relative weak effects , while gene expression data bridge the gap between genotypic variants and a phenotype as gene expression can be intermediate traits of multiple genetic markers . Besides the annotated body weight relevant genetic elements , such as QTL rs3662726 , genes Mogat1 , Igfbp2 , and Cyp2c37 , mgRF provided valuable hypotheses on putative , novel genetic elements and their interactions that are potentially important for body weight and obesity . In particular , the top-ranked SNPs and genes , which have similar levels of importance but are lack of known annotations , are excellent candidates for further validations . A key feature of mgRF is that it exploited splitting variables to incorporate non-linear interactions of variables into the model and to identify intriguing associations within and across two types of data . The proposed statistical test for variable associations aimed at extracting biological relevant markers and genes that might have been overlooked by individual variable importance ranking . The results of mgRF showed that several known obesity-related genes and loci were associate or even interacted with each other and genes that were strongly associated were indeed related to the traits of obesity and/or body weight , as these genes were enriched with biological processes on metabolisms . More importantly , the results revealed that many genetic elements , which have not been indicated previously to be associated with the traits , interacted with obesity-related genes and their associations may contribute more significantly to the traits than associations between genes that were known to be related to obesity . In addition , the results also included significant pairs of genes even though the predictive scores of individual genes whose predictive scores were insignificant . These results suggested that more obesity related genetic factors remain to be discovered and mgRF is potentially an enabling method for identifying genetic factors whose significance would not be appreciated unless their associations or interactions were taken into consideration . Another key advantage of RF is that at each splitting point , it only considers mtry ( k ) candidate variables for splitting , usually k<<m , where m is the number of variables . The time complexity of mgRF is the same as the conventional RF , in the order of . In practice , mgRF is usually more efficient than the conventional RF , thanks to the two-stage candidate variable sampling and the modified weighted sampling . In particular , the mtry ( k ) of mgRF is proportional to the number of modules instead of the actual number of variables . In our experiments , the average training time for mgRF was below 3 minutes for our C++ implementation on a desktop machine with an Intel Duo core 2 . 53 GHz CPU and 4 G memory ( see Table S2 for running time comparison ) . The most time consuming part of the mgRF framework is the network construction and module finding using HQCut , which took around 20 minutes on the same machine . For larger gene expression data , a common practice is to pre-filtering low variance genes to ∼10 , 000 most varying genes . For large SNPs data , to reduce the time of network construction and clustering , LD-pruning can be utilized to approximate the SNPs' module structure . Compared with conventional “genetics of gene expression” analysis , mgRF provided a complementary means to incorporate the knowledge of inherent structure of genetic elements . mgRF can also readily be applied to more than two types of data and can be efficient on large-scale applications as it can be easily parallelized to utilize the growing cloud computing environments . While mgRF includes statistical tests to identify significant pair-wise interactions among variables , there is amble room for identifying higher order interactions within the framework of variable importance .
The BxH mouse weight dataset consists of both categorical and continuous variables: gene expressions of 7 , 441 most varying genes and genotypes of 1 , 065 genetic markers that exhibit variation between two parental strains . The problem can be formulated as a regression problem of predicting the body weights of 132 F2 intercrossed female mice using these two types of variables . The detailed information regarding the experiment and data collection is in Ghazalpour et al . [12] . Given a high-dimensional dataset with multiple types of variables , we adopted a previously developed network construction and clustering method HQCut [27]–[29] to identify the intrinsic structures of variables . HQCut is able to group variables into clusters , i . e . network modules , using a parameter-free spectral clustering-based method to optimize a network modularity function [44] . HQCut has been applied to analyze complex human disease such as Alzheimer's disease [28] , [37] . Pearson's correlation was used to compute the correlation between two continuous variables ( e . g . gene expression ) . The correlation between two categorical variables ( e . g . genetic markers ) was defined as their normalized Mutual Information . For the correlation between a categorical variable and a continuous variable , we first discretized continuous variable X into three categories , given bywhere μX is the mean and δX is the standard deviation of X . We then calculated the correlation of the discretized variable with the categorical variable using Mutual Information . Given the correlations of all pairs of variables , we constructed the correlation network using the same method described in [27]–[29] . For two variables to be connected in the network , their correlation needs to satisfy at least one of the following criteria: ( 1 ) the correlation is greater than 0 . 5 and one of the variable is ranked among the top 5 most correlated variables of the other variable; ( 2 ) the correlation is greater than 0 . 8 and one of the variables is ranked among the top 50 most correlated variables of the other . These two criteria ensure a sparse weighted network structure while maintaining both local ( via rank-based threshold ) and global ( via value-based threshold ) properties [29] . Since different types of variables might have correlation values in different scale , the ranking threshold is independent of each pair of data types . For example , if genotypic and gene expression data are provided , above criteria will be separately applied on SNP-to-SNP , gene-to-gene , and SNP-to-gene ( or the other direction ) respectively . HQcut is then applied to identify the optimal partitioning that cluster genes into non-overlapping modules and automatically determine the number of modules based on the modularity function [44] . The basic building blocks of RF for regression problems are the regression trees [45] , which recursively partitioning the dataset into two subsets based on a specific variable among all variables to minimize the squared loss . RF is an ensemble of regression trees , where each tree is built using a set of bootstrap samples , which is a subset of the original sample . At each splitting ( or internal ) node of a tree only mtry ( k ) small randomly selected variables ( or attributes ) are evaluated . The overall prediction of a forest is the majority vote or the average over the predictions from all individual trees . In a bagging iteration , approximately one third of the observations are not used . These unused observations , the so-called out-of-bag ( OOB ) sample , can be used to estimate the generalization error . In general , three parameters are needed to be determined in the conventional RF algorithm and mgRF: ntrees ( n ) , the number of trees in the forest , mtry ( k ) , the number of candidate variables that each node considers to find the best split , and nodesize ( s ) , the minimum size of sample in a node where no further splitting is needed . Since a large ntrees usually stabilizes variable importance measures , we set it to a large number ( ntrees = 1000 ) in our experiments . We set mtry = m/3 , one third of the number of variables as recommended for regression problems . We set nodesize to 3 as opposed to the recommended 5 because the sample size of our datasets is usually small ( <200 ) . Our preliminary experiments have shown that the RF and mgRF are insensitive to parameter choices . To reduce previously mentioned bias on the variable importance and to incorporate priori knowledge of variables' structure , we adopted a two-stage candidate variable sampling procedure: we first selected a subset of modules , each of which captured a set of correlated variables , and then chose one representative variable from each module to form the candidate splitting variables for each node in RF . Given the module structure of a correlation network , we sampled a subset of modules as candidate modules , from which candidate splitting variables are selected in the second stage . By using the two-stage candidate variable sampling , we could significantly reduce the number of variables to be evaluated in each split as the number of modules was typically much smaller than the best mtry parameter in the original RF . We further defined the module importance ( MI ) of a module mi to be the sum of VIs of its member variables , . MI of a particular module summarized the contribution of all its member variables . We further defined the corrected variable importance ( cVI ) to estimate the importance of individual variables . The cVI was defined as a weighted sum of VIs of all its connected neighbors within the same module . where M ( vi ) is the module that vi belongs to and cij is the measure of correlation between variables vi and vj . We further enhanced mgRF by recursively building a series of RFs , where the selection of candidate splitting variables was not only guided by the module structure as in the two-stage candidate variable sampling procedure , but also guided by MIs and VIs that generated from previous RF . Except the first RF , where both modules and variables were uniformly chosen , the construction of successive RFs follows a modified weighted sampling scheme , which combined both uniform and weighted sampling , to favor informative variables . Take the sampling of modules as an example , given a set S of weighted modules , a subset S1 of size N1 was randomly chosen from S without replacement , where the probability of selecting S1 was proportional to the weights . We then uniformly selected N2 modules from S1 to form a smaller set S2 , which was the set of final candidate modules . The same procedure was applied to selecting variables from each module . The choices of N1 and N2 depended on the data analyzed . In principle , N1 should be large enough to cover truly relevant objects ( i . e . informative variables ) and N2 should be small enough to allow diversity of splitting variables . In the current study we set N1 = N/3 , where N is the size of modules ( or variables in one module ) and performed weighted sampling based on MIs ( or VIs ) for module ( or variable ) sampling . In module sampling , we set , where |M| was the number of modules , because we assumed that there should be multiple latent factors contributing to the trait variable . In variable sampling we let N2 = 1 as previously discussed to ensure the unbiased property of MI . Note that the variable sampling was weighted on VIs instead of cVIs during each run of RFs . In other words , we only corrected VIs generated in the last iteration because VIs of variables within a module are good estimation of their relative importance , which is sufficient for sampling a representative variable from a given module . We downloaded the latest MGI mouse Gene Ontology annotations from Gene Ontology consortium [46] . Given the 7 , 441 most varying genes as the background , and a list of genes to be test , each GO biological processes term was assigned a p-value to quantify the significance of gene-term enrichment using the Fisher's exact test . Regression trees are well suited for modeling non-linear effects such as epistatic interactions because of its conditional splitting method used . We expect that some variables in a regression tree are more likely to be split on when the tree has already been split on a corresponding interacting variable . We derived a simple but effective statistical test to assess the significance of such interactions based on the Fisher's exact test . In particular , given two variables u and v , we counted the number of times they appeared in an ensemble of N trees as n and m , respectively . Under the null hypothesis of u and v being independent , the number of times , k , that they both appears in the same tree should follow the hypergeometric distribution , where C ( x , y ) is a binomial coefficient that choosing y from x . The p-value from the test is computed by summing over the probability of right-tail extremes , where k is set to ( k , min ( m , n ) ) . Note that our definition is slightly different from the one implemented in the conventional RF [18] for classification , where a test based on the Gini importance is applied . In our test , we used the counts of variables being chosen to cope with continuous traits . In our experiment , we used N = 6000 to achieve a stable ranking of interactions . | Obesity has become a perilous global epidemic that can lead to complex diseases , such as diabetes and cardiovascular diseases . Much effort has been devoted to the studies of the genetic mechanisms that pillow the manifestation of obesity . Although a large quantity of experimental data has been accumulated lately using high-throughput techniques , our understanding of genetic mechanisms of obesity is still limited . The proposed method is motivated to address three critical issues that have impeded the existing methods . The first is the curse of dimensionality in selecting a subset of genetic elements related to the traits of interest from a large number of candidates . The second is genetic multiplicity underlying non-Mendelian traits , in which multiple genes are in interplay . The third issue is the integration of data from multiple sources in light of genetic multiplicity and curse of dimensionality . Here , we propose a new method , which augments the Random Forests method with a network-based analysis , to integrate genotypic and gene expression information and identify correlated multiple genetic elements underlying mouse weight . Our results shed light on complex genetic interactions underlying obesity , which can form viable hypotheses worthy of further investigation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"systems",
"biology",
"biology",
"computational",
"biology"
] | 2013 | Integrative Analysis Using Module-Guided Random Forests Reveals Correlated Genetic Factors Related to Mouse Weight |
While some human-specific protein-coding genes have been proposed to originate from ancestral lncRNAs , the transition process remains poorly understood . Here we identified 64 hominoid-specific de novo genes and report a mechanism for the origination of functional de novo proteins from ancestral lncRNAs with precise splicing structures and specific tissue expression profiles . Whole-genome sequencing of dozens of rhesus macaque animals revealed that these lncRNAs are generally not more selectively constrained than other lncRNA loci . The existence of these newly-originated de novo proteins is also not beyond anticipation under neutral expectation , as they generally have longer theoretical lifespan than their current age , due to their GC-rich sequence property enabling stable ORFs with lower chance of non-sense mutations . Interestingly , although the emergence and retention of these de novo genes are likely driven by neutral forces , population genetics study in 67 human individuals and 82 macaque animals revealed signatures of purifying selection on these genes specifically in human population , indicating a proportion of these newly-originated proteins are already functional in human . We thus propose a mechanism for creation of functional de novo proteins from ancestral lncRNAs during the primate evolution , which may contribute to human-specific genetic novelties by taking advantage of existed genomic contexts .
Although it is a generally accepted notion that gene duplication is the major way to create new genes [1–3] , numerous cases have been reported in recent years demonstrating in multiple different species the creation of new proteins out of ancestral non-coding DNAs [4–18] . Recent studies further suggest that this de novo mechanism for gene origination may account for a significant proportion of new genes [2 , 14] and contribute to lineage-specific genetic novelties [18–20] . Currently , several comparative transcriptome studies have proposed that a proportion of de novo genes may originate from ancestral long non-coding RNAs ( lncRNAs ) [7 , 16 , 21] , while the evolutionary mechanism underlying this lncRNA-protein transition remains elusive . First , it is unknown whether the differences in functional significance of lncRNAs , or some other sequence features , could explain the biased origination process of de novo genes from a specific subset of lncRNAs . Specifically , given that ancestral lncRNAs have precise splicing structures and tissue expression profiles similar to those of de novo proteins in human [16] , it is unclear whether they have already obtained certain biological functions on the RNA level: one reason for us to hypothesize that functional non-coding genes may be favorable precursors is because they might survive longer during evolution , providing a wider time window for the emergence and stabilization of ORF , assuming that the emergence of the protein coding part does not interfere with the original function . Second , it is unclear whether the human de novo genes have gained functional significance . Although it has been an established notion that de novo protein-coding genes could have important functions in Drosophila [17 , 18 , 22] , the functional significance of de novo genes in hominoid lineage is still controversial–given the smaller effective population size in hominoids [23] , the detection of these genes may be largely due to the weaker selection for removing the translational noises . Actually , for the dozens of human-specific de novo genes identified , only a few genes were linked to human diseases and regulations by circumstantial evidence [11 , 24–26] . While functional studies with transgenic monkeys could potentially characterize the functions of these hominoid-specific proteins , it is still technically challenging and could not provide a global view of the extent to which these genes are functional . Alternatively , a comparative population genetics approach , i . e . characterizing polymorphisms in the gene locus and comparing the pattern to that of the orthologous region in a closely related species , could provide evolutionary clues to the functional significance of the de novo genes . We thus performed a population genetics study in human and rhesus macaque to interrogate the origin and functional significance of these newly-originated de novo protein-coding genes . We noted that these proteins in human seem to have originated from ancestral GC-rich lncRNAs . Although these lncRNAs generally are not more selectively constrained than other lncRNA loci , and the existence of these newly-originated proteins is not beyond anticipation under neutral expectation , our results showed that at least a proportion of these de novo proteins should have acquired protein-level functions , based on the signatures of purifying selection detected specifically in human populations . We thus propose a mechanism for creation of functional de novo proteins from ancestral lncRNAs during the primate evolution .
To interrogate the genesis and functional implications of de novo proteins in primates , we firstly performed a comprehensive survey for newly-originated de novo protein-coding genes in the hominoid lineage . We devised a genome-wide pipeline integrating ab initio identifications [16] and meta-analysis of public datasets [9–11 , 13 , 16] ( Fig 1A; Materials and Methods ) . Briefly , we first inferred the locus ages on the basis of the syntenic genomic alignment generated by UCSC , and only retained human genes with high-quality alignments in the out-group species ( Discussion ) . With this approach , the potential bias in de novo gene identification introduced by blast-like alignments is well controlled [27] . Then , for each locus , the existence of the ORF in multiple out-group species was inferred separately ( Materials and Methods ) . Candidate de novo genes were then identified based on age assignments of ORFs , by summing up the information on the presence and absence of orthologous ORFs in vertebrate phylogeny with the principle of parsimony [2 , 16] . We further performed sequence similarity study to analyze these candidates against all annotated human proteins , further verifying that they originated through de novo evolution , rather than other mechanisms such as gene duplication ( Materials and Methods ) . The resulting 56 candidates , together with 99 literature-documented primate-specific de novo genes [9–11 , 13 , 16] , were then subjected to additional inclusion criteria ( Materials and Methods ) . Consequently , only genes with 1 ) reliable evidence for transcriptional and translational activities in human ( S1 Fig; Materials and Methods ) , and 2 ) detectable common ancestral “disablers” , disrupting the ORFs in all out-group species at the same sequence position [9 , 11] , as indication for newly-created but not old dying genes , were included ( Fig 1A; Materials and Methods ) . In total , 64 protein-coding genes were identified with recent origination in the hominoid lineage through de novo evolution ( Fig 1B; Tables 1 and S1 ) , with 43 encoding human-specific proteins ( Class I , the younger proteins ) , and another 21 encoding similar proteins in human and chimpanzee but not in rhesus macaque ( Class II , the older proteins ) . The transcript structure and expression of these genes at the transcriptional or translational levels in human are strongly supported by public genomics data . The transcriptional structure for all of these genes were supported by full-length mRNA or spliced EST evidence ( S2 Table ) , with 88% of the splicing junctions also supported by short RNA-Seq reads ( Fig 1E; Materials and Methods ) ; the full-length transcript structure for 17 of these genes were also verified by the Iso-Seq data , generated recently through the PacBio transcriptome sequencing ( S2 Table; Materials and Methods ) . In addition , the protein expressions for all of these genes were supported by large-scale mass spectrometry studies in human ( Tables 1 and S2; Materials and Methods ) . To infer the transcriptional capacity of the 64 de novo genes in the common ancestor of human and closely related species , we performed cross-species transcriptome analysis in human , chimpanzee and rhesus macaque . First , we found that 83 . 9% of the 64 genes , and 92 . 9% of the 43 human-specific genes transcribed in at least one tissue in rhesus macaque or chimpanzee as lncRNAs ( Fig 1C ) , with the expression levels significantly higher than the background expression levels ( S1 Fig; Materials and Methods ) . Second , the expression levels of the genic regions relative to upstream and downstream regions were comparable among the three species ( Fig 1D ) , and the majority of human splicing junctions were also detectable in chimpanzee and rhesus macaque orthologous regions ( Fig 1E ) . Third , the non-coding orthologs of human de novo genes in rhesus macaque and chimpanzee also show tissue expression profiles similar to human ( Fig 1F ) . The inter-species similarity of tissue expression profiles was further supported by clustering analysis , with the same tissue types from different species clustered together ( S2 Fig and S3 Table ) . By the parsimony principle , we conclude that the transcription structure and expression profile of these de novo genes had been shaped in the common ancestor prior to the acquisition of coding potential in the human lineage . It is thus interesting to investigate whether these lncRNA precursors with precise splicing structures and tissue expression profiles have already obtained certain biological functions on the RNA level , and may thus represent favorable precursors for new de novo proteins . Because lncRNAs could have a variety of functions , not all of which can be easily assayed , as an alternative , we sought an evolutionary approach by quantifying the level of selective constraints in the orthologous lncRNA loci of these de novo genes in rhesus macaque as a proxy for determining the functional status in the ancestor . This assumes that the selective constraints on these loci have remained unchanged in the macaque lineage since it had a common ancestor with human . We first compiled the whole lncRNAome in rhesus macaque using a similar strategy as described previously [31] , on the basis of strand-specific RNA-Seq in ten tissues from the same macaque animal [16 , 32] ( Fig 2A; Materials and Methods ) . A total of 5 , 641 lncRNA transcripts were assembled , verifying known features of transcripts such as the tight association with epigenetic markers and CpG islands [33 , 34] ( Figs 2B and 2C and S3 ) . Moreover , as positive control , we compiled a list of 89 non-coding genes in rhesus macaque . These non-coding genes are reportedly functional in human , as supported by experimental evidence [35] . Given the existence of the similar lncRNA transcripts in rhesus macaque , we assumed that these macaque lncRNA transcripts may also have functions and are under similar selective constraints as annotated non-coding genes in human ( Materials and Methods ) . To quantify the level of selective constraints , we then performed whole genome sequencing of 24 independent macaque animals and generated 23 . 7 billion paired-end reads with high quality , yielding high sequencing coverage of the macaque genome ( ranging from 26- to 70-fold ) . On the basis of these sequencing data , as well as seven public datasets for macaque genomes [36–38] , we profiled 54 , 079 , 575 single nucleotide polymorphic sites across the macaque genome , yielding a considerable number of polymorphisms located on the lncRNA loci ( S4 Table; Materials and Methods ) . On the basis of the polymorphism data in the population of 31 unrelated macaque animals , we measured the level of polymorphisms in the subset of macaque lncRNA loci , whose orthologous regions in human were de novo genes , and compared that to the same measures in all macaque lncRNAs , as well as the list of 89 established non-coding genes as a reference [35] ( Materials and Methods ) . As expected , we found that the list of 89 established non-coding genes are selectively constrained in rhesus macaque on the basis of the significantly decreased nucleotide diversity ( π ) , compared with the synonymous sites of known macaque coding genes as a neutral control ( Wilcoxon one tail test , p-value = 0 . 008 , Fig 2D ) . The selection on these non-coding genes seems to be moderate , as compared with that of the non-synonymous sites of known macaque coding genes as a benchmark ( Fig 2D ) . In contrast to this small repertoire of non-coding genes with functions , it seems that lncRNA transcripts are in general not selectively constrained in rhesus macaque , with the nucleotide diversity comparable with that of the synonymous sites across the macaque genome as a neutral control ( Wilcoxon test , p-value = 0 . 917 , Fig 2D ) . In addition , the orthologous loci for the lncRNA precursors of human de novo genes are not subjected to strong selective constraints as those non-coding genes , with the population genetics feature indistinguishable from that of the synonymous sites ( Wilcoxon test , p-value = 0 . 570; Fig 2D ) , as well as that of the whole lncRNA pool ( Wilcoxon test , p-value = 0 . 449; Fig 2D; Materials and Methods ) . In conclusion , we didn't find evidence for higher selective constraints for the orthologous lncRNA loci for human de novo gene precursors . Hence , it seems likely that the ancestor of de novo genes may not be particularly distinct in terms of functional importance before the proteins arise . Given that the lncRNA precursors for human de novo genes did not display particularly distinct functional status , it is interesting to investigate whether other features , such as sequence features , may explain why de novo genes originate from some lncRNAs but not the others . In addition , given the smaller effective population size in hominoids , the detection of de novo proteins might arise from translational noise that is not acted upon or not yet removed by purifying selection , rather than being positively selected for due to their newly-acquired protein-level functions . We thus performed comprehensive sequence analysis of these de novo genes to investigate whether any sequence features could underlie the biased origination process of de novo genes from a subset of lncRNAs , and whether the existence of these de novo proteins is beyond anticipation in terms of their theoretical lifespan . Interestingly , when analyzing the sequence features of these orthologous lncRNA precursors , we found that they have significantly higher GC contents in comparison with other lncRNAs and non-genic regions ( Fig 3A and 3B; Wilcoxon one-tailed tests , p-value<1 . 0e-7 for both comparisons ) . The ORFs of de novo genes derived from GC-rich lncRNAs were also observed to have significantly higher GC content when compared with functional proteins in RefSeq ( Fig 3B and 3C; Median of GC content = 0 . 57 vs . 0 . 53; Wilcoxon one-tailed test , p-value = 6 . 6e-6 ) . Further investigation revealed that such a GC-rich property endows these newly-originated ORFs with longer theoretical lifespan , even longer than their current age . As the stop codons are AT-rich , the higher GC content usually supports relatively stable ORF compared with GC-poor sequences [39 , 40] ( S4 Fig ) . As expected , compared with RefSeq proteins , these newly originated ORFs have less content of fragile codons–codons convertible to stop codon by a single point-mutation , and are thus less susceptible to non-sense mutations ( Fig 3D; Wilcoxon one-tailed test , p-value = 0 . 002 ) . Accordingly , we found that these ORFs have long half-life time under neutrality ( Materials and Methods ) , even significantly longer than other functional proteins in RefSeq ( Fig 3E; Wilcoxon one-tailed test , p-value<2 . 2e-16 ) . Especially , compared with the younger de novo genes , the older de novo genes have higher GC content ( Fig 3C ) , less content of fragile codons ( Fig 3D ) and longer half-life time ( Fig 3E ) . Overall , the theoretical lifespan of these newly-originated proteins is generally longer than their current age ( Figs 1B and 3E ) , thus indicating that the existence of these de novo proteins is not beyond anticipation even under neutral expectation ( Fig 3F ) . Overall , the de novo gene repertoire we identified in the hominoid lineage actually represents a snapshot for the steady-state representation of a dynamic turnover process of ORFs . The detection of these GC-rich de novo proteins with stable ORFs , together with the previous reports that many de novo genes have stable expression profiles possibly by sharing the transcriptional context with nearby protein-coding genes through cis-natural antisense or bi-directional promoters [16 , 41] , seems to favor the notion that a significant portion of the turnover is probably driven by genetic drift and those GC-rich "survivors" with long ORF lifespan and stable expression profile were retained and detected during a birth-and-death process . What we found above suggest that the emergence and retention of de novo genes are likely under neutral forces . However , considering these GC-rich “survivors” have been exposed to natural selection for relatively long time due to their theoretically longer lifespan , it is interesting to investigate whether some of these newly originated ORFs have been maintained by selective constraint in the current population , due to their newly-acquired protein-level functions . We thus performed population genetics study in human and rhesus macaque populations to assess whether selective constraints are applied to these ORF regions in human populations but not their non-coding counterparts in rhesus macaque . We first profiled a set of polymorphism sites in human populations , by re-analyzing whole genome sequencing data in 67 individuals from different sub-populations ( Fig 4A and S5 Table; Materials and Methods ) . We expect that if the de novo genes encode functional proteins and are maintained by purifying selection , the polymorphism level for exonic regions of these genes will be lower than intronic regions . The polymorphism level for non-synonymous sites should also be significantly lower than that of synonymous sites , as the former will be under much stronger selection . Moreover , we expect to find a difference in the frequency spectra at nonsynonymous vs . synonymous sites , resulting in a skew towards low frequency variants compared to the latter . These are indeed what we found: 1 ) The θw and π measures were significantly lower in the exonic region of the de novo genes compared to the intronic region of the same locus ( Monte Carlo p-values<1e-4; Figs 5A and S5; Materials and Methods ) . In addition , the UTR regions of these de novo genes showed θw and π measures that are lower than the intronic regions , while slightly higher than CDS regions ( Figs 5A and S5 ) . 2 ) Compared with synonymous sites , the nucleotide diversity for non-synonymous sites was significantly lower ( Wilcoxon one-tail test , p-value = 0 . 019; S6 Fig ) . Accordingly , the ratio of the nucleotide diversity for non-synonymous sites to synonymous sites was generally smaller than 1 ( Fig 5B ) . 3 ) The frequency spectrum of the derived alleles had an excess of low-frequency variants at the non-synonymous sites in the de novo genes compared to that at the synonymous sites ( Fig 5C ) , which is similar to known protein-coding genes ( Fig 5E ) . As a control , we classified mutations in human lncRNAs into synonymous or non-synonymous sites within the longest pseudo-ORFs , and didn’t observe any difference in their respective frequency spectrum ( Fig 5F; Materials and Methods ) . Accordingly , as a negative control , we performed population genetics study on macaque orthologous regions ( without coding potential ) of human de novo genes , in a population of 82 unrelated rhesus macaque animals ( Materials and Methods ) . Custom library with >135 , 000 120-bp DNA oligos were designed to capture the macaque orthologous regions . Ultra-deep sequencing was then performed ( Materials and Methods ) and 222 million 150-bp paired-end reads were generated and uniquely located on the macaque genome ( NCBI SRA accession number: SRP052932; S6 Table ) . The average effective coverage of the targeting regions reached to 94% in each sample ( Figs 4B and 4C and S7 ) , and a total of 10 , 162 single nucleotide polymorphisms were identified across the target genomic regions with high sensitivity and specificity , as verified by a follow-up whole genome sequencing with 30× coverage in one of these macaque animals ( Fig 4D and 4E; Materials and Methods ) . On the basis of the polymorphism data of the macaque orthologous regions , we found that both the θw and π measures were uniform across the length of these regions , in contrast to the clear differences observed in human ( Figs 5A and S5 ) . We further classified the polymorphism sites on macaque lncRNAs into synonymous or non-synonymous sites , according to codon-level alignments between human de novo proteins and their orthologous lncRNAs in rhesus macaque . No significant difference was detected for the nucleotide diversity of pseudo-non-synonymous and pseudo-synonymous sites in rhesus macaque ( Wilcoxon one-tail test , p-value = 0 . 607; S6 Fig ) , with the ratio of the nucleotide diversity between these two groups comparable to 1 ( Fig 5B ) . In addition , the resulting frequency spectrum of derived alleles at the pseudo-non-synonymous sites is indistinguishable from that at the pseudo-synonymous sites ( Fig 5D ) . The population genetics analyses thus suggest that these newly-originated de novo genes have gained new functions specifically in human . Taken together , although the de novo proteins seem to emerge from lncRNA precursors with no bias towards those functionally-constrained lncRNAs , and their existence is not beyond the anticipation under neutral expectation , at least a proportion of these proteins should have acquired protein-level functions specifically in human , as revealed by the species-specific signatures of purifying selection on these newly-originated de novo genes . We thus depicted a new mechanism for the origination of functional proteins from ancestral non-coding transcripts with precise splice structures and specific tissue expression profiles during the primate evolution .
Although our current study identified a list of 64 hominoid-specific de novo genes , more proteins are expected to originate through this de novo mechanism [42] . Several types of de novo genes might be underrepresented due to the computational pipelines currently used to identify these genes: 1 ) Genes with shorter ORF . Automatic gene annotation pipeline typically neglect genes with short ORFs by using arbitrary criteria to define the minimal ORF length . Consequently , considering de novo proteins generally have short ORFs [43–45] , a large proportion of de novo proteins with short ORFs were removed . 2 ) Genes without orthologous regions in out-group species were not included , due to the requirements of the high-quality alignments for accurate age assignments of ORFs in vertebrate phylogeny , as well as of the detection of common ancestral “disablers” as indication for newly-created rather than old dying genes . Although such a design effectively lowered the false-positives and potential bias introduced by blast-like alignments in de novo gene identification [27] , some false-negatives could still result from our stringent criteria . 3 ) Genes with non-stable accession numbers . Considering the difficulties in defining the coding potential , the Ensembl accession numbers assigned to de novo genes are typically not stable . Although our study has combined multiple versions of Ensembl databases to identify de novo genes , some genes may still be overlooked . 4 ) Genes with low expression . De novo genes typically express in low abundance [13 , 16 , 18] . As reliable evidence for transcriptional and translational expression is needed to define a de novo gene , especially considering the relatively low sensitivity of mass spectrometry technology to reliably detect peptides of low abundance , these genes may be missed . 5 ) Additionally , the current identification pipeline by comparative genomics approaches typically focuses on in-group ORFs that are missed in other out-group species . In such an occasion , a considerable proportion of de novo genes originated through lineage-specific expression of pre-existing ORFs might be neglected [17 , 18] . Overall , although these 64 genes may not fully recapitulate the true repertoire of de novo genes in the hominoid lineage , they should constitute a representative group for further analysis and elucidation of the evolving process of de novo genes from precursor lncRNAs . Although de novo genes are also regarded as “motherless” genes due to their lack of ancestral protein-coding genes as precursors , we and others have found that at least a proportion of lncRNAs might represent an intermediate stage of their origination , narrowing the gap between non-coding DNA and protein-coding genes . Such an origination process may take advantage of existed genomic contexts . For example , these lncRNA “precursors” usually share the transcriptional context with the nearby protein-coding genes through cis-natural antisense or bi-directional promoters [16 , 41] . These lncRNAs with stable expression profiles , although not more selectively constrained according to our population genetics study , may then lay the foundation for the emergence of new de novo genes . In addition , the GC-rich sequence property of these lncRNAs further supports stable ORFs of the newly-originated proteins ( Fig 3 ) . Overall , these genomic features provide a theoretically favorable foundation for the birth of some functional proteins–a notion well supported by our population genetics data , which revealed that some of these loci already encode human-specific functional proteins ( Fig 5 ) . Although such an origination process is plausible , currently both the definitions of de novo genes and lncRNAs are depending on some arbitrary criteria [9 , 13 , 31] . Additional lines of evidence are thus needed to fully support the mechanism through which de novo genes come from the lncRNA pools during the primate evolution . For example , since it is still technically challenging to fully annotate proteomes based on mass-spec studies across tissues , development stages and species , it is inadequate to directly identify human de novo genes on the basis of the presence or absence of peptides across different species . Alternatively , conceptual translation of ORFs between species is still the main strategy in the field to infer the existence of these de novo proteins in different out-group species [9 , 13] . In this context , although the orthologs of these human de novo genes could be defined as “lncRNAs” in chimpanzee and rhesus macaque by the current criteria , they may actually encode smaller version of these de novo proteins in out-group species . We thus performed cross-species analyses to test this “functional ORF expansion model” . Briefly , if the functional proteins were indeed absent from out-group species , we would expect similar substitution rates between non-synonymous and synonymous sites when performing comparative genomics analysis . When aligning the truncated forms of the human de novo proteins in non-human primates , we found that the merged dN/dS ratio does not deviate significantly from 1 ( dN/dS = 0 . 90 ) . In line with this finding , our population genetics study on macaque orthologous regions of human de novo genes also indicated that these macaque orthologs may not encode similar functional proteins as in human ( Fig 5 ) . Even considering these population genetics evidence , we still could not fully exclude the possibility that some of these so-called “lncRNAs” might actually encode fast-evolving or smaller version of the protein in out-group species . As such , these proteins might be under weak selection , and the signals for selective constraints could not be detected based on the population size of this study . Future mass-spec studies with high sensitivity may aid in clarifying these issues . Protein-coding genes have typically higher GC content than non-coding regions . It has been proposed that the increased GC content in genic regions could be maintained by natural selection , such as the GC preference on the wobble sites potentially shaped by adaptive evolution for the stability of mRNA secondary structure or the efficient protein synthesis [46] . However , here we found that other gene-associated genomic regions are also GC-rich , such as the intronic regions ( Fig 3B ) . Theoretically , considering the models for new gene origination , each protein-coding gene could be traced to an ancient origination event from non-coding DNA . There is thus a formal but as yet unexplored possibility that the biased inheritance from GC-rich lncRNAs could be another major factor underpinning the different extents of GC content between coding regions and genomic background . As we provided an evolutionary and functional connection between protein-coding genes and non-coding DNA regions in the hominoid lineage , we formally tested this “GC-rich inheritance model” . Correspondingly , we found that GC-rich lncRNAs are favorable precursors for new proteins . More importantly , the GC-rich features could be detected in all genomic regions associated with these newly originated de novo genes , even for the wobble sites with established GC preference , as well as their lncRNA precursors , which also resembled those of the well-known protein-coding genes ( Figs 3B and S8 ) . Besides being a consequence of adaptive evolution after the acquirements of the ORFs , the GC-rich feature of the protein-coding genes may also inherit from the ancestor lncRNAs , thus complementing previous theory on GC-rich feature for protein-coding genes [46–49] .
Rhesus macaque samples were obtained and manipulated from the internationally-accredited ( Association for Assessment and Accreditation of Laboratory Animal Care , AAALAC ) animal facility of the Institute of Molecular Medicine in Peking University . The present study was approved by the Institutional Animal Care and Use Committee of Peking University . De novo protein-coding genes in the hominoid lineage were identified using a genome-wide pipeline integrating ab initio identifications and meta-analysis of public datasets . On the basis of Ensembl gene annotations ( v68 ) , de novo genes were identified using a similar pipeline as we published previously [16] . Briefly , 1 ) we inferred the locus ages on the basis of the syntenic genomic alignment generated by UCSC , and only human genes assigned with specific locus age were retained; 2 ) for locus with high-quality alignment ( coverage >70% and identity >50% ) in the out-group species , the existence of the ORF in multiple out-groups ( chimpanzee , orangutan , rhesus macaque , mouse , guinea pig , dog , hedgehog and armadillo ) was inferred separately by Exonerate [50] , and if the sequence in particular out-group encoded at least one frame-disrupting indel or premature stop codon , with the subsequent maximum continuous ORF shorter than 70% of the human ORF length ( a cutoff based on the previous practice in this field [9 , 13 , 16] ) , the ORF was regarded as non-existent in this out-group; 3 ) we inferred the origination timing of ORFs for these de novo genes by summing up information of their presence or absence in multiple out-group species , along with the phylogenetic tree with the principle of parsimony , and subsequently retained only genes originated in the hominoid lineage; 4 ) sequence alignments were then performed against all human proteins ( BLAST e-value cutoff of 10−6 ) to ensure these new genes originated through de novo evolution other than gene duplications . Finally , 56 protein-coding genes were identified as candidate de novo genes in the hominoid lineage . The resulted 56 candidate genes , together with 99 literature-documenting primate-specific de novo genes [9–11 , 13 , 16] , were then subject to two additional inclusion criteria . First , genome-wide expression filters were introduced to ensure these genes had convincing evidence for transcriptional and translational expression in human . Public RNA-Seq data in 17 human tissues ( Human BodyMap 2 . 0 data from Illumina and data from references [51 , 52] ) were integrated and analyzed to estimate the gene expression level of each gene , according to a standardized pipeline [53] . To distinguish true transcription signals from the background expression , we first estimated the RPKM values for the genomic background represented by 10 , 000 randomly-selected intergenic regions . The expression levels of intergenic regions were significantly lower than 0 . 2 RPKM in all tissues ( S1A Fig; Monte Carlo p-values ranging from 0 . 002 to 0 . 028 ) . Therefore , a more conservative PRKM cutoff of 0 . 5 was arbitrarily set to confirm the transcriptional expression of these de novo genes in human . Two candidates ( ENSG00000205056 and ENSG00000198547 ) with low RPKM scores were also included due to their reliable experimental evidence for transcriptional expression [9 , 11] . Peptide evidences from large-scale mass spectrometry studies were then extracted from PRIDE [28] , PeptideAtlas [54] , ProteomicsDB [55] and Human Proteome Map [29] . A peptide was considered to support the protein expression of a de novo gene only if 1 ) when performing BLAT similarity searches against all human proteins ( Ensembl v68 , BLAT settings-t = prot-q = prot-stepSize = 5 ) , its whole sequence exactly match the CDS region of the de novo gene , with the second-best hit in the proteome ( if existing ) including at least one mismatch; and 2 ) when performing BLAT similarity searches against the human genome , its whole sequence identically and exclusively match the CDS region of the de novo gene ( hg19 , BLAT settings-stepSize = 5-stepSize = 5-t = dnax-q = prot ) . Only genes with 1 ) RNA-Seq RPKM >0 . 5 in at least one of the 17 human tissues , and 2 ) at least one convincing item of peptide evidence in support , were retained ( Fig 1A ) . Second , to verify that these genes are newly-originated rather than old dying genes , we manually checked the corresponding ORF regions in multiple out-group species ( chimpanzee , orangutan , rhesus macaque , mouse , guinea pig , dog , hedgehog and armadillo ) , and only genes with common ancestral disablers shared by multiple out-group species were retained ( Fig 1A ) . Here , a common ancestral disabler refers to a mutation disrupting the ORF in multiple out-group species at the same sequence position [9 , 11] . In such scenario , the mutation is more likely to be of an ancestral status according to the parsimony principle , thus indicating the gene is newly-originated rather than old dying . Totally , a list of 64 genes was identified to originate recently in the hominoid lineage through de novo evolution ( Fig 1B and Tables 1 , S1 and S2 ) . We also studied the characteristics of these de novo genes across primate species in the context of new genomics technologies . According to computational pipelines described previously [9 , 13 , 16] , mRNA and EST data from UCSC Genome Browser , RNA-Seq data archived in RhesusBase [53 , 56] , as well as single-molecule long-read sequencing data on human transcriptome [57] were downloaded and analyzed to investigate the transcriptional structure of these de novo genes in human . On the basis of public RNA-Seq data in human , chimpanzee and rhesus macaque [16 , 51 , 52] , comparative transcriptome studies were then performed to compare the transcription level , splicing structure and tissue expression profiles of these de novo genes with their non-coding orthologs in chimpanzee and rhesus macaque , according to a pipeline previously described by us [16] . Specially , an RPKM cutoff of 0 . 2 was set to distinguish convincing transcription and transcriptional noise as described above ( S1B and S1C Fig ) . Strand-specific , Poly ( A ) -positive RNA-Seq data in ten tissues ( adipose , prefrontal cortex , cerebellum , heart , kidney , liver , lung , muscle , spleen , testis ) of the same macaque animal were used to assemble the lncRNAome in rhesus macaque [16 , 32] , following a computational pipeline as described previously [31] . Briefly , RNA-Seq reads of each macaque tissue were aligned separately to the macaque genome ( rheMac2 ) with Tophat ( v2 . 0 . 6 ) [58] . Transcriptome assembly was then performed with both Cufflinks ( v2 . 0 . 2 ) and Scripture ( VPaperR3 ) [59 , 60] , and redundant transcripts were merged with Cuffcompare ( v2 . 0 . 2 ) [60] after boundary correction . To control for false-positives , only long , multi-exonic transcripts ( >200 bp ) with supportive evidences in ≥2 tissues or by both assemblers were retained [31 , 61] . To evaluate the performance of this transcriptome assembly , Cuffcompare ( v2 . 0 . 2 ) was also introduced to compare the assembled transcripts with multi-exonic protein-coding genes as annotated in RefSeq . Finally , a total of 90 , 322 multi-exonic transcripts were assembled , which represents transcript structures reconstructing for 95% known multi-exonic protein-coding genes , suggesting the feasibility of this assembly strategy ( S3 Fig ) . Several stringent criteria , such as proteome annotation- and comparative genomics-based filtering procedures , were incorporated to exclude protein-coding transcripts . Briefly , 1 ) transcripts with ≥1 splice junction overlapped with known protein-coding genes annotated in either Ensembl or RefSeq were discarded; 2 ) PhyloCSF was applied to score the coding potential of these candidates ( multiple sequence alignments of 9 mammalian genomes , —frames = 3—orf = StopStop3 ) [62] and only transcripts with PhyloCSFscore <65 were retained , corresponding to a false negative rate of 1% and a false positive rate of 5% on the basis of RefSeq annotation . 3 ) the nucleotide sequences or 3-frame stop-to-stop translation products were subjected to Blastx , Blastp and HMMER searching against all human proteins or known protein domains ( Pfam-A , Pfam-B ) [63] , and transcripts with significant hits ( e-value ≤10−4 ) were discarded; 4 ) transcripts with putative ORFs ≥100aa longer were also discarded . The strategy had a good performance in distinguishing lncRNAs from protein-coding transcripts ( S3 Fig ) and a total of 5 , 641 lncRNA transcripts were assembled . On the other hand , non-coding genes in human were searched and downloaded from lncRNAdb ( http://www . lncrnadb . org/ ) [35] and the macaque orthologs of these functional human lncRNAs were then retrieved by liftOver . For human de novo genes and their orthologs in chimpanzee and rhesus macaque , sequences of different genomic regions were retrieved and the GC contents were calculated by a customized Perl script ( https://github . com/Jia-Yu-Chen ) . Following a previous study [40] , we also calculated for each de novo gene the proportion of fragile codons that could become stop codons by single mutation . We further investigated whether the existence of these de novo proteins is beyond anticipation in terms of their theoretical lifespan . Briefly , an ORF would eventually be interrupted if it does not experience any functional constraint , and the rate of ORF interruption under neutrality is largely determined by point mutation rate and insertion/deletion mutation rate . We thus estimated the interruption rate in terms of half-life time of ORF ( t1/2 ) according to the computational simulation method developed by Zhang and Webb [64] ( Fig 3E ) . The half-life time of a given ORF is the time required for an ORF to be interrupted in one-half of 20 , 000 simulation replicates , with the rates of point and insertion/deletion mutations being set as 1 . 25 and 0 . 1 per site per billion years , respectively , as previously defined [64] . Given this half-life time ( t1/2 ) , the minimal probability of an ORF remaining intact today under neutrality was then determined by the following equation , by assuming that the lineage-specific gene emerged right after the species divergence from the most recent common ancestor ( T ) ( Fig 3F ) . We firstly profiled the polymorphism data in human populations , by re-analyzing whole genome sequencing data in 67 individuals from different sub-populations and archived with high sequencing coverage by the 1000 Genomes Project ( S5 Table ) . Briefly , deep sequencing reads were mapped to the human genome ( hg19 ) using BWA [65] , and the polymorphism sites of each sample were identified and evaluated according to the standard GATK pipeline with UnifiedGenotyper ( V2 . 7–4 ) [66] . After stringent filtering strategies to remove false-positives in variant calling , 18 , 186 , 523 highly reliable single nucleotide variants were identified across the human genome , with 85 . 4% supported also by the 1000 Genomes Project ( Fig 4A ) . Accordingly , we profiled the distributions of polymorphic sites in rhesus macaque populations as a reference . For each human de novo gene , we performed targeted capture and ultra-deep sequencing of the macaque orthologous regions ( and 1kb flanking regions ) in a population of 82 unrelated male animals ( S6 Table ) . Briefly , custom library with >135 , 000 120-bp DNA oligo probes were designed by Agilent SureSelect XT Target Enrichment System ( Agilent Technologies , Inc . , Santa Clara , USA ) , with a 3-folds tiling coverage , to capture the targeted regions in rhesus macaque . Genomic DNA from the macaque animals was isolated from 200–500μl whole blood using the QIAamp DNA Blood Mini Kit ( Qiagen , Venlo , Netherlands ) and 3μg DNA of each animal was sheared to fragments with a peak at 150–200bp using Covaris S220 . Then , the adaptor-ligated libraries were amplified , purified and hybridized with SureSelect Capture Library according to the manufacturer’s instructions . After 16h hybridization at 65°C , the captured targets were pulled down by Dynabeads MyOne Streptavidine T1 ( Life Technologies , Ltd . , Carlsbad , USA ) and amplified for the library preparation , which were then sequenced on Illumina Miseq system with 151-bp paired-end read mode . Totally 222 million 150-bp paired-end reads were generated and uniquely located on the macaque genome ( S6 Table ) . The average effective coverage of the targeted regions reached to 94% in each sample ( Fig 4B ) , and 86% of whole orthologous region ( or 95% of CDS regions ) were sequenced with coverage of ≥30 in all of the 82 macaque animals ( Figs 4C and S6 ) . A total of 10 , 162 highly-reliable single nucleotide polymorphisms were then identified , according to the standard GATK pipeline with UnifiedGenotyper ( V2 . 7–4 ) [66] . To evaluate whether allele dropout or other false-positives introduced by target capture may compromise our approach , we further performed whole genome sequencing with 30× coverage in one of these macaque animals ( Animal ID: 920653 ) for evaluation ( Fig 4B ) . Genomic DNA was obtained for the library preparation of whole genome re-sequencing , deep sequencing was performed on a HiSeq 2000 Sequencing System with a 151×2 paired-end read mode , and single nucleotide polymorphisms were identified according to the standard GATK pipeline . Finally , 96 . 2% polymorphic sites identified in the targeted sequencing were verified by the whole genome sequencing , with 99 . 5% showing the same genotype ( Fig 4D and 4E ) . Whole genome sequencing data from 24 macaque animals generated previously in our lab , as well as seven animals published previously [36–38] , were also analyzed to profile a genome-wide polymorphism dataset across the macaque genome , according to the pipeline as described above . All deep sequencing data in this study are available at NCBI SRA under accession numbers SRP052932 . On the basis of the polymorphism data from the population of 31 macaque animals , we measured the nucleotide diversity ( π ) for the orthologous lncRNA loci of human de novo genes , all lncRNAs , and a list of 89 functional non-coding genes in rhesus macaque . Non-synonymous and synonymous sites of macaque protein-coding genes as annotated by RefSeq were used as benchmarks for the extent of the selective constraints ( Fig 2D ) . Wilcoxon test was performed to test whether the nucleotide diversity between two groups are significantly different , with a p-value cutoff of 0 . 05 ( Fig 2D ) . On the basis of the polymorphism data obtained by analyzing the whole genome sequencing data of 67 human individuals and the targeted sequencing data of 82 macaque animals , we estimated the polymorphism levels ( θw and π ) for different genomic regions ( exon , intron , CDS and UTR ) of the de novo genes in human and their orthologs in rhesus macaque ( Figs 5A and S5 ) . We further performed statistical tests to determine whether the different polymorphism levels between exonic and intronic regions of human de novo genes are statistically significant , with the background estimated by 10 , 000 times of Monte Carlo simulations , assuming the polymorphic sites were randomly distributed in exonic and intronic regions of human de novo genes . Considering that the average nucleotide diversity in rhesus macaque is higher [67] , if the exonic regions are more selectively constrained than intronic regions , we should have greater statistical power to detect the difference . The observation of a comparable nucleotide diversity between macaque exonic and intronic regions then indicates that these macaque orthologs of human de novo genes may not encode similar functional proteins as in human ( Figs 5A and S5 ) . The ratio of the nucleotide diversity between non-synonymous sites to synonymous sites was also determined for each de novo gene , as well as its non-coding ortholog in rhesus macaque ( Fig 5B ) , in which the pseudo-non-synonymous and pseudo-synonymous sites in macaque orthologs were determined by codon-level alignment with human de novo proteins . For each polymorphic site , the derived allele was defined by the EPO pipeline [68 , 69] . The frequency spectra of derived alleles were then estimated , with 1 , 000 times of bootstrap performed to estimate the confidence intervals of the proportions of polymorphism sites ( Fig 5C–5F ) . Similar analyses were performed for known protein-coding genes as annotated by RefSeq , as well as human lncRNAs as annotated by GENCODE ( v19 ) as controls . | Although gene duplication has been believed as a predominant mechanism for creating new genes , recent reports suggested that new proteins could evolve “de novo” from non-coding DNA regions . These de novo genes are also named as “motherless” genes due to their lack of ancestral proteins as precursors , while recently we and others found that lncRNAs may represent an intermediate stage of their origination . To further elucidate this lncRNA-protein transition process , here we identified 64 hominoid-specific de novo genes and report a new mechanism for the origination of functional de novo proteins from ancestral non-coding transcripts: These non-coding “precursors” are generally not more selectively constrained than other lncRNA loci; and the existence of these de novo proteins is not beyond anticipation under neutral expectation; however , population genetics study in 67 human individuals and 82 macaque animals revealed signatures of purifying selection on these genes specifically in human population , indicating a proportion of these newly-originated proteins are already functional in human . We thus propose a mechanism for creation of functional de novo proteins from ancestral lncRNAs during the primate evolution . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Emergence, Retention and Selection: A Trilogy of Origination for Functional De Novo Proteins from Ancestral LncRNAs in Primates |
The study aimed to determine costs to the state government of implementing different interventions for controlling rabies among the entire human and animal populations of Tamil Nadu . This built upon an earlier assessment of Tamil Nadu's efforts to control rabies . Anti-rabies vaccines were made available at all health facilities . Costs were estimated for five different combinations of animal and human interventions using an activity-based costing approach from the provider perspective . Disease and population data were sourced from the state surveillance data , human census and livestock census . Program costs were extrapolated from official documents . All capital costs were depreciated to estimate annualized costs . All costs were inflated to 2012 Rupees . Sensitivity analysis was conducted across all major cost centres to assess their relative impact on program costs . It was found that the annual costs of providing Anti-rabies vaccine alone and in combination with Immunoglobulins was $0 . 7 million ( Rs 36 million ) and $2 . 2 million ( Rs 119 million ) , respectively . For animal sector interventions , the annualised costs of rolling out surgical sterilisation-immunization , injectable immunization and oral immunizations were estimated to be $ 44 million ( Rs 2 , 350 million ) , $23 million ( Rs 1 , 230 million ) and $ 11 million ( Rs 590 million ) , respectively . Dog bite incidence , health systems coverage and cost of rabies biologicals were found to be important drivers of costs for human interventions . For the animal sector interventions , the size of dog catching team , dog population and vaccine costs were found to be driving the costs . Rabies control in Tamil Nadu seems a costly proposition the way it is currently structured . Policy makers in Tamil Nadu and other similar settings should consider the long-term financial sustainability before embarking upon a state or nation-wide rabies control programme .
While rabies has been identified as a priority zoonoses that needs to be addressed globally [1] , it has a special relevance in South Asia . More than 55 , 000 rabies deaths have been estimated to occur among humans annually with little under half being contributed by India alone [2] , [3] . Experts from animal as well as human health sectors agree on the controllable nature of the disease and on the importance of joint population level interventions for restricting disease transmission among animals and humans [4] , [5] . Evidence from India and elsewhere demonstrates the efficacy of principle rabies intervention strategies . Indian researchers have studied the application of different post-exposure prophylaxis ( PEP ) regimens among humans [6] . Indian researchers have also used the experience of dog population control in specific urban settings to demonstrate the impacts of animal birth control strategies [7] , [8] . Of late there is mounting evidence produced by international researchers related to the efficacy of anti-rabies immunization among animals in reducing rabies transmission [9] . Economic assessments have also been conducted in different parts of the world which study the economic impact of rabies [2] , economics of rabies control [10] and cost effectiveness of different post-exposure prophylaxis regimens [11] . This body of work has been instrumental in development of national strategic plans for rabies control [12] . However , as previously documented , rabies researchers have not been able to satisfy the information needs of policymakers [13] and the economics of rabies control remains a “significant constraint” in rolling out rabies control programmes in low income countries [14] , [15] . A possible explanation could be that to date , only a handful of studies have looked at combined costs of rabies across human and animal sectors [2] , [10] , [16] . Most of these analyses have been conducted from the societal perspective that is of limited use to program managers . Additionally , because of the design of cost effectiveness analyses , their findings are always relative in nature and are difficult to generalise in absolute terms . Accordingly , we undertook a costing exercise building upon an earlier assessment [17] of rabies control initiative in the Southern Indian state of Tamil Nadu . Its objective was to determine the costs to the government of implementing different combinations of strategies for controlling rabies among human and animal populations in a state like Tamil Nadu . Tamil Nadu is the southernmost state in India having a population of 72 million [18] and is considered one of the better performing states in public health [19] . According to the results of a study based upon verbal autopsy of deaths between 2001–03 , it had 0 . 5 deaths or fewer per 100 , 000 human population due to furious rabies [20] . In response to calls for controlling dog bites and rabies , the state government formed a state level rabies coordination committee in 2008 to develop and manage a multisectoral response to dog bites and rabies in the state . This was the first time a large scale population level rabies control intervention was implemented in a large state in India [17] . As described in Table 1 , the human interventions consisted of ensuring availability of anti-rabies vaccine at all government-run health facilities in the state as well as promoting awareness about rabies control across the state . Rabies antibody was not provided universally due to perceived high costs . The animal interventions involved outsourcing of ABC-AR operations to private veterinarians; dog catching operations were handled by local animal welfare organizations in selected urban areas of the state . ABC-AR was conducted throughout the year as specified in the guidelines of Animal Welfare Board of India [21]; vaccination-only strategies , whether parenteral or oral , were not considered . The interventions were supposed to be implemented in a continuous fashion throughout the year and not conducted in a campaign mode . The animal and human sector interventions were implemented by different departments and coordinated at the state and district levels through formal multi-stakeholder coordination mechanisms [17] .
Based upon the existing interventions in Tamil Nadu [17] , it was assumed that the entire population ( rural as well as urban ) would be covered by the expanded intervention . Costs were estimated for two combinations of interventions . Based upon the existing intervention model , the first set of interventions consisted of increased surveillance and awareness , in addition to provision of anti-rabies vaccine ( ARV ) to all patients reporting dog bites at public health facilities . The second combination of interventions involved an additional component of antibody administration to patients with severe dog bites in addition to the ARV . Based upon the feedback received from local program managers [24] , it was assumed that dog bite cases that report at peripherally located and low-throughput health centres would be provided with rabies vaccine through the easier intramuscular route , while those that report at high-throughput hospitals with better trained personnel would be provided vaccination through the intradermal route . The procurement costs of intradermal and intramuscular vaccine formulations ( having different vial sizes ) and antibodies were estimated from the state level procurement records [25] and market data , respectively . A standard 30% wastage rate was assumed for both the vaccine formulations in the absence of specific reference points . A lesser wastage rate of 15% was used for the antibody . The annual number of outpatient visits for dog bites was calculated from the monthly dog bite visits reported by the state disease surveillance system over a twenty month period from January 2008 to August 2009 . This was divided by the expected number of hospital visits for each dog bite case , to arrive at the annual number of dog bites in the state . While the national guidelines [26] recommend vaccination only for category 2 and category 3 dog bites , in practice , the vaccine was being administered to all reported dog bite cases , which was factored into our analysis . The proportion of dog bites categorised as ‘severe’ and requiring antibodies was assumed to be 63% , using estimates from other national studies [27] . Based upon the feedback received from program managers , some program administration costs were included to cover expenditure related to awareness generation , training and surveillance related activities . The then-prevalent model of ABC-AR was selected as one of the intervention strategies . Parenteral vaccination using teams of dog-catchers and oral vaccination were selected as hypothetical intervention scenarios to determine the extent to which costs could be reduced by less resource-intensive exercises . Using dog population density figures from the livestock census [28] , the number of animal sheds ( having capacity for 30–45 animals ) required to cater to 100 , 000 human populations were calculated . The fixed and recurrent costs were then calculated for every 100 , 000 population . The costs for animal interventions were sourced from state program guidelines and adjusted for inflation . Subsequently , differential costing was conducted to include additional stay and veterinary fees for operating on female dogs . More vehicles were assumed to be required in rural areas because of the larger distances to be covered . Therefore , increased capital and fuel costs were considered for dog shelters in rural areas . All capital costs were depreciated over 5 years . Costing for dog catchers' and ambulance drivers' time was done on a monthly basis using state salary norms . Animal census costs were also included as an annual exercise and estimated accordingly . A base case scenario was constructed for each of the five different combinations of human and animal interventions using the existing or most likely estimates of key input parameters . The values of input data for our analysis were sourced from our review of program documents , published research literature and from our personal observations in the state . A sensitivity analysis was conducted by varying the values of principle input factors . More than 224 , 000 scenarios of animal and human interventions were tested . The values of input parameters for the base case and alternative scenarios have been described in Supplementary Files . These were refined based upon the feedback received from experts at two different national consultations of Indian rabies experts organized in 2011 [29] and 2013 [30] . Rabies control is a long term proposition , requiring sustained levels of high coverage of interventions in the animal populations [10] . Accordingly , in addition to estimating the annual costs on the basis of a one-time assessment , we also assessed the long term implications of the animal sector interventions . Given the limited data on the impact of parenteral animal vaccination campaigns in mixed ecological settings such as India , we used data from an Indian study [7] describing the impact of dog population management interventions to assess the long term implications of the animal sector interventions . We projected costs of four interventions—ABC-AR , injectable vaccination , oral vaccination , and a hypothetical intervention coupling injectable vaccination with injectable contraception for 20 years based on 2012 costs . For interventions involving contraception , a decrease in the dog population was estimated from a dog demographic model used earlier in India [7] . The model estimated the change in total stray dog population and the proportion of sterile dogs over a 20-year period given a sterilization rate of 62–87% in several mark-recapture study areas in Jodhpur city , from 2005 to 2007 . Since no other dog demographic models in the Indian context were available , we estimated the total number of stray dogs and its proportion that would be sterile for each year in Tamil Nadu assuming a similar setting and level of coverage . To project costs for future years , annualized capital costs ( for 5-year depreciation ) were assumed to be constant over 20 years , and recurrent costs were scaled to the projected dog population size in each year . Recurrent costs were calculated separately for the unsterile ( requiring vaccination and sterilization ) and sterile ( requiring only vaccination ) dog populations . Interventions which did not involve sterilization assumed a constant dog population . For the hypothetical injectable vaccination and contraception intervention , the additional cost of the injectable contraceptive was assumed to be negligible and the initial cost in 2012 was assumed to be the same as the cost of the injectable vaccine intervention alone in the base case scenario . The study is based upon one-time costs data collected from state programme managers . Therefore the analysis only considers those human cases that were reported to the public health surveillance system . This is likely to be an underestimate . Moreover , there is limited data on the completion of treatment; and it is possible that a small portion of patients might not complete their treatment , leading to a further underestimate of dog bite incidence rate . Data on categorization of dog bites , dog bite burden among animals and dog ecology is limited . In the absence of more data , the upper and lower bounds of the input parameters were taken from a range of sources , including expert opinion , summarised in Supplementary Files 1 & 2 . In the absence of longitudinal data , we used dog demographic projections from an Indian study [7] to estimate the long term resource requirements for different rabies control interventions . However , there is limited information related to the reliability of these findings in rural areas and other parts of India . While recent studies recommend canine vaccination in annual campaigns having coverage exceeding 60%[9] , the current analysis estimates the cost of an year-long continuous routine vaccination strategy which is likely to provide a conservative estimate of likely costs . More long-term efficacy studies for different interventions are required to better comment upon their cost effectiveness .
The annual costs of providing post exposure prophylaxis with antibodies for severe dog bites for Tamil Nadu was calculated to be $ 2 . 2 million ( Table 2 ) . This was more than three times the costs of rolling out a vaccine-only program and translates into costs of $ 11 and $ 3 , respectively for each dog bite patient vaccinated . Using base case scenarios , the annual costs of implementing ABC-AR , Injectable vaccination and oral vaccine programmes were calculated to be $ 44 million , $ 23 million and $ 11 million , resulting in each dog's vaccination costing $ 22 , $ 11 and $ 5 , respectively . On varying the key input parameters , we found that the costs of the human interventions ranged from $3–$82 million , while the costs for animal intervention ranged from $9–$98 million annually . In order to compare the relative effects of different cost components on individual set of interventions , a tornado chart was prepared ( Figure 1 & Figure 2 ) centred around the costs of the base case scenario for each combination of interventions . The value of each cost component was varied to its upper and lower bounds and the impact on the total program cost charted as red and blue bars , respectively . In the case of human interventions ( Vaccine only and Vaccine+Antibodies ) , health seeking patterns , cost and wastage rates of vaccine and antisera and the burden of dog bites were found to be the major cost drivers causing the greatest fluctuations in the cost of the program . In relation to the other drivers , the use of intradermal versus intramuscular vaccine regimes did not greatly influence program costs . Antibody procurement comprised around 70% of total costs of the human sector interventions , followed by vaccine procurement costs and training & health promotion costs . In case of animal-sector interventions , the dog population and number of dog catchers required per team appear to be important drivers of ABC-AR programme costs . On the other hand , vaccine costs have greater role to play in influencing costs of vaccination-only programmes . Sex distribution of dogs does not affect total program costs in the long term , even for ABC-AR in which different surgical procedures are required for male and female dogs . Assuming an average sterilization rate of 62–87% , the dog population size in Tamil Nadu would be expected to decrease by 70% over a 20 year period from an estimated 2 , 022 , 055 dogs in 2012 to 615 , 408 in 2032 ( Supplementary File 4 ) . The proportion of sterilized dogs would stabilize at 80% such that the number of dogs needing ABC would decrease by 94% from 2 , 022 , 055 in 2012 to 123 , 082 dogs in 2032 . Projected costs for ABC-AR , injectable vaccination , oral vaccination , and injectable vaccination-cum-contraception are shown in Figure 3 . While costs are highest for ABC-AR in 2012 , the cost drops quickly and is lower than that for injectable vaccination and similar to that for oral vaccination by 2032 ( Table 3 ) . Total costs over the 20-year period are highest for injectable vaccination and are comparable for oral vaccination and ABC-AR . The hypothetical joint injectable vaccination and contraception intervention would result in the lowest cost by 2032 and the lowest total cost .
Using base case scenarios , scaling up the existing animal interventions ( ABC-AR ) in Tamil Nadu would require 20 to 65 times the funds required for scaling up human post exposure prophylaxis alone . Moreover , a combination of human Post-exposure prophylaxis with ABC-AR would cost over 2 . 1% of the annual budgetary allocations for the departments of health , animal husbandry and municipal administration together in Tamil Nadu [32] . This is an important lesson for the proposed national rabies control programme in India which is currently structured around financing ABC-AR operations across selected cities [33] . Recent discussions have advocated parenteral vaccination of canines as a first step towards elimination of rabies [1] , [9] . This would require a high level of coverage ( >60% ) costing 27% of the annual budget of the state department of animal husbandry , the likely implementing agency for such an intervention , and would need to be sustained continuously for multiple years or even decades . Due to the challenges of achieving high vaccination coverage even among humans [34] and the high costs of existing animal interventions described above , the policymakers are unlikely to commit to a comprehensive rabies control programme yet . A more favourable case for rabies control among canines could be made by developing newer animal interventions that are not only efficacious but also affordable and effective , such as an inexpensive canine injectable contraceptive cum vaccine . In the absence of an intervention that promises long term sustainability , it is likely that ad hoc measures like post exposure vaccinations to economically productive animals continue . Antibodies were not made universally available for human vaccines because of the costs involved . Our calculations show that the costs of a combined ( antibody plus vaccine ) rabies programme would be three times the costs of vaccine only intervention and is likely to cost an additional expenditure of $ 1 . 5 million ( Rs . 82 . 5 million ) annually . Given the costs of different vaccine formulations in Tamil Nadu , choosing intradermal over intramuscular vaccine regimen is likely to result in annual savings of $ 13 , 000 ( Rs 700 , 000 ) only . This relatively small amount should not be a deterrent to state public health programme managers in choosing a vaccine regimen that is more appropriate to the clinical setting and qualifications of their staff [24] . Rabies control efforts in Tamil Nadu seem a costly proposition as they are currently structured in the state . This would necessarily require high levels of technical , political and financial commitments before the government chooses to embark upon a long-term rabies control strategy . Given recent recognition of the need for a national rabies control programme in India by the National Centre for Disease Control [33] and the FAO/WHO/OIE tripartite statement on inclusion of rabies as an ‘entry point’ for demonstrating zoonoses control efforts at the global level [1] , it is important that these discussions adopt a long term perspective and take local complexities into account before developing a national or a global rabies elimination strategy . | Rabies is a fatal viral disease . It is transmitted mostly through dog bites in greater parts of Asia and Africa . It is primarily a disease of the poorer population groups with children being the most vulnerable . Control of rabies among humans therefore requires interventions in the animal as well as the human sectors . Animal sector interventions include vaccination accompanied with or without sterilization of dogs . Human interventions are limited to individual vaccination following dog bites . We estimated the costs to the government of rolling out animal as well as human sector interventions across an entire state having a human population of 72 million . We also estimated the major drivers influencing program costs and the implications to the government of adopting such a strategy over a long time . We found that the animal sector interventions were many times more costly than the most expensive human interventions . We also found that in the absence of dog population control measures , it will require substantial financial commitment on the part of the government to be able to invest in dog vaccination strategies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"rabies",
"socioeconomic",
"aspects",
"of",
"health",
"disease",
"ecology",
"social",
"and",
"behavioral",
"sciences",
"economic",
"models",
"health",
"economics",
"veterinary",
"diseases",
"cost",
"models",
"global",
"health",
"z... | 2014 | Costs Analysis of a Population Level Rabies Control Programme in Tamil Nadu, India |
Previous studies have demonstrated that Marburg viruses ( MARV ) and Ebola viruses ( EBOV ) inhibit interferon ( IFN ) -α/β signaling but utilize different mechanisms . EBOV inhibits IFN signaling via its VP24 protein which blocks the nuclear accumulation of tyrosine phosphorylated STAT1 . In contrast , MARV infection inhibits IFNα/β induced tyrosine phosphorylation of STAT1 and STAT2 . MARV infection is now demonstrated to inhibit not only IFNα/β but also IFNγ-induced STAT phosphorylation and to inhibit the IFNα/β and IFNγ-induced tyrosine phosphorylation of upstream Janus ( Jak ) family kinases . Surprisingly , the MARV matrix protein VP40 , not the MARV VP24 protein , has been identified to antagonize Jak and STAT tyrosine phosphorylation , to inhibit IFNα/β or IFNγ-induced gene expression and to inhibit the induction of an antiviral state by IFNα/β . Global loss of STAT and Jak tyrosine phosphorylation in response to both IFNα/β and IFNγ is reminiscent of the phenotype seen in Jak1-null cells . Consistent with this model , MARV infection and MARV VP40 expression also inhibit the Jak1-dependent , IL-6-induced tyrosine phosphorylation of STAT1 and STAT3 . Finally , expression of MARV VP40 is able to prevent the tyrosine phosphorylation of Jak1 , STAT1 , STAT2 or STAT3 which occurs following over-expression of the Jak1 kinase . In contrast , MARV VP40 does not detectably inhibit the tyrosine phosphorylation of STAT2 or Tyk2 when Tyk2 is over-expressed . Mutation of the VP40 late domain , essential for efficient VP40 budding , has no detectable impact on inhibition of IFN signaling . This study shows that MARV inhibits IFN signaling by a mechanism different from that employed by the related EBOV . It identifies a novel function for the MARV VP40 protein and suggests that MARV may globally inhibit Jak1-dependent cytokine signaling .
Filoviruses , which include the genera ebolavirus ( EBOV ) and marburgvirus ( MARV ) , are enveloped negative-strand RNA viruses that cause highly lethal hemorrhagic fever in humans and in non-human primates . The ability of filoviruses to counteract innate antiviral responses of the host , particularly the IFNα/β response is thought to promote uncontrolled virus replication in vivo and thereby contribute to development of severe disease [1] . The IFNs , which include IFNα/β and IFNγ , are antiviral cytokines . IFNα/β are members of a family of proteins that interact with the same ubiquitous receptor to trigger innate antiviral defense mechanisms and promote adaptive immunity [2] . IFNγ also triggers expression of antiviral genes , however , its major function is to modulate adaptive immune responses [3] . IFNα/β signaling results in the tyrosine phosphorylation and activation of the Janus kinases Jak1 and Tyk2 . These phosphorylate STAT2 and STAT1 , which in turn heterodimerize and associate with interferon regulatory factor 9 ( IRF9 ) to form a complex that is translocated into the nucleus to activate genes involved in antiviral response ( reviewed in [4] ) . IFNγ signaling activates Jak1 and Jak2 , resulting in tyrosine phosphorylation of STAT1 . This induces STAT1 homodimerization and translocation to the nucleus such that IFNγ dependent gene expression is induced ( reviewed in [4] ) . Of note , Jak1 , a kinase involved in multiple cytokine signaling pathways , is critical for both IFNα/β and IFNγ signaling . For example , in cells lacking Jak1 , IFNα/β fails to trigger STAT1 or STAT2 tyrosine phosphorylation and Tyk2 tyrosine phosphorylation is greatly reduced or eliminated [5] , [6] . Similarly , in cells lacking Jak1 , IFNγ fails to trigger Jak1 , Jak2 or STAT1 tyrosine phosphorylation [5] , [7] , [8] . Filovirus genomes encode seven structural proteins . Four of these proteins , the nucleoprotein ( NP ) , the viral proteins VP35 and VP30 , and the L protein are tightly associated with the RNA genome , form the nucleocapsid and mediate replication and transcription ( reviewed in [9] ) . Besides its function as polymerase cofactor , VP35 acts as an inhibitor of antiviral pathways ( see below ) . Two of the filovirus structural proteins are matrix proteins , VP40 , the functional equivalent of the matrix ( M ) proteins of other non-segmented negative-stand RNA viruses , and the minor matrix protein VP24 that is unique to filoviruses . As a peripheral membrane protein VP40 is located at the inner side of the virion membrane . It is critical for viral budding and interacts with cellular proteins involved in vesicle formation to facilitate virus release [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] . The minor matrix protein VP24 is involved in nucleocapsid formation and assembly [19] , [20] , [21] , [22] , [23] . EBOV VP24 plays a crucial role in host tropism [24] , [25] and is able to counteract the type I IFN response ( see below ) . Filoviruses possess a single surface protein , the type I transmembrane glycoprotein GP that mediates attachment to target cells and virus entry . Besides EBOV VP35 and VP24 , EBOV GP is the third filoviral protein known to interfere with antiviral cellular functions [26] . Among filoviruses , IFN evasion strategies have been most thoroughly explored for EBOVs . The EBOV species Zaire ebolavirus ( ZEBOV ) suppresses production of IFNα/β and inhibits cellular responses to IFNα/β and IFNγ [27] , [28] , [29] , [30] . Inhibition of IFNα/β production appears to be mediated by the VP35 protein [31] , [32] , whereas cellular responses to IFNα/β and IFNγ are blocked by the EBOV VP24 protein [33] , [34] . EBOV VP24 prevents the IFN-induced nuclear accumulation of tyrosine phosphorylated STAT1 . This results in inhibition of IFN-induced gene expression and blocks the antiviral effects of IFNs . The inhibition of STAT1 nuclear accumulation is mediated by interaction of VP24 with NPI-1 subfamily of karyopherin α proteins that normally transport dimerized phospho-STAT1 to the nucleus [33] , [34] . MARVs have a genome organization similar to EBOVs , but they are phylogenetically distinct from EBOVs [35] . Despite their similar genomic organization , morphology and the similarity of MARV versus EBOV induced disease , several biological differences between the viruses have been noted , such as differences in their transcription strategies [36] , in the structure of their replication promoters [37] , the use of mRNA editing to express the surface glycoprotein by EBOVs but not MARVs [38] , [39] and differences in the protein requirement for nucleocapsid formation [40] , [41] . In terms of the capacity of EBOV and MARV to counteract host IFN responses , microarray analyses suggest that ZEBOV and MARV each efficiently suppress host IFN responses , and each virus effectively inhibits cellular responses to exogenously added IFNα [30] . However , examination of the phosphorylation status of STAT1 following addition of IFNα to infected cells revealed an intriguing difference between ZEBOV and MARV . While ZEBOV did not inhibit the IFNα-induced tyrosine phosphorylation of STAT1 , MARV infection resulted in an inhibition of both STAT1 and STAT2 tyrosine phosphorylation [30] . The present study demonstrates that MARV infection inhibits not only IFNα/β but also IFNγ and Jak1-dependent IL-6 signaling . Further , the MARV protein mediating these effects has been identified . We show that expression of the MARV matrix protein VP40 is sufficient to block IFN and IL-6 signaling pathways . Experiments in which either Jak1 or Tyk2 are over-expressed suggest that MARV VP40 targets Jak1 function . These observations identify an important difference in the biology of MARV and EBOVs , identify a novel function for a negative-strand RNA virus matrix protein and suggest that MARV may inhibit multiple Jak1-dependent cytokine signaling pathways .
Previous studies demonstrated that tyrosine phosphorylation of STAT1 and STAT2 is strongly reduced in MARV- but not in ZEBOV-infected Huh-7 cells treated with IFNα [30] . To confirm this observation and to determine whether MARV inhibition extends to other Jak-STAT signaling pathways , the impact of MARV infection on IFNα-induced STAT1 and STAT2 phosphorylation and on IFNγ-induced STAT1 phosphorylation was compared . As reported , MARV but not EBOV inhibited phosphorylation of endogenous STAT1 and STAT2 induced by IFNα ( Fig . 1A ) . MARV also inhibited IFNγ-induced STAT1 phosphorylation , whereas EBOV did not ( Fig . 1B ) . For these studies , immunofluorescence analyses were performed in parallel to confirm that more than 95% of cells were infected with either virus ( data not shown ) . These data show that MARV not only blocks type I but also type II IFN signaling by interfering with an early step of the Jak-STAT signaling cascade . Since previous studies indicated that the nuclear translocation of phosphorylated STAT1 is inhibited in EBOV-infected cells [33] , [34] , we examined the cellular localization of STAT1 in MARV-infected cells by immunofluorescence ( Fig . 1C ) . As expected , STAT1 was translocated into the nucleus in non-infected cells treated with IFNα ( left panels , red staining ) , whereas IFNα-induced translocation was inhibited in ZEBOV-infected cells ( right panels , infected cells shown in green ) . Please note that a single non-infected cell in the ZEBOV infection panel showed nuclear accumulation of STAT1 . Nuclear translocation of STAT1 was also blocked in MARV-infected cells treated with IFNα ( middle panels ) . Taken together , these results highlight a fundamental difference in the mechanisms by which MARV and EBOV counteract innate immune responses . Since our data suggested that MARV infection leads to the inhibition of IFN-induced STAT phosphorylation , we next sought to determine the activation status of Jak1 and Tyk2 , the Janus kinases involved in IFNα-induced phosphorylation of STAT proteins . Huh-7 cells were infected with MARV or ZEBOV , treated with IFNα and the phosphorylation state of endogenous Jak1 and Tyk2 was analyzed by western blot analysis . As shown in Figure 2A , both kinases were phosphorylated in ZEBOV-infected cells in response to IFNα , although phosphorylation of Jak1 was less pronounced compared to non-infected cells ( Fig . 2A , compare lane 2 and 4 ) . However , only background levels of Jak1 phosphorylation were detectable and Tyk2 phosphorylation was completely blocked in MARV-infected cells treated with IFNα ( Fig . 2A , lane 6 ) . From this we concluded that the inhibition of the Jak-STAT-signaling pathway by MARV takes place upstream of Jak phosphorylation or directly at the Janus kinases . As part of the innate immune response , the Jak-STAT signaling cascade acts as a first line of defense to prevent viral infections . Therefore , we determined at which time point of the MARV replication cycle the observed inhibition of Jak activation occurs . Further , we asked whether live virus and viral replication are needed to antagonize IFN signaling . Huh-7 cells were infected with live MARV or UV-inactivated MARV , treated with IFNα , harvested at different time points post-infection ( p . i . ) and subjected to western blot analysis to determine the phosphorylation state of Tyk2 . While Tyk2 was still efficiently phosphorylated in MARV-infected and IFN-treated cells at 1 hour and 2 . 5 hours p . i . , respectively , near complete inhibition of Tyk2 phosphorylation was achieved at 4 hours p . i . ( Fig . 2B ) . A single MARV replication cycle takes approximately 21 hours in Vero E6 cells [42] . Thus , it can be concluded that the observed antagonistic effect occurs early in infection . Additionally , since MARV infection did not lead to the inhibition of Tyk2 phosphorylation at time points earlier than 4 hours p . i . , it is assumed that binding of MARV to its receptor does not trigger its IFN antagonist function . Interestingly , infection of cells with UV-inactivated MARV prior to IFNα treatment did not lead to the inhibition of Tyk2 phosphorylation ( Fig . 2B ) , supporting the assumption that receptor binding does not play a role in the MARV-specific inhibition of the IFN signaling cascades . In addition , these data indicate that intracellular virus replication is required for the observed antagonistic effects . To examine whether MARV indirectly inhibits Jak1 phosphorylation via protein tyrosine phosphatases ( PTPs ) , we treated MARV-infected and IFN-treated cells with different PTP inhibitors prior to IFNα stimulation . Besides an inhibitor against PTP1B , which specifically dephosphorylates Tyk2 and Jak2 [43] , we tested the broad acting phosphatase inhibitor sodium orthovanadate . Our results show that even in the presence of PTP inhibitors Tyk2 phosphorylation was inhibited in MARV-infected cells ( Fig . 2C ) , suggesting that the observed inhibitory effects do not depend on active cellular PTPs . To identify the viral protein mediating the antagonistic effects observed in MARV-infected cells , individual EBOV or MARV proteins were assessed for their capacity to counteract the antiviral effects of IFNβ ( Fig . 3A ) . Vero cells were transfected with expression plasmids; one day post-transfection the cells were either mock-treated or treated overnight with IFNβ , and the cells were then infected with a Newcastle disease virus that expresses GFP ( NDV-GFP ) . Since NDV is IFN-sensitive , GFP expression in these cells provides a measure of virus replication , and suppression of GFP expression provides a read-out for the antiviral effects of IFNβ . While empty vector ( pCAGGS ) -transfected , mock-treated cells permitted NDV-GFP replication , IFNβ-treated , empty vector-transfected cells , in contrast , greatly suppressed GFP expression ( Fig . 3A , panels 1 and 2 ) . As previously described , expression of Nipah virus W protein , or ZEBOV VP24 , known inhibitors of IFN signaling , rescued replication of NDV-GFP in IFNβ-treated cells [33] , [44] ( Fig . 3A , panel 3 and 4 ) . Surprisingly , MARV VP24 did not detectably counteract the antiviral effects of IFNβ ( Fig . 3A , panel 6 ) . In fact , the only MARV protein tested that clearly permitted NDV-GFP replication in IFNβ-treated cells was the major matrix protein VP40 ( Fig . 3A , panel 9 ) . In contrast , the homologous ZEBOV protein , ZEBOV VP40 , did not support NDV-GFP replication ( Fig . 3A , panel 5 ) . To confirm the finding that MARV VP40 antagonizes IFN signaling , we analyzed the intracellular distribution of STAT2 in cells transiently expressing MARV or EBOV proteins VP35 , VP24 , or VP40 ( Fig . 3B ) . Since it has been shown by Brzozka et al . [45] that rabies virus phosphoprotein ( P ) efficiently blocks the nuclear translocation of STAT2 into the nucleus , P was used as a positive control ( Fig . 3B ) . Cells transfected with empty vector served as a negative control . While expression of either MARV VP40 or ZEBOV VP24 led to a significant inhibition of STAT2 accumulation in the nucleus , none of the other tested filoviral proteins including ZEBOV VP40 and MARV VP24 was able to inhibit nuclear translocation of STAT2 in response to IFNα ( Fig . 3B ) . From this , we concluded that MARV VP40 is the viral protein interfering with IFN signaling . Our results obtained with infected cells clearly show that MARV infection leads to the inhibition of STAT and Jak phosphorylation , whereas ZEBOV infection does not . To assess the impact of MARV VP40 on IFN-induced signaling in the absence of other viral proteins , STAT1-GFP or STAT2-GFP were co-transfected into Huh-7 cells with empty vector or with plasmids expressing ZEBOV VP40 , ZEBOV VP24 , MARV VP40 or MARV VP24 . The phosphorylation state of the STAT proteins in response to IFNα/β ( Fig . 4A ) and IFNγ ( Fig . 4B ) was examined by western blot analysis . Expression of the Langat virus NS5 protein ( LGTV NS5 ) , a protein previously demonstrated to inhibit STAT1 and STAT2 tyrosine phosphorylation [46] served as a control . Following addition of IFNα/β to transfected Huh-7 cells , MARV VP40 inhibited the IFNα/β -induced tyrosine phosphorylation of either STAT1-GFP ( Fig . 4A , left panel ) or STAT2-GFP ( Fig . 4A , right panel ) . In contrast , the ZEBOV VP40 , ZEBOV VP24 and MARV VP24 proteins failed to inhibit STAT1 or STAT2 tyrosine phosphorylation ( Fig . 4A ) . Relative to empty vector-transfected cells , LGTV NS5 reduced the IFNγ-induced phosphorylation of STAT1-GFP ( Fig . 4B ) , but ZEBOV VP24 , MARV VP24 and ZEBOV VP40 failed to inhibit STAT1 phosphorylation . In contrast , MARV VP40 expression led to a substantial reduction in IFNγ-induced STAT1 tyrosine phosphorylation ( Fig . 4B ) . Next , we analyzed the impact of MARV VP40 on the phosphorylation of Janus kinases in cells treated with IFNα/β or IFNγ . 293T cells were transfected with empty vector or plasmids that express LGTV NS5 , ZEBOV VP24 , ZEBOV VP40 , MARV VP24 or MARV VP40 , treated with IFNα/β and analyzed for phosphorylation of endogenous Jak1 and Tyk2 . MARV VP40 inhibited the IFNα/β-induced tyrosine phosphorylation of both kinases ( Fig . 4C ) . Interestingly , none of the other expressed proteins including LGTV NS5 detectably blocked Jak1 phosphorylation . Although Tyk2 phosphorylation was also reduced by LGTV NS5 and to a lesser extent by ZEBOV VP24 , this reduction was less pronounced compared to cells expressing MARV VP40 ( Fig . 4C ) . Similar results were obtained in cells treated with IFNγ . Inhibition of Jak1 and Jak2 phosphorylation in response to IFNγ treatment was only observed in cells expressing MARV VP40 ( Fig . 4D ) . Taken together , these results clearly confirm that MARV not only uses a different mechanism than EBOV to block IFN signaling , but an alternate viral protein carries out this function . To address the functional significance of the observed inhibition , the impact of MARV VP40 on IFNβ and IFNγ-induced transcription was assessed by reporter gene assay ( Fig . 5 ) . Two reporter constructs were used . One , activated by IFNα/β , possesses an ISG54 promoter and contains an interferon stimulated response element ( ISRE ) . The second , activated by IFNγ , possesses three gamma activated sequence ( GAS ) elements . 293T cells were transfected with either reporter plus expression plasmids for MARV VP40 or , as controls , MARV VP24 and ZEBOV VP24 . To control for non-specific or cytotoxic effects of the viral proteins , the results of these assays were normalized to a co-transfected constitutively-expressed Renilla luciferase reporter plasmid . MARV VP40 and ZEBOV VP24 inhibited ISG54 promoter activation , whereas MARV VP24 failed to inhibit its activation ( Fig . 5A ) . Similarly , MARV VP40 inhibited IFNγ-induced gene expression , consistent with its capacity to block IFNγ activation of STAT1 . As previously described , ZEBOV VP24 inhibited IFNγ-induced gene expression [33] , but MARV VP24 did not inhibit gene expression in this assay ( Fig . 5B ) . The impact of MARV VP40 upon IFNγ-induced production of the 10 kDa interferon-gamma-induced protein ( IP-10 ) , an immune cell chemoattractant protein secreted by several cell types in response to IFNγ , was also assessed . Human umbilical vein endothelial cells ( HUVECs ) were transfected with the indicated expression plasmids , treated with IFNγ , and cell supernatants were tested for the presence of IP-10 by ELISA . MARV VP40 and , to a lesser extent , ZEBOV VP24 inhibited IP-10 expression , whereas MARV VP24 did not ( Fig . 5C ) . To assess the specificity of this effect and exclude cell death or disruption of membrane signaling components , a similar assay was performed testing the impact of viral protein expression on TNFα-induced secretion of IL-8 which is NF-κB-mediated [47] , [48] . None of the expressed proteins , including MARV VP40 , detectably affected IL-8 production ( Fig . 5D ) . Therefore the impact of MARV VP40 seems to be specific for Jak-STAT signaling and does not extend to the induction of NF-κB-mediated signaling [47] . Interestingly , our observations are reminiscent of the phenotype seen in Jak1-deficient cells , where the absence of Jak1 results in loss of Jak1 , Tyk2 , STAT1 and STAT2 tyrosine phosphorylation in response to IFNα/β and loss of Jak1 , Jak2 and STAT1 tyrosine phosphorylation in response to IFNγ [5] , [7] , [8] . To examine whether the observed inhibitory effect of MARV on IFN signaling extends to other , non-IFN , Jak-STAT signaling pathways , we next analyzed the IL-6-induced activation of STAT1 and STAT3 in MARV-infected cells and cells expressing VP40 . IL-6 was chosen because , in Jak1-deficient cells , IL-6-induced STAT1 phosphorylation was absent , and STAT3 phosphorylation was greatly reduced [7] . Huh-7 cells were infected with MARV , treated with IL-6 at 24 hours p . i . and cell lysates were subjected to western blot analysis . As shown in Figure 6A , phosphorylation of endogenous STAT1 was not detectable and STAT3 phosphorylation was strongly diminished in MARV-infected , IL-6 treated cells , reflecting the phenotype of Jak1-deficient cells [7] . Similar results were obtained with transfected Huh-7 cells . MARV VP40 inhibited the IL-6 induced tyrosine phosphorylation of STAT1-GFP to undetectable levels , and FLAG-STAT3 tyrosine phosphorylation was highly reduced ( Fig . 6B and C ) . In contrast , ZEBOV VP40 , ZEBOV VP24 and MARV VP24 did not inhibit phosphorylation of either STAT1 or STAT3 ( Fig . 6B and C ) . To further assess the capacity of MARV VP40 to target the function of Jak1 , MARV VP40 , ZEBOV VP40 , MARV VP24 or ZEBOV VP24 were co-transfected with expression plasmids for STAT2-GFP and either HA-tagged Jak1 or HA-tagged Tyk2 . Over-expression of Janus kinases leads to their tyrosine phosphorylation [49] and to the phosphorylation of STAT proteins ( Fig . 7 ) . First , we determined the phosphorylation state of HA-Jak1 and STAT2-GFP in transfected cells by western blot analysis ( Fig . 7A ) . While MARV VP40 completely inhibited the phosphorylation of over-expressed HA-Jak1 and consequently , the phosphorylation of STAT2-GFP , ZEBOV VP40 , ZEBOV VP24 and MARV VP24 did not show any inhibitory effect ( Fig . 7A ) . Extending this observation , HA-Jak1 over-expression also led to tyrosine phosphorylation of endogenous STAT1 and STAT3 , and this was inhibited by MARV VP40 but not by the other tested viral proteins ( Fig . 7B ) . In contrast , none of the expressed filovirus proteins , including MARV VP40 , detectably inhibited tyrosine phosphorylation of over-expressed HA-Tyk2 or Tyk2-induced STAT2-GFP phosphorylation ( Fig . 7C ) . Further titration of HA-Tyk2 expression was performed , and phosphorylation of endogenous STAT1 was monitored ( Fig . 7D ) . Two-fold dilutions of HA-Tyk2 plasmid were transfected with either empty vector or MARV VP40 plasmid . When 500 ng of Tyk2 plasmid was transfected , less phospho-Tyk2 was detected in the MARV VP40-expressing cells than in cells receiving empty vector . Similarly , levels of phosphorylated endogenous STAT1 were decreased in the presence of MARV VP40 . However , the total levels of HA-Tyk2 were also decreased in the presence of MARV VP40 in these samples ( Fig . 7D ) . Therefore the bands were quantified by densitometry and the ratio of phosphorylated Tyk2 to total Tyk2 was calculated for each sample . In all samples the ratios were in the range of 0 . 85 to 1 . 05 , suggesting that the decreased levels of phospho-Tyk2 were due to reduced total levels of Tyk2 ( data not shown ) . These data support a model where MARV VP40 targets Jak1 function but do not completely exclude the possibility that MARV VP40 has a modest capacity to inhibit Tyk2 . MARV VP40 contains a late domain ( PPPY ) , positioned from residues 16–19 , that mediates VP40 interaction with the cellular protein Tsg101 , a component of the ESCRT I machinery , and contributes to its budding function [18] . To determine whether this late domain is critical for MARV VP40 inhibition of signaling , the 16-PPPY-19 motif was mutated to 16-AAAA-19 ( M40-AAAA ) . Relative to wild-type EBOV VP40 or wild-type MARV VP40 , M40-AAAA exhibited greatly reduced budding , in the form of virus-like particles ( VLPs ) , from transfected 293T cells , despite comparable expression in the whole cell extracts ( Fig . 8A ) . As expected , a separately expressed GFP was not released into the cell medium ( Fig . 8A ) . When tested for its ability to suppress IFNα/β-induced signaling , the mutant suppressed STAT1 phosphorylation comparably to either LGTV NS5 or wild-type MARV VP40 ( Fig . 8B ) . The mutant also suppressed IFNα/β-induced activation of the ISG54 promoter comparably to wild-type MARV VP40 ( Fig . 8C ) . Therefore we conclude that the MARV VP40 late domain is not required for inhibition of signaling .
Previous studies have shown that both members of the filovirus family , MARV and EBOV , impair cellular responses to IFNs [30] , [33] , [34] , [50] . While ZEBOV blocks the nuclear accumulation of tyrosine-phosphorylated STAT1 [33] , [34] , the present study demonstrates that MARV has evolved a different mechanism to counteract IFN signaling . We show that MARV inhibits the IFNα-induced tyrosine phosphorylation of not only STAT1 and STAT2 but also of the upstream kinases Jak1 and Tyk2 . This inhibition prevents the IFN-induced nuclear accumulation of STAT1 and STAT2 . Further , MARV infection inhibits the IFNγ-induced tyrosine phosphorylation of STAT1 . The inhibition extends even beyond the IFNα/β and IFNγ signaling pathways to another Jak1 dependent signaling pathway , the IL-6 pathway , where the phosphorylation of STAT1 and STAT3 was inhibited . Significantly , the study also identifies a single MARV protein , the matrix protein VP40 , sufficient to mediate these inhibitory effects , whereas ZEBOV-induced inhibition of IFN signaling is mediated by VP24 [33] . Emphasizing the specificity of the inhibitory function for MARV VP40 , neither ZEBOV infection nor ZEBOV VP40 expression impairs Jak or STAT phosphorylation . Moreover , MARV VP24 , including VP24s corresponding to the Musoke strain and the Angola strain , which caused an outbreak with a very high fatality rate [51] , did not detectably inhibit IFNα/β-induced gene expression ( Fig . 5B and data not shown ) . Musoke MARV VP24 was also unable to inhibit IFNα/β- , IFNγ- or IL-6-induced phosphorylation of Jaks or STATs . The striking differences in the strategies employed by filoviruses to block IFN signaling may have been driven by the different evolutionary paths taken by Marburg and Ebola viruses . Bayesian analysis of genome sequence differences indicates that Ebola and Marburg viruses diverged from a common ancestor several thousands of years ago ( S . T . Nichol , personal communication ) . Evolution in and adaptation to different host species might account for different immune evasion mechanisms . So far , there is only limited information available about the natural host spectrum of filoviruses . Various species of African fruit bats were found to be seropositive or RT-PCR-positive for EBOV [52] , [53] , however , as yet Ebola viruses have not been isolated from bats . In contrast , Towner and colleagues reported the successful isolation of MARV from the Egyptian fruit bat Rousettus aegyptiacus [54] . Since this bat species is also discussed as a potential reservoir for EBOV [53] , it remains unclear if Marburg and Ebola viruses differ in their host tropism . Recently , the Asian EBOV species Reston ebolavirus ( REBOV ) , which is thought to be non-pathogenic for humans , was isolated from pigs [55] . Phylogenetic analyses suggested that the REBOV clade has evolved separately from the African Ebola viruses [55] . Interestingly , REBOV VP24 was also shown to interfere with the nuclear translocation of STAT1 [34] , indicating that the ability of VP24 to counteract IFN signaling was evolved among Ebola viruses prior to the separation of the African and Asian species . Notably , VP24 contributes to the host specificity of ZEBOV [24] , [25] . Whether VP40 plays a similar role in MARV host tropism has yet to be determined; however , it is intriguing that a mouse-adapted MARV acquired amino acid changes in VP40 [56] . The effects of MARV infection and MARV VP40 expression on IFNα/β , IFNγ and IL-6 signaling mirror the impact of Jak1 knock-out on these pathways . In cells lacking Jak1 , no STAT or Jak phosphorylation was observed upon IFNα/β or IFNγ treatment [5] . Similarly , the absence of Jak1 profoundly affects the IL-6 pathway as elimination of Jak1 was sufficient to fully abrogate any detectable phospho-STAT1 and greatly reduce phospho-STAT3 following IL-6 addition [7] , [8] . Interestingly , MARV infection and individual expression of MARV VP40 closely mirror this phenotype , where following IL-6 addition , phospho-STAT1 was undetectable but residual phospho-STAT3 was present ( Fig . 6 ) . Further studies will reveal to what extent the observed residual STAT3 phosphorylation may mediate IL-6 signaling . Our data are consistent with a model in which MARV VP40 targets Jak1 function , either directly or indirectly , although the possibility remains that MARV VP40 can also impair signaling of other Jak family kinases . A possible indirect mechanism of the observed inhibition could be a modulating effect of MARV VP40 on PTPs targeting Jak kinases . Recently , it has been reported that transgenic mice with reduced expression of the PTP CD45 were protected against lethal EBOV infection [56] . Interestingly , CD45 acts as a negative regulator of Jak1 in cells of hematopoietic origin [57] . However , our data suggest that PTPs are not involved in MARV-mediated inhibition of Jak1 signaling in cells of non-hematopoietic origin . Therefore , it is of interest to further extend those studies and to analyze Jak/STAT signaling in human hematopoietic cells in the context of MARV and EBOV infection . The observed inhibitory effects of MARV VP40 on both IFNα/β-induced gene expression and the antiviral effects of IFNβ may explain the capacity of MARV to prevent cellular responses to exogenously-added IFNα [30] . In this respect , MARV VP40 appears to serve the same purpose as the EBOV VP24 proteins which also counteract IFNα/β signaling . It is likely that counteracting IFNα/β signaling has a significant impact on viral pathogenesis in vivo , because , despite the presence of viral VP35 proteins that suppress IFNα/β production [31] , [58] , [59] , [60] , filovirus replication in vivo results in significant IFNα production [61] . The presence of IFNα/β signaling inhibitors likely also contributes to the relative insensitivity of filoviruses to IFNα/β as an antiviral therapy [50] . IFNγ also has antiviral properties [62] , however , suppression of IFNγ signaling may also modulate adaptive immune responses to infection . For example , human cytomegalovirus down-regulates Jak1 expression in a proteasome-dependent manner , and although a specific viral gene product that mediates this effect has not been defined , this function prevents the IFNγ-induced upregulation of MHC class II on infected cells [63] . Another viral protein that interacts with Jak1 and blocks the type I IFN signaling pathway is the measles virus V protein , but the consequence of this function for adaptive immunity has not been defined [64] . The possible impact of MARV infection and MARV VP40 expression on other cytokine signaling pathways involving Jak1 remains to be defined . Given the prominent role of Jak1 in numerous pathways , the impact of MARV VP40 on cytokine signaling may be quite broad . Filovirus VP40 proteins are matrix proteins sufficient to drive budding of virus-like particles , and they are thought to be the driving force for the budding of infectious virus [11] , [13] , [18] , [65] , [66] , [67] . The finding that MARV VP40 also serves as an inhibitor of IFN signaling is surprising and novel . Another example of a negative-strand RNA virus matrix protein that inhibits IFN responses is the vesicular stomatitis virus ( VSV ) matrix protein ( M ) . VSV M inhibits innate immune responses , including IFNβ production , by a mechanism different from MARV VP40 , inhibiting host cell transcription as well as nucleo-cytoplasmic transport of cellular mRNAs [68] , [69] , [70] , [71] . Host factors that interact with filovirus VP40 proteins have been described [14] , [18] , [65] , [72] , [73] . The most fully characterized interactions occur via the VP40 late domain which facilitates budding and release of virus particles . ZEBOV VP40 possesses two late domains , a PTAP motif and an overlapping PXXP motif [11] , [65] . These mediate interaction with Tsg101 , Nedd4 , and Rsp5 [14] , [65] . MARV VP40 possesses a single PPPY motif that allows interaction with Tsg101 [18] . To address the potential role of these well-characterized motifs in MARV VP40 inhibition of Jak/STAT signaling , a 16-PPPY-19 to 16-AAAA-19 mutant MARV VP40 was generated . As previously described , this mutation severely impaired MARV VP40 budding ( Fig . 8A ) [18] . Yet this mutation had no detectable impact on MARV VP40 inhibition of IFNα/β signaling ( Fig . 8B and C ) . Therefore , the late domain is dispensable for the IFN signaling function of VP40 , and the budding and signaling functions of MARV VP40 appear to be separable . Of note , IFN-induced cellular inhibitors of filovirus VP40 budding have recently been described . These include the IFN stimulated ISG15 and tetherin [26] , [74] , [75] . ISG15 is an IFN-induced protein which inhibits budding of EBOV VP40 . ISG15 inhibits the ubiquitin ligase Nedd4 , which interacts with EBOV VP40 through the PPXY motif to promote VP40 ubiquitination and budding [65] , [75] , [76] . Tetherin is constitutively-expressed in some cell types but is also IFN-inducible . Its expression can prevent release of VLPs produced following expression of EBOV or MARV VP40 [26] , [74] . Co-expression of EBOV GP has been shown capable of counteracting this antiviral function [26] . Whether MARV GP can also inhibit tetherin has not yet been addressed; however , because MARV VP40 can inhibit IFN signaling , it appears to have a built-in capacity to resist IFN-induced mechanisms that target viral budding . This study has identified an important difference in the biology of MARV and EBOV , defined a novel function for the MARV VP40 matrix protein and suggests that MARV may inhibit multiple Jak1-dependent cytokine signaling pathways . Future studies will determine whether the different means by which EBOV and MARV counteract cell signaling pathways result in significant differences in the pathologenesis of these viruses . Determining the molecular mechanisms by which MARV VP40 blocks signaling may facilitate development of new anti-MARV therapies .
293T , Vero E6 , Vero ( ATCC , Manassas , VA ) and Huh-7 ( kindly provided by Dr . DiFeo , Mount Sinai School of Medicine ) cells were maintained in Dulbecco's modified Eagle medium ( DMEM ) supplemented with 10% fetal bovine serum and 10 mM HEPES pH 7 . 5 or in DMEM supplemented with penicillin ( 50 units/ml ) , streptomycin ( 50 mg/ml ) and 10% fetal bovine serum . HUVECs were maintained in F-12K medium ( ATCC ) supplemented with 0 . 1 mg/ml heparin ( Sigma-Aldrich , St . Louis , MO ) , 0 . 03 mg/ml endothelial cell growth supplement ( ECGS ) ( Sigma-Aldrich ) , and 10% fetal bovine serum ( HyClone ) . A previously-described Newcastle disease virus engineered to express green fluorescence protein ( NDV-GFP ) was propagated in 10-day-old embryonated chicken eggs [77] . ZEBOV strain Mayinga and MARV strain Musoke were grown and propagated as described previously [30] . All work with infectious filoviruses was performed under biosafety level 4 conditions at the Institute of Virology , Philipps University of Marburg , Marburg , Germany . PCR products corresponding to FLAG-tagged , HA-tagged or untagged viral proteins of EBOV ( Accession # NC002549 ) and MARV ( Accession # NC001608 ) were cloned into the pCAGGS or pcDNA3 . 1 expression vectors [78] . The Nipah Virus W ( NiV W ) protein expression plasmid was previously described [77] . The expression plasmid for V5-tagged Langat Virus NS5 was previously described [46] . Human Jak1 ( Accession # BAE02826 ) and Tyk2 ( Accession # NP_003322 ) were RT-PCR amplified from RNA isolated from 293T cells and cloned with an HA tag into the pCAGGS vector . For the generation of the late domain mutants , site directed mutagenesis was performed using the QuickChange XL II kit ( Stratagene , La Jolla , CA ) . A Flag-tagged Rabies P expression plasmid in a pCR3 background was kindly provided by Drs . Conzelmann and Brzozka ( Ludwig Maximilian University , Munich , Germany ) . 293T cells were transfected using Lipofectamine 2000 ( LF2K ) at a ratio 1∶1 with plasmid DNA ( µg DNA: µL LF2K ) . Vero cells were transfected using LF2K at a ratio 1∶2 . Huh-7 cells were transfected using LF2K at a ratio 1∶2 . 75 . HUVEC cells were electroporated using the AMAXA nucleofector II , nucleofection program V-001 and solution V according to the manufacturer's directions ( Lonza , Walkersville , MD ) . Cells were lysed with an IGEPAL lysis buffer ( 50 mM Tris [pH 8 . 0] , 280 mM NaCl , 0 . 5% IGEPAL , 0 . 2 mM EDTA , 2 mM EGTA , 10% glycerol , 1 mM dithiothreitol ( DTT ) supplemented with protease inhibitor cocktail ( Roche ) and 0 . 1 mM Na3VO4 ) [79] for 30 min on ice and spun at 13kRPM on a refrigerated tabletop centrifuge for 1 minute . Universal type I IFN ( a consensus IFNα/β ) , human IFNβ and human IFNγ ( PBL , Piscataway , NJ ) were used at 1000 IU/ml unless otherwise specified for 30 min in RPMI ( GIBCO ) or phosphate buffered saline ( PBS ) supplemented with 0 . 3% BSA . Human IFNα-2b ( Essex Pharma , Kenilworth , NJ ) diluted in PBS supplemented with 0 . 1% BSA was used at 1000–2000 IU/ml unless otherwise specified . Human TNFα ( Peprotech , Rocky Hill , NJ ) was used at 50 ng/ml for 24 hours in HUVEC culture medium as described above . Human IL-6 ( Peprotech , Rocky Hill , NJ ) was used at 50 ng/ml in RPMI supplemented with 0 . 3% BSA . 4×105 Vero cells per well were cultured in 24 well plates and transfected with 1 µg of each plasmid encoding viral proteins . At 24 hours post-transfection cells were treated with IFNβ ( 1000 IU/ml ) for 24 hours . Then cells were infected with 5 hemagglutinating units of NDV-GFP virus in a volume of 200 µl of 0 . 3% BSA in PBS for 1 h , washed twice and replaced with DMEM supplemented with 10% FBS . GFP expression was visualized at 16 hours post-infection with a fluorescence microscope . 293T cells ( 5×105 ) or Huh-7 ( 3×105 ) were transfected with 0 . 5 µg of a construct having an IFN-stimulated gene 54 promoter driving expression of a chloramphenicol acetyltransferase ( CAT ) reporter gene ( pISG54-CAT ) , 0 . 1 µg of a constitutively expressing Renilla luciferase reporter construct ( pCAGGS-luc ) , and the indicated amounts of the expression plasmids . Twenty-four hours post-transfection , cells were washed and treated with IFN ( as described above ) . Sixteen hours post-IFN treatment , cells were harvested using reporter lysis buffer ( Promega , Madison , WI ) and analyzed for CAT and luciferase activities by standard methods . CAT activity was quantified by using a PhosphorImager and normalized to the luciferase activity . Alternatively , an ISG54-firefly luciferase reporter plasmid ( pISG54-Luc ) ( 0 . 3 µg ) reporter was used , and a dual luciferase reporter ( DLR ) assay was performed according to the manufacturer's guidelines ( Promega ) . For IFNγ-dependent gene expression , a reporter having 3 copies of the gamma activated sequence driving the expression of firefly luciferase ( GAS-Luc ) ( 0 . 3 µg ) was transfected with 0 . 1 µg of a constitutively expressing luciferase reporter construct ( pCAGGS-luc ) , and the indicated amounts of the expression plasmids . Twenty-four hours post-transfection , cells were washed and treated with IFNγ ( as described above ) . Sixteen hours post-IFN treatment cells were harvested and analyzed using a DLR assay ( Promega ) . Assays were performed in triplicate and p-values were calculated by a two-tailed Student's t-test for unpaired samples using the software GraphPad Prism ( GraphPad Software , Inc . ) . For the detection of the overexpressed viral proteins , the anti-V5 ( Invitrogen ) , anti-HA and anti-Flag M2 ( Sigma ) antibodies were used at a 1∶5000 dilution in 1% non-fat dry milk in Tris-buffered saline ( TBS; 20 mM Tris-HCl , pH 7 . 4; 150 mM NaCl ) . As a loading control , anti beta-tubulin ( Sigma ) antibody was used at a 1∶10 , 000 dilution in 1% non-fat dry milk in TBS . Anti-GFP was used at a 1∶10 , 000 dilution in 1% non-fat dry milk in TBS ( Clontech , Mountain View , CA ) . Phosphorylated STAT1 was detected with a phospho-tyrosine specific antibody recognizing phospho-Y701 ( BD Transduction Laboratories , San Jose , CA ) , and total levels of STAT1 with an antibody recognizing the STAT1 C-terminus ( BD Transduction Laboratories ) diluted to 1∶1000 and 1∶500 , respectively , in 1% non-fat dry milk in TBS . STAT2 and its phosphorylated form ( pY689 ) were detected with polyclonal antibodies ( Santa Cruz Biotechnology , Santa Cruz , CA and Upstate , Lake Placid , NY respectively ) diluted 1∶500 in 1% non-fat dry milk in TBS . STAT3 , pY705-STAT3 , Tyk2 , pY1054/1055-Tyk2 , pY1022/1023-Jak1 , pY1007/1008-Jak2 ( Cell Signaling , Beverly , MA ) , Jak1 ( BD Transduction Laboratories ) and Jak2 ( Millipore , Billerica , MA ) were used at a 1∶500 dilution in TBS , 0 . 1% Tween and 5% BSA . For the detection of IP-10 and IL-8 , supernatants of transfected HUVECs treated with 100 IU/ml human IFNγ or 50 ng/ml TNFα for 24 hours were collected and diluted 1∶100 and 1∶1000 , respectively , in PBS supplemented with 5% fetal bovine serum . The BD OptEIA Human IP-10 and Human IL-8 kits were used ( BD Biosciences , Franklin Lakes , NJ ) . Huh-7 cells grown in six-well plates to approximately 50% confluence were infected with ZEBOV or MARV at an MOI of 5 . At 24 hours p . i . , cells were left untreated or treated with IFNα-2b ( concentrations indicated in the figure legends ) , 10 IU/ml IFNγ or 50 ng/ml IL-6 for 20 or 30 min , respectively . Where indicated , filovirus-infected cells were treated with the phosphatase inhibitors sodium orthovanadate ( Sigma; 167 µM , 4 h ) or PTPIB-Inhibitor ( Merck; 33 µM , overnight; addition of fresh inhibitor the next morning for 40 min ) , or DMSO ( Sigma ) prior to IFN treatment . These conditions were shown to be sufficient to block Tyk2 dephosphorylation in non-infected cells treated with IFNα for 60 minutes in the presence of phosphatase inhibitors ( data not shown ) . Thereafter , cells were washed twice with PBS and scraped into 2× protein loading buffer ( 114 mM Tris-HCl , pH 6 . 8; 2 . 5% SDS; 125 mM dithiothreitol; 25% glycerol; 0 . 25% bromphenol blue ) . Cell lysates were transferred to fresh tubes , boiled for 2 . 5 to 10 min and subjected to SDS-polyacrylamide gel electrophoresis . Proteins were blotted onto polyvinylidene difluoride membranes , and the membranes were blocked in 5% non-fat dry milk in TBS containing 0 . 1% Tween 20 for 1 hour at room temperature , followed by an incubation step with the appropriate primary antibody in TBS supplemented with 5% bovine serum albumin and 0 . 1% Tween 20 overnight at 4°C . To detect endogenous cellular proteins , the following antibodies were used: rabbit anti-STAT1-phospho Tyr 701 ( CST; dilution 1∶3000 ) , rabbit anti-STAT1-total ( BD transduction; dilution 1∶3000 ) , rabbit anti-STAT2-phospho Tyr 689 ( Biomol; dilution 1∶1000 ) , rabbit anti-STAT2-total ( Imgenex; dilution 1∶1000 ) , rabbit-anti-STAT3-phospho Tyr705 ( CST; dilution 1∶500 ) , rabbit-anti-STAT3-total ( Santa Cruz; dilution 1∶500 ) , rabbit-anti-Tyk2-phospho Tyr1054/1055 ( CST; dilution 1∶3000 ) , rabbit-anti-Tyk2-total ( Santa Cruz , dilution 1∶3000 ) , rabbit-anti-Jak1-phospho Tyr1022/1023 ( Biomol; dilution 1∶1000 ) and rabbit-anti-Jak1-total ( Santa Cruz; dilution 1∶1000 ) . Horseradish peroxidase-conjugated secondary antibodies ( Dianova ) were used and visualized by using either the chemiluminescence substrate SuperSignal West Dura Extended Duration or SuperSignal West Femto Maximum Sensitivity ( Pierce ) according to the manufacturer's instructions . To verify virus infection , infected and IFN-treated cells grown on glass coverslips were subjected to immunofluorescence analysis using virus-specific antibodies as described below . Huh-7 cells grown on glass coverslips were infected with ZEBOV or MARV at an MOI of 5 or left uninfected . At 24 hours p . i . , cells were washed twice with PBS and inactivated by treatment with 4% paraformaldehyde for 24 hours . Cells were then permeabilized with a mixture of acetone and methanol ( 1∶1 , vol/vol ) for 5 min at −20°C and treated with 0 . 1 M glycine . As primary antibodies , a rabbit antiserum directed against the nucleocapsid complex of MARV ( 1∶100 ) and a goat antiserum directed against ZEBOV ( 1∶500 ) ( kindly provided by Dr . Becker , Philipps University of Marburg , Marburg , Germany ) were used . To detect endogenous STAT1 or STAT2 proteins in filovirus-infected cells , the cells were fixed in 4% paraformaldehyde as described above , washed with 50 mM NH4Cl in PBS and permeabilized with 0 . 5% Triton X-100 in PBS . After incubation with primary antibodies ( rabbit anti-STAT1 or rabbit anti-STAT2 ( Santa Cruz; dilution 1∶100 ) along with filovirus-specific antibodies ) , the specimens were washed with PBS and incubated with fluorescence-labeled secondary antibodies . To analyze the intracellular localization of endogenous STAT2 in cells expressing individual viral proteins , Huh-7 cells were transfected with 2 µg of plasmid DNA encoding MARV or ZEBOV VP40 , VP35 , or VP24 using FuGene 6 ( Roche ) according to the manufacturer's instructions . VP24 proteins and ZEBOV VP35 were tagged with an HA epitope . As a control cells were transfected with 2 µg pCR3 Flag-tagged rabies virus P . Immunofluorescence analysis was performed by using antibodies directed against STAT2 , MARV VP35 ( mouse; 1∶100 ) , MARV VP40 ( mouse; 1∶100 ) , ZEBOV VP40 ( mouse; 1;100 ) , Flag- ( Sigma; dilution 1∶700 ) or HA-tags ( Sigma; dilution 1∶1000 ) . 293T cells were transfected with 2 µg of expression plasmid . At 48 hours post-transfection , cell culture supernatants were clarified by centrifugation at 200×g for 5 min and pelleted through a 20% sucrose cushion in NTE buffer ( 100 mM NaCl , 10 mM Tris [pH 7 . 5] , 1 mM EDTA [pH 8 . 0] ) at 160 , 000×g for 2 hours at 4°C . Supernatants were aspirated and the pellets containing the virus-like particles ( VLPs ) were resuspended in NTE buffer . Cells were washed with PBS and lysed in radioimmunoprecipitation assay buffer ( RIPA ) ( 50 mM Tris [pH 7 . 4] , 150 mM NaCl , 0 . 1% sodium dodecyl sulfate [SDS] , 0 . 5% deoxycholate , 1% NP-40 ) supplemented with protease inhibitor cocktail ( Roche ) . VLPs and lysates were analyzed by SDS-PAGE and visualized by western blotting , as described [80] . | The closely related members of the filovirus family , Ebola virus ( EBOV ) and Marburg virus ( MARV ) , cause severe hemorrhagic disease in humans with high fatality rates . Infected individuals exhibit dysregulated immune responses which appear to result from several factors , including virus-mediated impairment of innate immune responses . Previous studies demonstrated that both MARV and EBOV block the type I interferon-induced Jak-STAT signaling pathway . For EBOV , the viral protein VP24 mediates the inhibitory effects by interfering with the nuclear translocation of activated STAT proteins . Here , we show that MARV uses a distinct mechanism to block IFN signaling pathways . Our data revealed that MARV blocks the phosphorylation of Janus kinases and their target STAT proteins in response to type I and type II interferon and interleukin 6 . Surprisingly , the observed inhibition is not achieved by the MARV VP24 protein , but by the matrix protein VP40 which also mediates viral budding . Over-expression studies indicate that MARV VP40 globally antagonizes Jak1-dependent signaling . Further , we show that a MARV VP40 mutant defective for budding retains interferon antagonist function . Our results highlight a basic difference between EBOV and MARV , define a new function for MARV VP40 and reveal new targets for the development of anti-MARV therapies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/emerging",
"viral",
"diseases",
"virology",
"infectious",
"diseases/viral",
"infections",
"immunology/innate",
"immunity",
"virology/immune",
"evasion",
"virology/host",
"antiviral",
"responses"
] | 2010 | Marburg Virus Evades Interferon Responses by a Mechanism Distinct from Ebola Virus |
Overexpression of the xenotoxin transporter P-glycoprotein ( P-gp ) represents one major reason for the development of multidrug resistance ( MDR ) , leading to the failure of antibiotic and cancer therapies . Inhibitors of P-gp have thus been advocated as promising candidates for overcoming the problem of MDR . However , due to lack of a high-resolution structure the concrete mode of interaction of both substrates and inhibitors is still not known . Therefore , structure-based design studies have to rely on protein homology models . In order to identify binding hypotheses for propafenone-type P-gp inhibitors , five different propafenone derivatives with known structure-activity relationship ( SAR ) pattern were docked into homology models of the apo and the nucleotide-bound conformation of the transporter . To circumvent the uncertainty of scoring functions , we exhaustively sampled the pose space and analyzed the poses by combining information retrieved from SAR studies with common scaffold clustering . The results suggest propafenone binding at the transmembrane helices 5 , 6 , 7 and 8 in both models , with the amino acid residue Y307 playing a crucial role . The identified binding site in the non-energized state is overlapping with , but not identical to , known binding areas of cyclic P-gp inhibitors and verapamil . These findings support the idea of several small binding sites forming one large binding cavity . Furthermore , the binding hypotheses for both catalytic states were analyzed and showed only small differences in their protein-ligand interaction fingerprints , which indicates only small movements of the ligand during the catalytic cycle .
The development of multidrug resistance ( MDR ) is one major impediment in cancer and antibiotic therapies [1]–[3] . In 1976 Juliano and Ling were able to associate the occurrence of MDR with the presence of P-glycoprotein ( P-gp ) , the most prominent member of the adenosine triphosphate ( ATP ) binding cassette ( ABC ) transporter superfamily [4]–[6] . ABC proteins are energy dependent transporters with P-gp ( ABCB1 ) , multidrug resistance protein 1 ( MRP1 , ABCC1 ) and breast cancer resistance protein ( BCRP , ABCG2 ) playing an important role in the protection of cells from harmful xenotoxins . Additionally , ABC proteins are known for modulating the pharmacokinetic profile of drugs and therefore the food and drug administration ( FDA ) suggested that new drug candidates should be routinely screened for P-gp interaction [7] . In this respect reliable in silico methods to characterize P-gp interaction would be of great benefit and help to render the drug discovery process more efficient [8] . However , the polyspecificity of the transporter poses a remarkable challenge concerning this task [9] . A number of ligand based studies have been conducted and provide some insights into the molecular basis of ligand/transporter interaction [10] , [11] . With the help of biochemical studies like cysteine-cross linking , arginine scanning or photoaffinity labeling , amino acids contributing to binding of selected substrates were identified . On grounds of these experiments interaction sites for verapamil , rhodamine ( R-site ) , Hoechst ( H-site ) and of cyclic peptide P-gp inhibitors ( CPPI's ) in the transmembrane ( TM ) domains ( TMDs ) of P-gp have been postulated [12]–[16] . Following the ABC transporter topology , P-gp possesses two TMDs , each consisting of 6 TM helices ( TMHs ) , and two nucleotide binding domains ( NBDs ) . While the TMDs are generally responsible for ligand interaction , ATP binding and hydrolysis takes place at the highly conserved nucleotide binding domains ( NBDs ) [17] . In case of propafenone type ligands photoaffinity labeling studies proposed two symmetrical binding regions at the interfaces of TMHs 5/8 and TMHs 2/11 , respectively [18] , [19] . Nevertheless , due to the small number and the low resolution of crystal structures of ABC-exporters , concrete binding hypotheses remain to be elucidated [20] . The lack of high resolution structures can be explained by the fact that ABC efflux pumps are located in the membrane and that they are rather flexible proteins . As energy dependent transporters they undergo large structural changes during one catalytic cycle , comprising ligand and ATP binding , ligand release and nucleotide hydrolysis [17] , [21] , [22] . Up to now the structure of human P-gp could not be resolved , for which reason homology models relying on bacterial homologues had to be utilized . With respect to this the bacterial transporters Sav1866 and MsbA structures , representing different catalytic states of the transport cycle , were generally used as templates [20] . In 2009 the crystal structure of mouse P-gp [12] in complex with a cyclic tetrapeptide was resolved , thus representing a ligand binding competent conformation of the protein . With 88% sequence identity it is well suited for homology modeling of the human homologue and thus paves the way for structure-based approaches . The present study aimed at elucidating the binding mode of propafenone type inhibitors of P-gp using a combined homology modeling/docking approach . Propafenones show a clear structure-activity relationship ( SAR ) pattern [11] and thus represent versatile tool compounds to pursue this task . The wealth of ligand-based information available allows judging the reliability of docking poses on basis of the SAR pattern rather than by use of energetic terms derived from scoring functions . The selected compounds were the piperidine analogue GPV005 , the analogous des-hydroxy derivative GPV186 , the arylpiperazine GPV019 , the hydroxyphenylpiperidine GPV062 , and the benzoylamide GPV366 ( Figure 1 ) . All compounds bear a carbonyl group , which has been shown to be important for high P-gp inhibitory activity [23] . There are numerous studies showing that there is a basic underlying correlation between P-gp inhibitory activity and lipophilicity of the compounds . This accounts for several compound classes and has also been shown for propafenone analogues . However , propafenones which bear a 4-hydroxy-4-phenylpiperidine moiety are generally by a factor of 10 more active than equi-lipophilic derivatives without the hydroxy-group in 4-position of the piperidine moiety ( Figure S1 ) [24] . This points at a distinct additional interaction mediated by the 4-hydroxy group , most probably in form of a H-bond . This distinct SAR pattern in combination with the recently described common scaffold clustering [25] , [26] was used to guide the prioritization of docking poses .
In March 2009 Aller et al . published the crystal structure of mouse P-gp in the absence of a ligand ( PDB ID: 3G5U ) [12] and in complex with stereoisomeric CPPI's ( PDB IDs: 3G60 , 3G61 ) [12] . These structures represent the ligand binding competent state and were therefore the first choices for investigating drug/P-gp binding . As the structural difference between the apo protein and the co-crystallized structures was surprisingly low ( 0 . 61 Å of Ca atoms ) the higher resolved 3G5U structure was utilized as homology modeling template ( 3G5U_Pgp ) . With the modeling program MODELLER 100 different homology models were created and refined . All models were assessed with the geometry check tool implemented in MOE , which was used as a selection criterion for the final model . As additional measure for model quality the GA341 method was used , which relies on sequence identity , compactness and the combined statistical z-score . All models obtained the highest possible GA341 value of 1 . Furthermore , the final model was analyzed with the structure assessment program PROCHECK [27] . The Ramachandran plot showed that 84 . 6% of the residues lie in most favored , 12 . 5% in additionally allowed , 2 . 1% in generously allowed and 0 . 8% in disallowed regions . The 2 . 9% of residues in generously allowed or disallowed regions are located in the nucleotide binding domains ( NBD ) or extracellular loops ( ECL ) and are therefore not involved in drug binding ( Figure S2 ) . The QMEAN analysis [28] ( Figure S3 ) showed that residues lining the binding pocket are of satisfactory quality . In order to cover different catalytic states of the protein , a second homology model was generated on basis of the bacterial transporter Sav1866 in the nucleotide-bound state ( PDB code: 2HYD ) [29] ( 2HYD_Pgp ) . This crystal structure is the highest resolution ABC exporter structure and has therefore been frequently used as modeling template [20] . 100 different models were generated and refined with MODELLER , of which all obtained a GA341 score of 1 . The final model was selected on basis of the geometry check function in MOE . The Ramachandran plot statistics provided by the evaluation tool PROCHECK showed that 92% of all residues lie in most favored regions , while 6 . 5% were found in additionally allowed , 0 . 2% in generously allowed and only 0 . 6% in disallowed regions ( Figure S2 ) . Most of the 0 . 8% residues that are located in generously allowed or disallowed regions can be found in the NDB . Although residue Y116 lies within the TMDs and could therefore be involved in drug binding , this residue is oriented outside the cavity . A Ramachandran analysis performed by MolProbity and MOE detected no outliers in the TM region . Furthermore , this model shows also good quality in the binding site region according to QMEAN analysis ( Figure S3 ) . For the docking process five different propafenone derivatives were selected according to their differences in lipophilic efficiency and fit quality [30] , and were docked into both homology models . With the genetic algorithm based docking program GOLD [31] 100 poses for each of the five ligands were generated . To determine the ASN , GLN and HIS flips the web application MolProbity was utilized [32] . In order to avoid any bias , the binding site was defined as the complete TM region . According to the binding site assessment tool implemented in the software suite Schrödinger ( SiteMap ) , this region in 3G5U_Pg mainly shows highly hydrophobic characteristics , which prompted us to dock the ligands in their non-ionized state . This is also supported by previous findings of ligand-based QSAR studies which indicated that the nitrogen atom not necessarily interacts in its charged form . However , since there is evidence that the protein's pore is water filled , the ligands were also docked in their ionized state [33] . This is also in accordance with recently published data which show that mutation of two glutamine residues at the entry path of the transporter to positively charged arginines affected the inhibitory activity of an positively ionizable propafenone analog , whereas the activity of GPV366 remained unmodulated [34] . The resulting poses in both conformations were distributed largely within the TM region of P-gp ( Figure 2 ) , showing interactions with protein residues of multiple TM helices , located throughout the binding region . The calculation of protein ligand interaction fingerprints ( PLIF ) with MOE showed that in case of 3G5U_Pgp residues primarily located on TM helices 1 , 5 , 6 , 7 , 8 , 11 and 12 were involved in binding ( Figure 3 ) . According to this tool , residues involved either show direct interactions with docking poses or are located within 4 . 5 Å distance to the ligand . The unprocessed complexes were energetically minimized using LigX , a minimization tool implemented in MOE for further evaluation . The minimized poses were clustered according to the root-mean-square deviation ( RMSD ) of the heavy atoms of the common scaffold ( Figure 1 ) [35] . To follow the idea of a common binding mode only those clusters were kept that comprehend at least four out of the five compounds used ( common scaffold clusters , CSCs ) . Clustering the poses of the docking run with 3G5U_Pgp resulted in 114 clusters , which were subsequently reduced to 12 CSC . As can be seen in Figures 2a and b some clusters protrude into the central cavity , but most of the CSCs are found in the vicinity of helices 5 and 8 ( called the 5/8 interface ) . Previous photo-affinity labeling experiments suggested this region to be in involved in propafenone binding [36] . The position of the CSCs close to the 5/8 interface was also reflected in the PLIF pattern , as the involvement of residues L304 and Y307 located in TM helix 5 , F343 of TM helix 6 , L724 and I731 in TM helix 7 , A761 in TM helix 8 and V981 in TM helix 12 was increased ( Figure 3 ) . In case of 2HYD_Pgp , the RMSD clustering process resulted in 78 clusters , which were reduced to nine common scaffold clusters , containing 264 poses ( Table 1 ) . As can be seen in Figure 2c and d , also docking into the nucleotide-bound homology model results in CSCs that tend to accumulate closer to the 5/8 interface and thus in vicinity of the photo-affinity labeled residues ( Figure S4 ) . The clustering process did not change the general PLIF pattern . TM helices 5 , 6 , 7 and 8 show similar contributions before and after scaffold clustering , but more frequently interactions were observed with individual residues , like Y307 ( TM helix 5 ) , Y310 ( TM helix 5 ) , L724 ( TM helix 7 ) and T769 ( TM helix 8 ) ( Figure 4 ) . The model based on the murine 3G5U structure represents the binding competent state , whereas the model based on the nucleotide-bound 2HYD structure likely represents the off-state of P-gp ligands [37] . Since propafenones might show different affinities towards these two structures , final pose evaluation was carried out in different ways . In the hit-to-lead decision process as well as in lead optimization different efficiency metrics are applied to prioritise lead candidates . Briefly , in case of equi-potent compounds these parameters select for the smaller , more hydrophilic ones . As high lipophilicity correlates with promiscuity , poor solubility and poor metabolic clearance [38] , candidates with high lipophilic efficiency ( LLE = log ( potency ) - logP ) are preferred . Ligand based studies clearly demonstrate a correlation between lipophilicity of P-gp inhibitors and their biological activity . However , as P-gp is extracting its ligands directly out of the membrane bilayer , this is most probably a consequence of concentration in the membrane rather than of direct protein interaction . Calculating the LLE normalizes for this effect and aids in identifying ligands with increased activity as a result of direct interaction with the protein rather than higher biomembrane distribution . The 4-hydroxy-4-phenylpiperidine GPV062 shows by far the highest LLE ( Table 2 ) suggesting that in contrast to the other ligands , the higher activity of GPV062 is not due to a high logP value . While LLE normalizes for the lipophilic bias in potency description , LE simply corrects for the size of a molecule by dividing the activity of a compound by its heavy atom count . This approach is extensively used in fragment based drug design to select those fragments , which are worth being further investigated . As Reynolds et al . [30] concluded that LE generally is biased towards smaller molecules , the normalized size-independent fit quality ( FQ ) was assessed . Both , LE and FQ , clearly highlight the hydroxyphenylpiperidine GPV062 as being the most efficient compound ( Table 2 ) . The explanation for the increased LLE and FQ values seems to be the 4-hydroxy-group of GPV062 . As this group clearly reduces the lipophilicity of a molecule , the increase in activity was interpreted as a result of hydrogen bonding . Thus , those CSCs were prioritized in which GPV062 is able to form a hydrogen bond between the hydroxyl-group of the 4-hydroxy-4-phenyl moiety and the protein . With 3G5U_Pgp only one quarter of all twelve common scaffold clusters showed a hydrogen bond between the hydroxyl-group of GPV062 and the protein ( Table 1 ) ( GPV062-OH Clusters ) . These three clusters ( CSCs I , II , III ) are located very close to each other at the 5/8 interface ( Figure 5a ) , with an increased number of interactions formed by residues L304 , Y310 , L724 , A761 and V981 . Furthermore , the PLIF pattern showed that interactions with TM helices 1 and 11 are no longer present . The positions of CSCs I and III are very similar , since both are forming a hydrogen bond with Y310 and a π/π-interaction with F336 . In CSC II , on the contrary , a hydrogen bond interaction with A761 was observed . For further evaluation of the poses a pharmacophore search was performed , utilizing a model published by Langer et al . that based on a set of propafenone type P-gp inhibitors [39] . Only those two clusters that formed a hydrogen bond with Y310 matched this pharmacophore query . As depicted in Figure 5b , both clusters perfectly fit the photolabeling pattern observed in this half of the protein . Evaluation of the docking results with 2HYD_Pgp could not be based on ligand affinity data , since this structure represents the nucleotide-bound off-state and therefore is considered as the low-affinity state for substrates . This rules out prioritization on basis of SAR-information . All common scaffold clusters of 2HYD_Pgp are in close vicinity of the 3G5U_Pgp GPV062-OH poses ( Figure 6 ) .
The homology models generated in this study resemble two different states of P-gp: the open-inward or apo state and the open-outward or nucleotide-bound state . Since the publication of mouse P-gp in the absence ( PDB ID: 3G5U ) or in complex with ligands ( PDB Ids: 3G60 , 3G61 ) only a few homology models of the human homologue were published on the basis of these structures . Pajeva et al . presented two homology models that were based on the structure of 3G61 , chain A , which is complexed with QZ59-SSS [40] , [41] . The advantage of selecting this template for homology modeling is the presence of the complexed ligands . On the other hand , 3G5U is resolved at higher resolution ( 3 . 80 Å ) and shows only minor differences in the binding site ( RMSD of all atoms of QZ59-SSS surrounding residues: 1 . 251 Å ) . However , the still relatively low resolution of the template certainly needs to be taken into account when it is used for docking experiments . The open-outward model relied on the structure of the bacterial homologue Sav1866 ( PDB ID: 2HYD ) , which possesses the same domain architecture as P-gp [42] and therefore frequently served as modeling template . With a resolution of 3 . 0 Å it represents one of the best resolved full ABC transporters . The relevance of this nucleotide-bound structure is widely accepted , as experimental studies showed close association of the NBDs [43] , [44] . In contrast , the structures of mouse P-gp disagree with kinetic and FRET studies that report no complete dissociation of the NBDs [37] , [45] . In addition , a recent cross-linking study further strengthened this by showing that an M1M cross-link between L175C and N820C did not prevent verapamil and rhodamine B to be transported [46] . However , as P-gp is known to be highly flexible and to undergo large conformational changes during the catalytic cycle , the existence of a state with dissociated NBDs cannot be ruled out entirely . Additional evidence was presented by Sauna et al . , who demonstrated that ATP binding reduces the affinity for propafenone analogues [37] . Finally , the fact that the mouse P-gp structure ( 3G5U ) has been cocrystallized with two ligands strongly indicates that this structure represents a ligand-binding competent state of the protein . Thus it was considered as a versatile template for modeling the high-affinity state of the protein for subsequent docking studies . Although ligand docking is a commonly used tool for the identification of ligand-protein interactions , in case of P-gp it bears a lot of challenges: ( i ) P-gp possesses a large binding cavity that consists of several binding sites , ( ii ) is highly flexible , and ( iii ) is probably able to harbor more than one ligand simultaneously [47] , [48] . Finally , there is no high resolution structure of human P-gp available , which requires to work with protein homology models . Considering the low resolution of the templates , this adds additional layers of uncertainty . Thus , results from ligand docking runs have to be interpreted very carefully . In an attempt to combat all these uncertainties we applied an exhaustive docking protocol avoiding to a maximum possible extent the use of scoring functions and including all the knowledge present from SAR and QSAR studies . In docking experiments , the definition of the binding site is a key parameter of the docking protocol . As only little information is available about binding of propafenones into P-gp , the whole TM region was selected as a potential interaction region . In order to avoid any bias introduced by scoring functions , a large amount of docking poses was generated . While placement algorithms of docking programs are most of the time able to find the native pose of a ligand in the binding pocket , the correct estimation of the binding energy leading to a correct ranking of the poses is still unsatisfying . To overcome this uncertainty of scoring functions , we recently implemented experimental data guided docking/scoring . In this approach prioritization of docking poses is performed on basis of mutagenesis data , biochemical data , and/or information from ligand based studies [25] , [26] . The interaction of propafenones with P-gp follows a clear structure-activity relationship pattern ( for reviews see [11] ) . Based on these results and on calculation of lipophilic efficiency ( LLE ) and fit quality ( FQ ) we selected a small set of analogs for docking and subsequent common scaffold clustering . Both LLE and FQ as well as previously performed Hansch analysis stressed the importance of the hydroxyl-group of GPV062 for high activity . Clustering of all poses according to their common scaffold ( Figure 1 ) combined with pose selection based on H-bonding interactions of the OH group allowed a considerable reduction of docking poses . Although docking experiments have their limitations depending on the validity of the target structure , the results of docking into 3G5U_Pgp are very consistent . As shown in Figure 5 the three final clusters are located in close vicinity . Especially CSCs I and III are very similar , showing strong H-bonding interactions with Y310 and thus supporting the importance of the hydroxyl group of GPV062 . Both clusters also match the pharmacophore model of Langer et . al [39] . Due to previously performed ligand based studies also the importance of the carbonyl group of the propafenone scaffold became evident [49] . Although initial poses show no interaction with the carbonyl group , these become apparent after processing of data with the rotamer explorer implemented in MOE . When rotating amino acid residue Y307 towards the carbonyl group , an interaction can be generated ( Figure 7 ) . In a dynamic system H-bond formation thus might be observed . Interestingly , for CSC III a rotation of Y307 did not result in an interaction with the carbonyl group , most probably due to a small offset of the carbonyl group towards the cell interior . However , this assumption would need further investigations , since discussing possible interactions on atomistic detail has to be done with caution when working with a homology model , especially if the resolution is quite low . Nevertheless , the relevance of Y307 in ligand binding was also shown with cocrystallized CPPI's , where the R-stereoisomer forms an interaction with this residue [12] . Furthermore , this residue is in close vicinity to I306 , which was shown to lead to permanent activation of ATPase activity when mutated to cysteine and covalently linked with the thiol-reactive drug substrate verapamil [15] . CSC II forms a weak H-bond between the hydroxyl-group of GPV062 and the backbone of A761 . With respect to the ligand interaction tool in MOE the strength of this bond is only 1/10 compared to that in CSCs I and III . Applying the rotamer explorer results in either formation of a stronger hydrogen bond with the OH-group of GPV062 or formation of a new interaction with the carbonyl group ( with these interactions not being coexistent ) . Finally , with respect to residues photoaffinity labelled by benzophenone analogous propafenones , CSCs I and III show a better match ( Figure 5b ) , because the photoreactive carbonyl group is closer to the PAL region than in CSC I . In consideration of these findings the pose of CSC I was preferred over the other two clusters . It is also known that binding of propafenones to P-gp meets steric constraints in the vicinity of the nitrogen atom , because diphenyl moieties in this position lead to a log order decrease in activity [49] . In all three clusters the introduction of a diphenyl substituted nitrogen results in steric clashes and subsequent minimization of the binding pocket leads to the loss of H-bond interactions . Docking into 3G5U_Pgp with ionized ligands resulted in three different CSCs that show an interaction between the OH_group of GPV062 and the protein . While one is located very central in the pore ( CSC IV ) forming an H-bond between GP062-OH and A727 , the other two ( CSC V and VI ) exactly match CSC I of the docking with neutral ligands . For the latter an H-bond between the hydroxyl-group and Y310 could be observed . As can be seen in Figure 2 , the different CSCs of 2HYD_Pgp are located in the same binding site at the 5/8 interface . Regarding their different orientation within this region , docking poses can be separated into two distinct groups . Docking poses belonging to group 1 ( CSCs a , b , c and d ) frequently form interactions between the carbonyl group and Y307 . Furthermore , H-bond interactions between the piperidine nitrogen or the hydroxyl-group and Y310 can be observed . This interaction pattern is similar to the one of CSCs I and III of the docking run performed with 3G5U_Pgp . Individual GPV062 poses show additional H-bond interactions between the 4-hydroxy-group and Y310 , another frequently observed interaction in CSCs I and III . According to these observations the transformation of CSCs I and III in the apo state into CSCs of group 1 of the nucleotide-bound state seems possible . In contrast , group 1 and group 2 are in an up-side-down orientation when compared to each other . In this case the carbonyl group is located near Y310 and thus closer to the extracellular portion of the protein . The nitrogen atom , as well as the hydroxyl group , is oriented towards Y307 and N721 , which was also observed for CSC II of the 3G5U_Pgp docking run . Therefore , group 2 , comprising clusters e , f and g , corresponds to the nucleotide-bound conformations of CSC II of the apo-conformation . CSCs h and i cannot be clearly assigned to one of these groups and have to be regarded separately . The nitrogen atom of CSC h shows a similar location as the N of group 2 , however , due to a shift of the central phenyl ring downwards , H-bond interactions between the carbonyl oxygen and Y307 and the OH-group and N721 can be formed simultaneously . CSC i shares its carbonyl group orientation with group 2 , but the central phenyl ring lies in a perpendicular direction , which results in interactions between the ligand nitrogen and hydroxyl group with Q725 . Considering the docking run to 2HYD_Pgp with ionized ligands , group 1 could be clearly reproduced . Three out of seven CSCs form those characteristic H-bond interactions between the carbonyl oxygen and Y307 and the hydroxyl group and Y310 . In contrast to the unprotonated ligands , the nitrogen atom and Y310 form a pi/cation interaction and occur at higher frequency . Overall the clusters belonging to group 1 show high homogeneity and strong interactions . In contrast to this the poses of each of the four other clusters share no consistent pattern and therefore the common binding was only reflected in geometrically similar positioning . Interestingly , although the experimental data suggest two symmetrical binding sites , no common scaffold cluster and hardly any poses could be found at the second photoaffinity labeled site at the 2/11 interface . One possible explanation might be the asymmetry of the template crystal structure 3G5U . The region consisting of TM helices 4 , 5 , 7 , 8 , 9 and 12 in case of 3G5U_P-gp , and TM helices 3 , 4 , 5 , 6 , 7 and 8 in case of 2HYD_P-gp , in both cases showed larger sites when using the SiteFinder tool in MOE than their counterparts around the 2/11 interface . This demonstrates the limitations of docking experiments relying on one crystal structure that represents only a snapshot of a flexible protein . Thus , to rule out the possibility that every docked ligand will end up at the 5/8 interface just because of this asymmetry , a docking run with rhodamine 123 was conducted . In this case 21 of 39 clusters were found in vicinity of residues I340 , L975 and V981 , which are located on TM helices 6 and 12 and known to be involved in rhodamine binding [13] . In order to gain first insights into the potential ligand translocation pathways , the compounds were docked in two different catalytic states of P-gp . Interestingly , the docking results show similar interaction patterns . In both models , ligand poses are found in close vicinity ( 4 , 5 Å ) of residues Y307 and Y310 of TM helix 5 , F343 of TM helix 6 and L724 of TM helix 7 , which suggests involvement of both TM domains in drug binding . This is in accordance with Loo et al . , who showed that both TM domains are essential for drug translocation [50] . In Figure 8 the interacting amino acid residues of both docking approaches are depicted . In the 3G5U_Pgp structure the interactions seem to be very similar , concentrating on the 5/8 interface . Due to the conformational change and the resulting movement of TM helix 12 , interactions between propafenones and V977 and V981 are lost . Top views of the models indicate that the corresponding interacting residues ( 3G5U_Pgp: yellow , 2HYD_Pgp: blue , both: green ) face the central pore . It seems that the conformational change associated with nucleotide binding moves previously buried residues towards the binding pocket and therefore allows them to form new interactions with the ligands . In Figure 9 a Venn diagram compares residues in binding sites of CPPIs and verapamil with that of propafenones . As TM helices 5 , 6 and 7 are lining the central cavity in the murine P-gp structure , a considerable overlap of residues interacting with propafenones and that shown to interact with cocrystallized CPPIs can be found in this region . One residue of TM helix 7 , F728 , is suggested to interact with all four drugs and therefore plays a crucial role in ligand binding . This is in agreement with the finding of Loo et al . that TM helix 7 is part of the drug binding site [51] . Loo et al . also demonstrated that binding of vinblastine , cyclosporin A and rhodamine B could prevent the formation of a cross-link between L339C and F728C , suggesting that the ligands are at least partially located between these two residues [52] . This is also the case for the three docking clusters in 3G5U_Pgp , which are presented in this study . Furthermore , the diagram is consistent with the notion that P-gp possesses a large binding cavity , which harbors different partially overlapping drug binding sites for different ligands [39] , [40] . In the cocrystallized structures 3G60 and 3G61 the cyclopeptides are located at the interface of the two TMDs , which explains the high overlap between these ligands and verapamil or propafenones , respectively . Ligand docking into polyspecific antitargets such as the hERG potassium channel and the drug transporter P-glycoprotein requires thorough validation of the poses obtained . In this paper we describe the application of an SAR-guided docking protocol , which for the first time retrieves a binding hypothesis for propafenone-type inhibitors of P-gp . Although performing docking studies with homology models always bears a lot of risks the results are in agreement with experimental studies , which strengthens the applicability of the complex docking protocol we used for this study . This could pave the way for structure-based ligand design approaches .
Two homology models based on the bacterial homologue Sav1866 ( PDB ID: 2HYD , resolution: 3 . 0 Å [29] ) and murine P-gp ( PDB ID: 3G5U , chain A , resolution: 3 . 8 Å [12] ) were built . Both models were generated with the program MODELLER 9v7 using the automodel protocol [53] . In case of 3G5U_Pgp the alignment proposed by Aller et al . [8] was used ( Figure S5 ) . To correct the disruption in TM helix 12 of 3G5U a secondary structure constraint between residues 885 and 928 was applied . For 2HYD_Pgp the alignment was done according to Stockner et al . [54] ( Figure S6 ) . The linker region between the TM domains was modeled . Out of the 100 generated models those with the smallest number of outliers according to the geometry check function in MOE were selected for docking . For the docking study five propafenone derivatives were selected on basis of known SAR and differences in LLE and FQ . LLE was calculated by subtracting ClogP from experimentally determined IC50 values and FQ was calculated as outlined in [30] . To examine the quality of the ClogP calculation , the values were compared with previously published experimentally defined logP data of propafenone analogs [23] . A correlation of r = 0 . 92 could be identified . Minimization and protonation of the ligands was performed with MOE . For the correct determination of ASN/GLN/HIS flips the web application MolProbity was utilized [32] . The docking process was performed using the Gold Suite 1 . 2 . 1 [31] . Hydrogens were added and the binding site was defined as the entire TM region of the homology model . All side chains were kept rigid and the ligand was treated flexible by performing 100 genetic algorithm runs per molecule . The implemented Gold scoring function GoldScore was used for evaluation of the complexes . The final poses and the surrounding protein amino acid residues were minimized using LigX implemented in the MOE software package . Rescoring was performed with the empirical scoring function XSCORE . On basis of the common scaffold an RMSD matrix of all five ligands was generated and used for clustering . The dissimilarity matrix was clustered with the program R [55] , using complete linkage as clustering algorithm and a clustering height of 3 Å . Only those clusters were kept that inherited at least four out of the five ligands docked . In case of 3G5U_Pgp those clusters were selected for final assessment that were able to form a hydrogen bond between the OH-group of GPV062 and the protein , detected by the ligand interaction tool of MOE . | A major reason for the failure of cancer , antibiotic and antiviral therapies is the development of multidrug resistance ( MDR ) . P-glycoprotein ( P-gp ) , an ATP-dependent transport protein located in the membrane of epithelial cells of the kidney , liver , pancreas , colon and the blood-brain barrier , has been linked to the export of a broad variety of xenotoxins . Overexpression of P-gp leads to extrusion of therapeutic drugs and therefore triggers MDR . Thus , identification of potential P-gp inhibitors represents a promising concept for treatment of multiresistant tumours . However , due to lack of high resolution structural information and the polyspecific ligand recognition pattern only very limited information is available on the molecular basis of ligand/transporter interaction . Within this study we characterized the propafenone binding site of P-gp by docking a set of derivatives with known SAR into homology models of P-gp which represent both the apo and the nucleotide-bound state . Poses retrieved are in accordance with results from previous photoaffinity labeling studies and thus pave the way for structure-based in silico screening approaches . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computer",
"applications",
"medicine",
"medicinal",
"chemistry",
"drugs",
"and",
"devices",
"small",
"molecules",
"computer-aided",
"design",
"protein",
"structure",
"pharmacology",
"proteins",
"chemistry",
"biology",
"drug",
"discovery",
"biochemistry",
"computer",
"sci... | 2011 | Exhaustive Sampling of Docking Poses Reveals Binding Hypotheses for Propafenone Type Inhibitors of P-Glycoprotein |
Borrelia burgdorferi , the bacterial pathogen of Lyme borreliosis , differentially expresses select genes in vivo , likely contributing to microbial persistence and disease . Expression analysis of spirochete genes encoding potential membrane proteins showed that surface-located membrane protein 1 ( lmp1 ) transcripts were expressed at high levels in the infected murine heart , especially during early stages of infection . Mice and humans with diagnosed Lyme borreliosis also developed antibodies against Lmp1 . Deletion of lmp1 severely impaired the pathogen's ability to persist in diverse murine tissues including the heart , and to induce disease , which was restored upon chromosomal complementation of the mutant with the lmp1 gene . Lmp1 performs an immune-related rather than a metabolic function , as its deletion did not affect microbial persistence in immunodeficient mice , but significantly decreased spirochete resistance to the borreliacidal effects of anti-B . burgdorferi sera in a complement-independent manner . These data demonstrate the existence of a virulence factor that helps the pathogen evade host-acquired immune defense and establish persistent infection in mammals .
Lyme borreliosis , caused by Borrelia burgdorferi sensu lato , is the most prevalent tick-borne human disease in the United States , Europe and many parts of Asia [1] . Once the pathogen is deposited in the mammalian dermis by feeding Ixodes ticks , it establishes a localized infection at the bite site , then disseminates to distant cutaneous sites and various internal organs , including the spleen , bladder , joints , heart and central nervous system [1]–[3] . While B . burgdorferi persists in several tissue locations in mammals , only a limited set of organs , most frequently the joints and the heart , experience robust host-inflammatory responses resulting in clinical complications , such as Lyme arthritis and carditis . Antibiotic treatment is usually , but not always , successful , and some patients develop a form of antibiotic-resistant arthritis that is thought to be unrelated to persistent infection [4] . The B . burgdorferi transcriptome undergoes dynamic changes during the complex enzootic cycle of the spirochetes [5]–[8] . B . burgdorferi grown in laboratory medium or within host-implanted dialysis membrane chambers readily responds to altered environments , adapting to changes in temperature , pH , nutrients , and host immune responses [6] , [9]–[15] . A significant fraction of the B . burgdorferi genome ( 8 . 6% ) , or 150 genes , could be differentially expressed in vitro in response to physiochemical alterations in growth conditions , and a major proportion of these genes ( 46% ) encode proteins with predicted export signals [13] . However , while all B . burgdorferi lipoproteins have outer membrane export signals , some are retained in the periplasm by sequence-specific signals [16] . Studies have identified a few B . burgdorferi genes which are preferentially expressed in specific mammalian and arthropod environments and gene deletion studies [17] have confirmed that some of those differentially-expressed gene products support spirochete infectivity . For example , the B . burgdorferi genes bbk32 , dbpA/B and bmpA/B are selectively expressed in mammals and facilitate B . burgdorferi infection of the murine host [8] , [18] , [19] . In contrast , ospA/B , bb0365 and bb0690 are highly expressed during specific stages of B . burgdorferi persistence in ticks and support the spirochete life cycle in the arthropod [7] , [20] , [21] . Other genes , such as ospD is dispensable for infectivity [22] , [23] , regardless of tightly regulated expression in vivo [22] . As gene duplication is a crucial mechanism of evolutionary innovation and the B . burgdorferi genome harbors significant clusters of paralogous genes in addition to large numbers of unique genes with unknown functional annotations [24] , [25] , many spirochete proteins may have functional redundancy . Thus , despite selective expression and beneficial contribution to the spirochete life cycle , antigens could be functionally redundant and non-essential for infectivity . Therefore , further identification of virulence genes that have significant impact on B . burgdorferi survival in vivo and pathogenesis is important for the development of preventative strategies . The clinical complications of Lyme borreliosis are primarily triggered by B . burgdorferi-induced host inflammatory responses [26]–[28] . Although spirochetes colonize a wide variety of host tissues , the inflammatory response that results in pathology is observed in a limited set of host organs , most commonly in one or both mouse ankles or human knees and the heart . The diversity of host niches likely influences spirochete gene expression . While microbial antigens that are expressed at higher levels in a time- or tissue-specific manner may assist in B . burgdorferi persistence in local environments , antigens , especially those exposed on the microbial surface , could directly participate in host–pathogen interactions contributing to the genesis of organ-specific pathogenesis . Therefore , we assessed the expression levels of a selected set of B . burgdorferi genes in diverse murine tissues because of their putative membrane localization . We sought to determine if B . burgdorferi gene products that are preferentially expressed at high levels in clinically-relevant host microenvironments directly contribute to microbial virulence . The characterization of microbial ligands that are differentially expressed during the pathogen's life cycle is important for the identification of novel vaccine targets and the prevention of the multi-system disorders caused by B . burgdorferi .
B . burgdorferi persists in diverse tissue environments of the mammalian host . To identify B . burgdorferi genes that are expressed at high levels in vivo , particularly in a tissue-specific manner , we employed a sensitive quantitative RT-PCR ( qRT-PCR ) approach to compare spirochete transcriptomes in multiple murine tissues and in vitro . A total of 91 spirochete genes were selected for expression analysis , based on their putative association with the spirochete membrane as determined by the database annotation and in silico analysis for extracellular exposure ( Table S1 ) . Groups of C3H/HeN mice ( 5 animals/group ) were challenged with B . burgdorferi ( 105 cells/mouse ) and skin , joints , heart and bladder tissue were collected following 1 , 2 , 3 and 4 weeks of infection . Total RNA was isolated , and corresponding tissues from the indicated time points were combined into four separate pools of skin , joint , heart and bladder samples . qRT-PCR analysis was performed using gene-specific primers as detailed in the Materials and Methods section . Analysis of qRT-PCR data revealed that 44 B . burgdorferi genes ( out of 91 assessed ) were not transcribed at detectable levels in vivo . The remaining 47 genes displayed variable expression across different tissues , which is presented as fold increase in transcript levels relative to flaB , together with corresponding in vitro expression levels ( Figure S1 ) . B . burgdorferi bb0210 , annotated as surface-located membrane protein 1 ( lmp1 ) , which encodes an exported protein with type I signal peptide with unknown function [25] , displayed the most dramatic differential expression in murine tissues , with the highest level of expression found in the heart ( Figure S1 ) . We used the initial qRT-PCR screen as a guide to focus on genes of potential importance , and performed more detailed temporal and spatial expression analyses of lmp1 throughout the B . burgdorferi infection in mice . A similar qRT-PCR experiment ( Figure S1 ) , using separate RNA samples collected at weekly intervals , showed that lmp1 was expressed at high levels between weeks 1 and 2 , but produced at low levels between weeks 3 and 4 ( data not shown ) . Therefore , a detailed expression analysis of lmp1 focused on the early phases of B . burgdorferi infection in the murine host . To accomplish this , groups of C3H/HeN mice ( 5 animals/group ) were infected with B . burgdorferi , and tissues were isolated at 7 , 10 , 15 , and 20 days . Isolated total RNA was converted to cDNA and subjected to qRT-PCR to measure copies of lmp1 transcripts , relative to flaB expression . The expression of lmp1 was selectively upregulated in the heart at 7 and 10 days post-infection , compared to that in the skin , joints , bladder and infected ticks ( Figure 1 ) . Similar to syringe-based infection , the expression of lmp1 was also significantly higher in infected hearts compared to other tissues , when mice were infected via natural tick-borne B . burgdorferi infection . The expression of lmp1 in the murine heart , analyzed at day 7 following B . burgdorferi–infected tick challenge , was significantly higher than corresponding lmp1 expression levels in the infected skin , bladder and joints ( Figure S2 ) . Consistent with lmp1 expression data , mice infected with B . burgdorferi and human patients with diagnosed Lyme disease also developed detectable antibody responses to Lmp1 ( Figure S3 ) . These results suggest that lmp1 encodes an immunogenic antigen that is expressed in vivo , with dramatic expression in the murine heart in both syringe and tick-transmitted B . burgdorferi infection . Therefore , we further examined the role of Lmp1 in B . burgdorferi virulence in mice . Although the function of B . burgdorferi Lmp1 is unknown , the protein displays putative conserved domains including unique repeat modules housed in the central portion of the protein ( Figure 2A ) . Polyclonal antibodies that specifically recognized native B . burgdorferi Lmp1 were generated ( Figure S4 ) and used in a proteinase K accessibility assay , which indicated that Lmp1 is exposed on the microbial surface ( Figure 2B ) . To further study the role of the Lmp1 in B . burgdorferi infectivity , we created lmp1-deficient B . burgdorferi . An isogenic mutant of lmp1 was created by replacing part of the lmp1 open reading frame with a kanamycin resistance cassette via homologous recombination ( Figure 2C ) . A DNA construct was generated for the intended recombination , sequenced to confirm identity and transformed into B . burgdorferi as detailed in the Materials and Methods section . Transformants were further screened using PCR analysis to ensure that the antibiotic cassette was inserted into the intended chromosomal locus ( Figure 2D ) , and that the plasmid profiles of the wild type and mutant spirochetes were identical ( data not shown ) . RT-PCR analysis showed that while lmp1 mRNA was undetectable in the lmp1 mutant , similar to the parental isolate , the mutant was able to transcribe the surrounding genes , bb0209 and bb0211 ( Figure 2E ) . qRT-PCR analysis further confirmed that the transcript levels of bb0209 and bb0211 in the lmp1 mutant were 95 ( ±9% ) and 85 ( ±8% ) of the respective wild type levels ( data not shown ) . The lmp1 mutant spirochetes contained a similar protein profile to that of the wild type ( Figure 2F , left panel ) and , as expected , the lmp1 mutant did not produce Lmp1 protein ( Figure 2F , right panel ) . We next compared the murine infectivity of lmp1 mutant B . burgdorferi with that of the parental isolates . Groups of 10 C3H/HeN mice were inoculated intradermally with equal numbers of wild type or lmp1 mutant B . burgdorferi ( 105 spirochetes/mouse ) . qRT-PCR analysis ( Figure 2G ) and culture ( data not shown ) of murine skin biopsy and blood samples collected following one week of infection indicated that both lmp1 mutants and wild type spirochetes were readily detectable in skin and blood . When Ixodes ticks were allowed to feed on mice inoculated with lmp1-deficient B . burgdorferi after two-weeks of infection , the mutants were able to migrate into feeding ticks ( data not shown ) . Although these observations indicated that lmp1 mutant B . burgdorferi was infectious in the murine host , the mutant was unable to establish persistent infection in mice and failed to induce disease , as assessed by the development of arthritis and carditis . To rule out the possibility that the observed phenotypic defects of the lmp1 mutant B . burgdorferi to infect the murine host were the result of anomalous effects of genetic manipulation , we sought to complement the lmp1 mutant spirochetes with a wild type copy of the lmp1 gene in cis , and use this isolate in murine infection studies . As lmp1 lacks an obvious upstream promoter , we first fused the open reading frame of lmp1 with the B . burgdorferi flaB promoter . The flaB-lmp1 fusion , along with the streptomycin resistance cassette , aadA [29] was then inserted into pXLF14301 [20] for integration into the B . burgdorferi chromosome ( Figure 3A ) . lmp1 mutants were transformed and selected using antibiotics . PCR analysis confirmed that one of the lmp1-complemented spirochete isolates retained all of the B . burgdorferi plasmids present in the parental isolate ( data not shown ) . RT-PCR and immunoblotting showed that the lmp1-complemented isolate produced both lmp1 mRNA ( Figure 3B ) and Lmp1 protein ( Figure 3C ) . Lmp1 mRNA and protein production in the complemented and wild type isolates was further assessed by qRT-PCR and densitometric analysis of the immunoblot ( Figure 3C ) , respectively , and normalized against FlaB production , which indicated that both isolates produced comparable levels of Lmp1 ( data not shown ) . We then compared the ability of the wild type , lmp1 mutant and complemented spirochetes to establish infection and induce disease in the murine host . Groups of 10 C3H/HeN mice were separately inoculated intradermally with wild type , lmp1 mutant and lmp1-complemented B . burgdorferi ( 105 cells/mouse ) . The spirochete burdens in heart , skin , bladder and joints were evaluated at day 7 , 10 , 15 , 21 and 28 following B . burgdorferi infection . The results showed that , except for the initial time point ( day 7 ) , lmp1 mutants were severely impaired in their ability to colonize all murine tissues and were undetectable in the heart after 10 days of infection ( Figure 3D ) . Similarly , culture analysis of murine heart isolated after 15 days of B . burgdorferi infection showed that the lmp1 mutant spirochetes could not be recovered ( data not shown ) . In contrast , both wild type and lmp1-complemented B . burgdorferi readily persisted in all tested murine tissues throughout infection , with significantly higher burdens than lmp1-deficient spirochetes–heart ( P<0 . 001 ) , skin ( P<0 . 006 ) , bladder ( P<0 . 004 ) and joints ( P<0 . 002 ) . Both wild type and lmp1-complemented B . burgdorferi caused severe inflammation , but lmp1 mutants induced less severe disease , as reflected by the histopathological signs of carditis ( Figure 4A and 4B ) , development of swelling in the tibiotarsal joints ( Figure 4C ) , and histopathological signs of arthritis ( Figure 4D ) . In most cases , the tissue burdens of lmp1-complemented isolates ( Figure 3D ) and their ability to induce inflammation ( Figure 4B and 4C ) were comparable , yet significantly lower ( P<0 . 05 ) , than those of the wild type isolates , possibly due to the constitutive expression of lmp1 from the heterologous flaB promoter which might be detrimental to spirochetes . The analysis of spirochete growth in the culture media indicated that the wild type , lmp1 mutants and lmp1-complemented isolates follow a similar growth pattern without significant variation in motility or obvious morphological defects ( data not shown ) . However , lmp1 mutants were impaired to persist in the immunocompetent murine hosts following the first week of infection . We next assessed whether Lmp1 function in vivo is related to the metabolic or immune environment of the host . To accomplish this , we compared the infectivity of wild type and lmp1 mutant B . burgdorferi in the established immunodeficient murine model of Lyme borreliosis using severe combined immunodeficient ( SCID ) mice [30] . Groups of SCID mice ( 3 animals/group ) were inoculated intradermally with 105 wild type , lmp1 mutant , or lmp1-complemented B . burgdorferi . The spirochete burdens in the heart , skin , bladder , and joints were evaluated at day 7 , 14 and 21 following B . burgdorferi infection using quantitative PCR . In parallel , the viability of the spirochete was determined by culture of murine blood and spleen isolated at day 7 , 14 and 21 . The results showed that the wild type , lmp1 mutant and lmp1-complemented B . burgdorferi could be cultured from murine tissues at all time points ( data not shown ) and that there was no significant difference in B . burgdorferi burdens in all tested murine tissues throughout the infection ( Figure 5A ) . Consistent with similar burdens of wild type and genetically-manipulated pathogens in SCID mice , one of the innate immune mechanisms , the ability of macrophages to phagocytose invading pathogens , which is important in controlling B . burgdorferi infection , did not differ between wild type and lmp1 mutants ( Figure 5B ) . As neutralizing antibodies that develop in B . burgdorferi infected mammals are primarily responsible for controlling spirochete burden in the host , we explored whether reduced virulence of lmp1mutant spirochetes correlates with the susceptibility of B . burgdorferi to the borreliacidal activity of immune sera . To accomplish this , wild type , lmp1 mutant and lmp1-complemented B . burgdorferi were exposed to antisera collected from B . burgdorferi–infected C3H mice . lmp1 mutants were significantly more susceptible to the bactericidal activities of the anti-B . burgdorferi sera than wild type B . burgdorferi , and were protected by genetic complementation with lmp1 ( Figure 5C ) . The susceptibility of lmp1 mutants to the bactericidal activities of the immune sera did not differ significantly when using active or heat-inactivated serum ( data not shown ) , indicating that lmp1 deletion enhances the borreliacidal effects of antibodies in a complement independent manner . Like parental isolates , the lmp1 mutants were not susceptible to bactericidal activities by the non-immune serum collected from naïve mice ( data not shown ) , suggesting that Lmp1 is not required for serum resistance by spirochetes . Together , these data suggest that Lmp1 contributes to B . burgdorferi defense against host-acquired immune responses by enhancing resistance to the bactericidal antibodies that develop during infection .
In nature , B . burgdorferi is maintained through a complex enzootic cycle [1] . Once transmitted to mammals , B . burgdorferi can establish persistent infection in a variety of tissue locations . Limited studies suggest that spirochete genes expressed in higher levels in infected host tissues could be important for B . burgdorferi survival [8] . Since membrane proteins could have a direct contribution to the adaptation of pathogens to host environments , we assessed the expression of 91 B . burgdorferi genes encoding potential membrane proteins covering spirochete infectivity in multiple murine tissues . Our data show that few of the genes analyzed are differentially or highly expressed in the selected tissues . Targeted deletion of one of the spirochete genes that is highly expressed in cardiac tissue , lmp1 , while resulting in the initial clearance of pathogen burden in infected hearts , also affected overall virulence of B . burgdorferi in murine infectivity and reduced the outcome of Lyme disease . Our results show that lmp1 mutants persist in SCID mice at similar levels to parental isolates and are susceptible to borreliacidal antibody-mediated killing in vitro , suggesting that Lmp1 contributes to B . burgdorferi defense against host-acquired immune responses . Identification of hitherto unrecognized virulence genes of B . burgdorferi , such as lmp1 , that support pathogen infectivity in mammals could shed light on the pathogenesis and prevention of Lyme disease . Genes that are selectively expressed in vivo could be important for B . burgdorferi persistence in nature , possibly allowing spirochete adaptation to highly heterogeneous metabolic and immune environments . While assessment of pathogen gene expression in vivo is an important prerequisite to understanding microbial pathogenesis , microarray analysis is of limited use for the assessment of the B . burgdorferi transcriptome in vivo , primarily due to the low level of pathogen RNA in infected tissues [31] . Microarray studies were attempted after isolation and amplification of microbial RNA [7] , [31] , but this requires longer periods of RNA manipulation and risks degradation or clonal alteration during amplification and processing of transcripts . In contrast , optimal quantities of B . burgdorferi RNA can be isolated from spirochetes grown in vitro or in a host-implanted dialysis membrane , and microarray-based studies have been used to assess B . burgdorferi gene expression in these ‘host-like’ conditions that have yielded important information on the role of B . burgdorferi genes in pathogen infectivity [9] , [12] , [13] , [15] , [32] . However , the host environment is too complex and dynamic to be duplicated artificially . Therefore , we employed a sensitive qRT-PCR approach for the direct assessment of pathogen gene expression in vivo . This method is reproducible , as two independent sets of animal infection studies identified the same set of 47 B . burgdorferi genes as expressed in vivo , and a majority of them displayed higher expression levels in mice , than in vitro . The variable expression of these genes in multiple tissue locations possibly reflects the adaptive responses of the pathogen to local host environments , enabling immune evasion , adhesion , or nutrient uptake , among other possibilities . Furthermore , regardless of their functional role in pathogen persistence , antigens that are highly produced in certain host sites , such as the joints and heart , could participate in the genesis of inflammatory disease . Our qRT-PCR analysis also identified a set of 44 B . burgdorferi genes that may not be important for mammalian infectivity , as none of these displayed detectable transcription within the first 4 weeks of infection . On the other hand , a set of 26 genes encoding potential lipoproteins displayed detectable expression in vivo and , with the exception of bbo40 , expression of many genes ( bb0806 , bbb09 , bbd10 , bbj09 , bbl39 , bbm27 , bbn38 , bbn39 , bbq47 , bbs41 ) agreed with a previous study that evaluated the expression of B . burgdorferi lipoproteins in murine dermis [5] . Although our expression analysis identified selected B . burgdorferi transcripts highly expressed in vivo , rather than proteins , these mRNA are likely the signatures of translated antigens . This speculation is supported by the recent study that identified 103 spirochete immunogens by screening in vitro translated genome-wide proteomic arrays with B . burgdorferi-specific immune sera [33] . Our target list of potential membrane proteins ( Table S1 ) overlaps 10 of these identified immunogens ( BB0543 , BBB09 , BBH06 , BBL39 , BBM27 , BBN38 , BBN39 , BBO39 , BBO40 and BBS41 ) . Each of these immunogens was represented in our identified set of B . burgdorferi genes expressed in murine tissues ( Figure S1 ) . Our study established B . burgdorferi lmp1 as a spirochete gene that is highly expressed in the early stages of mammalian infection , most notably in the murine heart . Lmp1 was first described in an earlier study [34] which highlighted the protein as a potential B . burgdorferi adhesin and indicated that Lmp1-specific antibodies develop in B . burgdorferi–infected mammals . lmp1 does not belong to the paralogous gene families in B . burgdorferi [24] and is thought to encode a relatively large 128-kDa outer membrane protein with putative type I signal peptide . It is not clear if the leader peptide is indeed cleaved or if the hydrophobic N-terminus serves as a membrane anchor; however , the antigen is exposed on the spirochete surface . Lmp1 retains 86–88% amino acid identity across orthologs in the related infectious spirochetes B . afzelii and B . garinii , with the highest sequence conservation in the amino and carboxyl termini . Analysis of the lmp1 locus indicates that the gene overlaps with the immediate upstream gene , bb0209 , and shares a short intergenic region with the downstream gene , bb0211 , suggesting these three genes are likely part of an operon; however , the lmp1 mutant is able to express both bb0209 and bb0211 . Notably , bb0211 encodes for the DNA mismatch repair protein , MutL , and therefore likely bears a house-keeping function in B . burgdorferi biology . Although the functions of bb0209 and lmp1 are unknown , the amino terminus of BB0209 and the carboxyl terminus of Lmp1 both contain multiple tetratricopeptide repeat domains [35] , which are shown to mediate protein-protein interaction and the assembly of the pilus and multiprotein complexes [36] . In addition , the central region of Lmp1 harbors a unique cluster of seven repeat motifs , each consisting of 54 amino acids , which is highly conserved amongst Lmp1 orthologs and potentially participates in adherence to host components [34] . Our studies further establish that Lmp1 is a surface-exposed virulence factor of B . burgdorferi , as lmp1 deletion severely interferes with spirochete infectivity and pathogenesis . After its inoculation in the host , B . burgdorferi remains locally in the dermis for a few days and then disseminates to distant organs , likely via the bloodstream , during the first week of infection [37] , [38] . Our data indicate that Lmp1 function may not be important during early infection , including spirochete dissemination , since no apparent differences in the burdens of wild type spirochetes and lmp1 mutants were observed in murine blood samples collected at day 3 , 5 ( data not shown ) and 7 ( Figure 2G ) . Instead , lmp1 mutants displayed a severe reduction in numbers in the disseminated organs by day 10 , implying that Lmp1 is critical for the persistence of the spirochete in murine tissues . In agreement with the higher level of expression of lmp1 in the cardiac tissue , as compared to other organs , a greater effect of lmp1 deletion was also observed in the heart where the mutants were selectively eliminated after 10 days of infection , suggesting that Lmp1 plays a dominant role in spirochete infection of the heart . B . burgdorferi burdens in mice began to decline after first two weeks of infection [37] , which coincides with the development of the acquired immune response , such as neutralizing antibodies that controls spirochete infection [26] , [28] . This is consistent with the observation that lmp1 mutants persist in similar levels to the parental isolate in the heart at day 7 , despite dramatic wild type expression of lmp1 , but begin to decline during the second week . The ability of lmp1 mutants to survive in SCID mice further indicates that the function of the Lmp1 is not related to metabolic requirements of the spirochete survival in vivo , as shown for B . burgdorferi PncA [39] or AdeC [40] , but rather is associated with B . burgdorferi survival in host immune environments . B . burgdorferi isolates missing the lp28-1 plasmid [41] , which houses VlsE [42] , [43] , display similar impaired host persistence in immunocompetent mice , as VlsE likely confers protection against host-generated borreliacidal antibodies . Therefore , while lmp1 is variably expressed in diverse tissue environments and carries a predominant role in spirochete persistence in the heart , a basal level of lmp1 expression is noted throughout the murine infection which , based on our in vitro data ( Figure 5C ) , might contribute to the protection of B . burgdorferi against host-acquired immune responses , as was recently proposed for B . burgdorferi OspA in feeding ticks [44] . In summary , we have identified a select set of B . burgdorferi genes encoding potential membrane proteins that are expressed during murine infection . Many of these in vivo-expressed genes are differentially expressed in various host tissues , including joints and heart , and can participate in pathogen persistence and the genesis of disease . Here , we present direct evidence that one microbial gene expressed at higher level in the cardiac tissue , lmp1 , encodes an essential virulence factor that plays an important role in immune evasion and dramatically influences spirochete persistence in murine tissues and the genesis of inflammation . Whereas previously identified B . burgdorferi virulence antigens are mostly plasmid-borne , and thus have greater instability and sequence divergence , Lmp1 is chromosomally encoded and is relatively conserved among orthologs in related infectious spirochetes . Further identification of B . burgdorferi virulence determinants that actively support spirochete persistence in vivo could contribute to the development of effective therapeutic strategies against Lyme borreliosis .
Borrelia burgdorferi infectious isolate A3 [45] , a clonal derivative of the B . burgdorferi whole genome sequenced strain B31 M1 [24] , [25] , was used in this study . Four- to six-week old female C3H/HeN and pathogen-free NCr-SCID mice were purchased from the National Institutes of Health . Mice were inoculated with a single subcutaneous injection of 105 spirochetes per mouse . All animal procedures were performed in compliance with the guidelines and with the approval of the Institutional Animal Care and Use Committee . Ixodes scapularis ticks used in this study belong to a colony that has been reared and maintained in the laboratory . The identity and oligonucleotide primer sequences for the quantitative RT-PCR analysis of B . burgdorferi genes are indicated in Table S1 . The B . burgdorferi target genes [24] , [25]were selected based on their predicted localization on the spirochete membrane according to the database annotation ( www . tigr . org ) and PSORT in silico analysis [46] . Groups of mice ( 5 animals/group ) were infected with B . burgdorferi ( 105 spirochetes/mouse ) , and samples of skin , heart , tibiotarsal joint and bladder were collected and frozen in liquid nitrogen at one-week intervals between 1 and 4 weeks of infection . Total RNA was extracted from tissue samples using the TRIzol reagent ( Invitrogen ) . To reduce traces of contaminating DNA , samples were further digested with RNase-free DNaseI ( Qiagen ) , purified using the RNeasy kit ( Qiagen ) and reverse transcribed to cDNA using the AffinityScript cDNA synthesis kit ( Stratagene ) . The relative levels of B . burgdorferi cDNA in each sample were assessed by quantitative PCR ( qPCR ) , and DNA contamination in each sample was measured using an equal volume of purified RNA as a template . Samples from each time point were pooled by tissue type , and final pools of skin , heart , joints and bladder were used in the qPCR analysis . The primers used for qPCR reaction were designed using OligoPerfect Primer design software ( Invitrogen ) based on the B . burgdorferi B31 M1 genomic sequence [24] , [25] . All PCR primer pairs had a similar annealing temperature ( 60°C ) and spanned 100–300 base pairs of each of the target B . burgdorferi genes . Each primer pair was tested for efficiency and non-specific amplification by melt-curve analysis using B . burgdorferi genomic DNA as a template . In one case of paralogous genes , the same set of primers was assigned for the detection of both genes as indicated in Table S1 . To generate reliable in vivo gene expression data and to further ensure specific amplification of B . burgdorferi cDNA in murine tissue samples , the qPCR amplification in each well was followed by melt-curve analysis , and wells showing non-specific amplification were discarded from data analysis . The amplification cycle consisted of initial denaturation at 95°C for 5 min followed by 45 cycles each at 95°C for 10 sec , 60°C for 20 sec and 72°C for 30 sec and final melt curve analysis: 55°C for 30 sec , increase 0 . 5°C per cycle to 95°C . The amplification was performed in an iQ5 real-time thermal cycler ( Bio-Rad ) using SYBR Green Supermix ( Bio-Rad ) as detailed . For expression screening of B . burgdorferi genes , we simultaneously assayed 8 candidate genes in each 96-well PCR plate using duplicate wells of template cDNA ( skin , heart , joint and bladder samples ) with parallel positive ( B . burgdorferi genomic DNA ) and negative ( no template ) controls . Transcript levels of individual genes were assessed in spirochetes grown in vitro in BSK medium ( 107 cells/ml ) and in each of the murine samples , calculated using the 2−ΔΔCt method [47] , normalized against flaB transcripts and presented as fold increase in gene expression . Two independent mouse experiments used the same parameters of gene expression analysis to ensure the reproducibility of the assay . For detailed temporal and spatial analysis of lmp1 expression by qRT-PCR analysis , amounts of target transcripts were calculated from standard curves prepared from known quantities of flaB and lmp1 DNA as described [21] , [48] . Mice ( 5 animals/group ) were infected via single intradermal injection with B . burgdorferi ( 105 spirochetes/mouse ) or via tick feeding using B . burgdorferi–infected nymphs . Infected murine samples , the heart , skin , bladder and tibiotarsal joints , were removed at different timepoints and frozen in liquid nitrogen . B . burgdorferi–infected ticks were isolated by allowing ticks to feed on 15-day infected mice as described [7] . For quantitative measurement of B . burgdorferi burden in infected tissues , flaB transcripts were measured in infected samples and normalized to mouse or tick β-actin levels . The oligonucleotide primers used for mutagenesis and genetic complementation of B . burgdorferi are indicated in Table S2 . The lmp1-deficient B . burgdorferi was created by exchanging a 2068 base pair DNA fragment encompassing the 5′ terminus of the lmp1 gene with a kanamycin-resistance cassette via homologous recombination as described [8] . Briefly , DNA fragments flanking up- and downstream of the lmp1 gene were PCR-amplified using primers P1–P4 and inserted into two multiple-cloning sites flanking the kanAn cassette in plasmid pXLF10601 [49] . This plasmid was sequenced to confirm the identity of the insert and electroporated into wild type B . burgdorferi B31 isolate A3 . Transformants were selected for growth in the presence of kanamycin ( 350 µg/ml ) . Ten clones were isolated and PCR analysis was used to confirm the intended recombination event using primers P1–P9 . The presence of all endogenous plasmids contained in the parental A3 isolate was also assessed in the mutant clones as described [8] . One of the lmp1 mutant clones that retained the same complete set of plasmids as the wild type isolate was used in additional experiments . Genetic complementation of the lmp1 mutant was achieved by re-insertion of a wild type copy of the lmp1 gene in the B . burgdorferi chromosome [20] . The upstream of the lmp1 open reading frame ( ORF ) overlaps with the preceding gene by a few nucleotides , lacking an intergenic region with a discernible promoter . We , therefore , fused the lmp1 ORF with the B . burgdorferi flaB promoter [50] . Two B . burgdorferi DNA fragments encompassing the full-length lmp1 gene and the flaB promoter were PCR-amplified , fused and cloned into the BamHI and SalI sites of pKFSS1 housing a streptomycin-resistance cassette ( aadA ) [29] . A DNA element containing the flaB promoter-lmp1 gene fusion and the aadA cassette was cut with BamHI and SmaI from the recombinant plasmid pKFSS1-lmp1 and inserted into the corresponding restriction sites of the plasmid pXLF14301 [20] that contains the required 5′ and 3′ arms for homologous recombination in the B . burgdorferi chromosomal locus bb0444–0446 . The plasmid construct was sequenced to confirm its identity and 25 µg of the plasmid DNA was electroporated into the lmp1 mutant . Four clones were isolated by their ability to grow in the presence of both kanamycin and streptomycin . PCR analysis was used to confirm the intended recombination event , and one of the lmp1-complemented clones that contained the same plasmid profiles as the wild type was chosen for further study . For in vitro growth analysis , an equal number of wild type and genetically-manipulated spirochetes were diluted to a density of 105 cells/ml and grown at 33°C in BSK-H medium until they reached the stationary phase ( 108 cells/ml ) . Aliquots of spirochetes were assessed every 12 hours , under a dark-field microscope , for motility , cell clumping and numbers of spirochetes counted using a Petroff-Hausser cell counter . For phenotypic analysis of lmp1 mutants and lmp1-complemented isolates in vivo , B . burgdorferi were injected into groups of mice ( 10 animals/group ) via needle-inoculation ( 105 spirochetes/mouse ) intradermally on the back . Mice were sacrificed at 7 , 10 , 15 , 21 and 28 days following inoculation . Skin , heart , joints , bladder and blood samples were collected and B . burgdorferi burdens were measured by quantitative PCR analysis as previously described [7] , [8] . Development of joint swelling in the infected mice was also evaluated at similar time points as for spirochete burdens . For histological evaluation , heart and joint samples were collected at days 15 , 21 and 28 following inoculation . For tick acquisition studies , groups of mice ( 5 animals/group , 20 ticks/mouse ) were fed on by naïve ticks following two weeks of B . burgdorferi infection . The ticks were allowed to feed to repletion and were immediately analyzed for quantitative RT-PCR measurement of B . burgdorferi burden as detailed earlier [7] . Generation of murine polyclonal antibodies against recombinant Lmp1 , ELISA and immunoblotting were performed as described [7] , [8] . Recombinant Lmp1 protein was produced in E . coli using the bacterial expression vector pGEX-6P1 ( Amersham-Pharmacia Biotech ) with specific primers as indicated in Table S2 . Expression , purification and enzymatic cleavage of the glutathione transferase ( GST ) fusion proteins were carried out as detailed [7] , [8] . Sixteen serum samples from humans with a clinical history of Lyme disease , collected from the CDC Lyme patient serum panel were used in the ELISA . Five serum samples from normal individuals residing in non-endemic areas for Lyme disease were also collected from CDC and used as negative controls . For immunoblotting , recombinant Lmp1 ( 0 . 05 µg/lane ) was resolved on a SDS-PAGE gel and probed with 1∶ 1000 dilution of murine or human serum . Murine antiserum was collected from a group of 5 mice , 15 days after infection with B . burgdorferi via syringe inoculation ( 105 cells/mouse ) . Murine antiserum generated against recombinant Lmp1 was used in a Proteinase K accessibility assay to determine surface exposure of the Lmp1 as described [48] . Bone marrow derived macrophages were isolated from naïve C3H mice as described [51] and cultured for 5 days at 37°C ( 2×105 cells/well ) in L929-conditioned DMEM media . The cells were then washed and resuspended in serum-free DMEM and incubated with B . burgdorferi at a multiplicity of infection ( MOI ) of 10 at 37°C for 2 hours . Cells were washed with cold PBS to remove unbound B . burgdorferi and fixed in 3 . 7% paraformaldehyde , and processed for confocal immunofluorescence as described [52] . Spirochetes , cellular actin and nuclei were detected using FITC-labeled anti-B . burgdorferi goat IgG ( KPL ) , phalloidin-Texas Red ( Invitrogen ) , and DAPI ( Invitrogen ) , respectively . Susceptibility of wild type or genetically manipulated spirochetes to borreliacidal activities in infected mouse sera was performed as described [53] . Briefly , spirochetes were grown in BSK-H medium to a density of 107 spirochetes/ml and exposed to active or heat-inactivated 50% sera isolated from mice infected for 15 days with B . burgdorferi . Samples were incubated at 33°C for 24 to 48 hours and spirochete viability was assessed using dark-field microscopy . Susceptibility of spirochetes to borreliacidal activities in infected mouse sera were also tested in parallel using a combination of two vital stains that specifically label live and dead spirochetes , as described [8] . Briefly , spirochetes were incubated for 48 hours and labeled with the live/dead BacLight Viability kit ( Invitrogen ) according to the manufacturer's instructions . B . burgdorferi–infected mice were examined for swelling of the tibiotarsal joints as detailed earlier [8] . Ankle joints from each of the rear legs of each mouse were measured using a precision metric caliper in a blinded fashion . The thickest diameters of the tibiotarsal joints were measured in each mouse prior to B . burgdorferi infection , and the development of ankle swellings was monitored and tabulated on a weekly basis until the sacrifice of the mice . For histological evaluations of arthritis and carditis , at least 5 ankle joints and 5 hearts were collected from each group of mice ( 5 animal/group ) infected with the different isolates . For histology , joints and hearts ( cut in half through bisections across the atria and ventricles ) were fixed in 10% formalin and processed for Hematoxylin and Eosin staining . Twenty randomly chosen sections from each mouse group were assessed for histopathological comparisons . Signs of arthritis were evaluated as described [8] , based on a combined assessment of histological parameters of B . burgdorferi-induced inflammation [54]–[56] , such as exudation of fibrin and inflammatory cells into the joints , alteration in the thickness of tendons or ligament sheaths , and hypertrophy and hyperplexia of the synovium . Signs of carditis [54] , [57] were evaluated based on the cardiac inflammatory infiltrate , including the transmural infiltration of neutrophils in the blood vessels and infiltration of surrounding connective tissue with macrophages . Carditis was scored on a scale of 0 ( no inflammation ) , 1 ( mild inflammation with less than two small foci of infiltration ) , 2 ( moderate inflammation with 2 or more foci of infiltration ) , or 3 ( severe inflammation with focal and diffuse infiltration covering a large area ) . Both joint and heart tissue sections were blindly examined by two independent researchers . Results are expressed as the mean±standard deviation ( SD ) or standard error mean ( SEM ) . The significance of the difference between the mean values of the groups was evaluated by two-tailed Student t test . | The pathogen of Lyme borreliosis , Borrelia burgdorferi , causes disease in many parts of the world , resulting in multi-system complications in infected humans and animals . The microbe produces certain antigens in response to host environments that potentially allow it to persist and cause disease . Here , we analyzed the expression of B . burgdorferi genes encoding potential membrane proteins in infected hosts and show that one of them , termed Lmp1 , is dramatically expressed in infected mice , most prominently in cardiac tissue during early infection . Mice and humans diagnosed with Lyme borreliosis also develop antibodies against Lmp1 . Deletion of lmp1 in an infectious isolate of B . burgdorferi impairs the pathogen's ability to persist in murine tissues , especially the heart , and to induce disease , which was reversed when the gene was inserted back into the chromosome of the mutant . Lmp1 performs an immune-related , rather than a metabolic , function as its deletion does not affect microbial persistence in immunodeficient mice , but decreases the spirochete's ability to resist the borreliacidal effects of anti-B . burgdorferi sera . These data identify the existence of a surface-located antigen of B . burgdorferi that helps the pathogen evade host-acquired immune defense and establish persistent infection and disease in mammals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology/immunity",
"to",
"infections",
"microbiology",
"infectious",
"diseases/bacterial",
"infections",
"molecular",
"biology",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] | 2009 | A Chromosomally Encoded Virulence Factor Protects the Lyme Disease Pathogen against Host-Adaptive Immunity |
Once considered a phenotypically monomorphic bacterium , there is a growing body of work demonstrating heterogeneity among Mycobacterium tuberculosis ( Mtb ) strains in clinically relevant characteristics , including virulence and response to antibiotics . However , the genetic and molecular basis for most phenotypic differences among Mtb strains remains unknown . To investigate the basis of strain variation in Mtb , we performed genome-wide transposon mutagenesis coupled with next-generation sequencing ( TnSeq ) for a panel of Mtb clinical isolates and the reference strain H37Rv to compare genetic requirements for in vitro growth across these strains . We developed an analytic approach to identify quantitative differences in genetic requirements between these genetically diverse strains , which vary in genomic structure and gene content . Using this methodology , we found differences between strains in their requirements for genes involved in fundamental cellular processes , including redox homeostasis and central carbon metabolism . Among the genes with differential requirements were katG , which encodes the activator of the first-line antitubercular agent isoniazid , and glcB , which encodes malate synthase , the target of a novel small-molecule inhibitor . Differences among strains in their requirement for katG and glcB predicted differences in their response to these antimicrobial agents . Importantly , these strain-specific differences in antibiotic response could not be predicted by genetic variants identified through whole genome sequencing or by gene expression analysis . Our results provide novel insight into the basis of variation among Mtb strains and demonstrate that TnSeq is a scalable method to predict clinically important phenotypic differences among Mtb strains .
A hallmark of infection with Mycobacterium tuberculosis ( Mtb ) is the high degree of variability in disease course and response to therapy . This heterogeneity in Mtb infection outcome and treatment response has traditionally been attributed to variability in host determinants . Yet it is increasingly apparent that meaningful differences exist among Mtb strains in features that impact immunogenicity , virulence , and response to antibiotic treatment [1–4] . Despite the potential consequences of strain variation for the development of new diagnostics , drugs , and vaccines , predicting and defining the causal genetic determinants of biologically important phenotypic variation between strains remains a challenge . Our understanding of strain heterogeneity has been vastly improved through the advent of affordable next-generation sequencing technologies . This has facilitated the whole genome sequencing ( WGS ) of thousands of Mtb clinical strains and uncovered numerous sequence variants including single nucleotide polymorphisms ( SNPs ) , insertions-deletions ( in-dels ) , large sequences polymorphisms ( LSPs ) , and insertion element transpositions [5–11] . Sequence variation is used as a proxy for phenotypic diversity , yet the functional consequences of most polymorphisms in the Mtb genome are not known . The field’s ability to predict phenotype from genotype is most well developed for resistances to clinically important first and second line drugs , where large population-based studies have been successful in identifying genetic determinants of antibiotic resistance [6 , 8 , 11–13] . Yet extending such analyses to predict responses to new antibiotics or to predicting more complex phenotypes such as virulence remains difficult . In the setting of antibiotic development , sequence conservation is often used to indicate conservation of gene function , but does not reflect capacity for the emergence of resistance or plasticity in the underlying genetic and cellular networks that may create strain-to-strain differences in antibiotic efficacy . Thus , there is a need for systematic and high-throughput methods to predict phenotypic variation among Mtb strains , especially for clinically relevant phenotypes such as antibiotic susceptibility , likelihood of drug resistance , and virulence . TnSeq , genome-wide transposon mutagenesis coupled with next-generation sequencing , has emerged as a high-throughput approach to define the contributions of genes and genetic networks to microbial phenotypes [14–21] . TnSeq experiments reveal the fitness cost of gene disruption , and thus can identify both global and context-specific genetic requirements . In Mtb , TnSeq has been used to identify genes and pathways required across a variety of contexts , including defining the genes essential for in vitro growth , which has provided a roadmap for antibiotic development [17–19] . Most TnSeq studies are performed on reference strains , including those in Mtb , which , to date , have been conducted exclusively with laboratory-adapted reference strains . In the published literature , a small number of TnSeq studies in other pathogens have performed pairwise comparisons between a reference strain and a single clinical isolate to define differences in the genetic requirements for growth in vitro or under antibiotic stress [14 , 22 , 23] . These studies have identified strain-to-strain differences in requirements for key pathways including aspects of central metabolism . However , it is unclear how representative these findings are since the choice of a single clinical isolate is relatively arbitrary . Comparisons across larger panels of non-reference ( clinical ) strains have not been performed in part because there has been no systematic methodology to analyze TnSeq data across multiple genetic backgrounds . Differences in genome structure and gene content between strains create significant challenges for TnSeq analysis since existing approaches rely heavily on mapping sequencing reads to a reference genome . Gene deletions can result in false gene essentiality calls , gene duplications may result in false gene non-essentiality calls , and sequence variants that alter transposon insertion sites can create more subtle errors . Thus , an analytic approach that permits rigorous statistical analysis in the setting of genomic variation is needed . In this work we developed a pipeline to use TnSeq to comprehensively define genetic requirements for in vitro growth across a diverse panel of nine Mtb strains , with the goal of predicting strain-specific differences in the requirements for genes that are the target of antibiotic development . In contrast to sequence data , which has limited phenotypic predictive power without prior knowledge about causal variants , the genetic requirements identified through TnSeq directly reflect the cost of gene disruption on microbial growth . Therefore , differences in genetic requirements identified from our TnSeq data allowed us to successfully predict differences between strains in the responses to both a widely used and a novel antibiotic . Importantly , these differences could not be readily predicted from WGS or gene expression data . Our work demonstrates the power of high-throughput mapping of genetic requirements via TnSeq to uncover clinically relevant phenotypic differences between Mtb strains .
It has become clear from a growing number of in vitro , animal model , and epidemiological studies that important phenotypic differences exist among Mtb strains [24] . Yet how to predict phenotypic variation among strains systematically is unclear . Because TnSeq provides a direct read out of the genes required for growth under a given condition , we hypothesized that TnSeq might be a robust method to address this gap . We therefore sought to use TnSeq to define differences in the genetic requirements for in vitro growth across a panel of genetically diverse Mtb clinical strains . We focused on a panel of 8 Mtb clinical strains ( Fig 1A ) . These strains had previously been classified by large sequence polymorphisms ( LSP ) , a low-resolution method of genotyping , as belonging to three of the major Mtb phylogenetic lineages [25] , and the most prevalent lineages currently in circulation: Euro-American , East Asian , and Indo-Oceanic ( Fig 1A ) . To define the genetic requirements for in vitro growth for these strains , we performed transposon mutagenesis using the modified Himar1 transposon [26] and included the reference strain H37Rv as a control . Transposon libraries were generated in biological duplicate for each clinical strain , and in triplicate for the control strain , H37Rv . The libraries were then subjected to transposon-junction sequencing [27] . Because interpretation of TnSeq data relies on the accurate mapping of transposon-junction reads to the genome of the mutagenized strain , data analysis can be confounded by sequence variants . To account for the genetic variation in our panel of clinical strains , we performed WGS on each strain , followed by reference-based assembly , annotation , and variant calling ( Fig 1B , Materials and Methods ) . WGS revealed that , excluding repetitive elements , such as genes in the PE/PPE family and insertion elements ( S1 Table ) which are difficult to confidently describe due to the technological limitations of short-read next-generation sequencing , between 3 and 24 genes were deleted in each strain ( S2 Table ) . For the most part , these deletions corresponded to the previously described regions of difference [28] . We also identified large-scale duplications present in a subset of the clinical strains , encompassing up to 309 genes ( S3 Table ) . Similar duplications have previously been observed in Mtb strains and appear to be the result of unequal homologous recombination mediated by IS6110 elements [29–31] . In addition to structural variants , we identified genes disrupted by nonsense and frameshift mutations and insertion elements ( S4 Table ) . Several hundred SNPs between each strain and the reference genome , including non-synonymous polymorphisms in the coding regions of a number of genes , were also identified ( S5 Table ) . Transposon-junction reads were then mapped to each strain’s genome assembly ( Fig 1B ) , revealing that the Mtb clinical strains were amenable to transposon mutagenesis , with ~50 to 70% of Himar1 transposon insertion sites ( TA dinucleotides ) in each library containing one or more insertions ( S6 Table ) . This is comparable to the saturation of the H37Rv libraries generated in this study and previously published Mtb transposon libraries [17 , 19] . Library reproducibility was also high , as determined by comparing transposon insertion count across each gene between library replicates ( Spearman correlation coefficient = 0 . 81–0 . 99 , S6 Table ) . We then sought to develop a rigorous analytic approach to compare transposon-junction sequencing data across these genetically diverse strains in order to identify differences in their genetic requirements for in vitro growth . TnSeq is used to identify essential genes under a given growth condition , defined as genes unable to tolerate transposon insertions due to the severe fitness cost of disruption . Doing so involves statistical analyses of transposon insertion location and frequency across a coding region , ultimately making a qualitative classification ( essential v . non-essential ) from quantitative insertion count data . Yet performing discrete comparisons of essential genes between strains or conditions overlooks quantitative differences in transposon count between genes with the same qualitative classification . For example , a gene able to tolerate transposon insertions in two strains could be classified as non-essential in both strains , even if the relative abundance of these transposon mutants in their respective library pools differs significantly . Such differences in transposon mutant abundance reflect differences in the fitness cost of gene disruption in a particular genetic background , and indicate the relative dependency of a strain on the gene for growth and survival [16 , 32 , 33] . We sought to identify genes with such differential requirements among our panel of strains , incorporating the sequence variants identified through our WGS pipeline . To identify differences in the genetic requirements for growth among strains , we used a permutation test-based method to identify genes with statistically significant differences in transposon insertion count between strains [34 , 35] . Briefly , this method calculates the difference in transposon insertion count across a gene between two strains and compares this value to a null distribution of differences generated by randomly reshuffling the insertion counts among the data sets being compared , identifying genes that have a difference in insertion count more extreme than would be expected by chance . To account for the sequence variants identified through WGS and permit comparative analyses between strains , we placed the TA dinucleotides onto a common indexing system based on coordinates in H37Rv . Deleted genes , genes in the duplicated region , and repetitive elements , such as PE/PPE/PGRS genes and insertion elements , were excluded from subsequent analyses ( S1 Fig ) . Pairwise comparison of each clinical strain’s transposon libraries to the H37Rv control libraries identified between ~10 and ~50 genes with a statistically significant difference in transposon insertion count per strain ( adjusted P-value <0 . 05 and magnitude fold-change ≥ 2 ) ( S7 Table ) . Many of the genes with differential genetic requirements are involved in fundamental cellular processes ( Fig 2 ) . A number are involved in redox homeostasis , including the catalase-peroxidase ( katG , increased genetic requirement in all clinical strains compared to H37Rv ) , genes in the mycothiol synthesis pathway ( mshA , increased requirement in East Asian 631; mshC , increased requirement in Euro-American 630 , 663 , East Asian 621 , 631 , Indo-Oceanic 641 , 667 ) , and genes involved in sulfur uptake and assimilation ( cysW , cysT , subI , cysN , cysA1 , increased requirement in Euro-American 630 , 663 ) , processes that support synthesis of mycothiol and other biomolecules involved in redox homeostasis such as thioredoxins ( Fig 2 , S7 Table ) . There were also numerous differences among strains in the requirements for genes whose products mediate central metabolism ( Fig 3 , S7 Table ) . The genes in this group included genes in the pyruvate dehydrogenase complex ( dlaT , increased genetic requirement in Euro-American 630 , East Asian 631; lpdC , increased requirement in Euro-American 663 , East Asian 621 , 631 ) and malate synthase ( glcB , decreased genetic requirement in East Asian 621 , 662 ) . Some of the enzymes encoded by these genes are key regulators of entry into the tricarboxylic acid ( TCA ) cycle ( dlaT , lpdC ) , and may indicate differences in metabolic flux in these strains . Our results demonstrate that TnSeq is a scalable approach to identify functional genetic differences among Mtb strains . Yet TnSeq is more technically challenging than WGS or gene expression methodologies . Therefore , we sought to determine whether the differences in genetic requirements identified by TnSeq could have been predicted from WGS or expression data . We first assessed whether any of the differential genetic requirements could have been predicted by sequence variants identified through our WGS pipeline . A differential genetic requirement could be attributed to a specific polymorphism in only one case . Euro-American strain 630 had an ~12-fold decrease in transposon insertion count across ndh , the type II NADH dehydrogenase , relative to H37Rv , indicating an increased genetic requirement for this gene ( Fig 4A ) . Our WGS pipeline revealed that strain 630 has a nonsense mutation in ndhA , a paralog of ndh [36] , suggesting that a loss of functional redundancy created a relative requirement for the remaining NADH dehydrogenase ( Fig 4A , S4 Table ) . In another instance of apparent loss of redundancy , icl2 , encoding one of two isocitrate lyase enzymes in the TCA cycle [37] , is disrupted by a frameshift mutation in the reference strain H37Rv and Euro-American strain 630 , but is intact in the other clinical strains in our panel ( S4 Table ) . In the strains with an intact icl2 , there was a non-significant increase in transposon insertion count across the paralog , icl1 , compared to H37Rv ( Fig 4A ) . In many other instances the genetic basis for differences in genetic requirements between strains was not clear . For example , our WGS pipeline identified large-scale duplications in a subset of the clinical strains encompassing the TCA cycle gene icd1 , which encodes one of two isocitrate dehydrogenases ( S3 Table ) [38] . The functional redundancy created by this duplication might decrease the genetic requirement for the other isocitrate dehydrogenase isoform , encoded by icd2 , for strains with two copies of icd1 . However , we observed a decreased genetic requirement for icd2 across all clinical strains , even those without the duplication ( Fig 4A ) . Thus , in most cases the relationship between genetic requirements and sequence variants are not readily apparent , likely reflecting the complexity of genetic networks . Transcriptomic studies have identified differences in gene expression between Mtb clinical strains [39 , 40] , therefore , we next assessed whether genes with differential requirements for growth display corresponding differences in expression . We selected a set of 10 genes identified as differentially required between strains by TnSeq , prioritizing genes of known function , and determined expression by NanoString ( see Materials and Methods ) . For all but one gene , the sigma factor sigB , there was no correlation between transposon insertion count across a gene and gene expression relative to the reference strain H37Rv ( Fig 4B ) . These findings suggest that differences in genetic requirements between strains cannot be predicted by expression data , consistent with comparisons of gene expression and gene essentiality within a strain [41] . We then sought to validate some of the functional differences among strains predicted by our TnSeq data . Because strain diversity has the potential to impact the efficacy of drugs , we focused on genes whose products are the targets of antibiotics . Our TnSeq data indicated that all of the clinical strains have an increased requirement for several genes implicated in redox homeostasis including the gene encoding the catalase-peroxidase , KatG , as compared to H37Rv ( Fig 2 , S7 Table ) . KatG is the activator of isoniazid ( INH ) , a key first-line anti-tubercular agent that is converted to its active form by the bacterial catalase-peroxidase in a redox sensitive fashion . We sought to determine whether the differences in genetic requirement for katG would predict differences in response to INH . We first determined minimum inhibitory concentration ( MIC ) for H37Rv and the panel of clinical strains ( see Materials and Methods ) . The MICs for all strains were within 2-fold that of H37Rv , with the exception of Indo-Oceanic strain 667 , which had an MIC 4-fold lower than H37Rv ( S8 Table ) . Thus , the differential genetic requirement for katG did not correspond to differences in INH susceptibility . A major mechanism of acquired resistance to INH is through katG mutation . In laboratory studies , INH resistance occurs at a high rate , primarily through mutations that result in a loss of KatG function , including complete and partial deletions of the gene [42 , 43] . However , in clinical studies , a relatively small number of katG mutations account for most INH resistance . A single SNP ( S315T ) , which preserves some catalase-peroxidase activity , accounts for over two-thirds of INH resistance in some studies , and katG deletions are rare [44 , 45] . The increased genetic requirement for katG among clinical strains identified by TnSeq might explain these observations . We reasoned that the increased genetic requirement for katG among clinical isolates might reduce the number of viable mutations that confer INH resistance ( target size ) , resulting in a lower INH resistance rate [42] . To test this hypothesis , we measured the rate at which INH resistance is acquired by performing Luria-Delbruck fluctuation analysis on three clinical strains from our panel , one from each lineage , and H37Rv . Consistent with the prediction from our TnSeq data , all three clinical strains acquired resistance to INH ( 1 μg/mL ) at a significantly lower rate than H37Rv ( Fig 5A ) . Lineage-based differences in basal mutation rate that impact the likelihood of drug resistance have previously been described [3] , and additional work has found sub-lineage and strain-specific differences in mutation rate ( personal communication , Sarah Fortune and Nathan Hicks ) . Therefore , to confirm that the decreased rate of resistance to INH among the clinical strains was not due to a lower basal mutation rate , we also determined the rate of resistance to an antibiotic that targets a different cellular process , rifampicin . The rifampicin resistance rates were not different between H37Rv and any of the clinical strains tested ( S2 Fig ) . Together , these findings suggest that the fitness costs of disrupting genes that encode antibiotic activators or targets can also contribute to the frequency of emergence of drug resistance , consistent with findings from population-based studies [6 , 9] . We identified a number of central carbon metabolism genes with differential genetic requirements from our TnSeq data ( Fig 3 ) . Bacterial metabolism has emerged as a promising target for antibiotic development [46 , 47] but our data suggests that molecules targeting these pathways may have variable efficacy across clinical strains . For example , the clinical strains varied in their requirement for glcB , which encodes malate synthase ( GlcB ) , the second enzyme in the glyoxylate shunt ( Fig 3 , S7 Table ) . The glyoxylate shunt is an anaplerotic pathway important for central carbon and fatty acid metabolism and the enzymes in this pathway , malate synthase and isocitrate lyase , play multiple , essential roles in establishing and maintaining in vivo infection [48–51] . It is also a pathway present only in prokaryotes , plants , and fungi . These features make the glyoxylate shunt an attractive drug target . Indeed , a novel GlcB inhibitor has recently been developed that is effective in reducing the burden of H37Rv in a mouse model of infection [52] . This inhibitor was developed through a structure-guided optimization of phenyl-diketo acids [52] . The inhibitors of this series bind to the active site of GlcB by coordinating the catalytic Mg2+ ion with the ketoacid moiety , making hydrogen bonds with the catalytic acid , arginine , through both keto groups , and engaging in anion-π interactions with the catalytic base , asparagine , through the phenyl group . These amino acid moieties are conserved at the sequence level across the clinical strains ( the only SNPs in glcB are G104S in the four East Asian strains , and T191A in Indo-Oceanic strain 641 , S5 Table ) , but our TnSeq data suggested that the genetic requirement for glcB varied between strains , potentially altering sensitivity to the inhibitor . We hypothesized that strains with a decreased genetic requirement for glcB during in vitro growth would be less sensitive to GlcB inhibition in culture . To test this hypothesis , we used the 2-Cl-6-F-3-Me-Phenyl diketo acid member of the inhibitor series , which was developed for the favorable pharmacological qualities needed to chemically evaluate GlcB as a target in Mtb infection [52] . The acid form inhibits purified GlcB with an IC50 of 5 . 5 μM , and the esterified form , which has improved cellular uptake , kills H37Rv grown on acetate with an MIC99 of 2 μM and kills H37Rv grown on dextrose in the presence of fatty acids with an MIC99 of 8 μM . We determined MIC for the esterified form of the GlcB inhibitor by Alamar Blue assay for bacteria grown on acetate with dextrose ( see Materials and Methods ) for all strains ( Fig 5B ) . Strains differed in their MIC90 up to 30-fold , from 7 μM ( H37Rv ) to 212 μM ( East Asian 621 ) . The strains with a significant increase in transposon insertion count across glcB ( East Asian 631 , 662 ) had MIC90s ~25- and ~24-fold higher than H37Rv , respectively . The increase in MIC90 corresponded with reduced bacterial killing ( Fig 5C ) . Across all strains , the log2 fold-change in transposon insertion count across glcB correlated with MIC90 ( R = 0 . 692 , P = 0 . 0388 ) ( Fig 5D ) . We also determined the MICs for ofloxacin and streptomycin , which did not differ by more than 2-fold among strains ( Fig 5E , S3 Fig , S8 Table ) . Thus , the differences in sensitivity to the GlcB inhibitor are specific to this pathway and do not reflect general differences in drug susceptibility among strains . Finally , we assessed whether differences in sensitivity to the GlcB inhibitor could have been predicted by genetic variants identified through WGS or by gene expression . The first step in the glyoxylate shunt , hydrolysis of isocitrate into glyoxylate and succinate , is catalyzed by isocitrate lyases encoded by two paralogous genes: icl1 and icl2 ( Fig 3 ) . As described above , icl2 is disrupted by a frameshift mutation in H37Rv , yet many strains retain an intact open reading frame and express a second , functional isocitrate lyase [37] . The icl2 frameshift was present in only one of our clinical strains ( Euro-American strain 630 ) while the other strains in the panel possessed an intact icl2 ( Fig 5F , S4 Table ) . Thus , icl2 status could not be used to predict susceptibility to the GlcB inhibitor among these strains . Expression of glcB was not significantly different between H37Rv and any of the clinical strains ( Fig 4B , ANOVA with Tukey’s post test ) . Taken together , these results indicate that our TnSeq data can predict susceptibility to a novel antibiotic that could not be readily derived from sequence or expression data alone .
Despite the relatively conserved genomic content of Mtb , strains can vary dramatically in features that impact clinical outcomes , including capacity to acquire antibiotic resistance [3 , 6] . To investigate the basis of strain variation , we used TnSeq to define differences in the genetic requirements for in vitro growth across a panel of 9 genetically diverse Mtb strains . To date , this represents the largest comparative TnSeq study of a microorganism and provides an analytic framework for multi-strain TnSeq comparisons in bacteria . We identified as many as ~50 differentially required genes in each strain relative to the reference strain , H37Rv , for in vitro growth . These included differences in the requirements for genes involved in fundamental cellular processes , such as central carbon metabolism . Some of these differences occurred in all strains in the panel , perhaps reflecting adaptation of the reference strain to laboratory culture . These findings underscore the potential pitfalls of relying on reference strains to infer the common biology of a species . Other genes were differentially required in only a subset of strains or were strain-specific . Mtb has a geographically determined phylogenetic structure that divides into seven major lineages , yet we did not find that differences in genetic requirements tracked strictly with lineage . While it is tempting to think of Mtb strains within a lineage as having collective behavior , our data is consistent with other studies showing substantial genomic and phenotypic diversity within lineages [24 , 53] . As with any screening technique , not all biologically important differences can be captured by TnSeq . Two limitations of TnSeq are polar effects and the possibility for trans complementation . Polar effects , whereby transposon insertion disrupts not only the function of the gene containing the insertion , but also downstream genes in an operon , are mitigated by the design of the transposon , which contains a strong , outward facing promoter [54] . An earlier study found no insertion bias toward the 5’ end of operons , suggesting polar effects are negligible in this system . Complementation in trans , whereby a transposon mutant’s phenotype is suppressed by bacteria wild-type for the locus present in the transposon library , has been reported in some TnSeq studies [55] . Such effects are expected to be limited to secreted products , and could be identified in future studies by phenotypic screening of arrayed transposon libraries . A key advantage of TnSeq is that it uses the cell as a sensor to directly report on the genes and pathways required for microbial growth . This stands in contrast to WGS , where the selective pressures generating sequence variants may not be known . This feature of TnSeq allowed us to use differences in genetic requirements between strains to predict differences in antibiotic responses . Because many antibiotics target processes essential for growth , we reasoned that a change in genetic requirement for an antibiotic target could predict differences in antibiotic sensitivity . Our glcB data validate this approach , which could be used to prioritize targets for drug development , since antibiotics should ideally be effective against all strains of a pathogen . Another important consideration in antibiotic development is the likelihood that resistance will emerge [56] . Some mechanisms of antibiotic resistance , such as the acquisition of inactivating enzymes , activation of efflux pumps , or decrease in antibiotic uptake , may not be detected by TnSeq . Yet our katG data indicate that the likelihood of loss of function mutations in antibiotic activators , an important resistance mechanism , can be predicted by TnSeq . Our findings that TnSeq identifies biologically important differences among strains that are not readily predicted by WGS data highlights a challenge for the field . We are rapidly developing catalogues of bacterial sequence variants but largely lack data to link genotypes and phenotypes . For Mtb , such analyses are confounded by this pathogen’s clonal population structure . Studies in which Mtb genotype has been successfully linked to antibiotic phenotype have required WGS data from hundreds of strains [6 , 8 , 11–13] . With TnSeq , we were able to use a much smaller set of strains to predict antibiotic phenotypes . Yet our analyses suggest that attributing differences in genetic requirements to specific sequence variants will require much larger data sets to achieve the statistical power to dissect such interactions . Here we demonstrate that TnSeq is a scalable method to help close the gap between genotype and phenotype . Importantly , TnSeq can be used to predict strain-specific phenotypes that are not readily predicted from sequence or expression data alone . Given the potential consequences of strain variation for the efficacy of antibiotics and vaccines , this approach could be utilized under other in vitro and in vivo conditions to assess the generalizability of new therapeutic approaches . Moving forward , integrating TnSeq data with orthogonal approaches , such as proteomics , metabolomics , and transcriptomics , will doubtless provide further insight into the basis of Mtb’s phenotypic heterogeneity . This work provides a foundation for these future studies and highlights the cellular pathways that may be subject to the most variation in Mtb .
Clinical strains were identified as previously described and cultured from single colonies [25] . Strains were grown at 37°C and cultured in Middlebrook 7H9 salts supplemented with 10% OADC , 0 . 5% glycerol and 0 . 05% Tween-80 or plated on 7H10 agar supplemented with 10% OADC , 0 . 5% glycerol and 0 . 05% Tween-80 unless otherwise noted . Clinical strains were handled to minimize in vitro passaging . Genomic DNA was isolated from a 10-mL culture of each clinical strain using standard phenol-chloroform techniques . Libraries were prepared with the Illumina Nextera XT kit and sequenced on an Illumina MiSeq Desktop Sequencer with MiSeq Reagent Kit v2 . Paired-end sequencing with read lengths of 151 bases was performed for an average coverage of 150x ( range 62-240x ) . Genome sequences of the clinical strains were assembled with a custom comparative-assembly pipeline [57] using M . tuberculosis H37Rv ( GenBank accession number NC_000962 . 2 ) as the reference genome . Contig-building was used to define large-scale insertions and deletions ( up to 40kb ) , as well as to infer changes in genomic locations of IS6110 transposons . Variants including SNPs , insertions , and deletions , were identified by alignment with the reference sequence , except in repetitive elements , such as transposases and PE/PPE/PE-PGRS genes , due to the technological limitations of Illumina sequencing ( S1 Table , S1 Fig ) . Genes with disrupted open reading frames due to small insertions and deletions and nonsense mutations were identified ( S4 Table ) as were genes with at least 10% of their open reading frame deleted ( S2 Table ) . The H37Rv and clinical strain Himar1 transposon libraries were generated as previously described [26 , 27] . Libraries were selected for 21 days on 7H10 agar supplemented with 10% OADC , 0 . 5% glycerol , 0 . 05% Tween-80 and 0 . 2% Cas-amino acids ( Difco ) with 20 μg/mL kanamycin . Two independent libraries of ~100 , 000 mutants each were generated for each clinical strain and three independent libraries of ~100 , 000 mutants each were generated for H37Rv . Genomic DNA was extracted from each transposon library using standard phenol-chloroform techniques and prepared for transposon-junction sequencing essentially as described [27] . Libraries were sequenced on an Illumina HiSeq . Raw reads were mapped to TA sites in each strain’s genome assembly and read counts were reduced to unique template counts using Transit [35] . Statistical comparisons of insertion counts across open reading frames were performed using a permutation test-based method ( "resampling" method in Transit ) , after placing TA sites on a common indexing system and normalizing for library read-depth [34 , 35] . Insertions in the central 90% of each gene were considered and a LOESS correction for genome positional bias was performed . P-values were adjusted for multiple comparisons by the Benjamini-Hochberg procedure to control the false-discovery rate at ≤ 5% [35] . Deleted genes , repetitive genes , and genes in the duplicated region were excluded prior to performing comparisons of insertion counts between strains ( S1 Table , S2 Table , S3 Table , S1 Fig ) . RNA was isolated from triplicate mid-log cultures for each strain using standard techniques . Briefly , cells from each culture were harvested by centrifugation , resuspended in 1 mL of TRIzol ( Thermo Fisher ) , and disrupted by bead-beating . Chloroform was added to 25% of the total volume and total RNA was isolated using the Direct-zol RNA MiniPrep kit ( Zymo Research ) . Samples were then treated with TURBO DNase ( Ambion ) for 1 hour to remove residual contaminating genomic DNA , followed by clean-up with RNA Clean & Concentrator columns ( Zymo Research ) . RNA was quantified by Qubit RNA Assay ( Thermo Fisher ) and 25 ng of RNA was used as input in the NanoString nCounter assay ( NanoString Technologies ) with custom-designed probes to quantify gene expression . Target sequences are listed in S9 Table . Data were analyzed with nSolver version 4 ( NanoString Technologies ) ; raw NanoString counts were normalized to internal positive controls to correct for technical variation between assays , and normalized to a housekeeping gene ( sigA ) to correct for variation in RNA input amount . Background counts from no-input samples were subtracted . Fluctuation analysis and statistical analyses of the data were performed essentially as described previously , with 24 independent cultures per isolate [3] . Isoniazid MICs were determined by the microplate Alamar Blue assay [58] and the agar proportion method [59] . For the Alamar Blue assay , strains were grown in 7H9 supplemented with OADC , glycerol , and Tween-80 to an OD600 of ~1 , filtered with a 5 μM filter to obtain a single-cell suspension , then diluted in 7H9 with supplements to an OD600 of 0 . 003 and pipetted in replicate into 96-well plates in the presence or absence of serially diluted isoniazid as described [58] . Plates were incubated at 37°C with shaking for four days before adding Alamar Blue ( Bio-Rad ) , then incubated for an additional two days before visually scoring for color change . MIC was defined as the lowest concentration of antibiotic that prevented color change from blue or purple to pink . For the agar proportion method , 7H10 media supplemented with 10% OADC , 0 . 5% glycerol and 0 . 05% Tween-80 was prepared and isoniazid added to attain the following concentrations: 0 . 8 , 0 . 4 , 0 . 2 , 0 . 1 , 0 . 05 , 0 . 025 , and 0 . 0125 μg/mL . Serial dilutions of duplicate mid-log cultures for each strain were prepared and plated on isoniazid-containing or control plates without antibiotic . Three weeks after plating , CFU were enumerated , and the reported MIC is the isoniazid concentration which reduced CFU by an average of 99% compared to the control . To determine the MICs for the GlcB inhibitor and control antibiotics streptomycin and ofloxacin , strains were grown in 7H9 supplemented with OADC , glycerol , and Tween-80 to an OD600 of ~1 , filtered with a 5 μM filter to obtain a single-cell suspension , then diluted into testing media to an OD600 of 0 . 01 and pipetted into 96-well plates in the presence or absence of serially diluted drug as described above . Testing media consisted of 7H9 media supplemented with 0 . 5% dextrose , 0 . 1% sodium acetate , 0 . 085% NaCl , and 0 . 05% Tyloxapol . Plates were incubated at 37°C with shaking for six days before adding Alamar Blue , then incubated for an additional two days before reading OD570 ( resorufin ) and OD600 ( resazurin ) . The difference between these two OD values was used as a viability reporter . The MIC90 was derived from curve fit plots generated in Prism and the reported value is the average of two independent experiments . On the day that plates were read , an aliquot from each well was used to make serial dilutions and plate for CFU . MIC experiments were performed in technical replicate at least twice . Plating for CFU was performed in technical replicate once . | Tuberculosis , caused by the bacterium Mycobacterium tuberculosis ( Mtb ) , remains a serious global health problem , causing ~1 . 5 million deaths a year world-wide . Like other bacterial pathogens , diversity among strains of Mtb contributes to differences in infection outcomes , vaccine efficacy , and response to antibiotic treatment . Currently , the important genetic differences that contribute to variation among Mtb strains remain poorly understood . In this study , we applied a functional genomics technique called TnSeq to a panel of Mtb clinical strains to investigate the genetic basis of strain diversity . We identified a number of genes that are differentially required for growth in culture among these strains . Some of these genes are involved in the response to antibiotics , including the first-line antitubercular agent isoniazid and a novel antitubercular drug currently in development . We found that the genetic differences between strains uncovered by TnSeq predicted responses to these antibiotics . Our results demonstrate the utility of TnSeq for identifying clinically relevant differences among Mtb strains . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"transposon",
"mutagenesis",
"drugs",
"microbiology",
"antibiotic",
"resistance",
"antibiotics",
"genome",
"analysis",
"genetic",
"elements",
"pharmacology",
"molecular",
"biology",
"techniques",
"mutagenesis",
"and... | 2018 | TnSeq of Mycobacterium tuberculosis clinical isolates reveals strain-specific antibiotic liabilities |
Identification of dengue patients at risk for progressing to severe disease is difficult . Significant plasma leakage is a hallmark of severe dengue infection which can suddenly lead to hypovolemic shock around the time of defervescence . We hypothesized that the detection of subclinical plasma leakage may identify those at risk for severe dengue . The aim of the study was to determine the predictive diagnostic value of serial ultrasonography for severe dengue . Daily bedside ultrasounds were performed with a handheld ultrasound device in a prospective cohort of adult Indonesians with dengue . Timing , localization and relation to dengue severity of the ultrasonography findings were determined , as well as the relation with serial hematocrit and albumin values . The severity of dengue was retrospectively determined by WHO 2009 criteria . A total of 66 patients with proven dengue infection were included in the study of whom 11 developed severe dengue . Presence of subclinical plasma leakage at enrollment had a positive predictive value of 35% and a negative predictive value of 90% for severe dengue . At enrollment , 55% of severe dengue cases already had subclinical plasma leakage , which increased to 91% during the subsequent days . Gallbladder wall edema was more pronounced in severe than in non-severe dengue patients and often preceded ascites/pleural effusion . Serial hematocrit and albumin measurements failed to identify plasma leakage and patients at risk for severe dengue . Serial ultrasonography , in contrast to existing markers such as hematocrit , may better identify patients at risk for development of severe dengue . Patients with evidence of subclinical plasma leakage and/or an edematous gallbladder wall by ultrasonography merit intensive monitoring for development of complications .
Dengue virus infection is the most rapidly spreading mosquito-borne viral disease in the world and has the highest burden in often resource poor ( sub ) tropical countries . Symptomatic dengue patients usually present with a self-limiting febrile illness , called dengue fever . However , a proportion of dengue patients develops severe complications around the time of defervescence . Transient plasma leakage into serosal cavities , which may progress to life-threatening hypovolemic shock , is a hallmark of severe dengue [1] , [2] . In the former World Health Organization ( WHO ) classification ( 1997 ) , severe dengue characterized by fever , thrombocytopenia , plasma leakage and hemorrhagic tendency , is referred to as dengue hemorrhagic fever ( DHF ) and , when accompanied by circulatory failure , as dengue shock syndrome ( DSS ) ( table 1 ) [3] . The most recent WHO classification ( 2009 ) defines severe dengue infection in case plasma leakage leading to shock or respiratory failure , severe bleeding or severe organ failure is present [4] ( table 1 ) . However , in clinical practice severe bleeding and severe organ failure are relatively uncommon and rarely occur without plasma leakage [5] , [6] . Hence , plasma leakage is an essential element of severe dengue , regardless whether the WHO 1997 or 2009 classification is used . One of the most important challenges for clinicians caring for dengue patients is to identify those patients who will progress to severe disease . Currently , no accurate biomarkers are available and prediction of significant plasma leakage is difficult in clinical practice . Serial hematocrit values are frequently used to detect hemoconcentration as a sign of plasma leakage , but are hard to interpreted due to individual variations , absence of baseline hematocrit values , intravenous fluid administration and bleeding . Monitoring for pleural effusion and/or ascites - both predilection sites of plasma leakage in dengue - is an alternative . Chest X-rays can only detect significant pleural effusions and , therefore , ultrasonography of the chest and abdomen may be preferential . The main advantages of ultrasonography are its high sensitivity to detect small amounts of pleural effusion and ascites [7] , [8] , [9] , [10] , [11] , [12] and the possibility to visualize the gallbladder wall , which is frequently thickened in dengue [13] , [14] . So far , most dengue ultrasonography studies are cross-sectional in nature , whereas significant changes in vascular permeability may occur after the initial presentation of the patient . Serial ultrasonography may therefore be needed to detect the presence of plasma leakage . Despite the advantages of ultrasonography to detect plasma leakage in dengue , the clinical application is still limited in resource poor settings , because it requires sophisticated facilities and expertise . Lately , more affordable handheld ultrasound devices have become available , which can be used at the bedside . The aim of our study was to determine the predictive diagnostic value of serial ultrasonography for severe dengue . Measurements were done using a handheld ultrasound device at the bedside . We compared the ability to detect plasma leakage by serial handheld ultrasonography with conventional ultrasonography performed by a radiologist . The presence of plasma leakage detected by handheld ultrasonography was compared with frequently used clinical and laboratory parameters for plasma leakage and related to severe dengue as a clinical outcome .
For this prospective observational study , we included patients with dengue infection , aged 14 years or above , who were admitted to Rumah Sakit Hasan Sadikin , an academic referral hospital in Bandung West Java , Indonesia , from March 2011 to January 2012 . All patients had a clinically suspected dengue infection , which was retrospectively classified as non-severe or severe dengue according to WHO criteria from the guidelines edition 2009 ( table 1 ) [4] . According to these guidelines a narrow pulse pressure ( difference between systolic and diastolic blood pressure of 20 mmHg or less ) is also a sign of shock . The presence of plasma leakage was confirmed by a significant hematocrit change ( minimal 20% ) and/or single high hematocrit values ( >50 and >44% for men and women respectively ) , and/or by plasma leakage seen during ultrasonography . In case of a discrepancy in the presence or absence of plasma leakage between handheld or conventional ultrasonography , the latter was used as a golden standard . All patients had to be enrolled either in the febrile phase ( defined as a temperature of 37 . 5°C or more on that day ) or in the critical phase ( defined as the period within 48 hours after defervescence and before platelet counts increased again ) of dengue infection . Patients with concurrent chronic disease and pregnant patients were excluded . Demographic , clinical , laboratory and ultrasonography data were collected using a standardized data collection form . Ultrasonography was not part of routine clinical care . A full blood count , including hematocrit measurements , was daily performed as part of routine clinical care , until platelet counts had shown a substantial increase and the patient had improved clinically . Patients were asked to return 2–3 weeks after discharge for follow up , which included clinical evaluation , blood drawing for full blood count including hematocrit for baseline values , serology and handheld ultrasonography . The degree of hemoconcentration was calculated according to the formula; % hemoconcentration = [ ( peak hematocrit surrounding defervescence ) - ( convalescent hematocrit or minimum hematocrit during admission ) ]/ ( convalescent hematocrit or minimum hematocrit during admission ) ×100 . Samples of dengue suspected cases were tested for presence of viral RNA by reverse-transcriptase PCR and samples were tested for dengue specific IgM and IgG ( Panbio , Windsor , Australia ) , interpreted according to the manufacturer's instructions . Samples that were IgM positive , had an IgM or IgG conversion , a 4-fold titer increase in IgG and/or IgG levels comparable to a HI titer of at least 1∶2560 ( the IgG cut off point set by the manufacturer to detect secondary infection ) were considered dengue positive . Patients with both negative PCR and dengue serology results were excluded . Bedside ultrasound examinations were performed daily , using a handheld portable ultrasound device ( Signos personal ultrasound ) with a 3 . 5 MHz transducer . Measurements started on the day of study enrollment until the day of discharge . All handheld ultrasound examinations were carried out by study physicians , who first received a four-week practical training in systematically detecting ascites , pleural effusion and determining the gallbladder wall aspect and thickness . To validate the results of these portable ultrasonograms , one conventional ultrasound was made per patient by a radiologist . This was done in the critical phase , since the likelihood of finding plasma leakage is the highest at this time-point . Study physicians and radiologists were blinded for each other's results . Ultrasonography was performed after a minimum of 6 hours of fasting , with the patient in supine position . The abdomen was screened for the presence or absence of ascites in the hepato-renal pouch and spleno-renal area ( figure 1 A–D ) . The conventional ultrasonography by the radiologist additionally screened for free fluid around the urine bladder , which was not feasible with the portable ultrasound device . Presence of multiple layers in the gallbladder wall and/or a gallbladder wall thickness of 0 . 3 cm or more were considered as a thickened gallbladder wall ( figure 2 A and B ) . The presence or absence of pleural effusion was determined by visualizing bilateral costophrenic angles in supine position ( figure 3 A and B ) . The Study Research Ethics Committee of the Faculty of Medicine of the Padjadjaran University , Bandung , Indonesia , approved all legal , ethical , radiological aspects of the study . Written informed consent was obtained from all patients or from their parents , in case the patient was below 17 years of age . All patients were able to provide informed consent . Data are expressed as medians with interquartile ranges ( IQR ) or numbers with percentages . Differences in non-continuous data of 2 groups were analyzed by Pearson's Chi-square test or by Fisher's exact test in case of expected counts <5 . Continuous variables between 2 groups were analyzed by unpaired t-test in case of normally distributed data; and by Mann Whitney U test in case of non-parametric data . Relationships between continuous data were examined by Pearson's correlation for parametric data and Spearman's correlation for non-parametric data . A p-value<0 . 05 was considered significant . All analysis were performed using SPSS ( version 18 . 0 ) statistical program .
Among the 71 patients with clinically suspected dengue infection , 66 patients had a positive PCR and/or serology and were included . Fifty-five ( 83% ) patients were retrospectively classified as non-severe dengue and 11 ( 17% ) as severe dengue according to the 2009 WHO classification [4] . All patients with severe dengue had plasma leakage with vascular collapse; none of the patients had severe bleeding or organ failure ( table 1 ) . Patient characteristics and baseline data are presented in table 2 . The first day of fever was defined as the first day of illness . All patients were discharged from the hospital in good health . At enrollment , ascites and/or pleural effusion were found by handheld ultrasonography in 11/55 ( 20% ) patients in the non-severe dengue group and in 6/11 ( 55% ) patients of the severe dengue group ( p = 0 . 03 ) ( table 3 ) . As 6 of 17 patients with ascites or pleural effusion at enrollment developed severe dengue , the positive predictive value ( PPV ) of detection of plasma leakage by handheld ultrasonography for development of severe dengue was 35% . None of the patients had clinical signs or symptoms suggestive for plasma leakage at enrollment , as determined by careful history taking and physical examination by the study clinicians . Four of the 11 patients eventually classified as having severe dengue , already fulfilled criteria for severe dengue at the time of enrollment due to subclinical plasma leakage combined with a narrow pulse pressure . Five of the 49 patients without ultrasonographic evidence of plasma leakage at enrollment , eventually developed severe dengue , yielding a negative predictive value ( NPV ) of 90% . Two of these patients already had a narrow pulse pressure at enrollment without further clinical evidence for plasma leakage . After admission , another 10 patients developed ultrasonographic findings of plasma leakage , mainly in the critical phase ( 80% ) . In total , 17/55 ( 31% ) patients in the non-severe and 10/11 ( 91% ) in the severe dengue group developed ultrasonographic evidence of plasma leakage ( p<0 . 001 ) ( table 3 ) . The diagnosis of plasma leakage in the patient with severe dengue without ultrasonographic evidence of plasma leakage was based on an >20% increase in hematocrit values . Moreover , three patients without evidence of plasma leakage during follow up developed hypotension and/or a narrow pulse pressure . In 9/11 ( 82% ) severe dengue cases , shock developed without preceding clinical symptoms or signs of plasma leakage . The remaining two patients had symptomatic pleural effusions . Ascites was more frequently detected by ultrasonography than pleural effusion , which was found either right-sided or bilateral ( table 3 ) . A thickened gallbladder wall was found in 34/52 ( 65% ) of non-severe and 9/10 ( 90% ) of severe dengue patients at enrollment ( p = 0 . 15 ) ( table 3 ) . The median thickness of the gallbladder wall at enrollment was 0 . 36 cm ( IQR , 0 . 27–0 . 50 cm ) in patients with non-severe dengue versus 0 . 54 cm ( IQR , 0 . 34–0 . 75 cm ) in those with severe dengue ( p<0 . 01 ) . Gallbladder wall thickness could not be determined in six patients at enrollment , because these patients had taken food . Eventually , a thickened gallbladder wall was detected in 37/55 ( 67% ) of the non-severe dengue and 11/11 ( 100% ) of the severe dengue patients ( p = 0 . 27 ) . Detection of a thickened gallbladder wall at enrollment had a PPV of 21% and a NPV of 91% for severe dengue . In patients who developed ascites or pleural effusion , thickening of the gallbladder preceded the detection of ascites or pleural effusion by 1 to 3 days in 6/10 ( 60% ) of the non-severe and in 6/9 ( 67% ) of the severe patients . Multiple layers were frequently visible in thickened gallbladder walls suggesting the presence of a sub-serosal fluid layer ( figure 2 B ) . Patients with subclinical plasma leakage at enrollment had a thicker gallbladder wall than those without plasma leakage with median values of 0 . 59 cm ( IQR , 0 . 38–0 . 79 cm ) and 0 . 35 cm ( IQR , 0 . 27–0 . 44 cm ) respectively ( p<0 . 01 ) . There was an inverse correlation of gallbladder wall thickness with platelets counts at enrollment ( rs = −0 . 53 , p<0 . 001 ) . In contrast , no significant correlation existed with clinical symptoms ( nausea , vomiting or abdominal pain ) and laboratory values ( absolute value of hematocrit or its change in time , serum albumin , alanine aminotransferase ( ALAT ) ) ( data not shown ) . Regarding the timing of the ultrasonographic abnormalities in relation to the day of illness , in patients developing pleural effusion or ascites and gallbladder wall thickening during the study , these findings were already detectable on the first ultrasonography performed at enrollment in 17/27 ( 63% ) and 43/48 ( 90% ) of patients , respectively . The first ultrasonography was performed after a median duration of 6 days of illness ( IQR , 5–7 days ) . The median day plasma leakage was first detected by ultrasound was day 7 of illness ( IQR , 6–7 days ) . The maximum gallbladder wall thickness was also observed on this day ( IQR , 6–8 ) . Fifty-five conventional ultrasound reports were available to serve as the golden standard for the detection of plasma leakage by serial handheld ultrasonography . In 48/55 patients ( 87% ) , the handheld ultrasound scored the same results for the presence or absence of plasma leakage as the conventional ultrasound on that same day . In an additional 4/55 patients ( 7% ) portable ultrasonography missed the presence of plasma leakage on the same day , but still detected plasma leakage one day after the conventional ultrasound . In 3/55 patients ( 5% ) , handheld ultrasonography failed to detect plasma leakage seen by conventional ultrasonography . This was partly because conventional ultrasonography also screened for ascites in the pelvis . Gallbladder wall thickness measurements were available for 45 patients using the conventional ultrasonography . For 37/45 ( 82% ) of the handheld ultrasonography measurements , the presence or absence of gallbladder wall thickening was identical to the conventional ultrasonography . There were no cases in which conventional ultrasonography did not show plasma leakage while handheld ultrasonography on the same day did . An elevated hematocrit value was found in only 4/66 ( 6% ) of patients at enrollment . Single hematocrit values at enrollment were not helpful in discriminating between patients with and without ultrasonographic plasma leakage and between patients classified as non-severe or severe dengue ( figure 4 A ) . The change in hematocrit values was also not helpful for this purpose ( figure 4 B ) . In 13/21 ( 62% ) of the patients in whom hemoconcentration was observed , this was only first detected after the critical phase when patients already started to recover . Other variables showed larger differences between patients with and without ultrasonographic plasma leakage . Patients with plasma leakage had lower median platelets counts ( 21×109/L vs . 70×109/L; p<0 . 001 ) , serum albumin levels ( 36 g/L vs . 40 g/L; p<0 . 01 ) and higher ALAT levels ( 60 U/L vs . 26 U/L; p<0 . 05 ) .
This study shows that serial handheld ultrasonography at the bedside may be useful to identify patients at risk for severe dengue . More than 1 out of 3 patients ( 35% ) with ultrasonographic evidence of subclinical plasma leakage at enrollment developed shock , in contrast to only 1 out of 10 patients ( 10% ) without subclinical plasma leakage . Dengue-related shock is known to develop rapidly and it is currently not possible to identify those at risk for shock at an early stage of illness . Our findings suggest that careful monitoring of circulatory status is merited in those patients with ultrasonographic evidence of subclinical plasma leakage since there is a significant risk for progressing to shock . Despite much research effort , reliable biomarkers for predicting severe dengue have not yet been identified . Therefore , even though the positive predictive value of ultrasonography for predicting severe dengue in our study was still far from optimal ( 35% ) , serial ultrasonography may be a considerable advance in the clinical management of dengue patients . The high sensitivity of handheld ultrasonography to detect plasma leakage has implications for the interpretation of WHO classification of dengue severity from 1997 , which continues to be used frequently in research settings [6] . Ultrasonography findings of subclinical plasma leakage changed the diagnosis from non-severe dengue ( dengue fever ) to severe dengue ( dengue hemorrhagic fever ) in 9 cases of our study population if WHO 1997 guidelines had been applied . This is in line with recent studies that report subclinical plasma leakage in patients with clinically mild dengue infection [15] , [16] . Moreover , it contradicts the long-existing dichotomous view of dengue severity that is reflected in WHO classifications , in which plasma leakage is a hallmark of severe dengue infection only [8] , [9] . Also in the WHO 2009 guidelines , plasma leakage remains the main criteria for severe dengue , since the other two criteria – severe bleeding and severe organ failure- are relatively rare . Handheld abdominal ultrasonography showed a thickened gallbladder wall in 65% of the dengue cases at enrollment , and the degree of thickening was associated with dengue severity . Gallbladder wall thickening often preceded the first detection of pleural effusion or ascites and may therefore be an useful early warning sign for plasma leakage , as has also been suggested by others [7] , [8] , [13] , . Several case reports have described the occurrence of acalculous cholecystitis in dengue patients . In most of these cases , the diagnosis was based on an ultrasonographic demonstration of a thickened gallbladder wall . Some of these patients underwent cholecystectomy whereby histology showed congestion of the serosa with presence of mononuclear cell infiltrates and lymphocyte follicle formation [18] , [19] . We speculate that plasma leakage plays an important role in the pathogenesis of gallbladder wall thickening in dengue . Both ascites and gallbladder wall thickening appear and resolve around the same time and within a short time span [9] , [14] . In those gallbladders with the most outspoken wall thickening , multiple layers were visible due to sub-serosal edema , suggesting plasma leakage inside the gallbladder wall . Finally , gallbladder wall thickening was not associated with abdominal pain in our patients , which makes inflammation less likely . We demonstrated that the sensitivity for detecting plasma leakage and a thickened gallbladder wall of handheld ultrasonography performed by a non-radiologist , was comparable to conventional ultrasonography by a radiologist . Compared to ultrasonography , hematocrit proved to be a poor indicator of subclinical plasma leakage and changes in the hematocrit often occurred too late to be of clinical benefit . There are only few prospective ultrasonography studies in dengue [9] , [16] , [17] and our study is , to our knowledge , the first to employ a handheld ultrasound device operated by clinicians instead of radiologists . Heterogeneity in study populations , dengue classification , timing of ultrasonography and used definitions of gallbladder wall thickening and dengue classifications complicate a proper comparison between results of previous ultrasonography studies and ours . The reported proportion of dengue patients with gallbladder thickening in earlier studies varied from 28% to 100% , while pleural effusion and ascites was found in 32% to 100% and 15% to 96% , respectively [7]–[9] , [11] , [12] , [17] . So far , the clinical application of ultrasonography in dengue was hampered by financial and logistic constraints , but the recent availability of more affordable handheld ultrasound devices may help to overcome these restrictions . Serial bedside ultrasonography by clinicians may not only have a prognostic value by identifying dengue patients at risk for disease , but also a diagnostic value . The clinical features of dengue overlap with other endemic febrile illnesses , such as rickettsiosis and typhoid fever . The ultrasonographic detection of ascites , pleural effusion or gallbladder edema in a patient with a clinical picture of dengue supports the diagnosis of dengue [20] . Moreover , handheld ultrasonography might especially proof useful for triage during dengue outbreaks when health care facilities are overloaded . Our study has several limitations . First , our study population consisted of patients admitted to an academic referral hospital in whom severe disease was common . This explains why plasma leakage was already visible by ultrasonography in many patients at the time of enrollment . Moreover , it makes it impossible to determine the day of illness at which plasma leakage could have been detected initially . Studies in patients presenting at an earlier stage of dengue are required to answer this question . Second , adiposity was rare in our Indonesian study population and the diagnostic performance of handheld ultrasonography in populations with more overweight may be lower . Finally , despite the fasting instructions before ultrasonography , we cannot exclude that some patients had eaten before the examination which may have resulted in an over-estimation of gallbladder wall thickness . In conclusion , serial handheld ultrasonography holds promise as a diagnostic and prognostic tool in dengue . Compared to serial hematocrit values , serial ultrasonography had a better predictive value in identification of those patients at risk for severe dengue . Moreover , it may improve monitoring of circulatory status and enable timely adjustments in fluid balance to avoid hypovolemic shock or iatrogenic fluid overload . | Dengue virus infection ranges from a mild febrile illness to severe illness . Severe dengue is mainly characterized by transient plasma leakage , which may lead to a sudden onset of shock around the time of defervescence . Severe bleeding and organ impairment are less common features of severe dengue . In clinical practice it is difficult to predict which dengue patient will develop severe complications . Commonly used laboratory markers indicating plasma leakage—such as hematocrit—are rarely of clinical benefit . In contrast , ultrasonography can directly visualize plasma leakage . Because ultrasonography is not routinely used due to financial and logistical limitations , we used a more affordable handheld ultrasound device for daily bedside follow up , which had similar results as conventional ultrasonography . This study shows that a substantial proportion of mildly ill patients already had small amounts of plasma leakage before severe complications developed and—most importantly—that these patients had an increased risk for progression to shock compared to patients without plasma leakage . We conclude that ultrasonography can help to identify dengue patients at risk for shock and that patients with ultrasonographic evidence of plasma leakage should be monitored more carefully for circulatory status to timely recognize and possibly prevent shock . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"dengue",
"fever"
] | 2013 | The Predictive Diagnostic Value of Serial Daily Bedside Ultrasonography for Severe Dengue in Indonesian Adults |
The dynamics of the memory CD8 T cell receptor ( TCR ) repertoire upon virus re-exposure and factors governing the selection of TCR clonotypes conferring protective immunity in real life settings are poorly understood . Here , we examined the dynamics and functionality of the virus-specific memory CD8 TCR repertoire before , during and after hepatitis C virus ( HCV ) reinfection in patients who spontaneously resolved two consecutive infections ( SR/SR ) and patients who resolved a primary but failed to clear a subsequent infection ( SR/CI ) . The TCR repertoire was narrower prior to reinfection in the SR/SR group as compared to the SR/CI group and became more focused upon reinfection . CD8 T cell clonotypes expanding upon re-exposure and associated with protection from viral persistence were recruited from the memory T cell pool . Individual CD8 T cell lines generated from the SR/SR group exhibited higher functional avidity and polyfunctionality as compared to cell lines from the SR/CI group . Our results suggest that protection from viral persistence upon HCV reinfection is associated with focusing of the HCV-specific CD8 memory T cell repertoire from which established cell lines showed high functional avidity . These findings are applicable to vaccination strategies aiming at shaping the protective human T cell repertoire .
The capacity of CD8 T cells to recognize and respond to various pathogen-derived antigens is dictated by the diversity of their T cell receptor ( TCR ) repertoire . The TCR is a heterodimer of two chains , α and β , that comprise constant and variable regions . The most variable region in both chains is generated by somatic recombination involving variable ( V ) , junction ( J ) and diversity ( D ) gene segments that could theoretically generate ~1015–20 unique TCRs or T-cell clonotypes capable of recognizing peptide-MHC ( pMHC ) complexes [1] . Positive and negative selection in the thymus leaves ~ 2x107 T-cell clonotypes with unique TCR amino acid sequences that constitute the naïve human T-cell repertoire . Hypervariable complementarity-determining regions 1 and 2 ( CDR1 and CDR2 ) are formed by the germline V region sequences and interact mainly with MHC . The CDR3 of the TCR α and β chains , the most variable region of the TCR , are encoded by the V ( D ) J junction and interact primarily with peptide , thus determining antigenic specificity of the TCR [1] . Upon exposure to a viral infection , particular clonotypes recognizing virus-derived epitopes/pMHC are selected and expand into primary effectors that then contract to form a pool of long-lived memory T cells that are able to respond rapidly upon virus re-exposure . The size and diversity of the expanding effector T cell repertoire can vary according to the initial germline repertoire of naïve CD8 T cells and strength of interaction with pMHC ( affinity and avidity ) . In contrast , factors governing the size and diversity of the antiviral memory CD8 T cell pool are not well understood . Most importantly , determinants for selection and maintenance of CD8 T cell clonotypes exerting an efficacious and protective anti-viral immune response upon virus re-exposure or reactivation remain elusive . There is evidence to suggest that the memory CD8 T cell repertoire can be modulated by heterologous infections and age [2 , 3] but other host and viral factors could be involved . One key characteristic of the virus-specific TCR repertoire is diversity , which defines the number of unique clonotypes forming the repertoire . The repertoire can be characterized as «narrow» or «broad» depending on the number of unique clonotypes it contains . Distinctive molecular properties of T cell clonotypes that determine their functionality include affinity , avidity , functional avidity and flexibility . Affinity describes the strength of binding of a single TCR to cognate pMCH complexes , whereas avidity ( structural avidity ) is the sum of binding affinities of multiple TCRs to their pMHC complexes . Functional avidity depends on how this translates into measurable biological functions such as cytokine production [4] . Flexibility is the capacity to recognize multiple variants of the same epitope and cross-react with these variants . Although TCRs are generally very specific or «private» in their response to a pMHC complex , identical TCR sequence usage in response to a specific epitope across multiple individuals termed «public TCRs» were observed in a number of infections , tumors and even autoimmune conditions [5 , 6] . Studies in chronic cytomegalovirus ( CMV ) and Epstein-Barr virus ( EBV ) infections demonstrated focusing and increased affinities of the virus-specific CD8 TCR repertoire [7] . In human immunodeficiency virus ( HIV ) infection specific clonotypes dominated the response against an epitope in the p24 Gag ( KK10; residues 263–272 ) restricted by HLA B*2705 in patients who controlled viral replication . These clonotypes exhibited higher avidity and polyfunctionality and superior control of HIV replication in vitro [8–10] . Furthermore , certain public clonotypes were detected in several HIV controllers [11 , 12] . However , another study examining different epitopes did not observe preferential use of particular clonotypes [13] . Escape mutations in targeted epitopes could be recognized by some of these highly functional clonotypes [10] yet they also drove the expansion of alternate clonotypes with dual reactivity against both the original and mutated epitopes [14] . This expansion was inversely correlated with residual viral load , suggesting that these alternative clonotypes play a role in limiting replication of mutated viruses [15] . Hepatitis C virus ( HCV ) infection represents a unique model of a human viral infection with dichotomous outcomes , i . e . spontaneous clearance ( ~30% of infected individuals ) and persistent infection [16] . Control of primary HCV infection in the chimpanzee model was associated with a more diverse CD8 TCR repertoire than infections that became chronic and were associated with escape mutations [17] . In humans , selection of high-avidity CD8 T cells correlated with control of primary infection [18] . Analysis in individuals with long-term clearance or persistent infection demonstrated a biased TCR repertoire towards a common core of public TCRs irrespective of infectious outcome [19] . HCV is also thought to exploit a «hole» in the T cell repertoire to undergo escape mutation and evade detection by the immune system [20] . The dynamics of the repertoire upon re-exposure to HCV are less defined . Rechallenging chimpanzees following clearance of primary infection demonstrated that resolution temporally coincided with the expansion of dominant clonotypes that were associated with clearance of primary infection [21] . This is consistent with data in the lymphocytic choriomeningitis virus ( LCMV ) model [22] . Higher CDR3 diversity correlated with viral clearance and better control of escape mutations within the targeted epitope and temporary narrowing of the repertoire was observed at the peak of the recall response [17] . The majority of acute HCV infections in North America occur among people who inject drugs ( PWID ) . Individuals who spontaneously clear their primary infection but maintain risk behaviors associated with HCV re-exposure remain at high risk of reinfection . As such , HCV represents a unique model to study correlates of protective immunity in a real life challenge experiment . Follow-up of PWIDs through consecutive episodes of infection and reinfection can provide an insight into the dynamics of the virus-specific TCR repertoire and the functional properties associated with protective immunity . We have previously demonstrated that protection from chronicity upon reinfection with HCV correlated with expansion of polyfunctional HCV-specific effector T cells and increased breadth of T cell responses , suggesting the generation of de novo responses . In contrast , viral persistence was associated with limited expansion of virus-specific T cells and infection with variant viral strains that were not recognized by preexisting virus-specific memory T cells [23] . Here , we examined dynamics of the TCR repertoire during reinfection . Our objectives were to distinguish the role of pre-existing memory versus de novo T cell responses during reinfection and to establish the functional correlates of CD8 T cell clonotypes associated with protective immunity in real life exposure and reinfection . Our results suggest that dominant HCV-specific tetramer+ CD8 T cell clonotypes mobilized during the reinfection episode were exclusively recruited from the preexisting memory pool . Protection from chronic infection was associated with a narrower TCR repertoire that became more focused upon reinfection with preferential selection of TCR clonotypes with high functional avidity and polyfunctionality .
To examine the evolution and dynamics of the HCV-specific memory CD8 T cell repertoire upon virus re-exposure , we performed longitudinal analysis of this repertoire on virus-specific CD8 T cells during HCV reinfection episode in five patients with different outcomes ( S1 Table ) . Three patients resolved two successive HCV infections , hereinafter termed SR/SR . Two patients failed to spontaneously resolve the reinfection episode and developed chronic viremia despite clearing a previous HCV infection , hereinafter termed SR/CI . Epitope-specific CD8+ T cells identified by three different MHC class I tetramers were sorted and the TCR repertoire sequenced at three distinct time points for each patient: i ) pre-reinfection ( range: -55 to -20 weeks before detection of reinfection ) , ii ) peak reinfection ( range: 3 to 8 weeks post detection of reinfection ) , and iii ) late/post reinfection follow-up ( range: 12 to 24 weeks post detection of reinfection ) . The peak reinfection time point was selected based on our previously published longitudinal analysis of the frequency of tetramer+ CD8 T cells [23] . In addition , we examined the repertoire during primary acute infection in patient SR/SR-3 , for whom samples were available early at week 3 post primary infection . In patients SR/SR-1 and SR/SR-3 , tetramer frequency was high enough during peak reinfection and primary infection , respectively which allowed us to sort effector ( CD127- ) and memory ( CD127+ ) HCV-specific CD8+ T cells as well . As a control , total naïve CD8+ T cells from patients SR/SR-1 and SR/CI-2 defined as CD8+CD45RO- were sorted and sequenced for all three time points tested . A summary of the time points tested for each patient and the corresponding tetramer frequencies is presented in S1 Table . The gating strategy and post-sorting purity are presented in S1 and S2 Figs . A summary of the sequencing information including the number of cells , sequences and clonality is presented in S2 Table . TCRBV and TCRBJ gene usage , as well as the CDR3 amino acid sequence of all TCR β chains expressed by the different HCV-specific T cell clonotypes were analyzed ( S3–S7 Tables ) . The top 10 clonotypes for each patient and time point are presented in Fig 1 . Longitudinal analysis of the TCR β chain dynamics demonstrated a striking difference between SR/SR and SR/CI patients where there was high level expansion ( up to 3 . 5 fold ) of specific clonotypes in two out of three SR/SR patients upon reinfection but limited or no expansion in patient SR/SR-3 and both SR/CI patients . This is consistent with the limited expansion of tetramer+ cells observed in the SR/CI ( S1 Table and [23] ) . The analysis also demonstrated that the dominant ( frequency > 1% ) and sub-dominant ( frequency > 0 . 5% ) clonotypes at the peak reinfection time point were recruited from the pre-existing memory CD8 T cell populations and no new clonotypes were detected . This was also true in patient SR/SR-1 where we managed to examine the repertoire of effector ( CD127- ) CD8 T cells that could be indicative of a de novo T cell response . Comparison of the effector ( CD127- ) and memory ( CD127+ ) CD8 T cell repertoire showed the same clonotypes albeit at slightly different frequencies ( Fig 1 and S3–S7 Tables ) . Although , some new clonotypes that were not present at pre-reinfection were detected in all patients at the peak time point as shown in the Venn diagrams in S3 Fig , they were of low frequencies ( < 0 . 5% ) . Similarly , comparison of the repertoire during primary acute infection and reinfection in patient SR/SR-3 demonstrated that the repertoire generated during primary infection was relatively stable and a precursor to the memory repertoire ( S4 Fig ) . Although the clonotype composition of the T cell repertoire remained relatively unchanged upon reinfection , there was a change in the dominance and hierarchy of the different clonotypes where a more limited number of clonotypes dominated the response at the peak of reinfection . For example , patient SR/SR-1 exhibited preferential expansion of the V20 clonotype . This is even more evident in patient SR/SR-2 where the two most dominant clonotypes ( V19 . 1 and V11 . 2 ) switched hierarchies upon reinfection , suggesting a selection and amplification of particular clonotypes during reinfection . Our analysis demonstrated no overlap and no common clonotypes within the dominant and sub-dominant Vβ-CDR3-Jβ detected in the tested patients . This lack of common repertoire was observed among patients targeting the same epitope whether they belonged to different groups ( SR/SR-1 and SR/CI-2 recognizing the HLA-A2 restricted NS3-1073 epitope ) ( Fig 1 and S3 and S6 Tables ) or the same group ( SR/SR-2 and SR/SR-3 recognizing the HLA-B27 restricted NS5B-2841 epitope ) ( S4 and S5 Tables ) . Altogether , these results suggest that the dominant protective CD8 T cell clonotypes expanding at the peak of reinfection episode were recruited from the preexisting memory pool and at least within this set of individuals , no common or public TCRs were detectable . Very limited expansion was observed in HCV-specific CD8 T cells and specific clonotypes in the SR/CI group . Data from the chimpanzee model suggested that a more diverse repertoire , i . e . the presence of a larger number of unique clonotypes recognizing the same pMHC , was associated with spontaneous resolution of primary acute HCV [17] . Hence , we proceeded to examine repertoire diversity within our study subjects in a memory response . Clonotypes were classified into 4 categories according to their frequencies within the repertoire: ( i ) dominant clonotypes ( frequencies >1% ) ; ( ii ) sub-dominant clonotypes ( 0 . 5–0 . 99% ) ; ( iii ) low abundance clonotypes ( 0 . 1–0 . 49% ) ; and ( iv ) lowest abundance clonotypes ( <0 . 1% ) . Our analysis demonstrated that the clonotypic profile in SR/SR patients was less diverse than that observed in SR/CI patients at the pre-reinfection time point ( Figs 2 and 3 ) . In the SR/SR group , 75–83% of the repertoire was contributed by 14–23 TCR clonotypes . This repertoire became more focused at peak reinfection for patients SR/SR-1 and SR/SR-2 , where 92% of the repertoire of tetramer+ CD8+ CD127- cells in SR/SR-1 and 96% of the repertoire of tetramer+ CD8+ T cells in patient SR/SR-2 were contributed by 15 and 2 clonotypes , respectively ( Fig 2A and 2B ) . In these two patients , the repertoire retained a more focused status post reinfection when compared to the pre-reinfection time point , with 77% and 93% of the repertoire contributed by 16 and 4 clonotypes , respectively ( Fig 2A and 2B ) . Patient SR/SR-3 , for whom samples were available during primary infection , also demonstrated a focused repertoire of effector ( CD127- ) and memory ( CD127+ ) HCV-specific T cells early during acute primary infection where 85% and 82% of the repertoire were represented by 27 and 26 clonotypes , respectively ( Fig 2C ) . This repertoire remained more or less stable during the memory phase where 83% of the repertoire was represented by 21 clonotypes but did not sensibly change upon reinfection . This is consistent with our previous results , where no significant expansion of the tetramer+ population was observed in this patient upon reinfection ( S1 Table ) , and IFNγ Enzyme-Linked Immunospot ( ELISPOT ) assays suggested that he had been reinfected with a different HCV subtype [23] . In contrast , the repertoire in the two SR/CI ( unprotected ) patients was more diverse at the preinfection time point . Dominant clonotypes represented 37% and 55% of the repertoire ( 18 and 23 TCR clonotypes , respectively ) . At the peak of the immune response during reinfection , the repertoire remained more diverse than in the SR/SR patients ( the dominant category represented 28–64% of the repertoire ) ( Fig 3A and 3B ) . The remaining 40–60% of the repertoire was represented by at least 133 and as much as 568 clonotypes . In contrast , the number of minor clonotypes in the SR/SR patients was much lower ( Fig 2 ) . Furthermore , we did not observe increased focusing of the repertoire in SR/CI patients at the late reinfection time point . The presence of a less diverse repertoire that becomes more focused upon reinfection in the SR/SR group was further supported by the observation that the top three clonotypes in patients SR/SR-1 and the top two clonotypes in patient SR/SR-2 represented 28% and 59% of the repertoire pre-reinfection , respectively ( Fig 2A and 2B , dissected pie-charts in the lower rows ) . Upon reinfection , these clonotypes became more dominant , representing 51% and 96% of the repertoire , respectively . This increased dominance was maintained post-reinfection , where they formed 39% and 90% of the repertoire , respectively . On the other hand , the top three clonotypes in SR/CI patients never represented more than 27% of the repertoire , and were sometimes as low as 9% ( Fig 3A and 3B , dissected pie-charts in the lower rows ) . These results confirm that the SR/SR TCR repertoire is highly focused and becomes more focused upon re-exposure and reinfection . In order to establish a quantitative measure of the changes observed in the TCR repertoire during reinfection we examined diversity using the Simpson diversity index that takes into account the number of species present as well as the abundance of each species , with 0 defined as infinite diversity and 1 as no diversity [24 , 25] . The focusing of the repertoire in the SR/SR group was reflected by an increase in the Simpson diversity index for patients SR/SR-1 and SR/SR-2 indicating a less diverse repertoire but it remained stable or decreased in patient SR/SR-3 and in the SR/CI group where no expansion was observed ( Fig 4A and 4B ) . Next , we examined richness of the repertoire . This parameter measures the number of clonotypes per sample [26] . We observed decreased richness index in the SR/SR group ( mostly SR/SR-1 and SR/SR-2 ) indicating focusing of the repertoire while the same index remained stable in patient SR/SR-3 and slightly increased in the SR/CI group ( Fig 4C and 4D ) . Next , we examined evenness of the repertoire . This parameter measures the relative abundance of the different clonotypes [26] . Again , we observed decreased evenness at peak reinfection in patients SR/SR-1 and SR/SR-2 indicating focusing of the repertoire ( Fig 4E and 4F ) while the same index slightly increased in patient SR/SR-3 and in the SR/CI group . Finally , we used the Morisita Horn index to compare the repertoires from pre-reinfection and peak reinfection , as well as from peak reinfection and late reinfection . This index is indicative of the overlap between two samples as it takes into account both the number of clonotypes and their frequencies within the repertoire [27] . An index of 1 represents complete overlap or identical repertoire and an index close to 0 represents two very different repertoires with almost no clonotypes shared between the samples . The analysis for SR/SR patients showed a Morisita Horn index around 0 . 6 when comparing the pre-reinfection and peak reinfection time points , indicating changes in the repertoire between these two time points . The repertoire was stable between peak reinfection and late reinfection as indicated by a Morisita Horn index around 0 . 9 . We observed the opposite trend for SR/CI patients , where the repertoire underwent more changes between the peak reinfection and late reinfection ( Morisita Horn index of 0 . 6 ) time points as compared to the changes between pre-reinfection and peak reinfection ( Morisita Horn index of 0 . 8 ) suggesting increased diversification with establishment of chronic infection . Collectively , all four measures demonstrate focusing of the repertoire in the SR/SR group and unchanged or even slight increase in diversity in the SR/CI group . Next , we examined CDR3 amino acid ( aa ) length distribution . As a reference of CDR3 lengths distribution from an unselected repertoire , we analyzed TCR sequences of naive CD8 T cells sorted from patient SR/CI-2 ( Fig 5C ) . As shown by the bell-shaped curve , the naive repertoire displayed a normal Gaussian distribution . The distribution was identical in the naive compartment in the three time points tested ( preinfection , peak and follow-up ) . In contrast , this normal distribution was altered in HCV-specific T cells from all patients analyzed ( Fig 5A and 5B ) , reflecting the antigen-specific selected population . Furthermore , this analysis showed an expected bias towards the CDR3 lengths of the most dominant clonotypes . For example , the CDR3 length distribution for patient SR/SR-2 was highly biased towards a length of 14 aa , which reflects the fact that the 2 highly-dominant clonotypes ( representing 78–96% of the repertoire at the different time points ) possess CDR3 of that length ( Fig 5A , middle ) . In addition , for patient SR/SR-1 the CDR3 length distribution was biased towards CDR3s with a length of 13 aa ( clonotype TCRVB20-TCRVJ01 . 01 , with frequencies of 8–35% of the repertoire , Fig 5A , left ) . Another important observation from this analysis is that , in the SR/SR group , we could clearly see an increase in the dominance of CDR3 of certain length between the pre-reinfection time point and the peak reinfection time point ( 13 aa in SR/SR-1 and 14 aa in SR/SR-2 ) . This is in agreement with the focusing of the repertoire in those patients . In the SR/CI group ( Fig 5B ) , we also observed a bias towards the length of the most abundant clonotypes ( 13 and 17 aa for SR/CI-2 and 15 aa for SR/CI-3 , Fig 5B ) but the CDR3 length distribution remained similar across the different time points . This reflected our earlier observation of limited to no focusing of the repertoire during reinfection in these patients . Next , we compared average CDR3 lengths , nucleotides ( NT ) addition and Germline index measuring junctional diversity for all patients as described by Yu et al [28] , but there was no clear difference between groups or across time points ( S5 Fig ) further confirming our earlier observation of limited novel diversification in the repertoire upon reinfection . Studies in HIV infection have demonstrated that TCR avidity towards pMHC correlated with CD8 T cell functionality , better control of viral replication and overall lower viral loads [8 , 29 , 30] . So we sought to evaluate whether selection and expansion of specific clonotypes during HCV reinfection was associated with higher affinity or functional avidity that would endow them with a superior protective capacity . To establish individual cell lines specific to the HLA A2 restricted NS3 epitope ( A2/NS3-1073; CINGVCWTV ) , HCV tetramer+CD8+ T cell were sorted and cultured under limiting dilution . Cell lines were generated from two patients: SR/SR-1 ( 69 cell lines generated ) and SR/CI-2 ( 36 cell lines generated ) . Five cell lines from each patient were selected for detailed analysis termed hereinafter cell lines R1-R5 and cell lines C1-C5 generated from patients SR/SR-1 and SR/CI-2 , respectively . We first evaluated the avidity of individual cell lines , a parameter that measures both binding strength and the total number of interactions on the T cell surface using a tetramer dilution assay . Both the percentage of tetramer positive cells and the mean fluorescence intensity ( MFI ) were equivalent for all cell lines regardless if they originated from a SR/SR or a SR/CI patient ( Fig 6 ) . Next , we evaluated the functionality of individual cell lines or micropopulations in response to stimulation with the cognate NS3-1073 peptide in an intracellular cytokine staining ( ICS ) assay . We evaluated the surface expression of the degranulation marker CD107a and intracellular expression of TNFα , IFNγ and IL-2 ( Fig 7A–7D ) . Representative intracellular cytokine staining ( ICS ) data are presented in S6 Fig . As shown in Fig 7A and 7B , CD8 T cell lines generated from the SR/SR patient were more sensitive to lower peptide concentrations and showed enhanced CD107a expression ( average EC50 SR/SR = 1 . 8x10-7; SR/CI:5 . 5x10-6 ) and TNFα production ( average EC50 SR/SR = 1 . 5x10-7; SR/CI:3 . 5x10-5 ) . The capacity to produce IFNγ and IL-2 was clone dependent , and no clear difference was observed between cell lines from the SR/SR versus cell lines from the SR/CI patient ( Fig 7C and 7D ) . In general , production of IL-2 was low ( Fig 7D ) . The polyfunctionality of CD8 T cells measured as the capacity of single cells to produce multiple functions is a key determinant of spontaneous resolution of HCV infection [32 , 33] . Boolean gating and pie chart analysis demonstrated that cell lines generated from the SR/SR-1 patient had a higher degree of polyfunctionality compared to cell lines generated from the SR/CI-2 patient ( Fig 7E and Supplementary S6B and S6C Fig ) . At maximum peptide concentration , we observed that an average of 85% of the cells displayed at least two functions for cell lines generated from SR/SR-1 , compared to 30% of the cells for cell lines generated from SR/CI-2 ( Fig 7E and S6B and S6C Fig ) . Furthermore , at limited peptide concentration ( 0 . 01μg/ml ) , an average of 50% of the cells were positive for at least one function for SR/SR-1 cell lines compared to an average of 13% for SR/CI-2 cell lines . Polyfunctionality index was calculated as previously described [31] to compare the overall polyfunctionality of all cell lines . As shown in Fig 7F , the degree of polyfunctionality was higher for all SR/SR-1 cell lines compared to SR/CI-2 cell lines . These results suggest that secondary clearance upon reinfection is associated with the presence of polyfunctional CD8 T cell clonotypes . To determine whether the functionality of the individual cell lines reflected preferential expansion in vivo during reinfection , we have sequenced the TCR of all 10 cell lines tested from patient SR/SR-1 ( cell lines R1-R5 ) and SR/CI-2 ( cell lines C1-C5 ) . As demonstrated in S8 Table , all cell lines isolated from patient SR/SR-1 with the exception of clone R3 carried TCR clonotypes that were detectable at frequencies ranging from 0 . 7% to 33% at the peak of the immune response during reinfection although some of them were micropopulations composed of two different clonotypes . Nevertheless , cell line R1 , one of the best responding cell lines , was partially ( 7% ) composed of TRBV-20 that showed 33% expansion during peak reinfection . Similarly , cell line R5 , with the highest TNF-α production was partially ( 12 . 5% ) composed of TRBV27-01*01 that showed 8% expansion during peak reinfection . Altogether , these data demonstrate that secondary clearance upon HCV re-exposure and reinfection is associated with selective expansion of high functional avidity CD8 T cell clonotypes .
Defining the correlates of protective immunity at the clonotypic level is essential for fine-tuning the design of prophylactic vaccines against chronic viruses with highly variable sequences such as HIV and HCV . This study provides an insight into the dynamics of the virus-specific CD8 T cell repertoire during HCV re-exposure and reinfection in a real-life setting . Our results demonstrate that protective immunity and rapid virus clearance upon reinfection was associated with expansion of a limited number of polyfunctional CD8 T cell clonotypes selected from the memory pool . These CD8 T cells displayed a focused repertoire and cell lines established from one SR/SR patient were characterized by high functional avidity and polyfunctionality . In contrast , very little expansion of HCV specific CD8 T cell was observed in SR/CI patients who developed persistent viremia upon reinfection and their repertoire was more diverse . Cell lines established from one SR/CI patient displayed reduced functional avidity demonstrated by a much weaker production of cytokines and lower cytotoxic potential that could potentially have facilitated virus persistence and chronicity . Our observations with the cell lines reflected the ex vivo characterization of the functionality of HCV-specific CD8 T cells in these patients [23] . Our longitudinal analysis of the TCR repertoire , using three different HCV tetramers , showed conserved clonotype usage within the same individual where the dominant and sub-dominant clonotypes forming the effector population at the peak of the immune response during reinfection were recruited from the pre-existing memory T cell pool generated following clearance of the primary infection . New clonotypes were only detected at low abundance frequencies . Two out of three patients in the SR/SR group had dominant clonotypes following primary infection and the same clonotypes expanded upon reinfection with a homologous variant , suggesting that these particular clonotypes played an important role in viral clearance . The third SR/SR patient displayed a focused repertoire before reinfection but showed limited expansion during reinfection . We have previously shown that this patient was reinfected with a different HCV subtype ( genotype 1b after primary infection with genotype 1a ) which could affect his immune response . In contrast , there was very limited expansion of HCV specific CD8 T cells in the SR/CI group , even despite reinfection with the same HCV subtype for patient SR/CI-2 ( Fig 1 ) . It is noteworthy that our previous analysis of autologous virus sequences demonstrated that the SR/CI patients were infected with variant viruses that were not recognized by pre-existing memory T cells [23] . It was not possible to sequence virus from the primary infection in those patients , preventing a clear comparison between both infection episodes . That potential mismatch between the two infecting viral strains could be responsible for the lack of significant expansion of the pre-existing memory T cells or the generation of new clonotypes . Our results are in concordance with data from the LCMV model demonstrating that the TCR repertoire of the primary epitope-specific CD8 T cell response was conserved in the memory pool , and that after a secondary effector recall response , there was 60–100% identity between the clonotypes of the primary effector , memory and recall responses [22 , 34] . Rapid resolution of HCV infection upon rechallenge in chimpanzees also coincided with the expansion of T cell clonotypes that dominated the memory CD8 T cell pool [17 , 21] . Cross reactivity between NS3-1073–specific CD8 T cells and an influenza virus epitope ( NA-231 ) was previously reported [35 , 36] . In addition , CD8 T cells reactive to this epitope could be amplified from a significant proportion of healthy individuals with no prior exposure to HCV [35 , 36] . Thus , it is possible that this may have influenced the specific CD8 T cell repertoire analyzed in our study . However , in these previous studies , the overall HCV-specific immune response was narrowly focused towards the NS3 region , which is not the case in patients SR/SR-1 and SR/CI-2 as we have previously described using ELISPOT assays [23] . Furthermore , it was previously demonstrated that these influenza cross-reactive T cells were of low affinity/avidity , and are unlikely to play a major role against HCV infection [37] . Indeed , in our hands , individual T cell lines generated from either the SR/SR or the SR/CI patient exhibited comparable avidities . Finally , TCR repertoire of cross reactive CD8 T cells was previously shown to be “private” as it varied greatly from one patient to another , suggesting that these cells do not carry a specific dominant receptor [38] . Future analysis using a larger cohort of patients responding to this epitope will be required to clarify this point . The TCR repertoire was narrower and less diverse in SR/SR patients than in the SR/CI patients at the pre-infection time point . Furthermore , this repertoire became more focused at the peak of the immune response during reinfection in patients SR/SR-1 and SR/SR-2 and retained a highly focused composition post-reinfection . This data was confirmed by examining various measures of diversity , richness and evenness . It is possible that those patients , that were recruited as resolvers had previously cleared more than one infection and may have been exposed to more than one genotype , which could have selected the most efficient repertoire . Focusing of the CD8 TCR repertoire upon re-exposure was reported in various infection and vaccination models including LCMV [34] . Focused CD8 TCR repertoires with selection of high-avidity T cell populations were also reported in HIV-1 slow progressors [39] . In contrast , a previous study of an HCV epitope ( NS3-1406 ) suggested that higher diversity of the repertoire would be more advantageous in order to offset viral escape as a result of epitope mutation [20] . A chimpanzee challenge study reached a similar conclusion , since the majority of HCV epitopes that escaped immune recognition upon infection were targeted by a CD8 T cell repertoire with reduced CDR3 amino acid diversity , suggesting that limited TCR diversity facilitates CTL escape mutations in this animal model [17] . Results from the chimpanzee model may not accurately reflect the real life re-exposure for several reasons . First , chimpanzees in these studies were rechallenged with a specific and homologous viral sequence and therefore were exposed to a less diverse viral population as compared to humans that are typically exposed to a complex mixture of viral variants . Second , the chimpanzee data were generated using CD8 T cell clones derived from the livers of infected animals . Examining the CD8 TCR repertoire in the liver of humans is ethically difficult . Although previous analysis demonstrated that PBMCs are a good reflection of the intrahepatic TCR repertoire [21] , certain low frequency clonotypes could be enriched within the liver and the in vitro expansion step employed in these studies may have introduced some bias in the repertoire . The Morisita Horn index showed that there were more changes in the repertoire of SR/SR patients upon reinfection ( from pre-reinfection to peak reinfection ) as compared to SR/CI patients . This could reflect the efficient priming and focusing of the T cell response in SR/SR patients , leading to virus clearance as discussed above . Interestingly , although we could not detect significant expansion of HCV-specific tetramer+ CD8 T cells and associated clonotypes in the SR/CI patients between pre-reinfection and peak , the Morisita Horn index reflected more changes later in reinfection between the peak and late reinfection time points . This suggests that the repertoire has undergone changes following the establishment of chronic infection and the accumulation of viral variants that would prime the expansion of different clonotypes as we also observed that the repertoire became more diverse in those patients ( S3 Fig ) but obviously this diversification was not enough to clear the virus . Additional analysis accompanied by in depth sequencing of the infecting viral strain at each episode and of instances where the variant epitope is still recognized by the specific T cells are essential to elucidate the interaction between the T cell repertoire , the infecting virus sequence and the emergence of escape mutations . The presence of a common set of TCRs associated with protection , known as public repertoires , was associated with viral control in several infections including HIV and CMV [5 , 6 , 40] . Miles and colleagues have also reported a biased TCR repertoire towards a common core of public repertoires in individuals with long-term spontaneous clearance or persistent HCV infection [19] . Our longitudinal analysis of the dominant and subdominant Vβ-CDR3-Jβ clonotypes in the SR/SR and SR/CI patients revealed no overlap between patients with the same HLA background and targeting the same epitope . Given the limited number of patients included in this study , it was not possible to draw a definitive conclusion about whether or not specific public clonotypes were associated with secondary clearance . Additional analysis of a larger cohort of patients targeting the same pMHC is required . Establishing CD8 T cell lines enabled us to characterize the molecular determinants of functionality of the individual clonotypes . We did not observe different avidity patterns among the 10 cell lines that were analyzed using the tetramer titration assay . Testing more cell lines might be necessary to identify some with a range of TCR avidity as was observed in the HIV and LCMV models [30 , 41] . It is also possible that this particular epitope selected mostly CD8 T cells with high avidity and that examining other epitopes may yield different results [42] . Furthermore , this study was performed on individuals that have successfully cleared a previous infection . It is thus possible that the clonotypes that were selected during primary infection and formed the memory pool are those with the highest avidity . Indeed , studies in a mouse model of influenza infection and rechallenge demonstrated that the clonotypes expanding in the recall response were those with the highest avidity [43] . Establishing cell lines specific to the same epitope but from early primary infection samples would thus also be informative . Moreover , a more sensitive method such as surface plasmon resonance might provide a more accurate measure of binding affinity and/or avidity that might be different between cell lines [44] . Functional avidity , or the capacity of a particular clone to translate TCR binding into a functional response , was strikingly different between the cell lines from the SR/SR patient compared to the cell lines from the SR/CI patient , especially for the surface expression of the degranulation marker CD107a and TNFα production . The cell lines established from the SR/SR patient responded well to lower peptide concentration . The polyfunctionality level was also greater in the cell lines from SR/SR as compared to the cell lines from SR/CI and is therefore an important correlate of control of viral replication [8 , 45] . We have already demonstrated a broad IFNγ response in the SR/SR-1 patient as measured by ELISPOT assays but the polyfunctional CD8 T cell response to different epitopes was dominated by the production of TNFα and the surface expression of CD107a and that the CD8 T cell response had an increased magnitude and polyfunctionality in the SR/SR patients compared to the SR/CI patients [23] . Hence , the data from the individual cell lines reflected well the overall in vivo response . It is possible that in this individual , cytolytic effector functions ( CD107a ) leading to killing of infected cells provide an overall better antiviral effect as compared to non-cytolytic ( IFNγ mediated ) functions . Similarly , TNFα systemic levels increase during HCV infection [46] and it can have multiple antiviral and inflammatory effects . Specifically , it can induce the apoptosis of HCV infected hepatocytes and bystander cells in the liver , which could enhance viral clearance [47 , 48] . Our study focused on the CD8 T cell response to HCV reinfection . Another important component that remains to be evaluated is the role of the antibody response and the antibody repertoire in protection upon reinfection . Indeed , Osburn et al have demonstrated that reinfection is associated with the development of cross-reactive antibodies [49] . The recent development of novel E2-tetramers that allow sorting and characterization of HCV-specific B cells and antibody repertoire [50] represent an invaluable tool to dissect the role of T cells versus antibodies in protection against reinfection . In conclusion , our results demonstrate that epitope-specific CD8+ T cell clonotypes expanding at the peak of reinfection are recruited from the memory pool , rather than being de novo clonotypes mobilized from the naïve pool . The repertoire is narrower in the SR/SR patients who were protected against viral persistence in comparison to SR/CI patients , and it becomes more focused upon reinfection in 2/3 patients . Analysis of individual CD8 T cell lines from SR/SR versus SR/CI patients revealed that HCV-specific cells associated with resolution of the reinfection had a better functional avidity and polyfunctionality rather than improved avidity of the TCR . Vaccination strategies aiming at enhancing the expansion and polyfunctionality rather than the diversity of HCV-specific T cells by use of adjuvants or immune modulators could be an interesting strategy to follow .
Study subjects were enrolled among PWIDs participating in the Montreal Acute Hepatitis C Cohort Study ( HEPCO ) [51] . This study was approved by the Institutional Ethics Committee of CRCHUM ( Protocol SL05 . 014 ) . All samples were anonymized . Primary HCV infection was identified in HEPCO participants who were initially negative for both HCV RNA and anti-HCV antibodies for at least 6 months , then had a positive HCV RNA and/or antibody test as previously described [32 , 51] . Participants who resolved primary HCV infection or participants who tested HCV RNA-negative and HCV antibody-positive at recruitment were enrolled in the reinfection study and followed every 3 months thereafter . HCV reinfection was defined by an HCV-RNA positive test following two negative tests that were performed ≥ 60 days apart . The day of the first positive RNA test was defined as day zero post detection of reinfection . Five cases of reinfection were identified between 2009 and 2012 for whom samples collected prior to reinfection were available . Clinical outcomes and immunological responses in these patients were previously reported [23] . Three patients spontaneously resolved ( SR ) their second infection ( SR/SR group ) while two patients became chronically infected ( SR/CI group ) . Patients’ demographics and clinical characteristics are summarized in S1 Table . MHC class I tetramers were synthesized by the National Immune Monitoring Laboratory ( NIML ) , ( Montréal , QC , Canada ) or the NIH Tetramer Core Facility ( Emory University , Atlanta , GA ) . The following tetramers were used: HLA-A1 restricted HCV NS3 peptide amino acids ( aa ) 1436–1444 ( ATDALMTGY ) [A1/NS3-1436] , HLA-A2 restricted HCV NS3 peptide aa 1073–1081 ( CINGVCWTV ) [A2/NS3-1073] , and HLA-B27 restricted HCV peptide NS5B peptide aa 2841–2849 ( ARMILMTHF ) [B27/NS5B-2841] . Cryopreserved peripheral blood mononuclear cells ( PBMC ) were thawed and CD8+ T cells were purified using the negative selection MACS CD8+ T cell Isolation Kit ( Miltenyi Biotec Inc , Auburn , CA ) . Tetramer staining and cell surface staining for CD3 , CD8 , CD45RO and CD127 were performed as previously described [32] . Directly-conjugated monoclonal antibodies against the following molecules were used: CD3–FITC ( clone UCHT1 ) , CD8–Pacific Blue ( clone RPA-T8 ) and CD45RO–Alexa Fluor 700 ( clone UCHL1 ) were obtained from BD Biosciences ( San Diego , CA ) . CD127/IL-7Ra–Alexa Fluor 647 ( clone HIL-7R-M21 ) was obtained from eBioscience ( San Diego , CA ) . Live cells were identified using LIVE/DEAD fixable aqua dead cell stain kit ( Molecular Probes Thermo Fisher Scientific , Burlington , ON ) . Multiparameter flow cytometry was performed on a BD Aria II cell sorter equipped with blue ( 488 nm ) , red ( 633 nm ) , and violet ( 405 nm ) lasers or a BD LSRII instrument equipped with an additional yellow-green laser ( 561 nm ) using FACSDiva version 6 . 1 . 3 ( BD Biosciences ) . Data files were analyzed using FlowJo version 9 . 4 . 11 for Mac ( Tree Star , Inc . , Ashland , OR ) . Genomic DNA was extracted from sorted cells and the variable ( Vβ ) , diversity ( Dβ ) and joining ( Jβ ) regions of the TCR β chain were sequenced using an automated high-throughput method ( Adaptive Biotechnologies , Seattle , WA ) . Briefly , CDR3 regions were amplified using a two-step amplification bias-controlled multiplex PCR approach [52] . Amplified libraries were sequenced using an Illumina instrument according to the manufacturer’s instructions . Demultiplexed reads were then further processed to reduce amplification and sequencing bias [53] . The resulting CDR3 amino acid sequences were classified into correct families according to the IMGT database ( www . imgt . org ) . Data were analyzed using ImmunoSEQ software ( v2 . 0 ) . Clonotype lists ( CDR3 sequences and frequencies within the repertoire ) were cleaned to remove out of frame sequences and sequences with stop codons within the CDR3 region . Clonotypes with a frequency of less than 0 . 01% of the total repertoire were excluded from the analysis since the number of events/sequences would correspond to less than one cell . TCR sequences raw data are available at ( https://clients . adaptivebiotech . com/pub/shoukry-2017-plospathogens ) . S2 Table details the number of sorted cells , total / unique productive sequences and clonality for each sample . Repertoire Simpson diversity index , richness , Shannon entropy index and Morisita Horn index were provided by Adaptive Biotechnologies . Richness index was calculated as the observed richness divided by the input cell number . Evenness was calculated as the Shannon entropy divided by the log of the observed richness . Venn diagrams were generated by comparing the amino acids clonotypes sequences across time points for each patient , after excluding clonotypes with frequencies < 0 . 01% as explained above . CDR3 length and NT addition for each clonotypes were provided by Adaptive Biotechnologies and the Germline index was calculated as: ( Total CDR3 length–total NT additions ) / Total CDR3 length . HCV specific , tetramer positive ( A2/NS3-1073 ) CD8+ T cells from two patients ( SR/SR-1 and SR/CI-2 ) were enriched and sorted as described above . Cells were diluted in R10 ( RPMI 1640 + 10% heat inactivated fetal bovine serum ( FBS; Life Technologies ) supplemented with penicillin + streptomycin ( pen/strep , 1X , Wisent ) and 40U/ml rIL-2 ( NIH-AIDS Reagents Program ) ( R10-P/S-IL2 ) ; and plated at a concentration of 5 cells per well in 96 well plates in presence of 5 x 104 non-autologous irradiated ( 30 Gy ) PBMCs as feeder cells and 0 . 01μg/ml anti-CD3 ( Beckman Coulter ) . Cells were cultured for two weeks in 96 well plates and half of the medium was replenished every 3 days ( R10-P/S-IL2 ) . Growing cell lines were then transferred to 24 well plates with a new round of stimulation with feeder cells ( 2x106 cells/well ) and anti-CD3 ( 0 . 01μg/ml final ) . Cell lines were then cryopreserved in freeze mix ( FBS + 10% DMSO ) at a concentration of 5x106 cells/ml . Avidity of T cell lines was assessed by staining with serial dilutions of tetramers ( A2/NS3-1073; 10μg/ml to 0 . 02μg/ml , two fold dilutions ) for 30 min at room temperature in the dark . Surface staining included live/dead marker , CD3-PB ( clone UCHT1 ) , CD4-BV605 ( clone RPA-T4 ) , CD8-APC-H7 ( clone SK1; all from BD Biosciences ) and flow cytometry was performed as above . T cell lines were stimulated with autologous EBV transformed B cell line ( BLCLs ) at a ratio of 2:1 ( T cell: BLCLs ) . BLCLs were irradiated at 100 Gy and prepulsed with HCV NS3 peptide ( NS3-1073-1081; CINGVCWTV ) for 1 h at 37°C in R-10 medium . BLCLs were then washed and incubated with T cell lines for 6 hours in AIM-V medium ( Life Technologies ) supplemented with 10% human serum ( Wisent ) and anti-CD107a-BV786 antibody ( clone H4A3; BD Bioscience ) . 10 μg/ml Brefeldin A ( BFA , Sigma ) and 6 μg/ml monensin ( Sigma ) were added after 1 h of stimulation . After stimulation , cells were washed and surface staining was performed as described in the tetramer titration section . Cells were then permeabilized with CytoFix/CytoPerm ( BD Biosciences ) for 15 minutes at 4°C in the dark , washed again , and incubated with anti-IFNγ-PE-Cy7 ( clone B27 ) , anti-TNF-α-PerCP-Cy5 . 5 ( clone MAb11 ) and anti-IL-2-PE ( clone MQ1-17H12; all from BD Biosciences ) for 30 min at 4°C in the dark . Cells were then washed , fixed and analyzed as above . Polyfunctionality was analyzed using Boolean gating and SPICE software [54] . TCR sequences raw data are available at https://clients . adaptivebiotech . com/pub/shoukry-2017-plospathogens | In this study we examined the diversity and dynamics of the repertoire of receptors of CD8 T cells that are selected and enriched upon real-life multiple exposures to viral infections . Using hepatitis C virus ( HCV ) infection in a cohort of high risk people who inject drugs , we demonstrate that protection upon two subsequent infections was associated with a narrow repertoire of virus-specific CD8 T cells and selective expansion of cells with high polyfunctionality ( increased TNFα production and cytotoxic potential ) . Our results have important implications in vaccination programs aiming at shaping the CD8 T cell repertoire against viral infections and cancers . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"hepacivirus",
"pathogens",
"immunology",
"microbiology",
"cloning",
"viruses",
"rna",
"viruses",
"cytotoxic",
"t",
"cells",
"molecular",
"biolo... | 2017 | Selective expansion of high functional avidity memory CD8 T cell clonotypes during hepatitis C virus reinfection and clearance |
Aedes aegypti , the ‘yellow fever mosquito’ , is the primary vector to humans of dengue and yellow fever flaviviruses ( DENV , YFV ) , and is a known vector of the chikungunya alphavirus ( CV ) . Because vaccines are not yet available for DENV or CV or are inadequately distributed in developing countries ( YFV ) , management of Ae . aegypti remains the primary option to prevent and control outbreaks of the diseases caused by these arboviruses . Permethrin is one of the most widely used active ingredients in insecticides for suppression of adult Ae . aegypti . In 2007 , we documented a replacement mutation in codon 1 , 016 of the voltage-gated sodium channel gene ( para ) of Ae . aegypti that encodes an isoleucine rather than a valine and confers resistance to permethrin . Ile1 , 016 segregates as a recessive allele conferring knockdown resistance to homozygous mosquitoes at 5–10 µg of permethrin in bottle bioassays . A total of 81 field collections containing 3 , 951 Ae . aegypti were made throughout México from 1996 to 2009 . These mosquitoes were analyzed for the frequency of the Ile1 , 016 mutation using a melting-curve PCR assay . Dramatic increases in frequencies of Ile1 , 016 were recorded from the late 1990's to 2006–2009 in several states including Nuevo León in the north , Veracruz on the central Atlantic coast , and Yucatán , Quintana Roo and Chiapas in the south . From 1996 to 2000 , the overall frequency of Ile1 , 016 was 0 . 04% ( 95% confidence interval ( CI95 ) = 0 . 12%; n = 1 , 359 mosquitoes examined ) . The earliest detection of Ile1 , 016 was in Nuevo Laredo on the U . S . border in 1997 . By 2003–2004 the overall frequency of Ile1 , 016 had increased ∼100-fold to 2 . 7% ( ±0 . 80% CI95; n = 808 ) . When checked again in 2006 , the frequency had increased slightly to 3 . 9% ( ±1 . 15% CI95; n = 473 ) . This was followed in 2007–2009 by a sudden jump in Ile1 , 016 frequency to 33 . 2% ( ±1 . 99% CI95; n = 1 , 074 mosquitoes ) . There was spatial heterogeneity in Ile1 , 016 frequencies among 2007–2008 collections , which ranged from 45 . 7% ( ±2 . 00% CI95 ) in the state of Veracruz to 51 . 2% ( ±4 . 36% CI95 ) in the Yucatán peninsula and 14 . 5% ( ±2 . 23% CI95 ) in and around Tapachula in the state of Chiapas . Spatial heterogeneity was also evident at smaller geographic scales . For example within the city of Chetumal , Quintana Roo , Ile1 , 016 frequencies varied from 38 . 3%–88 . 3% . A linear regression analysis based on seven collections from 2007 revealed that the frequency of Ile1 , 016 homozygotes accurately predicted knockdown rate for mosquitoes exposed to permethrin in a bioassay ( R2 = 0 . 98 ) . We have recorded a dramatic increase in the frequency of the Ile1 , 016 mutation in the voltage-gated sodium channel gene of Ae . aegypti in México from 1996 to 2009 . This may be related to heavy use of permethrin-based insecticides in mosquito control programs . Spatial heterogeneity in Ile1 , 016 frequencies in 2007 and 2008 collections may reflect differences in selection pressure or in the initial frequency of Ile1 , 016 . The rapid recent increase in Ile1 , 016 is predicted by a simple model of positive directional selection on a recessive allele . Unfortunately this model also predicts rapid fixation of Ile1 , 016 unless there is negative fitness associated with Ile1 , 016 in the absence of permethrin . If so , then spatial refugia of susceptible Ae . aegypti or rotational schedules of different classes of adulticides could be established to slow or prevent fixation of Ile1 , 016 .
Aedes aegypti , the ‘yellow fever mosquito’ , is the primary vector to humans of dengue and yellow fever flaviviruses ( DENV , YFV ) [1]–[3] . Vaccines are not yet available against DENV [4] and , despite the presence of a safe and effective YFV vaccine [5]–[7] , the World Health Organization estimates there are 200 , 000 cases and 30 , 000 deaths attributable to yellow fever each year [8] . The principal means to reduce transmission of these arboviruses has therefore been through control or eradication of Ae . aegypti [9] . Historic eradication campaigns that combined source reduction to remove larval development sites with use of dichloro-diphenyl-trichloroethane ( DDT ) to kill adults were successful in eliminating the mosquito and its associated arboviruses , especially in the Americas , but these programs were not sustained and Ae . aegypti and DENV re-emerged in force [10] , [11] . In recent decades , pyrethroid insecticides have played a major global role in the control of Ae . aegypti adults , often in combination with the organophosphate insecticide temephos to control immatures . However , the evolution of resistance to these and other insecticides in Ae . aegypti may compromise the effectiveness of control programs [12]–[14] . Since 1950 , operational vector control programs in México have used a series of insecticides to control mosquito vectors and reduce arbovirus and malaria transmission ( Official Regulations of México , NOM-032-SSA ) [12] . The organochlorine insecticide DDT was used extensively for indoor house spraying from 1950–1960 and was still used in some locations until 1998 . Organophosphate insecticides with malathion as the active ingredient were later used for ultra-low volume ( ULV ) space spraying of wide areas from 1981 to 1999 . In 2000 , vector control programs in México then switched to permethrin-based insecticides for adult control . This has provided prolonged and intense selection pressure for resistance evolution in Ae . aegypti . Indeed , pyrethroid insecticides with active ingredients such as permethrin , deltamethrin , resmethrin and sumithrin are now commonly applied across the world to kill adult mosquitoes and reduce the burden of mosquito-borne diseases . The future global use of bednets , curtains and other household items treated with pyrethroids for personal protection will likely increase dramatically [15]–[18] . This underscores the critical need to monitor and manage resistance to pyrethroid insecticides to maintain their use for vector control . Pyrethroids act by structure-related interactions with specific regions of voltage-dependent sodium channels that prolong the opening of these channels , and produce instant paralysis [19] . Nervous system stimulation proceeds from excitation to convulsions and tetanic paralysis . Metabolic resistance and target site insensitivity are both major forms of pyrethroid resistance [19] , [20] . ‘Knockdown resistance’ ( kdr ) is a generic term applied to insects that fail to lose coordinated activity immediately following pyrethroid exposure . Typically kdr is unaffected by synergists that inhibit esterases and monooxygenases . Instead kdr arises through nonsynonymous mutations in the voltage-gated sodium channel transmembrane gene ( orthologue of the paralysis locus in Drosophila melanogaster ) [21] that reduce pyrethroid binding . Kdr usually limits the effectiveness of pyrethroids to varying degrees depending on whether the insecticide contains a descyano-3-phenoxybenzyl alcohol ( type I pyrethroid ) or an α-cyano-3-phenoxybenzyl alcohol ( type II ) . Thus detection of kdr in the field may have severe consequences for sustained use of pyrethroids in mosquito control . A homology model of the housefly para protein was developed [22] to predict the location of binding sites for the pyrethroid , fenvalerate and for DDT . The model addressed the state-dependent affinity of pyrethroid insecticides , their mechanism of action and the role of mutations in the channel that are known to confer insecticide resistance . Specifically , the sodium channel was modeled in an open conformation with the insecticide binding site located in the hydrophobic cavity delimited by the domain II subunit 4 ( IIS4 ) - IIS5 linker and the IIS5 and IIS6 helices . Five novel mutations in IIS6 , one in IIS5 and one in the P loop were described in the para orthologue in Ae . aegypti [23] . Assays on larvae from strains bearing these mutations indicated reduced nerve sensitivity to permethrin inhibition . Two of these mutations occurred in codons Ile1 , 011 and Val1 , 016 in exons 20 and 21 , respectively . A transition in the third position of Ile1 , 011 encoded a Met1 , 011 replacement and a transversion in the second position of Val1 , 016 encoded a Gly1 , 016 replacement . This same region of IIS6 was later screened in 1 , 318 mosquitoes in 32 additional strains; 30 from throughout Latin America [24] . The Gly1 , 016 allele was never detected in Latin America and instead we found two new mutations in these same codons . A transition in the first position of codon 1 , 011 encodes a valine replacement while a transition in the first position of codon 1 , 016 encoded an isoleucine replacement . We developed melting curve PCR assays for these four mutations . Selection experiments , one with deltamethrin on a field strain from Santiago de Cuba and another with permethrin on a strain from Isla Mujeres , México rapidly increased the frequency of the Ile1 , 016 allele [25] . In bioassays of F3 offspring arising from crosses of permethrin susceptible Val1 , 016 homozygous parents and permethrin resistant Ile1 , 016 homozygous parents , Ile1 , 016 segregated as a recessive allele conferring knockdown resistance to homozygous mosquitoes at 5–10 µg permethrin in bottle bioassays , 4 . 3–14 . 0% resistance in heterozygous mosquitoes [24] . All Val1 , 016 homozygous mosquitoes died . Herein we report on an analysis of the frequency of the Ile1 , 016 mutation in 3 , 808 Ae . aegypti from 78 collections made from 1996–2008 throughout México . The overall frequency was 0 . 04% from 1996–2001 , had climbed to 2 . 7% by 2003–2004 , and increased only slightly to 3 . 6% by 2006 . Then , as would be expected with a recessive allele , Ile1 , 016 frequency rapidly increased to 33% in 2007–2009 . We also document a great deal of spatial heterogeneity in Ile1 , 016 frequency during 2007–2009 . A linear regression analysis based on seven collections from 2007 revealed that the frequency of Ile1 , 016 homozygotes accurately predicted knockdown rate for mosquitoes exposed to permethrin in a bioassay ( R2 = 0 . 98 ) . These results have led us to speculate that widespread use of permethrin-based insecticides in México from 2000–2008 may have resulted in rapidly increasing frequencies of the Ile1 , 016 mutation in Ae . aegypti . Potential implications and solutions for operational vector control are discussed .
Table 1 lists the cities and years of collection for Aedes aegypti , and city locations are mapped in Figure 1 . Single collections were made in those cities marked by * in Figure 1 . Superscripts next to the year of the collection in Table 1 indicate in which of three prior studies [26]–[28] samples were collected . Collections from 2006–2009 have not been included in any prior studies . At each collection site , we collected immatures from at least 30 different containers in each of three different areas located at least 100 m apart . This included water storage containers and discarded trash containers such as plastic pails , tires , and cans . Larvae were returned to the laboratory where they were reared to adults and then identified to species [29] . All mosquitoes were stored at −80°C prior to examination for presence of Ile1 , 016 . DNA was obtained from individual adults by salt extraction [30] , suspended in 300 µl of TE buffer ( 10 mM Tris-HCl , 1 mM EDTA pH 8 . 0 ) , and stored at −80°C . Genotypes at the Ilel , 016 locus were detected using allele specific PCR . Genotypes were determined in a single-tube reaction using two different “allele-specific” primers , each of which contained a 3′ nucleotide corresponding to one of the two alleles and a reverse primer that amplified both alleles . Allele specific primers were manufactured ( Operon Inc . , Huntsville , AL ) with 5′ tails [31] , [32] ( shown in brackets below ) that were designed to allow discrimination between SNP alleles based on size or melting temperature . The Valine allele specific primer was Val1 , 016 ( 5′-[GCGGGCAGGGCGGCGGGGGCGGGGCC]ACAAATTGTTTCCCACCCGCA CCGG-3′ ) and the isoleucine allele specific primer was Ile1 , 016 ( 5′-[GCGGGC]ACAAATTGTT TCCCACCCGCACTGA-3′ ) . Brackets indicate the portion of the primer added for melting curve PCR . The reverse primer was Ile1 , 016r 5′-GGATGAACCSAAATTGGACAAAAGC-3′ [24] . An intentional transversion mismatch was introduced three bases in from the 3′ end of allele specific primers to improve specificity and each allele specific primer differed by a transition at this site [33] . Melting curve PCR was performed as previously described [34] . Ile1 , 016 frequencies ( ) were calculated in each collection as twice the number of Ile1 , 016 homozygotes plus the number of Ile1 , 016 heterozygotes and then divided by twice the number of mosquitoes analyzed . Wright's inbreeding coefficient FIS [35] was estimated as ( 1 ) Where Hobs is the observed number of heterozygotes and is the expected number of heterozygotes where n is the sample size and assuming Hardy-Weinberg proportions . The null hypothesis FIS = 0 was tested using the formula [36]: ( 2 ) The 95% confidence interval ( CI95 ) around was calculated as the Wald interval: ( 3 ) which was then adjusted by adding half of the squared Z-critical value ( 1 . 96 ) to the numerator and the entire squared critical value to the denominator before computing the interval [37] . Fisher's model of natural selection [38] was used to estimate the expected trajectories for the Ile1 , 016 allele for a single population in which: ( 4 ) where: p = Ile1 , 016 frequency in generation t . Following permethrin exposure wIle/Ile is the relative survival of Ile1 , 016 homozygotes , wIle/Val is the relative survival of Ile1 , 016 heterozygotes and wVal/Val is the relative survival of Val1 , 016 homozygotes . Knockdown rates were determined by releasing 40 adults , 3–4 days of age , into 250 mL Wheaton bottles in which the inside walls were coated with either 5 . 0 or 10 . 0 µg of permethrin ( technical grade; Chem Services , West Chester , PA ) [39] . Following a 1-hr exposure period , the number of inactive mosquitoes were recorded .
Table 1 lists the location , collection years , sample sizes and numbers of mosquitoes of each genotype , the frequency of Ile1 , 016 at each site , the 95% confidence interval around that frequency and the FIS estimate and its significance . If FIS was significantly >0 then an excess of homozygotes was present while if FIS was significantly <0 then an excess of heterozygotes was present . FIS was significantly greater or less than zero in 6 of the 40 collections in which Ile1 , 016 was present . FIS values>0 were recorded in three cases because of unexpected Ile1 , 016 homozygotes ( Tantoyuca and Poza Rica 2004 ) or a general deficiency of heterozygotes ( Escuintla 2008 ) . In three cases , FIS values<0 occurred because of an excess of heterozygotes ( Huixtla and Coatzacoalcos 2008; Mérida – Cholul 2009 ) . The map-based representation in Figure 2 shows frequencies of Ile1 , 016 by year of collection for all sites where the allele appeared at least once . Ile1 , 016 first appeared amongst our collections in Nuevo Laredo on the U . S . border in 1997 ( Table 1 , Figure 2 ) . Overall frequency of Ile1 , 016 was very low , 0 . 04% , from 1996–2000 ( CI95 = 0 . 12%; n = 1 , 359 mosquitoes examined ) . This included mosquitoes collected throughout México . No mosquitoes were collected in 2001–2002 . In 2003–2004 , collections were made exclusively in the state of Veracruz which is located along the central Atlantic coast of México . In 2003 , Ile1 , 016 appeared in four collections with an overall frequency of 3 . 49% ( 1 . 18% CI95; n = 487 mosquitoes ) . Notably , the Ile1 , 016 frequency in one site , Pánuco , reached 20 . 0% in 2003 ( 0 . 12% CI95 ) . In 2004 , Ile1 , 016 appeared in only two collections with an overall frequency of 1 . 40% ( 0 . 99% CI95; n = 321 mosquitoes ) . No mosquitoes were collected in 2005 . In 2006 , ten collections were made in the state of Chiapas in the far southwest . Ile1 , 016 appeared in five collections with a cumulative frequency of 3 . 91% ( 1 . 15% CI95; n = 473 ) . In 2007 , five collections were made in the states of Yucatán and Quintana Roo in the Yucatán Peninsula of southeastern México and Ile1 , 016 appeared in all collections with an overall frequency of 57 . 6% ( 4 . 94% CI95; n = 190 ) . These collections also revealed that frequencies were not uniform within cities . For example , the frequency of Ile1 , 016 for collections within the city of Chetumal in Quintana Roo State ranged from 38 . 3% ( 11 . 97% CI95 ) in the Calderitas neighborhood to 88 . 3% ( 8 . 50% CI95 ) in the Lagunitas neighborhood . In 2008 , collections were made at six sites in Veracruz , the same ten sites in Chiapas as in 2006 , two more sites in the city of Chetumal , and one site in Monterrey in Nuevo León State in northern México . In Veracruz , the overall frequency of Ile1 , 016 was 45 . 7% ( 3 . 97% CI95; n = 300 ) and , as for Chetumal in 2007 , there was a great deal of spatial heterogeneity with site-specific Ile1 , 016 frequencies varying between 27 . 0% ( 8 . 62% CI95 ) in the city of Coatzacoalcos to 70 . 0% ( 8 . 87% CI95 ) in the city of Poza Rica . In Chiapas , the overall frequency of Ile1 , 016 was 14 . 5% ( 2 . 23% CI95; n = 480 mosquitoes ) with site-specific frequencies varying between 5 . 0% ( 4 . 80% CI95 ) in the mountain town of Mapastepec to 36 . 0% ( 9 . 26% CI95 ) in the coastal city of Huixtla . In Chetumal , Ile1 , 016 varied between 35 . 0% ( 11 . 77% CI95 ) in the Antorchistas neighborhood to 26 . 7% ( 11 . 02% CI95 ) in the Solidaridad neighborhood . The frequency of Ile1 , 016 in Monterrey , was 50 . 0% ( 10 . 23% CI95 ) . Only 3 collections were made in satellite villages surrounding Mérida in 2009 . No comparisons were made of these collections because they are confounded by year and by their distance from other Mérida collections . Figure 2 illustrates the general trend in México that the frequency of Ile1 , 016 has increased over the last 12 years in the states of Nuevo León , Veracruz , Yucatán , Quintana Roo and Chiapas . This is even more evident in Figure 3 in which the overall frequency of Ile1 , 016 in each state is plotted by year . It should be noted that although Figure 3 suggests that the Ile1 , 016 frequency in Chetumal , Quintana Roo , declined between 2007 and 2008 , this could be related to sampling of sites in different neighborhoods in these years . Linear regression analysis was used to determine how well the frequency of Ile1 , 016 homozygotes in a collection of Ae . aegypti predicts knockdown rate in bioassays . Analyses included F3 adults from five collections from Chetumal , one collection from Isla Mujeres ( northeast of Cancún ) in which Ile1 , 016 is fixed , and one collection from Iquitos in Perú where Ile1 , 016 is absent . The Isla Mujeres strain arose from five generations of permethrin selection . We found strong associations between the frequency of Ile1 , 016 homozygotes and the knockdown rate for exposures of both 5 and 10 µg permethrin per bottle in the bioassay ( R2 = 0 . 98 and 0 . 88 , respectively; Figure 4 ) . Exclusion of Isla Mujeres and Iquitos collections reduced the R2 values from 0 . 98 to 0 . 97 for the 5 µg concentration and from 0 . 88 to 0 . 76 for the 10 µg concentration . Figure 5 shows the expected trajectories for the Ile1 , 016 allele for a population with an initial Ile1 , 016 frequency of 0 . 04% ( as observed in México during 1996–2000 ) . The first two trajectories indicate a rapid fixation of Ile1 , 016 when Val1 , 016 is partially dominant ( 4–14% survival in heterozygous mosquitoes ) . Rapid fixation occurs because the initial frequency of matings among adults carrying Ile1 , 016 is expected to be small ( 0 . 14% in Fisher's model ) . There is also an extreme selection differential among mosquitoes because while the frequency of Ile1 , 016 homozygous and heterozygous mosquitoes are low ( 1 . 37×10−7 and 7 . 4×10−4 respectively in Fisher's model ) , most of the population is killed off by the insecticide . Thus , the frequency of Ile1 , 016 in the next generation becomes high because only homozygous mosquitoes and from 4–14% of the heterozygous mosquitoes survived . Figure 5 also presents a third scenario wherein , following permethrin exposure , mosquitoes recover over the next few hours from the knockdown . This was measured by removing mosquitoes from the bioassay bottle to an insecticide free cage [24] , [25] . We observed that all Ile1 , 016 homozygotes , 58% of heterozygotes and 15% of Val 1 , 1016 homozygotes recovered . Under this set of fitness conditions , Ile1 , 016 goes to fixation more slowly than when Ile1 , 016 is almost completely recessive .
The primary findings of this study are that: ( 1 ) frequencies of Ile1 , 016 in collections of the dengue virus vector Ae . aegypti have increased dramatically in the last decade in several states in México including Nuevo León to the north , Veracruz on the central Atlantic Coast and Chiapas , Quintana Roo , and Yucatán in the south; and ( 2 ) there was a strong association between the frequency of Ile1 , 016 homozygotes in a collection and knockdown rate in a bioassay . This complements earlier work [24] , [25] which documented that , in bottle bioassays with 5 or 10 µg permethrin , Ile1 , 016 segregates as a recessive allele conferring complete knockdown resistance in homozygous mosquitoes , whereas there is 86–96% mortality in heterozygous mosquitoes and complete mortality in Val1 , 016 homozygous mosquitoes . The analysis of predicted Ile1 , 016 frequencies using Fisher's model ( Figure 5 ) was included only to illustrate that a simple model of selection predicts the rapid increases in Ile1 , 016 frequencies that we have observed . Fisher's model assumes a closed population of infinite size that is uniformly exposed to selection . In reality , we find extensive gene flow among all collections within 130 km of one another in northeastern México and within 180 km of one another in the Yucatán [27] . Susceptibility alleles are therefore probably continuously reintroduced into treated populations . Further , permethrin applications are not uniform in and among cities and towns in México . Thus through long distance transport of Ae . aegypti during human commerce , local mosquito migration , and variable levels of permethrin exposure in space and time , there is ample opportunity for recruitment of Val1 , 016 homozygotes into a population . Another major caveat to the model used in Figure 5 is that it assumed that Ile1 , 016 confers the same marginal fitness in the absence of permethrin . In fact , we have observed that it was easy to select Ile1 , 016 homozygous strains in the laboratory but very difficult to maintain them due primarily to egg and early larval instar mortality . It is also interesting that Ile1 , 016 frequency declined in the state of Quintana Roo between 2007 and 2008 and in Ciudad Hidalgo between 2006 and 2008 . These observations are by no means definitive evidence of reduced fitness of Ile1 , 016 in permethrin free environments; the same results could have occurred through genetic drift in the field or through the concentration of deleterious and lethal recessive genes in strains during selection for Ile1 , 016 homozygous strains in the laboratory . Nevertheless , several studies have documented negative fitness effects associated with single-point mutations in para that confer kdr to pyrethroids and DDT . For example , behavioral studies on peach–potato aphids ( Myzus persicae ) showed that a reduced response to alarm pheromone was associated with both gene amplification and a para target-site mutation [40] . In Musca domestica , flies with the identical para mutations showed no positional preference along a temperature gradient while susceptible genotypes exhibited a strong preference for warmer temperatures [41] . Studies of para in Drosophila melanogaster link point mutations to behavioral disturbances . For example , the temperature-sensitive paralysis exhibited by individuals carrying the point mutation napts ( no action potential , temperature sensitive ) resulted from a reduction in the expression of the para gene [42] . This is consistent with a model in which as temperature rises , an increasing fraction of the available sodium channels are required to maintain propagation of action potentials . Fewer channels cannot meet the demands of elevated temperature [43] . The mutation tipE ( temperature-induced paralysis locus E ) also disrupts para expression and confers temperature induced paralysis as a result of a decrease in sodium channel numbers [44] , [45] . The sbl ( smellblind ) mutation , is also an allele of the para gene [46] . This associates sodium channel mutations with olfactory and chemotactic defects [46]–[48] and with changes in sexual behavior [48] , [49] . A 1986 National Research Council report on strategies and tactics for pesticide resistance management [50] , concluded that insecticide susceptibility should be viewed as a “natural resource” at risk of depletion if not managed properly . This lead to the concept of insecticide resistance management ( IRM ) [51]–[54] . One example of the value of implementing IRM schemes for management of mosquito vectors comes from a large-scale field demonstration project in southern Mexico that compared insecticide rotations ( carbamates , organophosphates and pyrethroids ) and insecticide mosaics ( organophosphates and pyrethroids ) with single use of insecticides ( DDT or pyrethroids ) for indoor residual spraying against anopheline malaria vectors [55]–[58] . This project demonstrated that use of insecticide rotations and mosaics , compared to single use of pyrethroids , can reduce resistance to pyrethroid insecticides in anophelines . If the same protocol could be implemented in Ae . aegypti then spatial refugia of susceptible Ae . aegypti or rotational schedules of different classes of adulticides could be established to slow or prevent fixation of Ile1 , 016 . However we recognize that this strategy is ethically more applicable to agricultural pests than to disease vectors because people living in refugia may be at greater risk for acquiring DENV infections . A large literature exists on the repellant properties of pyrethroids [59]–[64] . Their repellency led to the invention and deployment of pyrethroid-treated materials ( curtains , screens and wall hangings ) in the household . Female Ae . aegypti are endophagic and endophilic vector vectors and are almost exclusively anthropophilic [65] . Pyrethroid-treated materials may repel infected female Ae . aegypti from households and thus block DENV transmission in the household both by preventing inhabitants from becoming infected and from allowing infected inhabitants from transmitting DENV to Ae . aegypti in the home . Pyrethroid-treated materials used as curtains dramatically reduced Ae . aegypti populations and reduced DENV transmission in intervention versus control homes in Viet Nam [66]–[68] , the Philippines [69] , and Mexico and Venezuela [70] . A critical question is whether the Ile1 , 016 mutation reduces sensitivity to the repellant effects of pyrethroids . Consequently it is unknown as to whether kdr impacts indoor abundance of dengue virus-infected Ae . aegypti and dengue incidence . It is also unknown whether pyrethroid-treated materials will promote evolution of kdr in Ae . aegypti . It is likely that the current rise in Ile1 , 016 was driven by space spraying of pyrethroids to control adults in and around homes and non-target application of agricultural pyrethroids . We have presented a retrospective study of the prevalence of the Ile1 , 016 mutation in natural populations of Ae . aegypti . It will now be important to begin prospective studies in México . Intensive studies of Ile1 , 016 at single sites may reveal the intensity of selection at these sites . Identification of cities or sites that are moving away from use of permethrin-based insecticides may enable us to explore negative fitness effects associated with the Ile1 , 016 mutation . These are not only academic exercises because as pesticides are applied and the target population becomes resistant , the susceptibility resource is depleted [71] , [72] . A key assumption of IRM is that resistance alleles confer lower fitness in the absence of insecticides . Thus when a specific insecticide is discontinued , resistance will decline , and renew susceptibility . With sufficient time , during which alternative types of insecticides are used , the original insecticide can once again be applied . Resistance surveillance is an essential part of IRM schemes that provide data to inform program planning and pesticide selection , especially by detecting developing resistance at an early stage so that alternatives can be implemented . The results presented here suggest that widespread spatial spraying of permethrin may be rapidly increasing the frequency of the Ile1 , 016 mutation in Ae . aegypti in México . This raises the question of whether permethrin-based insecticides should be replaced with other alternatives to maintain and restore susceptibility to pyrethroids . In addition to potentially improving vector control performance in the short term , this action also could protect the downstream potential for use of emerging vector control products impregnated with pyrethroids such as long-lasting textiles [15] , [16] , [73] . We do recognize the difficulties surrounding operational large-scale changes of insecticide use patterns but hope that our findings will help to inform the debate regarding the critical need to monitor and manage insecticide resistance in order to protect the limited options that are available to combat Ae . aegypti and reduce dengue . | Pyrethroid insecticides prolong the opening of voltage-dependent sodium channels in insect nerves to produce instant paralysis and “knock-down . ” Many insects have evolved knock-down resistance through nonsynonymous mutations that reduce pyrethroid binding in the channels . In 2006 we discovered one such mutation in the arbovirus mosquito vector Aedes aegypti , called Ile1 , 016 , that confers very high knockdown resistance to the pyrethroid insecticide permethrin in mosquitoes homozygous for this mutation . We examined collections of Ae . aegypti from México during 1996–2009 and found that the overall Ile1 , 016 frequency increased from <0 . 1% in 1996–2000 , to 2%–5% in 2003–2006 , to 38 . 3%–88 . 3% in 2007–2009 depending upon collection location . We also demonstrate a strong linear relationship between the frequency of Ile1 , 016 homozygotes and knockdown rate in bioassays and speculate that widespread use of permethrin-based insecticides in México may be impacting the frequency of Ile1 , 016 . Such a rapid increase is predicted by a simple model of positive directional selection acting on a recessive allele . Unfortunately this model also predicts rapid fixation of Ile1 , 016 unless there is a negative fitness associated with Ile1 , 016 in the absence of permethrin and if insecticidal pressure can be reduced . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/viral",
"infections",
"genetics",
"and",
"genomics/gene",
"function",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"genetics",
"and",
"genomics/population",
... | 2009 | Recent Rapid Rise of a Permethrin Knock Down Resistance Allele in Aedes aegypti in México |
Integrative and conjugative elements ( ICEs ) are agents of horizontal gene transfer and have major roles in evolution and acquisition of new traits , including antibiotic resistances . ICEs are found integrated in a host chromosome and can excise and transfer to recipient bacteria via conjugation . Conjugation involves nicking of the ICE origin of transfer ( oriT ) by the ICE–encoded relaxase and transfer of the nicked single strand of ICE DNA . For ICEBs1 of Bacillus subtilis , nicking of oriT by the ICEBs1 relaxase NicK also initiates rolling circle replication . This autonomous replication of ICEBs1 is critical for stability of the excised element in growing cells . We found a conserved and previously uncharacterized ICE gene that is required for conjugation and replication of ICEBs1 . Our results indicate that this gene , helP ( formerly ydcP ) , encodes a helicase processivity factor that enables the host-encoded helicase PcrA to unwind the double-stranded ICEBs1 DNA . HelP was required for both conjugation and replication of ICEBs1 , and HelP and NicK were the only ICEBs1 proteins needed for replication from ICEBs1 oriT . Using chromatin immunoprecipitation , we measured association of HelP , NicK , PcrA , and the host-encoded single-strand DNA binding protein Ssb with ICEBs1 . We found that NicK was required for association of HelP and PcrA with ICEBs1 DNA . HelP was required for association of PcrA and Ssb with ICEBs1 regions distal , but not proximal , to oriT , indicating that PcrA needs HelP to progress beyond nicked oriT and unwind ICEBs1 . In vitro , HelP directly stimulated the helicase activity of the PcrA homologue UvrD . Our findings demonstrate that HelP is a helicase processivity factor needed for efficient unwinding of ICEBs1 for conjugation and replication . Homologues of HelP and PcrA-type helicases are encoded on many known and putative ICEs . We propose that these factors are essential for ICE conjugation , replication , and genetic stability .
Integrative and conjugative elements ( ICEs ) , also known as conjugative transposons , are mobile genetic elements that play a significant role in bacterial evolution and the acquisition of new traits [1] . They contribute significantly to the spread of antibiotic resistances in pathogenic bacteria . ICEs or putative ICEs are found in all major bacterial clades [2] . They reside integrated in a host genome and are propagated along with the host chromosome . Under certain conditions , an ICE can excise from the chromosome , form a double-stranded DNA ( dsDNA ) circle , and transfer to a recipient . Like conjugative plasmids , ICEs encode a multi-component mating pore complex that mediates their transfer from donors to recipients . Most ICEs are thought to transfer linear ssDNA . Transfer is through a type IV secretion system in Gram negative bacteria [3] , [4] , or its counterpart in Gram positive bacteria [5] . ICEs that transfer ssDNA were generally thought to lack the ability to undergo autonomous replication . However , recent work [6]–[8] and findings presented here indicate that autonomous replication is a property of many ICEs and that the mechanisms are conserved . ICEBs1 is approximately 20 kb and normally found integrated in the tRNA gene trnS-leu2 of Bacillus subtilis [9] , [10] . ICEBs1 gene expression and excision can be induced in >90% of cells in a population by overproduction of the activator and cell signaling regulator RapI [9] . Following induction , ICEBs1 undergoes autonomous plasmid-like rolling circle replication [6] . Replication of ICEBs1 is needed for stability of the element after excision [6] . ICEBs1 replication and conjugation both begin with nicking of the ICEBs1 origin of transfer , oriT , by the ICEBs1-encoded relaxase NicK . The nicked DNA is then unwound by the host-encoded helicase PcrA , rather than the replicative helicase ( B . subtilis DnaC ) [6] . During rolling-circle plasmid replication , the free 3′-OH of the nicked strand acts as a primer for replication by the host DNA polymerase , followed by recircularization of and complementary strand synthesis from the unwound single strand . By analogy to other conjugative systems , the single strand of ICEBs1 DNA covalently attached to the relaxase can also be targeted to the mating machinery by the putative coupling protein ConQ and transferred into recipient cells . Although unwinding of ICEBs1 DNA by the PcrA helicase is essential for both replication and conjugation of ICEBs1 , replication of the element in donor cells is not required for its transfer to recipients [6] . PcrA-type helicases ( PcrA from Gram positive bacteria and UvrD and Rep from E . coli ) are required for rolling circle replication of many different plasmids and phages . The PcrA-type proteins are efficient and processive DNA translocases , but typically have poor helicase activity . For many of the characterized phages and plasmids , the element-encoded relaxase that is needed for initiation of replication interacts with the PcrA-type helicase to stimulate DNA unwinding {reviewed in [11] , [12] , [13]} . Unlike these other relaxases , we found that the ICEBs1 relaxase NicK was not sufficient for ICEBs1 replication . In addition to nicK , a second ICEBs1 gene ( helP , previously ydcP ) was necessary for replication from ICEBs1 oriT . Expression of both nicK and helP in B . subtilis was sufficient to support replication from oriT . helP encodes a protein of previously unknown function and is conserved in many ICEs . We found that HelP is required for both mating and replication of ICEBs1 , and that it stimulates the function of the helicase PcrA . We also found that the E . coli helicase UvrD ( a homologue of PcrA ) can substitute for PcrA in B . subtilis , to support both cell viability and ICEBs1 conjugation and replication . Based on in vivo and in vitro analyses , HelP is a helicase processivity factor that is needed for efficient unwinding of ICEBs1 . helP homologues are found in many ICEs , often in a module with genes encoding the relaxase and the putative coupling protein , indicating that these ICEs may also be capable of autonomous replication . PcrA homologues are also found on many extrachromosomal elements , either separately or as a helicase domain attached to the relaxase , indicating that these elements all share a need for DNA unwinding that is met in different ways .
Studies of several rolling-circle plasmids have shown that only one plasmid gene , encoding the plasmid relaxase , is required for replication from the cognate origin of replication {reviewed in [12]} . In contrast , we found that the ICEBs1 relaxase NicK was not sufficient for autonomous replication from the ICEBs1 origin of replication oriT . Instead , a second ICEBs1 gene , helP , was also needed . To determine which ICEBs1 genes are needed for replication from oriT , we constructed a series of plasmids that carry the ICEBs1 oriT along with various candidate ICEBs1 genes . We then tested each plasmid for its ability to replicate in B . subtilis . The parent plasmid , pUS19 [14] , carries a pUC-derived origin of replication that is not functional in B . subtilis , but is functional in E . coli , allowing purification of each test plasmid from E . coli . pUS19 also carries spcE , which allowed us to transform each plasmid into B . subtilis and select for spectinomycin-resistant transformants that stably acquired the plasmid . Transcription of the ICEBs1 genes was driven from the ICEBs1 promoter Pxis that was cloned onto the plasmid . Pxis is derepressed in cells lacking ICEBs1 . After analyzing various plasmids containing different combinations of candidate ICEBs1 genes ( data not shown ) , we found that helP and nicK were sufficient to support replication from ICEBs1 oriT ( Table 1 ) . A plasmid containing oriT , nicK , and helP ( pCAL1255 ) was capable of transforming a B . subtilis strain lacking ICEBs1 ( Table 1 , pCAL1255 ) . The plasmid copy number was between 25–30 ( Table 1 ) as indicated by the amount of spcE DNA ( plasmid ) relative to ydbT , a chromosomal gene adjacent to the ICEBs1 attachment site attB . Replication of pCAL1255 ( oriT , Pxis-helP-nicK ) in B . subtilis was dependent on expression of helP and nicK from Pxis . We repressed expression from Pxis by transforming pCAL1255 into B . subtilis carrying an intact integrated ICEBs1 , which expresses the ICEBs1 repressor ImmR . We were still able to obtain spectinomycin-resistant transformants of pCAL1255 in the ICEBs1-containing cells . However , in these transformants , the plasmid copy number was one , indicating that the plasmid was likely integrated into the chromosomal ICEBs1 by homologous recombination . We found that autonomous plasmid replication from oriT was dependent on both helP and nicK . We did not obtain any spectinomycin-resistant transformants in B . subtilis cells lacking ICEBs1 from plasmids containing nicK without helP ( Table 1 , pJT151 ) or helP without nicK ( Table 1 , pCAL1260 ) . However , we were able to transform these plasmids into strains that expressed the missing gene from a chromosomal locus , showing that the replication defects could be complemented . pJT151 ( oriT , Pxis-nicK ) was able to replicate and had a copy number of 70–80 in a strain expressing helP from the chromosome . pCAL1260 ( oriT , Pxis-helP ) was able to replicate and had a copy number of approximately 4 in a strain expressing nicK from the chromosome ( Table 1 ) . The different copy numbers of pCAL1255 ( oriT , Pxis-helP-nicK ) and the plasmids missing helP or nicK but complemented from a chromosomal copy of the gene may be due to different expression levels of HelP and NicK from the plasmid versus the chromosome , or the effects of the different plasmid sequences on replication efficiency . The low copy number when nicK was expressed from the chromosome is also consistent with the notion that the relaxase functions preferentially in cis [15] . Based on these results , we conclude that , unlike many rolling circle plasmid replicons , the ICEBs1 relaxase NicK is not the only element-encoded protein needed for autonomous replication . ICEBs1-encoded NicK and HelP are both needed to support replication from oriT in the absence of other ICEBs1 products . We found that helP is required for replication of ICEBs1 . We constructed an in-frame markerless deletion of helP ( ΔhelP ) that removed its entire coding sequence from ICEBs1 ( Figure 1B ) . After inducing ICEBs1 gene expression , we measured ICEBs1 copy number by quantitative real time PCR ( qPCR ) . The copy number of ICEBs1 relative to the chromosome is expressed as the relative amount of DNA from ICEBs1 oriT compared to ydbT , as described previously [6] . Consistent with previous findings , the relative copy number of ICEBs1 oriT was 3–4 per cell one hour after induction of ICEBs1 gene expression ( Table 2 , line 1 ) . Under similar conditions , the copy number of the ICEBs1 ΔhelP mutant was approximately 0 . 5 ( Table 2 , line 2 ) . A copy number less than one is consistent with previous findings that replication-defective ICEBs1 is progressively lost from a population of dividing cells [6] . In order to determine the stage at which HelP functions in ICEBs1 replication , we tested whether helP is required for transfer of ICEBs1 into a recipient . Although ICEBs1 replication is not required for mating , the initial steps of nicking and unwinding of ICEBs1 DNA are common to both mating and replication [6] , [15] . Wild type ICEBs1 had a mating efficiency of 3 . 7% ( 3 . 7 transconjugants per 100 donors ) ( Table 2 , line 1 ) , similar to previous results [9] , [16]–[18] . In contrast , transfer of the ICEBs1 ΔhelP mutant was undetectable ( Table 2 , line 2 ) . Expression of helP from a truncated ICEBs1 integrated into the chromosome at thrC ( strain JT335; Figure 1 ) largely restored conjugation , indicating that the primary mating defect was due to loss of helP and not due to polarity on downstream genes ( Table 2 line 3 ) . Although HelP primarily functions in the donor , it also appears to play a role in the recipient . We found that expression of helP in the recipient from the IPTG-inducible promoter Pspank ( hy ) increased the mating efficiency to 1 . 4% , from the 0 . 11% efficiency when helP was provided only in the donor . There was no detectable transfer of the ICEBs1 ΔhelP mutant unless helP was also expressed in the donor , and no increase in mating efficiency of wild type ICEBs1 when helP was also expressed in the recipient . We suspect that HelP is required for autonomous replication and increases the stability of ICEBs1 in recipient cells prior to its integration into the recipient chromosome . Since HelP was required for both conjugation and replication of ICEBs1 , we expected it to be involved in nicking or unwinding of ICEBs1 DNA . There was still nicking of the proper site in oriT in the absence of helP ( data not shown ) , consistent with previous findings that NicK is the only ICEBs1-encoded protein needed for nicking of oriT [15] . Therefore we suspected that HelP was involved in the unwinding of ICEBs1 DNA by the host-encoded helicase PcrA , a DNA translocase that has very limited processivity as a helicase [19] . The solution structure of a HelP homologue , SAG0934 from Streptococcus agalactiae [20] , which is identical to Orf22 of Tn916 from Enterococcus faecalis , indicates that HelP contains an oligonucleotide/oligosaccharide binding fold ( OB-fold ) that is present in many ssDNA binding proteins [21] , consistent with a possible role in binding ICEBs1 ssDNA . Although HelP homologues and some Ssb proteins share the OB-fold , there appear to be no other significant sequence similarities between these proteins . Using crosslinking and immunoprecipitation ( ChIP ) , we found that HelP was associated with ICEBs1 oriT in vivo ( Figure 2 ) . Following induction of ICEBs1 gene expression , protein and DNA were crosslinked with formaldehyde and HelP was immunoprecipitated with polyclonal anti-HelP antibodies . Preliminary experiments with DNA microarrays ( ChIP-chip ) indicated that HelP was strongly associated with the excised and replicating ICEBs1 DNA and that there was little or no specific association with the chromosome ( data not shown ) . We then measured association of HelP with ICEBs1 DNA following ChIP using quantitative real time PCR ( ChIP-PCR ) with primers specific to ICEBs1 oriT and normalized to the amount of DNA from a chromosomal region ( Materials and Methods ) . oriT DNA was enriched >500-fold in the anti-HelP immunoprecipitates ( Figure 2A ) indicating association of HelP with ICEBs1 oriT in vivo . The immunoprecipitation was specific to HelP as ICEBs1 oriT was not significantly enriched in a strain deleted for helP ( 2 . 4-fold relative association in ΔhelP mutant compared to >500-fold in wild type ) . We found that in vivo , association of HelP with ICEBs1 oriT required the activity of the relaxase NicK . In a strain containing a nicK deletion , and a functional oriT [15] , association of HelP with ICEBs1 was abolished ( Figure 2A ) . To test if NicK nicking activity was required for association of HelP at oriT , we made a mutation in the catalytic site of NicK that changes the conserved tyrosine at position 195 to phenylalanine ( nicKY195F ) . We also incorporated a C-terminal 3× c-Myc tag onto the wild type and Y195F mutant NicK proteins to use in ChIP experiments with monoclonal antibodies to c-Myc . Both NicKY195F-Myc NicK-Myc were associated with ICEBs1 oriT as determined by ChIP-PCR experiments ( 40±2-fold enrichment and 44±4-fold enrichment , respectively ) . As expected , there was no detectable nicking of oriT by NicKY195F-Myc ( <0 . 05%±0 . 03% nicked oriT ) whereas there was normal nicking of oriT by NicK-Myc ( 32%±3 . 7% nicked oriT ) , as determined by primer extension assays [15] . The NicK-Myc was also functional in conjugation ( Materials and Methods ) . We measured the association of HelP with ICEBs1 oriT in vivo in strains expressing either NicK-Myc or NicKY195F-Myc . As expected , HelP was associated with ICEBs1 oriT in the strain expressing functional NicK-Myc ( data not shown ) . In contrast , HelP was not detectably associated with ICEBs1 oriT in the strain expressing NicKY195F-Myc ( Figure 2A ) . These results indicate that the presence of NicK at oriT is not sufficient and that nicking of oriT is required for recruitment of HelP to oriT . We also found that HelP was associated with the rolling circle replicating plasmid pBS42 ( data not shown ) , consistent with the notion that HelP does not require specific interaction with NicK . The host-encoded helicase PcrA is required for both conjugation and replication of ICEBs1 [6] . Since HelP is also required for conjugation and replication of ICEBs1 , we tested for effects of HelP on PcrA . We measured association of PcrA with ICEBs1 oriT by ChIP-PCR using polyclonal anti-PcrA antibodies and primers specific to oriT as described above . PcrA association with oriT was not significantly affected in a ΔhelP mutant , compared to wild type ( Figure 2B ) . We conclude that association of PcrA with ICEBs1 oriT is not dependent on HelP . In contrast , we found that association of PcrA with ICEBs1 oriT appeared to be largely , but not entirely , dependent on nicking of oriT by NicK . Enrichment of oriT in the PcrA immunoprecipitates was reduced but not abolished in the nicKY195F-myc mutant ( approximately 6-fold enrichment in nicKY195F-myc vs 65–70-fold enrichment with nicK-myc ) . Association of PcrA with ICEBs1 DNA in the nicK catalytic site mutant indicates that there might be interactions between ICEBs1 NicK and the host PcrA , as there are with other replicative rolling circle relaxases and PcrA [22] , [23] . The immunoprecipitation was specific for PcrA; in a pcrA recF double mutant [6] , ICEBs1 sequences were not significantly enriched ( approximately 1 . 1-fold relative enrichment of oriT compared to 12 . 5-fold enrichment in the recF parent ) . ( Loss of recF suppresses the lethality caused by loss of pcrA [24] , [25] ) . Unwinding and replication of ICEBs1 DNA by PcrA proceeds unidirectionally from the nicked oriT [6] . Replication of ICEBs1 is also accompanied by association of the host-encoded single stranded DNA binding protein ( Ssb ) to the ICEBs1 DNA [6] . To examine the location and role of HelP during unwinding and replication of ICEBs1 DNA from the nicked oriT , we compared association of HelP , PcrA and Ssb with an oriT-proximal ( Figure 2A–2C ) to an oriT-distal region ( Figure 2D–2E ) . We found that although HelP was not needed for the initial recruitment of PcrA to oriT ( Figure 2B ) , it was required for the association of PcrA with oriT-distal regions ( Figure 2E ) . Association of HelP and PcrA with the oriT-distal ( conE ) region of ICEBs1 DNA was readily detectable in wild type ( helP+ ) cells ( Figure 2D , 2E ) . In contrast , association of PcrA with ICEBs1 DNA in the ΔhelP mutant was greatly reduced in this region ( Figure 2E ) . This reduction was due to loss of helP and not an unexpected secondary effect due to polarity or alterations in the DNA because association was restored when helP was expressed in trans from an ectopic locus ( Figure 2E ) . Together , our results indicate that HelP is not needed for the initial association of PcrA with ICEBs1 oriT , but that HelP is needed for PcrA to become associated with distal regions , perhaps by affecting helicase processivity . We monitored association of Ssb with ICEBs1 DNA ( indicative of unwound ssDNA ) using an Ssb-GFP fusion and ChIP-PCR with anti-GFP antibodies , essentially as described previously [6] . We found that Ssb-GFP was associated with both the oriT-proximal and oriT-distal ( conE ) region of ICEBs1 ( Figure 2C , 2E ) . In contrast , there was little or no association of Ssb-GFP with the oriT-distal region in the helP mutant ( Figure 2E ) , and association was reduced , but still appreciable , in the oriT region ( Figure 2C ) . Association of Ssb with ICEBs1 DNA was restored by expression of helP from an ectopic locus ( Figure 2C , 2E ) , indicating that the defect in the helP mutant was due to loss of helP and not some unexpected effect . Based on these results , we conclude that HelP is required for both PcrA and Ssb to associate with oriT-distal sequences in ICEBs1 , and that HelP is likely needed for the processive unwinding of ICEBs1 DNA . This function would be sufficient to explain the requirement for helP in both conjugation and replication of ICEBs1 . We wished to test directly the ability of HelP to facilitate unwinding of duplex DNA by PcrA . Our many different preparations of B . subtilis PcrA were of low concentration and rapidly lost helicase activity . Since most structural and biochemical analyses of PcrA have been done with a homologue from another organism {reviewed in [26]} , we decided to test for effects of HelP on UvrD from E . coli . UvrD is a well-studied homologue of PcrA and has 41% identity with B . subtilis PcrA . Like PcrA , UvrD is required for replication of several rolling circle replicating plasmids [27] . We found that expression of E . coli uvrD from the IPTG-inducible promoter Pspank ( Pspank-uvrD ) in B . subtilis suppressed the lethality caused by loss of pcrA . This suppression occurred in both the absence and the presence of IPTG , indicating that , in the absence of induction , expression from Pspank was leaky and sufficient levels of UvrD were produced . In addition , UvrD was able to support replication and conjugation of ICEBs1 nearly as well as PcrA ( Table 3 ) . In cells missing pcrA but expressing uvrD , the mating efficiency was approximately 1% ( transconjugants per donor ) and ICEBs1 was capable of replication and had a copy number of 2–3 ( Table 3 line 3 ) . Based on these results , we conclude that E . coli uvrD can replace pcrA in B . subtilis and provides the functions of PcrA needed for cell viability and those needed for ICEBs1 conjugation and replication . We found that HelP stimulated the ability of UvrD to unwind a partial duplex DNA substrate in vitro . We purified hexahistidine-tagged UvrD ( his-UvrD ) and HelP ( his-HelP ) and used two different partial duplex DNA substrates to monitor unwinding ( Materials and Methods ) . The substrates had either a small ( 22 nucleotides ) or large ( 81 nucleotides ) fluorophore-labeled oligonucleotide hybridized to single-stranded M13mp8 DNA . Helicase activity was measured by release of the labeled oligonucleotide , detected by gel electrophoresis and fluorometry . HelP alone had no effect on the partial duplex substrate . In both cases , <0 . 5% of the substrate duplex was unwound after 45 minutes . UvrD alone was able to unwind the 22 bp duplex: approximately 20% of the substrate was unwound after 45 minutes ( Figure 3A ) . The addition of HelP increased the amount of substrate that was unwound to approximately 40% ( Figure 3A ) . In contrast , there was little or no release of the large ( 81-mer ) oligonucleotide from the partial duplex by UvrD alone ( <2% unwound after 45 minutes ) . In the presence of HelP and UvrD , >25% of the 81 bp duplex substrate was unwound by 45 minutes ( Figure 3B ) . Based on the increase in unwinding and the difference in stimulation between the 22 bp and 81 bp duplex substrates , we conclude that HelP stimulates the ability of UvrD , and likely PcrA , to processively unwind duplex DNA . In addition , since there is no relaxase in this assay , relaxase , and specifically NicK , is not required , at least in vitro , for the function of HelP . The well characterized ICE Tn916 encodes two HelP homologues , Orf22 and Orf23 . Comparison of HelP to each of these indicates that HelP is more similar to each protein than they are to each other ( Figure 4 ) . In addition , Orf23 is about 20 amino acids shorter at the C-terminus than Orf22 and HelP . The gene organization in ICEBs1 near helP and in Tn916 near orf22 and orf23 is similar ( Figure 5 ) . helP and its homologues are usually grouped with two other genes: 1 ) a relaxase gene , nicK in ICEBs1 and orf20 in Tn916 , and 2 ) a gene encoding the predicted coupling protein , conQ in ICEBs1 [18] and orf21 in Tn916 ( Figure 5 ) . The coupling protein targets the relaxosome complex that is linked to ssDNA to the mating pore [28] . This genetic arrangement reflects a functional relationship as the relaxase and likely HelP are part of the relaxosome that interacts with the cognate coupling protein . When present , the majority of helP homologues are found in pairs , although ICEBs1 has only one helP . helP homologues are found in at least 72 ICEs or putative ICEs , primarily in firmicutes [29] . Phylogenic analysis revealed that these HelP homologues fall into seven clades ( Figure 6 ) . When an ICE encodes two HelP homologues , each one is in a separate clade: one clade contains the longer HelP homologue and other clade contains the shorter HelP homologue . For example the clade labeled “Tn916 Orf22” contains most of the longer HelP homologues and the clade labeled “Tn916 Orf23” contains most of the shorter HelP homologues . ICEBs1 HelP is similar in size to the longer Orf22 , but appears to be almost equidistant from the Orf22 and Orf23 clades ( Figure 6 ) . Together , our results indicate that HelP proteins are encoded by many different ICEs from Firmicutes . If the function of these homologues is conserved , then HelP proteins act as helicase processivity factors for many ICEs and these ICEs likely undergo autonomous rolling circle replication .
PcrA and its well-characterized homologues in E . coli , UvrD and Rep , are members of superfamily 1 ( SF1 ) of non-hexameric helicases {reviewed in [13] , [26]} . These proteins are highly processive 3′-5′ ssDNA translocases that bind to and move along strands of ssDNA . They are also weak 3′-5′ helicases that bind and destabilize dsDNA . They are involved in multiple cellular processes , including several types of DNA repair , replication restart , and clearing recombination proteins from DNA {reviewed in [26] , [30]} . Strains lacking SF1 helicases are usually nonviable [24] , [31] , presumably due to loss of the DNA translocase activity of these proteins [31]–[33] . Although PcrA , Rep and UvrD are efficient and processive DNA translocases , they have very poor helicase activity on their own [19] , [34] , [35] . Helicase activity requires oligomerization [36]–[38] , and these helicases interact with accessory factors that stimulate activity . For example , the DNA mismatch repair protein MutL facilitates loading and processivity of UvrD in E . coli [39] and the double-strand break repair protein Ku interacts with and stimulates processivity of UvrD1 in Mycobacterium smegmatis [40] . PcrA processivity is stimulated in vitro by YxaL from B . subtilis [41] , although the role of YxaL in vivo is not known . Although they are poorly processive , PcrA , UvrD , and Rep are used by many plasmids and phages that undergo rolling-circle replication . In E . coli , UvrD is needed for rolling circle replication of some plasmids [27] , [42] and Rep is used for replication of several ssDNA phages {reviewed in [11]} . In some Gram positive bacteria , PcrA is used for replication of many different rolling circle plasmids [12] , [31] , [32] . In addition , PcrA is required for unwinding of ICEBs1 DNA for both conjugation and replication [6] . For rolling circle replicating elements that use a host-encoded SF1-type helicase , efficient unwinding of duplex DNA is often stimulated by interaction with the element-encoded replication relaxase . Relaxases introduce a nick into dsDNA and mark the site for recruitment of the helicase for unwinding and replication of the plasmid or phage DNA . Replicative relaxases of ssDNA phages from E . coli , including gpA ( cisA ) from øX174 and the product of gene II of the f1 family of phages , increase Rep helicase processivity {reviewed in [11]} . The replicative relaxases of rolling circle replicating plasmids pT181 and the related pC221 , RepC and RepD , respectively , appear to interact with and stimulate PcrA [23] , [43]–[47] . In the case of RepD , this stimulation is thought to occur by increasing the affinity of PcrA for its DNA substrate and potentially by decreasing its rate of dissociation , rather than by altering its kinetic properties [46] , [48] . As far as we are aware , all known examples of plasmid or phage-encoded helicase-stimulating proteins are relaxases . By analogy to the related plasmid relaxases , we expected that the only ICEBs1 product needed for ICEBs1 rolling circle replication would be its relaxase NicK . However , we found that the ICEBs1 gene product HelP is a helicase processivity factor that is required , in addition to NicK , for replication and conjugation of ICEBs1 . We do not yet know how HelP stimulates helicase activity , but it appears to act at a step after association of PcrA with ICEBs1 oriT DNA . HelP could stimulate dimerization of PcrA since it is the dimer and not the monomer that has helicase activity [19] . HelP could also decrease the rate of dissociation of PcrA from DNA during unwinding , analogous to the activity demonstrated for RepD [46] , [48] . HelP could accomplish one or a combination of these stimulatory functions by direct protein-protein contact with PcrA . Alternatively , HelP could remodel the DNA substrate during unwinding , promoting helicase activity . helP homologues are found primarily in firmicutes . There appear to be >300 homologues in the non-redundant protein database , and at least 128 of these are found in known or putative ICEs that are included in the ICEberg database [29] , and we suspect that most of the others are in unrecognized ICEs . Although ICEBs1 has only one helP , most , but not all , helP homologues are found in pairs , with one member of the pair approximately 20 amino acids longer than the other . It is possible that each member of a HelP pair stimulates processivity of a dissimilar range of helicases , thereby broadening the host range of the mobile element . If the function of HelP homologues is conserved , then the ICEs with helP homologues likely undergo autonomous replication and that HelP proteins are important for stability and transfer of many ICEs . In contrast to the many ICEs and plasmids that utilize a host-encoded SF1-type helicase for replication and transfer and rely on an accessory factor to facilitate unwinding , some ICEs and plasmids encode their own SF1-type helicase . PcrA-like helicase domains are found attached to some plasmid relaxases . For example , TraI of the E . coli F plasmid and TrwC of R388 , have a C-terminal PcrA-like helicase domain that is required for conjugation [49]–[51] . The helicase activity is highly processive , potentially because it is tethered to the relaxosome {reviewed in [52]} . In addition , many known and putative ICEs encode discrete PcrA-like helicases [29] . It is not known if they are processive or require accessory proteins , although some of those ICEs also contain a helP homologue . An element that encodes its own helicase might have a broader host range by avoiding reliance on a host-encoded helicase . That HelP homologues , relaxases that also function as processivity factors , and PcrA homologues are encoded on conjugative elements indicates that the need for DNA unwinding can be met in different ways .
All B . subtilis strains ( Table 4 ) are derivatives of JH642 ( trpC2 pheA1 ) and were constructed by natural transformation . Strains are either cured of ICEBs1 ( designated ICEBs10 ) or contain ICEBs1 marked with the Δ ( rapI-phrI ) 342::kan allele to monitor conjugative transfer [9] . ICEBs1 gene expression was induced by overproduction of rapI from amyE::{ ( Pspank ( hy ) -rapI ) spc} , amyE::{ ( Pxyl-rapI ) spc} , or lacA::{ ( Pxyl-rapI ) tet} . Strains used as recipients in mating experiments contained the streptomycin resistance allele ( str-84 ) [9] . Other mutations that were previously described include: the deletion-insertion ΔpcrA1021::cat [6]; thrC1165::cat [6] , used as a control for threonine auxotrophy and chloramphenicol resistance; and thrC329::{ ( Pxis-nicK ) mls} [15] , used for ectopic expression of nicK . Plasmids pCAL1255 ( oriT , Pxis-helP-nicK ) , pCAL1260 ( oriT , Pxis-helP ) , and pJT151 ( oriT , Pxis-nicK ) contain DNA segments from ICEBs1 inserted into the BamHI restriction site of pUS19 , a spectinomycin-resistant derivative of pUC19 [14] . In addition to conferring resistance to spectinomycin in B . subtilis , each plasmid contains ICEBs1 oriT . pCAL1255 ( oriT , Pxis-helP-nicK ) contains the xis promoter ( Pxis ) driving transcription of helP and nicK . The Pxis-helP-nicK insertion is comprised of three non-contiguous segments of ICEBs1: 1 ) a 527 bp segment containing 254 bp of the 5′-end of immR and the entire 273 bp intergenic region between immR and xis; 2 ) a 419 bp segment that contains helP and extends 35 bp downstream of the helP stop codon; and 3 ) a 1076 bp segment containing nicK ( and oriT ) and extending 17 bp downstream of the nicK stop codon . The junctions between the DNA segments were designed to allow translation initiation of helP from the xis ribosome binding site and translation initiation of nicK from the conQ ribosome binding site , as the start codons of helP and of nicK were placed the same distance downstream of the Pxis and the helP-conQ intergenic regions as the native start codons of xis and conQ , respectively . Transcription of the Pxis-helP-nicK insertion is in the same orientation as the spectinomycin- and ampicillin-resistance genes on the vector backbone . pCAL1260 ( oriT , Pxis-helP ) contains the same Pxis-helP-nicK insertion present in pCAL1255 , but with an additional T inserted between the 5th and 6th codon of nicK . The single base insertion in pCAL1260 leads to premature termination after the 10th codon of the nicK ORF . pJT151 is essentially pCAL1255 ( oriT , Pxis-nicK ) but with the entire helP coding sequence deleted . N-terminal hexahistidine tagged HelP ( his-HelP ) and UvrD ( his-UvrD ) were overproduced in and purified from E . coli . helP was amplified by PCR from B . subtilis ( strain JH642 ) chromosomal DNA , digested with BamHI and HindIII and ligated into the same sites of the T7-expression vector pET28b ( Novagen ) to give plasmid pCAL1297 . uvrD was amplified from E . coli strain MC1061and cloned , by isothermal assembly , into the XbaI and BamHI sites of pET28b to give plasmid pJT370 . For conjugation experiments , all donor strains contained ICEBs1 with the Δ ( rapI-phrI ) 342::kan allele and ICEBs1 gene expression was induced by overproduction of the regulatory protein RapI from an ectopic locus [9] . When donors carried the Pspank ( hy ) -rapI allele , all strains were grown in rich medium and induction was for one hour with 1 mM IPTG . When the Pxyl-rapI was used , strains were grown in minimal S750 medium containing 1% arabinose and induction was for 2 hours after the addition of 1% xylose [56] . Recipients were streptomycin resistant ( str-84 ) . Donors and recipients were mixed in a 1∶1 ratio , filtered onto nitrocellulose membranes and incubated on agar containing minimal salts for 3 hours , essentially as described [9] . The mating efficiency was determined as the number of colony forming units ( CFU ) of streptomycin- and kanamycin- resistant transconjugants per CFU of the donor . Association of HelP , PcrA , NicK-Myc and Ssb-GFP with ICEBs1 DNA was measured by chromatin immunoprecipitation followed by quantitative PCR ( ChIP-qPCR ) essentially as described [6] , [57] . Polyclonal antibodies from rabbits ( Covance ) were used to precipitate HelP , PcrA and Ssb-GFP and monoclonal antibodies to c-Myc ( Invitrogen ) were used to precipitate NicK-Myc . qPCR was used to measure the relative amount of DNA in immunoprecipitates and to measure relative plasmid copy number . Primers to the ICEBs1 nicK/oriT region , designated oriT , were CLO280 , 5′-TGGCTACGTT GGCACGTATG-3′ , and CLO281 , 5′-AATTGACGGC AACCTTGACC-3′ . Primers to ICEBs1 conE , approximately 6 kb downstream from oriT , were oMMB238 , 5′-TGATGGTTCAAATCCTGCATTGTCAC-3′ , and oMMB239 , 5′-GAACTTACCT AGTGCAAACATGAC-3′ . Plasmid copy number was determined using primers oJT168 , 5′-GTGGAATCAT CCTCCCAAAC-3′ , and oJT169 , 5′-AATGGCTCTT CTCACATCAG-3′ , that are specific to spcE found on the plasmids and not in the chromosome . Values obtained in ChIP-PCR and plasmid copy number experiments were normalized to the chromosomal locus ydbT , approximately 15 kb upstream of the chromosomal attachment site for ICEBs1 , with primers CLO284 , 5′-CTTCCGCACA TGCTCCGAAC-3′ and CLO285 , 5′-TCGGCAGCAG GATCACTGAC-3′ . his-HelP and his-UvrD were purified using similar conditions . Expression strains were grown in 2 liters of 2×YT medium supplemented with 0 . 4% glycerol , 20 mM sodium phosphate buffer pH 7 . 0 , and 10 mM MgSO4 at 30°C until they reached an OD600 of 0 . 8 . Protein expression was induced by the addition of 0 . 2% arabinose and 1 mM IPTG followed by incubation for 3 hours . Cells were harvested by centrifugation , washed with 100 ml of phosphate-buffered saline , resuspended in 25 ml lysis buffer {50 mM Tris-HCl ph 8 . 0 , 0 . 3 M NaCl , 10 mM imidazole , 1 mg/ml lysozyme , 5 µg/ml DNase I and 1× CellLytic Express ( Sigma-Aldrich ) } and lysed by incubation at room temperature with gentle inversions . The lysate was cleared by centrifugation at 15 , 000× g and the cleared lysate was applied to 1 ml Ni-NTA agarose resin , washed with Tris-NaCl ( 50 mM Tris-HCl pH 8 . 0 , 0 . 3 M NaCl ) containing 10 and 20 mM imidazole and eluted with the same buffer containing 250 mM imidazole . The eluted protein was dialyzed against 10 mM MOPS , pH 7 . 5 , 200 mM NaCl and 1 mM tris ( 2-carboxyethyl ) phosphine . his-UvrD was approximately 93% pure and his-HelP was approximately 97% pure . Helicase activity was measured using two different partial-duplex DNA substrates , designed by analogy to previously described templates [41] . The substrates were generated by annealing M13mp18 circular ssDNA ( Affymetrix ) with 1 ) a 22-bp oligonucleotide ( oJT276 , 5′-ACTCTAGAGGA TCCCCGGGTAC-3′ ) , or 2 ) an 81-bp oligonucleotide ( oJT278 , 5′-GGCCAGTGCCA AGCTTGCATG CCTGCAGGTC GACTCTAGAG GATCCCCGGG TACCGAGCTC GAATTCGTAA TCATGGTCAT-3′ ) . Each oligonucleotide was labeled on the 5′-end with an infrared fluorescent dye , IRDye800 for oJT276 and TYE705 for oJT278 ( IDT ) . Helicase assays were performed in 200 µl reactions at 37°C . Mixtures containing 0 . 5 nM DNA substrate and 0 or 1 µM his-HelP , in 20 mM MOPS , pH 7 . 5 , 200 mM NaCl , 15 mM MgCl2 , 2 . 5 mM ATP , 1 mM tris ( 2-carboxyethyl ) phosphine were preincubated at 37°C for 10 min , and started by the addition of 25 nM his-UvrD . 20 µl aliquots were withdrawn at regular intervals and added to 5 µl stop buffer ( 5% Ficoll , 15% glycerol , 0 . 12% Orange G , 1% SDS and 50 mM sodium EDTA ) . Samples ( 10 µl ) were analyzed by electrophoresis on a 15% TBE-polyacrylamide gel containing 2 . 5% glycerol followed by visualization and quantitation using the Odyssey Infrared imaging system ( Li-Cor ) . We identified 128 homologues of ICEBs1 helP ( including helP ) in 72 bacterial ICEs using the HMMER3 and WU-BLAST2 search tools in ICEberg ( http://db-mml . sjtu . edu . cn/ICEberg/ ) , the web-based resource for bacterial ICEs . The ICEberg database ( last updated on August 14 , 2011 ) contains sequence information for 431 known and putative ICEs . All 130 homologues were aligned using ClustalX 2 . 1 [58] using multiple-alignment mode with default parameters . The neighbor-joining method was used to generate the phylogenic tree from the ClustalX PHYLIP output file at the Interactive Tree of Life ( http://itol . embl . de/index . shtml ) [59] . Seven clades were defined by grouping together homologues that had an average distance of <0 . 22 . An additional HMMER search [60] with default search settings ( hmmer . janelia . org ) identified >300 HelP homologues in the non-redundant protein database . All but five of these were in Firmicutes . The exceptions were one homologue in a Mycoplasma , Ureaplasma urealyticum , two in an uncultured bacterium MID12 , and two homologues in Klebsiella pneumoniae which carries Tn916 ( usually associated with Firmicutes ) on the composite ICE Tn6009 . Global sequence similarities between HelP and the two HelP homologues from Tn916 were analyzed by the Needleman-Wunsch alignment method [61] . | Integrative and conjugative elements ( ICEs ) are mobile DNA elements that transfer genetic material between bacteria , driving bacterial evolution and the acquisition of new traits , including the spread of antibiotic resistances . ICEs typically reside integrated in a bacterial chromosome and are passively propagated along with the host genome . Under some conditions , an ICE can excise from the chromosome to form a circle and , if appropriate recipient bacteria are present , can transfer from donor to recipient . It has recently been recognized that some , and perhaps many , ICEs undergo autonomous replication after excision from the host chromosome and that replication is important for stability and propagation of these ICEs in growing cells . Using ICEBs1 , an ICE from Bacillus subtilis , we found a conserved and previously uncharacterized ICE gene that is required for conjugation and replication . We found that this gene , helP , encodes a helicase processivity factor that associates with ICEBs1 DNA and enables the host-encoded helicase PcrA to unwind the double-stranded ICEBs1 DNA , making a template for both conjugation and DNA replication . Homologues of helP are found in many ICEs , indicating that this mechanism of unwinding is likely conserved among these elements . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology"
] | 2013 | A Conserved Helicase Processivity Factor Is Needed for Conjugation and Replication of an Integrative and Conjugative Element |
Chikungunya virus ( CHIKV ) is the most common alphavirus infecting humans worldwide , causing acute and chronically debilitating arthralgia at a great economic expense . To facilitate our study of CHIKV , we generated a mCherry tagged replication-competent chimeric virus , CHIKV 37997-mCherry . Single particle cryoEM demonstrated icosahedral organization of the chimeric virus and the display of mCherry proteins on virus surface . CHIKV 37997-mCherry is attenuated in both IFNαR knockout and wild-type mice . Strong anti-CHIKV and anti-mCherry antibody responses were induced in CHIKV 37997-mCherry infected mice . Our work suggests that chimeric alphaviruses displaying foreign antigen can serve as vaccines against both aphaviruses and other pathogens and diseases .
Chikungunya virus ( CHIKV ) is a mosquito-transmitted , enveloped positive stranded RNA virus that belongs to the Alphavirus genus of the Togaviridae family . CHIKV infection causes an acute febrile illness typically accompanied by severe arthralgia , with relapses for weeks to months [1] . In the past decade , CHIKV has spread from endemic areas of Africa and Asia to new populations in Europe and the Americas , making CHIKV a global threat and the most common alphavirus infecting humans . Millions of individuals were infected during the 2000s , resulting in thousands of deaths [2] . The CHIKV RNA genome encodes four nonstructural ( nsP1 to nsP4 ) and five structural ( C , E3 , E2 , 6K and E1 ) proteins . All five structural proteins are translated as a single polyprotein from which the capsid ( C ) protein is released via self-cleavage . The envelope polyprotein precursor ( E3-E2-6K-E1 ) is translocated into the endoplasmic reticulum ( ER ) and processed by host signalase into E1 , 6K and E3E2 polyprotein ( p62 ) . In the ER , E1 and p62 assemble into heterodimers and subsequently trimerize to form viral spikes . p62 is cleaved by host furin or furin-like proteases into E3 and E2 during trafficking through the Golgi and trans-Golgi network ( TGN ) . The mature CHIKV particles bud at the plasma membrane and have T = 4 quasi-icosahedral symmetry with 240 copies of the E1E2 heterodimer , assembled into 80 spikes on the viral surface; 240 copies of C form an icosahedral nucleocapsid core enclosing viral genomic RNA within the lipid membrane [3] . E1 is a type II membrane fusion protein and sits at the base of the trimeric spike with E2 positioned on top of E1 . The ectodomain of E1 consists of three domains [4] . Domain I links distal domain II and membrane proximal domain III . A fusion loop is situated at the distal end of E1 domain II , and is protected by domain B of E2 , located at the distal end of E2 [4 , 5] . After viral entry into target cells , the acidic environment of endosomes triggers conformational rearrangements within E2 , leading to domain B dissociating from the fusion loop [6] . E1 then forms a homotrimer , further exposing the fusion loops of each monomer at the end of the trimeric complex for insertion into the host membrane [7] . E3 limits the movement of E2 domain B until it is cleaved from E2 by furin or furin-like proteases to prevent accidental activation of the fusion peptide before the assembly and budding of mature virions [5 , 8] . Studies show alphaviruses can tolerate insertion of foreign antigens to N-terminal of E2 generating chimeric virus presenting foreign antigens [9–11] . There are currently no licensed vaccines or treatments for CHIKV infection . We hypothesize that an attenuated CHIKV carrying foreign antigens may induce immune responses against both CHIKV and the foreign antigen . In the current study , we report a replication-competent CHIKV with mCherry fused to the N-terminus of E2 . The resulting chimeric virus presents mCherry on the virion surface in a repetitive pattern and induces strong antibody responses against both CHIKV and mCherry . The virus is attenuated in vivo and provides a safe research tool to study CHIKV virology , as well as demonstrating a useful vaccine platform .
Vero and BHK21 cells ( ATCC CRL-1586 and CCL-10 ) were maintained at 37°C in a fully humidified atmosphere with 5% CO2 in DMEM ( Invitrogen ) medium supplemented with penicillin and streptomycin , 10 mM HEPES , non-essential amino acids , and 10% FBS ( Hyclone ) . The CHIKV 37997 and La Reunion ( LR ) strain virus clone was gift of S . Higgs ( Kansas State University ) and the CHIKV vaccine strain 181/clone 25 was gift of S . Weaver ( University of Texas Medical Branch ) . Viruses were produced from infectious cDNA clones as previously described [12 , 13] . The mCherry gene was inserted between E3 and E2 after the furin cleavage site in CHIKV 37997 genome to make CHIKV 37993-mCherry cDNA clone [14] . mCherry protein fused with 6xHis-tag at the N-terminus was expressed in bacterial cells and affinity purified through Cobalt resins ( ThermoFisher Scientific ) . Vero cells were infected with CHIKV 37993-mCherry at a multiplicity of infection ( MOI ) of 5 . At 8 h post infection , the cells were fixed and imaged with a Nikon Ti inverted fluorescence microscope equipped with a Yokogawa CSU-22 spinning disk confocal with a 561 nm laser . Purified viruses were subjected to electrophoresis with 4–12% sodium dodecyl sulfate-polyacrylamide gel ( ThermoFisher Scientific ) followed by Coomassie staining with SimpleBlue SafeStain ( ThermoFisher Scientific ) or Western blot analysis with rabbit polyclonal anti-CHIKV 181/25 ( IBT Bioservices ) and mouse monoclonal anti-mCherry ( Sigma ) . C57BL/6J mice were purchased from The Jackson Laboratory ( Sacramento , CA ) . Ifnar-/- mice [15] , on a C57BL/6J background , were a gift from Dr . Michael Diamond ( Washington University ) and bred in house under specific-pathogen-free conditions in microisolator cages ( Innovive Inc . , San Diego , CA ) . To study viral pathogenesis in wild type mice , three week-old C57BL/6J mice were inoculated with CHIKV subcutaneously in the left footpad with 103 or 106 PFU of CHIKV in PBS supplemented with 1% heat inactivated FBS . Joint swelling was monitored via left and right foot measurements at the peritarsal region ( width x height ) using digital calipers . Sera were collected at day 3 after infection and rear ankles were collected at day 7 after infection . To study viral pathogenesis in a lethal mouse model , 6–8 week-old Ifnar-/- mice were inoculated with CHIKV subcutaneously in the left footpad with 10 , 100 or 1 , 000 PFU of CHIKV in PBS supplemented with 1% heat inactivated FBS . Joint swelling was monitored via left and right foot measurements at the peritarsal region ( width x height ) using digital calipers , and mouse survival was monitored for at least 3 weeks . Immunization and challenge with CHIKV were performed by injection of virus subcutaneously in the left footpad . To immunize mice with mCherry , 8 week-old C57BL/6J mice were administered with purified mCherry protein via intraperitoneal injection . Mice were prime immunized with 100 μg of mCherry protein mixed with complete Freund’s adjuvant ( Sigma-Aldrich ) followed by boosting with 50 μg of mCherry proteins mixed with incomplete Freund’s adjuvant ( Sigma-Aldrich ) 2 weeks later . All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee at PMI Preclinical , LLC ( San Carlos , CA ) , protocol number IAC 1705 . Mice received humane care according to the criteria outlined by the National Research Council’s Institute of Laboratory Animal Resources in the “Guide for the Care and Use of Laboratory Animals” . All injections of mice with virus were performed under anesthesia with isoflurane or ketamine and xyalzine . Viral RNA in culture supernatant or mouse blood was extracted by the QIAamp Viral RNA Mini kit ( Qiagen ) following the manufacturer’s protocol . To extract viral RNA in mouse tissues , tissues were first lysed in TRIzol ( ThermoFisher Scientific ) using Qiagen Tissue Lyser II ( Qiagen ) . RNA was extracted following the manufacturer’s protocol . Isolated RNA was analyzed by qRT-PCR and compared to a standard curve generated from CHIKV-GLuc plasmid [16] with the following primers: CKV_For , 5’-TGGCCACCTTTGCAAGCTC-3’; CKV_Rev , 5’-GGGATGAACTCCATTGTAGC-3’; and CKV_Probe , 5’/56-FAM/AGGTACGCACTACAGCTACC/36-TAMSp/3’ . Anti-CHIKV and anti-mCherry antibodies in mouse sera were quantified using virion-based ELISA . To detect anti-CHIKV antibodies , Immulon 4HBX plates ( 96-well , Thermo Scientific ) were coated with CHIKV 181/25 virus purified through sucrose cushion ( 2 . 5 x 108 genome copies/well ) . To detect anti-mCherry antibodies , Immulon 4HBX plates were coated with purified 6xHis-mCherry protein ( 1 μg/well ) . Serial dilutions of mouse serum were added to the plates . Serum of naïve mice was used as the background control . Next , plate-bound antibodies were detected with biotin-conjugated goat anti-mouse IgG antibodies ( Southern Biotech ) , followed with streptavidin-conjugated horseradish peroxidase ( Southern Biotech ) . Binding was detected with 3 , 3’5 , 5’-tetramethylbenzidine substrate ( Neogen ) . Endpoint titers were defined as the reciprocal of the last dilution to have an absorbance two times greater than the background control . Neutralizing activity of mouse serum was quantified using a plaque reduction neutralization test ( PRNT ) . Serial dilutions of mouse serum were incubated with 50 PFUs of challenge virus for 1 hour at 37°C , followed by infection of Vero cells for 1 hour at 37°C . To quantify virus by plaque assay , Vero cells were incubated with serial dilutions of virus containing supernatant for 1 hour at 37°C . Next , cells were overlaid with medium containing 2% FBS and 0 . 8% agarose , followed by culture at 37°C . Plaques were counted 2 days later . To coat ELISA plates , CHIKV 181/25 was produced in BHK21 cells virus and was pelleted through a 20% sucrose cushion . To make gradient-purified viruses for single particle cryoEM study , we adapted a published protocol [17] . CHIKV 37997 and CHIKV 37997-mCherry viruses were produced in BHK21 cells and first concentrated by pelleting the culture supernatant through a 20% sucrose cushion . Then , the resuspended virus was layered onto 2 ml of 60% OptiPrep ( Sigma-Aldrich ) and spun at 50 , 000 x g for 1 . 5 hour using a SW28 rotor . After ultracentrifugation , the supernatants were removed to leave 4 ml above the virus band . The remaining 4 ml supernatant , the virus band and the underlay of 2 ml of 60% OptiPrep were mixed to reach a final concentration of 20% OptiPrep . The mixture was spun at 360 , 000 x g for 3 . 5 hours with a NVT65 . 2 rotor . The virus band was extracted and buffer exchanged to NTE buffer ( 20 mM Tris , pH 8 . 0 , 120 mM NaCl , 1 mM EDTA ) using an Amicon Ultra-2 Centrifugal Filter Unit with Ultracel-100 membrane ( Millipore ) . Purified CHIKV 37997-mCherry virus was flash-frozen on C-Flat copper grids ( R2/2 , 200 mesh ) in liquid ethane . The movie-mode data of CHIKV 37997-mCherry was taken on a DE-20 direct electron detector ( Direct Electron , LP , San Diego , CA ) on a 200-kV JEOL 2200FS microscope under low-dose conditions . We used a frame rate of 25 frames/s and 1 . 52 s exposure that corresponded to ∼38 e/Å2 dose/image . Individual frames in each image were aligned using DE_process_frames . py script provided by Direct Electron ( Direct Electron , LP , San Diego , CA ) using radiation damage compensation to increase low-frequency content in the final images , thus facilitating subsequent particle alignment . Particle images were automatically boxed out using the E2BOXER program from the EMAN2 suite [18] . EMAN2 was used for CTF determination and correction; the subsequent processing ( particle alignment , orientation search and refinement , and 3D reconstruction ) was done in IMAGIC-5 [19] . All data were analyzed using Prism software ( La Jolla , CA ) and statistical significance was assigned when P values were < 0 . 05 . Neutralization curves were calculated using non-linear regression . Mouse survival curves were compared with Gehan-Breslow-Wilcoxon test . Viral titers and mouse joint swelling were analyzed using a one-way or two-way ANOVA test .
To make a fluorescently tagged CHIKV , we inserted DNA encoding mCherry into a cDNA clone immediately downstream of the furin cleavage site between E3 and E2 in CHIKV 37997 ( Fig 1A ) . The resulting chimeric CHIKV 37997-mCherry virus was then produced as previously described [16] . In CHIKV 37997-mCherry infected cells , mCherry-E2 displayed a typical glycoprotein expression pattern: from Golgi apparatus to endosomes to the plasma membrane , and alphavirus-induced Env positive membrane extensions were observed ( Fig 1B ) . Next , we compared viral growth curves of CHIKV 37997-mCherry with the parental CHIKV 37997 virus in vitro . Viruses released in the culture supernatants were quantified by viral RNA level measured by qRT-PCR and infectious particle number measured by plaque assay . In BHK21 cells , CHIKV 37997-mCherry demonstrated slightly slower growth kinetics compared to parental wild-type virus at early time points ( before 12 hours post-infection ) , however no differences were seen at later time points ( after 18 hours post-infection ) ( Fig 1C ) . CHIKV 37997-mCherry formed plaques similar to parental viral plaques ( Fig 1D ) and had a similar viral genome copy to PFU ratio ( 2 . 49 x 103 ) as the parental CHIKV 37997 ( 2 . 75 x 103 ) . In gradient purified CHIKV 37997-mCherry virus E1 and capsid proteins were present at the expected size ( Fig 1E ) . The double bands that react with both anti-mCherry monoclonal antibody and anti-CHIKV polyclonal antibody are probably E3-mCherry-E2 precursor and mCherry-E2 fusion proteins . Insertion of mCherry between E3 and E2 might slow down the furin cleavage of E3 because the parental CHIKV 37997 virus contains very little uncleaved E3-E2 ( p62 ) precursor . The smaller sizes of E3-mCherry-E2 and mCherry-E2 than expected may due to some glycosylation difference with mCherry insertion rather than truncation within the fusion protein as the viral RNA genome was intact in gradient purified CHIKV 37997-mCherry ( S1 Fig ) . To understand the presentation of tagged mCherry proteins on chimeric CHIKV 37997-mCherry virus , we performed single-particle cryoEM analysis of purified CHIKV 37997-mCherry virus particles ( Fig 2 ) . CHIKV 37997-mCherry virions display a typical T = 4 icosahedral structure similar to wild-type CHIKV and other alphaviruses ( Fig 2A and 2B ) . Compared with the cryoEM map of CHIK virus-like particles that we previously obtained ( Fig 2C ) [16] , CHIKV 37997-mCherry displayed extra density on the particle surface representing mCherry proteins fused at the N-terminal of E2 ( Fig 2D ) . The mCherry proteins were well exposed , occupying the space between viral spikes on the particle surface ( Fig 2D ) . After we confirmed that CHIKV 37997-mCherry replicated to a similar level as wild-type CHIKV 37997 in vitro , we next compared their growth and pathogenesis in vivo . In an IFNαR-/- lethal infection model , injection of 10 PFU of CHIKV 37997 into the mouse footpad resulted in death in 100% of the mice in 4 days . In contrast , all the mice infected with 10 PFU of CHIKV 37997-mCherry survived through the period of observation ( 23 days ) ( Fig 3A ) . Surprisingly , infection with the CHIKV vaccine strain 181/clone 25 ( CHIKV 181/25 ) also resulted in death in 40% of the mice within 9 days ( Fig 3A ) . Infection of all three strains caused footpad swelling on the ipsilateral side , but with a delay observed with CHIKV 37997-mCherry and CHIKV 181/25 infections compared to CHIKV 37997 ( Fig 3B ) . Because CHIKV 37997-mCherry was found to be attenuated compared with the parental virus at a 10 PFU dose , we next compared its pathogenesis with the vaccine strain CHIKV 181/25 in IFNαR-/- mice over a range of different inoculation doses . Survival curves of mice infected with 10 PFU to 104 PFU of CHIKV 37997-mCherry vs CHIKV 181/25 showed no significant difference ( Fig 4 ) , indicating CHIKV 37997-mCherry is similarly attenuated as CHIKV 181/25 in IFNαR-/- mice . Because CHIKV 37997-mCherry is attenuated like the vaccine strain , we next tested if infection with CHIKV 37997-mCherry provides a protection similar to the vaccine strain ( Fig 5 ) . The IFNαR-/- mice that survived infection with 10 PFU to 103 PFU of CHIKV 37997-mCherry or CHIKV 181/25 were challenged with 10 PFU of pathogenic CHIKV LR strain . All the mice pre-exposed to either CHIKV 37997-mCherry or CHIKV 181/25 survived ( Fig 5A ) , although it was apparent sterilizing immunity was not induced in some animals . However , viral loads of challenge virus were limited compared to CHIKV LR replication in the control mice ( Fig 5B ) , which demonstrated death in 100% of the mice within 4 days ( Fig 5A ) . We next compared antibody responses in CHIKV 37997-mCherry vs CHIKV 181/25 infected IFNαR-/- mice before ( Fig 6A ) and after ( Fig 6B ) re-challenge with CHIKV LR . Infection with 10 , 100 or 1 , 000 PFU of CHIKV 37997-mCherry and CHIKV 181/25 induced similar levels of antibodies against CHIKV in mice at 3 weeks post-infection and 11 days post-challenge with CHIKV LR . As expected , antibodies against mCherry were induced in mice infected with 37997-mCherry . Neutralizing activities of the antiserum in CHIKV 37997-mCherry and CHIKV 181/25 infected mice were next tested using a plaque reduction neutralization assay ( Fig 7 ) . Serum from CHIKV 37997-mCherry and CHIKV 181/25 vaccinated mice at both 3 weeks post-infection ( Fig 7A ) and 11 days post-re-challenge ( Fig 7B ) neutralized virus entry of CHIKV 181/25 at similar efficiency . Interestingly , serum from CHIKV 37997-mCherry vaccinated mice neutralized virus entry of CHIKV 37997-mCherry more efficiently than serum from CHIKV 181/25 vaccinated mice , with IC50s for the former ~30-fold higher than those for the latter ( Fig 7C and 7D ) . Serum from CHIKV 181/25 infected mice neutralized CHIKV 181/25 and CHIKV 37997-mCherry at similar efficiency ( Fig 7 ) , excluding the possibility that serum from CHIKV 181/25 infected mice only weakly neutralized CHIKV 37997-mCherry . After we found CHIKV 37997-mCherry was attenuated in the lethal mouse model of immunocompromised IFNαR-/- mice , we next compared its pathogenesis with the parental and vaccine strains in a chronic infection mouse model ( Fig 8 ) . Three weeks old wild-type mice were infected with 103 PFU of CHIKV 37997 , CHIKV 37997-mCherry or CHIKV 181/25 by injection in the rear footpad . As expected , CHIKV 37997 infection caused significant swelling of the footpads on the ipsilateral side at day 3 and day 7 post-infection ( Fig 8A ) . In contrast , infection with CHIKV 37997-mCherry did not cause significant footpad swelling on day 3 and significantly less swelling than CHIKV 37997 on day 7 ( Fig 8A ) . Tissue inflammation induced by infection of the vaccine strain CHIKV 181/25 in this chronic infection model was also attenuated to a similar level as CHIKV 37997-mCherry ( Fig 8A ) . Consistently , plasma RNA viral loads at 3 days post-infection of CHIKV 37997-mCherry or CHIKV 181/25 were significantly lower than those in CHIKV 37997 infected mice ( Fig 8B ) . At 28 days post-infection , in the chronic phase of CHIKV infection , persistent viral RNA in ankle tissues was compared ( Fig 8C ) . As we , and other researchers , have reported before [16 , 20] , CHIKV persists in both the ipsilateral and contralateral ankles after CHIKV 37997 infection . In contrast , viral RNA in the ankles on the contralateral side of CHIKV 37997-mCherry infected mice were not significantly higher than the assay limit of detection , and viral loads in ankles on the ipsilateral side were significantly lower than those in parental CHIKV 37997 infected mice ( Fig 8C ) . In vaccine strain CHIKV 181/25 infected mice viral loads varied extensively between individual mice in ankles of both sides ( Fig 8C ) , consistent with the large variation of plasma viral load in CHIKV 181/25 infected mice ( Fig 8B ) . In some CHIKV 181/25 infected mice , virus replicated and persisted at high loads relative to those in CHIKV 37997-mCherry infected mice , although no statistical significant difference could be demonstrated between the two groups . Similar attenuation of CHIKV 37997-mCherry at a dose as high as 106 PFU was observed in 3-weeks old mice ( Fig 9 ) . We observed presentation of mCherry on the surface of chimeric CHIKV 37997-mCherry in a repeating pattern with 240 copies of mCherry on each virion . Strong anti-CHIKV responses were induced by infection of CHIKV 37997-mCherry , so we next tested if strong antibody responses against mCherry could be induced by CHIKV 37997-mCherry . In adult mice , infection with 104 or 106 PFU of CHIKV 37997-mCherry induced high antibody responses against mCherry when assessed at day 14 post-infection and the high levels of antibody lasted for at least 34 days ( Fig 10A ) . Immunization with 100 μg of purified mCherry protein in complete Freud’s adjuvant induced antibodies against mCherry at more than 20 fold lower level than induced by CHIKV 37997-mCherry infection at day 14 ( Fig 10A ) . Boosting with 50 μg of mCherry enhanced antibody responses to the level induced by CHIKV 37997-mCherry infection in only 2 of 7 mice ( Fig 10A ) . Interestingly , serum from mCherry immunized mice could neutralize CHIKV 37997-mCherry virus entry , although at a lower efficiency than the serum from CHIKV 37997-mCherry infected mice ( Fig 10B ) . This is consistent with the results that the serum from CHIKV 37997-mCherry infected IFNαR-/- mice neutralized CHIKV 37997-mCherry at about a ~30-fold higher efficiency than the serum from CHIKV 181/25 infected IFNαR-/- mice ( Fig 7C & 7D ) .
CHIKV is the most common alphavirus infecting humans–with millions of individuals infected during the 2000s , resulting in thousands of deaths [2] . Currently there are no licensed vaccines or treatments for CHIKV infection . An attenuated live vaccine strain ( CHIKV 181/25 ) was developed and reported to be safe , highly immunogenic and produced well-tolerated side effects in a phase II clinical trial [21–23] . Two point mutations at I12 and R82 in the E2 glycoprotein of CHIKV 181/25 are responsible for its attenuation of acute disease in mice [24 , 25] . Acquisition of R82 in E2 enhances the affinity of CHIKV 181/25 particles for glycosaminoglycans ( GAGs ) in vitro [26] , which may limit the capacity of CHIKV to disseminate from early sites of primary replication in vivo [24 , 27] . A recent study reported that R82 in E2 also renders CHIKV 181/25 vulnerable to neutralization by antibodies targeting E2 domain B and therefore more likely to be cleared [28] . In our study , CHIKV 37997-mCherry was more efficiently neutralized by serum from CHIKV 37997-mCherry infected mice than from CHIKV 181/25 infected mice ( Fig 7 ) . This may explain the quick clearance of CHIKV 37997-mCherry in blood and ankle tissues ( Figs 8 & 9 ) and therefore the attenuation of CHIKV 37997-mCherry in vivo . Interestingly , we found antiserum against mCherry neutralized virus entry of CHIKV 37997-mCherry although at a lower efficiency than the antiserum raised in CHIKV 37997-mCherry infected mice ( Fig 10 ) . mCherry proteins displayed on the surface of CHIKV 37997-mCherry particle was not expected to mediate either the virus-receptor interaction or membrane fusion during virus entry . How anti-mCherry antibodies neutralize CHIKV 37997-mCherry warrants further studies . We can speculate that binding of antibodies to mCherry may limit the flexibility of E2 domain B and therefore prevent the activation of the fusion loop for virus entry . This additional neutralizing epitope and the presence of anti-mCherry antibodies provides the likely explanation for the ~30-fold more efficient neutralization of CHIKV 37997-mCherry by the antiserum raised in CHIKV 37997-mCherry infected mice compared to antiserum raised in CHIKV 181/25 infected mice , and therefore the attenuation of CHIKV 37997-mCherry ( Fig 7 ) . Whether or not chimeric CHIKV virus carrying other foreign antigens is attenuated depends on whether the foreign antigens are presented as neutralizing epitopes . Reversion of attenuating mutations was reported in CHIKV 181/25 infected wild-type mice [24 , 25] . The insertion of mCherry between E3 and E2 in CHIKV 37997 genome may not be so easy to revert compared to reversion of single amino acid mutation . This could explain the large variation of virus titers in blood and ankle tissues in mice infected with CHIKV 181/25 compared to consistent low virus titers in mice infected with CHIKV 37997-mCherry ( Figs 8 & 9 ) . Future studies on virus stability in vitro and in vivo as well as in vivo tropism of CHIKV 37997-mCherry are needed to understand the dynamics of the virus populations . On each chimeric CHIKV 37997-mCherry virion , 240 copies of mCherry are presented in a repetitive pattern between viral spikes ( Fig 2 ) . The high density of mCherry can be easily recognized by B cells and , therefore , activate B cells efficiently . Consistently , a single infection of CHIKV 37997-mCherry induced long-lasting and high titer antibody responses against mCherry that can only be reached after boosting in mice conventionally immunized with mCherry proteins ( Fig 10 ) . Our results suggest that fusions to E2 may be a promising vaccine platform . Fusion of foreign antigens to E2 generates replication-competent chimeric virus able to present these antigens in a repetitive array on virion surface [9 , 10] . Live attenuated chimeric aphaviruses are able to induce strong antibody responses against the foreign antigen as well as anti-alphavirus response . Recently , a study was reported using CHIKV virus-like particles as a vaccine platform to present NANP repeats from the circumsporozoite protein ( CSP ) of the Plasmodium falciparum malaria parasite [11] . Similarly the dense array of NANP repeats on VLPs induced strong antibody responses that protected mice from malaria infection . In that study NANP repeats of 58 amino acid were inserted on E3 and E2 separately . In our study , a much larger antigen , mCherry ( 236 amino acid ) was fused to E2 with little effect on E3 cleavage efficiency , which permits infectivity of the chimeric virus . Virus replication generates viral RNA replication intermediates and viral RNA that strongly activate host innate immune responses , which may explain the strong immunogenicity of our replication-competent chimeric virus[29] . In summary , we report an attenuated mCherry tagged replication-competent CHIKV that may suggest an additional risk mitigation should an individual be accidentally infected . In addition our work suggests similar chimeric replication-competent alphaviruses can serve as vaccines against other pathogens and diseases , in addition to the chimeric VLP-based vaccines . | Chikungunya virus ( CHIKV ) is an alphavirus capable of causing long term debilitating joint and muscle pain at a great economic expense . Currently there are no licensed vaccines or treatment for CHIKV infection . We generated a modified version of the virus , termed CHIKV 37997-mCherry , stably expressing a fluorescent tag on the surface of the virus . To achieve this , a red fluorescent protein , mCherry , was fused to the virus envelope protein E2 . Structural studies demonstrated the presence of mCherry on the virus surface . Infection of mice with CHIKV 37997-mCherry caused less severe disease in the animals compared to wild-type virus . Infection with CHIKV 37997-mCherry induced immune responses against both the mCherry protein and the virus . Furthermore , CHIKV 37997-mCherry is as attenuated as the vaccine strain CHIKV 181/clone 25 in different mouse models , causing less joint swelling and reduced persistence of viral genomes in tissue . Our work suggests that chimeric alphaviruses carrying foreign antigen on virus particles may serve as vaccines against both aphaviruses and other pathogens and diseases . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"alphaviruses",
"vir... | 2018 | An attenuated replication-competent chikungunya virus with a fluorescently tagged envelope |
In this study , a Burkholderia mallei tonB mutant ( TMM001 ) deficient in iron acquisition was constructed , characterized , and evaluated for its protective properties in acute inhalational infection models of murine glanders and melioidosis . Compared to the wild-type , TMM001 exhibits slower growth kinetics , siderophore hyper-secretion and the inability to utilize heme-containing proteins as iron sources . A series of animal challenge studies showed an inverse correlation between the percentage of survival in BALB/c mice and iron-dependent TMM001 growth . Upon evaluation of TMM001 as a potential protective strain against infection , we found 100% survival following B . mallei CSM001 challenge of mice previously receiving 1 . 5 x 104 CFU of TMM001 . At 21 days post-immunization , TMM001-treated animals showed significantly higher levels of B . mallei-specific IgG1 , IgG2a and IgM when compared to PBS-treated controls . At 48 h post-challenge , PBS-treated controls exhibited higher levels of serum inflammatory cytokines and more severe pathological damage to target organs compared to animals receiving TMM001 . In a cross-protection study of acute inhalational melioidosis with B . pseudomallei , TMM001-treated mice were significantly protected . While wild type was cleared in all B . mallei challenge studies , mice failed to clear TMM001 . Although further work is needed to prevent chronic infection by TMM001 while maintaining immunogenicity , our attenuated strain demonstrates great potential as a backbone strain for future vaccine development against both glanders and melioidosis .
Melioidosis and glanders are severe zoonotic diseases caused by two closely related Gram-negative pathogens known as Burkholderia pseudomallei and B . mallei , respectively [1 , 2] . The genomic relatedness between these two pathogens suggests that B . mallei is a host-adapted clone of B . pseudomallei , which evolved from a process of reductive evolution . Genes retained by B . mallei share 99% sequence identity with their B . pseudomallei orthologs and of those , 650 genes have been identified as putative virulence determinants via in silico genomic subtraction from non-pathogenic Burkholderia species [3] . In addition , the presence of very few B . mallei specific genes suggest it’s possible to generate a live attenuated vaccine with a B . mallei backbone that can cross-protect against both melioidosis and glanders [4] . Where B . pseudomallei is an environmental saprophytic pathogen ubiquitous in soil and fresh water surfaces , B . mallei is an obligate mammalian pathogen that typically infects solipeds ( horses , donkeys , etc ) [1 , 5] . Despite epidemiological differences , the clinical and pathological manifestations of B . pseudomallei or B . mallei infections bear striking resemblance . Both pathogens can be contracted via the cutaneous , oral and/or inhalational routes . Depending on the dose and route of transmission , B . pseudomallei or B . mallei infection may result in an acute or chronic disease . Clinical manifestations of acute infection from either disease , which include fever , malaise , abscess formation , pneumonia and sepsis , are non-specific . The lack of pathognomonic symptoms , in addition to their ability to cause silent infection , makes rapid and accurate diagnosis problematic for these Burkholderia infections . Since mortality rates among severe infections are high , and there are no reliable antibiotic therapy or licensed pre- and post-exposure vaccines , both pathogens remain top candidates for bioterrorist use and thus have been classified as category B , tier 1 , biothreat agents [1] . The destructive potential of B . pseudomallei and B . mallei has heightened concerns among public health officials due to the increased potential of opportunistic infection among growing populations of diabetic and immunocompromised people [2] . For military personnel and susceptible individuals , the availability of a vaccine would be the most efficacious and cost-effective way to protect from disease . Progress in vaccine development shows formulations consisting of subunits or live-attenuated strains are the most effective in conferring protection against both pathogens . Subunit vaccines consisting of purified protein [6]; recombinant Hcp proteins [7]; lipopolysaccharide ( LPS ) [8 , 9]; truncated recombinant proteins LolC and PotF [10]; and outer membrane vesicles ( OMV ) [11] have achieved the greatest protection to date . While encouraging , subunit vaccines provided only partial protection , which is attributed to their inability to generate broad protective immunity , specifically cell mediated immunity [12] . Live attenuated vaccines are recognized for their ability to elicit strong broad immune responses that provide long-lasting protection [12] . Thus far , attenuated mutants lacking a functional purN , purM , aroB , ilvl , or bipD genes in B . pseudomallei , and ilvl or DD3008 ( capsule ) genes in B . mallei have been evaluated for their protective potential [12] . Although these candidates have proven capable of providing significant protection during the acute stage of infection , none have yet to afford full protection during the chronic stage of infection . To create a live attenuated B . mallei mutant that will generate a protective immune response against chronic infection , we focused on iron transport systems as a target of mutagenesis . For a majority of bacterial pathogens , the acquisition of iron and iron complexes has long been recognized as major pathogenic determinant and thus also represent a promising target for vaccine development . In the host environment , free iron is too scarce and iron complexes are too large to diffuse effectively through porin channels . To survive in these growth-limiting conditions , bacteria utilize siderophores and/or high-affinity outer-membrane receptors to uptake iron and iron complexes [13] . In the case of B . pseudomallei and B . mallei , very little information exists concerning iron uptake mechanisms in the host and their roles in virulence . In one study , Kvitko el at . , generated single , double and quadruple B . pseudomallei mutants defective in siderophores and/or hemoglobin utilization [14] . While mutants defective in these systems are often attenuated , the B . pseudomallei mutants remained fully virulent in a murine model of acute melioidosis [14] . Failure to eliminate virulence was attributed to redundancy in the iron transport system , citing a reliance on alternative iron sources and acquisition mechanisms . To negate this redundancy , we targeted the inner membrane energy transfer protein TonB , an essential component that interacts with all outer membrane receptor proteins that carry out high-affinity binding and energy dependent iron uptake [15 , 16 , 17] . When assessed in multiple models of infection , tonB mutants displayed severe attenuation compared to their wild-type homologs [18 , 19 , 20 , 21] . In the case of K . pneumoniae , Hsieh et al . showed 100% protection in challenge mice previously vaccinated with the tonB mutant homolog [20] . Thus , Burkholderia TonB-dependent iron-transport systems , specifically their contribution to survival , persistence and potential as targets for attenuation , should be investigated further . In this communication , we describe the construction and characterization of a B . mallei tonB mutant as a backbone strain for subsequent vaccine development against acute inhalational murine glanders and melioidosis .
The bacterial strains and plasmids used in this study are listed in Table 1 . All E . coli strains were grown in Luria-Bertani ( LB ) media at 37°C or 30°C , as required . For all the experiments , all B . mallei strains were taken from freezer stocks , plated on LB agar containing 4% glucose ( LBG ) , and incubated at 37°C for 3 days . For liquid cultures , a few colonies ( 2–3 ) were inoculated into 20 mL of LBG broth and incubated overnight with agitation at 37°C . When employing antibiotic selection , we used kanamycin and polymyxin B at concentrations of 50 μg/mL and 30 μg/mL , respectively . For counter-selection , co-integrates were grown in YT broth ( 10 g of tryptone and 10 g of yeast extract ) and then plated on sucrose agar ( YT agar supplemented with 5% sucrose ) , as described by Hamad et al . [22] . When appropriate , LBG broth and agar were supplemented with FeSO4 at a concentration of 200 μM . Unless otherwise stated , wild type B . mallei ATCC 23344 or CSM001 ( B . mallei Lux ) , B . mallei TMM001 ( tonB mutant ) , and TMM002 ( pTonB-comp ) were used in all experiments . Cloning methods were performed as previously described [22] . Chromosomal and plasmid DNA were isolated by using the DNeasy Qiagen Blood and Tissue kit , and the QIAGEN Plasmid Mini Kit , respectively ( Qiagen , Inc . , Valencia , CA ) . Polymerase chain reaction ( PCR ) products were purified with either the QIAquick PCR purification kit or QIAquick gel extraction kit ( Qiagen ) . Restriction enzymes and T4 DNA ligase were purchased from NEB and used in accordance with the manufacturers’ instructions ( New England Biolabs Inc . , Ipswich , MA ) . Primers used in this study were purchased from Sigma-Aldrich Co ( St . Louis , MO ) . DNA fragments obtained for cloning were amplified with Phusion High-Fidelity DNA polymerase ( New England Biolabs ) by using the following touchdown PCR protocol: 1 cycle of 95°C for 5 min , 29 cycles of 95°C for 30 sec , 70°C to 55°C ( -5°C/cycle ) for 30 sec , 72°C for 2 min , 29 cycles of 95°C for 30 sec , 55°C for 30 sec , 72°C for 30 sec , and 1 cycle of 72°C for 7 min . Matched adaptamers containing 3’ enzyme restriction sites and 5’ complementary sequences were amplified via touchdown PCR . The sequences of the PCR primers were as follows: ΔtonB US forward primer ( AAG CTA GCC CTC GGC GCG GCG ATC CGC GAC GT ) ( underlined sequence indicates NheI site ) ; ΔtonB US reverse primer ( CGG TAT TGC CGA GAT TAA CGG TGC GGC ACG TCG T ) ; ΔtonB DS forward primer ( ACG ACG TGC CGC ACC GTT AAT CTC GGC AAT ACC G ) ; and ΔtonB DS reverse primer ( CCA AGC TTT ACG AGC ATG ACG TCG ACG AGC GGC GTC ATG TTG ) ( underlined sequence indicates HindIII site ) . The adaptamers were fused together via splicing by overlap extension ( SOE ) PCR to create a 1794-bp chimeric fragment containing sequences flanking the tonB gene plus its first 33 codons . The chimeric fragment was digested with NheI and HindIII and ligated into the pMo130 vector to create the allelic exchange plasmid pTonB-allex . The pTonB-allex plasmid was then transformed into E . coli S17-1 and introduced into B . mallei via conjugal transfer . Merodiploids were selected based on their growth on kanamycin ( Km ) and polymyxin B ( Pbx ) agar plates and ability to turn yellow after exposure to pyrocathecol . Single deletion mutants were counter selected on YT agar supplemented with 5% sucrose and 200 μM FeSO4 . After the tonB mutant was screened for Pxb resistance and Km sensitivity , deletion was confirmed via PCR amplification , followed by sequencing of the tonB gene and flanking DNA regions by using the following primers: confirmation forward primer ( 5’ GCG CCA CGC GGC CGA TTG CCG CTT TCT ) and confirmation reverse primer ( ACA GAA CCG TGC CGT CGC TTT ) . To restore the tonB mutant ( renamed TMM001 ) to wild-type function , pMo168 carrying a functional tonB gene plus its native promoter was used for complementation . Briefly , a fragment containing the wild-type tonB gene plus approximately 120 bp of its upstream sequence flanked by enzyme restriction sites was amplified by using the following PCR primers: complementation forward primer ( CCG CTA GCC TGA TTT TCC GCA AGT GAT GCA GCA CT ) ( underlined sequence indicates NheI site ) and complementation reverse primer ( CCA AGC TTT TAA TCG GTC AGA GTG AAG TCA TAA GGC ) ( underlined sequence indicates HindIII site ) . The fragment was then digested with NheI and HindIII and ligated into the pMo168 plasmid to create pTonB-comp . After transformation into E . coli S17-1 , pTonB-comp was introduced into B . mallei via conjugal transfer . TMM001 containing pTonB-comp was isolated via selection on LBG + Km agar plates and confirmed by PCR amplification , followed by sequencing of the region flanking the tonB gene by the same primers used to confirm the tonB mutation . Overnight cultures were used to inoculate 50 mL of LBG with 6 x 106 CFU of each strain . Inoculated cultures were then incubated with agitation at 37°C . At the indicated time points , 1 mL aliquots from each culture were taken to measure optical density at 600 nm . Individual data points represent the OD600 mean ± standard deviation ( SD ) of three independent experiments . A significant difference due to treatment over time was ascertained via two-way ANOVA . Significant differences ( p ≤ 0 . 05 ) of each OD600 reading was determined at every time point compared to wild type using one-way ANOVA followed by Dunnett’s multiple comparison test . Overnight cultures were diluted to 1 x 105 CFU/ml in LBG + 200 μM of 2 , 2’-dipyridyl and poured onto plates , as previously described [23] . Disks containing iron sources were placed on the surface of the LBG plates , which were incubated at 37°C for 48 h . Disks contained 10 μL of the following compounds at the specified concentrations: hemin , 8 . 0 μM; hemoglobin , 4 . 5 μM; myoglobin , 4 . 5 μM; transferrin , and lactoferrin , both at 30 μM or FeSO4 , 10 mM [23] . Iron utilization was quantified by measuring the diameter of growth around the disk . Ten μl samples of overnight cultures , grown in LBG or LBG + 200 μM FeSO4 or 200 μM 2 , 2’-dipyridyl , were spotted onto CAS agar plates and incubated at 37°C . Halos were then monitored and the diameter of color change was measured over the course of the next 4 days . For the CAS agar , solutions were prepared as previously described [24] . An unpaired t test with equal standard deviation was performed on halo measurements to ascertain a significant difference ( p ≤ 0 . 05 ) between the strain-specific halos produced . Female , 6- to 8-week-old BALB/c mice obtained from Harlan Laboratories ( Indianapolis , IN , USA ) were housed in microisolator cages under pathogen-free conditions . Animals were provided with rodent feed and water ad libitum and maintained on a 12 h light cycle . Before experiments , mice were afforded an adaption period of at least 1 week . Humane endpoints were strictly observed and time of death was recorded upon death of the animal or at the study’s end . Animals were observed closely throughout the study for clinical symptoms ( immobility , dyspnea , paralysis ) and moribund animals were anesthetized and then euthanized via cervical dislocation . Anesthetized BALB/c mice ( n = 8 per treatment ) were inoculated i . n . with the indicated CFU of TMM001 , grown in LBG ± 200 μM FeSO4 and diluted in phosphate-buffered saline ( PBS ) in a total volume of 50 μL ( 25 μL/ naris ) . Mice were monitored and deaths recorded over a period of 14 days . Survival curves were generated and analyzed by using the Kaplan-Meier method . A significant difference ( p ≤ 0 . 05 ) in survival curves was ascertained via a log-rank test . Anesthetized mice ( n = 8 per treatment ) were challenged i . n . with 1 . 5 x 104 CFU/50 μL of the B . mallei bioluminescent reporter strain CSM001 , and TMM001 in LBG ± 200 μM FeSO4 . At 24 , 48 and 72 h post challenge , BALB/c mice were euthanized and necropsied for organ collection . The lungs , liver and spleen were homogenized in 1 mL of PBS by using a tissue grinder ( Covidien , Mansfield , MA ) , and then the bacteria were enumerated by standard plate counts on LBG + 200 μM FeSO4 . Significant differences ( p ≤ 0 . 05 ) in colonization at 24 and 48 h were individually determined via one-way ANOVA followed by Tukey’s multiple comparisons test . Significant difference ( p ≤ 0 . 05 ) in colonization at 72 h was extrapolated by using an unpaired t test with equal standard deviation . Anesthetized mice ( n = 8 per treatment ) were immunized i . n . with PBS or the indicated CFU of TMM001 diluted in PBS in a total volume of 50 μL ( 25 μL/ naris ) . Mice were challenged 21 days post immunization with 1 . 5 x 105 CFU ( ~220 LD50 ) of CSM001 ( LD50 of 6 . 81x102 CFU ) or 9 x 102 CFU ( ~3 LD50 ) of wild-type B . pseudomallei K96243 ( LD50 of 3 . 12x102 CFU ) [25 , 26] , diluted in PBS in 50 μL ( 25 μL/ naris ) . Mice were monitored , and deaths were recorded until the end of the study . Survival curves were generated and analyzed by the Kaplan-Meier method . A significant difference ( p ≤ 0 . 05 ) in survival curves was ascertained via log-rank test . To find significant differences in individual treatment , when compared to the PBS-treatment control , an additional log rank test was employed in which an adjusted definition of significance ( p ≤ 0 . 05/ the number of pair wise comparisons ) was used . Bioluminescent images were acquired on an IVIS Spectrum ( Caliper Corp . , Alameda , CA , USA ) , as previously described [25] . Briefly , anesthetized BALB/c mice placed in the isolation chamber were transferred to the imaging chamber , which was then connected to an internal anesthesia delivery system that maintained 1–2% isoflurane . Bioluminescence signaling was measured after three minutes’ exposure with no excitation ( filters blocked ) and an open emission filter to capture all luminescent signals from labeled bacteria . To depict the differences in intensity of the signal , bioluminescence was represented in the images with a pseudo-color scale ranging from red ( most intense ) to violet ( least intense ) . Scales were manually set to the same values for every comparable image to normalize the intensity of the bioluminescence across time points . Serum extracted from PBS or TMM001-vaccinated BALB/c mice at 21 days post treatment , was evaluated for B . mallei-specific IgG1 , IgG2a and IgM using the Ready-Set-Go ! ELISA Kit ( Affymetrix eBioscience , San Diego , CA ) as instructed by the manufacturer . Briefly , microplates ( Costar , Cambridge , MA ) were coated with 10 μg/ml of heat inactivated B . mallei and incubated overnight at 4°C . Wells were then washed twice with PBS , 0 . 05% Tween-20 , and then blocked over night with the Assay Buffer provided in the kit . After the wells were washed , a 1:10 , 000-fold dilution of sera samples was added to the appropriate wells , followed by the detection antibody provided by the kit . After 3 h incubation , the wells were washed four times before 100 μL of the substrate solution was added . After 15 min incubation , 100 μL of stop solution consisting of 2 N H2SO4 was added , and absorbance was measured at 450 nm with the Epoch microplate spectrophotometer ( Winooski , VT ) . An unpaired Student’s t test was performed to ascertain a significant difference ( p ≤ 0 . 05 ) in B . mallei-specific Ig levels between the PBS and TMM001-treated mice . At the indicated time points , necropsies were performed to collect the lungs , liver and spleen . Organs were instilled with 10% formalin , paraffin-embedded , and processed for histopathology . Hematoxylin and eosin-stained slides were examined and blindly scored by a pathologist for the follow observations: perivascular and peribronchial inflammatory infiltrates , necrosis and microabscesses in the lung; granulomas , necrosis and histocytosis in the spleen; and inflammation and necrosis in the liver . Severity of pathology was scored using the following combined scale: 0 ( unremarkable ) , 1 ( minimal ) , 2 ( mild ) , 3 ( moderate ) and 4 ( severe ) . Pathology scores were combined with a percent factor associated with the extent of the damage ( 0–25% , 25–50% , 50–75% , 75–100% ) and added together to give the total score for each organ . Each image is representative of three replicates per treatment . A two-way ANOVA was performed on each organ individually to assess a significant difference in treatment over time . Student’s t test was performed to ascertain a significant difference ( p ≤ 0 . 05 ) between the treatments of each organ , individually , at 0 and 48 h . At the indicated time points following challenge , whole blood was collected by cardiac puncture . The blood was stored in microvette tubes without anti-coagulant and incubated at room temperature for 20 min to permit clotting . Serum was collected after centrifugation of the tubes and stored at -80°C . Samples were inactivated as previously described [27] and verified for sterility . Serum chemokine/cytokine levels were measured by using the murine bioplex ELISA kit ( BioRad Bio-Plex Pro Mouse Cytokine 23-plex Assay ) according to the manufacturer’s specification . Serum sample were diluted 1:4 in PBS and expression of the following molecules was determined: interleukin ( IL ) -1α , IL-1β , IL-2 , IL-3 , IL-4 , IL-5 , IL-6 , IL-9 , IL-10 , IL-12 ( p40 ) , IL-12 ( p70 ) , IL-13 , IL-17A , eotaxin , granulocyte–colony-stimulating factor ( G-CSF ) , granulocyte–macrophage colony-stimulating factor ( GM-CSF ) , gamma interferon ( IFN-γ ) , keratinocyte-derived chemokine ( KC ) , monocyte-chemotactic protein ( MCP-1 ) , macrophage inflammatory protein ( MIP ) -1α , MIP-1β , RANTES , and ( tumor necrosis factor ) TNF-α . Data values represent the mean ± the SEM of three animals per treatment and were ascertained as previously described [27] . Out of range values above the asymptote of equation ( > OOR ) we set to the highest extrapolated value to provide a conservative estimate that allowed statistical analysis . A significant difference ( p ≤ 0 . 05 ) in individual serum cytokine levels in PBS vs . TMM001-treated mice was determined by using the Mann-Whitney test .
A previously described method for genetic manipulation via allelic exchange was used to create an unmarked tonB mutant in the B . mallei strain ATCC 23344 [22] . To ensure the mutant phenotype was not due to polar effects incurred during mutagenesis , the TMM001 was transformed with the plasmid pTonB-comp , which carries the intact tonB gene plus its native promoter ( Table 1 ) . Unlike the wild type , TMM001 appears as bright yellow colonies that discolor the surrounding media when grown in Luria Bertani + 4% glycerol ( LBG ) plates ( S1 Fig ) . This phenotype in iron transport mutants has been attributed to the unregulated production and accumulation of iron-bound siderophores , which are yellow-to-brown in color , in contrast to uncolored iron-free siderophores [28 , 29 , 30] . The wild-type phenotype was restored when TMM001 was complemented , which grew as muted yellow-beige colonies with no media discoloration . To determine the effect of the tonB deletion on growth rate and iron requirement , growth curves were performed with the following strains and broth conditions: wild type in LBG , TMM001 in LBG ± 200 μM FeSO4 ( Fig 1 ) . When grown in LBG , TMM001 exhibited a reduced growth rate , displaying a longer lag phase , compared to that of the wild type . When grown in LBG + 200 μM FeSO4 , the growth rate of the TMM001 increased substantially approaching that of the wild type . Notably , TMM001 grown in iron-supplemented media maintained wild-type growth rates showing statistically significant differences only after 25h of growth . To determine if the deletion of tonB in B . mallei resulted in differential siderophore production , both the wild-type and TMM001 were seeded on CAS agar . The CAS media was used because when strong iron chelators , such as siderophores , are secreted , they are able to strip the dye complex of iron , which results in the formation from blue to orange/yellow zones ( S2 Fig ) . Siderophore secretion zones were measured after 96 h and calculated as the diameter of the halo minus the diameter of bacterial colony on the filter disk . TMM001 produced significantly larger halos ( 33 . 3 ± 0 . 5 mm ) compared to those of the wild type ( 12 . 3 ± 0 . 6 mm ) . These results are in line with previously studies that show iron transport mutants hypersecrete siderophores in a futile attempt to acquire iron [28 , 30 , 31 , 32 , 33 , 34 , 35] . A disk diffusion assay was performed to examine the ability of the TMM001 to utilize the following sources of iron: FeSO4 , hemoglobin , hemin , lactoferrin , and transferrin . Iron assimilation was determined by measuring the diameter ( mm ) of bacterial growth around the disk containing specific iron sources placed on iron-depleted media ( S1 Table ) . The wild-type strain was able to grow by utilizing all iron sources , while TMM001 was only capable of utilizing FeSO4 , the only iron source acquired by a TonB-independent process . In previous characterization studies of our acute respiratory murine inhalational glanders model , we observed that the 50% lethal dose using B . mallei strain ATCC 23344 was 7 . 4 x 104 CFU/50 μL ( Torres lab experimental data ) . To establish the role of tonB in B . mallei virulence , we challenged BALB/c mice intranasally ( i . n . ) with 1 . 5 x 105 CFU , 1 . 5 x 106 CFU and 1 . 5 x 107 CFU of TMM001 grown in LBG ± 200 μM FeSO4 and monitored them for survival up to day 14 . The Kaplan-Meier curve shows an inverse correlation between the dose and/or iron concentration and the mouse survival rate ( Fig 2 ) . Despite growth conditions , all BALB/c mice challenged with 1 . 5 x 107 CFU of TMM001 succumbed to infection by 4 days post challenge . At lower doses , the effect of supplementing TMM001 with 200 μM FeSO4 on survival was still apparent . At day 14 , survival increased from 62 . 5% to 100% and 0% to 12 . 5% when BALB/c mice received a challenge dose of 1 . 5 x 105 CFU and 1 . 5 x 106 CFU of the TMM001 , respectively , which was grown in LBG alone . We next enumerated bacterial counts in infected organs to determine the role of TonB in B . mallei’s ability to disseminate and colonize target tissues . BALB/c mice challenged i . n . with 1 . 5 x 104 CFU of the wild-type CSM001 or TMM011 grown in LBG ± 200 μM FeSO4 were euthanized at 24 , 48 and 72 h post challenge . At each time point , the lungs and spleen were processed and plated for CFU quantification . Compared to CSM001 , the numbers of TMM001 recovered from the lungs were significantly reduced at 24 h ( ★ p ≤ . 05 ) and 48 h ( ★★★★ p ≤ . 0001 ) , independent of growth conditions ( Fig 3A ) . A similar trend was observed in the spleen with significantly reduced numbers of TMM001 compared to the CSM001 at 24 h ( ★ p ≤ . 05 ) and 48 h ( ★★★ p ≤ . 001 ) ( Fig 3B ) . When grown in LBG + 200μM FeSO4 prior to challenge , TMM001 resembled CSM001 , showing no statistical difference in the number of bacteria recovered from the lungs . However , a statistical difference was seen in the recovery of TMM001 grown in FeSO4 in the spleen at 72 h ( ★ p ≤ . 05 ) ( Fig 3A and 3B ) . BALB/c mice challenged with the CSM001 expired before the 72 h time point and data are not presented . To evaluate the protective efficacy of TMM001 against CSM001 challenge , BALB/c mice received PBS , 1 . 5 x 104 CFU or 1 . 5 x 105 CFU of TMM001 ( grown in LBG only ) , via the i . n . route . At 21 days post-immunization , vaccinated mice were challenged i . n . with 1 . 5 x 104 CFU of CSM001 . The wild-type homolog CSM001 , containing a luminescent reporter , was used to assess the protective potential of TMM001 via real-time in vivo monitoring . All infected PBS-treated BALB/c mice died by day 4 , presenting with a calculated median survival of 3 days post challenge ( S3 Fig ) . In contrast , infected mice immunized with TMM001 at a dose of 1 . 5 x 105 CFU or 1 . 5 x 104 CFU showed 100% ( ★★★ p = 0 . 0003 ) and 87 . 5% ( ★★★ p = 0 . 0003 ) survival , respectively . Dissemination and colonization of CSM001 was monitored in TMM001-treated and naïve BALB/c mice using IVIS at 72 h post challenge and every 7 days thereafter until the experiment ended . At 72 h post challenge , PBS-treated BALB/c mice exhibited a luminescent signal associated with anatomical locations corresponding to the lungs , liver , spleen and brain . However , this signal was not detected at similar locations in BALB/c mice immunized with TMM001 ( S4 Fig ) . To evaluate whether TMM001 immunization resulted in the production of sterile immunity , BALB/c mice surviving the experimental challenge were euthanized and organs harvested to be analyzed for gross pathology and bacterial persistence . Although the lungs and livers showed no signs of evident pathology , BALB/c mice presented with splenomegaly accompanied by multiple splenic abscesses ( S5 Fig , panels D-F ) , which mirrors spleens at stage 3 of murine melioidosis infection , as we previously described [27] . Bacterial counts were only recovered from the spleens of mice immunized with a dose 1 . 5 x 105 ( 334 , 666 ± 70 , 465 CFU per spleen ) and 1 . 5 x 104 ( 61 , 917 ± 18 , 217 CFU per spleen ) CFU of TMM001 . Based on the phenotypic yellow pigment of the colonies , polymyxin B resistance and kanamycin sensitivity , we were able to conclude that all bacteria recovered were TMM001 and not CSM001 . In an attempt to eliminate persistence of the attenuated TMM001 strain , as well as to reduce organ pathology , an attenuated strain titration study was initiated to identify the lowest immunization dose that still provided 100% protection . The TMM001 titration study used the following CFUs for immunization: 1 . 5 x 104 , 1 . 5 x 103 and 1 . 5 x 102 . Twenty-one days post-immunization , three mice from each immunization group were euthanized , and organs and serum were harvested for histopathological and cytokine analysis . Forty-eight hours after CSM001 ( 1 . 5 x 104 CFU ) challenge , an additional 3 mice from each treatment were euthanized , and organs and serum were harvested for histopathological and cytokine analysis . As previously observed , all PBS-treated mice challenged with B . mallei CSM001 died by day 4 , with a median survival of 3 days ( Fig 4 ) . The titration curve exhibits a significant dose-dependent increase in survival in TMM001-treated mice challenged with CSM001 . All mice immunized with 1 . 5 x 102 CFU expired by day 15 , with an increased mean survival of 9 days ( ★★ p = 0 . 0016 ) . Mice immunized with 1 . 5 x 103 CFU or 1 . 5 x 104 CFU , survival up to 28 days increased to 62 . 5% ( ★★ p = 0 . 0016 ) and 100% ( ★★★ p = 0 . 00016 ) , respectively . Assessment of bacterial burden in surviving animals showed the spleen and , to a lower extent , the liver chronically infected in the TMM001-treated but not the CSM001 strain . The generation of murine humoral immune responses to B . mallei following treatment with mock ( PBS ) or TMM001 was determined by analysis of sera using ELISA . Compared to mock-vaccinated mice , sera from TMM001-treated mice had significantly higher titers of B . mallei-specific IgM and IgG antibodies ( Fig 5 ) . Mean differences in absorbance for IgG1 , IgG2a , and IgM were 5 . 4-fold ( p = 0 . 0009 ) , 4 . 8-fold ( p = 0 . 0106 ) , and 10 . 9-fold ( p = 0 . 0028 ) higher , respectively , in TMM001-vaccinated mice . The mouse tissues ( lungs , liver and spleen ) from the TMM001 titration study ( n = 3 per treatment ) at 0 h and 48 h post-challenge were processed for histology . Representative images of the lungs , liver and spleen from PBS- and TMM001 ( 1 . 5 x 104 CFU ) -immunized mice are presented in S6 Fig . At 0 h , the lungs , livers and spleens of PBS-treated mice were unremarkable , presenting as normal healthy tissue with normal architecture ( S6 Fig , panels A-C ) . BALB/c mice immunized with TMM001 presented with mild-to-moderate changes in pathology: perivascular and peribronchial inflammatory infiltrates in the lung sections ( S6 Fig , panel D ) , hepatitis with multifocal necrosis and scattered abscesses in the liver sections ( S6 Fig , panel E ) , and necrosis of follicles and accumulation of neutrophils in spleen sections ( S6 Fig , panel F ) . At 48 h post challenge with CSM001 ( 1 . 4 x 104 CFU ) , PBS-treated mice showed moderate-to-severe pathological changes , such as abscesses and multifocal inflammatory infiltrates in the lungs ( S6 Fig , panel G and Fig 6A ) , areas of hepatocellular necrosis , occasional abscesses with necrotic cores and areas of focal necrosis in the liver ( S6 Fig , panel H ) , and congestion of the red pulp , proliferation of large foamy macrophages ( inset of Fig 6 , panel C ) and necrosis affecting the mantle zone ( S6 Fig , panel I and Fig 6 , panel C ) . Similarly , TMM001-immunized mice showed moderate-to-severe changes in pathology , but with a few differences . In the lungs , large , multifocal inflammatory infiltrates , as well as abscesses , were present with focal consolidation observed as well ( S6 Fig , panel J and Fig 6 , panel B ) . The liver presented with hepatitis and multiple foci of hepatocellular necrosis ( S6 Fig , panel K ) , and large granulomas were formed in the spleen ( S6 Fig , panel L and Fig 6 , panel D ) . Histopathology scores showed significant differences due to treatment over time in the lungs ( ★★★★ p ≤ 0 . 0001 ) , liver ( ★★★★ p ≤ 0 . 0001 ) and spleen ( ★★ p ≤ 0 . 001 ) . When comparing the differences in treatment at 0 h and 48 h , the lungs ( ★ p = 0 . 05 ) , liver ( ★ p = 0 . 05 ) and spleen ( ★ p = 0 . 05 ) showed a robust trend toward significance ( Fig 6 , panels E-G ) . Overall , TMM001-immunization alone does cause some histopathology as evident by the histopathology at 0 h . That being said , PBS-immunized animals exuded more extensive pathology 48 h after CSM001 challenge compared to TMM001-immunized animals . Sera that was collected at 48 h post challenge from PBS- and TMM001-treated mice was used to identify pro-inflammatory cytokine and chemokine responses that correspond with disease outcome . Prior to challenge , a similar baseline expression of cytokines and chemokines was detectable in serum of representative animals from both the PBS- and TMM001-immunized animals ( Fig 7A ) . Following CSM001 challenge , the overall cytokine/chemokine expression increased markedly in both PBS- and TMM001-immunized animals ( Fig 7B ) compared to baseline , consistent with our previous observations of innate immune responses to Burkholderia species [27] . An attenuation of the pro-inflammatory serum cytokine/chemokine response to challenge was observed in the TMM001-treated compared to PBS control . The reduction of several pro-inflammatory mediators due to TMM001-treatment was significant , including IL-6 ( p = 0 . 049 ) , GM-CSF ( p = 0 . 037 ) , MCP-1 ( p = 0 . 022 ) , and RANTES ( p = 0 . 032 ) ( Fig 7 ) . A trend for reduction of several other pro-inflammatory IL-1β ( p = 0 . 097 ) , G-CSF , ( p = 0 . 067 ) and KC ( p = 0 . 05 ) due to TMM001-treatment was also observed ( Fig 7 ) . We next tested TMM001 for its protective potential against B . pseudomallei in an acute inhalational model of murine melioidosis . Mice received 1 . 5 x 104 CFU of TMM001 and at 21 days post-immunization , they were challenged with 9 . 0 x 102 CFU ( 3 LD50 ) of B . pseudomallei strain K96243 [26] . All PBS-treated BALB/c mice died by day 5 post-challenge and displayed a median survival of 5 days ( Fig 8 ) . In mice immunized with TMM001 , survival was increased to 75% ( ★★★ , p ≤ 0 . 001 ) at the end point of 36 days . As with the previous B . mallei study described above , the TMM001 strain , but not the wild-type B . pseudomallei , were recovered from immunized mice who presented with splenomegaly accompanied by abscesses .
To date , immune correlates of protection for B . mallei and B . pseudomallei are not clearly defined . Due to their intracellular lifestyle , these pathogens use an array of virulence factors to invade , replicate , and cause pathogenesis from within host cells , which can impede immune detection and , in some cases , protection . An extensive review of the literature suggested to us that an ideal vaccine for both pathogens would induce robust humoral and cell-mediated responses [1 , 36] . Thus , we decided to examine live attenuated vaccines , as these are often cited as the most efficacious approach to vaccine development against intracellular pathogens because they mimic natural infection , inducing both humoral and cell-mediated immunity , without causing disease . Moreover , exposure to the live attenuated strain allows the immune system to customize a protective response , in addition to generating an immune memory for lifelong protection against infection . In growth curve experiments , it was found that TMM001 was unable to maintain wild-type growth kinetics ( Fig 1 ) . Upon supplementing the culture with free iron , TMM001 exhibited increased growth rates more reminiscent of the wild-type , which is illustrated by a shorter lag phase and prolonged maintenance of wild-type growth kinetics . In a separate growth curve study , full rescue of the wild type phenotype in TMM011 was achieved after both the starter and sub-culture were supplemented with free iron . The correlation between free iron concentration and the growth rate of TMM001 illustrates the importance of TonB as a facilitator of iron transport , which has a direct impact on bacterial fitness . The results of our survival study show an inverse correlation not only between TMM001 dose and survival but also between concentration of free iron and survival ( Fig 2 ) . Compared to the wild-type B . mallei strain ( LD50 of 7 . 5 x 104 CFU ) , TMM001 is approximately 3-fold more attenuated when grown with FeSO4 ( LD50 of 2 . 38 x 105 CFU ) ; and when grown in LBG alone , the tonB mutant is attenuated by approximately 7-fold ( 5 . 59 x 105 CFU ) . Differences in virulence are consistent with the data of the dissemination study which showed the lowest burden in animals infected with TMM001 and higher bacterial burdens in animals infected with TMM001 grown with FeSO4 ( Fig 3 ) . In both experiments , FeSO4 supplementation failed to fully reverse attenuation of TMM001 in vivo . This outcome was not unexpected as the concentration of free iron in the host ( 10-24M [37] ) is well below that is needed to sustain bacterial replication . While FeSO4 supplementation would prolong its survival and therefore increase its virulence , TMM001 would be unable to sustain the wild-type phenotype once its internal stores of iron are exhausted . Lack of full complementation by FeSO4 supplementation can also be attributed to the role of TonB in the import of other substrates . While the majority of TonB-dependent transport systems function to uptake iron , vitamin B12 , nickel chelates , and carbohydrates can also be transported by this mechanism . Overall , decreased mortality observed in animals challenged with the TMM001 grown in LBG alone illustrates the importance of iron and its TonB-mediated acquisition to virulence . In a series of TMM001 titration studies , it was empirically determined that a dose of 1 . 5 x 104 CFU of TMM001 resulted in 100% protection ( Fig 4 ) and CSM001 clearance following challenge . Protected animals developed strong B . mallei-specific IgG1 , IgG2a , and IgM responses ( Fig 5 ) , which we attribute to TMM001-mediated protection . The observation and correlation of strong IgG and IgM elicitation and protection are cited often in Burkholderia vaccine studies [38 , 39 , 40 , 41] . In human cases , it was found that patients with less severe , localized infection produced detectable Burkholderia-specific IgM antibody titers , whereas none were detected in patients suffering from acute disseminated infection [39] . Thus , it is plausible to suggest TMM001 treatment protects against lethal infection by neutralizing bacteria and/or preventing their dissemination to target organs via antibody-mediated mechanisms . TMM001 immunization resulted in pathological differences that may explain increased survival and protection . In general , histopathological scoring shows a robust trend toward significant differences in the pathology seen in the lungs , liver and spleen of PBS- vs . TMM001-immunized animals ( Fig 6 , panels E-G ) . Further analysis of these tissues revealed two discriminatory elements of pathologic damage between vaccine treatments . First , despite the finding that the lungs and livers from both PBS- and TMM001-immunized animals displayed some degree of tissue damage , the pathological changes in TMM001-immunized mice were much less severe ( S6 Fig , panel A ) . Second , the differential alteration in spleen architecture implied that the PBS- and TMM001-immunized animals responded differently to infection . For example , splenic tissues from PBS-treated mice show a diffuse response to injury ( i . e . diffuse severe histiocytosis ) , while splenic tissue from TMM001-immunized mice showed a focal response to injury ( i . e . granuloma formation ) ( S6 Fig , panels C-D ) . These histological observations suggested to us that immunization with TMM001 may result in the induction of an immune response that produces a different type of tissue damage , in addition to confining infection to prevent disseminated disease , an important cause of morbidity and mortality in many diseases [39 , 42 , 43 , 44] . The histopathological differences between PBS- and TMM001-vaccinated mice suggest that TMM001 reduces disease by attenuating immune-mediate pathology at sites of bacterial proliferation . Our observation that pro-inflammatory cytokine/chemokine responses are attenuated following TMM001 treatment further supports this conclusion . In models of murine melioidosis , it has been established that increased expression of IL-1β and IL-6 follow B . pseudomallei dissemination and coincide with acute sepsis and mortality [45 , 46] . Clinical evidence further suggests a correlation between elevated serum levels of IL-1β and IL-6 and poor prognosis in patients with septic melioidosis [47 , 48 , 49] . Our previous studies demonstrated that pre-treatment with CpG oligonucleotides protected mice from B . pseudomallei exposure and reduced pro-inflammatory cytokine/chemokine ( e . g . IL-1β , IL-6 , G-CSF , KC , MCP-1 ) expression in the lung [35] . In the study by Judy , et al . , a moderate pro-inflammatory response was associated with protection while excessive inflammation caused pulmonary pathology [26] . Further , the protective effects of CpG treatment to reduce lung pathology were attributed to a reduction in neutrophil and inflammatory monocyte recruitment [26 , 50] . Similarly , we have previously shown that the virulence of B . pseudomallei strains in direct comparisons corresponds with excessive production of pro-inflammatory cytokines and chemokines that recruit neutrophils and monocytes [27] . These observations in clinical and animal model studies support a role for exacerbated pro-inflammatory responses to mediate lung pathology in disease due to Burkholderia species . Thus , treatment with TMM001 activates a moderate pro-inflammatory cytokine/chemokine response associated with protective immune responses and attenuates the exacerbation of this response that is associated with neutrophil infiltration and immune-mediated tissue damage . Lastly , since B . mallei and B . pseudomallei are genetically closely related , TMM001 was further tested for its potential to provide protection in an acute inhalational model of murine melioidosis . The significant cross protection seen in TMM001-treated mice provides an optimistic outlook for the development for a single vaccine for both pathogens . Immunization with TMM001 resulted in full protection and clearance of CSM001 when tested in an acute respiratory model of murine glanders . This live attenuated strain is unique not only because it provided full protection against both acute and chronic stages of infection but also because it imparted significant cross protection against B . pseudomallei infection . It is hypothesized that the persistence of viable bacteria is key for protective potential . In previous vaccination studies , the failure to provide successful long-term protection has often be attributed to the quick removal of live attenuated strains from the host [51 , 52] . Thus , we believe the resulting long-term protection is link to the ability of TMM001 to evade rapid clearance . This notion is supported by vaccination studies which reported long-term survivors to be generally colonized at the end of the study [51 , 52 , 53] . It is plausible to propose that this persistence increases the accessibility of the immune system to protective antigens or it might contribute to the development of an environment adverse to wild-type colonization via chronic elicitation of the immune response . Although its persistence is important for its protective potential , TMM001 is able to colonize and maintain in the host . Before securing approval from the division of select agents and toxin ( DSAT ) for removal of this strain from the Health and Human Services select agent list and becoming a legitimate vaccine candidate , the ability of TMM001 to cause chronic infection needs to be addressed . Studies are now focused on using TMM001 as a backbone to generate a further attenuated strain with the introduction of additional mutations . As TMM001 is only 7-fold less virulent than the wild-type B . mallei strain , we believe this to be the best strategy for optimization and do not anticipate problems with over-attenuation . Currently we are targeting genes that are contributing to persistence with the intention of developing a more attenuated strain that can persist long enough to elicit a protective immune response without establishing chronic infection . For example , we are focused on genes involved in TonB-independent mechanisms of iron assimilation . Bacterial transport systems that are shown to transport iron in this manner include the following: FbpABC transport system of Neisseria gonorroeae¸ SfuABC transport system of Serratia marcescens , VctPDGC ABC cytoplasmic membrane transport system of Vibrio cholera , etc [Ref] . Looking for homologs of these systems in B . mallei could provide optimal targets for further attenuation . Overall , we believe the present study represents a significant advancement in the battle against pathogenic Burkholderia infections , in which TMM001 could be further optimized to become an effective vaccine against glanders , melioidosis , or other Burkholderia infections . | Burkholderia mallei and B . pseudomallei are the causative agents of glanders and melioidosis , respectively . In addition to the recent rise in cases of glanders and the endemicity of melioidosis worldwide , these pathogens have gained attention as potential bioweapons . Further , these pathogens have huge potential for aerosol delivery and often produce fatal infection amongst untreated individuals . Both pathogens are difficult to treat , and even with antibiotic intervention , patients relapse or get re-infected . A big challenge for vaccine development against these pathogens includes identification of broadly protective antigens and a better understanding of the correlates of protection from both acute and chronic infections . Our study is the first to demonstrate significant protection against a lethal challenge with both Burkholderia species . Because TMM001 persists in immunized mice , we propose that this attenuated organism is a promising backbone-based strain from which a legitimate vaccine candidate can be generated . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Characterization of the Burkholderia mallei tonB Mutant and Its Potential as a Backbone Strain for Vaccine Development |
Type IV pili are expressed by a wide range of prokaryotes , including the opportunistic pathogen Pseudomonas aeruginosa . These flexible fibres mediate twitching motility , biofilm maturation , surface adhesion , and virulence . The pilus is composed mainly of major pilin subunits while the low abundance minor pilins FimU-PilVWXE and the putative adhesin PilY1 prime pilus assembly and are proposed to form the pilus tip . The minor pilins and PilY1 are encoded in an operon that is positively regulated by the FimS-AlgR two-component system . Independent of pilus assembly , PilY1 was proposed to be a mechanosensory component that—in conjunction with minor pilins—triggers up-regulation of acute virulence phenotypes upon surface attachment . Here , we investigated the link between the minor pilins/PilY1 and virulence . pilW , pilX , and pilY1 mutants had reduced virulence towards Caenorhabditis elegans relative to wild type or a major pilin mutant , implying a role in pathogenicity that is independent of pilus assembly . We hypothesized that loss of specific minor pilins relieves feedback inhibition on FimS-AlgR , increasing transcription of the AlgR regulon and delaying C . elegans killing . Reporter assays confirmed that FimS-AlgR were required for increased expression of the minor pilin operon upon loss of select minor pilins . Overexpression of AlgR or its hyperactivation via a phosphomimetic mutation reduced virulence , and the virulence defects of pilW , pilX , and pilY1 mutants required FimS-AlgR expression and activation . We propose that PilY1 and the minor pilins inhibit their own expression , and that loss of these proteins leads to FimS-mediated activation of AlgR that suppresses expression of acute-phase virulence factors and delays killing . This mechanism could contribute to adaptation of P . aeruginosa in chronic lung infections , as mutations in the minor pilin operon result in the loss of piliation and increased expression of AlgR-dependent virulence factors–such as alginate–that are characteristic of such infections .
Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen , recently listed as one of the highest priority antimicrobial-resistant threats by the World Health Organization , due to its intrinsic antibiotic resistance and recalcitrance to therapy [1] . Among its virulence factors are filamentous surface appendages called type IV pili ( T4P ) , sophisticated biological nanomachines that are broadly distributed among bacteria and archaea [2 , 3] . In P . aeruginosa , T4P facilitate surface and host cell adhesion , colonization , biofilm maturation , virulence , and twitching , a form of surface-associated motility facilitated by cycles of extension , adhesion , and retraction of T4P fibres [3–11] . T4P are composed of hundreds to thousands of copies of small proteins called major pilins ( PilA in P . aeruginosa ) along with the low abundance minor pilins ( MPs ) FimU-PilVWXE [12–16] . The MPs are encoded in a polycistronic operon with the pilY1 gene that codes for a large ~125 kDa non-pilin protein . The operon is positively regulated by the virulence factor regulator Vfr , and the two-component system ( TCS ) FimS ( AlgZ ) -AlgR . FimS is a predicted histidine sensor kinase while AlgR is a response regulator that promotes expression of genes important for biofilms and chronic cystic fibrosis ( CF ) lung infections [17–21] . The N-termini of immature pilins are cleaved and methylated at the cytoplasmic face of the inner membrane by the prepilin peptidase , PilD , while PilY1 may be processed by signal peptidase 1 [22–25] . Mature pilins are polymerized into a T4P fibre via an envelope-spanning assembly machinery , where individual PilA subunits are added or removed at the platform protein , PilC , via action of the ATPases PilB and PilT , respectively [2 , 26] . The MPs and PilY1 are required for T4P function in several bacterial species , including P . aeruginosa , Escherichia coli , Neisseria meningitidis , N . gonorrhoeae , and Myxococcus xanthus [12–15 , 27–30] . PilY1 and the MPs were originally proposed to oppose pilus retraction , as a few surface pili remain in pilY1 and MP mutants when retraction is blocked via deletion of pilT [23 , 28 , 29 , 31 , 32] . We recently showed that when T4P MPs are missing , the equivalent minor pseudopilins of the Xcp type II secretion system can pilus prime extension in the pilT background , and that deletion of both sets of minor components abolishes pilus assembly [24] . We also demonstrated that PilY1 and the MPs are present in sheared pili , and that the loss of PilV , PilW , PilX , or PilY1 excludes the other three components from the pilus [24] . Thus , PilVWXY1 are proposed to form a core assembly-initiation subcomplex , while FimU and PilE are thought to connect this complex to PilA . Initiation of assembly with subsequent addition of multiple PilA subunits would place the MPs at the pilus tip , with PilY1 –the largest component–at the distal position , supporting the hypothesis that PilY1 is a T4P-associated adhesin [31] . PilY1 and the MPs ( and their regulators FimS-AlgR ) are required for T4P biogenesis , and therefore T4P-mediated functions [12–15 , 17 , 19] . However , recent studies hinted at more enigmatic roles of PilWXY1 in virulence . Bohn et al . [33] showed that in a non-piliated P . aeruginosa background , subsequent loss of pilY1 reduced virulence in a Caenorhabditis elegans fast killing assay and in a mouse airway infection model , and increased resistance to killing by neutrophils . Thus , PilY1 has a role in virulence that does not require functional pili . Other studies using C . elegans infection models suggested that MP and pilY1 mutants had attenuated virulence relative to WT , and in one case , to a non-piliated mutant [34–37] . Recently , Siryaporn et al . [38] showed that PilWXY1 were required for surface-activated virulence towards amoebae , while other non-piliated mutants had WT virulence . The N-terminal region of PilY1 has weak sequence similarity to the eukaryotic von Willebrand factor A ( VWFa ) domain , which can be deformed by shear forces [39] . In-frame deletion of this domain from PilY1 allowed normally avirulent planktonic cells to kill amoebae [38] . PilY1 was therefore proposed to be a mechanosensor , where deformation of its VWFa domain upon surface interaction led–by an as-yet unknown mechanism–to increased expression of virulence factors . One important caveat of that study was that an algR mutant ( which lacks PilY1 and the MPs ) had WT virulence towards amoebae [38] . Deformation of PilA subunits by tensile forces acting upon surface-attached pili was also proposed as a possible way to signal attachment . Detection of partly unfolded pilins by the Pil-Chp chemotaxis system could lead to increased cyclic adenosine monophosphate ( cAMP ) synthesis via the CyaB adenylate cyclase [40 , 41] . cAMP is bound by Vfr , a key transcription factor that promotes expression of virulence factors involved in motility , attachment , and secretion [20 , 40 , 41] . fimS-algR transcription is activated by Vfr , leading to increased transcription of fimU-pilVWXY1E [40] . PilVWXY1 were proposed to repress their own expression in an AlgR-dependent manner , as the loss of pilV , pilW , pilX , or pilY1 led to elevated expression of the MP operon and fimS-algR [23 , 33 , 38 , 40] . The mechanism of this putative feedback inhibition is largely uncharacterized , but was speculated to involve FimS [40] . Once expression of the MP operon is activated , extracellular PilY1 may sense surface association and transduce this information through the T4P assembly machinery [38 , 40] . This signal is thought to activate an inner membrane-localized diguanylate cyclase , SadC , to increase levels of c-di-GMP , promoting expression of genes associated with a biofilm lifestyle , while repressing early-phase virulence traits such as swarming motility [40 , 42] . This model was supported by studies demonstrating that loss of pilW , pilX , or pilY1 in a high-c-di-GMP background resulted in hyper-swarming and reduced c-di-GMP levels , as measured by liquid chromatography-mass spectrometry of extracts from surface-grown cells [39 , 43] . Rodesney et al . [44] showed that c-di-GMP levels increased in response to shear forces , and that functional T4P were required for this phenomenon , further supporting this hypothesis . However , unlike pilW , pilX , and pilY1 mutants , a sadC mutant had WT virulence towards amoebae , suggesting the PilWXY1-SadC pathway may be important for surface sensing , but not necessarily for surface-activated virulence [38] . Although PilY1 and the MPs clearly influence virulence , the underlying mechanism remains to be established [33–36 , 38 , 45] . We hypothesized that a subset of these components represses FimS activity , such that loss of pilW , pilX , or pilY1 activates FimS-AlgR , shifting the bacteria to a less pathogenic phenotype typically associated with chronic infection . We found that slow killing ( SK ) of C . elegans by pilW , pilX , and pilY1 mutants was significantly delayed compared to WT or a pilA mutant , and this delay was dependent on FimS-AlgR , because double mutants had WT killing kinetics . Hyperactivation ( via phospho-mimetic point mutation ) or overexpression of AlgR alone was sufficient to delay killing . Together , these data are consistent with a model where loss of PilWXY1 relieves feedback inhibition on expression of the AlgR regulon , resulting in dysregulation of virulence factors that are important for C . elegans pathogenesis .
Specific genes in the MP operon were reported to be important for virulence in amoebae , nematodes , and mouse models , but those studies were done using different strains of P . aeruginosa [33–36 , 38 , 45] . We first confirmed these results in the C . elegans SK model , using two well-studied strains . SK assays were performed using PA14 with deletions of pilA , fimU , pilV , pilW , pilX , pilY1 , or pilE ( Fig 1A ) . An E . coli OP50 plate was included as a negative control for pathogenicity; worms began to senesce on these plates around day 7–8 , consistent with published data regarding temperature-dependent effects on lifespan [46] . As worms at later time points were at increased risk of death due to ageing in addition to P . aeruginosa infection , statistical significance was assessed using the Gehan-Breslow-Wilcoxon test , which places greater weight on earlier time points [47] . A pilA ( major pilin ) mutant was slightly less pathogenic than WT; subsequent comparisons were made relative to pilA , since all mutants lack pili . fimU and pilE mutants were more pathogenic than the pilA mutant , similar to WT . In contrast , pilW , pilX , and pilY1 mutants were less pathogenic than the pilA mutant , suggesting that delayed killing was not due to loss of functional T4P . Virulence of the pilV mutant was similar to the pilA mutant . The twitching and virulence defects of pilW , pilX , and pilY1 mutants could be partially complemented by expression of the relevant gene in trans ( S1 Fig ) . The stoichiometry of PilY1 and the MPs is important for optimal T4P function [23] , which may explain the lack of full complementation . To verify that these phenotypes were not strain-specific , we tested PAO1 transposon-insertion mutants of pilA , fimU , pilV , pilW , pilX , pilY1 , and pilE in the SK assay ( Fig 1B ) . Similar to the results in PA14 , PilWXY1 were important for T4P-independent virulence . However , the fimU and pilV mutants killed nematodes more slowly than pilA; the PA14 and PAO1 MPs are divergent ( 61–75% amino acid similarity ) , so it is possible that FimU and PilV function slightly differently in PAO1 versus PA14 [48] . To focus on genes that were generally important for virulence of P . aeruginosa , we undertook studies of the mechanism responsible for delayed killing of C . elegans by the pilW , pilX , and pilY1 mutants . PilWXY1 were previously proposed to increase c-di-GMP production by SadC , such that loss of pilW , pilX , or pilY1 resulted in a biofilm-deficient phenotype , indicative of low intracellular c-di-GMP [39 , 40 , 43] . Therefore , we hypothesized that biofilm defects of pilW , pilX , and pilY1 might impede their ability to colonize the C . elegans gut , leading to delayed killing . The PA14 and PAO1 parent strains and their cognate pilA , fimU , pilV , pilW , pilX , pilY1 , and pilE mutants formed negligible levels of biofilm in liquid SK medium , chosen to approximate the growth conditions used for the SK assay ( S2 Fig ) . To assess the levels of cyclic-di-GMP in these strains , we constructed a luminescence-based cdrA promoter reporter based on an extensively-characterized green fluorescent protein-based reporter system [44 , 49–54] . cdrA promoter activity has been positively correlated with c-di-GMP levels , as measured by liquid chromatography-mass spectrometry [49 , 51 , 53 , 54] . We verified that overexpression of SadC led to a ~60-fold increase in cdrA promoter activity , while overexpression of AlgR , which positively regulates genes that promote c-di-GMP production [55 , 56] , led to a ~2-fold increase in promoter activity that was enhanced to ~4-fold when algR expression was increased with 0 . 05% L-arabinose ( Fig 2A ) . Deletion of sadC or algR led to a ~2-fold decrease in cdrA promoter activity relative to WT . cdrA promoter activity in WT is expected to be relatively low in liquid media because c-di-GMP levels increase upon surface attachment [43] . Compared to WT , pilW , pilX , and pilY1 had ~3-fold lower cdrA promoter activity , indicative of reduced c-di-GMP ( Fig 2B ) . These results are consistent with reports that PilWXY1 promote c-di-GMP production via SadC [39 , 40 , 43] . We next investigated whether SadC was required for virulence , as would be predicted if decreased virulence in pilW , pilX , and pilY1 mutants was due to dysregulation of SadC activity . A small decrease in virulence towards C . elegans was previously reported for a PA14 sadC mutant [57]; however , we saw no difference between WT and sadC mutants in the PA14 and PAO1 backgrounds ( S3 Fig ) . Further , overexpression of SadC led to a hyper-biofilm phenotype in vitro in SK medium , but a slight delay in killing , demonstrating that the amount of biofilm formed in vitro does not correlate with virulence in C . elegans ( Fig 3 ) . Although the exact mechanisms of P . aeruginosa pathogenesis in C . elegans are not fully understood , biofilms were suggested to be important for establishment of infection [57–59] . Our in vitro data suggests that biofilms may not be a major contributor to P . aeruginosa pathogenesis in this model , but direct visualization and quantification of biofilms in the nematode gut will be required to support this conclusion . After ruling out involvement of the SadC pathway , we explored the potential role of FimS-AlgR in PilWXY1-mediated modulation of killing kinetics . Informed by previous work in our laboratory showing that the sensor kinase PilS of the PilSR TCS interacts directly with PilA in the inner membrane to decrease PilR-dependent major pilin expression [60] , we hypothesized that FimS interacts with one or more MPs , and that loss of that interaction could lead to activation of AlgR and subsequent upregulation of the MP operon . Bacterial two-hybrid ( BACTH ) assays were used to identify potential interactions between FimS and PilA , FimU , PilV , PilW , PilX , or PilE ( Fig 4A ) . We also screened for interaction of FimS and AlgR , which has been inferred but never demonstrated [19] . Interactions between FimS and each pilin were identified; however , based on our experience with PilS [60] , binding is necessary but not sufficient for inhibition . We also demonstrated interaction of FimS and AlgR ( Fig 4A ) , providing further support for the hypothesis that FimS is the sensor kinase for AlgR . To decipher which MPs might modulate expression of the operon , we monitored expression from the fimU promoter using a luxCDABE reporter . Compared to WT PA14 , there was a ~25-fold increase in luminescence in pilV , pilW , pilX , and pilY1 mutants , which was restored to WT by expressing the corresponding pilin in trans ( Fig 4B , S4 Fig ) . fimU and pilA mutants had ~5-fold increased promoter activity , while that of a pilE mutant was comparable to WT . fimS and algR mutants had low baseline luminescence , ~10-fold lower than WT . To determine whether the increased promoter activity in pilV , pilW , pilX , and pilY1 mutants depended on FimS-AlgR , either fimS or algR was deleted in the pilY1 mutant background . The luminescence was ~10-fold lower than WT in the pilY1 algR double mutant , consistent with AlgR acting as a positive regulator of the MP operon [40] . Loss of fimS in the pilY1 mutant background also abolished fimU promoter activity ( ~10-fold lower than WT ) , supporting the idea that FimS may monitor PilVWXY1 and activate AlgR when their levels drop . Based on these data , PilA , FimU , and PilE are unlikely to modulate FimS-AlgR activity even though they can interact with FimS . PilVWXY1 were previously proposed to form a complex in the inner membrane , such that loss of any one component destabilizes the others [24] . Since PilY1 is thought to be cleaved on the periplasmic side of the inner membrane , it is unlikely to interact directly with the transmembrane domains of FimS [24] . Thus , we suspected that high fimU promoter activity in the pilY1 mutant was due to reduced levels of one or more of the other pilins . To address this , we overexpressed FimU , PilV , PilW , PilX , or PilE in the pilY1 mutant and measured fimU promoter activity . All these strains had luminescence comparable to the pilY1 mutant ( S4 Fig ) . Conversely , distinct effects have been observed in other studies upon overexpression of PilY1 [39 , 40 , 43] . Therefore , we overexpressed PilY1 in the pilW and pilX ( high-luminescence ) backgrounds; but PilY1 alone was insufficient to alter fimU promoter activity . Together , the data suggest that no individual component of the PilVWXY1 subcomplex is capable of modulating FimS activity when others are absent . We also tested whether PilD processing of PilVWX was required for modulation of FimS activity . We constructed a pilD mutant , which lacks twitching motility because unprocessed pilins remain in the inner membrane [23 , 61] . The absence of pilD had no impact on fimU promoter activity ( S5 Fig ) , and a pilD mutant had virulence equivalent to a pilA mutant , likely attributable to its lack of T4P . Thus , PilVWX can modulate FimS activity in their unprocessed form . Because the results suggested that loss of PilWXY1 relieves feedback inhibition on FimS-AlgR , resulting in AlgR activation , we tested whether hyperactivation of AlgR alone could delay killing of C . elegans . We made chromosomal algRD54E phospho-mimetic point mutants [62] in both PA14 and PAO1 backgrounds . We also made algRD54A point mutants , as AlgR phosphorylation is required for transcription of a subset of genes in its regulon , including the MP operon [17 , 62 , 63] . We verified that the algRD54A mutant was defective for twitching motility , while the algRD54E mutant had WT twitching ( S6 Fig ) . Unexpectedly , a fimS mutant retained ~50% twitching motility , in contrast to previous reports [18 , 62] . In the absence of FimS , AlgR might be phosphorylated by small phosphate donors [64] . Based on the fimS data , we also questioned the assumption that AlgR phosphorylation was necessary for expression from the fimU promoter . When we overexpressed WT AlgR or AlgRD54A in the algR mutant ( S6 Fig ) , its twitching defect was fully complemented by AlgR , and partially complemented ( 25% ) by AlgRD54A . Thus , although it increases binding to the fimU promoter [17 , 62] , phosphorylation of AlgR is not essential for transcription of the MP operon . SK assays were then performed for PA14 and PAO1 algRD54A and algRD54E mutants , plus PA14 and PAO1 fimS and algR deletion mutants . PA14 and PAO1 algRD54E mutants killed more slowly than the corresponding WT strains , while fimS , algR and algRD54A mutants had WT virulence ( Fig 5A and 5B ) . Loss of FimS-AlgR decreases expression of the MPs and PilY1 and prevents pilus assembly [17 , 40] . Because our data show that loss of FimS-AlgR ( and thus MP expression ) had no impact , we conclude that delayed killing of nematodes by pilW , pilX , and pilY1 mutants is due to inappropriately timed FimS-AlgR activation . Increased transcription of fimS-algR in a pilY1 mutant relative to WT has been reported [38] , suggesting that delayed killing could arise through expression of increased amounts of the FimS-AlgR TCS , as well as its activation . Therefore , we asked whether increased AlgR levels would attenuate virulence , as previously demonstrated in a mouse infection model [65] . When algR was expressed in trans from a multicopy plasmid in PA14 algR , killing was delayed compared to the vector control ( Fig 6A ) . Because un-phosphorylated AlgR can also affect transcription of a subset of genes [66 , 67] , we tested the same mutant complemented with AlgRD54A . Complementation of the algR mutant with AlgRD54A resulted in a severe delay in killing relative to the vector-only control . Thus , AlgR hyperactivation and overexpression independently diminish P . aeruginosa virulence towards C . elegans . Lastly , as AlgR is a positive regulator of biofilm formation [17 , 55 , 56] , we performed biofilm assays for PA14 algR complemented with AlgR or AlgRD54A . Expression of either variant led to hyper-biofilm formation ( Fig 6B ) , further emphasizing that the ability of a strain to form biofilms in SK medium does not correlate with virulence in worms . Instead , we suggest that virulence factors repressed by FimS-AlgR are important for C . elegans SK , and an increase in AlgR levels and/or activity at the wrong time delays killing . To provide further support for this model , we asked whether the virulence defects of PA14 pilW , pilX , and pilY1 mutants required FimS-AlgR . We deleted fimS or algR in the pilW , pilX , and pilY1 backgrounds , and tested virulence of the double mutants ( Fig 7 ) . We also deleted pilW , pilX , and pilY1 in the algRD54A background , to test if AlgR activation was necessary for the delayed killing by pilW , pilX , and pilY1 mutants . In all cases , the double mutants had WT virulence , equivalent to that of the fimS , algR , or algRD54A single mutants . These results demonstrate that the delay in killing that results from loss of PilWXY1 requires both FimS and AlgR . Although overexpression of AlgRD54A in trans repressed virulence ( Fig 6A ) , the chromosomal mutation was sufficient to alleviate delayed killing by pilW , pilX , and pilY1 mutants , suggesting that AlgR phosphorylation is important for modulation of virulence when PilWXY1 are missing . The sigma factor AlgU ( AlgT/σ22/σE ) acts upstream of FimS-AlgR to promote algR transcription [68–70] , thus we tested its potential involvement in modulation of virulence by PilWXY1 . An algU mutant killed more rapidly than WT ( Fig 8 ) , as previously demonstrated in mouse models [71] , while pilW algU , pilX algU , and pilY1 algU double mutants had near-WT virulence ( less than an algU mutant , but more than pilW , pilX , and pilY1 single mutants ) . Although AlgU promotes algR transcription [69] , loss of AlgU alone does not prevent AlgR expression [68] . Given the reduced virulence of the pilW algU , pilX algU , and pilY1 algU double mutants relative to algU , PilWXY1 modulation of FimS-AlgR signalling appears to be intact in the algU mutant . These data are consistent with studies showing that mucA and mucD mutants , in which algR and algU are highly transcribed [69 , 72–74] , are less virulent towards C . elegans [75–77] .
P . aeruginosa uses T4P to attach to surfaces and host cells , for biofilm maturation , and to move across surfaces via twitching motility [2] . The MPs and PilY1 are important players in T4P biogenesis and function , but also in regulation of swarming motility , surface attachment , mechanosensation , and virulence [38–40 , 43] . The MP operon is positively regulated by FimS-AlgR , a TCS implicated in regulation of chronic P . aeruginosa lung infections [17–19] . Here , we explored the connection between loss of PilWXY1 ( and thus , loss of T4P ) and AlgR activation in virulence towards C . elegans , as summarized in Fig 9 . We showed that pilW , pilX , and pilY1 mutants kill nematodes more slowly than WT or a pilA mutant , supporting the idea that PilWXY1 modulate virulence independently of their role in T4P assembly . We confirmed previous reports [23 , 33 , 40] that in the absence of pilV , pilW , pilX , or pilY1 , expression of the MP operon is significantly increased , and that this requires FimS-AlgR . Either hyperactivation or overexpression of AlgR delayed killing , while loss of fimS or algR in pilW , pilX , or pilY1 reverted virulence to WT levels . These data–coupled with BACTH data showing that the MPs interact directly with FimS in the inner membrane ( Fig 4 ) –suggest that FimS may act as a molecular thermostat to monitor MP levels , and in their absence , activates AlgR to upregulate expression of the MP operon . A similar inventory control mechanism was recently described for the PilSR TCS , where PilS phosphorylates PilR when PilA levels are low , and dephosphorylates PilR when PilA levels are high [60] . It is not yet clear if FimS responds to changes in levels of the PilVWXY1 subcomplex , thought to prime assembly of T4P [24 , 78 , 79] . When overexpressed individually in trans , each of the MPs inhibited twitching motility in PAO1 [23] , but since the others were still expressed from the chromosome , the exact nature of the signal detected by FimS remains to be determined . When expressed in trans , no single component of the PilVWXY1 subcomplex reduced fimU promoter activity if others were absent ( S4 Fig ) . The specific signal that inhibits FimS activity remains to be deciphered . Whether the FimS-inhibitory signal is the same in PA14 and PAO1 also remains unknown . Though PilWXY1 modulated virulence of PA14 and PAO1 , FimU and PilV influenced virulence only in PAO1 ( Fig 1A and 1B ) . Given the MPs are divergent , FimU and PilV may play different roles in PAO1 versus PA14 [48] . It is possible that FimU and PilV are more important for stability of the PilWXY1 subcomplex in PAO1 than in PA14 , and/or that PAO1 FimU and PilV can directly modulate FimS activity . Kuchma et al . [39 , 43] reported that loss of pilW , pilX , or pilY1 increased swarming motility and decreased biofilm formation , both indicative of low c-di-GMP levels . As biofilms were proposed to contribute to P . aeruginosa pathogenesis in C . elegans , we investigated whether the reduction in virulence in the absence of PilWXY1 was linked to decreased biofilm via loss of SadC activation [57–59 , 80] . In our hands , levels of sadC had no impact on virulence even though they clearly modulated the amount of biofilm produced in SK media ( Fig 3A and 3B , S3 Fig ) . Irazoqui et al . [59] examined the C . elegans gut during P . aeruginosa infection and described extracellular material that they suggested might indicate presence of a biofilm . Anti-biofilm compounds reduced P . aeruginosa virulence towards C . elegans , but a mechanism of action for those compounds has not been described [58] . Recently , the small RNA SrbA was shown to modulate both biofilm and virulence towards C . elegans; however , deletion of srbA led to altered transcription of at least 26 other genes that may also affect virulence [81] . Rather than using standard biofilm media , we performed these assays in liquid SK media to more closely mimic the conditions to which bacteria are exposed in the SK assay . To our knowledge , this is the first report to use SK media for biofilm assays . As we found no correlation between biofilm formation and virulence , we suggest that acute-phase virulence factors may be more important for C . elegans pathogenesis in the SK model . However , we recognize that in vitro biofilm assays may not replicate the conditions within the C . elegans gut; direct visualization of bacteria in worms will be needed to clarify the role of biofilm formation . PilY1 and the MPs have been implicated in surface detection and activation of virulence , via signalling through SadC [38 , 40] . Because loss of PilY1 or the MPs prevents T4P assembly and function , it is crucial to distinguish phenotypes resulting from lack of specific proteins versus loss of piliation [24] . Luo et al . [40] suggested that association of PilY1 with surfaces transduces a signal through the T4P machinery to stimulate c-di-GMP production by SadC , while Rodesney et al . [44] showed that loss of pilA , pilY1 , or pilT prevents surface-activated c-di-GMP production . Rodesney et al . [44] proposed that both PilY1 and functional T4P are required for mechanosensation; however , it is not possible to delete pilY1 without ablating T4P assembly . Our cdrA promoter reporter data support the idea that PilWXY1 promote cyclic-di-GMP production by SadC , as loss of pilW , pilX , or pilY1 decreased cdrA promoter activity ( Fig 2B ) . However , we argue that the PilWXY1-SadC pathway–though important for c-di-GMP signalling–is not critical for virulence towards C . elegans . Instead , our data show that PilWXY1-FimS-AlgR signalling axis is responsible for T4P-independent changes in virulence of pilW , pilX , and pilY1 mutants . Thus , surface attachment may induce c-di-GMP production via PilWXY1-SadC [40 , 43] , while the brief trapping of T4P outside the cell upon contact with a surface might transiently deplete PilVWXY1 levels in the inner membrane , resulting in increased FimS-AlgR activity and transition towards a sessile , biofilm lifestyle . Whether the loss of pilW , pilX , or pilY1 leads to increased amounts of AlgR , its increased phosphorylation via FimS , or both , remains to be clarified . Okkotsu et al . [62] showed that AlgR and AlgRD54E levels are comparable , suggesting that the delay in killing we observed for PA14 algRD54E is attributable to the D54E phospho-mimetic mutation alone . Overexpression of AlgRD54A in trans delayed killing ( Fig 6A ) , but the same mutation on the chromosome reverted virulence of pilW , pilX , and pilY1 mutants to WT levels ( Fig 7 ) . Therefore , we suspect that it is primarily AlgR phosphorylation ( or lack of AlgR dephosphorylation ) that leads to delayed killing . However , it is possible that both increased AlgR protein levels and phosphorylation contribute . Kong et al . [55] showed that AlgR binds fimS-algR , suggesting that the TCS could positively regulate its own transcription in response to reduced PilWXY1 levels . In addition to being essential for T4P function , FimS and AlgR control alginate production in the context of chronic CF infections , where algR transcription is high [18 , 82] . Phosphorylation of AlgR increases binding affinity at some–but not all–of its target sequences [17 , 62 , 63 , 67] . For example , AlgRD54N failed to support twitching motility , but did not affect alginate production [17 , 63] . Our twitching motility data suggests that AlgRD54A is capable of binding to the fimU promoter , albeit less efficiently than WT AlgR ( S6 Fig ) . FimS is an unorthodox histidine kinase , with four transmembrane domains instead of the typical two , and lacks both a periplasmic sensing domain and the canonical motif involved in ATP coordination that mediates auto-phosphorylation [19 , 83] . Direct interaction and/or phospho-transfer between FimS and AlgR have not been reported . Rather , the idea that FimS acts as a kinase for AlgR comes from this and other studies demonstrating similar phenotypes for fimS , algR , and algRD54N mutants [17 , 18 , 84] . Here , we demonstrated that FimS and AlgR interact in the BACTH assay ( Fig 4 ) lending further support to this model . FimS and AlgR promote expression of genes important for production of alginate , biofilms , and c-di-GMP , and inhibit expression of the T3SS , pyocyanin , and quorum sensing [55 , 56 , 74 , 85 , 86] . The observation that the loss of algR had no impact on virulence towards amoebae [38] or nematodes ( Fig 5A and 5B ) suggests that the AlgR-activated genes may not contribute to virulence , although the mechanisms of killing could differ . In mouse models , fimS and algR deletion mutants are attenuated , though overexpression of AlgR also markedly reduces virulence [55 , 65 , 87] . Further , Little et al . demonstrated that PAO1 algRD54E had WT virulence in Drosophila melanogaster and mouse infection models , while an algRD54A mutant was highly attenuated [87] . The outcomes that result from interaction of P . aeruginosa with different hosts will depend on a combination of factors including host defenses , site of infection , available nutrients , and virulence repertoire of a particular strain . However , our results suggest that changes in the specific repertoire of bacterial virulence factors , or the timing of their production , can tip the balance in the host’s favour . The subset of AlgR-regulated virulence genes important for C . elegans pathogenesis is not defined . Screening of a PA14 transposon library for loss of virulence implicated several genes encoding regulators rather than individual virulence factors , suggesting that C . elegans pathogenesis is multifactorial [35] . Consistent with this hypothesis , a study of 18 WT P . aeruginosa strains revealed no correlation between pathogenicity and any specific virulence factors [88] . We saw WT or greater levels of virulence for algR and algU mutants , respectively , consistent with a role for AlgRU in repression of acute phase virulence factors ( Figs 5 and 8 ) . Factors under positive control of AlgRU ( chronic-phase virulence factors ) may be important during later stages of infection in more complex mammalian infection models , but not crucial for pathogenesis in nematodes [89 , 90] . In support of this hypothesis , past studies have demonstrated that increased mucoidy , via mutation of mucA or mucD , reduced nematode killing [75–77] . While important for the initial stages of infection , T4P are often lost over time in chronic CF lung infections [5 , 91 , 92] . P . aeruginosa CF isolates frequently become mucoid via activation of AlgR , and production of many virulence factors is reduced [82 , 93 , 94] . Although the two outcomes are not necessarily temporally or mechanistically linked , mutations that achieve both may be advantageous during chronic CF lung infections . Specifically , loss of PilWXY1 may be adaptive in the context of CF , leading to AlgR activation . To test this idea , it will be interesting to examine the genotypes of mucoid CF isolates for these types of mutations . In conclusion , our results suggest that PilWXY1 promote virulence towards C . elegans by inhibiting FimS-AlgR activation . These data demonstrate how loss of one virulence factor ( T4P ) may activate others ( via AlgR ) . Because the interplay between virulence factors in P . aeruginosa is complex and dynamic , careful consideration will be required when designing potential anti-virulence therapeutic strategies .
Strains and plasmids used in this work are listed in S1 Table . Bacteria were grown at 37°C for 16 h in 5 ml lysogeny broth ( LB ) Lennox , or on 1 . 5% agar LB plates , unless otherwise specified . Plasmids were transformed into chemically-competent E . coli by heat-shock , and into P . aeruginosa by electroporation [95] . Where appropriate , gentamicin ( Gm ) was added at 15 μg/ml for E . coli , and 30 μg/ml for P . aeruginosa . Kanamycin ( Kan ) was added at 50 μg/ml for E . coli , and 150 μg/ml for P . aeruginosa . Ampicillin ( Amp ) was added at 100 μg/ml for E . coli . L-arabinose was added at 0 . 05% where indicated to induce expression from the pBADGr promoter [96] . Vectors were constructed using standard cloning procedures , using the primers listed in S2 Table . Deletion constructs were designed to contain 500–1000 bp homology upstream and downstream the gene to be deleted . Deletion constructs for PA14 fimU , pilV , pilW , pilX , pilY1 , and pilE were synthesized by Genscript in the pUC57Kan vector . pEX18Gm-sadC was created by amplifying the sadC deletion region from PA14 sadC roeA [42] , followed by digestion and ligation into pEX18Gm . pEX18Gm-fimS , pEX18Gm-algRD54A , and pEX18Gm-algRD54E were made by overlap extension PCR [97] . Restriction digestion followed by ligation of the upstream and downstream fragments was used to create the deletion constructs pEX18Gm-algR , pEX18Gm-algU , and pEX18Gm-pilD . pMS402-PfimU and pMS402-PcdrA were created by amplifying and digesting the promoter regions of the PA14 MP operon and cdrA gene , respectively . Digested pBADGr was treated with alkaline phosphatase prior to ligation to avoid re-circularization of the vector . Constructs were verified by Sanger sequencing ( MOBIX lab , McMaster , Hamilton , ON ) . Allelic exchange was used to remove or alter specific genes [98] . pEX18Gm suicide plasmid derivatives ( see Cloning procedures and S1 Table ) were used to create all mutants in this work . After heat-shock transformation into E . coli SM10 cells , pEX18Gm constructs were conjugated into corresponding PA14 or PAO1 parent strains . Cells were then transferred to Pseudomonas isolation agar ( PIA ) Gm100 plates and incubated for 18 h at 37°C , to select for integration of pEX18Gm derivatives into the chromosome . Colonies were streaked onto LB/sucrose and incubated at 30°C for 18 h to select against merodiploids . Resultant colonies were patched onto LB and LB Gm30 to identify gentamicin-sensitive colonies . Regions flanking the desired mutations were amplified and sequenced to confirm success . Twitching motility assays were performed as previously described [99] , with the following modifications . Individual colonies were stab-inoculated in triplicate into 1% agar LB solidified in plasma-treated tissue culture-grade plates ( Thermo Fisher ) and incubated at 30°C for 48 h . Agar was carefully removed and plates were stained with 1% crystal violet for 5 min . Unbound dye was removed by rinsing with water , then stained twitching areas were measured using ImageJ . Twitching zones were normalized to WT ( 100% ) . Biofilm assays were performed as previously described , with modifications [100] . P . aeruginosa cultures were grown for 16 h at 37°C , diluted 1:200 in fresh LB , and grown to OD600 ~0 . 1 . Cultures were then diluted 1:500 in liquid SK media ( 50 mM NaCl , 0 . 35% peptone , 1 mM CaCl2 , 1 mM MgSO4 , 5 μg/ml cholesterol in EtOH , 20 mM KH2PO4 , and 5 mM K2HPO4 ) , then 96-well plates were inoculated with 150 μl each strain , in triplicate . Sterility controls ( liquid SK media ) were included throughout the plate to check for contamination . Plates were covered with peg lids ( Nunc ) then wrapped in parafilm and incubated at 37°C for 24 h , shaken at 200 rpm . After incubation , the OD600 of the plate was measured to check for uniform growth and lack of contamination . Peg lids were washed for 10 min in 200 μl/well 1X phosphate-buffered saline ( PBS ) , then stained with 200 μl/well 0 . 1% ( w/v ) crystal violet for 15 min . Unbound crystal violet was removed by washing lids in 70 ml distilled water 5 times at 10 min intervals . Crystal violet was solubilized from lids in 200 μl/well 33 . 3% acetic acid , then the absorbance at 600 nm was measured . Optical density and absorbance at 600 nm were plotted for growth and biofilm formation , respectively , then analyzed by one-way ANOVA followed by Dunnett post-test to compare each mutant to the WT control , p = 0 . 05 . Error bars indicate standard error of the mean . Representative wells of acetic acid-solubilized crystal violet were imaged . SK assays were performed as described previously [101] . SK plates ( 0 . 35% peptone , 50 mM NaCl , 2% agar , 1 mM CaCl2 , 5 μg/ml cholesterol , 1 mM MgSO4 , 20 mM KH2PO4 , 5 mM K2HPO4 , 100 μM FUDR ) were seeded with 100 μl of an overnight culture and incubated overnight at 37°C . The following day , plates were enriched with 1 ml of an overnight culture concentrated to 100 μl . Synchronized L4 worms were collected from E . coli OP50 plates , washed twice in M9 buffer , and then >50 worms were seeded onto each bacterial lawn on individual SK plates . SK plates were incubated at 25°C and scored for dead worms every 24 h . Worms were considered dead when they did not respond to touch , and were removed from the plate . OP50 was included as a negative control for virulence . Percent survival was plotted as a function of time . Survival curves were plotted on GraphPad Prism 5 . 00 for Windows , then compared using the Gehan-Breslow-Wilcoxon test , p = 0 . 05 . Given that larvae were synchronized at 20°C then transferred at L4 to 25°C for the duration of the assay , worms were at risk of death due to senescence , rather than direct killing by P . aeruginosa , before day 10 [46] . Therefore , the Gehan-Breslow-Wilcoxon test , which gives weight to earlier timepoints , was used in favour of the standard log-rank test ( notably , all reported differences were also significant by the standard log-rank test ) . To correct for multiple analyses , the critical p-value of 0 . 05 was divided by the number of pairwise comparisons made within an individual trial , as per the Bonferroni method [102] . Each assay was performed at least 3 times , and differences were only considered significant if they were reproducible in the majority of trials . Representative trials are shown; all replicates can be viewed in the Supplemental Material ( S1 File ) . Luminescent reporter assays were performed as previously described , with minor modifications [60] . Various strains harbouring the pMS402-PfimU or pMS402-PcdrA plasmids , encoding the luciferase genes under control of the fimU or cdrA promoters , respectively , were grown for 16 h at 37°C in LB Kan150 , then diluted 1:50 in fresh liquid SK media with Kan150 , in addition to Gm30 and 0 . 05% L-arabinose where appropriate . Subsequently , 100 μl of each culture was added to white-walled , clear-bottom 96-well plates ( Corning ) in triplicate , and incubated with shaking at 37°C in a Synergy 4 microtiter plate reader ( BioTek ) . Luminescence readings were taken every 15 min for 5 h , and normalized to growth ( OD600 ) at each time point . Readings that exceeded the limit of detection ( >4 000 000 luminescence units ) were discarded . At least 3 individual trials were performed . Error bars indicate standard error of the mean . To test for interactions between FimS and AlgR or individual pilins , BACTH assays were performed as previously described [103] . pUT18C and pKT25 derivatives , encoding the T18 and T25 domains of the Bordetella pertussis CyaA adenylate cyclase fused to the N-terminus of FimS , AlgR , PilA , FimU , PilV , PilW , PilX , or PilE [24 , 60 , 104] , were co-transformed into E . coli BTH 101 to screen for pairwise interactions . Single colonies were inoculated in 5 ml LB Amp100 Kan50 and grown overnight . The following day , 100 μl was inoculated into 5 ml fresh media and grown to OD600 = 0 . 6 , then 5 μl was spotted onto MacConkey plates ( 1 . 5% agar , 100μg/ml ampicillin , 50μg/ml kanamycin , 1% ( w/v ) maltose , 0 . 5mM isopropyl b-D-thiogalactopyranoside ) ( Difco ) or LB Amp100 Kan50 plates supplemented with 100 μl of 20 mg/ml X-gal . Plates were incubated at 30°C for 24 h . An interaction was considered positive when colonies appeared pink or blue on MacConkey and LB + X-gal plates , respectively . BTH 101 expressing pUT18C-fimS and pKT25-fimS was used as a positive control [49] . Negative controls included BTH 101 expressing the empty vectors pUT18C and pKT25 , and BTH 101 expressing pKT25-fimS and pUT18C ( empty vector ) . | Pseudomonas aeruginosa causes dangerous infections , including chronic lung infections in cystic fibrosis patients . It uses many strategies to infect its hosts , including deployment of grappling hook-like fibres called type IV pili . Among the components involved in assembly and function of the pilus are five proteins called minor pilins that—along with a larger protein called PilY1—may help the pilus attach to surfaces . In a roundworm infection model , loss of PilY1 and specific minor pilins delayed killing , while loss of other pilus components did not . We traced this effect to increased activation of the FimS-AlgR regulatory system that inhibits the expression of virulence factors used early in infection , while positively regulating chronic infection traits such as alginate production , a phenotype called mucoidy . A disruption in the appropriate timing of FimS-AlgR-dependent virulence factor expression when select minor pilins or PilY1 are missing may explain why those pilus-deficient mutants have reduced virulence compared with others whose products are not under FimS-AlgR control . Increased FimS-AlgR activity upon loss of PilY1 and specific minor pilins could help to explain the frequent co-occurrence of the non-piliated and mucoid phenotypes that are hallmarks of chronic P . aeruginosa lung infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"bacteriology",
"biofilms",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"caenorhabditis",
"pathogens",
"microbiology",
"operons",
"animals",
"pseudomonas",
"aeruginosa",
"animal",
"models",
"caenorhabditis",
"elega... | 2018 | Pseudomonas aeruginosa type IV minor pilins and PilY1 regulate virulence by modulating FimS-AlgR activity |
Many everyday estimation tasks have an inherently discrete nature , whether the task is counting objects ( e . g . , a number of paint buckets ) or estimating discretized continuous variables ( e . g . , the number of paint buckets needed to paint a room ) . While Bayesian inference is often used for modeling estimates made along continuous scales , discrete numerical estimates have not received as much attention , despite their common everyday occurrence . Using two tasks , a numerosity task and an area estimation task , we invoke Bayesian decision theory to characterize how people learn discrete numerical distributions and make numerical estimates . Across three experiments with novel stimulus distributions we found that participants fell between two common decision functions for converting their uncertain representation into a response: drawing a sample from their posterior distribution and taking the maximum of their posterior distribution . While this was consistent with the decision function found in previous work using continuous estimation tasks , surprisingly the prior distributions learned by participants in our experiments were much more adaptive: When making continuous estimates , participants have required thousands of trials to learn bimodal priors , but in our tasks participants learned discrete bimodal and even discrete quadrimodal priors within a few hundred trials . This makes discrete numerical estimation tasks good testbeds for investigating how people learn and make estimates .
To characterize how people make continuous estimates , first we outline Bayesian decision theory , which prescribes how to maximize expected rewards . Bayesian decision theory is composed of three components that each need to be specified: the prior probability , the likelihood , and the decision function . There are a variety of distributions and functions that can be used for each component , but how the components are combined is fixed by the laws of probability [26] . The decision maker begins with a prior P ( S ) , which gives the prior probability of each state of the world , S . For simplicity , we assume that the states of the world all are arranged along a single dimension and each state has a one-to-one mapping to a response . On each trial , the decision maker observes some data , X , that are noisy or ambiguous , and the likelihood , P ( X|S ) , is the probability of the observed data given each of the possible states of the world . The prior and the likelihood are combined via Bayes’ rule to determine the posterior probability of the states of the world having observed the data , P ( S | X ) ∝ P ( X | S ) P ( S ) ( 1 ) where equality is achieved if the right-hand side is divided by P ( X ) . In a sequential task , this posterior distribution is used as the prior distribution for the next trial , so the prior reflects a participant’s accumulated experience throughout the task . The best estimate depends not only on what is believed to be true about the world; because S given X is uncertain , it also depends on what happens if an incorrect response is made . The dependence of rewards on the response is given by the loss function L ( R; S ) , which captures the loss ( negative reward ) for making response R if the state of the world is S . The decision function , DL , then maps the posterior probabilities onto the response with the smallest expected loss D L ( P ( S | X ) ) = arg min R ∫ S L ( R ; S ) P ( S | X ) . ( 2 ) Continuous estimates have been modeled using a particular set of priors , likelihoods , and decision functions . Likelihoods are often assumed to be Gaussian because this density is a good match to the perceptual noise in many tasks , and participants have been shown to correctly adapt to the amount of noise in their perception . For example , in multi-sensory integration and sensorimotor tasks , the normative weight applied to each sensory cue depends on the variance of Gaussian-distributed perceptual noise , and participants’ weights come close to matching these normative weights [8 , 10 , 27] . For the prior , a standard choice for the training distribution in continuous estimation tasks is a Gaussian density because it makes the analysis analytically tractable: combining a Gaussian prior with a Gaussian likelihood results in a Gaussian posterior . However more flexible schemes for learning priors exist , and a common way to introduce flexibility is to use a non-parametric prior which grows in complexity as more data are observed . Kernel density estimation is a standard choice for building a non-parametric prior for continuous training data [28] and this kind of representation has been used in many models of human categorization [29 , 30] . In kernel density estimation , the nonparametric prior is constructed from a weighted sum of component parametric densities , one for each previously observed data point . Mixture priors , which have also been used in models of categorization [7 , 31 , 32] , provide another representation that allows more flexibility in the number of parametric components than kernel density estimation . This representation operates between the simple parametric and the kernel density cases , grouping similar data together into the same component , but allowing different components for data that are dissimilar . In continuous estimation tasks , participants learn Gaussian and other unimodal training distributions quickly: in various continuous estimation tasks , unimodal training distributions have been learned in hundreds of trials [10 , 19 , 21 , 23] . Participants can also learn bimodal training distributions , demonstrating that they do not use just simple parametric priors , but they are slower to do so . Experimenters have had to train participants on bimodal distributions for thousands of trials [10 , 22] , because fewer training trials do not result in clear evidence of learning [19] . Participants were able to use bimodal distributions when presented with an explicit summary , and this work also showed participants are better described as using a mixture prior rather than kernel density estimate with a very narrow kernel [15] . For the decision function in continuous estimation , a small number of simple functions have been considered , each of which can be motivated by one or more loss functions . An all-or-none loss function leads to choosing the response with the highest posterior probability , a quadratic loss function leads to taking the mean of the posterior , and a linear loss function yields the posterior median as the best response . A fourth decision function , drawing a sample from the posterior , requires a more complex motivation . One route is to speculate that participants assume that the computer is adaptively responding to their input in a competitive fashion , so that a stochastic decision function can increase their expected reward [33 , 34] . Another is to assume that participants are maximizing expected reward subject to particular computational costs: if participants draw samples from the posterior and these samples require time or effort to generate , it can be better to make a quick and less accurate decision rather than a slow and effortful accumulation of enough samples to calculate the maximum of the posterior [35 , 36] . Comparison of these decision functions in continuous estimation has yielded task-dependent conclusions . Some researchers have found evidence for the mean decision function , finding that participants used a loss function that was quadratic near the correct value but more linear far from the correct value , giving it robustness to outliers [24] . Other work has found evidence for a mean decision function despite feedback which did not encourage this decision function [23] , but a later analysis showed that this particular task does not discriminate well between decision functions [14] . Recent research has incentivized the max decision function and then investigated the decision function actually used . Work using Gaussian priors found that instead of the max , participants were performing an interpolation between drawing a single sample and taking the max of the posterior [20] . There are various mechanisms that could produce this interpolation: participants could be drawing a number of samples from the posterior distribution and taking the mean of these samples as their estimate , they could be raising the posterior to a power greater than one and then sampling their estimate from this exponentiated posterior distribution , or perhaps they combine the two by taking the mean of a number of samples from an exponentiated posterior . While these explanations are indistinguishable for Gaussian posteriors [20] , work using a bimodal training distribution has successfully tested two of these possibilities , the mean of a number of samples versus a sample from an exponentiated posterior , and found that participants were drawing a single sample from an exponentiated posterior distribution [15] . Though researchers have occasionally investigated discrete numerical estimation , it has not received much attention , possibly because it has been viewed as no different from continuous estimation . However if we look at the work that has been done , there are suggestions that people make these two types of estimates differently . Bayesian decision theory is useful here for cataloguing the similarities and differences . The likelihood in discrete numerical estimation is similar to that found in continuous estimation . As in continuous tasks , participants find it more difficult to discriminate stimuli that are closer physically even though they are naturally discrete . Indeed , in perceptual numerosity tasks , discrimination performance nearly follows Weber’s law , implying that the standard deviation of the perceptual noise distribution is proportional to its mean [37] , which has been modeled as Gaussian noise on the log-transformed numbers [38–41] . Like in continuous tasks , participants appear to use knowledge of their own perceptual noise to set their likelihood in numerosity tasks [38] , a point we also address in the S1 Methods . Investigations into the priors used in discrete numerical estimation have shown that participants can learn unimodal distributions of stimuli quickly , as in continuous estimation . Participants are able to reconstruct the frequency of events from unimodal distributions from just a few trials [42] and their estimations of new events quickly show an influence of the mean in a changing sequence of numbers [43] . In the similar task of absolute identification , in which participants are asked to identify a series of perceptual stimuli with numerical labels [44] , participants are also influenced by unimodal distributions of stimuli [45] . However , tasks training bimodal prior distributions point to potential differences between continuous and discrete numerical estimation . The first potential difference is in the speed of learning bimodal priors . In one task , participants asked to reconstruct bimodal prior distributions were able to do so within a few hundred training trials [18] , and in another participants could do so for some bimodal distributions after only 12 trials [46] . Though this suggests that participants have a speed advantage in learning priors for discrete numerical estimates , these priors were assessed through reconstruction and it needs to be established whether the same priors are used in estimation . A potential difference in the decision function was also found in [18] , in an experiment in which participants were asked to estimate the revenues associated with trained and novel company names . Participants’ estimates for novel companies were either at the lower edge of the range of trained revenues or were in the middle of the range , results which were modeled as drawing a set of samples from the prior combined with a mixture of two decision functions . One decision function was to use the lowest sample in the set as the estimate ( because unknown companies are likely to have low revenue ) , and the other was to take the mean of the set of samples as the estimate [18] . The use of the mean of a small number of samples as the decision function contrasts with the exponentiated posterior supported by work in continuous estimation , and this is another potential difference . However , revenue estimation is very different from the perceptual tasks used in continuous estimation , so it would help to investigate the decision function in a perceptual discrete numerical estimation task . Here we investigate these potential differences in two different discrete numerical estimation tasks: estimating the number of dots on a screen and estimating the area of a rectangle . We are particularly interested in whether participants can quickly learn complex multimodal prior distributions and what decision function they use to make their estimates . In exploring this , we go further than previous studies by investigating whether participants use a kernel density estimate , a mixture model , or a categorical distribution as a prior . Through both combinatorial model comparison and fitting of nested models we examine the decision function , investigating whether the mean , max , the mean of a number of samples from the posterior , a sample from an exponentiated version of the posterior distribution , or perhaps a more complex decision function best explains participants’ estimates . We compare our findings to the results from continuous estimation in the discussion as well as explore the implications for what priors people can learn and what decision functions they use .
To characterize how participants make discrete numerical responses , we ran a new experiment on numerosity estimation in which participants were trained on a bimodal distribution . Participants were asked to estimate the number of dots that briefly appeared on a screen in a series of trials , receiving feedback about whether they were correct and what the correct answer was after each trial as shown in Fig 1A . They were not told anything about which numbers to expect in addition to the feedback . On each trial , the number of dots on the screen was drawn from a sharp bimodal distribution with two peaks on either side of a region of lower probability values ( e . g . , the distribution shown in the top left corner of Fig 2 ) . A sharp bimodal distribution allows us to better identify the prior used . If participants are not generalizing beyond the numbers that were given as feedback , then their prior should eventually match the training distribution and they will not respond outside the range of the stimuli . However , if participants are using a parametric or kernel density prior distribution , then the prior distribution will have some spillover outside the range of stimuli , and participants will respond outside this range even after hundreds of training trials . Using a mixture prior will also result in responses outside the range , but they will likely be fewer in number . Examples of these four possibilities are shown in the top row of Fig 2 . Before and after the main task , we included a separate discrimination task that allowed us to characterize the likelihood distribution for each participant . This removed a degree of freedom from the process of characterizing the prior and decision function . The noise in numerosity judgments is well-known to follow Weber’s law with a standard deviation proportional to the mean [37] , and has been modeled in past research as a lognormal distribution [38–41] . We assumed that the likelihood distribution was accurately calibrated and thus equivalent to the noise distribution . The scale of the lognormal distribution σ ( i . e . , the standard deviation of the natural logarithm of a lognormally distributed variable ) , which can be determined from the Weber fraction w using the formula σ = log ( w 2 + 1 ) , was estimated in a discrimination task in which participants were asked which of two screens contained more dots ( shown in Fig 1C ) . Participants’ discrimination judgments were well fit with a standard deviation that ranged from 0 . 18 to 0 . 53 , with a median of 0 . 22 . These estimates are in reasonable agreement with previous research that found for numerosity estimates that discriminability was equivalent to σ ≈ 0 . 16 [37 , 47] . After fixing the lognormal standard deviation , we can make predictions for the responses after the training distribution has been learned for each pairing of prior and decision function . Using the median estimate of σ = 0 . 22 , predictions from pairings of possible priors and decision functions are shown in Fig 2 in the form of conditional response distributions ( CRDs ) : for trials on which a particular number of dots are presented , each panel shows the distribution over the responses expected by the combination of prior and decision function . Two qualitative features stand out in these plots . The first is the number of modes in each CRD . If the mean decision function is used or if the prior is a Gaussian distribution then the CRD will be unimodal , otherwise it will be bimodal . The second qualitative feature is whether responses occur outside the range of values that was presented . Responses will always be within the range of presented values for the categorical prior , while for other priors responses can occur outside of the range if participants sample from the posterior distribution . In the main task , participants were assigned to one of three groups , with each group participating in a series of trials in which the sharp bimodal prior distribution covered a larger or smaller range . This was done to ensure that the results were not strongly dependent on the distance between peaks or on the particular numbers assigned to the modes of the distributions . The average and individual CRDs for each of the three groups are shown in Fig 3 , along with the distribution of the training trials given to participants . In order to plot stable performance , the first 300 trials were not included in these plots . The empirical CRDs for each group show a strong bimodality , implying that neither the mean decision function nor the Gaussian prior characterize human data . Responses are not made only at the modes of the training distribution , however , as a large number of responses are found between the two peaks . These middle responses are evidence against the max decision rule ( which would be very unlikely to result in an intermediate response ) , so from qualitative inspection posterior sampling is left as the best characterization of the decision function . Further inspection of the average data shows very few responses outside of the range of presented values: the narrow group shows no such responses , and the medium and wide groups show few responses of these types . For the medium and wide groups , the responses outside the range of presented values only appear for a small subset of the participants: the third and fifth participant in the medium group , and the third and fourth participant in the wide group . The lack of responses outside of the range of presented values , combined with the identification of participants’ decision function as consistent with posterior sampling , implicates the use of a categorical prior , though this is not easy to distinguish from a mixture prior , as shown in Fig 2 . We fit a set of computational models ( see Methods; Model comparison ) to provide quantitative evidence that individual participants were using categorical priors and sampling from the posterior . Each model was fit to all of the trials and the prior was updated after each instance of feedback was given . We specifically tested the combinations of prior updating ( categorical ( Dirichlet ) or Gaussian kernel ) and decision functions ( mean ( Average ) , Max or Sample ) . We performed a model comparison using the Bayes Information Criterion that adjusts the fit of the model with a penalty for complexity [48] . Eighteen of the twenty participants in this experiment were best described by the categorical prior and a decision function that drew a single sample from the posterior . For the remaining two participants , one was best described by a Gaussian kernel and a max decision function and the other by a categorical prior and a max decision function . The best models for each participant are indicated in Fig 3 and the BIC values transformed into approximate posterior probabilities are shown in the S1 Methods . To allow for a wider range of possible behaviors , we also fit computational models that allowed for “trembling hand” noise and models that allowed the posterior distribution to be raised to a power before the decision function was applied ( see Methods ) . Once we included this set of models , we found that nineteen of the twenty participants were best described by raising the posterior distribution to an exponent larger than one before the decision function was applied , while the remaining participant was best described with the original model ( implying an exponent of one ) . Once the posterior distribution was raised to a power , behavior was best described as a single sample for ten participants and as the mean of the exponentiated posterior for nine participants . This generalization elaborates on what was found with the first set of computational models: exponentiating the posterior means that participants lie between sampling and the max decision function , and the individual differences in using a single sample or the mean reflect individual differences in the amount of stochasticity in the estimates and in the tendency to sometimes respond near the middle of the presented range of stimuli . A final generalization was to fit a ‘super-model’ to each participant’s data ( see Methods ) that allows us to further investigate the individual differences in stochasticity that participants have in their estimates by quantifying the number of samples they use . The individual best fits and an exercise showing that these parameters are identifiable are given in the S1 Methods . Reinforcing the model comparison above , nineteen of the twenty participants used an exponentiated posterior distribution to make their decisions: the exponent was well above 1 . 0 for all but one participant . This one participant was best fit by a single sample , so there was no evidence that any participants were taking the mean of a small number of samples from the untransformed posterior , instead participants were sampling from an interpolation between the posterior and a distribution that was entirely on the max of the posterior . This analysis allows us to further investigate those participants found to be using the mean of an exponentiated posterior . Half of these participants were best fit by taking the mean of between 2–4 samples , while the remainder were taking the mean of a larger number ( i . e . , about 30 ) samples . In summary , this experiment established that the great majority of participants could learn essentially categorical priors when using discrete numerical responses and tended to respond using a decision function that was either a single sample or the average of multiple samples drawn from an exponentiated posterior distribution . A question then arises about the generality of the results . Are people using a categorical prior distribution because the number of dots is necessarily a discrete quantity ? Or are they using a categorical prior distribution because of the discreteness of the responses ? To test whether the results of Experiment 1 were driven by the discreteness of dots or the discreteness of the response , we ran essentially the same experiment , but instead of asking participants to estimate the numbers of dots we asked them to estimate the area of rectangles . Like in the example of buying paint to cover a wall , rectangle area is a continuous quantity but we forced participants to make discrete responses: they were required to estimate the area of rectangles in whole square centimeters . Two groups of participants were run in this experiment and the results are shown in Fig 4 . When we fit their discrimination data , we found a median of σ = 0 . 41 . As in Experiment 1 , the average results in the estimation task show a bimodal distribution . Responses often fall in the middle but hardly ever fall outside the range of presented values . This pattern again is qualitatively most consistent with a categorical prior and a decision function that draws a single sample from the posterior . We used the same analysis approach of successive generalization with the same models as we used in Experiment 1 , with all individual results given in the S1 Methods . In the simplest model comparison , we found that every participant was best described by a categorical prior distribution and a decision function that was a single sample from the posterior . This result is given next to each individual in Fig 4 . When we generalized the comparison to allow the posterior distribution to be raised to a power before the decision function was applied , we found that every participant was better described by exponentiating their posterior distribution , bringing it closer to a distribution that was entirely on the max . As in Experiment 1 , half of participants were best described by taking a sample from this exponentiated posterior , while the other half were best described by taking the mean of the exponentiated posterior reflecting a tendency to sometimes respond near the middle of the presented range of stimuli . The more general ‘super-model’ analysis provided more detail on how many samples were being taken from the exponentiated posterior , and thus the amount of stochasticity in the estimates . All of the participants were best fit by taking the mean of between 3 and 100 samples from an exponentiated posterior distribution , with the participants best described as taking the mean of the exponentiated posterior tending to be on the higher end of this range . In both Experiments 1 and 2 participants used near-categorical prior distributions and either take a single or multiple samples from an exponentiated posterior when they are asked to make discrete responses , regardless of whether the underlying quantity was discrete or continuous . They clearly were able to learn a bimodal distribution surprisingly quickly , so these results lead to the question of how flexible this representation is . As shown in Fig 2 a mixture of Gaussians prior can quite closely imitate the categorical prior that was best supported by the data . As the mixture model interpolates between a Gaussian prior and kernel density estimation , it is difficult to provide evidence against this model . More generally , allowing mixtures of other types of distributions , such as uniform distributions , makes the problem even more difficult . In order to test whether participants were using mixture models , Experiment 3 investigates whether participants can learn a more complex prior distribution . In order to further investigate how complex a prior distribution participants can learn within a few hundred trials , participants in this experiment were trained on a quadrimodal prior distribution . As shown in Fig 5 , this distribution was designed to test whether participants were using simple mixture models . If participants are assigning all of the trials with 23–25 dots to one mixture component and all of the trials with 29–31 dots to a separate mixture component , then the predictions of the categorical prior and the mixture prior are clearly distinguishable: the mixture model always predicts a peak in response frequency at numbers 24 and 30 , while the categorical prior distribution predicts that these numbers will be selected less often than the peaks . These same predictions would also be made if participants are using other distributions in a mixture model , such as uniform distributions , with the same assignments also leads to the prediction that numbers 24 and 30 will be selected at least as often as the other presented numbers . Three groups of participants were run in this experiment: one group completed a perceptually easier numerosity task , one group a perceptually more difficult numerosity task , and one group completed a rectangle task . When we fit the discrimination data , we found medians of σ = 0 . 19 , σ = 0 . 14 , and σ = 0 . 14 for the Difficult Numerosity , Easier Numerosity , and Rectangle groups respectively ( first estimation task , see Methods ) and used the latter value for generating Fig 5 . The mean results for each of the groups are shown in Fig 6 and the qualitative results in this experiment again look like a combination of a categorical prior with sampling from the posterior distribution . We ran the same analysis as was done for Experiments 1 and 2 ( model comparison and fitting parameters of the ‘super-model’ with individual results given in the S1 Methods ) to determine which model explained participants responses best . Using the simplest model comparison , 15 of 21 participants were best described by a categorical prior and by sampling their estimates from the posterior . The remaining six participants were better described by using a Gaussian kernel prior , with four of them taking the max of the posterior and two sampling . The correspondence of these results to the individual data is shown in Fig 6 . The less restrictive model comparison again showed that a categorical prior and a single sample from an exponentiated posterior distribution was the best description of the largest number of participants ( 12 of 21 ) . The other participants were best fit by a variety of models . For the ‘super-model’ analysis , which allows us to better investigate the stochasticity in the decision function , 11 out of the 21 participants drew a single sample from a posterior distribution that had an exponent clearly above 1 . 0 , while the remainder either drew a single sample from a posterior distribution with an exponent not much different from 1 . 0 or took the mean of a larger number of samples from an exponentiated posterior distribution . No participants were best fit by averaging multiple samples from the unexponentiated posterior distribution . To test whether participants were using a simple mixture model which assigned the trials with dots 23–25 to one mixture component and trials with dots 29–31 to separate mixture component , we looked at whether participants produced fewer responses of 24 and 30 compared to the peaks of the distribution . If participants were equally likely to respond with any of the presented numbers ( after trial 300 and ignoring what the actual presented value was ) , then participants should have picked numbers 24 or 30 on at least 1/3 of trials . Using 1/3 of trials as a null hypothesis we ran binomial tests to determine if the actual number of responses was significantly lower than this value for each participant . Overall , 14 of 21 participants produced significantly fewer responses than the null hypothesis predicted ( p < 0 . 05 ) . The participant showing significant differences are marked with stars in Fig 6 . Clearly a number of participants were not using this simple mixture model as their prior distribution . Mixture models that are closer to the categorical prior are harder to rule out . For example , mixture model components might consist of separate components for every number , except for a single pair of numbers that are represented with the same component . For the prior distribution trained in this experiment , this would be a mixture model consisting of five component densities to represent the six presented responses . We simulated how often responses just outside the presented range would appear if there were separate mixture components for every number except for one single pair of adjacent numbers for a variety of choices of the adjacent pairs and values of σ and found that participants would be expected to produce a response just outside the range on perhaps as few as 0 . 6% of trials . This low rate means that it is not possible to say that any individual participant produced significantly fewer responses: the probability of producing zero of these responses on 200 trials assuming a 0 . 6% probability is 0 . 3 . However , we do note that 11 of 21 participants did not produce any of these just-outside-the-range responses ( and four additional participants produced only one ) . If all participants were consistently grouping adjacent numbers together the probability of observing this many participants with zero responses of this type is low ( p = . 027 ) . Overall , Experiment 3 demonstrates that participants are accurately learning very complex quadrimodal prior distribution within a few hundred trials . The complexity of the prior learned allowed us to even rule out for most participants a simple mixture model that could have explained behavior in the first two experiments .
Previous investigations had suggested that the decision function used in discrete numerical estimation might be different from that used in continuous estimation tasks . In tasks in which the max decision function was incentivized , work with continuous estimation tasks has shown that participants exponentiated the posterior distribution and drew a sample from this exponentiated distribution [15 , 20] . This result contrasts with the findings from discrete numerical tasks showing that even when incentivized to use the max rule , participants still appeared to use the mean of a small number of samples [18] . Our results for discrete numerical estimation were different . We encouraged the max decision rule ( by giving participants feedback of ‘correct’ or ‘incorrect’ ) , and we found that participants were using an exponentiated version of the posterior distribution . Our ‘super-model’ analysis allowed for both exponentiation of the posterior and for the mean to be taken of a number of samples , and we found strong evidence for exponentiation in a large majority of participants and individual differences in whether a single or multiple samples were used . Despite the individual differences , no participants were best fit by the mean of a small number of samples from the untransformed posterior . The divergent results between our task and the revenue estimation task of [18] need explanation . One potential key difference is that the likelihood was characterized in our task but not in the revenue estimation task . Pronounceable company names induce different expectations about company stock performance than non-pronounceable names [49] , and these expectations could be reflected in a variety of likelihoods that cause the resulting estimates to resemble those coming from a mixture of decision functions . This is of course speculative , and the divergence may be due to other task differences , but it highlights the importance of characterizing all of the components of Bayesian decision theory . Overall , our findings for the decision function roughly correspond with those found in continuous estimation , so there seems to be no strong dividing line between the decision function used to make these two types of estimates . Of course decision functions have usually not been characterized in much detail nor have been characterized across a range of tasks , so later investigations could reveal subtle differences in how the task shapes the decision function used . In terms of the prior , across three experiments we found the novel result that participants were better characterized as using a categorical prior than by a simple parametric distribution or by a kernel density estimate with any appreciable width . Mixture priors could possibly explain the results of Experiments 1 and 2 , but Experiment 3 showed that a simple implementation of a mixture prior did not match the data as well for most participants . The use of a categorical prior was supported by participants’ ability to learn complex multimodal distributions very quickly . The speed and flexibility of participants’ prior learning stands in contrast to work in continuous tasks , where it is difficult to find evidence for quick learning of bimodal priors . One task required 4 , 000 feedback training trials to teach participants a bimodal distribution [10] , and another required 1 , 700 trials [22] . However , when [19] used 1 , 500 training trials in an interval timing task , there was some suggestion of bimodality if the peaks were well-separated , but the data could also be explained by a uniform prior . The only example of equally fast learning of a bimodal prior without giving participants hints comes from other experiments using discrete numerical responses . In the revenue estimation experiments of [18] , it was found that a bimodal prior distribution could be reconstructed within 400 training trials . More impressive was the demonstration that bimodal priors could be reconstructed after as little as 12 trials by individual participants [46] . However , these demonstrations come from tasks in which participants are asked to reconstruct the distribution rather than make an estimate , so this work is the first to show that participants do use bimodal priors when making estimates , and can learn to do so more quickly than in continuous tasks . In addition the priors that participants learned were impressively accurate: we showed that quadrimodal prior distributions could be learned , and that their priors were better described by a categorical distribution rather than a kernel density estimate or some forms of mixture models . Given these differences in speed of learning , it is interesting to speculate whether there are particular properties of tasks that require discrete numerical responses that make it easier to learn a complex prior . One difference between discrete numerical and continuous estimates is that it is easier to provide clear feedback for discrete numerical estimates . Both [19] and [10] used visual position as feedback in their sensorimotor and interval timing tasks , and noise in vision and memory makes this feedback less certain . In the orientation estimation task of [22] participants were told their average deviation every 20 trials , while this feedback is digital it does not provide as much information as the true orientation used on each trial . It is difficult to see how the feedback could be improved for tasks that require continuous responses: feedback either needs to be susceptible to noise ( perhaps both sensory noise and noise in encoding and remembering the feedback ) or it is not directly mapped to the responses . Participants cannot perfectly be shown what response they should have made . In contrast , our experiments and the experiments of [18] showed participants the correct response after every trial in essentially a noise-free fashion . This is a real advantage of using discrete numerical responses and feedback because feedback can be given uncorrupted by sensory noise after every trial and it is easily mapped to the responses than participants make . In fact , [46] explicitly showed this difference when participants were asked to reproduce a distribution: for experiments in which numbered stimuli were replaced by circles of various sizes , participants required more trials and greater separation between the modes to learn the bimodal distributions . It is possible that the differences in clarity of feedback explain the rates at which participants learn bimodal priors in different tasks . If participants are using a form of Occam’s razor when constructing their prior distribution , then the more informative trials would more quickly convince them to abandon a simpler prior in favor of a more complex representation . The priors learned in our experiments , especially Experiment 3 , were much more complex than those taught to participants in other estimation tasks [10 , 15 , 18 , 19 , 22 , 46] . In addition to having four modes , our prior had a pattern of low-probability and no-probability responses that participants’ responses matched . Participants were not just representing the prior as a mixture of two parametric components , but were learning the prior probabilities associated with individual responses . Work using other tasks has demonstrated fairly complex prior learning , but in other tasks it is generally not clear whether participants are learning a prior or a mapping . For example using a categorization task , a subset of participants learned to discriminate a multidimensional quadrimodal distribution from a multidimensional mixture of two Gaussian distributions [50] . While participants were able learn these complex discriminations and their behaviour could be described by a model that approximates Bayesian inference , this work did not rule out a complex decision bound model ( i . e . , a mapping ) as an alternative [50 , 51] . In Experiment 3 we ran additional trials to test whether participants were actually learning and using a prior or if instead they were learning a mapping from the stimuli to the responses . As discussed in the S1 Methods , we found that the responses of more than half of participants were best explained by a prior rather than a mapping . Use of a prior is also supported by recent work that demonstrated that participants take into account the reliability of various senses in a multisensory numerosity task [38] . Our results contrast with other work showing that participants do not learn a categorical prior . In a continuous estimation task with a wide range of possible responses , a categorical prior did not explain the data as well as a mixture [15] . Likewise in a numerosity task that showed participants a much wider range of numbers than our experiments , a mixture model provided a better fit to their participants’ data than just using the trained examples as a prior [52] . The key difference is likely the variety of correct responses in each experiment . As the number of potential responses increases it is hard to imagine that participants would precisely track the frequency with which every single number appeared . For example , if every number from 100 to 200 appeared in a random order with the exception of number 134 , it is implausible that participants would notice . This contrast raises questions about where the transition between a categorical prior and a mixture model occurs , and even if there is a distinction between the two . It is possible that participants represent a small set of numbers symbolically and use a categorical prior , but represent a large set of numbers as a mixture prior over a continuous variable . Alternatively , it could be that our categorical prior is simply a mixture with a separate component for each response . In this case , there would likely be a smoother transition between representations of the prior for small and large sets of responses . Very few of our participants were best fit by a simple decision function: the max or mean of the untransformed posterior distribution . Instead it appeared that the large majority of participants were performing some kind of approximate inference by drawing one or more samples . Previous work has put forward mechanisms with which this could be done . For example , [20] showed that participants’ responses were consistent with either taking the mean of a number of samples in a continuous estimation task or drawing a single sample from an exponentiated posterior distribution . Later , [15] disambiguated these two operations with a bimodal prior , showing that raising the posterior distribution to an exponent was the better description . Both of these mechanisms have been touted as tradeoffs between effort and accuracy , and possibly a rational use of cognitive resources [35] , though there is always the possibility that participants have particularly complex hypotheses about the computer’s behavior instead . Drawing a sample may take time or effort , and a small number of samples may provide the best tradeoff between effort and accuracy to yield the highest overall reward [36] . Similarly , raising a posterior distribution to a power has also been cast as a tradeoff between effort and accuracy , but one that assumes effort is required to perform the exponentiation that transforms the belief distribution into a response distribution [53] . While this makes for a nice contrast , the picture is complicated by two additional mechanisms that are essentially indistinguishable from exponentiating the posterior distribution , even for bimodal priors: taking the maximum of a number of samples drawn from the posterior distribution , and taking the maximum of a posterior distribution that has been corrupted with noise [15] . This last mechanism may well differ from the others if it is assumed that the amount of noise in the posterior is not under the control of the participant; in this case sampling-like behavior would not be a tradeoff between effort and accuracy . We add to this literature by showing that while an exponentiated posterior distribution is necessary to explain the data as in [15] , additionally a large number of participants appear to be taking the mean of a number of samples drawn from this exponentiated posterior distribution . Despite the fact that the maximum of the posterior was asked for by identifying only exactly correct responses as ‘correct’ , participants still showed some tendency to produce some responses near the mean of the posterior . It is interesting to speculate what sort of mechanism could support both a tendency to respond with the max and a tendency to respond with the mean . Our best fitting combination of the mean of samples from an exponentiated posterior distribution is one possibility . It could be that participants are using an exponentiated posterior helps to emphasize the mode which is most likely under the posterior distribution . The later sampling operation helps to select the best response in that mode , trading off the need to pick the highest posterior with the uncertainty introduced by having several highly likely responses in close proximity . It may even be that responses near the mean of the posterior are an accidental byproduct of this two-stage process . However , the difficult-to-distinguish alternatives to an exponentiated posterior point toward alternative combinations . One of these is a pure sampling approach: participants draw samples from the posterior distribution and sometimes take the maximum of the samples and sometimes take the mean . Another alternative combination is to ascribe all the variability to noise in the posterior distribution: using a noisy posterior distribution , sometimes participants take the maximum and sometimes they take the mean . To gain additional purchase on this question , we correlated the average response times of participants with the model parameters . It might be expected that if participants were using any of tradeoffs between effort and accuracy that there would be correlations between each participant’s average response time and the number of samples or the exponent that the super-model recovered . This kind of correlation has been found in previous work when looking two-alternative responses [36] . However , both within each experiment and across experiments , we found no reliable relationship between the either of these model parameters and the response times of participants ( see S1 Methods for details ) . On the surface , this null result could be considered evidence that participants use the maximum or mean of a noisy posterior distribution to produce their estimates and that the amount of noise in the posterior does not depend on participant effort . However , it could also be that participants have such differing goals for effort / accuracy tradeoffs that this washes out whatever correlations there are between response time and model parameters . Future work would provide stronger tests of these mechanisms using within-participant designs that manipulate rewards and time-pressure , along with emphasizing that the computer is not responding to participant behavior . There are many examples of discrete numerical tasks in everyday life , such as the examples of painters quickly assessing the size of a wall in order to buy the right number of paint cans or of the party hosts assessing the hunger of their guests when buying a discrete number of pizzas . In our experiments , we used a numerosity task and an area estimation tasks because both of these tasks are well studied and the likelihood distributions have been well characterized . This allowed us to quickly measure the standard deviation of the likelihood for each participant . If we had used less controlled stimuli , then we might have had to measure the full distribution of responses that each individual stimulus evoked in order to characterize the likelihood . Our laboratory tasks are similar to some everyday tasks . The numerosity task we used is similar to estimating the number of visible stars in the sky ( which does vary depending on the time of day and light pollution ) , and estimating the size of a rectangle shares some similarities to the example of the painter who needs to assess the area of wall . However , there are differences as well: the stars in the sky do not differ in size from night to night and painters can view a wall from many distances and angles before producing an estimate . With the right stimuli , it would be interesting to investigate real-life performance in discrete numerical estimation tasks . Our results demonstrate that people represent a surprising amount of complexity in their prior distribution with relatively few training trials and use this complex prior when making new estimates . Training complex priors has multiple benefits: we can more easily observe how people represent priors and we can investigate some of the more complex schemes describing how people convert the posterior distribution into a single estimate . This work raises many questions about how prior and posterior distributions are represented and how estimates are made . Discrete numerical estimation tasks , which are simple to implement and quick to train , are well suited for future work in this area .
Twenty-one University of Warwick students participated in this experiment for course credit . Participants gave written informed consent and the experiment was approved by the University of Warwick Humanities and Social Sciences Research Ethics Committee . Participants were divided into three groups and each participated in one version of the experiment , as outlined below . One participant was excluded because of computer error and second was only given one block of discrimination trials but was included in the analysis . The stimuli consisted of displays of a number of identical dots . Each display of dots consisted of white dots on a black background , visible for 500 ms . Dot radius and dot density were randomized for each display to encourage participants to make numerosity judgments instead of judging the amount of light produced by the display , the density of the display , or the area occupied by the dots . In a single display all dots had a single common radius of between 3 and 9 pixels that was chosen randomly with equal probability on each trial . Dots were positioned randomly within a circular available region which was centered on the display , subject to the constraint that no dot could lay within one-dot-diameter of another dot . The available region randomly varied between 150 and 450 pixels in radius . A uniform draw was made over the possible values of dot density ( where density equaled the maximum number of dots that could appear in that block divided by the area of the available region ) , which determined the radius of the available region on a trial . The experiment consisted of a single session with three blocks . The estimation trials , in which participants saw a single display of dots and responded with their estimate of the number of dots in each display , were presented in the second block . The estimation trials differed for the three groups of participants: the narrow group , the medium group , and the wide group . The narrow group consisted of eight participants who saw 800 estimation trials in which the number of dots varied between 23 and 29 . For this group , displays with 23 and 29 dots each appeared with probability 0 . 3 and the displays with the remaining numbers appeared with probability 0 . 08 . The medium group consisted of six participants who saw 700 estimation trials in which the number of dots varied between 23 and 32 . For this group , displays with 23 and 32 dots each appeared with probability 0 . 3 and the displays with the remaining numbers appeared with probability 0 . 05 . The wide group consisted of six participants who saw 700 estimation trials in which the number of dots varied between 23 and 35 . For this group , displays with 23 and 35 dots each appeared with probability 0 . 28 and the displays with the remaining numbers appeared with probability 0 . 04 . Every participant saw 10 practice estimation trials displaying between one and four dots before beginning the main phase of the experiment . The first and third block consisted of 128 discrimination trials each ( always proceeded by 4 practice trials ) , in which participants saw two sequential displays of dots and picked the display that contained the larger number of dots . On every discrimination trial , one of the displays had a specific high or low number of dots . These anchor numbers were set to be either 11 dots below or above the lowest number seen in the estimation trials . The other display consisted of a number of dots that was equal to the anchor plus an offset . The offset was randomly chosen with equal probability from the set of {-8 , -4 , -2 , -1 , 1 , 2 , 4 , 8} . Because of computer error one participant , in the group that had a range of 23 to 29 in the estimation trials , was given anchor trials of 18 and 54 , in his or her first block . This participant received the correct anchor trials ( 12 and 40 ) in the third block . On estimation trials , after the dot display disappeared , participants were asked to enter the number of dots that they saw . After entering their response , participants received feedback about whether they were correct and the actual number of dots that were shown . On discrimination trials , participants only received feedback about whether they were correct or not . Twelve University of Warwick students participated in this experiment for £6 apiece . Participants gave written informed consent and the experiment was approved by the University of Warwick Humanities and Social Sciences Research Ethics Committee . Participants were divided into two groups and each participated in one version of the experiment , as outlined below . The stimuli consisted of displays of rectangles of particular areas . For each display , the width of the rectangle and its position were randomized to encourage participants to judge the area of the rectangles without exclusively relying on its length , width , or position on the screen . Rectangle width was chosen from a continuous uniform distribution between 2cm and 10cm , with length chosen to achieve the desired area given the width . A fixed positional jitter was chosen for each trial uniformly from a 6cm square . On estimation trials the rectangle was on average in the center of the screen and appeared for 500ms . On discrimination trials , the first rectangle appeared 10cm left of center plus positional jitter for 500ms , and after a 500ms delay the second rectangle appeared 10cm right of center plus positional jitter for 500ms . The procedure was identical to Experiment 1 with the following exceptions . Participants responded in the estimation trials with the area of the rectangle in square cm . A narrow group and a medium group was run in this experiment consisting of six participants each , with equivalent numbers and probabilities to the groups with the same names in Experiment 1 . Every participant saw 700 estimation trials in this experiment . Twenty-four University of Warwick students participated in this experiment for £6 apiece . Participants gave written informed consent and the experiment was approved by the University of Warwick Humanities and Social Sciences Research Ethics Committee . Participants were divided into three groups and each participated in one version of the experiment , as outlined below . Two participants were excluded from the rectangle group and one from the easier numerosity group because of computer error . Two additional participants from the easier numerosity group saw between ten and fifteen additional non-feedback trials at the beginning of the second estimation block and their data ( excluding these trials ) were included in the analyses . For all groups the experiment consisted of four blocks: the first discrimination block of 128 trials , the first estimation block of 500 trials , the second estimation block of 200 trials , and finally the second discrimination block of 128 trials . Feedback was given for all blocks except for the second estimation block which served as a test of whether a prior had been learned . The first discrimination and estimation blocks used easier-to-see displays than the second discrimination and estimation blocks . The three groups of participants differed in the details of the displays they were shown . The difficult numerosity group were run with the same display parameters as in Experiment 1 during the first discrimination and estimation blocks . During the second estimation and discrimination blocks , the first numerosity group was shown each dots display for 50ms and at a much reduced luminance . The easier numerosity group was given different display parameters during the first discrimination and estimation blocks: the range of the common radius of dots was from 6 to 9 pixels , while the available region randomly varied between 225 and 375 pixels . The easier numerosity group had a shorter display ( 50ms ) as well as more variability on the second discrimination and estimation blocks: the range of the common radius of dots was from 1 to 11 pixels , while the available region randomly varied between 150 and 450 pixels . The third group , the rectangle group , was given rectangles that randomly varied in width between 4 . 8 and 6 . 5cm during the first discrimination and estimation blocks . In the second discrimination and estimation blocks , this group was given a shorter ( 50ms ) and slightly dimmer ( gray instead of white ) display with rectangles that randomly varied in width between 2 and 10cm . All participants in this experiment were given the same trial structure during the estimation trials . The distribution that generated the trials was quadrimodal with a 20% chance of drawing each of the numbers 23 , 25 , 29 , or 31 . In addition , there was a 10% chance of drawing each of the numbers 24 and 30 . To estimate the variability in participants’ internal estimates , X , we analyzed the discrimination data in order to utilize fitted parameters for the estimation task . Specifically we assumed that the internal estimate was distributed according to a log-normal distribution ( in accordance with Weber’s law ) log ( X ) ∼N ( log ( X ) ; log ( S ) , σ2 ) where X and S are positive integers . Participants were presented with two stimuli , S1 and S2 ( as in the 2AFC discrimination trials ) and had to estimate which one was larger . In order to fit the variable σ we maximized the log-likelihood across trials ( or rather minimized the negative log-likelihood ) using Matlab’s fminbnd function σ ^ = arg max Σ i log P ( R i | S i , 1 : 2 , σ ) . The model likelihood P ( Ri|Si , 1: 2 , σ ) was estimated numerically for each trial and condition by sampling X1 and X2 10 , 000 times and for each set generating a fictitious response . P ( Ri|Si , 1: 2 , σ ) = ( 1/10000 ) Σl H ( Xl , 1 − Xl , 2 ) where H is the Heaviside function and log ( Xl , 1 ) ∼N ( log ( Xl , 1 ) ; log ( S1 ) , σ2 ) and log ( Xl , 2 ) ∼N ( log ( Xl , 2 ) ; log ( S2 ) , σ2 ) are samples from the generative model above . This analysis assumed that participants chose the most likely response on each trial , which is what was found in a recent analysis of 2AFC choice tasks [54] . The purpose of our analysis for the Estimation data is to compare and rule out different models of human decision making ( see Experiments 1–3 above ) . One common way of comparing perceptual models of different number of parameters ( see e . g . [55] ) is to fit the parameters of each model through maximum likelihood and compensate for differences in model complexity by calculating the Bayesian Information Criterion ( which penalizes models with large number of parameters ) . Our secondary analysis instead created a single model that encompasses all of the candidate models as special cases , that is for certain parameter sets the larger model is equivalent to each of the nested candidate models . The best fit of the parameters therefore shows which of the models best describes the data . We will first describe the generative model , how to perform inference over it , how to perform model comparison , and lastly we describe the ‘super-model’ and explain what specific models it encompasses . A traditional way of comparing models is by maximizing the likelihood of each model and correcting for number of parameters using the Bayesian Information Criterion [48] . We factorially combined two different priors ( Dirichlet or Gaussian kernel ) , three types of decision function ( mean , max or sampling ) and three types of noise models ( none , trembling hand or softmax ) to produce 16 different models for each participant ( note that softmax is not defined for a max decision function removing two of the 2x3x3 = 18 combinations ) . The priors were updated using either a variable width Gaussian kernel or “zero width” Dirichlet updating . Combined with the likelihood , this creates the posterior distribution P ( St|Xt ) upon which the subject bases their decision . A softmax noise model performs a transformation of the posterior P n ( S t = i | X t ) = P ( S t = i | X t ) β Σ j P ( S t = j | X t ) β ( 6 ) where β < 1 leads to a widening ( or flattening ) of the posterior , while β > 1 leads to a sharpening of the posterior . For noise models none or trembling hand we set β = 1 . Choices ( as discussed above ) are then made based on either max , mean ( average ) , or sampling of the exponentiated posterior . Finally , the trembling hand noise model , included as an alternative to the softmax model , states that participants have a small probability ϵ , of performing a random choice . I . e . S t ^ ∼ U [ 1 : 100 ] if e < ϵ , where e is randomly sampled from U[0: 1] . While the trembling hand noise model is structurally implemented after the decision is made , it has a similar effects as the softmax noise model in increasing the variability of the responses . As the internal variable Xt , was unknown to the experimenters we used ancestral sampling [28] , drawing 10 , 000 samples from the generative process for Xt ∼ P ( Xt|St ) followed by the inference by the subject as described above ( based on any model parameters ) . This generates 10 , 000 independent estimates of S ^ which we can use to numerically approximate the probability of response S t ^ . As they are essentially discrete counts we describe them as a discrete categorical distribution which provides us with the model likelihood P m ( S t ^ | S t , p a r ) for trial t and any model m and its associated parameters par . For each model and parameter set for each subject the log-likelihood was thus calculated as log L = Σ t log P ( S t ^ = R t | S t , p a r ) . Note that the parameter σ had been fit independently for each subject through the discrimination experiment above and was thus not a free parameter . Any trials with subject responses of less than 5 dots ( Rt < 5 ) were ignored as erroneous key presses given that the true number of dots presented were at least 23 . Furthermore to avoid any singularities in calculating the likelihood we allowed for a small probability ( 0 . 001 ) that participants would make a random response in the range [1:100] ( similar to a slight Trembling Hand ) . Any parameters par = ( β , ϵ , or ψ ) were fit through maximum likelihood ( Matlab’s fminsearch ) . For model comparison we calculated BIC for each subject and each model: B I C = - 2 * Σ t log P ( S t ^ = R t | S t , p a r ) + N * log ( M ) ( 7 ) where N is the number of model parameters in par ( 0 , 1 or 2 ) and M is the total number of trials . As an alternative to the model comparison we can compare models factorially based on how they update the prior ( given by parameter ψ ) , how many samples are drawn ( parameter n ) , and how exponentiated the posterior is ( parameter β ) . The fitted parameter set for each subject encapsulates which model aspects best explain the subject’s behavior . Thus the specific models compared above become special cases within this larger parameter space , allowing us to extrapolate between the models . We put all of these into a unified framework , which we refer as the ‘Super-model’ ( as other models are sub-sets of it ) . Given the posterior P ( St|Xt ) we assume that participants choose their response by averaging over samples: S t ^ = 1 n Σ i n s k ( 8 ) where the samples sk are given by s k ∼ P n ( S t = i | X t ) = P ( S t = i | X t ) β Σ j P ( S t = j | X t ) β ( 9 ) where the parameters n and β are fitted for each subject ( see below ) . The summation over samples allows us to approximate properties of specific decision functions . For n = 1 a single sample is drawn , equivalent to the sampling decision function . The averaging approximates the mean of the distribution for very large n ( thus approximating the mean decision function ) . The sampling of sk is from a softmax function ( also known as the exponentiated Luce choice rule [56 , 57] ) which causes all the probability density to be sharpened at the peaks of the posterior for larger values of β . For large values of β the number of samples ( n ) becomes of little consequence ( for example for β > 2 . 7 , with Xt = 23 and σ = 0 . 22 after learning for 300 trials , more than 95 percent of the probability is at the maximum a posteriori ) . In this way specific parameters emulate the mean ( average , n = 10000 ) , max ( β = 1000 ) and sampling ( n = 1 ) decision functions . In order to fit the variables ( ψ , n , β ) we performed log-likelihood maximization on ψ , β using Matlab’s fminsearch function ( on −logL with 5 random initializations ) , for each of n = [1 , 2 , 3 , 4 , 5 , 10 , 30 , 100 , 1000] . For each subject this allowed us to find the parameter set with the maximum likelihood and , given that the models of interest are nested models of this model-parameter set , indirectly find the model that best describes the data . | Studies of human perception and decision making have traditionally focused on scenarios where participants have to make estimates about continuous variables . However discrete variables are also common in our environment , potentially requiring different theoretical models . We describe ways to model such scenarios within the statistical framework of Bayesian inference and explain how aspects of such models can be teased apart experimentally . Using two experimental setups , a numerosity task and an area estimation task , we show that human participants do indeed rely on combinations of specific model components . Specifically we show that human learning in discrete tasks can be surprisingly fast and that participants can use the learned information in a way that is either optimal or near-optimal . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"learning",
"decision",
"theory",
"statistical",
"noise",
"decision",
"making",
"applied",
"mathematics",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"kernel",
"functions",
"cognitive",
"psychology",
"mathematics",
"probability",
"distribution",
"st... | 2016 | Fast and Accurate Learning When Making Discrete Numerical Estimates |
The host response to dengue virus infection is characterized by the production of numerous cytokines , but the overall picture appears to be complex . It has been suggested that a balance may be involved between protective and pathologic immune responses . This study aimed to define differential immune responses in association with clinical outcomes by gene expression profiling of a selected panel of inflammatory genes in whole blood samples from children with severe dengue infections . Whole blood mRNA from 56 Indonesian children with severe dengue virus infections was analyzed during early admission and at day −1 , 0 , 1 , and 5–8 after defervescence . Levels were related to baseline levels collected at a 1-month follow-up visit . Processing of mRNA was performed in a single reaction by multiplex ligation-dependent probe amplification , measuring mRNA levels from genes encoding 36 inflammatory proteins and 14 Toll-like receptor ( TLR ) -associated molecules . The inflammatory gene profiles showed up-regulation during infection of eight genes , including IFNG and IL12A , which indicated an antiviral response . On the contrary , genes associated with the nuclear factor ( NF ) -κB pathway were down-regulated , including NFKB1 , NFKB2 , TNFR1 , IL1B , IL8 , and TNFA . Many of these NF-κB pathway–related genes , but not IFNG or IL12A , correlated with adverse clinical events such as development of pleural effusion and hemorrhagic manifestations . The TLR profile showed that TLRs were differentially activated during severe dengue infections: increased expression of TLR7 and TLR4R3 was found together with a decreased expression of TLR1 , TLR2 , TLR4R4 , and TLR4 co-factor CD14 . These data show that different immunological pathways are differently expressed and associated with different clinical outcomes in children with severe dengue infections .
Dengue disease is emerging in the developing world at an alarming rate [1] . Moreover , the disease is potentially lethal , but only supportive measures are available . A major obstacle in the development of novel therapeutic strategies is that the pathophysiology of dengue disease is poorly understood [2] , [3] . Clinically , dengue disease is a mosquito-borne disease especially affecting children in endemic , mostly tropical regions and is caused by infection with dengue virus , a member of the Flaviviridae family [4] . Though most patients only suffer from a mild febrile illness called dengue fever ( DF ) , a relatively small group of patients experiences more severe forms of disease which are characterized by increased vascular permeability leading to bleeding manifestations and shock , such as dengue hemorrhagic fever ( DHF ) and dengue shock syndrome ( DSS ) [5] . The mechanism leading to severe , critical disease is insufficiently understood . Antibody-dependent enhancement of viral replication is generally thought to play a role since most cases of DHF/DSS occur after secondary infection by a dengue virus serotype different from the first infection [2] , [6]–[8] . Severe disease is characterized by high viremia titers , coagulation abnormalities and activation of the immune system [5] , [9] . The host defense strongly depends on interferon ( IFN ) -γ production , but other cytokines are also involved , as well as antibody production and T cell responses [10] , [11] . Typically , cytokine levels are highest among the day of defervescence , which is defined as the first day after the start of fever that the body temperature returns to normal . At the same day viremia titers drop to low or undetectable levels , whereas the risk of shock development is increased [12] . The relationship between cytokine production and severe disease is not fully understood , but it is increasingly recognized that a balance is involved between protective and pathologic immune responses [2] . In general , cytokine production depends on the recognition of pathogens via Toll-like receptors ( TLRs ) . In viral disease , TLR3 , TLR7 , TLR8 and TLR9 are commonly involved , but specific knowledge about TLRs in dengue infections is limited [13]–[15] . RNA-based gene expression profiling methods offer the opportunity to examine complex biological processes in a large context . We previously developed and validated an RT-PCR based assay for the measurement of multiple mRNA levels in a single reaction , multiplex ligation-dependent probe amplification ( MLPA ) [16]–[19] . The MLPA panel includes 36 target genes of various mediators of inflammation , such as cytokines , chemokines and intracellular signalling molecules and in addition a second panel has been developed to measure 14 TLR associated genes [20] . In order to investigate patterns of innate immune responses in severe dengue infections in children at the level of gene transcription , blood samples were collected from a group of children with severe dengue disease in Indonesia . The samples were analyzed using MLPA technology and associated with clinical outcomes .
The study was designed as a prospective , observational cohort study and was conducted in the Dr . Kariadi University Hospital in Semarang , Indonesia , simultaneously with an other study investigating the pathophysiology of hemorrhagic tendencies in dengue virus infections [21]; from this cohort , only patients were included for whom RNA was successfully isolated , which was available for all patients consecutively admitted from February 2002 until March 2003 . The ethics committee/ institutional review board of the dr Kariadi University Hospital in Semarang , Indonesia , approved all legal , ethical , clinical and laboratory aspects of the study including the informed consent procedure , which was obtained , after informing the patients and their parents and/or guardians in local language , as a written form from children's parents or guardians before inclusion . Children , aged 3 to 14 years , admitted to the paediatric intensive care unit or a ‘high nursery care unit’ with a clinical diagnosis of suspected DHF or DSS were included . Demographic and clinical data were collected using a standardized data collection form . A Tourniquet test was performed to detect a bleeding tendency together with chest- and abdominal X-rays and ultrasonography to detect ascites and/or pleural effusion . Children were classified as having DF , DHF or DSS according to WHO criteria [1] . In brief , the clinical diagnosis of DHF included the presence of fever , a hemorrhagic tendency , thrombocytopenia ( <100×103/mm3 ) , evidence of pleural effusion and/or a significant ( >20% ) rise , or drop after volume replacement therapy , of hematocrit values . DSS was defined as DHF with evidence of circulatory failure . All other cases not meeting these criteria were diagnosed DF . Citrated blood samples were obtained on the day of admission , days 1 , 2 and 7 after admission and at a 1-month follow-up visit . Because in dengue disease each patient experiences a different clinical course and is hospitalized at different time points in that course , these blood samples were referred to a common clinical time point , which is the day of defervescence according to international agreement . In addition , samples obtained at admission from patients presenting >1 day before defervescence were clustered into a time point designated ‘Early’ , which indicates early immune activation before defervescence . Paired blood samples were tested for serologic evidence of acute dengue infection . Dengue virus specific IgG and IgM antibodies were measured by ELISA ( Focus Technologies , Cypress , Ca , USA ) . The sensitivity and specificity of these tests were evaluated previously [22] . Cases were considered serologically-confirmed if the IgM ELISA was positive during the acute phase of disease ( optical density of the sample higher than the optical density of the cut-off serum provided by the manufacturer ) and/or if a four-fold increase in IgG titre was demonstrated in paired acute and convalescent sera . For some patients , a definitive serodiagnosis was not possible because no convalescent sample was obtained . Dengue virus antigen and/or viral RNA was detected in these cases using a dot blot immunoassay and/or a dengue serotype specific reverse transcription PCR , respectively [23] . Patients with definitive serodiagnosis and/or positive dot blot and/or positive PCR were considered to have confirmed dengue virus infection . Blood samples were centrifuged within 1–2 hours after retrieval at 15°C for 20 minutes at 1600×g . Subsequently , plasma was separated and red blood cells were lysed using Red Blood Cell lysis Buffer ( Roche , Mannheim , Germany ) according to the manufacturer's protocol . Total white blood cells were stored at −80°C in mRNA lysis/-binding buffer ( Roche ) and transported to the Netherlands on dry ice where total mRNA was isolated using High Pure RNA Isolation Kit ( Roche ) . The mRNA of multiple inflammatory molecules was analyzed by MLPA as described previously [16] . In addition a second panel was added measuring the involvement of multiple TLR related genes [20] . MLPA is insensitive to the total amount of mRNA that is included in the reaction; therefore , the profile is independent of the total white blood cell ( WBC ) count . All samples were tested with the same batches of reagents , and a negative and endotoxin-stimulated control sample were included on each plate . The final polymerase chain reaction ( PCR ) fragments amplified with carboxyfluorescein-labeled primers were separated by capillary electrophoresis on a 16-capillary ABI-Prism 3100 Genetic Analyzer ( Applied Biosystems , Nieuwerkerk aan de IJssel , The Netherlands ) . Peak area and height were processed using GeneScan analysis software ( Applied Biosystems ) . The levels of mRNA for each gene were expressed as a normalized ratio of the peak area divided by the peak area of a control gene , beta 2 microglobulin ( B2M ) , resulting in the relative abundance of mRNAs of the genes of interest . All calculations were carried out using SPSS version 13 . 0 ( SPSS Inc . , Chicago , IL , USA ) . The relative mRNA levels of subjects from time-point ‘Early’ and day −1 , 0 , 1 and 5–8 post defervescence were compared to baseline levels collected at a 1-month follow up visit using Wilcoxon signed ranks test . mRNA's with >80% of values below the detection limit were excluded from analysis and considered not detectable . P-values for each time point were corrected for multiple comparisons according to the method of Benjamini and Hochman , with q = 0 . 1 [24] . Associations with clinical variables were calculated by univariate logistic regression analysis and odds ratios±SE are presented . A P<0 . 05 was considered statistically significant .
Samples for mRNA analysis were collected from 56 children with confirmed dengue virus infections . All children presented with clinically severe disease; the patients were classified according to WHO criteria [1] as having DF ( n = 7; 13% ) , DHF ( n = 29; 51% ) or DSS ( n = 20; 36% ) . Twenty patients presented more than 1 day before defervescence ( median 3 days before defervescence; interquartile range 2–3 ) . These patients had experienced symptoms since 4 days before defervescence ( median; interquartile range 3–5 ) . Further patient characteristics at admission are given in Table 1 . During stay at the hospital , 10 patients ( 18% ) developed one or more complications related to their dengue infections . Profound shock was noted in 7 children ( 13% ) and recurrent shock in 5 children ( 9% ) . Other complications included disseminated intravascular coagulation ( n = 3; 5% ) , encephalopathy ( n = 3; 5% ) or pulmonary edema ( n = 1; 2% ) . Eventually , 4 patients ( 7% ) died despite treatment . These patients had all been classified as having DSS; two of them died from profound shock; one patient also had recurrent shock and encephalopathy and one patient suffered in addition from disseminated intravascular coagulation . The surviving patients all made a complete recovery within a maximum of 12 days of stay in the hospital . Gene expression profiles of 36 inflammatory genes during hospital stay of each fever day were compared to baseline values at a 1-month follow up visit ( Table 2 ) . The profile showed up-regulation of 8 genes , down-regulation of 7 genes and no effect on 13 genes , whereas 8 genes were not detectable . Most gene expression profile changes occurred on the day of defervescence ( day 0 ) . The expression patterns revealed up-regulation of the Th1 type cytokines interferon-γ ( IFNG ) and interleukin-12a ( IL12A ) on consecutive days , indicating an antiviral response . Another gene strongly induced by dengue infection was MIF , a marker of macrophage activation . No other cytokine genes from the MLPA panel were up-regulated , but some other genes were involved , including cell cycle control genes such as cyclin-dependent kinase inhibitor 1A ( CDKN1A ) at consecutive days and polyadenylate-specific ribonuclease ( PARN ) and protein-tyrosine phophatase nonreceptor-type 1 ( PTPN1 ) at day 0 only . Furthermore , glutathione S-transferase ( GSTP1 ) , involved in detoxification metabolism , was significantly up-regulated at day −1 and day 1 and MD2 , a co-factor of TLR4 , was activated early before defervescence . Of note , in contrast to the enhanced levels of antiviral cytokine mRNAs , down-regulation of genes was observed at consecutive days for a whole cluster of genes associated with the nuclear factor ( NF ) -κB pathway ( Figure 1 ) . Transcription levels of NF-κB subunit 2 ( NFΚB2 ) and tumor necrosis factor receptor 1 ( TNFR1 ) were decreased as well as levels of the cytokine genes IL1B at day −1 and IL8 , TNFA and NF-κB subunit 1 ( NFΚB1 ) at day 0 . Furthermore , transcription was inhibited during the early phase of PTPN4A2 . The gene expression profiles of 14 TLR associated genes during admission showed up-regulation of 2 genes , down-regulation of 4 genes and no effect on 6 genes , whereas 2 genes were not detectable ( Table 3 ) . Most gene expression profile changes occurred at day −1 and day 0 . Up-regulation of TLR gene transcription was observed for TLR7 and TLR4 transcript variant 3 ( TLR4R3 ) during the early phase and at day −1; in addition , TLR7 was up-regulated at day 0 and TLR4R3 at day 1 . Decreased expression of TLRs was noted for TLR1 , TLR4R4 and TLR4 co-factor CD14 at day −1 and day 0 , whereas decreased expression of TLR4R4 was also noted during early activation and for TLR2 at all time-points . No detectable levels were measured of the viral recognition associated TLR genes TLR3 and TLR9 . Clinical associations of gene expression levels with parameters of disease severity were calculated at day −1 , 0 and 1 from defervescence . Odds ratios and p-values are presented in Table 4 . The development of DSS correlated significantly with down-regulation of IL4R2 . Pleural effusion as diagnosed by either chest X-ray or ultrasonography was associated with up-regulation of members of the NF-κB pathway such as TNFA and TNFR1 , but also with IL1RA , MIP1A and TLR1 and was inversely related to MIF and TLR7 . Hemorrhagic manifestations including mild and/or severe bleeding were associated with the NF-κB pathway members NFKB1 and TNFR1 . The occurrence of severe complications such as profound or recurrent shock , disseminated intravascular coagulation , encephalopathy and/or pulmonary edema was not significantly associated with genes from the panel but CD14 , IL1B , IL1RA , NFKB1 , PDE4B , TLR1 and TLR10 showed a trend towards significance , particularly during defervescence .
The pathogenesis of severe critical disease following dengue virus infection involves the activation of multiple inflammatory pathways . This is the first study to report high throughput gene expression profile changes of inflammatory genes in children with severe dengue virus infections . The study revealed two points of interest . First , the profile showed characteristics of a general antiviral response with up-regulation of IFNG , which is an established major antiviral cytokine in dengue disease , together with up-regulation of IL12A , a potent stimulator of IFN-γ production [11] , [25] . In contrast , no other cytokine genes from the panel were up-regulated , but a cluster of genes related to the NF-κB pathway was down-regulated , including NFΚB1 , NFΚB2 , TNFR1 , IL1B , IL8 and TNFA . In addition , the NF-κB pathway linked expression patterns of NFΚB1 , TNFR1 and TNFA correlated with adverse clinical events , such as the development of pleural effusion and hemorrhagic manifestations . A second finding of interest in this study is that this is the first report of in vivo differential activation of multiple TLRs in dengue disease , which may implicate a role for these receptors in dengue infections . The intracellular NF-κB pathway is a major route for inflammatory stimuli to the release of most pro-inflammatory mediators , including TNF-α , IL-1ß and IL-8 [26] . Most of these mediators were reported to be elevated in plasma from patients with severe dengue disease , especially around the day of defervescence [27] , [28] . In contrast , studies showed poor NF-κB related cytokine production after stimulation of peripheral blood mononuclear cells which were isolated from patients with dengue [29]–[31] . Also in the current study most in vivo NF-κB pathway associated gene transcription levels were down-regulated . These results may be limited to the site of production , i . e . located to circulating leucocytes , because in vitro data showed that dengue virus was able to stimulate the NF-κB pathway in other , non leukocyte cells , such as human hepatoma cells and endothelial cells [32] , [33] . Yet , no studies have directly compared different cell types or investigated the mechanism or function of NF-κB down-regulation by dengue . Down-regulation of inflammatory genes during the day of defervescence may also be the result of time effects such as depletion- or negative feed-back mechanisms following previous activation of these genes . Time effects play an important role in dengue disease; the immune response can be roughly divided into early inflammatory changes , changes during defervescence and late changes after defervescence [12] , [34] . An exploration of ‘early’ time effects was carried out in our study by clustering all early admission samples from patients entering the hospital more than one day before defervescence . However , because no major differences were noted in comparison to changes observed during defervescence , it shows no evidence that early time effects affecting the investigated genes interfered with later changes during defervescence . The production of cytokines such as INF-γ indicates activation of an appropriate , protective immune response to dengue virus infection . However , overwhelming production of cytokines and/or cross-activation of pathological immune responses has been postulated as a mechanism to endothelial damage and subsequent vascular leakage , shock and death . Previous studies showed direct associations between mortality and increased protein plasma levels of IL-6 , IL-8 , IL-10 , IL-1RA and MIF [31] , [35] , [36] . In addition , increased plasma levels of TNF-α , TNFR1 , TNFR2 , IL2R , IL-8 , IL-13 , IL-18 , soluble CD8 and TGFß1 protein plasma levels were associated with development of DSS , whereas levels of TNF-α also correlated with hemorrhagic manifestations and levels of TNFR1 were associated with plasma leakage [27] , [29] , [35] , [37]–[40] . Within this context , it is important to define which immune responses are protective and which responses play a mere pathological role . Clustering of the confusingly large amount of mediators which appear to be involved in dengue into collective pathways can help to retrieve a clearer picture , if possible . A previous microarray based study revealed this way that activation of interferon-I dependent pathways was associated with beneficial outcomes in adult patients with dengue [41] . The current study did not investigate interferon-I dependent pathways , but included the genes IFNG and IL12A and found no evidence of an association of these genes with adverse outcomes . In contrast , the current study found associations of various other inflammatory genes with adverse outcomes and it showed that at least for some part these associations could be clustered to members of the NF-κB pathway: TNFA and TNFR1 were highly associated with pleural effusion ( odds ratios 118 . 3±59 . 6 and 31 . 4±15 . 9 , respectively ) and NFKB1 and TNFR1 correlated with hemorrhagic manifestations ( odds ratios 14 . 0±6 . 4 and 14 . 0±6 . 9 , respectively ) . In addition , a trend towards significance was found for association of the NF-κB pathway related genes IL1B and NFKB1 with the occurrence of severe complications and for IL8 and NFKB2 with pleural effusion . Of note , TNF-α is one of the major NF-κB pathway effector molecules and others showed that in two independent mouse models of dengue infection , antibodies to TNF-α significantly decreased dengue disease severity and mortality in vivo [42] , [43] . Also , it was demonstrated that addition of serum from patients with acute dengue infection to endothelial cells in vitro induced increased expression of the endothelial cell activation marker ICAM-1 and that this effect could be blocked by antibodies to TNF-α [44] . Based upon these data all together , activation of TNF-α in particular , but possibly of the NF-κB pathway as a whole , may be involved in the mechanism of vascular leakage and transition into severe disease in patients with dengue . Pathogen recognition is an important step in the activation of inflammatory pathways and TLRs play a major role in this process [26] . Each TLR is associated with specific ligands and response patterns , though interactions occur and in fact each pathogen is recognized by not only one TLR but by a set of TLRs , usually together with also other pattern-recognition receptors [45] . In the current study , the in vivo role of ten TLRs during dengue infection was explored and differential gene expression of TLRs was found: increased expression of TLR7 and TLR4R3 together with a decreased expression of TLR1 , TLR2 , TLR4R4 and TLR4 co-factor CD14 . These results show that the transcription profile of the total white blood cells is perturbed in children with severe dengue disease; the lack of clustered data hampers further interpretation of these results . The in vivo role of TLR7 in particular though , which was up-regulated in our panel and associated with a lack of pleural effusion in patients during fever day -1 and day 0 ( odds ratios 4 . 05±1 . 98 and 5 . 24±2 . 55 , respectively ) may be of interest to be further explored , also because previous investigators identified TLR7 as an endosomal pattern-recognition receptor for single-stranded RNA viruses including influenza virus and vesicular stomatitis virus [14] , [46] and a recent study showed that TLR7 could be triggered by dengue virus in vitro [47] . Gene profiling methods offer opportunities to examine biological processes in great detail . Yet , due to the amount of data and because mRNA levels may be subject to posttranslational processing there is risk of over interpretation of results . Moreover , because the current study is an observational study , it has difficulty to distinguish between primary and secondary effects [48] . Still , the current MLPA panel was validated before and proved useful for the study of molecular mechanisms during various inflammatory conditions [16]–[20] . The statistical error in the present study was reduced as much as possible by applying Benjamini-Hochman statistical correction for multiple testing [24] . Furthermore , the error was reduced by testing at several days independently from each other: still , most gene expression changes occurred at several days consecutively . In addition , most changes were noted at the day of defervescence with a more or less gradual decline of changes during recovery; this pattern closely resembles data from others [27] , [34] . As such , the data suggest that the observed effects indeed represent effects related to the dengue infection . The children included in our study were all clinically severely ill; only these patients were included because gene array profiling studies investigating intra-individual gene transcription changes , such as differential activation of differential immunological pathways , are best performed in extreme cases . Not all children in our study could be identified as severe cases by WHO criteria: 7 patients ( 13% ) were diagnosed as having DF . The limitations of the WHO criteria to identify all patients suffering from clinically severe disease have been observed before [49] , [50]; to circumvent this problem , clinical associations were calculated not only for fulfilment of DSS criteria , but also for other disease severity parameters such as occurrence of complications , pleural effusion and hemorrhagic manifestations . No associations were calculated for mortality because of the low incidence ( n = 4; 7% ) . With the limitations of an observational , gene array profiling study in mind , the data from our study provide a first insight into the molecular basis of inflammatory gene expression patterns in peripheral blood leukocytes from children with severe dengue infections in vivo . The profile showed up-regulation of the antiviral cytokines IFNG and IL12A and down-regulation of the NF-κB pathway . The function of these differential gene expression patterns is not precisely clear , but may be related to multiple associations between the NF-κB pathway and adverse clinical outcomes . In addition to these data , it was found that the expression of several TLRs was affected during severe dengue disease . | Dengue virus infection is an impressively emerging disease that can be fatal in severe cases . It is not precisely clear why some patients progress to severe disease whereas most patients only suffer from a mild infection . In severe disease , a “cytokine storm” is induced , which indicates the release of a great number of inflammatory mediators ( “cytokines” ) . Evidence suggested that a balance could be involved between protective and pathologic cytokine release patterns . We studied this concept in a cohort of Indonesian children with severe dengue disease using a gene expression profiling method . The study showed that the immune response to severe dengue infection was characterized by up-regulation of interferon pathway–associated cytokines and down-regulation of nuclear factor ( NF ) -kappaB pathway–associated cytokines . Many of the NF-kappaB pathway–related genes , but not the interferon pathway–associated genes , correlated with adverse clinical events such as development of pleural effusion and hemorrhagic manifestations . In addition , the study showed that the expression of specific pathogen recognition receptors called Toll-like receptors was differentially regulated in patients with dengue . Together , the data show that different immunological pathways may indeed be differently expressed and associated with different clinical outcomes in children with severe dengue infections . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"critical",
"care",
"and",
"emergency",
"medicine/pediatric",
"critical",
"care",
"pediatrics",
"and",
"child",
"health/pediatric",
"critical",
"care",
"microbiology/immunity",
"to",
"infections",
"genetics",
"and",
"genomics/gene",
"expression",
"infectious",
"diseases/neg... | 2008 | Differential Gene Expression Changes in Children with Severe Dengue Virus Infections |
Helminth parasites can cause considerable damage when migrating through host tissues , thus making rapid tissue repair imperative to prevent bleeding and bacterial dissemination particularly during enteric infection . However , how protective type 2 responses targeted against these tissue-disruptive multicellular parasites might contribute to homeostatic wound healing in the intestine has remained unclear . Here , we observed that mice lacking antibodies ( Aid-/- ) or activating Fc receptors ( Fcrg-/- ) displayed impaired intestinal repair following infection with the murine helminth Heligmosomoides polygyrus bakeri ( Hpb ) , whilst transfer of immune serum could partially restore chemokine production and rescue wound healing in Aid-/- mice . Impaired healing was associated with a reduced expression of CXCR2 ligands ( CXCL2/3 ) by macrophages ( MΦ ) and myofibroblasts ( MF ) within intestinal lesions . Whilst antibodies and helminths together triggered CXCL2 production by MΦ in vitro via surface FcR engagement , chemokine secretion by intestinal MF was elicited by helminths directly via Fcrg-chain/dectin2 signaling . Blockade of CXCR2 during Hpb challenge infection reproduced the delayed wound repair observed in helminth infected Aid-/- and Fcrg-/- mice . Finally , conditioned media from human MΦ stimulated with infective larvae of the helminth Ascaris suum together with immune serum , promoted CXCR2-dependent scratch wound closure by human MF in vitro . Collectively our findings suggest that helminths and antibodies instruct a chemokine driven MΦ-MF crosstalk to promote intestinal repair , a capacity that may be harnessed in clinical settings of impaired wound healing .
Infections with intestinal helminths affect more than 2 billion people globally [1 , 2] and drug-resistant helminths pose a considerable threat to agricultural livestock [3 , 4] . Helminth infections thus present a major global health concern particularly due to the propensity of these parasites to form chronic and repeated infections [5 , 6] . The potent immune-modulatory capacities of helminth parasites have been shaped by their long co-evolution with their hosts’ immune systems , which has resulted in a fine-tuned balance between inflammation on the one side , and parasite control on the other side [7] . In addition to their anti-inflammatory potential , helminths are increasingly recognized for their capacity to induce rapid tissue repair [8–10] . Considering the tissue migratory potential of the macroscopic larval stages , it is surprising that helminth infections are rarely associated with severe bleeding or bacterial sepsis [11] . This may partially be explained by parasite-mediated immune-modulation including the induction of IL-10 production in settings of bacterial translocation [11] . In addition , tissue dwelling nematodes typically initiate type 2 responses , which have been implicated in promoting tissue repair [12] . Helminths can trigger type 2 responses either directly [13] , or through the release of alarmins ( IL-25 , IL-33 , TSLP ) or adenosine from epithelial cells following invasion [14 , 15] . Although the mechanisms by which helminth invasion initiates type 2 responses are increasingly well described , relatively little is known about the effector mechanisms that promote repair of host tissues following larval migration . Models of acute lung damage or liver fibrosis caused by the helminths Nippostrongylus brasiliensis [8 , 9] or Schistosoma mansoni , respectively have demonstrated important roles for Arginase-1 ( Arg1 ) and insulin-like growth factor 1 ( IGF-1 ) expressing MΦ , and IL-9R-dependent production of innate lymphoid cells ( ILCs ) in controlling inflammation and tissue damage [9 , 16 , 17] . However , intestinal helminth infections are often chronic and proceed without the development of excessive fibrosis , suggesting that distinct mechanisms might contribute to homeostatic wound healing in the intestine . Moreover , little is known about a potential contribution of the immune memory to these processes , yet those living in endemic regions are continuously exposed to helminths and typically suffer repeated infections . Our previous work identified the helminth-antibody-driven activation of Arg1 expressing MΦ as an important mechanism to limit tissue disruption at early timepoints following repeated infection with the intestinal nematode Hpb [10] . Hpb is a natural parasite of mice with a strictly enteric life cycle that establishes chronic infections upon primary exposure [18] , but which can be rapidly targeted by immune antibodies following repeated infections [10 , 19 , 20] . Important overlaps between mechanisms of protective immunity and wound healing during helminth infection are beginning to emerge [8] , but the role of antibodies in wound healing has not been investigated . Our previous findings suggested that the combination of immune antibodies and helminth antigens can induce the expression of several MΦ genes with potential roles in tissue repair [10] . This included the CXCR2 ligands Cxcl2 and Cxcl3 , which have been implicated in cutaneous wound healing [21 , 22] . Here we demonstrate that intestinal wound containment was impaired in antibody- and antibody-receptor deficient mice and correlated with reduced expression of CXCR2 ligands . We further show that Fcrg-chain-signaling triggered the release of CXCL2 from MΦ and intestinal MF during helminth infection . Fcrg-dependent chemokine release by MΦ required stimulation by antibodies and helminths , whilst helminth products alone elicited the release of chemokines by MF through the dectin2 pathway , which is known to require Fcrg chain signaling . CXCR2 ligation was required for wound containment during murine helminth infection in vivo , and for efficient wound closure by human MF in vitro . Our findings reveal a dual role for helminths and immune antibodies in the trapping of invasive larvae and the promotion of tissue repair .
We have previously reported an important role for antibody-mediated MΦ activation in limiting parasite-induced tissue disruption and promoting larval trapping following secondary challenge infections with Hpb [10] . We also noted that the same MΦ exhibited a transcriptional profile similar to the one of wound healing MΦ [10 , 23] . To investigate whether antibodies also impacted on tissue repair we quantified the size of intestinal lesions at various time points post secondary challenge infection ( p . i . ) . A scheme describing the timing of Hpb infections , larval maturation ( L3–L5 ) and antihelminthic treatments is shown in Fig . 1A . At day 10 p . i . we could not observe significant differences in the size of intestinal lesions between wild-type ( WT ) mice and mice deficient in activation induced cytidine deaminase ( Aid ) which fail to produce affinity-matured , class-switched antibodies [24] ( Fig . 1B , C ) . We further observed a tendency of Aid-/- mice to develop increased numbers of intestinal lesions ( S1A Fig ) , which might be explained by the higher migratory capacity of larvae in the absence of antibodies . The cellular infiltration in the lesion underwent a dramatic change over time characterized by a shift from monocytic cells ( day 4 ) to granulocytes ( day 10 ) ( S1B Fig ) . However , in agreement with our earlier findings , lesions from Aid-/- and WT mice exhibited similar cellular compositions with the exception of reduced basophil numbers in the absence of Aid [10 , 25] ( S1C-H Fig ) . The finding that lesions were of similar size in Aid-/- and WT mice was rather surprising as larvae had already left the intestinal mucosa of Aid-/- mice at this time point , whilst they remained trapped within antibody sufficient WT lesions ( Fig . 1B ) . We therefore additionally quantified lesion sizes in challenge-infected WT and Aid-/- mice , or mice deficient in activating antibody receptors ( Fcrg-chain-/- ) , at later timepoints . Intact larvae were absent from the lesions of most WT mice by day 14 p . i . , possibly as a result of degradation by the large numbers of infiltrating granulocytes ( Fig . 1D , S1B Fig ) . Interestingly , lesion size had decreased slightly but not significantly ( by around 10% ) by day 14–21 p . i . as compared to day 10 p . i . in WT mice whilst Aid-/- and Fcrg-/- mice showed significantly larger lesions ( Fig . 1D , E ) . Impaired lesion contraction in antibody and antibody receptor deficient mice correlated with a reduced accumulation of α smooth muscle actin ( aSMA ) expressing cells , which include MF , the major cell type involved in wound contraction [26] ( Fig . 1F , G ) . To further investigate the impact of antibodies on the resolution of intestinal lesions we set up experiments to investigate lesion size at even later timepoints ( day 42 ) post secondary challenge infection . However , most of the experiments had to be halted at day 21 post infection as the majority of challenge-infected Aid-/- and Fcrg-/- mice showed symptoms of morbidity ( most likely severe peritonitis ) around this time point . Indeed only 3 Aid-/- mice survived symptom free until day 42 post challenge infection . When comparing intestinal lesions in WT and these surviving Aid-/- mice , an increased tendency to develop necrosis or fibrosis as well as impaired lesion resolution became apparent in the antibody deficient mice ( S2A , B Fig ) . Increased inducible nitric oxide synthase expression in peritoneal cells in Aid-/- as compared to WT mice already at day 14 post challenge infection hinted at increased peritonitis [27] in the absence of antibodies , possibly explaining the considerable morbidity in these mice late after challenge infection . Taken together these findings suggest that antibodies act not only to trap helminth larvae but also to promote the accumulation of aSMA+ cells that are important for mucosal lesion contraction . In addition , the efficient ( antibody-dependent ) resolution of intestinal lesions might by associated with the control of peritoneal inflammation during challenge infection with intestinal helminths . To investigate if type 2 immunity was intact in antibody and FcRg deficient mice , we analyzed peritoneal eosinophil accumulation and type 2 cytokine ( IL-4 and IL-13 ) production in the duodenum as key parameters of the local memory TH2 response . As shown in S2D Fig , large numbers of eosinophils infiltrated the peritoneal cavity during challenge infection ( day 14 and 21 ) in WT , Aid-/- and Fcrg-/- mice . Moreover , IL-4 and IL-13 were both detectable in duodenal culture supernatants of all challenge-infected mice with minor reductions in Aid-/- or Fcrg-/- mice , respectively ( S2E , F Fig ) . In contrast , during primary infection the levels of both cytokines were close to the levels in naïve mice ( limit of detection 4 pg/ ml ) ( S2E , F Fig ) . Together these data indicate that the TH2 response in challenge-infected antibody and FcRg deficient mice resembles that in challenge-infected rather than in primary infected WT mice . We reasoned the increased susceptibility of challenge-infected antibody and FcRg deficient mice to peritonitis may be due to defective lesion repair resulting in increased bacterial translocation across the intestinal epithelium . To test this hypothesis , we quantified soluble CD14 levels in the peritoneum of primary infected WT and challenge-infected WT , Aid-/- or Fcrg-/- mice . As shown in S2G Fig , sCD14 levels did not increase above those found in naïve mice following challenge infection of WT , Aid-/- or Fcrg-/- mice , whilst high levels of peritoneal sCD14 could be detected following primary infection of WT mice ( day 14 ) . These data suggest that immune memory may function to prevent bacterial translocation , but through mechanisms distinct from the impact of immune antibodies on wound contraction or peritonitis . As our data suggested a central role for antibodies in lesion contraction during challenge Hpb infection , we sought to confirm a role for “immune memory” in this process . We first compared lesion sizes in primary and secondary Hpb infected WT mice . At day 14 and 21 , primary lesions were less dense with fewer infiltrating cells ( S2H Fig ) , but of similar size ( S2I Fig ) to those seen at the same time-points following challenge infection . However , it should be noted that significantly more lesions were observed in challenge-infected mice compared to primary-infected mice ( S2J Fig ) , probably as a result of the greater type 2 immune response observed following challenge infection ( S2D Fig ) . To determine whether lesion resolution differed between primary or challenge infected mice we next analyzed lesion size at a later time-point , day 42 . By this time-point lesions had started to contract in challenge-infected , but not primary infected , mice ( S2I Fig ) , indicating that immune memory can result in more rapid wound contraction . To further support a role for ‘immune memory’ in lesion contraction we determined lesion size in antibody deficient ( Aid-/- ) mice receiving serum from naïve or challenge-infected WT mice . As shown in Fig . 1H , intestinal lesions in Aid-/- mice that received immune serum from challenge-infected WT mice exhibited lesions of a similar size to that seen in WT mice . In contrast , transfer of naïve serum into Aid-/- recipients did not result in a significant improvement of intestinal wound repair ( Fig . 1H , I top panel ) . Moreover , the improved lesion contraction in immune serum treated Aid-/- mice correlated with an increased accumulation of aSMA+ cells within the lesions of these mice as compared to lesions in untreated Aid-/- mice ( Fig . 1I bottom panel , J ) . These data indicate that immune memory acts to promote both lesion containment ( day 14–21 post-infection ) and lesion contraction ( day 42 post-infection ) . Our previous work had identified the chemokines Cxcl2 and Cxcl3 as being amongst the most highly up-regulated MΦ genes following culture of these cells together with helminth larvae and IS [10] . As CXCL2 and CXCL3 have been reported to promote repair processes , including smooth muscle cell migration [21 , 28] , we investigated the expression of these chemokines during helminth infection in vivo . We observed that at day 14 following primary infection or day 10 following challenge infection , CXCL2 & 3 were both expressed at low levels within intestinal lesions ( S3A , B Fig ) . However by day 14 post challenge infection , CXCL2 was abundantly expressed , particularly at the lesion edge where it co-localized with FGFR1 expressing MF and F4/80 expressing MΦ ( Fig . 2A ) . As FGFR1 is also expressed by fibroblasts , we confirmed the co-expression of aSMA in FGFR1 expressing cells in the periphery of intestinal lesions , which identified these cells as myofibroblasts ( S3C Fig ) . We further quantified CXCL2 expression intensity in areas with a comparable MΦ and MF infiltrate ( S3D , E Fig ) to exclude a potential bias due to lower numbers of MF in lesions of antibody ( receptor ) deficient mice . As shown in Fig . 2B , Aid-/- and Fcrg-/- mice presented a significantly reduced CXCL2 staining intensity on a per cell basis , which was due to lower intensities in both MΦ and MF ( Fig . 2C ) . Interestingly , CXCL2 expression in lesions of Fcrg-/- mice was significantly reduced as compared to that in Aid-/- lesions , suggesting that antibody independent but FcRg chain dependent signaling events might contribute to CXCL2 production during helminth infection in vivo . In keeping with the observed reduction in aSMA+ cells in lesions of Aid-/- and Fcrg-/- mice ( Fig . 1F , G ) , MF in these lesions expressed lower levels of FGFR1 ( S3F Fig ) , which is required for MF differentiation , including the upregulation of aSMA expression [29] . As our immunofluorescence ( IF ) analysis could not distinguish between chemokine production or surface binding , we additionally analyzed CXCL2 secretion into supernatants from duodenal organ cultures from challenge-infected mice . Small-intestinal tissue from infected WT mice secreted 3-fold more CXCL2 as compared to naïve mice , whilst Aid-/- or Fcrg-/- tissues secreted levels similar to that of naïve controls ( Fig . 2D ) . Thus , increased levels of CXCL2 might be secreted locally in response to antibody-trapped larvae . To investigate whether antibodies and antibody receptors additionally impacted on the production of the close homologue CXCL3 , we performed IF stainings on the corresponding serial sections of small intestinal lesions . As shown in S3G , H Fig , CXCL3 showed a similar pattern of expression as CXCL2 with a significant reduction in Fcrg-/- mice . Finally , we analyzed CXCL2 in Aid-/- mice that had received WT immune sera during challenge infection . As shown in Fig . 2E , F , transfer of immune but not naïve sera could partially restore chemokine production in antibody deficient mice . These data suggest that the CXCR2 ligands , CXCL2 & 3 , are upregulated in intestinal lesions by mechanisms involving Fcrg-chain-mediated activation of MΦ and/or MF by immune antibody-trapped helminth larvae . We had previously demonstrated that the IS-induced trapping of infective helminth larvae ( ”L3” ) by bone marrow derived MΦ ( BMM ) was associated with a strong transcriptional activation of Cxcl2/3 , which occurred independently of Fcrg-chain signaling [10] . To verify a direct effect of larvae and IS on CXCL2 secretion by MΦ , we quantified CXCL2 in supernatants from BMM after co-culture with helminth larvae in the presence or absence of IS . IS alone had no effect on CXCL2 release by WT BMM , whilst larvae induced CXCL2 secretion , which was enhanced by IS ( Fig . 3A ) . Moreover , Fcrg-/- BMM showed a considerably reduced CXCL2 release compared to WT BMM , when co-cultured with larvae in the presence of IS . This suggested that Fcrg-mediated activation of MΦ was responsible for the enhanced release , but not transcription of Cxcl2 . Indeed , unstimulated BMM showed considerable amounts of intracellular CXCL2 ( Fig . 3B ) , but no secretion ( Fig . 3A ) , whilst intracellular CXCL2 levels in helminth-IS activated BMM did not reflect the strong mRNA upregulation previously observed ( Fig . 3B ) [10] . Thus , surface IgG receptors likely play an important role in triggering CXCL2 release by MΦ during helminth infection . Crosstalk between MΦ and fibroblasts or MF is integral to the wound healing process [30] . Together , our in vitro and in vivo data suggested , that MΦ , which populate the wound early after infection [31] , might be activated by antibody-trapped helminth larvae to enhance MF recruitment to intestinal lesions by secreting CXCL2/3 . Hence , we tested a direct effect of MΦ-secreted CXCL2/3 by investigating scratch wound closure by MF in the presence of conditioned medium from helminth-IS activated MΦ . Addition of conditioned media from stimulated BMM significantly improved MF migration into the scratch wound area ( Fig . 3C ) . Both CXCL2 and CXCL3 bind with high affinity to the chemokine receptor CXCR2 [32] . Thus , we studied the involvement of CXCR2 signaling in the positive effect of BMM-secreted factors on scratch wound closure by MF using the CXCR2 receptor antagonist SB265610 . As shown in Fig . 3C , addition of the CXCR2 antagonist reduced the improvement of in vitro wound closure in response to conditioned media from helminth-IS activated MΦ ( see also S1 Movie ) . These findings support a paracrine role for MΦ-secreted CXCR2 ligands in MF-mediated wound containment during helminth infection . However , once exposed to larval components , MF may also produce autocrine CXCL2/3 . Previous studies have shown that pro-inflammatory stimuli can elicit CXCL2 expression by stromal cells [33] . To clarify if MF represent a potential additional source of CXCL2/3 during helminth infection , we isolated small intestinal MF and performed co-cultures with helminth larvae in the absence or presence of IS . As shown in Fig . 3D , larvae potently induced Cxcl2 gene expression in MF , which in contrast to MΦ , was not augmented by the addition of IS . Moreover , MF secreted considerable levels of CXCL2 protein when co-cultured with helminth larvae ( Fig . 3E ) , indicating that helminth larvae alone can activate MF to produce CXCL2 independently of antibodies . This was consistent with the finding that only low levels of surface IgG bound to MF cultured in the presence immune serum ( Fig . 3F ) . We also investigated a potential role of Fcrg-chain signaling in promoting CXCL2 production by MF . Whilst larvae alone , or larvae in combination with immune serum , lead to similar increases in CXCL2 production by WT MF , neither condition could significantly upregulate CXCL2 production by Fcrg-/- MF ( Fig . 3E ) . This suggested that another γ-chain dependent ( antibody independent ) event was required for the upregulation of MF CXCL2 production by helminth larvae . In addition to surface Fcr-mediated recognition of antibody-coated helminths , helminthic glycan structures can be recognized via pathogen recognition receptors such as dectin1 or 2 , the mannose receptor ( MR ) or DC-SIGN [34 , 35] . Amongst these receptors , dectin2 is unique in its requirement of the Fcrg-chain for signaling [36] . Thus , we studied a potential involvement of dectin2 in the helminth-driven induction of CXCL2 production by intestinal MF . As shown in Fig . 3G , H dectin2-/- MF showed a significant reduction in helminth-induced Cxcl2 mRNA expression and protein release , suggesting that dectin2 recognition of larval components contributes to the upregulation of chemokine production by MF . Interestingly a more dramatic defect was observed for Fcrg-/- as compared to dectin2-/- MF , which may indicate that additional Fcrg-chain coupling C type lectins are involved [37] . Finally , the pattern of expression of CXCL3 in wildtype and Fcrg-/- and dectin2-/- MF closely paralleled the one of CXCL2 ( Fig . 3I-K ) . Our earlier data suggested increased expression of CXCR2 ligands in helminth-induced lesions in vivo and a role for CXCR2 in MF migration in response to MΦ-secreted chemokines in vitro . To assess a potential role for CXCR2 in intestinal repair during helminth infection in vivo , we treated mice by the orally active CXCR2 antagonist SB265610 [22] . As CXCR2-driven neutrophil recruitment has been reported to contribute to protection against the helminth Strongyloides stercoralis [38] , we first enumerated intestinal lesions and luminal parasites at day 14 post Hpb challenge infection . No significant effect of CXCR2 antagonist treatment was noted for either lesion or parasite numbers ( Fig . 4A , B ) . We next compared the size of intestinal lesions between vehicle and SB265610 treated mice at day 14 post challenge infection . As shown in Fig . 4C , D , inhibition of CXCR2 signaling resulted in the development of larger and more irregular intestinal lesions . When analyzing intestinal aSMA expression , we observed a striking defect in the accumulation of aSMA+ cells within intestinal lesions of SB265610 treated mice ( Fig . 4E , F ) , suggesting that CXCR2-ligands might also contribute to MF accumulation and wound containment in vivo . This phenotype observed in SB265610 treated mice closely paralleled the phenotype of Aid-/- and Fcrg-/- mice suggesting that defective CXCL2/3 expression was responsible for impaired wound healing in the absence of immune antibodies or Fcrg chain signaling . In addition to their contractile function , MF promote wound healing by depositing extracellular matrix components , including collagen . Thus , we investigated whether the reduced accumulation of MF associated with CXCR2 blockade also impacted on collagen deposition within intestinal lesions . Mice treated with the CXCR2 antagonist showed normal collagen levels following helminth infection ( S4A , B Fig ) suggesting that this pathway is not involved in collagen deposition . However antibody ( receptor ) deficient mice displayed increased lesional collagen levels ( S4C , D Fig ) . This may suggest that collagen production on a per cell basis is increased in Aid-/- and Fcrg-/- mice , which could be explained by the potential of trapped larvae to suppress collagen expression , whilst inducing collagen-remodeling matrix metalloproteinase and anti-fibrotic cyclooxygenase-2 ( Cox2 ) enzymes in intestinal MF ( S4 E-I Fig ) . Several studies have implicated CXCR2 in granulocyte recruitment during helminth infection [38 , 39] . We therefore determined the number of neutrophils and eosinophils in lesions of SB265610 treated mice and in Aid-/- or Fcrg-/- mice . We observed considerable levels of neutrophil myeloperoxidase ( MPO ) , a widely used neutrophil marker , in the center of WT lesions at day 14 p . i . ( Fig . 4G ) . However , when comparing MPO expression between WT , SB265610-treated or antibody ( receptor ) deficient mice , we observed an increase rather than a decrease in neutrophils ( Fig . 4G upper panel , H ) , indicating that CXCR2 ligands are not required for neutrophil recruitment to intestinal lesions during challenge infection . As neutrophils can contribute to tissue damage during helminth infection [9 , 40] , we attempted to deplete neutrophils in the small intestine of Fcrg-/- mice to test a potential contribution of the increased neutrophil infiltrate to the impaired repair observed in these . Whilst neutrophils were successfully depleted in the peripheral blood of challenge infected Fcrg-/- mice ( S5A Fig ) , high numbers of MPO expressing cells were still observed in intestinal lesions of mice that had been treated with a neutrophil-depleting anti-Ly6G antibody ( S5B Fig ) . However , we could not observe a clear correlation between the quantity of MPO staining within intestinal lesions and lesion size ( S5C Fig ) , suggesting that the increased neutrophil accumulation in antibody deficient mice does not play a major role in their reduced capacity to repair helminth-induced lesions . Using 12/15-lipoxygenase ( 12/15LO ) as a marker for eosinophils ( see S1 Text , S5D Fig ) , we observed high numbers of eosinophils , many of which appeared de-granulated , within lesions of all mice ( Fig . 4G lower panel , I ) . Similarly to neutrophils , mice treated with a CXCR2 antagonist showed increased eosinophil numbers . Taken together , these findings implicate CXCR2 as an important immune-modulating pathway , which functions to promote wound contraction following helminth infection but does not contribute to granulocyte recruitment . To determine whether helminth- and antibody-induced CXCL2/3 production plays a role in wound contraction in response to other helminth species we performed co-cultures of Ascaris suum ( A . suum ) larvae and IS from challenge-A . suum- infected pigs with human monocyte-derived MΦ ( MDM ) . The pig parasite A . suum and the human parasite A . lumbricoides are closely related and cross-infections are common in endemic areas [41] . When incubated with IS , MDM adhered to A . suum larvae ( Fig . 5A upper panel , B ) , which was associated with partial immobilization of infective larvae ( Fig . 5A lower panel ) . This finding was in keeping with old literature , demonstrating the interaction of MΦ and A . suum after stimulation with IS [42] . Upon stimulation with phorbol esters and calcium ionophores or LPS , human and pig monocytes or MΦ can produce CXCL3 , which is homologue to mouse CXCL2 and which binds to CXCR2 with high affinity [43 , 44] . Thus , we quantified CXCL3 production by human MDM after co-culture with A . suum larvae in the absence or presence of IS by ELISA . As shown in Fig . 5C MDM which had been co-cultured with A . suum larvae and IS secreted significantly larger amounts of CXCL3 as compared to culture with larvae alone . Next , we performed an in vitro wound-healing assay using primary human fibroblasts , which had been differentiated into MF by culture in the presence of TGFβ1 for 48h . Human MF ( hMF ) cultured in the presence of conditioned media from unstimulated MDM exhibited a reduction in the wound area by around 50% after 48h following scratching ( Fig . 5D ) . In contrast , when conditioned media from A . suum-IS-activated MDM were added , hMF closed the scratch wound significantly faster , reaching complete wound closure after 48h ( Fig . 5D , S2 Movie ) . We also tested if A . suum secreted products directly impacted on scratch wound closure by hMF . However , addition of A . suum products to hMF monolayers after scratch wounding had no significant effect on wound closure ( S5E Fig ) , suggesting that mainly MΦ-secreted factors were responsible for accelerated scratch closure in the presence of conditioned media . Moreover , addition of the CXCR2 antagonist SB265610 abrogated the positive effect of A . suum-IS-activated MDM supernatants on in vitro wound closure by hMF ( Fig . 5D , S2 Movie ) , suggesting that CXCR2 signaling is largely responsible for this effect . Taken together these findings suggest that helminth-antibody-MΦ interactions can promote CXCR2 dependent wound closure during helminth infection across distinct host and parasite species .
The capacity of helminths to induce rapid and efficient tissue repair has long been recognized , but effector mechanisms responsible for wound healing have only recently begun to be determined . Here we demonstrate that antibodies , which contribute to protective immunity against several helminth species [19 , 45 , 46] , additionally contribute to wound contraction . Helminths damage host tissues not only by breaching physical barriers but also by driving strong inflammatory responses including the recruitment of tissue-destructive granulocytes [9 , 47] . Inflammatory cell recruitment is particularly large and rapid during memory type 2 responses , when tissue migrating larvae are also trapped and killed [31 , 48] . Our study reveals a novel mechanism by which the protective antibody response also participates in wound healing . Thus , when mounting a protective memory response , the mammalian host uses parallel mechanisms to ensure the limitation of infection and the repair of tissue damage at the same time . Of note , delayed type 2 inflammation during primary helminth infection was associated with higher soluble CD14 levels in the peritoneal cavity , thus implicating the rapid and strong inflammatory response during challenge infection in the control of bacterial translocation ( in the face of helminth triggered tissue damage ) . However , the striking morbidity of antibody and FcRg deficient mice at late time points after challenge infection suggested that this strong inflammatory response can ( potentially ) be harmful if not properly controlled or resolved . Thus , helminth products and antibody-FcRg signaling possibly contribute to wound healing on several levels by promoting lesion containment and contraction whilst also limiting excessive peritoneal inflammation . Although we only detected low levels of CXCL2/3 following primary helminth infection , our study does not rule out a role for antibodies or CXCL2/3 in promoting tissue repair in the absence of immunological memory . The identification of dectin-2 as an innate signaling event mediating the induction of CXCL2/3 production by helminth products may suggest that these chemokines can indeed be produced even in the absence of antibodies . The finding that FcRg deficient mice showed a stronger reduction in chemokine levels during helminth infection further supports a role for dectin-2/ FcRg chain triggered CXCL2/3 . However the similar wound healing phenotype in antibody and FcRg deficient mice suggests that timely , CXCR2-driven lesion contraction during challenge infection may require the concerted activation of MΦ and MF by both antibodies and helminth products . Previous studies have identified several innate immune mechanisms that contribute to the repair of helminth-induced injury following primary infection [9 , 17 , 49 , 50] . Here , we observed that lesions in primary infected mice were of similar size as compared to those in secondary challenge-infected mice at day 14 and 21 post infection , which might be explained by the delayed inflammatory response during primary infection . However , primary lesions failed to contract to the same extent as secondary lesions by day 42 post infection , thus supporting a role for adaptive TH2 immune memory in intestinal wound repair . Moreover , the transfer of immune serum from challenge Hpb infected , but not from naïve , WT mice into challenge-infected Aid-/- recipients could rescue the wound healing defect in these mice . This finding supports the hypothesis that immune antibodies play a major role in the containment and contraction of intestinal lesions following helminth infection . Whether the wound healing-promoting effect of immune serum mainly depends on specificity towards the worm or on different Fc functionality ( e . g . through isotype switching or glycosylation ) is an open question , which should be addressed in future studies . Taken together , our work suggests that immune memory activates particularly potent and timely repair responses . It also represents the novel identification of an adaptive immune mechanism , antibody production , that can promote wound healing following helminth infection [8] . Two studies , including our own work , showed that antibody-activated Arg1 expressing MΦ limit the motility of tissue-dwelling helminth larvae , thereby preventing tissue damage during early challenge infection with Hpb or N . brasiliensis , respectively [10 , 48 , 51] . Our current study now demonstrates that antibody-trapped helminths modulate gene expression in MΦ and intestinal MF to improve wound contraction also at later stages of helminth infection . Reduced peripheral MF in Aid-/- and Fcrg-/- mice were associated with increased lesion size and enhanced granulocyte infiltration at late timepoints post-infection , which could be reproduced by CXCR2 blockade . In contrast to the importance of CXCR2 for neutrophil recruitment to other tissues [21 , 39 , 52] , CXCL2/3 thus appear to play redundant roles in neutrophil recruitment during enteric nematode infection . Instead , other chemoattractants such as complement component C5a , leukotriene B4 or platelet activating factor ( PAF ) may recruit neutrophils to intestinal lesions [53] . The increased granulocyte infiltrate in CXCR2-inhibited or Aid-/- and Fcrg-/- mice might further suggest that MF , which are recruited to helminth-induced intestinal lesions , can exert an anti-inflammatory function by limiting granulocyte accumulation . Interestingly , intestinal MF cultured with helminth larvae upregulated the expression of cyclooxygenase-2 , which serves anti-fibrotic and anti-inflammatory functions in the gut , including limiting neutrophilic inflammation [54 , 55] . Even if we did not observe a correlation between the extent of neutrophil infiltration and lesion size , our study does not rule out the possibility that the increased granulocyte accumulation in the lesions of antibody and Fcrg deficient or CXCR2-inhibited mice contributes to the impaired repair response in these mice . Future studies could thus address how helminths may modulate MF and granulocyte function to prevent excessive inflammation and instruct homeostatic wound healing . Our study describes a previously unrecognized function of helminth-induced CXCR2-ligands in intestinal wound containment , thus implicating the CXCL2/3-CXCR2 chemokine axis in repair of mucosal tissues . Whilst little was known about a potential involvement of CXCL2/3 in intestinal repair , the role of CXCR2 signaling in cutaneous wound healing after chemical damage or incision wounding is well-described [21 , 22] . In such settings of sterile damage , CXCR2 contributed to several of the central processes of skin repair , including keratinocyte migration , re-epithelialization and neovascularization . Moreover , CXCR2 was involved in the IL-33-induced repair of Staphylococcus aureus-infected wounds [56] . Our findings build on this work by additionally demonstrating that CXCR2 ligands act on MF to promote wound healing in the small intestine . Thus , it is likely that CXCR2 signaling is a conserved repair mechanism involved in the restoration of barrier integrity at different body surfaces in the absence or presence of infectious agents . Of note , CXCR2 can bind other ligands ( CXCL1 , CXCL5 , CXCL7 and CXCL8/ IL-8 ) in addition to CXCL2 and CXCL3 . Although these chemokines were not induced in macrophages in response to helminth larvae and immune serum [10] , our data does not rule out a contribution of CXCR2 ligands other than CXCL2/3 to lesion contraction during helminth infection in vivo . It should also be noted that we cannot entirely exclude an impact of antibody-FcRγ signaling on the local secretion of type 2 cytokines , e . g . by basophils [57] , which might additionally contribute to intestinal wound repair during secondary helminth infection . Our findings have important clinical implications as CXCR2 has been suggested as a drug target for inflammatory diseases [58] . Together with other studies demonstrating crucial roles for the CXCL2/3-CXCR2 axis in repair and homeostasis [21 , 22 , 59] , our findings argue for a careful evaluation of the therapeutic inhibition of this pathway . Future studies could address the potential of CXCR2 agonists and/or antibody-helminth-preparations to improve repair in settings of impaired wound healing responses ( e . g . diabetic ulcers ) or after surgery .
All mouse experiments were approved by the office Affaires vétérinaires ( 1066 Epalinges , Canton Vaud , Switzerland ) with the authorization number 2238 according to the guidelines set by the service de la consummation et des affaires vétérinaires federal ( Canton Vaud , Switzerland ) . Mice were euthanized using carbon dioxide . All procedures involving pigs were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health after review and approval by the Beltsville Area Animal Care and Use Committee under protocol number 10–012 . Mice ( strains see S1 Text ) were bred and maintained under specific pathogen free conditions at the École Polytechnique Fédérale ( EPFL ) de Lausanne , Switzerland . Mice were infected with 200 H . polygyrus bakeri larvae and two courses of antihelminthic cobantril were administered at days 28 and 30 p . i . Mice were re-infected with 200 larvae at day 44 and two courses of cobantril were administered at days 52 and 54 p . i . in some experiments . The CXCR2 antagonist SB265610 [22 , 60] ( Tocris Bioscience , Bristol , UK ) was administered ( 3 mg/ kg ) once daily from day 0 p . i . by oral gavage . Mice were sacrificed at day 4–42 post challenge infection . Ascaris suum larvae were obtained from infective eggs according to published procedures [61] . See also S1 Text . A . suum-challenge infected pigs ( see also S1 Text ) were bled at day 14 p . i . for serum collection . Murine bone marrow derived MΦ ( BMM ) or human MDM were prepared as described previously [10 , 62] . Mouse small intestinal MF were obtained by culturing lamina propria cells isolated from the small intestine of naïve C57BL/6 mice according to previously published methods [63 , 64] . Human primary gingival fibroblasts were differentiated into MF by culture with TGFβ1 for 48h . See also S1 Text . Serial paraffin sections of small intestines ( “Swiss rolls” ) were stained with hematoxylin and eosin or Sirius red . Lesions were identified by light microscopy and serial sections were used for immune staining and imaging ( see also S1 Text ) . Cells isolated from lesions were stained with fluorescently labeled monoclonal antibodies ( see S1 Text ) and acquired on a BD LSRII flow cytometer ( BD , Franklin Lakes , NJ ) . RNA extraction and qPCR analysis was performed as described elsewhere [10] ( for qPCR primer sequences see Table S1 in S1 Text ) . Concentrations of mouse CXCL2 , IL-4 or IL-13 in cell or intestinal ( upper duodenum ) culture supernatants were quantified by using an anti-mouse CXCL2 ELISA kit ( Sigma Aldrich , Buchs , Switzerland ) or eBbioscience Ready-SET-Go ! IL-4 or IL-13 kits ( eBioscience , San Diego , CA ) . Human CXCL3 was quantified using an anti-human CXCL3 ELISA Kit ( LuBioScience , Luzern , Switzerland ) . A scratch was introduced into MF monolayers using a 200-μl pipette tip according to published methods [65] and images were recorded using an Olympus Cell’R system with a UPLAN FL 10x phase objective . The scratch area was quantified using ImageJ and normalized to the area in the same position at T = 0 . Mann-Whitney test or one-way ANOVA followed by Holm-Sidak’s multiple comparisons test was used to compare the means of two groups or multiple groups , respectively . | To complete their lifecycles , helminth parasites have to migrate through tissues such as the skin , lung , liver and intestine . This migration causes severe tissue damage , resulting in the need for rapid repair to restore the integrity and function of damaged tissues . Protective type 2 immune responses against helminths can repair acute lung damage , but they can also promote liver fibrosis . However , how protective immune mechanisms might contribute to wound healing during enteric nematode infection has remained unclear . Here we show that during a protective antibody response , where helminth larvae are trapped in the intestinal mucosa , macrophages and myofibroblasts secrete chemokines , which promote the repair of helminth-caused lesions . Chemokine secretion by macrophages was triggered by antibodies and helminth products , whilst myofibroblasts produced chemokines directly in response to innate recognition of helminth products . The same chemokines that instructed intestinal repair in mice were also secreted by human macrophages , when co-cultured with immune serum and helminths . Finally , human myofibroblasts closed in vitro scratch wounds more rapidly , when stimulated with the chemokine secretions of helminth-antibody activated human macrophages . Thus , our findings reveal a novel mechanism , by which a protective antibody response can promote the repair of intestinal injury during helminth infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Immune Antibodies and Helminth Products Drive CXCR2-Dependent Macrophage-Myofibroblast Crosstalk to Promote Intestinal Repair |
HIV virions assemble on the plasma membrane and bud out of infected cells using interactions with endosomal sorting complexes required for transport ( ESCRTs ) . HIV protease activation is essential for maturation and infectivity of progeny virions , however , the precise timing of protease activation and its relationship to budding has not been well defined . We show that compromised interactions with ESCRTs result in delayed budding of virions from host cells . Specifically , we show that Gag mutants with compromised interactions with ALIX and Tsg101 , two early ESCRT factors , have an average budding delay of ~75 minutes and ~10 hours , respectively . Virions with inactive proteases incorporated the full Gag-Pol and had ~60 minutes delay in budding . We demonstrate that during budding delay , activated proteases release critical HIV enzymes back to host cytosol leading to production of non-infectious progeny virions . To explain the molecular mechanism of the observed budding delay , we modulated the Pol size artificially and show that virion release delays are size-dependent and also show size-dependency in requirements for Tsg101 and ALIX . We highlight the sensitivity of HIV to budding “on-time” and suggest that budding delay is a potent mechanism for inhibition of infectious retroviral release .
HIV incorporates an aspartic protease that requires homo-dimerization for activation and is the target of numerous FDA approved inhibitors [1–3] . The monomeric form is encoded within the immature virion as part of the Gag-Pol precursor which includes Matrix ( MA ) , Capsid ( CA ) , Spacer Peptide 1 ( SP1 ) , Nucleocapsid ( NC ) , Transframe ( TF ) , Protease ( PR ) , Reverse Transcriptase ( RT ) , and Integrase ( IN ) domains [4] . There are ~120 Gag-Pol proteins packaged in each immature HIV virion along with ~2 , 000 Gag proteins . Gag and Gag-Pol are synthesized from the same messenger RNA via a ribosomal slippage , therefore Gag has the same N terminal sequence as Gag-Pol with MA , CA , SP1 , NC , plus the Gag-specific Spacer Peptide 2 ( SP2 ) and the unstructured p6 domain that is essential for budding of infectious virions [5–7] . Protease activation is vital for auto-processing of Gag-Pol , which in turn is essential for maturation and infectivity of HIV virions [8 , 9] . The protease activity within Gag-Pol is highly regulated and the release from its boundaries in Gag-Pol , especially the TF domain , substantially increases its activity [10–12] . There are eleven canonical protease sites on Gag and Gag-Pol precursors , and in vitro experiments using recombinant PR and HIV Gag as substrate , have characterized the affinities of PR to these sites ( from high to low affinity: SP1/NC , SP2/p6 , MA/CA , NC/SP2 and CA/SP1 sites ) [4 , 13 , 14] . Once Gag is processed , the newly released CA assembles within the virion cavity to form the HIV mature capsid which encapsidates the RNA bound to Gag NC along with RT and integrase [7] . While the HIV protease has been studied extensively , the mechanism and timing of its initial activation in vivo has remained elusive , and the putative connection between protease activation and the endosomal sorting complexes required for transport ( ESCRTs ) , which support HIV budding [15] , remains unexplored . ESCRTs are implicated in cellular processes which require fission of budding membranes and are shown to play a major role in multivesicular body formation [16] , enveloped virus budding [15] , cytokinesis [17–19] , exosomal vesicle generation [20] , and plasma membrane repair [21] . Likely , the most studied of these processes is the impact of ESCRTs on HIV budding . The unstructured p6 domain of Gag hosts two major ESCRT interaction motifs , PTAP and YP [22–24] . The PTAP motif directly interacts with Tsg101 [25–28] , and its mutation has a severe effect on HIV virion infectivity . The YP motif interacts with ALIX [29–33]; ALIX also interacts with the upstream Gag NC domain however the exact function of this interaction is still not fully clear [34 , 35] . The PTAP and YP motifs are collectively known as HIV late domains; indeed , many enveloped viruses interact with early ESCRTs through specific domains within their matrix protein termed late domains [15] . The late domain terminology stems from the observed phenotype of late budding arrest visualized by electron microscopy of budding viruses with altered late domains [5 , 15 , 27 , 36 , 37] Here we found that late budding arrest of HIV , due to mutations of its late domains , is transient . We have characterized the budding kinetics starting with Gag virus like particles ( VLPs ) . HIV Gag protein is sufficient for assembly of Gag VLPs with the same size as HIV virions [38] . Using Gag VLPs , we show that Gag with mutated PTAP ( ΔPTAP ) or YP ( ΔYP ) motif releases out of plasma membrane with ~1 hour and ~20 minutes delay compared to WT , respectively . To analyze the effect of the same mutations on VLPs incorporating both Gag and Gag-Pol , we generated a fully functional Gag . Pol vector that incorporates both Gag and Gag-Pol and is sufficient for budding mature VLPs with similar efficiency to HIV-1 full-length virus . We show that with an active protease , the Gag . Pol VLP budding is delayed when introducing ΔPTAP and ΔYP mutations . Indeed , Gag . Pol VLPs with ΔPTAP mutation are released with ~10 hours delay and are void of HIV RT and PR . Gag . Pol VLPs with ΔYP mutation are released with ~75 minutes delay , which results in significant reduction of RT and PR incorporation within released VLPs . Budding of Gag . Pol VLPs with an inactive protease and either ΔPTAP or ΔYP mutations is dramatically slowed down with similar sensitivity to the involvement of Tsg101 and ALIX . Using Gag proteins with multiple GFP fusions as cargo , we further show that budding is sensitive to the size of cargo proteins , and this effect is reproduced when a PR inactive truncated Pol protein is used as cargo . Finally , modeling these data using MonteCarlo simulations show that the protease activation after complete assembly of HIV virions on plasma membrane can quantifiably explain the loss of Pol specific proteins to host cell cytosol before VLP release .
In parallel experiments , budding of HIVR9 and HIVR8 . 2 were compared to Gag VLP release 24 hours after transfection . Using Gag domain mutagenesis , we observed that while HIV release and maturation are affected by p6 late domain mutations , VLP production by Gag remains almost unaffected ( Fig 1A ) . Shown are the following mutations in Gag p6 , HIVR9 and HIVR8 . 2: ΔPTAP incorporate a 7LIRL10 instead of 7PTAP10 [6] , ΔYP include 36SR37 instead of 36YP37 [31] , and ΔPTAP . ΔYP has both PTAP and YP sequences altered ( 7LIRL10 plus 36SR37 ) . 24 hours post-transfection , cells and VLPs were collected as described in Materials and Methods and analyzed by immunoblotting using p24 , ALIX and Tsg101 specific antibodies . We found that incorporation of early ESCRTs in released Gag as well as HIVR9 and HIVR8 . 2 VLPs was sensitive to late domain mutagenesis; Tsg101 was fully sensitive to ΔPTAP mutation , and ALIX was only slightly affected by ΔYP mutation ( Fig 1A ) . ALIX migrates as two separate bands with the upper band likely related to a post-translational modified form . We don’t know yet at this stage the nature of this ALIX modification . ALIX background level corresponds to exosome release ( Fig 1A and 1B ) . As commonly reported , HIV virion release ( here shown HIVR9 and HIVR8 . 2 ) was detectably reduced under the ΔPTAP mutation in addition to a clear defect in Gag processing . Also , a slight change in release and maturation profiles was observed in ΔYP mutant HIVR9 and HIVR8 . 2 . In contrast to HIVR9 and HIVR8 . 2 , production of Gag VLPs was only slightly affected by mutations within p6 ( Fig 1 ) . Indeed , expression of Gag with alteration in late domains either as humanized , non-humanized co-expressed with Rev or within R9 with an abrogated ribosomal slippage leads to the same results ( Fig 1B , 1C and 1D ) . Aside from interacting with the p6 domain , ALIX also binds to the NC domain and mutations affecting NC have recently been implicated in HIV virion release [39] . We found that under NCΔC6S ( replacement of each NC cysteine by serine ) plus ΔYP and ΔPTAP mutations , ALIX retention in released Gag VLPs is further abrogated and a reduction in TSG101 retention was observed in Gag NCΔC6S VLPs ( Fig 1B ) , however , none of the amino acid substitutions and/or truncations had a marked effect on Gag VLP release ( see also S1A , S1B and S1C Fig ) . The observations related to Gag versus HIVR8 . 2 VLP release were confirmed by pulse/chase 35S-labeling experiments ( S2A Fig ) . The Tsg101/ALIX engagement-independent release of Gag VLPs was tested on different cell types with no major changes in the VLP release except for the apparent cell type specific NC effect ( S3 Fig ) . Plasma membrane binding requirement was tested by G2A mutation [40] which abrogated VLP budding ( S1A Fig ) . Given that compared to HIVR9 and HIVR8 . 2 , Gag VLP production 24 hours post-transfection shows differential dependence on late domains , we set out to test the requirements of higher ESCRT factors in release of Gag VLPs . To investigate ESCRT recruitment ( Tsg101 , ALIX , CHMP4b , and VPS4A ) to budding Gag VLPs , we used HA-tagged forms under tight control of their expression levels along with Gag NC and/or p6 mutants ( Fig 2A ) . We found that Gag VLPs were released with similar yield even when recruitment of Tsg101 and ALIX were compromised due to p6 and/or NC mutations , however surprisingly these VLPs retained both CHMP4b and VPS4A independently of p6 and NC alterations that are inhibiting the early ESCRTs recruitments . Fluorescently tagged ESCRT-III components have been previously localized within budding wild type Gag VLPs , however not in the presence of p6 mutations [41] . Even if our observation is based on a mild over-expression , it clearly shows that ESCRT-III and VPS4 have the potential to be recruited independently of ESCRT-I/ALIX . We further tested the requirement for VPS4 engagement in production of VLPs with compromised interactions with early ESCRTs . To this end , we expressed a dominant negative VPS4 ( ΔE228Q ) during production of HIV Gag VLPs . As shown in Fig 2B , the expression of VPS4ΔE228Q had a substantial negative effect on all Gag VLP production which confirms a requirement for VPS4 in eventual Gag VLP production . While our data show that Gag VLPs with compromised interactions with Tsg101 and ALIX were released with similar efficiencies 24 hours post-transfection , we further investigated the effect of these interactions on the kinetics of Gag VLP production . Shown in Fig 3 is the VLP production comparing WT , ΔPTAP , ΔYP , Δp6 and control ΔG2A . U2OS cells were used for both immunoblotting and microscopy ( left and right panels , respectively ) . Our analysis shows that the kinetics of VLP release is delayed by ~20 minutes ( ΔYP ) to ~1 hour ( ΔPTAP and Δp6 ) which is consistent with the similar VLP release observed 24 hours post-transfection ( see also Supporting Information section ) . To confirm that the Gag variants detected in VLPs using immunoprobing indeed originate from VLPs produced by cells , we visualized the released VLPs by total internal reflection microscopy ( TIRF ) on individual cells . Using TIRF and Gag p6 variants fused to mCherry , we followed the assembly and release of VLPs in live cells . We confirmed the similar VLP assembly on cellular plasma membrane between all Gag variants , and 12 hours post-transfection , we artificially detached the cells to visualize released VLPs as described in Materials and Methods . VLPs were indeed observed immobilized on the cell-free surface accordingly as shown in Fig 3 ( left panels ) . Having established that the HIV Gag VLPs with abrogated interactions with Tsg101 and ALIX are delayed in their release , we set out to investigate the discrepancy in budding of HIV versus Gag VLPs . Upon transfection in cells , the HIVR9 or HIVR8 . 2 express Gag along with Gag-Pol and all other HIV co-factors aside from ENV for HIVR8 . 2 . We chose to generate a system that only express Gag and Gag-Pol proteins for more accurate comparison with Gag to investigate whether the observed differences between Gag and HIV VLPs can be sufficiently explained by the packaging of Gag-Pol . We constructed the Gag plus Gag-Pol open reading frames in a single encoding cassette using humanized Gag and preserving the HIV ribosomal slippage ( S4A Fig ) ; the Gag plus Gag-Pol VLPs produced are referred to as “Gag . Pol” . We further generated variants of Gag . Pol by mutating p6 as described for Fig 1 ( ΔPTAP , ΔYP , and ΔPTAP . ΔYP ) , with either active ( PRwt ) or inactive protease ( PRΔD25N; [42] ) . As shown in Fig 4A , expression of these plasmids in absence of any HIV accessory protein , promoted the production of VLPs incorporating both Gag and Gag-Pol proteins , and Gag processing was only observed in VLPs with PRwt . PRwt VLPs release and mature similarly to HIV virions ( Fig 4B , WT lanes ) . Gag . Pol with ΔPTAP and ΔYP mutations resulted in formation of VLPs with defects in terms of VLP yield and maturation ( Fig 4 ) . Interestingly , PRΔD25N VLPs showed dramatic release defect in all p6 mutants ( Fig 4A ) . Over-expression of ALIX substantially rescued the maturation defect due to ΔPTAP mutation ( Fig 4A ) , as commonly reported . Immunoprobing for Gag and Pol domains indicates that ΔPTAP VLPs are devoid of any detectable RT , while an average of 70% RT loss is observed in ΔYP VLPs ( Fig 4A and 4B and S5 Fig ) . The RT loss is reversed in ΔPTAP VLPs by over-expression of ALIX as shown in Fig 4A . Interestingly , we observed that while ΔPTAP mutation induces identical RT loss in both Gag . Pol and HIV VLPs , the RT loss induced by ΔYP mutation in Gag . Pol VLPs is not occurring in HIVR9 and HIVR8 . 2 ( Fig 4B ) . These data suggest the potential engagement of an HIV effector ( s ) missing in the minimal Gag . Pol system , that is likely capable of supporting ALIX function in the context of ΔYP mutation . There is a reduction in the amount of incorporated RT within Gag . Pol p6 mutants when compared to incorporated RT in WT Gag . Pol . We hypothesized that delayed VLP release in addition to activation of PR before closure of the VLP neck would result in Pol auto-processing and subsequent diffusion of Pol products back to the host cytosol . Indeed , PR was also lost equivalently to RT in Gag . Pol p6 mutants , and follows the same profile in HIVR9 and HIVR8 . 2 ΔPTAP variants ( Fig 4B ) . Supporting the notion of a race between VLP neck closure and PR activation , we also found that WT Gag . Pol showed a ~25% RT loss when compared to ΔPR Gag . Pol ( S5C Fig ) . Based on the yields of Gag . Pol VLP production ( comparing both PRwt and PRΔD25N to Gag VLPs ) , we suspected a longer delay in release of Gag . Pol VLPs with altered p6 compared to Gag VLPs . VLP release kinetics of Gag . Pol variants were analyzed as shown in Fig 5 . As expected , Gag . Pol VLPs budded out at a slower rate compared to Gag VLPs , likely due to the Pol cargo size . To this end , all delays related to p6 mutations were extended in time . Unlike Gag VLPs , which were released with a constant delay measured with respect to the cytosolic Gag concentration , the delay in Gag . Pol VLPs did not follow the same curve as the cytosolic fraction . These kinetics indicates the occurrence of parallel processes during Gag . Pol VLPs production . Interestingly , in the context of PRwt ( Fig 5A , top panels ) , we observed that the appearance of mature p24 versus p55 precursor and related products ( p48 and p41 ) were not necessarily synchronized . Indeed , ΔYP mutation shows a delay in release of mature VLPs however their production does not continue to the same extent as for WT , instead , it saturates earlier while budding follows with VLPs enriched with Gag precursors . ΔPTAP mutation releases VLPs with mainly Gag precursors , especially Gag p48 , and with a substantial delay . To test the effect of packaging full length Pol we performed kinetics on PRΔD25N ( Fig 5A , bottom panels ) with p6 mutations . VLP production kinetics in Gag . Pol PRΔD25N with p6 mutants were all significantly affected , strongly suggesting the importance of early ESCRT engagement ( both Tsg101 and ALIX ) when large cargo is loaded . Importantly , in any case , no full abrogation of VLP release was observed under any p6 mutation . Our data show that p6 mutations create a delay in production of HIV Gag . Pol VLPs , which in turn results in premature activation of PR and diffusion of Pol components from budding VLPs . Also , the delay in VLP release was longer than the one measured for HIV Gag VLPs . To further dissect the mechanistic basis of the observed delay , we hypothesized that the delay length is associated with cargo size defined as domains added after HIV Gag protein , which are naturally present as Pol within HIV . Our observations in budding kinetics of Gag . Pol VLPs demonstrated that when protease activation is inhibited and VLPs incorporate the full length Gag-Pol protein , the kinetics of VLP release is further delayed and becomes strongly dependent on early ESCRTs . These observations suggest a dependence of VLP release on cargo size . To evaluate the influence of cargo size on VLP production , we artificially fused GFPs in frame and in tandem to Gag C-terminus ( in these experiments , every expressed Gag is in tandem with GFPs ) . We found that indeed VLP release by Gag-GFPx variants is proportionally reduced depending on cargo length ( x = 1 , 2 or 3 GFPs ) . The p6 late domain mutation directly dictates the efficiency based on the severity of p6 alterations ( Fig 6 ) . These observations were confirmed by pulse/chase 35S-labeling experiments ( S2B Fig ) . We further confirmed that intact Gag p6 is required for efficient VLP production with large cargo through rescue of p6 mutant Gag-3x . GFP VLP release by co-transfection of Gag with wt p6 ( Fig 7A ) . There is a predominant impact for PTAP and at lower extent for YP . We further modulated the cargo size using Pol truncations in the context of PRΔD25N for maintaining the integrity of Pol cargo . Experiments were performed both under physiological frame shifted expression of Gag . Pol along with Pol proteins expressed in frame with Gag which resulted in 10 fold increase of Pol incorporation in released VLPs . Under both conditions , we observed the same effect of p6 late domain mutations on VLP release ( Fig 7B ) . In both cases , VLP production is negatively affected depending on the length of cargo and nature of p6 alteration . In the context of truncated Gag . Pol with wild type protease , VLP production profile is more complex as deletions in Pol also influence timing of PR activation , as shown for Pol truncations ( S4C Fig ) . The effect of p6 alteration on VLP release by Gag . Pol full length was similar to our findings above , however , in both cases of “in frame” or “frame shift” expression of Gag-Pol , PR activation appeared to be tightly regulated by Pol C-terminus , likely the IN domain . Indeed , when IN is deleted , PR was activated before VLP budding and this activation accounted for a substantial loss in VLP yield . This premature activation of Gag . Pol PRwt ΔIN ( deletion of IN domain ) seems likely to occur before Gag . Pol clusters on the plasma membrane as a ΔG2A mutation of the same Gag . Pol constructs showed exactly the same profile of PR premature processing . These findings are in line with the absence of VLP release when Gag-Pol full length is expressed in frame due to drastic delay of Gag-Pol VLP production ( S4B Fig ) . Kinetics of Gag . Pol VLP release were analyzed using a Gellipsie stochastic algorithm [43] as detailed in Materials and Methods . This analysis incorporated a ) the VLP release rates , b ) protease activation kinetics , and c ) diffusion of protease byproducts out of the open VLPs on the plasma membrane . The simulated data were fitted into the experimental data extracted from Gag . Pol VLP kinetics as shown in Fig 8A . Simulations allowed separation of the three underlying processes . As shown in Fig 8A , the delay in release of VLPs behaves along a poissonian curve with average delay times for WT , ΔYP and ΔPTAP alterations of 5 min , 75 min and 620 min , respectively . The delay in release of Gag . Pol PRΔD25N is substantially longer . During the simulations , the rates of protease activation and diffusion of protease byproducts were held constant while various p6 alterations were analyzed with varying VLP release rates; these rates are shown in Fig 8B . All together , the simulations support our hypothesis that a delay in release of VLPs , all other events constant , results in substantial loss of Pol associated enzymes from the VLPs .
Three major points emerge from our results: i ) Late domain mutations of HIV Gag result in transient delay of virion release from the plasma membrane . ii ) HIV protease is activated following full assembly of virions on the plasma membrane and delays in virion release result in loss of Pol associated enzymes to the cell cytosol and budding of non-infectious virions . iii ) Size of cargo attached to the C-terminus of Gag modulates the speed and requirements for early ESCRT factors during HIV budding . While small cargo sizes rely mostly on Tsg101 , larger cargo sizes are similarly dependent on both Tsg101 and ALIX for efficient VLP budding . We show that alteration of Gag p6 late domains do not inhibit the release of HIV VLPs but rather result in delayed release . We characterized this effect for both VLPs that package HIV Gag only and for VLPs packaging both Gag and Gag-Pol ( Gag . Pol ) . For the Gag . Pol VLPs , the delay ranges from ~70 minutes for the ΔYP mutants that lose proper interaction with ALIX to more than 10 hours for the ΔPTAP mutants which completely lose Tsg101 recruitment . Since the assembly of VLPs takes approximately 45 minutes , a ~10 fold delay in release of the budding VLP will result in a substantial accumulation of ΔPTAP VLPs at the cell surface when analyzed 12 to 24 hours post-transfection . ΔYP mutation has a much shorter delay of ~70 minutes and therefore would result in lesser fold increase in budding VLPs at the cell surface . Importantly , these accumulation levels of VLPs are consistent with the observed phenotypes of HIV late domain mutagenesis [5 , 27 , 15] . Interestingly , we observed that a pool of budding Gag . Pol and HIV VLPs undergo cellular endocytosis especially when release is slowed down due to p6 alteration; indeed , specifically inhibiting endocytosis substantially rescued the yield of VLP release by late domain mutants , especially for large cargo driven by Gag ( Gag . Pol and HIVR8 . 2 ) ( S6 Fig ) . Our results rationally explain the infectivity assays previously reported on progeny virions lacking engagement of ESCRTs . Specifically , infectivity experiments using HIVR8 . 2 pseudotyped with VSV-G have shown that VLPs produced by HIVR8 . 2 ΔYP have a decreased infectivity of approximately 50% compared to wild type HIVR8 . 2 while HIVR8 . 2 ΔPTAP VLPs are non-infectious [44] . While these results could also indicate an alternate effect on particle release , a mismatch between released VLPs and their infectivity has been previously reported [45] . Analysis of the Gag . Pol VLP release kinetics suggests that activation of the protease is occurring immediately after completion of VLP assembly followed by Pol-associated enzymes diffusion out of VLPs in p6 mutants . The rates of PR activation and Pol product diffusion would result in the loss of all Pol enzymes ~60 minutes post-assembly as the VLPs remain open . Also , our analysis indicates that the VLP release times are distributed along a poissonian curve with an average of 5 minutes for WT , 70 minutes for ΔYP and 10 hours for ΔPTAP . This distribution of budding times correlates with percentage of Pol products lost in released ΔYP VLPs compared to WT VLPs . The ΔPTAP mutation which has a ~10 hours delay does not show Pol product incorporation . HIV Gag protein alone is capable of budding from the plasma membrane . We found that Gag still efficiently buds out under severe p6 mutations but with delay at the cell surface for periods of ~20 minutes to ~1 hour . There is some minimal endocytosis of VLPs assembled under mutated Gag compared to Gag . Pol and HIVR8 . 2 VLPs , as shown in S6 Fig . The observed reduction of VLP release due to endocytosis is in agreement with a balance between fast budding and endocytosis of delayed VLPs . Prior to our observations it was shown that HIV Gag with mutated or even deleted p6 releases VLPs from cells [46–49] . These observations were interpreted as related to an ESCRT-independent release of Gag VLPs . In the context of HIV , the mismatch between the levels of VLP release and infectivity was also investigated as an indication of ESCRT-independent budding process and/or budding through intracellular vesicles and exocytosis [48] . Here , our data indicate that HIV virions defective in ESCRT recruitment mainly bud out from the plasma membrane but with proportional delays according to the severity of p6 late domain alterations . Aside from the mutations within the p6 domain , we have conducted extensive mutations within the NC domain of Gag . We found that in the context of Gag expression , VLP budding is independent of NC engagement with ALIX and/or indirectly Tsg101 . Interestingly , using a slight over-expression of CHMP4 and VPS4 , we observed the incorporation of these proteins within released VLPs even in the context of severe p6 and NC mutations , and expression of VPS4DN markedly reduced the efficiency of VLP release . Engagement of Tsg101 and ALIX during the HIV budding is generally assumed to allow the recruitment of downstream ESCRT-III proteins which polymerize at the neck of the budding VLPs before release [15 , 50–54] . Based on our finding , we hypothesize direct recruitment of ESCRT-III and VPS4 to the neck of budding VLPs defective in early ESCRT engagement . To this end , we believe that , if this hypothesis is correct , the neck diameter formed in budding Gag VLPs is small enough to allow effective Tsg101/ALIX-independent CHMP recruitment and VLP release . In vitro , direct recruitment of CHMPs onto negatively curved membranes has been recently observed [55] . We cannot however rule out the possibility that ESCRT-III and VPS4 would be recruited in a Tsg101/ALIX-independent mechanism , possibly through engagement with AMOT and Nedd4 ubiquitin ligases [55–60] . In case of Gag . Pol and HIV VLP production , the Tsg101/ALIX-independent effective CHMP recruitment is substantially delayed due to the large cargo ( Pol ) . We hypothesize that incorporation of Gag-Pol results in wider neck diameters . This hypothesis can rationally explain the different VLP release delays with altered p6 accordingly . The timing of recruitment of ALIX into HIV and EIAV has been investigated using Gag VLPs [61 , 62] , based on our results we suggest that the recruitment may also be sensitive to cargo and therefore the recruitment should be further investigated in HIV virions incorporating both Gag and Gag-Pol . Finally , it is also possible that Tsg101/ALIX-independent CHMP recruitment to the neck of budding VLPs is naturally occurring , however , when Tsg101 and/or ALIX are involved during the CHMP recruitment , the process is faster and functions at maximum velocity to promote fast VLP release which promotes infectivity . In line with the above , we found that ESCRT engagement during VLP budding grows more critical by addition of cargo to the Gag C-terminus . We have measured the kinetics of release for Gag . Pol VLPs with inactivated protease . The Pol protein has a large protein mass ( 150 kDa ) and , in absence of processing , the full length Pol incorporates within the VLP . We found that under these conditions , Gag . Pol VLP production is similarly sensitive to Tsg101 as well as ALIX interactions as shown with ΔPTAP and ΔYP p6 mutants . These results are surprising since typically Tsg101 is the primary interaction during HIV VLP budding , however they agree with the increased importance of ALIX when the budding neck diameter is large like during cytokinesis [17 , 19 , 63] . Also , while Gag VLPs can still release efficiently even in the absence of functional late domains , addition of artificial cargos ( GFPs in tandem ) at the Gag C-terminus inhibits the VLP release in a cargo length-dependent manner . While these Gag-GFP experiments demonstrate the concept of cargo dependent ESCRT requirements , it does not directly reflect on effect of Pol in HIV-1 budding since the Gag-Pol comprises only 5–10% of Gags in the forming HIV virion . The rescue experiments with co-expression of Gag and GagΔp6-3xGFP are the closest comparison to the role of Pol in HIV budding . These experiments demonstrate efficient release of GagΔp6-3xGFP VLPs only when co-expressed with Gag which has a functional late domain . All these observations together support the mechanistic role of ESCRTs in accelerating the closure of budding VLPs with large necks , and that cargo size is the primary regulatory factor that dictates the early ESCRT requirement level . Compared to the minimal Gag . Pol system , when HIVR8 . 2 PRΔD25N VLP release is tested ( Fig 5 ) , full length Pol ( large cargo ) release is more sensitive to PTAP integrity than to YP . We hypothesize that there is an HIV factor that is absent in the Gag . Pol and is promoting the efficient VLP release in absence of functional PTAP/YP sites . This factor is likely acting to mimic some of the Tsg101/ALIX function in accelerating ESCRT-III recruitment and/or promoting Pol packaging before PR activation . The activation of HIV protease immediately post-assembly on plasma membrane is supported by some experimental evidence suggesting that increased packaging of Gag-Pol results in premature activation of PR [64] . Rapid maturation of HIV VLPs within 1 minute post-release has also been reported [65] , although our results predict at least 30 minutes delay between release and full maturation . Also , processing was shown to be essential for HIV VLP release [60] , however the rate of HIV assembly is not affected by PR inactivation [66] . The observed kinetics of Gag precursor release from budding virions analyzed using computer simulations support activation of PR immediately post-virion assembly . Early biochemical characterization of PR cleavage sites showed that Gag and Gag-Pol SP1/NC and SP2/p6 sites are the first to get cleaved by PR [67] . Therefore , if the VLP neck closes before PR activation ( as for WT p6 VLPs ) , soluble PR-containing fragments are trapped within the VLP and continue processing which results in virion maturation . In the case of delayed neck closure , soluble PR-containing fragments diffuse to host cytosol and the progeny virions produced lose Pol products based on the severity of p6 alteration . In agreement with our model ( Fig 9 ) , ΔPTAP and to lesser extent ΔYP VLPs are enriched mainly of Gag p48 and p41 forms , accordingly , clearly suggesting a loss of PR activity in these released VLPs . The first report identifying the importance of the PTAP sequence within Gag p6 used RT activity within the released HIV virions as a measure of viral fitness [6] . In these pioneering experiments , HIV ΔPTAP virions lost RT activity , however , inactivation of PR restored RT activity within released HIV ΔPTAP virions . Our data explain this observation as shown in Fig 4 and demonstrates that this phenotype is due to delayed release of ΔPTAP PRΔ VLPs with intact Pol domains . All together , our observations suggest that the engagement of early ESCRTs during HIV budding is obligatory for speeding up the closure of budding virions and release of fully formed particles before the HIV protease activation occurs , which is fundamental for safeguarding the infectivity of progeny HIV virions . Other viruses and cellular processes whose cargo are not as time sensitive may forego some interactions with ESCRTs therefore possibly explaining the diverse requirements of ESCRTs in these processes [68 , 69] . Our observations show that ‘budding delay’ is a potent mechanism for inhibition of infectious retroviral release and suggest that this mechanism can be used for developing antiviral treatments that would not block ESCRT-dependent cellular processes but slow them to the point of infectious retroviral release inhibition . We also speculate that such mechanism maybe exploited by host cells to inhibit spread of infection .
HIV-1ΔR8 . 2 ( HIV-1NL4-3 R9ΔApa [70] ) and HIV R9 were used . Its late domain mutants , ΔPTAP and ΔYP were previously described [43] . Humanized Gag was produced as previously described [71] . ALIX ( NM_013374 ) , Tsg101 ( NM_006292 ) , CHMP4b ( NM_176812 ) and VPS4A ( NM_013245 ) were kindly provided by Dr . Wesley Sundquist ( university of Utah ) and were all HA N-terminally tagged . GFP ORF was cloned from peGFP-N1 ( Clontech ) . Point mutations were introduced using the Quick Change site directed mutagenesis kit ( Stratagene ) . All cell lines used were grown in complete DMEM medium under standard conditions , excepted for TIRF experiments where cells were incubated in CO2-independent medium ( LifeTechnologies ) . Anti-ALIX [44] , anti-Tsg101 ( C-2 , Santa Cruz Biotech . ) , anti-HA ( HA . 11 clone 16B12 , Covance ) , anti-p24 ( 183-H12-5C , NIH AIDS Reagent Program ) , anti-p17 ( 17–1 , Santa Cruz Biotech . ) , anti-RT ( MAb21 , NIH AIDS Reagent Program ) , anti-PR ( 1696 , Santa Cruz Biotech . ) , and infrared dye coupled secondary antibodies ( LI-COR ) were used for immunoprobing . Scanning was performed with the Odyssey infrared imaging system ( LI-COR ) in accordance with the manufacturer’s instructions at 700 or 800 nm , accordingly . All cell lines used were transfected using lipofectamine 2000 ( LifeTechnologies ) , except for 293T cells using standard CaPO4 precipitation technique . Both cells and media were collected for analysis . Cells were lysed in RIPA buffer ( 140 mM NaCl , 8 mM Na2HPO4 , 2 mM NaH2PO4 , 1% NP-40 , 0 . 5% sodium deoxycholate , 0 . 05% SDS ) , and after removal of residual cell debris by centrifugation , VLPs were pelleted from cell supernatants by centrifugation for 2 hours through 10% ( w/v ) sucrose cushion at 15 , 000 x g . Final VLP samples were re-suspended in PBS . VLP release yields/ratio were calculated as VLPs-associated Gag forms per cell-associated Gag forms based on either CA or MA probing , after densitometry analysis of the immunoblotting data using the Image Studio Lite software ( LI-COR ) . HIV Gag kinetics were fit using a boltzman equation to calculate the delay times for various mutants as described in Supporting Information . Live images were acquired using iMIC Digital Microscope made by TILL photonics controlled by TILL’s Live Acquisition imaging software ( see also Supporting Information ) . U2OS cells were transfected with Gag-mCherry variants and observed by TIRF imaging . At 12 hours post-transfection , cells were gently detached using TryplE ( LifeTechnologies ) . Detachment was achieved by removing the medium and washing once with PBS; a thin layer of TryplE was added to cover cells to allow cell to detach . Images of cells before detachment and afterwards with released VLPs left on the glass support are shown in Fig 3 ( right panels ) . Simulations were setup following the Gillipsie algorithm [43] . Processing , diffusion of Pol and budding were simulated for a single VLP and repeated 500 times to generate a population . The expected p24 and p55 proteins were calculated based on the simulated VLP release . Three essential reactions were considered within each VLP as follows: δ[Gag . Pol]δt=−kp[Gag . Pol]*[Gag . Pol] δ[Pol]δt=+kp[Gag . Pol]*[Gag . Pol]−kd[Pol] δ[VLP]δt=−kr[VLP]−kr*[VLP*] The concentration shown in brackets is the number of molecules within one VLP . At time t = 0 therefore [Gag . Pol] ( t=0 ) =120 ( moleculesVLP ) and[Pol] ( t=0 ) =0 . In these equations kp is the processing rate , kd is the diffusion rate of Pol from the formed VLP with open neck , kr is the rate of VLP release before processing , and kr* is the rate of release after processing . The concentrations of p24 and p55 were calculated based on the following equations: if ( [Gag . Pol]+[Pol]<2 ) then[p24]=0and[p55]=[Gag]+[Gag . pol] if ( [Gag . Pol]+[Pol]>2 ) then[p24]=[Gag]+[Gag . Pol]and[p55]=0 Simulated curves of p24 and p55 ( for this analysis , we did not distinguish between p41 , p48 and p55 , summing all products and representing them as p55 ) are used in Fig 8A to fit the experimental p24 and p55 concentrations measured in Gag . Pol kinetics experiments . In these simulations , kp and kd rates are kept constant while each specific p6 mutation is simulated with a corresponding kr . The simulated internal Pol ( Red ) and Gag . Pol ( Blue ) concentrations in three VLPs are shown in Fig 8B . | ESCRTs are implicated in cellular processes which require fission of budding membranes . Likely the most studied of these processes is the HIV-ESCRT interactions . The canonical view is that interference with ESCRT recruitment results in a late budding arrest of virions at the plasma membrane and this mechanistic view of ESCRTs has shaped our understanding of their function in almost all cell biology . In this manuscript , we present a full kinetic analysis of HIV virion release under all known mutations in Gag that affect HIV-ESCRT interactions . Our data show that contrary to the canonical view , a defect in ESCRT recruitment does not inhibit virion budding , however it creates a delay . We further show that during budding delay , activated proteases release critical HIV enzymes back to host cytosol , leading to budding of non-infectious progeny virions . We suggest that budding delay is a potent mechanism for inhibition of infectious retroviral release and can be the basis for developing antiviral treatments which slow the budding process and therefore disproportionally affect infectious retroviral release . We also suggest that such budding delay may be one of the mechanisms underlying cellular innate immune responses which inhibit the spread of retroviral infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"transfection",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"293t",
"cells",
"biological",
"cultures",
"microbiology",
"enzymology",
"viral",
"structure",
"retroviruses",
"viruses",
"immunodeficiency",
"vi... | 2016 | The Race against Protease Activation Defines the Role of ESCRTs in HIV Budding |
Actin is an abundant protein that constitutes a main component of the eukaryotic cytoskeleton . Its polymerization and depolymerization are regulated by a variety of actin-binding proteins . Their functions range from nucleation of actin polymerization to sequestering G-actin in 1∶1 complexes . The kinetics of forming these complexes , with rate constants varying at least three orders of magnitude , is critical to the distinct regulatory functions . Previously we have developed a transient-complex theory for computing protein association mechanisms and association rate constants . The transient complex refers to an intermediate in which the two associating proteins have near-native separation and relative orientation but have yet to form short-range specific interactions of the native complex . The association rate constant is predicted as ka = ka0 , where ka0 is the basal rate constant for reaching the transient complex by free diffusion , and the Boltzmann factor captures the bias of long-range electrostatic interactions . Here we applied the transient-complex theory to study the association kinetics of seven actin-binding proteins with G-actin . These proteins exhibit three classes of association mechanisms , due to their different molecular shapes and flexibility . The 1000-fold ka variations among them can mostly be attributed to disparate electrostatic contributions . The basal rate constants also showed variations , resulting from the different shapes and sizes of the interfaces formed by the seven actin-binding proteins with G-actin . This study demonstrates the various ways that actin-binding proteins use physical properties to tune their association mechanisms and rate constants to suit distinct regulatory functions .
Actin is an abundant protein that constitutes a main component of the eukaryotic cytoskeleton . Actin polymerization and depolymerization drive essential cellular processes such as cell motility . Nucleation , growth , and disassembly of actin filaments allow cells to rapidly respond to external stimuli . It is known that addition of actin monomers at the barbed end of actin filaments is diffusion-limited [1] and assisted by electrostatic interactions [2] . Actin dynamics is regulated by a variety of actin-binding proteins ( ABPs ) . The functions of ABPs include the nucleation of actin polymerization , promotion of nucleotide exchange in G-actin , sequestration of G-actin , and severance and capping of actin filaments . Many of these functions involve formation of 1∶1 complexes with G-actin . The kinetics of forming these complexes undoubtedly is critical to the regulatory functions of the ABPs . Many of the bimolecular rate constants have been determined experimentally [3]–[9] , and the values cover at least three orders of magnitude . Recently we have developed a method for computing protein association mechanisms and rate constants , and applications to a set of 49 complexes , including two ABP:G-actin complexes , showed that the calculated rate constants are highly accurate [10] . Here we carried out a systematic computational study on the actin-association kinetics of seven ABPs in order to gain better understanding on how their regulatory functions are linked to their structures . The seven ABPs studied here span a range of regulatory functions ( Figure 1 ) . The Wiskott-Aldrich syndrome protein ( WASP ) stimulates the actin nucleation activity of the Arp2/3 complex . This function of WASP resides in the C-terminal region ( hereafter referred to as WCA ) . It is believed that addition of a G-actin molecule , recruited by the WASP WCA , to the Arp 2 and Arp3 subunits of the Arp2/3 complex , creates the nucleus for a new actin filament [11] , [12] . Filament growth occurs only when G-actin is present above a “critical” concentration . Typically , the barbed end grows with the addition of ATP-G-actin; ATP is then hydrolyzed on the filament , and the pointed end shrinks with the departure of ADP-G-actin . In the cytoplasm , G-actin is sequestered by ABPs such as profilin , ciboulot , thymosin β4 ( Tβ4 ) , and twinfilin; the first three favor ATP-G-actin [8] , [13] , [14] whereas the last favors ADP-G-actin [7] . The total G-actin pool is much higher than the critical concentration . G-actin sequestered by Tβ4 and twinfilin is incapable of adding to the barbed end of actin filaments , but ATP-G-actin bound to profilin and ciboulot is as competent as free ATP-G-actin for filament growth at the barbed end . The rapid exchange of G-actin molecules among these monomer-sequestering proteins , along with the promotion of the exchange of ATP for ADP in G-actin by profilin , ensures a proper portion of G-actin ready for filament growth [15] . Actin filaments can be severed and capped at the new barbed end by gelsolin . Actin can be released by tissue injury or cell death to the bloodstream , where polymerization is lethal . This ill fate is prevented by the severing and capping function of a plasma isoform of gelsolin , in conjunction with monomer sequestration by vitamin-D binding protein ( DBP ) . The seven ABPs all bind to the barbed end of G-actin , each with a helix lying in the cleft between subdomains 1 and 3 [8] , [16]–[21] ( Figure 2 ) . The cleft-lying helices in six of these structures lie in approximately the same position , while that of profilin is more to the back of the G-actin molecule . In WASP WCA , ciboulot domain 1 , and Tβ4 , the cleft-lying helices run from the back to the front of the G-actin molecule , but the direction of the helices is reversed in the other four ABPs . Beyond the cleft-lying helices , the seven ABPs span a significant range of structural diversity . The twinfilin C-terminal actin-depolymerizing factor homology domain ( ADF-H 2 ) and gelsolin domain 1 dock to G-actin from the front , whereas profilin docks from the base . WASP WCA , ciboulot domain 1 , and Tβ4 are disordered in the free state [6] , [8] , [22]–[24] and form extended structures upon binding G-actin , hanging over the latter's nucleotide-binding cleft ( which separates the subdomains 1 and 2 from subdomains 3 and 4 ) . Finally the three domains of DBP engulf G-actin tightly from the front , base , and back , respectively . Given their structural diversity , it can be anticipated that the ABPs exhibit a variety of association mechanisms and a range of association rates . Recently we developed the transient-complex theory for calculating protein association mechanisms and association rate constants [10] , [25] . Two proteins reach a transient complex by translational and rotational diffusion , and then form the final , native complex by conformational rearrangement . In the transient complex , the two proteins have near-native separation and relative orientation , but have yet to form most of the stereospecific native interactions . When the conformational rearrangement is fast , the whole association process is rate-limited by the diffusional approach to the transient complex . The rate constant can then be calculated aswhere ka0 is the basal rate constant , i . e . , the rate constant for reaching the transient complex by free diffusion , and the Boltzmann factor captures the bias of inter-protein electrostatic interactions . We have demonstrated that , by adaptively applying the transient-complex based approach , we can study the association of not only relatively rigid proteins but also intrinsically disordered proteins and proteins whose breathing motions are essential for accommodating the incoming partners [10] . Our large-scale application of the transient-complex based approach to demonstrate its prediction accuracy happened to include two ABPs: profilin and gelsolin domain 1 [10] . They piqued our interest in the structure-function relations of ABPs in general , in particular the roles of their actin-association kinetics in linking structure and function . The present study was aimed at elucidating these roles by computing the association mechanisms and quantifying the physical determinants of association rate constants . The results demonstrate the versatility of ABPs in using molecular flexibility and surface charges to tune association mechanisms and rate constants to suit distinct regulatory functions .
The transient-complex based approach for computing the association rate constant consists of three components [25]: generation of the transient complex-ensemble; determination of the basal rate constant ka0 by Brownian dynamics simulations without any biasing force [29]; and calculation of the electrostatic interaction energy in the transient complex by solving the full Poisson-Boltzmann equation . Our algorithm for locating the transient complex is based on the following observation . The native-complex energy well is characterized by a large number of contacts ( Nc ) between interaction loci across the interface but very restricted relative rotation between the proteins . Once outside the native-complex energy well the two proteins lose most of the specific short-range interactions while gaining nearly complete rotational freedom . The transient complex is then located at the midpoint of the sharp transition between these two regimes , where the value of Nc is denoted as [10] . The profilin:G-actin native complex has a large , relatively flat interface , involving nearly the full base of G-actin ( Figures 2B and S1A ) [17]; 22 interaction loci on either side of the interface form 58 contacts . The transient complex is defined with = 16 ( Figure S2A ) . The calculated basal rate constant is 3 . 6×105 M−1 s−1 . The values of Nc in the native complex and in the transient complex are listed in Table 2 for easy comparison among the seven ABPs . Profilin and G-actin show a high degree of charge complementarity across the interface in the native complex ( Figure S1A ) . Five cationic residues of profilin form ion pairs with seven anionic residues of G-actin ( Table 2 ) . Toward the back , the signs of the charges are reversed on both sides of the interface , with E82 of profilin paired with K113 and R372 of G-actin . Corresponding to the high degree of charge complementarity , the electrostatic interaction energy in the transient complex is −2 . 0 kcal/mol at ionic strength = 110 mM . Combining the contributions of the basal rate constant and the electrostatic interactions , the overall association rate constant is calculated to be 1 . 0×107 M−1 s−1 , which agrees well with the measured value of 1 . 4×107 M−1 s−1 at the same ionic strength [5] . At a low ionic strength ( 5 mM ) , the measured rate constant is higher , at 4 . 5×107 M−1 s−1 [13] , [30] . We can explain the increase in ka by a decrease in salt screening of the inter-protein electrostatic attraction . At the lower ionic strength , our calculations give = −3 . 0 kcal/mol and ka = 6 . 0×107 M−1 s−1 . Twinfilin is comprised of two ADF homology domains . Both isolated domains bind to G-actin [7] , but only the structure for the complex formed between ADF-H 2 and G-actin has been determined [21] . Relative to profilin , the binding site for twinfilin ADF-H 2 on G-actin is shifted from the base toward the front ( Figure 2C ) . The interface is shaped like a slightly folded rectangle ( to ∼130° ) , with the fold line , corresponding to the cleft-lying helix of twinfilin ADF-H 2 , just off the diagonal of the rectangle ( cf . Figure S1D–F ) . Interaction loci form 44 contacts across the interface . The transient complex is defined with = 13 , and the calculated basal rate constant is 7 . 9×105 M−1 s−1 . Twinfilin ADF-H 2 also shows a high degree of charge complementarity with G-actin across the interface ( Figure S1B ) . The binding site for twinfilin ADF-H 2 on G-actin is covered mostly by a negative electrostatic surface , delimited by E167 from the base side and E334 on the front side . A corner of this binding site , over subdomain 3 , has a positive electrostatic surface , due to a cluster of cationic residues including R147 , K291 , and K328 . Across the interface , twinfilin ADF-H 2 has a mostly positive electrostatic surface , delimited by K276 and K294 on the base side and R269 on the front side . The cationic corner of G-actin is paired with an anionic corner of twinfilin ADF-H 2 , including E296 , D298 , and E311 . The opposite charges form multiple ion pairs across the interface ( Table 2 ) . Correspondingly , the transient complex has a significant favorable electrostatic interaction energy: = −1 . 6 kcal/mol at an ionic strength of 110 mM . The resulting overall association rate constant is 1 . 2×107 M−1 s−1 , which agrees well with the measured value of 2 . 4×107 M−1 s−1 at the same ionic strength [7] . Gelsolin is comprised of six homologous domains; domain 1 and domain 4 bind G-actin , whereas domain 2 binds to the side of F-actin [31] . Gelsolin domain 1 is structurally similar to twinfilin ADF-H 2 , and their interfaces with G-actin are also very similar , except for a ∼20% reduction in interface area for the former ABP ( Figures 2C , D and S1B , C ) . The number of inter-protein contacts is reduced commensurately from 44 for twinfilin ADF-H 2 to 36 for gelsolin domain 1 . A similar reduction in , from 13 to 9 , is obtained for the transient complex . Correspondingly there is a slightly increase in the basal rate constant , from 7 . 9×105 M−1 s−1 for twinfilin ADF-H 2 to 1 . 1×106 M−1 s−1 for gelsolin domain 1 . While the interfaces of twinfilin ADF-H 2 and gelsolin domain 1 with G-actin are structurally similar , the charge distributions of the two ABPs in their G-actin binding sites are almost the opposite of each other ( Figure S1B , C ) . D85 and D86 of gelsolin domain 1 takes up the locations occupied by K276 and K294 of twinfilin ADF-H 2; the C-terminal carboxylate of the former swaps for R267 of the latter; and R96 of one exchanges for E296 , D298 and E311 of the other . Consequently like charges are matched across the gelsolin domain 1:G-actin interface ( Table 2 ) , and the transient complex has a significant unfavorable electrostatic interaction energy . At an ionic strength of 16 mM , = 1 . 0 kcal/mol , resulting in overall association rate constant of 2 . 0×105 M−1 s−1 . This result matches well with the measured value of 3 . 0×105 M−1 s−1 at the same ionic strength [3] . WASP WCA ( residues 431–502 ) can be further divided into the WASP homology 2 ( WH2 , also known as verprolin homology; residues 431–447 ) segment , linker ( residues 448–461 ) , central segment ( residues 462–483 ) , and acidic segment ( residues 484–502 ) [23] . The WH2 and central segments bind G-actin [6] , [23]; the central segment together with the acidic segment also binds Arp2/3 [6] , [22] , [23] . In the free state , WASP WCA is disordered [6] , [22] , [23] . Chereau et al . [20] determined the structure of a WASP peptide ( residues 430–458 ) encompassing the WH2 segment and most of the linker . The WH2 segment consists of the cleft-lying helix and an extended C-terminal tail , whereas the linker portion is still disordered ( Figure 2E ) . In the context of the full-length WASP in the free state , the central segment forms an amphipathic helix that has its nonpolar face docked to the GTPase binding domain ( GBD ) , resulting in auto-inhibition of WASP [32] . Binding of a Rho-family GTPase , Cdc42 , to the GBD releases the central segment , leading to the activation of WASP ( Figure 1 ) . In the complex with G-actin , the central segment is also likely to form an amphipathic helix that has its nonpolar face docked to G-actin [6] , [23] . The likely binding site for the central-segment amphipathic helix is at the top of G-actin , in the cleft between subdomains 2 and 4 ( Figure 2E ) . This is the site where a C-terminal helix of Tβ4 binds , as found in the structure of a gelsolin domain 1-Tβ4 chimera bound to G-actin [19] ( see below ) . The distance between the C-terminus of the WH2 segment and the N-terminus of the central segment is then 30–35 Å , which is spanned by the 14-residue linker running along the nucleotide-binding cleft separating the subdomains 1 and 2 from subdomains 3 and 4 of G-actin . We can model WASP WCA as a bivalent ligand , with the WH2 and central segments binding to separate sites on G-actin and connected by a linker . The equilibrium constant for simultaneous binding of the two segments can be written as [33]where Ka1 and Ka2 are the association constants for the two isolated segments , and Ceff is the effective concentration . If the linker is modeled as a worm-like chain that does not adversely affect the interactions of the two segments with their respective binding sites , thenwhich is the probability density of the linker end-to-end vector when the latter is the displacement vector d from the C-terminus of the WH2 segment to the N-terminus of the central segment . The measured association constants are 3 . 2×105 and 8 . 2×104 M−1 , respectively , for the isolated WH2 and central segments , and 3 . 1×106 M−1 for WASP WCA [23] . The effective concentration calculated from the experimental association constants is 0 . 1 mM . In comparison , the value of p ( d ) calculated with the 14-residue linker modeled as a worm-like chain with d = 30–35 Å is 0 . 08–1 mM . So the linker model appears quantitatively reasonable . Given that WASP WCA is intrinsically disordered and forms an extended conformation on the surface of G-actin , it is unlikely that WASP WCA forms its contacts with G-actin all at once . It is more likely that the binding starts with the initial docking of one segment and continues with subsequent coalescing of another segment . This dock-and-coalesce mechanism formed the basis of calculating the association rate constants of intrinsically disordered proteins [10] , [34] . The docking segment was identified with the one yielding the highest association rate constant , based on the following reasoning . First , multiple pathways could contribute to the binding , but the one yielding a much higher overall rate constant for forming the final complex than all alternative pathways would dominate . So we can just focus on the dominant pathway . Second , the coalescing step is likely to be fast so that the docking step becomes rate-limiting . So we can further narrow our consideration down to just the docking step , which allows for the treatment of our transient-complex based approach . In Figure 4 we display the rate constants calculated with six fragments of the WH2 segment proposed as the docking segment . The R431-K446 fragment gives the highest rate constant , 1 . 7×107 M−1 s−1 ( at ionic strength = 65 mM ) . This calculated result compares well with the measured rate constant , 4 . 3×107 M−1 s−1 [6] . While the reasoning behind our approach seems well justified and the predicted ka is validated by the experimental data , coarse-grained simulations of WASP WCA:G-actin association could yield direct evidence for the dock-and-coalesce mechanism . We also modeled the structure of the WASP central segment bound to G-actin ( see Methods for details ) . Based on this structure , our transient-complex based approach predicts a rate constant ∼104 M−1 s−1 . This is three orders of magnitude lower than the rate constant calculated with the WH2 fragment as the docking segment , thus justifying our contention that the dominant binding pathway of WASP WCA consists of the docking of the WH2 segment and the subsequent coalescence of the central segment ( Figure 3B ) . We now examine the physical determinants of the docking rate constant to provide a rationalization for its relatively high value , 1 . 7×107 M−1 s−1 . This value comes from a combination of a basal rate constant of 2 . 5×106 M−1 s−1 , the highest among all seven ABPs , and an electrostatic interaction energy = −1 . 1 kcal/mol , the third most favorable . The high basal rate constant can be attributed to a relatively small interface ( Nc = 26 in the native complex ) , formed by the docking of a 10-residue helix plus a six-residue extension to an open cleft ( Figure S1D ) . With = 7 , the transient complex is reached with relatively mild orientational restraints between the WASP WH2 segment and G-actin . The negative electrostatic surface over the cleft of G-actin has been noted above ( also see Figure S1D ) . The WH2 segment complements this with a positive electrostatic surface facing the cleft . In particular , R431 at the start of the cleft-lying helix and K446 at the end of the C-terminal extension form ion pairs with E167 and D25 of G-actin , respectively ( Table 2; Figure S1D ) . These favorable interactions explain the significant negative value of . Ciboulot domain 1 and Tβ4 are homologous , with sequence identities of 25% and 58% , respectively , for the N-terminal half ( ciboulot domain 1 residues 14–33 and Tβ4 residues 1–20 ) and C-terminal half ( ciboulot domain 1 residues 34–52 and Tβ4 residues 21–39 ) . Several lines of evidence suggest that the binding of ciboulot domains 1 ( and Tβ4 ) to G-actin also follows the dock-and-coalesce mechanism , with the N-terminal half as the docking segment and the C-terminal half as the coalescing segment . First , like WASP WCA , these two proteins are intrinsically disordered and adopt extended conformations upon binding G-actin [8] , [24] . Second , in the crystal structure of the complex with G-actin , the N-terminal half of the ciboulot domain 1 is resolved whereas the C-terminal half is still disordered ( Figure 2F ) [8] . Third , X-ray scattering data of the ciboulot domain 1:G-actin complex at low ionic strength could be fitted with the N-terminal and C-terminal halves bound to the barbed end and pointed end of G-actin , respectively , but not the data at physiological ionic strength [35] . The latter data was consistent with a model in which the N-terminal half is bound but the C-terminal half is dissociated . Fourth , for G-actin-bound ciboulot domain 1 , 1H-15N NMR cross peaks of the C-terminal half disappeared or attenuated upon a temperature increase from 25°C to 35°C , indicating either dissociation from or weakened interactions with G-actin [8] . Finally , ciboulot-bound G-actin must have its pointed end free for it to be competent for filament growth at the barbed end . Using the transient-complex based approach , we calculated the rate constants of the docking step with 10 fragments of the N-terminal half as the possible docking segment ( Figure 4 ) . The D10-N32 fragment gives the highest rate constant . The value at low ionic strength ( 10 mM ) , 0 . 9×106 M−1 s−1 , is close to the measured rate constant , 1 . 2×106 M−1 s−1 [8] . We note that ciboulot N32 aligns to WASP K446 , which is the last residue of the putative docking segment of WASP WCA . For the docking of the D10-N32 fragment , the basal rate constant is 0 . 3×106 M−1 s−1 and the electrostatic interaction energy is −0 . 6 kcal/mol . This basal rate constant is 10-fold lower than that for docking the corresponding WASP fragment , due to additional contacts . In ciboulot , the cleft-lying helix is longer by 1 turn of helix at the N-terminus , and the sequence linking this helix and the conserved “LKK” motif ( 30LKN32 in ciboulot and 444LNK446 in WASP ) is longer by two residues ( Figure 4 ) . The transient complex for the ciboulot fragment is defined with = 11 ( Figure S2B ) , an increase of four contacts from the WASP counterpart . Ion pairs with G-actin are maintained at the start and end of the ciboulot fragment ( Table 2 ) , but the electrostatic surface at the end of the fragment is not as strongly positive as that of the WASP fragment ( Figure S1E ) . This accounts for the moderation in relative to WASP . Unlike ciboulot domain 1 , Tβ4 sequesters G-actin and prevents its addition to the barbed end of actin filaments . This difference can largely be explained by a higher affinity of the Tβ4 C-terminal half , relative to the ciboulot domain 1 counterpart , for the pointed end of G-actin [8] , [35] . Nevertheless the complex formation with G-actin by Tβ4 , with both the N-terminal and C-terminal halves bound , is expected to follow the same dock-and-coalesce mechanism as ciboulot domain 1 . Didry et al . [35] designed a chimera by combining the N-terminal half of ciboulot domain 1 and the C-terminal half of Tβ4 . The change in intensities of 1H-15N NMR cross peaks suggested that the C-terminal half of the chimera became dissociated upon raising the ionic strength from low to physiological range . Even at low ionic strength , exchange between unbound and bound states on the 10-ms timescale was observed in NMR experiments for residues in the C-terminal half of the chimera . Following Irobi et al . [19] , we built the structure of Tβ4 bound to G-actin by combining models for the first 16 residues and for residues 17–39 ( Figure 2G ) . The N-terminal portion was a homology model based on the G-actin-bound ciboulot domain 1 [8] , and the C-terminal portion was taken from the structure of the gelsolin domain 1-Tβ4 chimera bound to G-actin [19] . This structural model for the full-length Tβ4 allowed us to exhaustively search for the docking segment in implementing the dock-and-coalesce mechanism . As explained above , the docking segment is selected to yield the highest rate constant for the docking step . Figure 4 displays the rate constants for the docking step calculated with fragments ending at residues 14 to 36 ( at ionic strength = 5 mM ) . The highest rate constant , 5 . 2×106 M−1 s−1 , is for the fragment ending at residue T20 , and a very close second , at 4 . 0×106 M−1 s−1 , is obtained for the fragment with one less residue , ending at K19 . That residue aligns with the last residues of the docking segments obtained above for the G-actin binding of WASP WCA and ciboulot domain 1 . To maintain consistency among the three proteins , we propose the fragment ending at K19 as the docking segment for Tβ4 . The calculated rate constant , 4 . 0×106 M−1 s−1 , for the docking step only slightly overestimates the overall association rate constant measured at the same low ionic strength [9] . The higher rate constants for the fragments ending at K19 and T20 relatively to those for shorter and longer fragments can largely be attributed to differences in . In particular , K18 and K19 make very strong favorable electrostatic interactions with G-actin D24 and D25 , respectively ( Table 2; Figure S1F ) . The transient complex of our proposed docking segment ( residues M0 to K19 ) has = −1 . 0 kcal/mol ( at 5 mM ionic strength ) . For shorter fragments , is positive . For increasingly longer fragments , first has a diminished magnitude and then switches the sign to positive . To further highlight the electrostatic contribution of K18 and K19 , we neutralized them by mutation to alanine . The mutant docking segment has = 0 . 7 kcal/mol . Correspondingly the calculated rate constant reduces by 19-fold to 2 . 1×105 M−1 s−1 . This compares favorably with the measured 10-fold reduction in the G-actin binding rate constant of the Tβ4 K18A/K19A mutant [9] . DBP sequesters G-actin in the plasma . Its three domains tightly clamp around the barbed end of G-actin ( Figure 2H ) . The structure of DBP in the free state , while slightly more open [18] , is still too narrow for G-actin to enter . We used normal mode analysis based on an elastic network model [36] to mimic the breathing motion that DBP likely undergoes during the association process ( Figure 3C ) . The lowest-frequency mode of DBP would widen the opening between domains 1 and 3 ( along with a shear motion between the two domains; Figure S3 ) , yielding a transient “open” conformation that is ready for binding G-actin . We treated the structure in which G-actin is bound to the transient open conformation of DBP as the native complex and applied the transient-complex based approach to calculate the association rate constant . The result , 1 . 6×104 M−1 s−1 , agrees well with the measured value of 2 . 2×104 M−1 s−1 ( at ionic strength = 12 mM ) [4] . The calculated rate constant came from a low basal rate constant of 7 . 4×104 M−1 s−1 and an unfavorable electrostatic interaction = 0 . 9 kcal/mol . The G-actin binding site on DBP is shaped like a scoop , with a deep basin for the base of G-actin . The large interface area ( Nc = 72 in the native complex ) along with the highly curved shape results in a very restricted transient complex ( = 17; Figure S2C ) , accounting for the low basal rate constant . The unfavorable is somewhat unusual , since the DBP:G-actin interface features multiple attractive ion pairs ( Table 2 ) . The overall positive can be explained by the large net negative charges carried by both proteins ( net charges of −13 and −12 , respectively , for DBP and G-actin; see Figure S1G ) . Engulfing of G-actin by the three domains of DBP ensures that the numerous anionic residues on the two proteins are not very distant . Their repulsion trumps the attraction of the ion pairs in the interface . The transient open conformation of DBP used in the ka calculation was from a 200-ps molecular dynamics refinement of a structure along a normal mode ( see Methods for details ) . To get some sense on how conformational dynamics may affect the calculation results , we also carried out ka calculations on the snapshots at 100 and 150 ps of the refinement . Compared to a value of 17 for the 200-ps snapshot , decreases to 15 and 10 for 100- and 150-ps snapshots , respectively . Correspondingly , ka0 increases from 7 . 4×104 M−1 s−1 to 1 . 5×105 and 2 . 6×105 M−1 s−1 , and increases from 0 . 9 kcal/mol to 1 . 3 and 1 . 4 kcal/mol , respectively . As a result of the compensatory changes in ka0 and , the overall rate constant is unchanged for the 100-ps snapshot ( ka still at 1 . 6×104 M−1 s−1 ) and minimally changed for the 150-ps snapshot ( ka = 2 . 4×104 M−1 s−1 ) . So all the snapshots from the molecular dynamics refinement give essentially the same ka .
We have carried out rate calculations for the association of seven ABPs with G-actin , and provided quantitative rationalization for the 1000-fold rate variations among the seven ABPs . The results demonstrate that ABPs can use their physical properties , in particular molecular flexibility and surface charges , in a variety of ways to modulate both the mechanisms of association and the magnitudes of association rate constants . The widely-varying rate constants of the ABPs appear to be tuned for their distinct regulatory functions . WASP WCA has the highest association rate constant , and this is fitting because WASP's recruitment of G-actin to the Arp2/3 complex is critical for the nucleation of new filaments . Deleting the WH2 segment , which according to our calculation is responsible for the high association rate constant with G-actin , leads to a significant slowing down of actin polymerization [23] . Profin and twinfilin , both having rate constants exceeding 107 M−1 s−1 , are responsible , respectively , for sequestering G-actin newly dissociated from the pointed end of filaments and for bringing G-actin to the barbed end for filament growth . Rapid G-actin association by these two ABPs would allow for rapid remodeling of actin cytoskeleton in response to external stimuli . The rate constant of DBP is the lowest . Presumably , tight , not rapid , binding of G-actin is key to ensure the disassembly of potentially lethal actin filaments in the bloodstream . Spontaneous nucleation of ADP-G-actin is extremely slow [37] , so DBP:G-actin association faces only poor potential kinetic competition . The seven ABPs exhibit three classes of association mechanisms ( Figure 3 ) , dictated by molecular shapes and flexibility . Relatively rigid globular proteins like profilin can reach the transient complex with G-actin by diffusion and then rapidly form their specific interactions nearly all at once . In contrast , an intrinsically disordered protein that adopts an extended conformation in the native complex with its target is unlikely to form their specific interactions all at once . It is more likely different segments of the protein contact the target at different times . In principle multiple pathways , each with the different segments contacting the target in a specific sequence , can lead to the native complex . Often one pathway dominates , leading to the dock-and-coalesce mechanism . This apparently is the case for the binding of WASP WCA , ciboulot domain 1 , and Tβ4 to G-actin . Finally , the fork-shaped DBP has an opening that is too narrow , both before and after G-actin binding , for G-actin to enter . Therefore DBP must make excursions to conformations with a wider opening before G-actin can enter . In the usual sense , the last two mechanisms would be referred to as induced folding and conformational selection , respectively . The diverse shapes/flexibility and association mechanisms of the ABPs provide a nice illustration of “form dictates function . ” Regardless of the association mechanism , the association rate constants can span a wide range . Although the ABP ( i . e . , DBP ) that follows the transient-opening mechanism in the present study has the lowest association rate constant , our previous study [10] identified the association of ribonuclease A with ribonuclease inhibitor as following the same mechanism , and yet the association rate constant in this case is as high as 3 . 4×108 M−1 s−1 [38] . So the association mechanism does not dictate the association rate constant . Rather , according to our transient-complex theory [10] , [25] , the association rate constant is determined by the basal rate constant , modeling the approach to the transient complex by free diffusion , and the electrostatic interaction energy in the transient complex . The basal rate constant usually falls in the range of 104 to 106 M−1 s−1 [10] and the variation is determined by the extent of orientational restraints between the proteins in the transient complex [39]–[41] . The extent of orientational restraints can be traced to the shape and size of the binding interface , and seems to be captured well by , the number of contacts in the transient complex . There is good anti-correlation between ln ( ka0 ) and ( Figure S4 ) . In particular , the binding interface of DBP with G-actin has a large area and a highly curved shape . Correspondingly , DBP has the highest and lowest ka0 . That low ka0 contributes to the low overall association rate constant of DBP . The electrostatic interaction energy in the transient complex can modulate the association rate constant by over four orders of magnitude [42] and largely explains the wide variation in rate constant among the seven ABPs studied . Its sign and magnitude are determined by the amount of charges carried by the proteins and degree of their complementarity across the interface [2] , [26]–[28] . The seven proteins studied here all bind to the same site on the same protein , yet they exhibit such diversity in association mechanisms and wide variation in rate constants . Dissecting the physical determinants of this diversity in association kinetics has now provided better insight into how the different structures of the ABPs allow them to achieve their distinct regulatory functions . The structures for two other proteins bound to G-actin are found in the Protein Data Bank ( PDB ) . A RPEL motif from the serum response factor coactivator MAL forms two helices on the G-actin surface , in a location similar to those occupied by the seven ABPs ( PDB entry 2V52 ) [43] . The N-terminal helix lies in the cleft between subdomains 1 and 3 , and the C-terminal helix interacts with subdomain 3 at the base . On the other hand , DNase I binds to G-actin at its top , interacting predominantly with subdomain 2 but also with subdomain 4 ( PDB entry 2A42 ) [20] . We predict association rate constants of 5 . 8×105 and 6 . 6×105 M−1 s−1 , respectively , for these two complexes . These results await experimental verification . Many other ABPs have yet to have structures determined for their complexes with actin . Some of these structures , including those for G-actin-bound complexes of ADF and of twinfilin ADF-homology domain 1 , can be modeled . Application of our transient-complex based approach for characterizing association kinetics to these new targets will further advance our understanding of how ABPs regulate actin dynamics .
The input to our transient-complex based approach for calculation protein association rate constants is the structures of native complexes . For the G-actin-bound complexes of profilin , twinfilin ADF-homology domain 2 , and gelsolin domain 1 , we directly used the structures of PDB entries 2BTF [17] , 3DAW [21] , and 1EQY [16] , respectively . A calcium ion coordinated by D85 carboxyl oxygens and G90 and A92 carbonyl oxygens of gelsolin domain 1 as well as an E167 carboxyl oxygen of G-actin was included as part of the gelsolin molecule . All hydrogen atoms were added and energy minimized by the AMBER program . The structures of the G-actin-bound complexes of the WASP WH2 segment and ciboulot domain 1 were from PDB entries 2A3Z [20] and 1SQK [8] , respectively . The C-terminal portions of the two ABPs were trimmed to various extents to produce putative docking segments for rate calculations . The structure of Tβ4 bound to G-actin was built according to Irobi et al . [19] . The N-terminal 16-residue portion was a homology model using PDB entry 1SQK ( the G-actin-bound ciboulot domain 1 ) [8] , with the sequences aligned according to Figure 4 . The C-terminal 23-residue portion was taken from PDB entry 1T44 , which is the structure of a gelsolin domain 1-Tβ4 chimera bound to G-actin [19] . The two portions were merged after superimposing the G-actin molecules in the two parent structures . The Tβ4 sidechains were then refined by energy minimization . The structure of the WASP central segment ( residues S462–S479 ) bound to G-actin was modeled as follows . The initial conformation of this segment , taken from the auto-inhibited structure ( PDB entry 1EJ5 [32] ) . The amphipathic helix in this segment was aligned to the C-terminal helix of Tβ4 bound to G-actin [19] such that five WASP residues implicated as being buried in the interface with G-actin by NMR experiments [23] were positioned toward G-actin . The three N-terminal residues preceding the amphipathic helix was manually adjusted to roughly follow the corresponding residues in the Tβ4 structure . The G-actin-bound complex of the WASP central segment was then refined by energy minimization and molecular dynamics simulation for 100 ps in explicit solvent . The initial structure for the G-actin-bound complex of vitamin-D binding protein was from PDB entry 1KXP [18] . The open DBP conformation was built on the lowest-frequency normal mode ( obtained by running the ElNemo program [36] ) . The amplitude of the motion along this mode was set with DQ = 300 . G-actin was then brought back and the complex was refined by energy minimized and molecular dynamics simulation for 200 ps in explicit solvent . Our transient-complex based approach for calculating protein-protein association rate constants has been described previously [25] and has been implemented into a web server ( http://pipe . sc . fsu . edu/transcomp/ ) [10] . All the ka calculations reported here were done via the TransComp server , without human interrogation . TransComp consists of three steps . The first is the generation of the transient complex , the late intermediate located at the rim of the native-complex energy well , by generating configurations of two associating proteins around the native complex . In the native-complex energy well the two proteins have a large number of contacts , Nc , between interaction loci across the interface but a small standard deviation , σχ , in the values of the relative rotation angle χ in the sampled configurations ( Figure S2 ) . As the two proteins move outside the native-complex energy well they immediately gain nearly full freedom in relative rotation . Hence there is a sharp increase in σχ as Nc is decreased . We fit the dependence of σχ on Nc to a function used for modeling protein denaturation data as a two-state transition , and identify the midpoint of the transition , where Nc is denoted as , as the transient complex . Once the transient complex is determined , the second step is to calculate the basal rate constant ka0 from Brownian dynamics simulations , using an algorithm developed previously [29] . In these simulations , there is no force or torque acting on the diffusing proteins , except when they encounter steric clash , which is treated as a reflecting boundary condition . The reaction surface ( i . e . , condition for forming the native complex ) is identified as the transient-complex ensemble ( i . e . , Nc = ) . The third step is to calculate the electrostatic interaction energy in the transient-complex ensemble , by solving the full , nonlinear Poisson-Boltzmann equation by the APBS program [44] . The transient-complex based approach treats the associating proteins as rigid . In some cases reaching the native complex by rigid-body docking always encounters steric clashes . Then there would be a large gap in the values of Nc calculated from the sampled configurations . The gap in Nc indicates that at least one of the proteins must be flexible or undergo conformational fluctuation during the association process . In particular , large gaps were found in docking the WASP R431-T447 fragment to G-actin and in the G-actin association of DBP using either the free or bound conformation . As we propose here , only a segment of WASP WCA first docks to G-actin; the C-terminal portion subsequently coalesces to its sub-site on G-actin . And DBP first undergoes breathing motion to produce an open conformation before G-actin enters . | Actin polymerization and depolymerization drive cell motility and are regulated by a variety of actin-binding proteins . The widely-varying rate constants ( ka ) of the actin-binding proteins associating with G-actin , spanning at least three orders of magnitude , appear to be tuned for their distinct regulatory functions . Here we applied our previously developed transient-complex theory to study the association kinetics of seven actin-binding proteins with G-actin . These proteins exhibit three classes of association mechanisms , due to their different molecular shapes and flexibility . The 1000-fold ka variations among them can mostly be attributed to disparate inter-protein electrostatic interactions . By computing the association mechanisms and quantifying the physical determinants of association rate constants , the present study reveals critical links between the structure and function of the actin-binding proteins . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"biomacromolecule-ligand",
"interactions",
"cell",
"motility",
"macromolecular",
"assemblies",
"biophysics",
"theory",
"biology",
"biophysics",
"simulations",
"biophysics"
] | 2012 | Prediction and Dissection of Widely-Varying Association Rate Constants of Actin-Binding Proteins |
Rift Valley fever virus ( RVFV ) , a Phlebovirus with a genome consisting of three single-stranded RNA segments , is spread by infected mosquitoes and causes large viral outbreaks in Africa . RVFV encodes a nucleoprotein ( N ) that encapsidates the viral RNA . The N protein is the major component of the ribonucleoprotein complex and is also required for genomic RNA replication and transcription by the viral polymerase . Here we present the 1 . 6 Å crystal structure of the RVFV N protein in hexameric form . The ring-shaped hexamers form a functional RNA binding site , as assessed by mutagenesis experiments . Electron microscopy ( EM ) demonstrates that N in complex with RNA also forms rings in solution , and a single-particle EM reconstruction of a hexameric N-RNA complex is consistent with the crystallographic N hexamers . The ring-like organization of the hexamers in the crystal is stabilized by circular interactions of the N terminus of RVFV N , which forms an extended arm that binds to a hydrophobic pocket in the core domain of an adjacent subunit . The conformation of the N-terminal arm differs from that seen in a previous crystal structure of RVFV , in which it was bound to the hydrophobic pocket in its own core domain . The switch from an intra- to an inter-molecular interaction mode of the N-terminal arm may be a general principle that underlies multimerization and RNA encapsidation by N proteins from Bunyaviridae . Furthermore , slight structural adjustments of the N-terminal arm would allow RVFV N to form smaller or larger ring-shaped oligomers and potentially even a multimer with a super-helical subunit arrangement . Thus , the interaction mode between subunits seen in the crystal structure would allow the formation of filamentous ribonucleocapsids in vivo . Both the RNA binding cleft and the multimerization site of the N protein are promising targets for the development of antiviral drugs .
The Bunyaviridae family comprises more than 330 viruses that affect vertebrates and plants . La Crosse virus , a member of the Orthobunyavirus genus , causes pediatric viral encephalitis in North America . The Bunyaviridae family also includes several other emerging human pathogens , such as the Hantaan and Sin Nombre viruses ( genus Hantavirus ) and the Crimean-Congo hemorrhagic fever virus ( genus Nairovirus ) . Viruses of the Tospovirus genus infect plants [1] . Bunyaviridae have either arthropods- or rodent-borne vectors and are amplified by vertebrate hosts . The Rift Valley fever virus ( RVFV ) , a Phlebovirus within the Bunyaviridae family , is transmitted by Aedes and Culex mosquitoes and is a medically and agriculturally important cause of epizootics in Africa . Although this virus primarily affects livestock , humans can be infected as well , and infections can lead to several syndromes ranging from a febrile illness to blindness , encephalitis and lethal hemorrhagic fever . The virus is currently found in the sub-Saharan area , as well as in Egypt , Yemen , Saudi-Arabia , Mayotte and Madagascar [2] . The continuing geographical expansion of RVFV draws concern for Europe , where the virus is considered to be an emerging threat [3] , [4] . Current vaccines to prevent RVFV epizootics are only partially attenuated , expensive and only induce short-lived immunity [5] . No specific drugs are available to cure an infection , and preventive efforts to avoid new outbreaks are mostly based on weather monitoring [6] . The genome of RVFV consists of three single-stranded RNA segments of either negative or ambisense polarity designated as L ( 6 , 404 nucleotides [nt] ) , M ( 3 , 885 nt ) , and S ( 1 , 690 nt ) . Within each of these three segments , coding regions are flanked at their 5′ and 3′ termini by non-translated regions that comprise two stretches of complementary nucleotides , leading to the formation of RNA panhandle structures [7] . The L and M segments are of negative polarity , while S has ambisense polarity , encoding the nucleoprotein ( N ) in antisense and the non-structural protein NSs in sense orientation . The L segment expresses a multifunctional protein that comprises an N-terminal endonuclease [8] and a large RNA-dependent RNA polymerase domain [1] . The M segment codes for glycoproteins GN and GC that are inserted in the virus lipid envelope and are responsible for cell tropism and membrane fusion . The endodomain of GN interacts with N , and this interaction is critical for genome packaging into infectious virus particles . As in other negative-stranded viruses , the genomic RNA ( vRNA ) in RVFV is packaged with two virally expressed proteins , N and L , into a ribonucleoprotein ( RNP ) complex that is competent for ( + ) RNA synthesis and transcription . Contrary to RNPs of Mononegavirales , RVFV N does not assemble into a tube-like structure [9]–[14] but rather forms a flexible serpentine-like structure [15] . The precise organization of RNA , N and L in this macrostructure is unknown . In addition to its critical role in protecting the vRNA and the antigenome ( cRNA ) , the N protein also plays an active role in RNA transcription and replication [1] , as well as in virion assembly [16] . Biochemical studies have shown that RVFV N forms dimers through aromatic residues located in the N terminus of the protein [17] . Recently , a crystal structure was reported for the RVFV N protein [15] , which revealed the basic fold of the protein , but raised a number of questions . For example , the crystal structure provided little insight into the mechanism of N multimerization into an RNP complex , and it was unclear how to relate the crystal structure to EM images of N polymers . Furthermore , the RNA binding site identified in the crystal structure of RVFV N differed from that seen in other viral N proteins . Here , we present the crystal structure of RVFV N forming a hexameric ring . The structure reveals the likely binding site for vRNA , and comparison with the previous crystal structure of RVFV N allows us to speculate on the mechanism that underlies the multimerization of N and its encapsidation of viral RNA .
To produce sufficient amounts of protein for structural studies in the absence of other viral proteins , we expressed RVFV N in E . coli with an N-terminal cleavable thioredoxin tag and purified it under non-denaturing conditions to preserve its structural integrity . The final gel filtration column showed two peaks , denoted as N1 and N2 ( Figure 1A ) . SDS-PAGE analysis revealed that both peaks contained a protein of the size expected for N ( 27 kDa ) , suggesting that N was the only protein present and ruling out protein contaminants that could have influenced the oligomeric state of N ( Figure 1A , inset ) . The position of peak N1 corresponds to a protein species with an apparent molecular mass of 300 kDa , suggesting that N formed higher-order oligomers . The position of peak N2 corresponds to a protein species with an apparent molecular mass of 94 kDa and would thus suggest the presence of smaller oligomers . This notion was confirmed by cross-linking experiments that indicated the presence of dimers , trimers and tetramers in fraction N2 ( Figure 1B ) . We also measured the OD260nm/OD280nm ratio for the two peak fractions to test for the presence of bound nucleic acids . Peak N1 had an OD260nm/OD280nm ratio of 1 . 19 , clearly indicating that the higher-order N oligomers co-eluted with nucleic acids [18] , presumably RNA from the expression host . In contrast , the OD260nm/OD280nm ratio of the N2 peak was 0 . 72 , showing that this fraction contained much less RNA than fraction N1 [18] . The variability in the oligomeric state of N expressed in E . coli is consistent with previous studies that used either N purified from infected cells or recombinant N expressed in insects cells , although in the latter case multimers with a higher MW were observed [16] , [17] . We next used surface plasmon resonance experiments to test whether the recombinant protein in fraction N2 retained its capacity for non-specific RNA binding . We measured the interaction of N with a 20-nucleotides-long RNA , and determined that the Kd of N for RNA is 3 . 8 µM ( Figure S1 ) . This result demonstrated that the recombinant N protein present in peak N2 can still bind RNA and therefore presumably has the native fold . We therefore used this protein for 3D crystallization and structure determination . Using the N2 fraction , crystals were obtained in the P6 space group with unit cell parameters of a = b = 180 . 9 Å and c = 47 . 4 Å . The selenomethionyl protein crystallized in the same space group with similar unit cell parameters , a = b = 175 . 5 Å and c = 47 . 4 Å ( Table 1 ) . The structure was determined using the SAD technique with data recorded at the Se absorption edge from crystals of the selenomethionyl protein that diffracted to 2 . 3 Å resolution . The structure was subsequently refined using a native data set that extended to 1 . 6 Å resolution ( Table 1 ) . The asymmetric unit contains three N molecules , labeled α , β and γ in Figure 2A , that form two distinct hexameric rings in the crystal , labeled I and II in Figure 2B . Hexamer I is formed by six copies of subunit α that surround the crystallographic 6-fold axis , whereas hexamer II is formed by three β , γ dimers that surround the crystallographic 3-fold axis ( Figure 2B ) . The two sets of hexamers , which face in opposite directions and are offset by 10 Å in the direction of the crystallographic c axis ( Figure 2C ) , form layers along the [a , b] plane of the crystal . Stacking of the layers in the crystal results in the formation of two sets of tubes , one set formed by hexamers I and the other by hexamers II , that both run along the crystallographic c axis but in opposite directions ( Figure S2 ) . The crystal structure reveals that the N monomer consists of an orthogonal bundle of thirteen α-helices ( Figure 3 ) . The structure can be divided into three domains . Residues 1–32 form a flexible N-terminal arm containing two α-helical segments that extends away from the globular core of the protein . The globular core itself consists of two domains , one formed by six helices spread over residues 36–90 , 110–122 , 211–220 and the other one formed by four helices spread over residues 103–110 and 130–204 ( Figure 3 ) . The core domain in our structure is virtually identical to that in the previously reported crystal structure of N [15] with an rms deviation between the backbone atoms of the two structures of∼0 . 7 Å ( Figure S3B ) . The position and conformation of the N-terminal arm , however , are very different in the two structures ( Figure S3 ) , a finding that will be discussed below . The fold of RVFV N is currently unique in the PDB , but considering the high level of conservation in their amino-acid sequences ( average identity>30% ) ( Figure S4 ) , other Phlebovirus N proteins are likely to adopt a similar fold . In our crystals , N forms ring-shaped hexamers ( the subunits are denoted A to F as shown in Figure S2 ) that have a thickness of 45 Å , an external diameter of approximately 100 Å , and a central funnel-like aperture with a diameter that narrows from 50 to 30 Å ( Figure 2C ) . Multimerization appears to be driven by the extended N-terminal arm , which wraps around the external surface of the globular core of the adjacent subunit , fitting snuggly into a hydrophobic groove and burying a surface of 1456 Å2 ( Figure 4A ) . In particular , the aromatic rings of residues Y3 , F11 , W24 , F28 and Y30 and the aliphatic side chains of residues L7 , V9 , V16 , I21 and V25 project from the N-terminal arm and fill up the hydrophobic groove formed by regions 36–82 , 108–126 , and 207–210 of the core domain of the adjacent molecule ( Figure 4B/C ) . This arm-core interaction is repeated in a directional manner , such that the arm of subunit A extends into the hydrophobic groove of subunit B , B into C , C into D , D into E , E into F and F into A , creating the hexameric rings seen in the crystal . This mode of multimerization is consistent with mutagenesis data that mapped the interacting domain of the Phlebovirus N protein to its N-terminal arm [17] . Hexamers I and II in the crystals of the native protein superimpose very well , but the two rings in the crystals of the selenomethionyl protein are slightly different ( Figure S5A ) . While hexamer II formed by seleniated N is the same as the two hexamers formed by native N ( Figure S5B ) , the subunits in hexamer I are more closely packed about the 6-fold symmetry axis . The domain of N that is near the center of the ring , comprising the loop connecting helices α10 and α11 , occludes part of the central aperture , suggesting a twist in the assembly of the ring subunits ( Figure S5A ) . Superimposition of native and seleniated hexamers I based on subunits A , creates an 11° deviation between the planes of the two rings ( Figure S5C ) . Furthermore , comparison of the subunits in hexamers I formed by native and seleniated protein shows that the contraction of the ring is due to a lateral slippage between adjacent subunits ( Figure S6 ) . As a result of the slightly different subunit organization , the asymmetric unit is shorter in crystals of the seleniated protein and the length of the crystallographic a and b axes is decreased by about 5 Å ( Table 1 ) . The existence of two types of rings in the crystals demonstrates the natural ability of N to form oligomers with different subunit organizations , providing a basis for the formation of serpentine-like RNP structures . The core of the N protein has a concave crescent shape and the relative orientation of its two domains is reminiscent of a head of pliers , suggestive of a role in grabbing genomic RNA ( Figure 3C ) . This cleft is sandwiched between three helices on one side ( α4 , α5 , α7 ) and two 310-helices ( η4 , η5 ) followed by three α-helices on the other ( α9 , α10 , α11 ) , a fold in accordance with the “ ( 5H+3H ) ” structural motif for RNA binding [19] . Furthermore , analysis of the electrostatic surface potential reveals a positively charged patch located within the inner part on one side of the hexamers ( Figure 5 ) . This patch includes residues R64 , K67 and K74 that are evolutionary conserved across Phleboviruses ( Figure S4 ) . To test whether this positively charged patch indeed constitutes the RNA binding site , we expressed and purified a triple RVFV N mutant ( R64D , K67D , K74D ) . The triple mutant eluted from the gel filtration column as a single peak , corresponding to N2 ( Figure S7A ) . SDS-PAGE analysis revealed that the peak fractions contained a protein of the size expected for N ( 27 kDa ) ( Figure S7A , inset ) , and mass spectrometry confirmed the protein to be RVFV N . The OD260nm/OD280nm ratio of the peak fraction was 0 . 52 , indicating that this fraction contained only protein [20] . Binding studies using surface plasmon resonance spectroscopy with a 20-nucleotides-long RNA showed that the triple mutant lost its ability to bind RNA , supporting the notion that the positively charged patch serves as the RNA binding cleft ( Figure S7B ) . Taking as a guide the structure of the rabies virus N protein bound to single-stranded RNA ( PDB code: 2GTT [11] ) , we could position an RNA molecule in the concave surface between the two core domains of the RVFV N protein , such that the RNA sugar phosphate backbone interacts with the positive charges in the basic cleft . The model of RVFV N protein bound to RNA further showed that each N subunit can accommodate approximately six RNA bases ( Figure S8 ) . The crystal structure showed that RVFV N forms hexameric rings . To assess whether N also forms hexamers in solution , we prepared negatively stained samples for analysis by electron microscopy ( EM ) . EM images of fraction N2 , which contained only protein and was used for 3D crystallization , did not show any ring-shaped complexes ( data not shown ) , consistent with the SEC result that showed that this fraction contains only small oligomers . By contrast , EM images of fraction N1 , which contained both protein and RNA , revealed distinct circular structures with diameters ranging from∼70 to 100 Å , which were stable over a period of one month ( Figure 6A ) . The images thus suggest that the formation of stable higher-order N oligomers requires the protein to associate with RNA and that the resulting higher-order oligomers have a ring-shaped structure . To obtain a better understanding of the structure of N-RNA complexes , we calculated 3D reconstructions of the ring-shaped complexes seen in fraction N1 . The small size of the complexes prevented us from using vitrified specimens for EM imaging , and we therefore prepared samples by cryo-negative staining . This specimen preparation method provides the high contrast of stain but minimizes the artifacts associated with conventional negative staining [21] . The N-RNA complexes adsorbed to the carbon support film preferentially with the flat side of the ring , making it necessary to use the random conical tilt approach to calculate 3D reconstructions [22] . We recorded a total of 30 image pairs at tilt angles of 50° and 0° , from which we selected 10 , 764 particle pairs . The particles from the images of the untilted specimen were classified into 100 classes , which revealed a variety of oligomers , ranging from tetramers to octamers ( Figures 6B and S9 ) . About 57% of all the particles were ring-shaped oligomers . The hexamer was the most abundant species with 24% , followed by the pentamer ( 22% ) , the heptamer ( 7% ) , and finally the octamer ( 4% ) . The averages of the various oligomers revealed a large variability in the ring shape , pointing to structural flexibility in the various N-RNA complexes . Because the hexamer was most prevalent and because N alone formed hexamers in the 3D crystals , we focused on calculating a 3D reconstruction of the hexameric N-RNA complex . We combined the particles from classes that produced the most similar averages ( 399 particles from 2 classes ) and calculated a 3D density map using the particles selected from the images of the tilted specimen and the best 10% of particles selected from the untilted specimen . According to the Fourier shell correlation ( FSC ) = 0 . 5 criterion , the final density map had a resolution of 25 Å ( Figure S10 ) . With a diameter of about 100 Å and a thickness of about 45 Å ( Figure 6C ) , the EM density map of the N-RNA complex has virtually identical dimensions as the crystal structures of the N hexamer . Accordingly , the EM density map nicely accommodated the crystal structure of the hexamer , illustrating that the hexamers , and by extension also the other ring-shaped oligomers , are compatible with RNA binding ( Figure 6D ) .
The N protein is the most abundant viral protein in the Phlebovirus virion and plays a key role in encasing vRNA in a protective coat . We have determined the crystal structure of RVFV N in a hexameric form , which shows largely the same fold that was previously seen in a crystal structure of monomeric N [15] . The two structures differ , however , in the position of the N-terminal arm . In the previous structure , the N-terminal arm packs closely against the core domain , while it extends away from it in our structure ( Figure S3 ) . Extension of the N-terminal arm is crucial for the oligomerization of N , as it mediates the interaction with the adjacent subunit in the crystallographic hexamer . We believe that the RVFV N hexamer is biologically relevant , because oligomers have been observed for many other N proteins [9] , [10] , [11] , [12] , [13] , [14] and EM revealed that the RVFV N-RNA complex also forms ring-shaped oligomers in solution ( Figure 6 ) . The different position of the N-terminal arm in our and the previous structure is intriguing as it may reflect the structural change that has to occur for N to multimerize , thus potentially providing a clue to the mechanism underlying the formation of a ribonucleocapsid . In the hexamer , the N-terminal arm lies in a hydrophobic pocket of the adjacent subunit , thus mediating an inter-molecular interaction ( Figure 4 ) . By contrast , in monomeric N , the N-terminal arm makes an intra-molecular interaction and binds to the same hydrophobic pocket but in its own core domain , burying a surface area of 1179 Å2 ( Figure S11 ) . The inter- and intra-molecular interactions with the N-terminal arm are mediated largely by the same residues of the core domain ( Figure 4C and Figure S11B ) . Interestingly , the intra-molecular interaction of the N-terminal arm not only fills the hydrophobic pocket of its own core domain , thus preventing oligomerization , but also covers the RNA binding cleft , so that N in this conformation is incapable of binding RNA . For a monomer , the closed conformation is presumably more favorable , because it reduces the hydrophobic surfaces on both the N-terminal arm and oligomerization groove . In case the closed conformation is a “waiting” conformation before oligomerization; residues involved in the molecular interaction would have to compete for the oligomerization groove and expose the hydrophobic side of the arm ( Video S1 ) . Our SEC analysis shows that peak N2 , which lacks RNA , contains only small oligomers , suggesting that the inter-molecular interactions mediated by the N-terminal arm are not very strong on their own , potentially because the intra-molecular interactions outcompete the inter-molecular interactions , and thus do not support large oligomer formation . The weak interactions between N proteins would allow easy addition and removal of subunits . The fact that we see hexamers in our crystals may be explained by the high protein concentration used for crystallization trials that drives the small units of nucleoproteins to assemble into larger stable oligomers . In solution , however , stabilization of the oligomers may require the additional association of the subunits with RNA . Binding to RNA would align N proteins to each other and increase their local concentration , thus stabilizing the inter-molecular interactions of the N-terminal arms and resulting in the stable , ring-shaped oligomers seen in SEC peak N1 ( Figure 6 ) . This model of RNA-stabilized oligomers provides an elegant molecular explanation for why N proteins can have an inherent tendency to multimerize without forming undesired , large oligomers in the absence of RNA . With only six subunits and a diameter of 100 Å , the RVFV N ring is the smallest one among the ring-shaped oligomers seen in crystal structures of N proteins from negative strand viruses ( Figure 7 ) . Although there are clearly common structural features in the oligomers , the mode of how the subunits interact with each other varies . In rabies virus ( RV ) , vesicular stomatitis virus ( VSV ) , respiratory syncytial virus ( RSV ) and influenza virus , extensions at both the N and C termini of the polypeptide are involved in organizing adjacent subunits into an ordered assembly [11] , [12] , [14] . By contrast , it is only the interaction of the N-terminal arm of RVFV N with the hydrophobic pocket of the neighboring subunit that mediates the contacts between adjacent subunits in oligomers . While this interaction appears sufficient to promote efficient protein polymerization , it leaves a significant degree of freedom at the level of lateral interactions . This plasticity is illustrated by the slightly different positions of the N-terminal arm on the core domain of the neighboring subunit seen in hexamers I and II in the crystals of the native and seleniated proteins ( Figure S5 ) . As a result , like the N proteins of RV and RSV , RVFV N can form rings with deformed shapes and a variable number of subunits ( Figure 6B ) and , although not yet observed , N may even have the capacity to form oligomers with a superhelical arrangement of the subunits . Although EM of RVFV N-RNA complexes also showed ring-shaped oligomers ( Figure 6 ) , it is not clear whether rings are the building block of the native ribonucleocapsid . The RNPs of several Mononegavirales have a superhelical subunit arrangement , including those of RV [11] , VSV [12] , RSV [14] , measles [9] , [10] and mumps [13] . However , the RNPs of Phleboviruses do not assemble into a highly ordered structure , but rather into flexible filamentous assemblies [15] , [23] , [24] , [25] , [26] . In particular , EM images of RNPs from RVFV [15] and other Bunyaviridae [23] , [24] , [25] display an extended filament-like structure , but they do not rule out some degree of symmetry in the way the vRNA is packaged . While it thus remains uncertain whether the RVFV ribonucleocapsid is formed by stacked rings or a superhelical oligomer or even a mixture thereof , the flexibility in the interaction between adjacent subunits would allow great variability in the architecture of the ribonucleocapsid . Flexibility in the contacts between N subunits allows the assembly to readily adapt to distortions introduced by external constraints or signals within the infected cells , while maintaining the connectivity of the RNA . The crystal structure of the hexamer reveals that several positively charged residues are clustered in a cleft that can accommodate a single molecule of RNA ( Figures 5 and S8 ) . This finding is in agreement with the proposal of Luo and collaborators that N RNA binding site is formed by two domains that contain a “ ( 5H+3H ) ” structural motif [19] . Genomic RNA would thus run like a belt inside the ring and be completely concealed from the innate immune system of the host cell , in a manner similar to the ribonucleocapsid of the rabies virus [11] . Each N subunit can accommodate up to six bases , so that one turn of the RNA inside the hexameric ring would translate to∼36 bases . Our single-particle EM reconstruction of the hexameric RVFV N-RNA complex is consistent with the crystallographic hexamer of N , but it does not show the RNA inside the ring ( Figure 6D ) . Considering the limited resolution of the EM density map of 25 Å , the small mass of 30 RNA bases and the negative charge of RNA , which would favor positive staining of the RNA , it is not surprising that the RNA is not visible in the EM map . By considering that the thickness of the hexameric N-RNA complex is about 45 Å ( Figure 6C ) and making the simplifying assumption that the entire nucleocapsid is formed by stacked hexamers , the ribonucleocapsid of the S segment of 1690 nt would span a total linear distance of about 0 . 25 µm . This value is consistent with the size of 0 . 27 µm of the ribonucleocapsid of Uukuniemi virus seen in EM images [27] , whose genome is only slightly larger than that of RVFV . Earlier studies showed that transcription and replication require not only the polymerase L but also the N protein [28] , implying that naked vRNA cannot be transcribed [1] . While it has been established that the two proteins are positioned in close proximity to each other , as L is recruited to the vRNA through a panhandle structure [29] and N through a short region in the 5′ region of the ORF [30] , how the two proteins interact with each other remains unclear . A recent study found a conserved region in the second domain of N consisting of helices α4 , α5 , α6 [31] . This domain may well play a role in the stacking of N subunits in the oligomer , but it could also mediate a transient interaction with L and promote a temporary release of N , thus liberating the RNA to become accessible for transcription by L . Additionally , helices α1 , α12 and α13 are located at the periphery of the hexameric ring , and residues projecting from these helices are also likely to form a significant part of the L-binding surface on the ribonucleocapsid . Given the substantial contribution of the N-terminal arm to the buried interface ( 1456 Å2 out of a total of 1640 Å2 buried surface between two adjacent N subunits ) , it is conceivable that interactions of α1 of the N-terminal arm with L could lead to a local unwinding of the filament structure and the exposure of vRNA while avoiding complete disassembly of the ribonucleocapsid . In conclusion , the structure of the hexamer formed by the RVFV N protein presented here shows that oligomerization is mediated by a flexible N-terminal arm , which binds a hydrophobic pocket in the adjacent subunit . The different hexamers seen in our crystals and the variability in the oligomeric state of N-RNA complexes seen in EM images demonstrate substantial flexibility in the interaction between subunits . Furthermore , comparison with a previous structure of the RVFV N protein suggests an elegant mechanism that allows the formation of stable N oligomers only in the presence of RNA . Finally , the nucleoprotein structure identifies potential sites that could be targeted for drug development . For instance , compounds blocking ribonucleocapsid assembly either by interfering with RNA binding or by trapping the N-terminal arm of N in a conformation that is not compatible with oligomerization , could serve as starting points to design specific antiviral molecules .
cDNA corresponding to the RVFV N protein ( strain Smithburn DQ380157 . 1 ) was cloned by recombination ( Gateway , Invitrogen ) into the pETG20A vector ( kindly provided by Dr . Arie Geerlof ) , which adds a cleavable N-terminal thioredoxin-hexahistidine tag , and used to transform E . coli strain C41 ( Avidis ) carrying the pRARE plasmid ( Novagen ) . Bacteria were grown in TB medium ( Athena Enzyme ) at 37°C to an OD600nm of 0 . 5 . Expression was induced with 0 . 5 mM IPTG , and bacteria were grown overnight at 17°C . Cells were pelleted , resuspended in 30 ml of lysis buffer ( 50 mM Tris , pH 8 , 300 mM NaCl , 5 mM imidazole , 5% glycerol , 0 . 1% Triton X-100 , 2 mM EDTA ) , frozen , and stored at -80°C . N was purified at 4°C . A frozen pellet was melted on ice , sonicated , and the lysate was cleared by centrifugation at 20 , 000 rpm for 30 min . The protein was first purified by metal affinity chromatography using a 5 ml HisPrep column ( GE Healthcare ) . The tag was then removed by cleavage with TEV protease , and the protein was further purified with a second metal affinity column followed by size exclusion chromatography ( SEC ) using a Superdex 200 column ( GE Healthcare ) in 10 mM HEPES , pH 7 . 5 , 300 mM NaCl . The R64D/K67D/K74D triple mutant was generated by first simultaneously introducing the R64D and K67D mutations followed by introducing the K74D mutation into the RVFV N cDNA using the QuickChange Site-Directed Mutagenis Kit ( Agilent ) . The sequences of the primers used to introduce the point mutations were: R64D/K67D forward: CTGGCTCTAACTgaTGGCAACgAcCCCCGGAGGATG , R64D/K67D reverse: CATCCTCCGGGGgTcGTTGCCAtcAGTTAGAGCCAG , K74D forward: CGGAGGATGATGATGgAcATGTCAAAAGAAGGC , and K74D reverse: GCCTTCTTTTGACATgTcCATCATCATCCTCCG . The complete coding region of each mutant was sequenced to confirm the desired modification . The triple mutant was expressed and purified analogous to the wild-type protein , and the expressed protein was verified by mass spectrometry . Analytical SEC was performed on a KW 803 column ( Shodex ) using a High Pressure Liquid Chromatography Alliance 2695 system ( Waters ) , and absorbance was measured at both 260 nm and 280 nm . The SEC column was calibrated with Kit LMW markers ( GE Healthcare ) . The protein eluted in two peaks , with apparent molecular weights of∼300 kDa ( N1 ) and 94 kDa ( N2 ) . The N1 peak was used for EM analysis of N-RNA complexes and the N2 peak for 3D crystallization screens of N . Binding affinities of wild-type and mutant N protein for ssRNA were determined using a ProteOn XPR36 instrument ( Bio-Rad Laboratories , Inc ) . NeutrAvidin ( Thermo Scientific ) was amine-coupled to a carboxylated sensor surface ( GLM sensor chip ) to a final immobilized level of 6000 RU . To test non-specific binding by N , biotin-labeled ssRNA oligonucleotides with a non-relevant sequence from the dengue virus 5′ non-translated region ( RNA20-3′biotin: GAGUUGUUAAUCUUUUUUUU-biotin; Sigma ) were diluted to 10 nM in sodium acetate ( pH 5 . 5 ) and injected for two minutes at a flow rate of 25 µl/min . Association and dissociation phases were measured for 240 sec and 600 sec , respectively . Measurements were performed in buffer containing 10 mM HEPES , pH 7 . 5 , 300 mM NaCl , 0 . 005% NP-20 . Data were analyzed in ProteOn Manager version 2 . 0 . N protein in 10 mM HEPES , pH 7 . 5 , 300 mM NaCl collected from fraction N2 was concentrated to 7 . 8 mg/ml , and 2 µl of protein solution was mixed with 2 µl of reservoir solution containing 200 mM MgNO3 and 17% ( w/v ) PEG 3350 for crystallization at 20°C using the hanging drop method . SDS-PAGE analysis of dissolved crystals confirmed that they contained full-length N protein . The crystals were flash-frozen in liquid nitrogen using 5% glycerol as cryo-protectant . The RVFV N protein crystallizes in space group P6 with unit cell parameters of a = b = 180 . 9 Å , c = 47 . 7 Å for the native protein and a = b = 175 . 5 Å , c = 47 . 4 Å for the seleniated protein . A native data set extending to 1 . 6 Å resolution and a Se-Met data set extending to 2 . 3 Å resolution were collected on beamline ID14–4 at the ESRF ( Grenoble , France ) . The Se-Met data set was collected at the Se absorption edge . Data were processed using the program XDS [32] . Of a total of 33 Se sites for the three monomers in the asymmetric unit , the position of 27 sites were identified using the program SHELXD [33] to analyze anomalous data ranging from 10 to 2 . 3 Å . After initial phase calculation and modifications with the SHELX suite , a readily interpretable map was obtained with an overall figure of merit of 0 . 62 . The program ARP/wARP [34] was used to generate an initial model , and a complete model for the three independent monomers was built using COOT [35] . Using this model , the native data set was subsequently solved by molecular replacement using the program Phaser [36] . The program REFMAC5 with the TLS option was used for crystallographic refinement [37] . The final models were assessed with PROCHECK [38] . Surface electrostatics were calculated using DELPHI [39] . Sequences were aligned using Muscle [40] and seaview [41] . Intermediate structures for the morphing were generated using LSQman [42] . Figures and movie were generated with the programs ENDscript , ESPript [43] and PyMOL ( http://www . pymol . org ) . Samples were prepared by negative staining and cryo-negative staining with uranyl formate as described [44] . For specimens prepared by conventional negative staining , images were taken using Philips CM10 electron microscope equipped with a tungsten filament and operated at an acceleration voltage of 100 kV . Images were recorded on a 1 k×1 k Gatan CCD camera at a magnification of 52 , 000×using a defocus value of –1 . 5 µm . For cryo-negative staining specimens , images were recorded using a Tecnai F20 electron microscope ( FEI ) , equipped with a field emission gun and operated at an acceleration voltage of 200 kV . Grids of cryo-negatively stained specimens , used to collect image pairs of specimens tilted to 50° and 0° , were loaded on an Oxford cryo-transfer holder and maintained at liquid nitrogen temperature during image acquisition . Images were taken at a magnification of 50 , 000× , with a defocus value of –2 . 0 µm for images of untilted specimens and –1 . 8 µm for specimens tilted to 50° . All images were recorded using low-dose procedures on Kodak film SO163 and developed for 12 min with full-strength Kodak D-19 developer at 20°C . Electron micrographs were digitized with a SCAI scanner ( Zeiss ) using a step size of 7 µm , and 3×3 pixels were averaged to obtain a pixel size of 4 . 2 Å on the specimen level for cryo-negatively stained specimens . 3D reconstructions from the cryo-negatively stained preparations were calculated using the SPIDER software package [45] . 10 , 764 particle pairs were interactively selected from a total of 30 image pairs using WEB , the display program associated with SPIDER , and windowed into small images of 60×60 pixels . The particles from the images of the untilted specimens were classified over 10 cycles of K means classification and multi-reference alignment specifying 100 output classes . 3D density maps of individual classes were calculated with the corresponding particles selected from the images of the tilted specimen and using the back-projection , back-projection refinement , and angular refinement procedures implemented in SPIDER . The final 3D reconstruction of the hexameric N-RNA complex included 439 particles ( 399 particles from images of tilted specimens and 40 particles from images of untilted specimens ) and its resolution was estimated by Fourier shell correlation ( FSC ) to be 25 Å according to the FSC = 0 . 5 criterion . The crystal structure of hexamer I formed by native RVFV N was first manually docked into the EM density map and then refined using the program UCSF Chimera [46] . | The Rift Valley fever virus ( RVFV ) , a negative strand RNA virus spread by infected mosquitoes , affects livestock and humans who can develop a severe disease . We studied the structure of its nucleoprotein ( N ) , which forms a filamentous coat that protects the viral RNA genome and is also required for RNA replication and transcription by the polymerase of the virus . We report the structure of the RVFV N protein at 1 . 6 Å resolution , which reveals hexameric rings with an external diameter of 100 Å that are formed by exchanges of N-terminal arms between the nearest neighbors . Electron microscopy of recombinant protein in complex with RNA shows that N also forms rings in solution . A reconstruction of the hexameric ring at 25 Å resolution is consistent with the hexamer structure determined by crystallography . We propose that slight structural variations would suffice to convert a ring-shaped oligomer into subunits with a super-helical arrangement and that this mode of protein-protein association forms the basis for the formation of filamentous ribonucleocapsids by this virus family . Both the RNA binding cleft and the multimerization site of the N protein can be targeted for the development of drugs against RVFV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biomacromolecule-ligand",
"interactions",
"medicine",
"macromolecular",
"assemblies",
"microbiology",
"protein",
"folding",
"protein",
"structure",
"infectious",
"diseases",
"biology",
"biophysics",
"macromolecular",
"complex",
"analysis",
"virology",
"viral",
"diseases",
"c... | 2011 | The Hexamer Structure of the Rift Valley Fever Virus Nucleoprotein Suggests a Mechanism for its Assembly into Ribonucleoprotein Complexes |
The somatic genome of the ciliated protist Tetrahymena undergoes DNA elimination of defined sequences called internal eliminated sequences ( IESs ) , which account for ∼30% of the germline genome . During DNA elimination , IES regions are heterochromatinized and assembled into heterochromatin bodies in the developing somatic nucleus . The domesticated piggyBac transposase Tpb2p is essential for the formation of heterochromatin bodies and DNA elimination . In this study , we demonstrate that the activities of Tpb2p involved in forming heterochromatin bodies and executing DNA elimination are genetically separable . The cysteine-rich domain of Tpb2p , which interacts with the heterochromatin-specific histone modifications , is necessary for both heterochromatin body formation and DNA elimination , whereas the endonuclease activity of Tpb2p is only necessary for DNA elimination . Furthermore , we demonstrate that the endonuclease activity of Tpb2p in vitro and the endonuclease activity that executes DNA elimination in vivo have similar substrate sequence preferences . These results strongly indicate that Tpb2p is the endonuclease that directly catalyzes the excision of IESs and that the boundaries of IESs are at least partially determined by the combination of Tpb2p-heterochromatin interaction and relaxed sequence preference of the endonuclease activity of Tpb2p .
Transposons represent harmful genetic elements because they potentially rearrange their host's genome , and their integration into important coding or regulatory regions can have deleterious effects . Transposons are therefore considered “junk” DNAs [1] , and hosts have evolved genome defense mechanisms to counteract these selfish elements [2] . However , transposons may not be just junk because they potentially contribute to the evolution of the host by genome rearrangements , alternating gene expression networks , or providing new genes from transposons to the host [3] . Therefore , host organisms must evolve by balancing the harmfulness and usefulness of transposons . An evolutional product likely created by such a balance is the programmed DNA elimination in the ciliated protist Tetrahymena , in which the transposon-related sequences are eliminated by a domesticated piggyBac transposase [4] . Most ciliates display nuclear dimorphism [5] . The germline micronucleus ( Mic ) is transcriptionally inert during vegetative growth , whereas the somatic macronucleus ( Mac ) provides the cell with most if not all RNA . When nutrients are scarce , Tetrahymena undergoes sexual reproduction ( conjugation ) , in which two mating partners form a pair ( Fig . 1A ) and their Mics undergo meiosis ( Fig . 1B ) . Three of the meiotic products are degraded , and the remaining product divides mitotically ( Fig . 1C ) . One of the products is exchanged with the mating partner and afterwards , the two pronuclei fuse to form the zygote ( Fig . 1D ) . The zygotic nucleus divides twice mitotically ( Fig . 1E ) ; of the four mitotic products , two become the new Mics , and the other two develop to become the new Macs ( Fig . 1F ) . The parental Mac is degraded at the end of this process , and the progeny resume vegetative growth when nutrients are supplied ( Fig . 1G ) . Two major types of programmed genome rearrangements occur in the developing new Mac of Tetrahymena . The first type is chromosome breakage , which leads to the fragmentation of germline chromosomes . The chromosome breakage occurs at conserved 15-nt sequences called chromosome breakage sequences ( CBSs ) . It has been estimated that there are ∼250 CBSs in the Mic genome [6] , [7] . The sites of chromosome breakages are healed by de novo telomere formation [8] . The second type of genome rearrangement in Tetrahymena is DNA elimination of internal eliminated sequences ( IESs ) , followed by ligation of their flanking sequences [9] by the non-homologous end joining ( NHEJ ) pathway [10] . The indispensability of the NHEJ pathway for DNA elimination was also demonstrated for another ciliate , Paramecium [11] . It has been estimated that there are over 8 , 000 different IESs , which represent ∼30% of the Mic genome [12] , [13] . Because many IESs contain transposon-related sequences , it is assumed that DNA elimination is a process that removes potentially harmful transposons from the transcriptionally active somatic genome [14] . Moreover , because some IESs are in regulatory regions and exons of genes , DNA elimination is necessary to create the streamlined functional somatic genome [15] . Despite the fact that different IESs do not share any detectable common sequences within themselves and in their flanking regions , invariable sets of IESs are eliminated from the Mac , and the majority of their boundaries occur within a few to several base pairs . The identities of IESs are most likely determined epigenetically by an RNAi-related mechanism [16] , [17] . While the Mic is transcriptionally inert during the vegetative growth , non-coding RNA transcription occurs in the Mic during the early stages of conjugation . The ∼28–29-nt siRNAs produced from the non-coding RNAs are selected for IES specificity by selective degradation of siRNAs complementary to the parental Mac genome [12] . The selected IES-specific siRNAs eventually induce the establishment of heterochromatin specifically on IESs in the developing new Mac . This heterochromatin comprises tri-methylated histone H3 at lysine 9 and lysine 27 ( H3K9me3 , H3K27me3 ) [18] , [19] and the chromodomain protein Pdd1p [20] , which binds to the histone H3 modifications . Although both H3K9me3 and H3K27me3 have been shown to play an important role in DNA elimination [18] , [19] , the functional distinction between H3K9me3 and H3K27me3 in the DNA elimination process , if any , is not clear . The heterochromatinized IESs are assembled into heterochromatin bodies located at the nuclear periphery [21] , and each IES is eventually excised as one linear or circular piece of DNA [22]–[24] . Artificially tethering Pdd1p to DNA is sufficient to induce DNA elimination [25] , indicating that heterochromatin but not the RNAi-related mechanism is the immediate signal inducing DNA elimination . Previous studies have shown that , in some IES elements , the deletion boundaries are determined by flanking cis-acting sequences located 40–50 bp away , and the precise nature of the deletion is dependent on the sequences at the boundaries [26]–[28] . However , because no sequence homology has been observed across different elements , it is unclear how the boundaries of IESs are determined . In addition it is not known whether and how heterochromatin is involved in the boundary determination . We previously reported that the domesticated piggyBac transposase-like protein Tpb2p ( Tetrahymena piggyBac-like transposase 2 ) localizes to the heterochromatin bodies and is essential for DNA elimination [4] . Furthermore , we reported that Tpb2p has the ability to produce DNA double-strand breaks at a boundary sequence of an IES in vitro [4] . Therefore , we hypothesized that Tpb2p is recruited to the IESs by directly interacting with a heterochromatin component and then inducing a DNA double-strand break at its preferential DNA sequence near the heterochromatin to execute DNA elimination . To validate this hypothesis , we analyze the roles of the individual domains of Tpb2p genetically and biochemically to understand 1 ) how Tpb2p interacts with heterochromatin; 2 ) whether the Tpb2p-heterochromatin interaction is necessary for DNA elimination; 3 ) what is the sequence preference of the endonuclease activity of Tpb2p; and 4 ) whether the sequence preference contributes to the choice of boundary sequence of IESs . Based on the results , we discuss how Tpb2p is involved in reproducible DNA elimination and how a domesticated transposase has evolved to catalyze the DNA elimination of transposons .
Recombinantly expressed Tpb2p has been demonstrated to have the ability to produce DNA double-strand breaks in vitro [4] . However , it is unclear whether the endonuclease activity of Tpb2p is necessary for DNA elimination in vivo because Tpb2p is also necessary for the formation of the heterochromatin bodies , which is believed to be a prerequisite of DNA elimination . To test whether the endonuclease activity of Tpb2p is needed for DNA elimination in vivo , we attempted to express a catalytically inactive Tpb2p mutant in a TPB2-null background . RNAi knockdown , which has been previously used to study the function of TPB2 [4] , only partially down-regulates TPB2 expression and is difficult to use for genetic rescue experiments . Therefore , we attempted to obtain knockout ( KO ) strains of TPB2 but without success . This lack of success might be because TPB2 is a haplo-insufficient gene , and heterozygous TPB2 KO strains are not viable . This is consistent with the fact that RNAi knockdown of TPB2 caused nearly complete lethality of sexual progeny [4] . To overcome this problem , we created TPB2 conditional knockout ( cKO ) strains . To make TPB2 expression conditional , we produced a TPB2 cKO construct in which the endogenous TPB2 promoter was replaced with the cadmium-inducible MTT1 promoter ( Fig . 2A ) . The TPB2 cKO construct was first introduced into the TPB2 locus in the Mic by homologous recombination to produce heterozygous TPB2 cKO strains , and then , two heterozygous TPB2 cKO strains were mated to obtain homozygous TPB2 cKO strains ( hereafter referred to as TPB2 cKO strains ) ( Fig . 2B ) . In these genetic crosses , TPB2 expression was continuously induced during conjugation to obtain viable progeny . It is known that the MTT1 promoter is leaky in the standard culture conditions [29] . Therefore , we used a metal-depleted medium ( see Materials and Methods for details ) to minimize the basal level activity of the MTT1 promoter . Western blot analysis and immunofluorescent staining using an anti-Tpb2p antibody demonstrated that in our culture conditions , Tpb2p was undetectable in the absence of cadmium during conjugation ( see Fig . 3B −Cd2+ lanes ) of the TPB2 cKO strains . In contrast , when cadmium was added , Tpb2p expression was clearly detected ( see Fig . 3B +Cd2+ lanes ) . Therefore , Tpb2p expression can be conditionally knocked out in the TPB2 cKO strains . It has been reported that the formation of heterochromatin bodies and DNA elimination can be inhibited by the RNAi knockdown of TPB2 [4] . To determine if the cKO of TPB2 phenocopies the RNAi knockdown of TPB2 , the formation of heterochromatin bodies in the TPB2 cKO cells was observed by immunofluorescent staining of the heterochromatin component Pdd1p . When TPB2 expression was induced in the presence of cadmium , the TPB2 cKO strains formed Pdd1p-containing heterochromatin bodies in the new Macs ( Fig . 2C , “induced” ) . In contrast , in the absence of TPB2 induction , Pdd1p-stained heterochromatin did not form large bodies but remained as dispersed small foci in the new Mac ( Fig . 2C , “non-induced” ) . Next , DNA elimination in TPB2 cKO strains was observed by DNA fluorescence in situ hybridization ( FISH ) against Tlr1 IESs , which are moderately repeated ( ∼30 copies ) in the Mic genome [30] . DNA elimination in wild-type cells is normally completed by ∼16-hr post-mixing [31] . In the presence of cadmium , Tlr1 IESs were undetectable in the new Macs at 36-hr post-mixing and detected only in the Mic ( Fig . 2D , “induced” , na = new Mac , i = Mic ) , indicating that these IESs were removed completely from the new Macs . In contrast , the Tlr1 IESs remained in the new Mac even at 36-hr post-mixing when TPB2 expression was not induced ( Fig . 2D , “non-induced” ) . These results indicate that the TPB2 cKO strains exhibit defects in heterochromatin body formation and DNA elimination in the absence of the induction of TPB2 , as it was previously reported for the TPB2 RNAi knockdown strains . Next , we attempted to establish a genetic rescue system in which the non-essential MTT1 locus [29] of the parental Mac in the TPB2 cKO strains was replaced with a MTT2 cassette expressing a gene of interest under the control of the copper-inducible MTT2 promoter [32] ( Fig . 3A ) . Before starting the rescue experiments , we first tested whether the TPB2 cKO locus could be kept silent in the presence of copper . We incubated the conjugating TPB2 cKO cells either with cadmium or copper , and Tpb2p expression was analyzed by western blot using an anti-Tpb2p antibody . Although Tpb2p expression was induced in the presence of cadmium , it was undetectable from the cells incubated with copper ( Fig . 3B ) . These results indicate that the expression of TPB2 in the TPB2 cKO locus , which is under control of the cadmium-inducible MTT1 promoter , is not induced by the addition of copper . Next , the MTT2 cassette containing the wild-type TPB2 tagged with HA epitope ( Fig . 3A , MTT2-HA-TPB2 ) was introduced into the TPB2 cKO strains ( Fig . 3C ) , and the cells were incubated with or without copper . We analyzed HA-Tpb2p expression by western blot using an anti-HA antibody and observed that HA-Tpb2p was detected only when copper was added to the medium ( Fig . 3D , compare WT +/− Cu2+ ) . Therefore , the expression of a gene in the MTT2 cassette is induced in the presence of copper . When the MTT2 cassette containing the wild-type TPB2 was introduced into the TPB2 cKO strains , heterochromatin body formation ( Fig . 3E , WT ) and DNA elimination ( Fig . 3F , WT ) were restored in the presence of copper . Also , expression of the wild-type TPB2 partially restored the progeny viability of the TPB2 cKO strains ( Fig . 3G , WT-rescue +Cu2+ ) . These results indicate that the wild-type TPB2 expressed from the MTT2 cassette in the parental Mac is sufficient for all essential steps of DNA elimination , although it might not be enough to restore some non-essential steps of DNA elimination . Therefore , the rescue system using the TPB2 cKO strains and MTT2 cassette can be used to assay functionalities of Tpb2p mutants . DNA elimination process was also analyzed by observing circularized excised IESs by PCR ( see Fig . 3H left ) . DNA elimination events in the wild-type cells release IESs in two different forms: the major linear form and the minor circular form [23] , [24] . The circular form of two different IESs , mse2 . 9 and R elements , were detected when the wild-type TPB2 was expressed in the TPB2 cKO background ( Fig . 3H , WT rescue +Cu2+ ) . However , the appearance of the excised mse2 . 9 IES was delayed in the conditional TPB2 KO cells expressing the wild-type TPB2 compare to the wild-type cells ( Fig . 3H , see Wild-type and WT rescue +Cu2+ ) , possibly because the ectopic expression of TPB2 from the MTT2 promoter in the parental Mac could not fully restore the function of endogenous TPB2 . This may explain why the progeny viability was much lower in the conditional TPB2 KO cells expressing the wild-type TPB2 than in the wild-type cells ( Fig . 3G ) . Tpb2p has the endonuclease catalytic domain that contains three aspartic acids that form the DDD catalytic triad ( Fig . 3A ) . We previously reported that the replacement of these three aspartic acids with leucines compromises the endonuclease activity of Tpb2p in vitro [4] . To understand the role of the endonuclease activity in vivo , the MTT2 cassette expressing the TPB2-CD mutant , in which the catalytic triad of Tpb2p was replaced with leucines ( D297L; D379L; D495L , “CD” in Fig . 3A ) , was introduced into the TPB2 cKO strains . The amount of Tpb2p-CD expressed from the MTT2 cassette after induction with copper was comparable to that of the wild-type Tpb2p from the cassette ( Fig . 3D , compare +Cu2+ lanes of WT and CD ) , indicating that the mutations do not significantly affect the stability of Tpb2p . The expression of the TPB2-CD mutant did restore heterochromatin body maturation , based on the localization of the heterochromatin component Pdd1p ( Fig . 3E , “CD” ) , but did not support the elimination of Tlr1 IESs from the new Macs , as evaluated by the fact that FISH using the probes complementary to Tlr1 IESs stains the new Mac ( Fig . 3F , “CD” ) . The circular form of two different IESs , mse2 . 9 and R elements , could not be detected by PCR when TPB2-CD expression was induced ( Fig . 3H , CD rescue +Cu2+ ) , indicating that no detectable IES excision was induced by the TPB2-CD expression . Consistent with the fact that DNA elimination is essential for the production of viable sexual progeny , expression of TPB2-CD mutant could not restore the progeny viability of the TPB2 cKO strains ( Fig . 3G ) . Altogether , we conclude that the endonuclease activity of Tpb2p is necessary for DNA elimination but dispensable for the heterochromatin body formation . The results above clearly indicate that the necessity of Tpb2p for heterochromatin body formation ( Fig . 2C ) should be attributed to an activity of Tpb2p other than its endonuclease activity . The endonuclease domain of piggyBac transposases is followed by a PHD finger-like domain [33] , [34] ( Supplemental Fig . S1 ) . Tetrahymena Tpb2p and its Paramecium homolog Pgm [35] also have a cysteine-rich domain downstream of their endonuclease domains ( Fig . 3A , Supplemental Fig . S1 ) . Although this cysteine-rich domain displays similarity to the PHD finger domain , it lacks a potential metal-binding residue in one of the two intermingled zinc-fingers ( Supplemental Fig . S1 ) . Therefore , it is unclear whether the cysteine-rich domain of Tpb2p has any biological role or if it is only a non-functional remnant of the PHD finger domain of the ancestral piggyBac transposase . To determine if the cysteine-rich domain of Tpb2p has any role in DNA elimination , a MTT2 cassette containing the TPB2-CRM mutant , in which two of the seven potential metal-binding core cysteine/histidine residues of the cysteine-rich domain were replaced with alanines ( C618A; C629A , “CRM” in Fig . 3 , Supplemental Fig . S1 ) , was introduced into the TPB2 cKO strains . The amount of Tpb2p-CRM expressed after induction with copper was comparable to that of the wild-type Tpb2p from the cassette ( Fig . 3D , compare +Cu2+ lanes of WT and CRM ) , indicating that the mutations did not significantly affect the stability of Tpb2p . We observed that the expression of TPB2-CRM did not restore heterochromatin body formation ( Fig . 3E , CRM ) , the elimination of Tlr1 IESs from the new Macs ( Fig . 3F , CRM ) , the formation of circularized excised IESs ( Fig . 3H , CRM rescue +Cu2+ ) , and the progeny viability ( Fig . 3G , CRM rescue +Cu2+ ) . Therefore , we concluded that the cysteine-rich domain of Tpb2p is essential for both heterochromatin body formation and the following DNA elimination . The cysteine-rich domain of Tpb2p resembles the PHD-finger domain ( Supplemental Fig . S1 ) , and some PHD finger-containing proteins bind to the N-terminal tail of histone H3 with methylated lysine residues [36] . Because histone H3 on the heterochromatinized IESs is specifically tri-methylated at lysine 9 and lysine 27 ( H3K9me3/K27me3 ) , we reasoned that Tpb2p might interact with heterochromatin through the interaction between its cysteine-rich domain and H3K9me3/K27me3 . This hypothesis is consistent with the fact that although the wild-type Tpb2p and Tpb2p-CD tightly co-localized with the heterochromatin component Pdd1p ( Fig . 3E WT , CD ) , the localization of Tpb2p-CRM did not completely overlap with that of Pdd1p ( Fig . 3E , CRM ) . Therefore , we aimed to investigate if the cysteine-rich domain of Tpb2p interacts with H3K9me/K27me . We prepared C-terminally biotinylated peptides representing the amino acids 1–19 or 16–35 of Tetrahymena histone H3 ( Supplemental Table S2 ) . The peptides were either unmodified or tri-methylated at lysine 4 , 9 or 27 . In addition , unmodified peptides with scrambled amino acid orders were prepared . The peptides were bound to avidin-coated beads and incubated with a recombinant cysteine-rich domain ( aa 566–aa 657 ) of Tpb2p fused to a maltose-binding protein ( MBP ) -tag ( MBP-Tpb2p-CRD ) , and the proteins co-precipitated with the peptides were analyzed by western blotting using an anti-MBP antibody . We observed that more MBP-Tpb2p-CRD was precipitated with the unmodified peptides than with the beads only ( compare lanes 2 and 4 , lanes 2 and 7 or lanes 10 and 12 of Fig . 4 , top ) , whereas the amount of MBP-Tpb2p-CRD precipitated with these unmodified peptides was comparable to the amount precipitated with the corresponding unmodified peptides with scrambled amino acid orders ( compare lanes 3 and 4 , lanes 6 and 7 or lanes 11 and 12 of Fig . 4 , top ) . Therefore , some physical property of the peptides , such as charge , likely mediates the co-precipitation of MBP-Tpb2p-CRD with the unmodified peptides . Importantly , significantly more MBP-Tpb2p-CRD was co-precipitated with the peptides having tri-methylated lysines 9 or 27 than with the corresponding unmodified peptides ( compare lanes 4 and 5 or lanes 7 and 8 of Fig . 4 , top ) . Therefore , the presence of tri-methylations at lysine 9 or lysine 27 enhances the interaction between MBP-Tpb2p-CRD and the peptides . In contrast , tri-methylations at lysine 4 did not enhance co-precipitation of MBP-Tpb2p-CRD with the peptide ( compare lanes 12 and 13 of Fig . 4 ) . The mutations ( C618A , C629A ) at the putative metal-binding core of the cysteine-rich domain , which abolish the heterochromatin body formation in vivo ( Fig . 3E ) , inhibited the co-precipitation of MBP-Tpb2p-CRD with any of the peptides in vitro ( Fig . 4 , bottom ) . We conclude that Tpb2 can interact with the histone H3 tail through its cysteine-rich domain , and this interaction is significantly enhanced by the presence of tri-methylated lysines 9 or 27 . We have previously shown that recombinant Tpb2p , which is expressed from E . coli , produces DNA double-strand breaks ( DSB ) possessing 4-base 5′ overhangs at the left boundary sequence of Tetrahymena R-IES ( 5′-AGTGAT-3′ ) in vitro [4] ( see also Fig . 5B ) . However , it is unclear what sequence feature , if any , is recognized by Tpb2p . To better understand what sequence feature of an IES boundary is recognized by Tpb2p , the left R-IES boundary sequence ( 5′-AGTGAT-3′ ) was placed in the middle of otherwise artificial sequence , and every position of the boundary was substituted with every other possible nucleotide ( Fig . 5B ) . The radiolabeled oligo DNA duplexes were incubated with recombinant Tpb2p and analyzed as described above . As previously observed , when Tpb2p was incubated with an oligo DNA duplex having the wild-type R-IES boundary , the major 50-nt product , which is produced by the endonucleolytic cleavage between the first A and first G of the boundary ( see Fig . 5A ) , was detected ( Fig . 5B , the leftmost lane ) . Base substitutions at positions −1 , +1 , +4 and +5 did not significantly affect the choice of cleavage position by Tpb2p in this in vitro assay ( Fig . 5B ) . In contrast , when position +2 was substituted , the major 50-nt product was greatly reduced , and instead , a few nucleotide longer products were detected ( Fig . 5B ) . Similarly , when position +3 was substituted , a few nucleotide longer products were detected , although the 50-nt product was not significantly reduced ( Fig . 5B ) . Importantly , no obvious cleavage products were detected when the substrates were incubated with the catalytically inactive Tpb2p ( Fig . 5C ) , indicating that the observed products were produced by Tpb2p , but not by any contaminated bacterial endonucleases . All together , these results suggest that most nucleotides of the left R-IES boundary sequence are replaceable without disturbing the boundary recognition by Tpb2p , whereas the 5′-TG-3′ sequence at positions +2 and +3 are important for Tpb2p to execute precise cleavage of the left R-IES boundary in vitro . Because not all boundary sequences of IESs share 5′-TG-3′ at their +2 and +3 positions [22] , [37] , [38] , this dinucleotide sequence cannot be the sole sequence feature recognized by Tpb2p . Tpb2p may recognize multiple different sequence features , including 5′-TG-3′ . Alternatively , Tpb2p may recognize some complex combinatorial sequence feature , which is shared in all boundary sequences of IESs and was not completely disrupted by the single-base substitutions in this study , and 5′-TG-3′ may be a part of this combinatorial sequence feature . From this study , we can conclude that the endonuclease activity of Tpb2p has a relaxed but not completely identified sequence preference for its substrates . Although our genetic analyses indicated that the endonuclease activity of Tpb2p is required for DNA elimination ( Fig . 3 ) , none of the results directly demonstrated that Tpb2p is the “excisase , ” the enzyme that cuts out IESs in vivo . Above , we observed that the base substitution of +2 or +3 positions of the left R-IES boundary force Tpb2p to cleave a few bases downstream in vitro ( Fig . 5B ) . If the same base substitutions at the left R-IES boundary forced a shift in the position of DSB downward in vivo , it would support the assignation of the excisase function to Tpb2p . A previously established , in vivo IES elimination assay was used to test this possibility . It has been demonstrated that IESs introduced into a non-coding region of the ribosomal DNA ( rDNA ) are removed precisely as their endogenous counterparts , albeit with lower efficiency , when the rDNA construct is introduced into the developing new Mac [26] . We prepared three different rDNA constructs ( Fig . 6A , right top ) having the R-IES and its flanking regions with 1 ) no base substitution ( WT construct ) ; 2 ) T to G substitution at position +2 of the left boundary ( T+2G construct ) ; and 3 ) G to T substitution at position +3 of the left boundary ( G+3T construct ) . These constructs were introduced into the new Mac of conjugating wild-type cells by electroporation ( Fig . 6A ) . Twenty-four sexual progeny possessing the transgenic rDNA were pooled for each construct , and their genomic DNA was analyzed by PCR to observe elimination of the R-IES of the introduced rDNA ( Fig . 6A , bottom ) . First , the PCR products were analyzed by gel electrophoresis . Two major PCR products were detected from the cells transformed with the WT construct ( Fig . 6B , WT ) . The shorter ( 1 . 3 kbp ) and longer ( 2 . 4 kbp ) major products correspond with the R-IES locus on rDNA with or without IES elimination , respectively . A minor PCR product ( 1 . 8 kbp ) , in which a part of R-IES was eliminated ( data not shown ) , was also detected ( Fig . 6B , WT , closed arrowhead ) . Similar PCR products were also detected from cells transformed with the T+2G construct ( Fig . 6B , T+2G ) and the G+3T construct ( Fig . 6B , G+3T ) . In the cells transformed with the T+2G construct , 1 . 8-kbp product ( s ) became as prominent as the other two products ( Fig . 6B , T+2G , closed arrowhead ) . In the cells transformed with the G+3T construct , an extra 1 . 5-kbp product was detected ( Fig . 6B G+2T , open arrowhead ) . Sequencing analysis revealed that this product had a short deletion in the IES ( data not shown ) . These results indicate that the substitutions at position +2 or +3 of the left boundary change frequencies of occurrence of alternative boundaries in vivo . Next , the 1 . 3-kbp products from cells transformed with the different constructs ( *1 , *2 and *3 in Fig . 6B ) were extracted from gel and cloned , and DNA sequences of 20 clones each were analyzed . All of the sequenced 1 . 3-kbp products from the cells transformed with the WT construct had the same elimination boundary ( Fig . 6C , WT ) , which exactly corresponded with the reported boundary of endogenous R-IES [39] ( Fig . 6C , red arrows ) . In contrast , in all of the sequenced 1 . 3-kbp products from the cells transformed with the T+2G and G+3T constructs , the left elimination boundaries shifted one to several bases downstream ( Fig . 6C , T+2G , G+3T ) . Interestingly , in the T+2G and G+3T constructs , the choice of the right boundary , which had no base substitutions , was also affected ( Fig . 6C , T+2G , G+3T ) . These results may suggest that there is crosstalk between the ends of an IES during DNA elimination . This crosstalk could be established before DNA cleavage , as would be expected if Tpb2p acts as a typical DNA transposase [40] , or during the repair of the excision site as suggested by Saveliev and Cox [22] . Importantly , the observed shifts of the left boundary by the T+2G and G+3T substitutions in the in vivo assay ( Fig . 6C , T+2G , G+3T ) correlate well with , although not identical to , the patterns of shifts of the cleavage position by the corresponding base substitutions in the in vitro Tpb2p endonuclease assay ( Fig . 5B ) . These results , together with the facts that Tpb2p has an endonuclease activity with wide substrate specificity and is necessary for DNA elimination , strongly suggest that Tpb2p is the excisase , the enzyme that cuts IES boundaries in vivo .
We demonstrated that the cysteine-rich domain of Tpb2p directly interacts with the N-terminal tail of histone H3 , and the interaction is significantly enhanced by the heterochromatin-specific histone modifications H3K9me3 and H3K27me3 in vitro ( Fig . 4 ) . Because these histone modifications specifically occur on IESs in the developing new Mac [18] , [19] , Tpb2p can be recruited to IESs through its direct interaction to H3K9me3 and H3K27me3 , and this recruitment may limit the occurrence of Tpb2p-endonuclease cleavage to the near surrounding heterochromatic regions . The interaction between Tpb2p and H3K9/K27me3 may specifically activate the Tpb2p-endonuclease to inhibit Tpb2p to form DNA DSB at non-IES loci . An IES is removed as one piece of DNA [22]–[24] . Therefore , although H3K9me3 and H3K27me3 likely occur throughout an IES [18] , [19] , the endonucleolytic cleavages of Tpb2p must be restricted to the ends of an IES . Because H3K9me3 and H3K27me3 are also bound by one of the most abundant heterochromatin components , Pdd1p [18] , [19] , competition of chromatin-binding sites with Pdd1p may exclude Tpb2p from the body of heterochromatin and only allow Tpb2p to bind the edges of heterochromatin regions . Alternatively , Tpb2p may localize throughout the heterochromatin segment , whereas the heterochromatin structure or some heterochromatin proteins may inhibit the action of Tpb2p endonuclease at the body of heterochromatin . Future research should clarify the spatial localization of Tpb2p on chromatin , which will help with understanding how Tpb2p acts only at the ends of IESs . Regardless of what molecular mechanism limits the nucleolytic action of Tpb2p to the IES ends , the heterochromatin-Tpb2p interaction does not appear to be sufficient to explain the reproducible occurrence of the border of IESs because 1 ) histone modifications are able to determine a chromatin segment only at the level of a size of a nucleosome , whereas most of the boundaries of IESs occur within a few to several nucleotides , and 2 ) there are IESs in Tetrahymena that have sizes similar to a single nucleosome [15] . This study demonstrated that the Tpb2p endonuclease has a relaxed sequence preference for its substrate . For the longer IESs , the combination of the heterochromatin localization and the sequence-biased action likely allows Tpb2p to precisely determine the boundaries of IESs at a sub-nucleosomal level . For the shorter IESs , it is possible that the substrate preference of the Tpb2p endonuclease alone is sufficient to determine the precise boundary . Consistent with this idea , many of the shorter IESs have a common 5′-TTAA-3′ sequence at their boundaries [15] on which DNA DSB is efficiently introduced by the endonuclease Tpb2p in vitro [4] . In addition to the heterochromatin-Tpb2p interaction and sequence-biased action of the Tpb2p endonuclease , the cis-acting sequences adjacent to IESs may also play a role in reproducibly determining the border of IESs . Some IESs have cis-acting sequences located 40–50 nucleotides outside of the boundaries of IESs that are necessary in cis for the precise occurrence of DNA elimination boundaries [26] , [28] , although the mechanism explaining how cis-regulatory elements are involved in DNA elimination is unclear . Because Tpb2p can induce DSB at the IES boundary sequences in oligo DNAs without cis-regulatory elements in vitro ( Fig . 5 ) , cis-regulatory elements are not necessary for Tpb2p to recognize the IES boundary sequences , at least on naked DNA . The cis-acting sequences may set nucleosome positioning , and thus , Tpb2p can be recruited to a fixed chromosomal location . Alternatively , the cis-acting sequences may create a nucleosome-free region where DNA is accessible for Tpb2p . The fact that the endonucleolytically inactive Tpb2p still supports heterochromatin body formation ( Fig . 3E , “CD” ) suggests that the heterochromatin bodies are not a product of DNA elimination but can be formed prior to the initiation of DNA elimination . In contrast , it has been reported that , in the absence of TKU80 , the excision of IESs occurs without the formation of the heterochromatin bodies [10] . One fact we should consider to reconcile these seemingly incompatible observations is that the endonucleolytically inactive Tpb2p mutant was expressed in the conditional TPB2 KO background . In the conditional TPB2 KO locus , TPB2 expression was under control of the MTT1 promoter , which can be activated by addition of the cadmium ion . It is known that the MTT1 promoter is leaky [26] . Therefore , although we used a medium containing minimum metals and we could not detect the wild-type Tpb2p by western blotting in the absence of the cadmium ion ( Fig . 3B ) , it is still possible that undetectable amount of Tpb2p from the conditional TPB2 KO locus induces some DNA elimination even without the cadmium ion in the medium . Nonetheless , because no circularized IES products were detected without inducing the wild-type TPB2 expression in the conditional TPB2 KO cells ( Fig . 3H , WT rescue −Cu2+ ) , leaky expression of TPB2 , if any , causes no or very little IES excision event . Therefore , we conclude that the heterochromatin bodies can be formed without massive DNA elimination . On the other hand , the results reported by Lin et al . ( 2012 ) [10] indicate that massive DNA elimination is not sufficient to induce the formation of the heterochromatin bodies in the absence of TKU80 . Because TKU80 encodes a KU80 homolog , which binds and senses the end of DNA double-strand breaks , the formation of the heterochromatin bodies may be triggered by a signaling cascade down stream of the KU80-mediated DNA double-strand break sensing . A leaky expression of the wild-type Tpb2p may be enough to activate this signaling cascade , and together with the expression of the endonucleolytically inactive Tpb2p , it may induce the heterochromatin body formation . DNA elimination pathways in ciliates are believed to have evolved as a genome defense mechanism against pathogenic invaders , such as transposons [41] . In Tetrahymena , there are several different types of transposons that do not share common boundary sequences , and the previous study demonstrated that Tpb2p is necessary for the elimination of all IESs tested , including the Tlr1 retrotransposon-like element [4] . Therefore , a single molecular mechanism including Tpb2p most likely executes the elimination of all transposons in Tetrahymena . Tpb2p is evolutionarily related to piggyBac transposases . Tpb2p and piggyBac transposases share a common molecular architecture: the endonuclease domain possessing a DDD catalytic core and the zinc-finger-like cysteine-rich domain that is related to the PHD-finger motif . Although Tpb2p induces DSB at a variety of sequences ( Fig . 5 ) , piggyBac transposases specifically cut the 5′-TTAA-3′ sequence [34] . Because insertions of piggyBac transposase in vivo are negatively correlated with the presence of heterochromatin [42] , the cysteine-rich domain of piggyBac transposase , in contrast with the domain of Tpb2p , is unlikely to interact with heterochromatin . Therefore , two important changes must have occurred in a piggyBac transposase during the evolution of DNA elimination in ciliates: loss of the strict sequence specificity for its substrate and gaining the ability to interact with the heterochromatin-specific histone H3 modifications . Although the oligohymenophorean ciliates Paramecium and Tetrahymena use piggyBac transposases for DNA elimination [4] , [35] , the spirotrich ciliate Oxytricha uses Tc1/mariner class transposases for DNA elimination [43] . Therefore , the involvement of domesticated piggyBac transposases in DNA elimination has most likely evolved in the lineage of oligohymenophorean ciliates . Although IESs in Tetrahymena are not flanked by any common sequence , IESs in Paramecium are flanked by a 5′-TA-3′ [44] , [45] . This indicates that the Paramecium piggyBac transposase-like protein Pgm still maintains a certain sequence specificity derived from an ancestral piggyBac transposase that cuts the 5′-TTAA-3′ sequence , whereas the Tetrahymena Tpb2p has evolved to recognize a much greater variety of sequences . Therefore , it appears that the piggyBac transposase has gradually lost its sequence specificity to the substrate during the evolution of oligohymenophorean ciliates . Relaxation of the substrate specificity of Tpb2p might be compensated for by the heterochromatin-binding ability of the cysteine-rich domain of Tpb2p because heterochromatin is specifically formed on IESs prior to DNA elimination [18] , [19] . Heterochromatin formation on IESs is targeted by an RNAi-related mechanism in Tetrahymena [18] . Because transposons in many different eukaryotes are silenced by heterochromatin formation induced by RNAi-related pathways [46] , we speculate that heterochromatin formation on IESs in the Tetrahymena lineage evolved from an ancient transposon-silencing mechanism and co-existed in parallel with the DNA elimination , even before the involvement of a piggyBac transposase in the DNA elimination . Such ancient DNA elimination system might be operated by transposases encoded by eliminated transposons , like we see today in Oxytricha [41] . A piggyBac transposon in the Tetrahymena lineage might have first evolved to target heterochromatin , and then , its transposase might have been domesticated to overtake the roles of transposon-encoded transposases in DNA elimination . Although DNA elimination is widely observed among most ciliates studied , the enzymes used for DNA elimination in different groups of ciliates have distinct properties and even have different transposon origins . Future studies of Tetrahymena Tpb2p and DNA elimination enzymes of other diverse groups of ciliates would help to further understanding of how a transposon has been domesticated for use in regulating a eukaryotic genome .
The wildtype strains B2086 and CU428 were provided by the Tetrahymena Stock Center ( http://Tetrahymena . vet . cornell . edu/ ) . Cells were cultured in 1×SPP medium with 2% proteose peptone at 30°C [47] . TPB2 cKO and TPB2 rescue strains were grown in the metal-depleted medium 1×SPPCT to minimize gene expressions from metal-inducible promoters . To make 1×SPPCT , 1×SPP medium was incubated with 5 g/100 ml of Chelex-100 resin ( BioRad ) with stirring for 2 hr . After filtration to remove the resin , essential trace metals were added ( 100×stock solution: FeCl3*H2O: 1 mg/ml , Co ( NO3 ) 2*6H2O , MnSO4*4H2O: 0 . 16 mg/ml ) . Mating was induced by mixing equal numbers of starved cells . Starvation was achieved in 10 mM Tris buffer ( pH 7 . 5 ) for 12–16 hr at 30°C . The oligo DNAs used in this study are shown in Supplementary Table S1 . A total of 475 bp of the TPB2 5′ flanking region and the first 1132 bp of the TPB2 genomic sequence were amplified from genomic DNA with the primer pairs TN5MT 5′ fw/rv and TN5MT3′ fw/rv , respectively . The neo5 cassette fused to an MTT1 promoter was amplified from a pMNMM3 vector with the primers TN5MTneo fw/rv [48] . After PCR purification , the fragments were combined using overlapping PCR as described previously [49] , resulting in the TPB2 cKO construct , which was directly used for germline transformation of mating Tetrahymena UMPS strains at 3 hr post-mixing . The UMPS strains were created by introducing a uridine monophosphate synthase ( UMPS ) gene from Dictyostelium into the Mac of Tetrahymena . The expression of this gene makes Tetrahymena cells sensitive to 5-fluoroorotic acid ( 5-FOA ) . After conjugation , these strains should lose the UMPS gene because it is only in the Mac , and thus , the progeny are 5-FOA resistant . Biolistic transformation was performed as previously described [50] . After transformation of the TPB2 cKO construct into the UMPS strains , 0 . 1 µg/ml cadmium chloride was added to the cells to induce TPB2 expression from the MTT1 promoter . One paromomycin and 5-FOA resistant clone was obtained , which was confirmed as a heterozygous TPB2 cKO strain . After mating to the WT strain , the heterozygous conditional KO strains were genotyped by PCR and crossed to each other to obtain homozygous conditional KO strains . During all matings , 0 . 1 µg/ml cadmium chloride was added . A blasticidin-resistance cassette was amplified by PCR from pBla1 vector using the primers Bra1_OL_5RACErv and Bra1_OL_fw . In parallel , the MTT2 promoter was amplified from the pDET2 vector with the primers MTT2_OL_fw and MTT2_HA_AvrII_rv . The two PCR constructs were combined with overlapping PCR as described previously [49] . The overlapping PCR product was cloned into pMNMM1 with AvrII and SalI restriction enzymes , resulting in pMBM2M . The TPB2 open reading frame was amplified from genomic DNA ( strain B2086 ) using the primers TPB2ORF_AvrII_fw and TPB2ORF_MluI_rv and subsequently cloned with the enzymes MluI and AvrII in pMBM2M , resulting in pMBM2M-TPB2 . Via site-directed mutagenesis , a catalytically dead version and cysteine-rich mutant were created from the template pMBM2M-TPB2 using the primers TPB2_D297L_fw/rv , TPB2_D379L_fw/rv and TPB2_D495L_fw/rv or TPB2_C618A_fw/rv or C629A_fw/rv , respectively . These rescue vectors were then transformed into the somatic nucleus of two different mating types of conditional TPB2 KO strains by ballistic transformation as described previously [50] . The transformants were selected with blasticidin and phenotypically assorted ( up to 10 mg/ml ) . To assess the phenotype of the TPB2 rescue strains , they were crossed with each other , and the expression of the respective rescue construct was induced by the addition of copper sulfate in a two-step procedure . Equal amounts of CuSO4 were added at 7 and 8 hr post-mixing , for a final concentration of 100 µM . The Mic genomic region containing the R-IES was amplified as two overlapping pieces from the Tetrahymena total genomic DNA using the primers R-leftFW/R-midRV and R-midFW/R-rightRV . The two pieces were combined by overlapping PCR and cloned into the ribosomal vector pD5H8 [51] using the NotI restriction site . pD5H8 containing the R-IES , including the flanking regions , was electroporated into mating wild-type strains . Electroporation was performed as described previously [52] with slight modifications . Mating WT cells in 10 mM Tris ( concentration: 7×10∧5 cells/ml ) were used for transformation at 8 . 5 hr post-mixing . The cells were washed in 10 mM HEPES pH 7 . 5 and resuspended in 120 µl of 10 mM HEPES and then mixed with 120 µl of plasmid DNA in 10 mM HEPES ( 300 ng/µl ) and electroporated ( 220 V , 50 Ω , 50 µF , exponential pulse ) using a BioRad Gene Pulser MXcell . After transformation , the cells were incubated in 1×SPP medium at 30°C overnight without shaking . Transformants were selected in 100 µg/ml paromomycin . Cells were fixed in 3 . 7% formaldehyde and 0 . 5% Triton-X 100 for 30 min at room temperature . The cells were resuspended in 3 . 7% formaldehyde and 3 . 4% sucrose and dried on Superfrost Ultra Plus slides ( Thermo Fisher ) . The samples were blocked for 2 hr with 3% BSA , 10% normal goat serum ( Invitrogen ) and 0 . 1% Tween 20 in PBS , followed by overnight incubation at 4°C in blocking solution containing a 1∶1000 dilution of anti-HA ( Covance ) , 1∶2000 dilution of rabbit anti-Pdd1p ( Abcam ) , 1∶2000 dilution of guinea pig anti-Pdd1p or 1∶2000 dilution of anti-Tpb2p antiserum . The guinea pig anti-Pdd1p antibody was obtained by immunizing a guinea pig with a peptide ( CTAHRSGSRLSQIQSNANQV ) . The anti-Tpb2p antibody was obtained by immunizing a rabbit with N-terminal half ( 1 aa to 556 aa ) of Tpb2p . After washing , the samples were incubated with a 1∶2000 dilution of secondary antibody against mouse or rabbit conjugated to Alexa 488 or Alexa 568 ( Invitrogen ) . The samples were washed , incubated with 10 ng/ml of DAPI ( Sigma ) in PBST , and observed by fluorescent microscopy . The DNA elimination assay using FISH was performed as previously described [53] . The plasmids Tlr1IntB , Tlr1 2 and Tlr1 4C1 [30] were mixed as templates to make probes against the Tlr1 IES . The labeling of the DNA with Cy3 was achieved by nick translation . Cells were fixed at 36 hr post-mixing as described above for immunofluorescence analysis . The excision of the mse2 . 9- and the R-IES elements were examined by detecting circularized IESs by nested PCR using the primers listed in Supplementary Table S1 . Cells were mated at the cell density of 5×105 cells/ml and single pairs were picked into a drop of metal depleted SPP medium ( 1×SPPCT ) . Pairs from the WT mating were picked at 7 h post-mixing . Tpb2 expression from the MTT2 cassette was induced in the rescued strains as described earlier and pairs were picked at around 10 h post mixing . As a control , WT rescue strain without addition of copper was used . At 24 h post-mixing 1×SPP medium was added to the drops to recover normal growth speed of the surviving cells . Sexual progeny formation was confirmed either by their 6-methylpurine resistance in the wild-type cells or by their blasticidin sensitivity in the rescue strains . A codon-optimized TPB2 coding region was amplified by PCR from the previously created vectors pGEX-TPB2 or pGEX-TPB2-CD [4] using the primers T2CPfw/T2ECrv . The PCR products were cut with EcoRI and XhoI and cloned into the EcoRI and SalI sites of the pMalC2X vector to obtain pMAL-TPB2 and pMAL-TPB2-CD . To generate pMAL-TPB2-CRM , C618 to A and C629 to A mutations were introduced into pMAL-TPB2 by site-directed mutagenesis using the DNA oligos T2EC C618A fw/rv and T2EC C629A fw/rv . To create the cysteine rich domain or its point mutant fused to the MBP protein , the primers TPB2_CRD_fw/rv were used with pMAL-TPB2 or pMAL-TPB2-CRM . The plasmids were introduced into the E . coli strain BL21 ( DE3 ) , which was cultivated to an A600 of ∼0 . 8 and then incubated with 0 . 5 mM IPTG for 10 hr at 16°C . The cells were lysed in 500 mM NaCl , 80 mM Tris , pH 8 . 0 , 0 . 2 mM PMSF and 1×complete proteinase inhibitor cocktail ( Roche ) . The lysate was incubated with Amylose resin ( NEB ) at 4°C , washed with 500 mM NaCl and 80 mM Tris pH 8 . 0 and finally eluted with 500 mM NaCl , 80 mM Tris , pH 8 . 0 , and 20 mM maltose , followed by dialysis in 20 mM Tris-HCl , pH 7 . 5 , 100 mM KCl , 4 mM MgCl2 , 4 mM MnSO4 and 10% glycerol . The endonuclease assay was performed as previously described [4] . The oligomeric DNA substrates used for the experiments are listed in Supplementary Table S1 . Peptides corresponding to the N-terminal tail of histone H3 were synthesized and biotin-tagged at their C terminal with a PEG linker ( Supplementary Table S2 ) . In addition , 10 µl ( bed volume ) of streptavidin-coupled Dynabeads ( Invitrogen ) were blocked with 2% BSA in interaction buffer ( 10 mM Tris pH 7 . 5 , 0 . 1 mM ZnSO4 , 0 . 05% NP40 and 250 mM NaCl ) for 1 hr at RT and then incubated with 1 µg of peptide in the blocking solution for 30 min at RT . After washing with interaction buffer and blocking again with 2% BSA 1 µg of MBP-Tpb2p-CRD was added to the beads and incubated overnight at 4°C . After five washing steps with interaction buffer , followed by two washing steps in PBS-T , the beads were resuspended in SDS-PAGE loading buffer , transferred to a fresh tube and boiled for 5 min . The samples were then separated on a 10% SDS-polyacrylamide gel , followed by a western blot . Detection was accomplished using anti-MBP antibody ( NEB ) . The secondary antibody was coupled with an infrared dye , which was visualized with an Odyssey scanner ( LI-COR Biosciences ) . | Transposons are not just threats to genome integrity but also potentially contribute to the evolution of the host . Therefore , host organisms must evolve by balancing the harmfulness and usefulness of transposons . An evolutional product likely created by such a balance is the programmed DNA elimination in the ciliated protist Tetrahymena , in which the transposon-related sequences are eliminated by the domesticated piggyBac transposase Tpb2p . In this study , we demonstrate that the cysteine-rich domain of Tpb2p interacts with the heterochromatin-specific histone modifications and is necessary for DNA elimination . Furthermore , we demonstrate that the endonuclease activity of Tpb2p in vitro and the endonuclease activity that executes DNA elimination in vivo have similar substrate sequence preferences . These results strongly indicate that Tpb2p is the endonuclease that directly catalyzes the excision of IESs and that the boundaries of IESs are at least partially determined by the combination of Tpb2p-heterochromatin interaction and relaxed sequence preference of the endonuclease activity of Tpb2p . These findings provide a molecular basis of the DNA elimination mechanism as well as of the evolution of a domesticated transposase-mediated genome defense against transposons . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | A Domesticated PiggyBac Transposase Interacts with Heterochromatin and Catalyzes Reproducible DNA Elimination in Tetrahymena |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.