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 |
|---|---|---|---|---|---|
Control of enzyme allosteric regulation is required to drive metabolic flux toward desired levels . Although the three-dimensional ( 3D ) structures of many enzyme-ligand complexes are available , it is still difficult to rationally engineer an allosterically regulatable enzyme without decreasing its catalytic activity . Here , we describe an effective strategy to deregulate the allosteric inhibition of enzymes based on the molecular evolution and physicochemical characteristics of allosteric ligand-binding sites . We found that allosteric sites are evolutionarily variable and comprised of more hydrophobic residues than catalytic sites . We applied our findings to design mutations in selected target residues that deregulate the allosteric activity of fructose-1 , 6-bisphosphatase ( FBPase ) . Specifically , charged amino acids at less conserved positions were substituted with hydrophobic or neutral amino acids with similar sizes . The engineered proteins successfully diminished the allosteric inhibition of E . coli FBPase without affecting its catalytic efficiency . We expect that our method will aid the rational design of enzyme allosteric regulation strategies and facilitate the control of metabolic flux .
Living cells coordinate metabolic flux through the allosteric regulation of enzymatic activity [1]–[3] . Allosteric sites provide a molecular platform for allosteric regulators , which are spatially apart from but energetically coupled with catalytic sites [4] , [5] . Binding of allosteric regulators induces an interaction rearrangement of allosteric residues and regulates enzymatic activity [6] . One of the challenges of rational allosteric control is to design mutants that do not impair catalytic function but change binding specificity to allosteric regulators . In the emerging era of engineering enzymatic substrates through active-site remodeling [7]–[9] , deregulating allosteric inhibition is necessary for removing allosteric behavior of template enzymes to obtain commodities . For example , the productivity of ethanol fermentation process is dramatically increased by a mutation in pyruvate dehydrogenase complex which leads to the complex being less sensitive to the allosteric inhibition by NADH [10] . Also , lysine production is increased by deregulation of the allosteric inhibition of aspartokinase [11] . Structures of enzyme-ligand complexes have provided the molecular details of enzymatic regulation; however , the underlying principles of allosteric regulation still need to be uncovered in order to engineer allosterically controllable enzymes . Allosteric and catalytic sites are similar in the sense that they both bind to specific ligands , but have been exposed to different evolutionary constraints . For example , the allosteric sites of mammalian phosphofructokinase were arisen from gene duplication and fusion , and differ across orthologues [12] . Also , the inhibitor binding site of glycogen phosphorylases has changed over the course of evolution from yeast to vertebrates , while the residues of catalytic sites are conserved [13] . Catalytic sites are responsible for substrate binding and conversion , thus mutations in catalytic sites usually demolish the catalytic function of enzymes [14] . Therefore , catalytic sites are usually highly conserved [15] , [16] . On the other hand , allosteric sites provide binding platforms for ligands but are not involved in the catalytic conversion of ligands . Moreover , allosteric regulation mechanisms often varied across species living in different environments at organism level , suggesting that residues in allosteric sites have evolved to adapt to their environments . For instance , adenosine monophosphate ( AMP ) can synergistically inhibit porcine FBPase with fructose 2 , 6-bisphosphate , but this synergism has not been found in E . coli FBPase [17] . Therefore , amino acid residues in allosteric sites may be subject to change to control the allosteric behavior of enzymes in different species . In this study , we systematically analyzed the molecular evolution of enzyme allosteric sites and discovered that allosteric sites have evolved along different evolutionary pathways compared to the highly conserved catalytic sites . We also compared the amino acid compositions of catalytic and allosteric sites , and discovered that allosteric sites have lower numbers of charged residues than do catalytic sites . We then introduced mutations into the allosteric sites of E . coli FBPase to control allosteric regulation without impairing catalytic activity . We confirmed that even at high doses of allosteric inhibitors , mutant E . coli FBPase maintained its catalytic activity . Understanding the evolutionary basis of enzyme allosteric control will provide a method to efficiently engineer tailor-made enzymes to control allosteric regulation .
To understand the differences in the evolutionary properties of catalytic and allosteric sites , we investigated sequence conservation of catalytic and allosteric sites in 56 enzyme structures . Each enzyme has single catalytic and allosteric sites , which are composed of total 212 and 490 residues respectively . The annotations for catalytic and allosteric residues were obtained from databases , experimentally determined enzyme-ligand complex structures , and biochemical studies [18] , [19] . Sequence conservation scores were calculated from multiple sequence alignments of homologous sequences collected from various species ( Fig . 1A , see Materials and Methods for details ) . As shown in Fig . 1B , residues of allosteric sites ( average conservation score = 0 . 58 ) are significantly less conserved than are residues of catalytic sites ( average conservation score = 0 . 94 , P = 1 . 3×10−67 , Table S1A ) , although both sites are significantly more conserved than the rest of the surface ( P = 6 . 2×10−75 ) . To confirm that these differences in evolutionary conservation were true for each protein sequence , we analyzed the average conservation score distributions of allosteric and catalytic sites and confirmed that the distributions were significantly different ( P = 9 . 4×10−18 , Fig . 1C , Table S1B ) . Furthermore , as shown in Fig . 1D , our observation is not biased toward certain enzyme commission classes ( Table S2 ) . This indicates that evolutionary constraints are significantly different on allosteric and catalytic sites . We also tested our analysis on different criteria for selecting 522 catalytic and 782 allosteric site residues based on the 56 enzyme-ligand complex structures to confirm our observation . Catalytic and allosteric residues were selected as those within 6 Å of the respective ligands/substrates in the complex structures and compared with the annotation of catalytic and allosteric sites in databases ( Table S1C , see Materials and Methods for details ) . We found that sequence conservation patterns of catalytic and allosteric residues selected from structures were similar to those of residues selected from the annotated database . Specifically , allosteric sites ( average conservation score = 0 . 58 ) were significantly less conserved than catalytic sites ( average conservation score = 0 . 82 , P = 3 . 2×10−58 , Fig . S1A , Table S1C ) . As before , we confirmed that conservation ratios between catalytic and allosteric sites in enzyme structures were significantly different in individual proteins ( Text S1 ) . Interestingly , allosteric sites were found to have a broader range of conservation scores compared to catalytic sites ( P = 1 . 0×10−61 , F-test , Fig . 1B ) . Allosteric sites are composed of evolutionarily more variable residues than catalytic sites , even though these interact with ligands just as in the case of catalytic sites . We were intrigued by these observations because ligand-binding sites are generally known to be conserved across species [16] , [20] , [21] . We speculated that the naturally variable residues in allosteric sites might constitute potential targets for engineering the allosteric regulation of enzymes without impairing their catalytic activities . Next , we investigated the physicochemical properties of the catalytic and allosteric sites of 56 enzymes by comparing their amino acid compositions . We found that charged residues such as lysine , histidine , glutamic acid , and aspartic acid were highly enriched in catalytic sites ( P = 5 . 5×10−14 , Fig . 2 , Table S3A ) , whereas hydrophobic residues such as proline , tryptophan , leucine , valine , isoleucine , phenylalanine , methionine , and tyrosine were highly enriched in allosteric sites ( P = 2 . 9×10−26 , Fig . 2 , Table S3A ) . We confirmed these observations with a different set of catalytic and allosteric sites derived from enzyme-ligand complex structures ( Fig . S1B , Table S3B ) . These differences in amino acid composition could be due to the different functional roles of the sites . In catalytic sites , ligands are subject to the heterolytic breakage and formation of covalent bonds , but such bond breakage and formation do not occur in allosteric sites . Hydrophilic residues often participate in a hydrogen-bonded network within the active site to facilitate bond breakage or formation [22] , [23] . On the other hand , hydrophobic residues are enriched in allosteric sites to provide a binding pocket for the ligand , with only a small fraction of charged residues that are present to facilitate specific interactions during ligand binding . Thus , we postulated that the allosteric behavior of the enzyme can be changed by mutating the evolutionarily variable residues in allosteric sites . We chose FBPase as a model system in which to test our hypothesis of allosteric site evolution . FBPase is a key metabolic enzyme in the gluconeogenic pathway , and has one catalytic site and two distinct allosteric sites that provide binding platforms for AMP and glucose-6-phosphate ( Glc-6-P , Fig . 3A ) [17] . Activation of the gluconeogenic pathway changes the carbon flux toward the pentose phosphate pathway and increases the level of NADPH that can be utilized to produce various desirable metabolites such as amino acids , fatty acids , and hydrogen [24]–[26] . To activate the gluconeogenic pathway under high glucose concentrations , the allosteric inhibition of FBPase by both AMP and Glc-6-P should be eliminated while maintaining its catalytic activity . We examined the sequence evolution of FBPase and found that the allosteric sites were significantly less conserved than the catalytic site ( P = 1 . 5×10−7 , Fig . 3B ) , but more conserved than rest of the surface residues ( P = 1 . 0×10−5 , Fig . 3B ) . Based on the enzyme-ligand complex structure , we selected residues from the catalytic ( 21 residues ) and allosteric ( 27 residues ) sites that were within 6 Å of the substrates , fructose-1 , 6-bisphosphate ( FBP ) , and the allosteric inhibitors ( AMP and Glc-6-P , Table S4 ) . The rest of the surface area was selected from solvent accessible residues of the enzyme . To alleviate the allosteric regulation of FBPase , we selected residues that have favorable binding interactions with AMP or Glc-6-P ( Fig . 4A and D ) . For the inhibitory effect of AMP with respect to FBP , less conserved and charged residues ( R132 , K104 ) were mutated ( Fig . 4A , Text S2 ) . As shown in Fig . 4B , the single mutants R132I and K104Q showed 15-fold ( P = 4 . 7×10−5 ) and 40-fold ( P = 1 . 9×10−4 ) higher inhibition constants ( Ki ) , respectively , than the wild type ( Table S5 ) . As both residues are directly involved in electrostatic interactions with AMP , each single mutant that diminished the interaction showed only modest effects on the binding of the anionic allosteric effector . However , in the case of the double mutant ( K104Q/R132I ) , the Ki was 140-fold ( P = 3 . 1×10−4 ) higher compared to the wild type , indicating that these mutations had a synergistic effect on disturbing the binding of AMP . We confirmed that mutations in the AMP binding pocket did not affect the regulatory control of Glc-6-P ( Fig . S2B ) . Next , the less conserved residues Y210 and K218 in the Glc-6-P binding pocket were also mutated ( Fig . 4D , Text S3 ) . Both the Y210F ( P = 1 . 7×10−3 ) and K218Q ( P = 4 . 0×10−3 ) mutants had about 17-fold higher Ki values than the wild type ( Fig . 4E ) . In case of the double mutant ( Y210F/K218Q ) , the Ki value was 25-fold higher than the wild type ( P = 9 . 9×10−3 , Fig . 4E ) . Mutations in the Glc-6-P binding pocket did not affect the regulatory properties of AMP to FBPase ( Fig . S2C ) . Notably , the catalytic efficiency ( kcat/Km ) of each mutant was sustained or slightly increased than that of wild-type FBPase , although allosteric regulation by AMP and Glc-6-P was perturbed ( P>0 . 1 , Fig . 4C and F ) . Next , we combined the four mutations ( K104Q/R132I/Y210F/K218Q ) to test whether they were effective in diminishing inhibition by both AMP and Glc-6-P simultaneously . The Ki of the quadruple mutant was 170-fold higher for AMP ( P = 6 . 3×10−4 ) and 25-fold higher for Glc-6-P ( P = 2 . 8×10−4 ) than that of the wild type ( Fig . 5A and Text S4 ) . Additionally , as shown in Fig . 5B , the catalytic efficiency of the quadruple mutant was similar to that of the wild type ( P = 0 . 78 ) . Finally , we investigated the activity profiles of wild-type and the quadruple mutant FBPase by simultaneously changing AMP and Glc-6-P concentrations . At concentrations higher than 100 µM of AMP and 1000 µM of Glc-6-P , the relative activity of wild-type FBPase dramatically decreased to lower than 30% ( Fig . 5C , left panel ) . However , the quadruple mutant FBPase was highly resistant to inhibition even in the presence of high concentrations of both AMP and Glc-6-P . Remarkably , the quadruple mutant FBPase retained >70% relative activity in the presence of 300 µM AMP and 3000 µM Glc-6-P ( Fig . 5C , right panel ) . These results suggest that our mutation strategy succeeded in deregulating the allosteric inhibition of E . coli FBPase without impairing its catalytic efficiency . We found that the mutations of conserved residues led to the loss of FBPase catalytic activity . We mutated the five conserved residues that interact with AMP or Glc-6-P and discovered that all the mutations had the complete loss of catalytic activity ( Table 1 ) . Two conserved residues in AMP binding site that are known to have favorable binding interactions with AMP via hydrogen bond ( T23 and Y105 ) were selected [27] , [28] . We mutated them into valine ( T23V ) and isoleucine ( Y105I ) , respectively , to remove their hydrogen bonds , expecting that those mutations would lead to the loss of allosteric regulation only . In case of Glc-6-P binding sites , three conserved residues , Q225 , E207 , and Y203 , which are corresponding to the formation of ionic and hydrogen bond with phosphate groups of Glc-6-P were selected . When we mutated these residues into isoleucine , leucine , or phenylalanine , all the mutations caused loss of catalytic function . These results indicate that conserved residues in allosteric sites are important for maintaining functional or structural constraint as well as binding allosteric regulator and propagating allosteric signal into catalytic site .
We analyzed the sequence evolution and amino acid compositions of catalytic and allosteric sites of enzymes . The evolution of ligand binding sites has been extensively investigated; however , catalytic and allosteric sites were not separately considered in many analyses of enzyme-ligand interactions [16] . Until recently , allosteric sites were expected to be conserved during the course of evolution , just as catalytic sites of enzymes are highly conserved to maintain their function [14] . However , several lines of evidence suggest that allosteric sites might be less conserved than catalytic sites , since allosteric regulation evidently evolved later than catalytic activity in enzymes along the course of evolution [29] , [30] . Thus , different regulatory mechanisms may exist across species that result in sequence variations in regulatory sites [31] . Further , sequence variations at allosteric sites may be directly linked to the fine-tuning of regulation . Environmental conditions encountered by various species may dictate the necessity to alter the control of metabolic flux . Therefore , residues in allosteric sites would accordingly change from one species to another to allow variations in the specificity of allosteric regulators . Indeed , AMP does not activate yeast glycogen phosphorylase ( GPb ) , but does activate vertebrate GPb , because the yeast enzyme lacks the residues that hydrogen bond with adenine [13] . Moreover , the physicochemical properties of catalytic and allosteric sites may differently affect their amino acid compositions . For instance , catalytic sites preferentially contain charged residues that help to stabilize the intermediate forms of substrates to promote bond formation or breakage [32]; these charged residues tend to be highly conserved across species to sustain the catalytic function of the enzyme . On the other hand , allosteric sites have larger numbers of hydrophobic residues to form binding pockets for allosteric ligands and a few charged residues for specific ligand interactions [33] . Therefore , amino acids in allosteric sites are more tolerant of mutations , because hydrophobic residues can be more easily replaced with similar sized residues compared to charged residues that are involved in specific interactions . We found that allosteric sites have more hydrophobic residues and less charged residues than catalytic sites . It might be possible few polar residues in allosteric sites might be crucial for the ligand specificity . Thus , those residues' changes have greater effect on the binding of polar ligands . For example , the replacement of polar amino acids to hydrophobic amino acids reduced the ligand binding in allosteric sites , which had been shown as increasing inhibition constants ( Ki ) . Although charged states of amino acids may be changed by the biochemical environment of enzymes , we note that sequence conservation analysis only takes into account the sequence variation or conservation of homologues . Environmental variations might already be reflected on the evolutionary constraints on functionally important sites . The AMP binding site of E . coli FBPase is well characterized and several mutations which lead to the loss of allosteric inhibition are known . Therefore , we tried not to repeat the same mutations . In the AMP binding site , K104 and R132 are positively charged amino acids that directly contact with AMP [34] and these residues form hydrogen bonds with the phosphoryl groups of AMP . Thus , the mutations of these residues would disrupt their interactions that stabilize AMP binding . Meanwhile , Y210 and K222 provide large contact surfaces to Glc-6-P interaction ( 62 . 1 Å2 and 51 . 8 Å2 , respectively ) and form hydrogen bonds with the hydroxyl and phosphoryl groups in Glc-6-P [34] . Thus , the mutations of Y210 and K222 to other hydrophobic amino acids should perturb the interaction with Glc-6-P . We searched for previous experimental data evaluating the effects of mutations in allosteric sites and compared with our analysis . We found that most successful allosteric deregulating mutations with no loss of catalytic activity correspond to residues that were less conserved ( average conservation score = 0 . 47 , Table S6 ) , whereas mutations leading to a loss of catalytic activity correspond mostly to residues that were conserved ( average conservation score = 0 . 88 ) . These two group of residues were found to have significantly different conservation scores ( P = 3 . 3×10−5; Mann-Whitney U test ) . However , we found some residues that did not follow the trend . For example , K42 and G191 in porcine FBPase are highly conserved but their mutation did not perturb the catalytic activity , whereas A54 is less conserved and its mutation perturbs the catalytic activity . We provide the frequency of naturally occurring amino acids in Table S7 . Amino acid substitutions that have successfully deregulated the allosteric control of enzyme were less frequently found in multiple sequence alignment ( Average 3% ) . Because frequently occurring amino acids may work in allosteric ligand binding , we selected the less frequently occurring amino acids for the mutation experiments . Although allosteric site residues are more varied than active site residues , we found that allosteric sites are generally more conserved than surface residues . It has been suggested that allosteric sites are localized near protein-protein interfaces which are generally more conserved than surface residues [35] . Furthermore , residues in allosteric sites are also known to serve an important functional role in information propagation from the allosteric site to the active site . Allosteric sites are energetically connected with catalytic sites and coevolved during evolution [11] . These functional roles of allosteric sites might be one of the reasons that allosteric site residues are more conserved than surface residues . Based on our sequence evolution analysis , we propose a novel engineering strategy to rationally modulate enzyme allosteric regulation . First , evolutionarily variable residues may be good targets for mutation because these residues tend to vary during evolution without losing a protein's activity . Mimicking natural evolution minimizes the probability of disrupting the catalytic activity of the enzyme [36] . In this study , we observed that all mutations in conserved residues invariably led to the loss of FBPase catalytic activity ( Table 1 ) , whereas mutations in variable residues generally did not result in loss of catalytic activity ( 0 out of 7 versus 10 out of 14 , P = 3 . 8×10−3; Fisher's exact test ) . Second , amino acids that are likely to give selectivity by forming specific interactions with allosteric regulators via ionic or hydrogen bonds should be considered to mutate . Third , target residues should be substituted with less frequently occurring residues in nature , since frequently occurring residues might still play a role in allosteric regulation . We noted that further experimental validations are needed to establish the generality of our method . This residue selection strategy based on our evolutionary analysis , when combined with current protein engineering approaches , can facilitate the effective control of enzyme allosteric regulation . In addition , redesign of catalytic function would require the removal of the allosteric regulation of template enzymes to get rid of unwanted inhibition . Understanding the evolutionary history of allosteric sites helped us to rationally design mutants for the allosteric control of FBPase . We successfully engineered an allosteric inhibition-resistant E . coli FBPase without impairing its catalytic efficiency . When E . coli is grown in minimal media containing glucose as a carbon source , intracellular concentrations of AMP and Glc-6-P are reported to reach concentrations of 280 µM and 2000 µM , respectively [37] , [38] . Wild-type FBPase is inhibited to less than 20% by these inhibitor concentrations , but the quadruple mutant can maintain its enzyme activity at >80% of these inhibitor concentrations ( Fig . 5C ) . In other words , the quadruple mutant FBPase engineered in this study can potentially enhance gluconeogenesis flux to regenerate reducing power ( NADPH ) through the pentose phosphate pathway even in the presence of elevated intracellular concentrations of AMP and Glc-6-P . Furthermore , our results have implications on the identification of disease-causing mutations . Identification of disease-causing mutations from genome-wide association studies or next-generation sequencing studies currently focus on sequence conservation [39] , which is based on the assumption that functionally important sites are conserved during evolution . Our findings support that mutations in allosteric sites may be responsible for deregulating enzyme allosteric control . Considering that dysfunction in allosteric regulation is highly associated with human disease , such as Alzheimer's disease and diabetes [40]–[42] , our study provides a possible explanation of why mutation of evolutionary variable residues in allosteric sites can cause diseases . In fact , more than 20 disease-causing mutations in the allosteric regulator binding domain of pyruvate kinase are found to be evolutionarily less conserved [43] , [44] . For the first time , we have systematically analyzed the evolutionary properties of enzyme allosteric sites . We found that residues in allosteric sites tend to be less conserved and more hydrophobic compared to those in highly conserved catalytic sites . Furthermore , we successfully deregulated the allosteric inhibition of FBPase without impairing its catalytic efficiency and propose a novel strategy for protein engineering . Recently , computational studies were shown to be quite powerful for identifying residues that deregulate allosteric behavior . For instance , a method combining molecular dynamic simulation and residue coevolution [11] was successfully applied to identify residues that are important for allosteric transition . Integrating such methods might improve the rational design of allosteric enzymes . We also expect that the sequence differences between allosteric and catalytic sites identified in this study will help to detect allosteric sites among potential ligand binding pockets , which currently relies on large-scale screening or serendipity [8] , [35] .
We built two types of catalytic and allosteric site datasets . First , we constructed annotated catalytic and allosteric sites collected from hand-curated databases . The catalytic site atlas ( CSA ) contains experimentally confirmed catalytic sites and the allosteric database ( ASD ) has manually curated allosteric sites with at least three cases of experimental evidences in 3D structure or biochemistry [18] , [19] . We found 56 allosteric proteins that have both catalytic and allosteric site annotations with solved 3D structures . Second , we constructed an alternative catalytic and allosteric site dataset . A hand-curated dataset contains high-quality data but might possess annotation bias or false negative residues . To overcome these limitations , we selected catalytic or allosteric residues from those within 6 Å of the substrates . This dataset contains more permissive residues with no annotation bias . Surface residues were defined as those that were highly solvent-accessible ( >50% of relative solvent accessible surface area ) . NACCESS [45] was used to calculate solvent-accessible area . We obtained the conservation scores and alignment files from the ConSurf server ( http://consurfdb . tau . ac . il/ ) with default options . The server collected homologous sequences for calculating conservation scores from the UniProtKB/SwissProt database [46] using an E-value cutoff of 10−3 with three iterations of PSI-BLAST as previously described [47] . Then , it filtered out sequences with more than 95% identity to the query sequence and those that were shorter than 60% of the query sequence length . Lastly , redundant sequences were removed using CD-HIT [48] . Homologous sequences showed moderate sequence identities ( 36 . 9±12 . 6 , Fig . S3 ) . Conservation scores of residues were calculated from the resulting homologous sequence set using the Rate4Site algorithm [49] . We normalized the conservation scores by using the percentile normalization method to compare conservation scores of different enzymes . Using the same strategy , we collected homologous sequences of E . coli FBPase to calculate conservation scores . Also , we compared our results with other conservation score methods and found that results were similar . The results were statistically significant in the methods that we tested ( from 10−7 to 10−55; Mann-Whitney U test , Table S8 ) . We compared conservation scores between catalytic and allosteric sites in Fig . 1 and 3 based on Mann-Whitney U test which is suitable for assessing whether one of two samples have larger values than the other when data does not follow a normal distribution . We also measured the statistical significance ( P-value ) by Fisher's exact test to compare each amino acids composition in Fig . 2 with the null hypothesis as catalytic and allosteric sites have the same amino acid proportions . In addition , we performed t-test analysis to assess the difference of inhibition constants ( Ki ) and catalytic efficiency ( kcat/Km ) respectively between wild type and mutants in Fig . 4 and 5 . We obtained reproducible results from replicate experiments . A P<0 . 05 was considered to be statistically significant . All the statistical tests were done by scipy in python module . PrimeSTAR HS DNA polymerase and pET101/D-TOPO were purchased from Takara Bio Inc . ( Shiga , Japan ) and Invitrogen ( Carlsbad , CA , USA ) , respectively . Oligonucleotides used for the construction of pET101/D-TOPO-FBPase and variants were synthesized by Bioneer ( Daejeon , Korea ) . All other reagents were obtained from Sigma-Aldrich ( St Louis , MO , USA ) . pET101/D-TOPO-FBPase was constructed by inserting the fbp gene , amplified from E . coli K-12 MG1655 genomic DNA using the FBP_CACC_F ( CACCATGAAAACGTTAGGTGAATTTATTGTCGAAAAG ) and FBP_B ( CGCGTCCGGGAACTCACGGATAAA ) primers , into pET101/D-TOPO following the manufacturer's instructions . This construct was then used as a template for amino acid substitutions by PCR-based site-directed mutagenesis . The PCR mixture for site-directed mutagenesis consisted of 50 ng pET101/D-TOPO-FBPase plasmid , 10 pmol of each primer , 1 . 25 U PrimeSTAR HS DNA polymerase , 250 µM of each dNTP , and 10 µl 5× buffer supplied by Takara Bio Inc . H2O was added to bring the final volume to 50 µl . PCR was carried out on an Applied Biosystems 2720 Thermal Cycler ( Applied Biosystems , Foster City , CA , USA ) under the following conditions: 98°C for 30 s , 12 cycles ( point mutations ) or 18 cycles ( multiple nucleotide changes ) of 98°C for 10 s , the appropriate primer Tm-dependent annealing temperature for 15 s , and 68°C for 7 min , followed by a final extension at 68°C for 10 min . After thermocycling , the original template DNA was eliminated by treating with DpnI at 37°C for 1 h , and PCR products were isolated using a QIAquick PCR Purification Kit ( Qiagen GmbH , Hilden , Germany ) . The E . coli Mach1™-T1R strain ( Invitrogen ) was transformed with 50 ng of the PCR product . Purified plasmids were sequenced by Solgent ( Daejeon , Korea ) using an ABI 3730XL capillary DNA Sequencer . The E . coli BL21 ( DE3 ) strain ( Invitrogen ) was transformed with 50 ng of plasmid isolated and purified from the Mach1™-T1R strain , and clones containing the selected pET101/D-TOPO-FBPase and its variants were grown in LB medium containing 50 µg/ml ampicillin . After cultures had reached an OD600 of 0 . 4–0 . 6 , determined using a UV-1700 spectrophotometer ( Shimadzu , Kyoto , Japan ) , protein synthesis was induced by adding isopropyl-beta-D-thiogalactopyranoside ( IPTG ) to a final concentration of 1 mM . The induced cells were harvested by centrifugation at 4 , 000× g for 10 min at 4°C after incubating for an additional 4 h . Cells were resuspended in lysis buffer ( 50 mM HEPES , 300 mM NaCl , 10 mM imidazole , pH 8 . 0 ) containing lysozyme ( Epicentre , Madison , WI , USA ) and protease inhibitor ( Sigma-Aldrich ) and sonicated 30 times for 2 s each time with 8-s pauses in between , at a 20–30% duty cycle . Soluble lysates were separated by centrifugation at 10 , 000× g for 20 min at 4°C . FBPases with the polyhistidine-tag were purified by affinity chromatography using Ni-NTA agarose ( Qiagen ) . Proteins attached to Ni-NTA agarose were washed twice with wash buffer I ( 50 mM HEPES , 300 mM NaCl , 20 mM imidazole , pH 8 . 0 ) and wash buffer II ( 50 mM HEPES , 300 mM NaCl , 40 mM imidazole , pH 8 . 0 ) and eluted with elution buffer ( 50 mM HEPES , 300 mM NaCl , 250 mM imidazole , pH 8 . 0 ) . Eluted proteins were desalted using a PD-10 desalting column ( GE Healthcare , Piscataway , NJ , USA ) . The concentrations of purified proteins were determined by Bio-Rad Protein Assay Kit ( Bio-rad , Hercules , CA , USA ) using BSA as a standard . The activities of the purified FBPases were measured by monitoring the evolution of Pi from FBP using the Malachite Green Phosphate Assay Kit ( BioAssay Systems , Hayward , CA , USA ) . Assay mixtures ( 50 mM HEPES , pH 7 . 5 , 0 . 1 mM EDTA , 0 . 17 µg enzyme , and varying amounts of FBP , AMP , and Glc-6-P in a total volume of 80 µl ) were incubated in microtiter plates ( Oy Growth Curves Ab , Helsinki , Finland ) at 25°C for 1 h prior to the initiation of the reaction by the addition of varying amounts of MgCl2 . The addition of 20 µl of Working Reagent ( BioAssay Systems ) quenched the reactions , and the plates were incubated at 25°C for 30 min to allow color development . Absorbance ( λ = 600 nm ) was measured on Bioscreen C MBR ( Oy Growth Curves Ab ) . A linear standard curve relating A600 to [Pi] was drawn according to the manufacturer's guidelines ( R2>0 . 99 ) . Kinetic parameters such as kcat , Km ( FBP , Mg2+ ) , and the Hill coefficient were determined by fitting initial velocity data to Equation 1 ( R2>0 . 95 ) , ( 1 ) where v is the velocity , S is either the concentration of FBP or Mg2+ , Vmax is the velocity at saturating FBP and Mg2+ , Km is the Michaelis constant for either FBP or Mg2+ , and H is the Hill coefficient . Inhibition data by AMP and Glc-6-P with respect to FBP were fit to the following equation ( Equation 2 ) for nonlinear noncompetitive inhibition ( R2>0 . 95 ) , ( 2 ) where v is the velocity , Vmax is the velocity with saturating FBP and Mg2+ with no inhibitor present , Km is the Michaelis constant for FBP , S is the concentration of FBP , Ki is the inhibition constant , I is the concentration of either AMP or Glc-6-P , and H is the Hill coefficient for FBP derived from Equation 1 . All data fitting were performed using GraphPad Prism 5 . 0 software ( GraphPad Software Inc . , La Jolla , CA , USA ) . All kinetic parameters measured in this study are listed in Table S5 . | Design of allosterically regulatable enzyme is essential to develop a highly efficient metabolite production . However , mutations on allosteric ligand binding sites often disrupt the catalytic activity of enzyme . To aid the design process of allosterically controllable enzymes , we develop an effective computational strategy to deregulate the allosteric inhibition of enzymes based on sequence evolution analysis of allosteric ligand-binding sites . We analyzed the molecular evolution and amino acid composition of catalytic and allosteric sites of enzymes , and discovered that allosteric sites are evolutionarily variable and comprised of more hydrophobic residues than catalytic sites . We then experimentally tested our strategy of enzyme allosteric regulation and found that the designed mutations effectively deregulated allosteric inhibition of FBPase . We believe that our method will aid the rational design of enzyme allosteric regulation and help to facilitate control of metabolic flux . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bioengineering",
"biotechnology",
"biological",
"systems",
"engineering",
"genomics",
"evolutionary",
"biology",
"evolutionary",
"processes",
"synthetic",
"biology",
"biology",
"computational",
"biology",
"evolutionary",
"selection",
"structural",
"genomics",
"engineering"
] | 2012 | Rational Engineering of Enzyme Allosteric Regulation through Sequence Evolution Analysis |
New microbial genomes are sequenced at a high pace , allowing insight into the genetics of not only cultured microbes , but a wide range of metagenomic collections such as the human microbiome . To understand the deluge of genomic data we face , computational approaches for gene functional annotation are invaluable . We introduce a novel model for computational annotation that refines two established concepts: annotation based on homology and annotation based on phyletic profiling . The phyletic profiling-based model that includes both inferred orthologs and paralogs—homologs separated by a speciation and a duplication event , respectively—provides more annotations at the same average Precision than the model that includes only inferred orthologs . For experimental validation , we selected 38 poorly annotated Escherichia coli genes for which the model assigned one of three GO terms with high confidence: involvement in DNA repair , protein translation , or cell wall synthesis . Results of antibiotic stress survival assays on E . coli knockout mutants showed high agreement with our model's estimates of accuracy: out of 38 predictions obtained at the reported Precision of 60% , we confirmed 25 predictions , indicating that our confidence estimates can be used to make informed decisions on experimental validation . Our work will contribute to making experimental validation of computational predictions more approachable , both in cost and time . Our predictions for 998 prokaryotic genomes include ∼400000 specific annotations with the estimated Precision of 90% , ∼19000 of which are highly specific—e . g . “penicillin binding , ” “tRNA aminoacylation for protein translation , ” or “pathogenesis”—and are freely available at http://gorbi . irb . hr/ .
Many computational methods for functional annotation of genes are based on a search for sequences with common evolutionary descent—homologs . One possible encoding of homology is the use of phyletic profiles: each row in the phyletic profile represents one gene , and the columns represent the presence or absence of homologs in sequenced genomes [1] , [2] . There are two main ways in which phyletic profiles can be used for annotation of gene function . Both of them involve propagating the annotation label . First , one could create phyletic profiles and propagate the annotation label within the profile—from genes with known function to their homologs included in the profile . This is homology-based annotation , and many schemes for doing so are possible [3] . Second , one could propagate labels between the profiles by finding similar profiles: assuming that genes that are inherited together tend to work together , one transfers annotation from a better-studied group of homologs to a profile that is similar but contains genes that are not as well studied . Again , this can be done in many ways . For example , phyletic profiles can be grouped by similarity using a variety of distance measures [e . g . ] , [ 1 , 4] possibly involving a machine learning framework [e . g . ] , [ 4]–[6] . Rows in the phyletic profile can stand for genes or groups of genes [e . g . ] , [ 1] , [7 , 8]; functional annotation can be assigned using a range of vocabularies , e . g . , UniProt controlled vocabulary of keywords [9] , Enzyme Commission numbers [10] , or arguably the most widespread vocabulary , the Gene Ontology [11] . In addition , one could employ some hybrid between the first two approaches , e . g . , when the evidence in favour of within-profile label propagation is used to improve the confidence of between-profile propagation and vice versa . Refinements of homology-based annotation include making a distinction between two types of homologous relationships: orthologs—sequences derived from the same gene in the last common ancestor , and paralogs—sequences derived from a duplication event [12] . Because orthologous pairs are expected to keep the same function [13]–[15] and paralogous pairs are expected to diverge in function [16] , the canonical approach to functional annotation relies on transfer of function between orthologs . However , the latest evidence suggests that , relative to pairs of paralogs , the conservation of function between pairs of orthologs is not as strong as the standard model would imply [17] . Our goal was to create a functional annotation model that learns to associate gene function with specific patterns in phyletic profiles—the presence and absence of different types of homologs in different organisms . To create the phyletic profiles , we combined ortholog cliques—fully interconnected groups of orthologs—with both additional orthologs and additional paralogs . We found that , instead of reducing the predictive accuracy , paralogs provide valuable information: compared to the model that includes only orthologs , the model that includes both orthologs and paralogs gave more predictions at the same average correctness . In addition , we performed experimental assays in the model organism Escherichia coli , showing that the annotation model provides a realistic assessment of confidence for the predicted annotations: a growth phenotype screen on E . coli knockout mutants indicated an overall Precision of 66%—out of 38 tested genes , we confirmed predictions for 25—agreeing with the expected Precision of 60% . We predict Gene Ontology annotations at various levels of specificity for about 1 . 3 million poorly annotated genes in 998 prokaryotes at a stringent threshold of 90% Precision: about 19000 of those are highly specific functions . In addition to these , we provide many more predictions at less stringent cut-offs in a Web resource GORBI ( http://gorbi . irb . hr/ ) .
In one OMA clique , inferred orthologous relations connect each protein to every other protein , so it is not surprising that they group proteins with mostly the same function ( see Materials and methods ) . However , OMA cliques leave out many of the existing orthologous relations . Consequently , phyletic profiles of OMA cliques are incomplete , leading to poor performance in our classification model: many of the true orthologous relations are missing , and the model can successfully annotate using only the most general GO terms ( Figure 2 , model a; Figure S1 in Text S1: panels A , B , and C ) . If we compensate for the missing orthologous relations in OMA cliques by adding all inferred one-to-one , one-to-many , many-to-one , and many-to-many orthologs left out when constructing the cliques , the model improves: the mean AUPRC is 0 . 8 ( Figure 2 , model b; Figure S1 in Text S1: panels D , E , and F ) . We also tested whether adding paralogs to phyletic profiles of OMA cliques improves the mean AUPRC: it does , showing that the functional information we obtain from paralogs is far from useless ( Figure 2 , model c; Figure S1 in Text S1: panels G , H , and I ) . Still , the mean AUPRC is 0 . 65—lower than if we enrich phyletic profiles with orthologous relationships . However , it is the combined information from orthologs and paralogs that provides us with the best model for functional annotation ( Figure 2 , model d; Figure S1 in Text S1: panels J , K , and L ) : the mean AUPRC increases to 0 . 85 . In the above experiments , adding only orthologs improved AUPRC more than adding only paralogs ( Figure 2 , models b and c , respectively ) . To test whether accounting for the ortholog/paralog distinction would further increase AUPRC , we encoded the phyletic profiles with three levels: presence of an OMA clique member or another ortholog ( 2 ) , presence of a paralog ( 1 ) , or absence of any of these ( 0 ) . We found a small gain in AUPRC resulting from the ortholog/paralog distinction ( Figure S2 in Text S1 , panel B ) , but we also found that increasing the number of levels in the dataset from the original two to the above-described three decreases the AUPRC ( Figure S2 in Text S1 , panel A ) . Taken together , accounting for the ortholog/paralog distinction did not yield an overall gain in AUPRC in the current machine learning setup ( Figure S2 in Text S1 , panel C ) , so we chose the binary model as our principal result . In this binary model that includes both orthologs and paralogs , most of the general GO terms have high AUPRC . More specific GO terms span a wide range of AUPRC ( Figure S3 in Text S1 ) . Nevertheless , both specific and general GO terms benefit from the inclusion of orthologs and paralogs . Specific GO terms such as “lysine biosynthetic process via diaminopimelate , ” “organic acid∶sodium symporter activity , ” or “bacterial-type flagellum basal body” are used in less than 0 . 1% of annotations in the 07-02-2012 UniProt-GOA release ( their Information Content is higher than 10 ) : the mean AUPRC of this subset of specific GO terms rises from 0 . 78 in the model that includes orthologs ( Figure 2 , model b ) to 0 . 83 in the model that includes both orthologs and paralogs ( Figure 2 , model d ) . For the general GO terms such as “protein transport , ” “kinase activity , ” or “plasma membrane , ” each used in more than 3% of annotations in the 07-02-2012 UniProt-GOA release ( their Information Content is lower than 5 ) , the corresponding change in AUPRC is from 0 . 80 to 0 . 88 . Intuitively , phyletic profiling should perform best for the Biological Process ( BP ) GO terms: proteins with similar profiles are expected to be involved in the same BP but not necessarily to have the same Molecular Function ( MF ) . For example , one kinase and one glucosidase may be involved in the same process of sporulation despite having different MF . As a result , one would expect phyletic profiling to be more appropriate for assigning BP GO terms than MF GO terms . Here , we report high predictive accuracy for all three ontologies ( Figure 2 , model d ) . In fact , among the best performing and most specific predictions are those for Molecular Function ( MF ) GO terms “acyl-CoA dehydrogenase activity , ” “transposase activity , ” “organic acid∶sodium symporter activity” and its parent term “solute∶sodium symporter activity , ” “penicillin binding” and its parent term “drug binding” ( Figure S3 in Text S1 ) . The AUPRC provides us with a view on predictive accuracy that values both the comprehensiveness of predicted annotations for a given GO term ( Recall ) and their correctness ( Precision ) across the entire range of model stringency cut-offs . To further explore the relationship between Precision and Recall at specific levels of model stringency , we chose three cut-offs—0 . 1 ( permissive cut-off ) , 0 . 3 ( medium cut-off ) , and 0 . 7 ( stringent cut-off ) , for the two best models—the model including orthologs ( corresponding to AUPRC values in Figure 2 , model b ) and the model including both orthologs and paralogs ( corresponding to AUPRC values in Figure 2 , model d ) . The combination of data and cut-offs resulted in six plots ( Figure 3 ) . For any of the cut-offs , the mean Precision for GO terms between the two models is similar ( Figure 3 , horizontal lines between A and D; B and E; C and F ) . However , there is a difference for Recall , in particular for the more stringent cut-offs ( Figure 3 , vertical lines between B and E; C and F ) . It is this increase in Recall that increases AUPRC , as we observed before ( Figure 2 ) . For example , at the most stringent cut-off the model including only orthologs predicts annotations with 414 GO terms for at least 50 poorly characterized genes in the 998 genomes , while the model including both orthologs and paralogs predicts annotations with 573 GO terms for at least 50 genes . To each unnannotated OMA clique , the model assigned a cut-off that indicates the probability of being annotated with a GO term . To have an interpretable measure of confidence for each prediction , we transformed this cut-off to the corresponding Precision ( see the Materials and methods section ) . We then propagated the function of each OMA clique to the member genes and obtained the functional annotations , along with the estimates of Precision for each annotation . As a consequence of the increased Recall , the model that includes both orthologs and paralogs provides more annotations at the same Precision ( Figure 4A ) . The increased Recall allows us to assign specific annotations at a very stringent threshold of 90% Precision . For example , we predict new annotations for E . coli , both using the most general , as well as many specific GO terms ( Figure 4B , Figures S4 and S5 in Text S1 ) . In the comparisons above , we obtained the best predictive performance for the model based on cliques of orthologs enhanced by both inferred orthologs and paralogs . We evaluated the ability of each model to generalize to novel data , the poorly characterized genes , with an out-of-bag method for testing predictive performance: we measured accuracy on a random subset of the annotated phyletic profiles left out when inferring the functional annotation model . This method was shown to give unbiased estimates of predictive performance [20] . To validate how realistic are these out-of-bag performance estimates , we chose annotations for 38 genes in Escherichia coli K-12 having at least 60% expected Precision , for three GO terms that were straightforward to investigate experimentally using readily available antibiotics: “DNA damage response , ” “translation , ” and “peptidoglycan-based cell wall biogenesis . ” The 38 E . coli strains , each with the deletion of one among the 38 selected genes , were grown in the presence of antibiotics that target the above Biological Processes: nalidixic acid ( causes severe DNA damage , including double-strand breaks ) , kasugamycine ( inhibitor of translation initiation ) , and ampicillin ( inhibitor of cell wall synthesis ) ( Figure 5 ) . To each of the 38 genes the model assigned a Precision , as explained in the Materials and methods section . For example , Precision associated with the E . coli gene yfgI for “DNA damage response” was 62%; for “translation” and “peptidoglycan-based cell wall biogenesis” it was lower than 1% . We would therefore predict this gene to be involved in “DNA damage response” with a probability of being a false positive of 38% ( 100-62 ) . For the other two GO terms the probability of being a false positive would be over 99%: the annotation model inferred that these are unlikely functions for this gene . To experimentally evaluate a predicted annotation , we used the E . coli mutant deleted in the gene whose function we predicted . We compared the mutant to the E . coli wild type when grown in the presence of the antibiotic that inhibits the predicted function . If the gene is indeed involved in the predicted function , the survival of the mutant is expected to be lower than the survival of the wild type . For example , we predicted “DNA damage response” for the E . coli yfgI gene , so the corresponding mutant and the wild type were grown in the presence of DNA-damaging nalidixic acid; we expect the mutant to have lower survival than the wild type because its DNA repair capabilities are diminished . We might predict a particular function , such as “DNA damage response , ” for an important gene that is indirectly involved in many biological processes . Deleting such a gene may lower survival non-specifically and thus appear to validate our prediction . To control for this , we also grew each mutant in the presence of the two additional antibiotics . For the above example of the yfgI gene , if our prediction is correct , the survival of the mutant should not be different from the survival of the wild type when grown on kasugamycin or ampicillin . Therefore , we considered a prediction confirmed only if both of the following criteria were satisfied: 1 ) the survival of a mutant was lower than 25% of the wild type when grown with the addition of the antibiotic inhibiting the process predicted by our model , and 2 ) the survival of the mutant was higher than 50% of the wild type when grown on the other two antibiotics . For example , we predicted “DNA damage response” for the E . coli yfgI gene: when grown on DNA-damaging nalidixic acid , the yfgI mutant had 7% survival of the wild type , but when grown on kasugamycine or ampicilin , the survival was much higher: 98% and 74% of the wild type , respectively ( Table S1 ) . We therefore consider the prediction for the involvement of the yfgI gene in DNA repair processes confirmed: the yfgI mutant is sensitive to a DNA-damaging agent , while exhibiting wild type-like resistance to other stresses . With these criteria , 25 out of 38 genes had confirmed predictions , which is equivalent to the experimental Precision of 66% ( Figure 5 ) . Since the selected genes had an expected Precision of 60% , the experiments show that the estimates of accuracy provided by the model are realistic . In fact , 14 of the 38 tested genes have Precision ≥85% . For these genes , the experiments have shown 11 out of 14 ( 79% ) to be correct , approximately matching the expected precision of 85% . Consequently , these estimates can be used to guide decisions when prioritizing genes for an in-depth experimental investigation of function in the wet lab . In addition to the systematic experimental verification of novel annotations in three GO categories as described above , here we highlight individual predictions for which we found supporting evidence in the publicly available databases . This information was not available to the classifiers at the time when the models were constructed . The following examples are for E . coli K12 , as this is by far the best-studied model prokaryote [21] . We predict genes hypC and hybG to have “nickel cation binding . ” These genes had no GO terms assigned in the 07-02-2012 UniProt-GOA release ( http://www . uniprot . org/uniprot/P0AAM3 and http://www . uniprot . org/uniprot/P0AAM7 ) , and we therefore considered them unannotated . In the meantime , hypC was annotated with “metal ion binding” using experimental evidence: this is a parent GO term of our prediction . Moreover , when examining the literature , we found evidence that these two genes are involved in the biosynthesis of the [NiFe] cluster [22] . Another prediction is for gltL: we predicted it is annotated with “ATP-binding cassette ( ABC ) transporter complex . ” In line with our predictions , PortEco , a portal that includes information from 14 different E . coli data resources [23] , labels the gene as “ATP-binding component of ABC superfamily . ” Note that more general electronic GO annotations were available for this gene , e . g . “ATP binding , ” “ATPase activity , ” and “ATP catabolic process” ( http://www . uniprot . org/uniprot/P0AAG3 ) . A similar prediction of a more detailed function is for ybgI , where we predict GO terms from both BP and MF ontologies . This gene is known to be a conserved metal-binding protein [24] , having an electronic GO annotation “metal ion binding”; we predict it is annotated with the BP GO term “Mo-molybdopterin cofactor metabolic process . ” Based on the structure of the protein , Ladner et al . hypothesize this protein is a hydrolase-oxidase enzyme [24]; we predict this protein is annotated with the MF GO term “hydrolase activity , acting on acid anhydrides , in phosphorus-containing anhydrides . ” Because we showed our functional annotation model gives realistic estimates of predictive accuracy , we made our predictions freely available in a Web site GORBI ( http://gorbi . irb . hr/ ) . Our predictions can be queried either using GO accession number , NCBI taxonomy ID , or gene/protein ID ( Figure 6 ) . For example , one can focus on more general or more specific GO terms , depending on their position in the “Gene Ontology DAG” ( Figure 6 , insert A ) . In addition , an experimenter can tune the search parameters to get a small number of high-confidence candidates , or a larger number of candidates that potentially have more false predictions , depending on the availability of annotations for the desired function and the available resources for experimental validation .
The intuition of phyletic profiling is that corresponding genes gained and lost together in different genomes are likely to share function: they could be involved in the same metabolic pathway , which is therefore incomplete without all the members in a genome [1] . Additionally , even if the two genes are parts of separate pathways and don't strictly require each other for function , they could both share a role beneficial for survival in a particular environment [25] . The standard way of finding corresponding genes in different genomes is via sequence homology: in addition to inferring function via homology , a phyletic profile allows to infer function based on the presence or absence of the corresponding genes in a range of organisms . In functional annotation , we often differentiate between two subtypes of homologs , orthologs and paralogs [e . g . , 26] . According to the standard model of genome evolution , paralogs—pairs of genes diverged through a duplication event—could obtain a new function [e . g . , 27] . Conversely , orthologs are pairs of genes diverged through a speciation event and should be more likely to retain function; they are therefore expected to be more informative in functional annotation [15] . However , the exact quantification of the functional divergence in a pair of orthologs and a pair of paralogs is not fully resolved . It was observed that the search for homologs using the best bidirectional hit approach , without explicitly distinguishing orthologs from paralogs , produces a higher level of functional compactness via Gene Ontology ( GO ) terms [28] than is present in the ortholog databases Homologene [29] and OMA [30] . In addition , Studer and Robinson-Rechavi list scenarios where the standard model—predicting that paralogs diverge in function more than orthologs—is invalid; for example , cases where paralogs share function , and orthologs do not [31] . A recent large-scale study further challenged the veracity of the standard model: the authors compared mouse and human ortholog and paralog pairs and surprisingly found that paralogs tend to conserve function more than orthologs [32] . This finding caused a stir in the community—demonstrating the relevance of the topic—but was subsequently challenged in two publications [17] , [33] . Nevertheless , a recent systematic survey showed that the divergence in function between paralogs is not as strong as the standard model would imply [17] . In addition , we know that homologs—orthologs and paralogs combined—are useful in functional annotation , especially when their sequence similarity is above the “twilight zone” [34] . Further , orthologs and paralogs share a common ancestor: paralogs , as well as orthologs , could carry functional information useful for annotation . In line with the recent research [17] , our results show that the standard model , when viewed in the functional annotation context , tends to draw too strong of a line between orthologs and paralogs . When we enriched clique-only annotation models with additional orthologs and additional paralogs , we obtained a model that outperformed both the model that was enriched only with orthologs and the model that was enriched with refined homologs at different evolutionary distances ( Figure S6 in Text S1 ) . The improvement is most notable in the number of new annotations we were able to assign: while keeping the Precision at the same high level , our best model increases Recall ( Figure 3 ) , and consequently gives us more predictions at the same level of correctness ( Figure 4A ) . Even so , our results do not contradict the standard model in two major points: 1 ) cliques of orthologs—groups where all genes are connected with orthologous relations—are indeed functionally very similar ( Figure 7 ) , and 2 ) our results support the current widespread annotation efforts that use homology: even when we disregard the orthology/paralogy relationships to enhance cliques , we obtain high predictive accuracy ( Figure S6 in Text S1 ) . The OMA algorithm infers paralogs ( i . e . , non-orthologs ) among genes linked as the best bidirectional hits in the respective genomes: a witness to non-orthology breaks the link between two genes [35] . Because only one witness is enough to break the orthologous relationship , the OMA algorithm produces orthologous groups with high specificity [36] . As a trade-off , the set of inferred paralogs might contain pairs whose orthologous link was erroneously broken; the probability for this to happen increases with the addition of new genomes ( A . Altenhoff , personal communication ) . Therefore , our set of paralogs might contain orthologs that were misclassified as paralogs . Even so , when we enriched clique-only annotation models with the inferred paralogs , predictive accuracy increased less than when we enriched clique-only annotation models with the missing orthologs ( Figure 2 , models c and b , respectively ) . We obtained these results despite enriching with a larger number of paralogous pairs than orthologous pairs: it is not the number of added pairs that improves the predictive accuracy , but the genome they are located in . An important output of any computational annotation model is an estimate of confidence for the annotations: it can subsequently be used to guide decisions about experimental validation . In fact , one project that provides a framework for the exchange of information between the computational and experimental communities is COMBREX [37] , [38] . To meaningfully contribute to growing resources such as COMBREX , we wanted to evaluate whether our annotation model provides realistic estimates of confidence for the individual annotations . Probing growth profiles of knockout E . coli mutants with sub-lethal concentrations of antibiotics is an established method of functional annotation [39] , [40] . Here , we experimentally validate whether a gene is involved in the predicted Biological Process by growing the respective knockout E . coli mutant in a medium containing the antibiotic that targets the Biological Process we predicted . The experimental results support the estimates of Precision obtained from a cross-validation procedure , serving as a proof of principle that our phyletic profiling-based model is useful when searching for novel functions of poorly annotated genes in a microbiology lab . Our annotation model assigns GO terms from across the GO hierarchy , for both general and specific terms . Overall , more general terms tend to have a higher cross-validation Area Under the Precision-Recall Curve ( AUPRC ) ( Figure S3 in Text S1 ) and consequently the annotations assigned with these terms are more likely to be correct . The AUPRC scores such as the one we use serve as a test of the internal consistency of the model . On the one hand , the model captures the similarities of the phyletic profiles of the OMA cliques ( and the enriched OMA cliques ) of orthologs; on the other hand , the model captures the GO terms assigned to the OMA cliques of orthologs . For a given GO term , the AUPRC scores will be low if the phyletic profiles' features cannot be used to transfer the function between the profiles . Thus , we make no prior assumptions whether a GO term at a certain level of specificity can be transferred across profiles , but rather infer this from the data itself in a systematic manner . An experimenter can focus on more general or more specific GO terms depending on the trade-off of reported Precision and the cost/time required of experiments; when using the GORBI Web site , GO terms can be selected depending on their position in the Gene Ontology hierarchy ( Figure 6 , insert A ) . To facilitate the use of the generated computational annotations , we provide them in a Web site GORBI ( http://gorbi . irb . hr/ ) where each prediction is accompanied by the annotation model's estimate of confidence . We contribute a solution for computational annotation of genes that utilizes a distinction between two types of homologs—orthologs and paralogs—to yield an innovative annotation model: phyletic profiles derived from cliques of orthologs enriched with both orthologs and paralogs have shown the best predictive accuracy . Our results are in line with related recent research: while it is generally accepted that pairs of orthologs have a lower rate of functional divergence , the divergence in paralogous pairs is not as strong as the standard model would imply [17] . In addition , we performed validation experiments in knockout mutants of E . coli , showing that our annotation model reports realistic measures of predictive performance . The agreement with the experimental results implies that our functional annotations—and the corresponding confidence estimates—can be used to narrow the search space for potential function candidates and thereby help to bridge the widening gap between the sequenced and characterized proteins . For successful annotation of newly sequenced proteins , we need contributions from both the computational community—a large number of credible annotations—and the experimental community—validating the most interesting computational annotations . In turn , the validated findings from the wet-lab can be fed into the computational annotation pipelines , helping to propel a virtuous circle that increases the number of experimentally annotated genes . Our research aims to contribute to the understanding of the deluge of data we face , whether from complete microbial genomes for which we provide annotation predictions ( http://gorbi . irb . hr/ ) , or from the metagenomics projects , in particular the emerging human microbiomes , to which we can apply our annotation model .
We downloaded all annotation data from the FTP site of the UniProt-GOA database [41] . We used the Gene Ontology ( GO ) vocabulary for functional annotation [28] . We included all annotations assigned by a curator ( evidence codes EXP , IMP , IGI , IPI , IEP , IDA , ISS , RCA , IC , NAS , TAS ) , and from the non-curated annotations ( evidence code IEA ) , we included those inferred from UniProtKB keywords , UniProt Subcellular Location terms , Enzyme Commission numbers , and InterPro ( reference codes GO_REF:0000004 , GO_REF:0000023 , GO_REF:0000003 , and GO_REF:0000002 , respectively ) . Despite not being curated , a recent report showed these electronic annotations are of high quality , in particular for the only analysed Prokaryote , E . coli [42] and Figure S7 in Text S1 We express the specificity ( opposite of generality ) of a GO term GOi with respect to its Information Content:where freq ( GOi ) is the frequency of GOi among all annotations for the twelve Reference genomes [43] . The OMA algorithm is a graph-based method of orthology inference [35] . Roth et al . provide full details of the algorithm , and we summarize the main points relevant to our work . The algorithm starts with an all-against-all sequence alignment: proteins from two species are connected if they are best bidirectional hits , within a confidence interval , in the compared species . The connections between a pair of proteins are broken when one of them is the best bidirectional hit with one of the proteins in a connected pair in some third species , and the other is the best bidirectional hit with the second protein in the same pair; the broken pairs are inferred paralogs . The remaining connections are inferred orthologs . Finally , OMA cliques of orthologs are sub-graphs where all proteins are connected by orthologous relationships ( Figure 1 ) . In this work , we only use OMA cliques that group at least 10 members . The OMA algorithm is available as a stand-alone version; the results can also be browsed on the OMA web site [44] . Because one essential component of our work is annotating OMA cliques of orthologs based on the proteins they contain , we first checked whether OMA cliques contain proteins with the same function . First , unannotated OMA members were labelled with the GO terms of annotated OMA members at four thresholds: if 30 , 50 , 70 , or 90% of OMA members have the respective function . To assign these labels , we used only annotations available in the 16-01-2008 UniProt-GOA release . Next , we checked the annotations in the more recent 17-10-2011 UniProt-GOA release . For each unannotated protein , we consider the labelled function to be confirmed if the protein holds the respective annotation in the more recent release; we consider the labelled function to be rejected if the protein holds the same annotation alongside a ‘NOT’ qualifier ( explicit rejection ) or a new annotation that is not the propagated one ( implicit rejection ) . To make a more robust measure , we summarize the confirmed and rejected annotations for each GO term . We named this measure ‘Coherence of a GO term . ’ More formally , where is the set of confirmed annotations associated with term GOi and is the set of rejected annotations associated with term GOi . We account for the definition of the GO: the assignment of any GO annotation assumes the assignment of all the GO parent terms . This is a conservative estimate of Coherence: we consider as rejected an annotation that might not have been added to the database yet . Annotations are continuously being added to UniProt-GOA database , and the annotation update interval for a gene can be as long as 12 years [42] . To compensate for this bias , we evaluated coherence on a three-year interval , as most genes in E . coli are updated within that time frame . For each GO term , the functional coherence depends on the imposed annotation threshold ( Figure 7 ) : when a larger fraction of OMA members in 2008 supported the GO annotation , we found more newly annotated proteins that support this propagated GO annotation in 2011 . The drawback of the increasing threshold was a smaller number of GO terms that can be used in annotation and consequently a smaller number of annotated OMA groups used in training the annotation model . We chose the threshold of 50% as a compromise: for most GO terms , the newly annotated proteins are in accordance with the propagated functions—fraction of correctly predicted newly arrived annotations is greater than 0 . 9—and we are left with enough specific GO terms for functional annotation ( Figure 7 , panel C ) : 422 GO terms from the Biological Process ontology , 48 GO terms from the Cellular Component ontology , and 264 GO terms from the Molecular Function ontology . We use the 50% threshold throughout this work . The phyletic profile of an OMA clique of orthologs is encoded as a vector of binary values . The vector's length is 998 items—the number of prokaryotic genomes included in our work . Each position in the vector indicates the presence or absence of an OMA clique member in the respective genome . There are 64052 annotated and unannotated OMA phyletic profiles in our dataset ( Figure 1 ) . We enriched the phyletic profiles , first by connecting the missing orthologs to OMA clique members ( Figure 1 , full lines ) , and second by connecting the paralogs ( Figure 1 , dashed lines ) to OMA clique members . Orthologs include one-to-one orthologs , one-to-many orthologs , many-to-one orthologs , and many-to-many orthologs . The Clus-HMC [7] algorithm builds decision trees for hierarchical multi-label classification ( HMC ) . In contrast to ordinary classification trees [45] , which can be used for single-label annotation , Clus-HMC is able to deal with multiple , hierarchically organized class labels , such as terms from the Gene Ontology . It builds decision trees for HMC by extending the standard decision tree learning algorithm: It splits the training data into subsets based on attribute values , in order to minimize the weighted sum of variances for all class labels within the subsets resulting after the split [7] . In this weighted sum , a parameter w0 can be used to place more weight on either the more specific , or the more general GO terms . The default value of this parameter is 0 . 75 , which places more weight on more general terms . Changing the default value of the w0 parameter to place more weight on the specific terms will favour them , possibly trading off the accuracy of the more general terms for a gain in accuracy of the more specific terms . To test for possible gain , we experimented with different values of the w0 parameter to place higher weight on either the more general GO terms ( default value , w0 = 0 . 75; w0 = 0 . 5 ) or on the more specific GO terms ( w0 = 1/0 . 75 = 1 . 33; w0 = 1 . 75; w0 = 2 . 0; w0 = 3 . 0 ) . Clus-HMC-Ens proved to be robust to the value of the w0 parameter ( Figure S8 in Text S1 ) : we did not record a significant change in the AUPRC values ( p-value was not lower than 0 . 28 in the five tested values of the w0 parameter , Wilcoxon signed-rank test ) , and we therefore used the default value in all our computational experiments . In addition , the hierarchy of class labels introduces dependencies between the classes: Clus-HMC is aware of the hierarchical relationships between the multiple labels and uses this information to improve predictive performance . The Clus-HMC algorithm was extended to an ensemble setting ( Clus-HMC-Ens ) [18] , where a forest of decision trees for HMC is learned: The predictions of the individual trees are combined to obtain the overall prediction of the ensemble . Clus-HMC-Ens implements , among other methods , the Random Forest ( RF ) ensemble [19] approach , where the individual trees are constructed by using a randomized version of Clus-HMC . Each tree is constructed from a different sample of the training dataset: The bagging ( Bootstrap aggregating ) methodology of resampling the dataset [46] is used to construct the different samples . One bootstrap sample consists of the same number of examples as the original dataset , but they are randomly drawn with replacement; consequently a bootstrap sample contains about two thirds of unique examples . A model—Clus-HMC decision tree—is produced from each of the bootstrap samples . When estimating the classification error , out-of-bag estimates are calculated . The examples that were omitted from the bootstrap sample—one third of the original dataset—are used in calculating Precision , Recall , and Area Under the Precision-Recall Curve ( AUPRC ) . The estimates are based on the random sample , and the measures are therefore unbiased . To check whether adding paralogs improves the functional annotation model regardless of the machine learning algorithm used , we inferred functional annotation models with the standard approach used in phyletic profiling: transfer of function via pairwise distance measures between phyletic profiles , as implemented in a kNN classifier ( Figure S9 in Text S1 ) . The conclusions presented above do not change: the model that includes both orthologs and paralogs outperforms the model that includes only orthologs . Because Clus-HMC-Ens outperforms kNN in computational efficiency and predictive accuracy , we used Clus-HMC-Ens throughout this work . We compare models of functional annotation using Precision-Recall curves: in the Precision-Recall space , Recall is on the x-axis , and Precision is on the y-axis . Traditionally , Precision and Recall are defined for binary classification: an instance either has or does not have the label; in our case , each OMA clique either has or does not have a GO annotation . Precision and Recall are defined for each GO term:where is the number of correctly predicted true annotations ( “True Positives” ) , is the number of incorrectly predicted true annotations ( “False Positives” ) , and is the number of missed true annotations ( “False Negatives” ) . Precision stands for the fraction of correctly predicted examples out of all the predictions , and Recall stands for the fraction of correctly predicted examples out of all known to be true . Here , we are dealing with a multi-class problem: each OMA clique can be annotated with multiple GO terms . The classifier we are using is adapted for such a problem and assigns a probability that each OMA group is assigned each of the GO terms . By varying a cut-off for the probability form 1 . 0 to 0 . 0 , we are relaxing the stringency of the predictions: an increasing number of OMA groups are assigned an increasing number of GO terms . Fixing this cut-off at the three values and calculating Precision and Recall for each GO term created visualizations in Figure 3 . The probabilities allow us to have a ranking of GO annotation predictions for OMA cliques and proteins therein . In addition to the ranking , we wanted to have an intuition for the number of candidates we need to experimentally examine in order to get confirmed annotations . Therefore , we translated the probabilities to Precision for each GO term . Similarly as above , we varied the cut-off for the probability , and calculated the corresponding Precision for each GO term at each probability cut-off: out of all the OMA clique annotations that pass the threshold , we counted the number of true positives , and the number of false positives . To compare models in Figure 2 , we used a single measure of performance that combines Precision and Recall: Area Under the Precision-Recall Curve ( AUPRC ) . To calculate AUPRC , we first varied the probability cut-off from 1 . 0 to 0 . 0 and obtained the Precision-Recall curve . We then calculated the area that is enclosed between the Recall axis and the curve . The closer AUPRC is to 1 . 0 , the better the model . All deletion mutants used herein were derived from wild-type sequenced Escherichia coli MG1655 by P1 transduction . P1 phage was grown on a series of Keio collection deletion mutants listed in Table S1 . Successfully transduced mutants were selected on LB plates supplemented with kanamycine . Bacteria were grown in LB broth at 37°C , to the exponential phase ( OD600 = 0 . 2–0 . 3 ) . Viable cell counts were estimated by plating serial dilutions on LB plates , as well as LB plates supplemented with 400 ug/mL kasugamycine ( inhibitor of translation initiation ) , 4 ug/mL nalidixic acid ( causes severe DNA damage , including double strand breaks ) , and 3 ug/mL ampicillin ( inhibitor of cell wall synthesis ) . Plates were incubated overnight at 37°C . The concentrations of antibiotics used in this study were selected as the concentrations that lead to ∼10% survival of the wild type E . coli . | While both the number and the diversity of sequenced prokaryotic genomes grow rapidly , the number of specific assignments of gene functions in the databases remains low and skewed toward the model prokaryote Escherichia coli . To aid in understanding the full set of newly sequenced genes , we created a computational model for assignment of function to prokaryotic genomes . The result is an innovative framework for orthology and paralogy-aware phyletic profiling that provides a large number of computational annotations with high predictive accuracy in train/test evaluations . Our predictions include annotations for 1 . 3 million genes with the estimated Precision of 90%; these , and many more predictions for 998 prokaryotic genomes are freely available at http://gorbi . irb . hr/ . More importantly , we show a proof of principle that our functional annotation model can be used to generate new biological hypotheses: we performed experiments on 38 E . coli knockout mutants and showed that our annotation model provides realistic estimates of predictive accuracy . With this , our work will contribute to making experimental validation of computational predictions more approachable , both in cost and time . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computer",
"applications",
"computer",
"science",
"web-based",
"applications",
"genomics",
"functional",
"genomics",
"bio-ontologies",
"genetics",
"biology",
"computational",
"biology",
"comparative",
"genomics",
"genetics",
"and",
"genomics",
"gene",
"function"
] | 2013 | Phyletic Profiling with Cliques of Orthologs Is Enhanced by Signatures of Paralogy Relationships |
Canine hip dysplasia is a common , non-congenital , complex and hereditary disorder . It can inflict severe pain via secondary osteoarthritis and lead to euthanasia . An analogous disorder exists in humans . The genetic background of hip dysplasia in both species has remained ambiguous despite rigorous studies . We aimed to investigate the genetic causes of this disorder in one of the high-risk breeds , the German Shepherd . We performed genetic analyses with carefully phenotyped case-control cohorts comprising 525 German Shepherds . In our genome-wide association studies we identified four suggestive loci on chromosomes 1 and 9 . Targeted resequencing of the two loci on chromosome 9 from 24 affected and 24 control German Shepherds revealed deletions of variable sizes in a putative enhancer element of the NOG gene . NOG encodes for noggin , a well-described bone morphogenetic protein inhibitor affecting multiple developmental processes , including joint development . The deletion was associated with the healthy controls and mildly dysplastic dogs suggesting a protective role against canine hip dysplasia . Two enhancer variants displayed a decreased activity in a dual luciferase reporter assay . Our study identifies novel loci and candidate genes for canine hip dysplasia , with potential regulatory variants in the NOG gene . Further research is warranted to elucidate how the identified variants affect the expression of noggin in canine hips , and what the potential effects of the other identified loci are .
Many hereditary disorders appear in both humans and dogs with gene variants from common ancestral genes affecting the disease [1] . Canine hip dysplasia ( CHD ) is a non-congenital disease , causing skeletal abnormalities in growing dogs , the first signs appearing at the age of three to four months [2] . CHD is defined as the laxity of the joint , resulting in instability and subluxation [2 , 3] . The femoral head is not completely covered by the acetabulum , which then leads to increased force over smaller surface area due to incongruence of the coxofemoral joint [2 , 3] . This in turn causes microfractures in the acetabulum and the femoral head [2 , 3] , detrition of the articular cartilage , inflammation of the synovial membrane and secondary osteoarthritis ( OA ) [3] . CHD can be painful up to a point where it poses a serious welfare problem . Hip dysplasia appears in humans in divergent forms from infancy to adolescence and to adulthood [4] . Of these , the adolescent hip dysplasia is clinically and developmentally closest to CHD [5] . Official scoring of CHD varies by country: in Finland it is scored categorically from A ( normal ) to E ( severely affected hip joints with possible OA ) using ventrodorsal extension radiographs as standardized by the Finnish Kennel Club under the Fédération Cynologique Internationale ( FCI ) [6] . The scoring is based on the following features: level of congruence between the femoral head and the acetabulum , degree of subluxation of the joint , Norberg angle , shape of the femoral head and neck , shape and depth of the acetabulum , and signs of secondary OA [6 , 7] . Other traits also add into hip morphology , but they are not routinely evaluated [8] . In the end , only the categorical score for each hip joint is recorded and made available for later use . The severity of CHD depends on environmental and genetic factors [2 , 3 , 9–23] . CHD prevalence varies also by breeds and breed groups [5] , but is common for example in German Shepherds with a reported 37% prevalence in Finland between 2000–2017 ( 6413/16433 ) [24] . The heritability ( h2 ) estimates for the hip score are generally moderate across breeds ( 0 . 20–0 . 38 ) [12–14] , although in German Shepherds the h2 estimates have varied from 0 . 1 to 0 . 6 as summarized in [18] . In a Finnish German Shepherd population the estimates were 0 . 31–0 . 35 [25] . In earlier studies the h2 estimates for the traits determining the hip score ( e . g . Norberg angle , secondary OA , articular congruence and dorsolateral subluxation ) have varied considerably from low h2 = 0 . 10 to high h2 = 0 . 73 [10 , 11 , 14 , 15 , 19] . Some evidence of major genes affecting CHD exist from studies utilizing variance estimates and Bayesian modeling [16 , 18 , 26 , 27] . Also , a recent study investigating quantivative trait locus ( QTL ) associations with CHD revealed a large effect locus on chromosome 28 , which had a 6° additive effect on Norberg angle values in Golden Retrievers and Labrador Retrievers [28] . Fels and Distl suggested QTLs on chromosomes 19 , 24 , 26 and 34 , which associated with CHD in German Shepherds [29] . In addition , a number of other small effect loci and potential candidate genes have been found [15 , 21 , 23] . The reported QTLs and candidate genes are inconsistent between studies , however , as both Sánchez-Molano et al . [15] and Zhu et al . [22] have discussed . Ultimately , different study populations and methods affect the results substantially , which must be recognized when reviewing the data . FBN2 encoding for fibrillin 2 is to our knowledge the only gene in which a mutation has been demonstrated to be associated with CHD using gene expression analysis [30] . However , Lavrijsen et al . [23] found no evidence of association between the region harboring FBN2 and CHD , but they suggested this discrepancy may be caused by differences in Dutch and U . S . Labrador Retriever populations . We have carried out genome-wide association studies ( GWAS ) in case-control cohorts , revealing a total of four associated loci on two chromosomes . Subsequent sequencing of the underlying region on one chromosome identified putative regulatory variants of NOG , which downregulated a reporter gene expression in vitro , and were associated with the healthy and the mildly dysplastic phenotypes .
To map CHD loci , we originally performed GWAS using the Illumina 173K chip with 160 controls ( with A/A hip scores , left/right hip ) and 132 cases ( D/D , D/E , E/D or E/E ) , which revealed a suggestive association on canine chromosome 9 ( Fig 1 ) . Subsequently , we genotyped 233 more individuals and analyzed the data from all 525 dogs . However , one control was dropped from the subsequent meta-analyses due to a missing genotyping batch covariate . Therefore , the first meta-analysis cohort comprised 277 controls and 132 cases ( with D/D , D/E , E/D or E/E hip scores ) ( Fig 2 ) , and the second , less stringent meta-analysis included the same controls and 247 cases ( with C/C , C/D , D/C , D/D , D/E , E/D or E/E hip scores; dogs with C/C are later on referred as mild cases ) ( Fig 3 ) . None of the top SNPs from the different analyses reached genome-wide significance . The data were analyzed using two different statistical methods within the R package "GenABEL": FASTA and QTSCORE . In the original association analysis and the first meta-analysis QTSCORE was used with environmental residual to be comparable with FASTA . In the second meta-analysis QTSCORE was used with standard genomic control [31] . We decided to use FASTA because it is effective in handling highly stratified data of related individuals , while it usually is not as conservative as QTSCORE . However , for the second meta-analysis we used only QTSCORE with standard genomic control , because FASTA ended up losing more power ( S1 Fig ) . QTSCORE was used to obtain the permuted P-values for genome-wide significance in all analyses , because FASTA cannot be used for this task . Inflation factor lambda , describing possible inflation of test statistics due to population stratification , was estimated as 1 . 02 for FASTA , deflation of 0 . 76 for QTSCORE for the original analysis and 1 . 01 and 0 . 72 for the first meta-analysis respectively . For the second meta-analysis lambda was 1 . 43 . When lambda was < 1 , the deflation was corrected with reverse genomic control [32] . The P-values after inflation and deflation factor corrections , after permutation tests with QTSCORE , and genotypic and allelic odds ratios ( OR ) with the respective 95% confidence intervals ( CI ) for all of the top SNPs are shown in Table 1 ( original GWAS ) , Table 2 ( first meta-analysis ) , and Table 3 ( second meta-analysis ) . The association on chromosome 9 in the original genome-wide association analysis was 14 times stronger , and in the first meta-analysis over 45 times stronger than in any other loci in the genome ( Fig 1 and Fig 2 ) . In the second meta-analysis the association on chromosome 1 was over seven times stronger than what was observed on chromosome 9 , and over 14 times stronger than for any other loci in the genome ( Fig 3 ) . The SNPs on chromosome 9 represent two separate loci; there was moderate to high linkage disequilibrium ( LD ) between the markers within each locus ( r2 = 0 . 64–0 . 99 ) , but limited LD between the loci ( r2 = 0 . 34–0 . 56 ) ( S1 Table ) . Clumping analysis , a tool to estimate the number of independently associated loci , corroborated this by revealing two loci within the targeted region on chromosome 9 ( S2 Fig ) . OR for the SNPs within the first locus ( Chr9: 30993502–32382532 ) were all less than one ( Table 1 ) , and for the markers within the second locus ( Chr9: 36543581–36579921 ) OR varied between 2 . 25 and 4 . 90 ( Table 1 ) . The results from the first meta-analysis were similar to the original GWAS ( Table 2 ) . We observed a 2 . 6x difference in the smallest P-values , indicating a small gain in power , when we compared the SNPs with the strongest association in both analyses: 7 . 8x10-6 in the original GWAS ( BICF2P742007 ) and 3 . 0x10-6 in the first meta-analysis ( BICF2G630837405 ) . The LD-structure of the top SNPs from this analysis ( S1 Table ) also resembled the respective values seen in the original GWAS , indicating two independent loci . The odds ratios implied a protective role for the top SNPs within the first locus as in the original GWAS . In the second locus , BICF2G630837240 locating at Chr9:36579921 had OR of 2 . 00–4 . 78 ( Table 2 ) , which again was comparable to the values observed in the original GWAS . The other SNPs in this second locus , locating within 36837067–36886621 , were only observed in this meta-analysis , although they exhibited the strongest association to the disorder ( Table 2 ) . These three SNPs were in moderate LD ( 0 . 68–0 . 71 ) with BICF2G630837240 , but had odds ratios < 1 ( Table 2 ) . As in the original GWAS and the first meta-analysis with the stringent phenotype definition , some of the top SNPs demonstrated notable LD between each other ( r2 = 0 . 80–1 . 00; S1 Table ) in the second meta-analysis using the relaxed phenotype . The clumping procedure indicated two loci on chromosome 1 and two loci on chromosome 9 ( S3 Fig ) . The first associated locus on chromosome 1 spans over ~1 . 1 Mb region , and all of the ORs imply a protective association ( OR = 0 . 26–0 . 62; Table 3 ) . The second locus is ~367 kb long and ORs for all the three SNPs are also < 1 ( Table 3 ) . The two loci on chromosome 9 corresponded to the same regions that have been defined before . However , the association of the second locus on chromosome 9 reached a P-value of 1 . 1x10-4 at best ( BICF2G630837209 ) and was therefore not included in Table 3 . Table 4 summarizes the four loci and their representative SNPs that displayed the strongest association with the disorder over the analyses . All of the top SNPs in the ~1 . 1 Mb locus on chromosome 1 were in high LD with each other ( S1 Table ) . One of the associated SNPs is intronic to NADPH Oxidase 3 ( NOX3 ) ( BICF2S23248027 ) , and the rest of the SNPs lie in an intergenic region between NOX3 and AT-Rich Interaction Domain 1B ( ARID1B ) ( BICF2P468585 , BICF2P1037296 , and BICF2P357728 ) ( Table 4 ) . The SNPs closest to ARID1B were BICF2P357728 and BICF2P1037296 , located 91234 bp and 101945 bp upstream . The second locus on chromosome 1 included many genes ( Table 4 ) . The corresponding SNPs were also in high LD with each other ( r2 = 0 . 84–0 . 96 , S1 Table ) and were located either within MAM Domain Containing 2 ( MAMDC2 ) ( BICF2S23329752 and BICF2S23660342 ) or within Protein Prenyltransferase Alpha Subunit Repeat Containing 1 ( PTAR1 ) ( BICF2P1129598 ) . The top SNPs on chromosome 9 span over a region of ~ 5 . 3 Mb . BICF2S23027935 locates to an intron of Ankyrin-Repeat and Fibronectin Type III Domain Containing 1 ( ANKFN1 ) and 153764 bp upstream from NOG , a known bone morphogenetic protein ( BMP ) inhibitor ( Table 4 ) . BICF2P742007 is intergenic and lies close to NOG ( 66839 bp upstream ) . The SNP representing the second locus on chromosome 9 ( BICF2G630837240 ) is situated between the genes for Mitochondrial rRNA Methyltransferase 1 ( MRM1 ) and LIM Homeobox 1 ( LHX1 ) ( Table 4 ) . We compared the genotype frequencies of the top markers ( Table 5 ) between the different phenotype groups to assess if any of the loci behave differently in these comparisons . Here we did the following comparisons: controls ( hip scores A/A ) to mild cases ( hip scores C/C ) and mild cases to moderate-to-severe cases ( hip scores C/D , D/C or worse ) , because other comparisons were covered in the in the meta-analyses described above . The allele and genotype frequencies of healthy dogs and mildly dysplastic dogs did not differ significantly on either chromosome ( Table 5 ) . Interestingly , the allele and genotype frequencies between mild cases and moderate-to-severe cases were significantly different on chromosome 9 , but not on chromosome 1 ( Table 5 ) . Odds ratios for all the significant comparisons indicated a protective association for the locus near NOG on chromosome 9 ( Table 5: BICF2S23027935 and BICF2P742007 ) . The second locus on chromosome 9 , near LHX1 , increased the odds for hip dysplasia about 2- to 5-fold in all the significant comparisons ( Table 5: BICF2G630837240 ) . The three most common genotypes for the locus with the strongest association on chromosome 9 in the original GWAS ( BICF2S23027935 and BICF2P742007 , Table 1 ) were the homozygous GA ( both SNPs represent the non-risk allele ) and AG ( both SNPs represent the risk allele ) , and the heterozygous RR ( IUPAC coding ) ( Table 1: BICF2S23027935 and BICF2P742007 ) . The AG genotype differentiates moderate-to-severe from mild hip dysplasia ( Table 6 ) . Given the GA and AG genotypes , the odds ratio [95% confidence interval] for mild cases and controls is 0 . 90 [0 . 41–2 . 07] , for mild , moderate or severe cases and controls 0 . 27 [0 . 16–0 . 45] , and for moderate-to-severe and mild cases 0 . 16 [0 . 09–0 . 29] . To identify additional CHD-associated variants , we resequenced a 7 Mb genomic target ( corresponding to bases 30620001–37620000 on chromosome 9 ) in 24 control and 24 affected dogs representing the most common homozygous SNP genotype combinations ( SNPs BICF2S23027935 , BICF2P742007 , BICF2G630834826 , BICF2G630837209 , BICF2G630837240 , BICF2G630837405 and BICF2P272135; see Tables 1 and 2 and methods ) . We used a custom pipeline to systematically screen the target area in comparison with the CanFam3 . 1 annotation . Altogether we found 30197 unique variants in 21140 positions ( S2 Table ) and classified them based on the associated gene , the predicted functional effect of the variant and the phenotype of the individual . The study design , however , does not permit the direct assessment of the association between the phenotype and the genotype as the case and control animals were selected based on both the phenotype and an opposing homozygous combined genotype of seven top SNPs on chromosome 9 ( see Tables 1 and 2 and the methods ) . We therefore screened for variants that segregated completely or nearly completely with either group of dogs . The difference in counts of each variant between the two groups were determined ( S2 Table ) . 61 variants remained after excluding those displaying an absolute difference of 21 or smaller , those with an intergenic or intronic location , and those leading to a synonymous mutation ( S3 Table ) . An upstream variant was found in the immediate vicinity of SMG8 encoding for a nonsense mediated mRNA decay factor . However , there is a gap in the dog/human alignment at this position . A DNase I hypersensitivity site and an H3K27Ac signal reside in the human genomic regions homologous to those harboring upstream variants of benzodiazepine receptor ( peripheral ) associated protein 1 ( BZRAP1 ) and ring finger protein 43 ( RNF43 ) . A strong H3K27Ac signal was also seen in an area on the human chromosome 17 corresponding to a variant upstream of RAD51 paralog C ( RAD51C ) . Several potential splicing mutations were seen in RNF43 and testis expressed gene 14 ( TEX14 ) . Two variants downstream of ANKFN1 were in close proximity ( within 10 kb ) to a SNP with a statistically significant association with the phenotype ( S3 Table ) . See S4 Table , S5 Table , S4 Fig and methods for the calculation of the association between the 217 target area SNPs and the phenotype , N = 426 . We also assessed how the variants targeted specific genes so that the target genes segregated with the risk or non-risk SNP genotypes leading to various functional effects ( Table 7 ) . A missense variant rs852180586 in Apoptosis antagonizing transcription factor ( AATF ) is 26 . 5 kb away of BICF2G630837405 ( Tables 2 and 7 and S3 Table ) but is predicted to be tolerated . The two variants downstream ANKFN1 are close to BICF2G630834765 ( Table 7 , S3 Table ) but there is no evidence for a functional effect . The potentially deleterious missense variant rs24532262 in the myeloperoxidase ( MPO ) gene was connected with the cases . The mutation does not target the mature protein , however . A missense variant in PCTP corresponds to a location near the carboxy-terminus of the phosphatidylcholine transfer protein that is outside of any known protein domain . All other coding variants were predicted to be tolerated . A potentially regulatory variant was discovered 364 bp upstream of RAD51C . This variant was found in 22 cases ( 16 homozygous , 6 heterozygous animals ) and none of the controls . The corresponding site on human chromosome 17 displays a strong H3K27Ac ( acetylation of lysine 27 of the histone H3 protein ) chromatin mark signal . No other evidence for gene regulatory variants was found . Intronic variants close to splice regions were discovered in the gene for ring finger protein 43 ( RFN43 ) and TEX14 . Twenty-eight SNPs in the resequenced target region associated with the phenotype ( S5 Table ) . These SNPs concentrated on two loci ( S5 Fig , S5 Table ) corresponding to those found in the LD-analysis ( S1 Table , S3 Fig ) . As the sequencing depth was variable we combined the reads from cases and controls to separate pools . Visual inspection of two pools of sequences revealed a deletion variant at chr9:31453837–31453860 in the first locus ( vertical black line in S5 Fig ) . This 24-bp deletion variant at resided within an AGG-triplet repeat region in close proximity to the NOG gene in eight control dogs . In addition , one dog had a 27 bp deletion . NOG and its upstream sequence are conserved across species [33] ( S6 Fig ) . The corresponding region on the human chromosome 17 is placed within a putative gene regulatory element upstream of NOG gene with binding sites for several transcription factors [34] ( Fig 4 ) . Additionally , there are H3K4Me1 ( mono-methylation of lysine 4 of the H3 histone protein ) and H3K4Me3 ( tri-methylation of lysine 4 of the H3 histone protein ) histone mark peaks linked to this region; H3K4Me1 marks associate with enhancers and H3K4Me3 with active promoters [34] ( Fig 4 ) . The corresponding region on mouse chromosome 11 overlaps with binding sites for Suz12 ( OREG1916695 ) and Mtf2 ( OREG1828914 ) transcription factor binding sites . The presence of the deletion variant upstream of NOG ( Fig 4 ) was directly assessed by PCR in the whole population of dogs in this study . The fragment sizes were analyzed by gel electrophoresis for all the samples . PCR failed to give a product in nine samples and the product was ambiguous in one sample . The deletion genotype counts and frequencies for each phenotype category for the remaining 516 dogs are shown in Table 8 . The deletion genotype correlated with the genotypes of the SNPs BICF2P742007 and BICF2S23027935 in all three phenotype categories . Spearman’s rank correlation coefficient rho [with 95% confidence intervals] was 0 . 64 [0 . 56–0 . 71] for controls , 0 . 69 [0 . 56–0 . 79] for mild cases and 0 . 67 [0 . 58–0 . 75] for moderate-to-severe cases ( S6 Table ) . The significance of the protective effects of the NOG deletion and GA SNP genotype in various subsets of dogs was next investigated using logistic regression . The odds ratios for the corresponding generalized linear model ( GLM ) coefficient estimates are presented in Table 9 . In contrast to the effect of the GA SNP genotype , the protective effect of the NOG deletion was most significant between the mild and moderate-to-severe cases ( Table 9 ) . To assess the effect of the deletion , we compared the full ( with the NOG deletion ) and the reduced ( without the NOG deletion ) GLM models using chi-squared test . There was a statistically significant difference between the full and reduced models on controls and moderate-to-severe cases ( P < 0 . 05 ) and on mild and moderate-to-severe cases ( P < 0 . 001 , Table 10 ) . Finally , the receiver operating characteristic curve was used to assess the discrimination potential between the full and reduced models . We argue , based on the results from this comparison , that the full rather than reduced model better discriminates the controls and moderate-to-severe cases ( P < 0 . 01 ) , as well as the mild and moderate-to-severe cases ( P < 0 . 001 , Table 10 ) . We investigated the effects of the deletions on the expression of a luciferase reporter gene in vitro . We designed three constructs ( S7 Table ) , where the longest construct A with 14 AGG-triplet repeats corresponds to resequencing data from Dog2 ( S7 Fig and Fig 4 ) . Construct B had a deletion of eight AGG-triplets , and construct C had a deletion of seven AGG-triplets when compared to construct A ( Fig 4 ) . The sequences corresponding to the constructs A and C were common in the cohort , whereas we recovered the sequence corresponding to variant B in only one individual . The constructs were cloned to a plasmid containing a luciferase reporter under the control of a minimal promoter . We used two experimental setups with HEK293 human embryonic kidney and U-2 OS human osteosarcoma cell lines: HEK293 cells with 50 ng and U-2 OS cells with 10 ng DNA . The results are expressed as mean ± SD of four technical replicates from three independent experiments for each cell line and treatment . The firefly luminescence control was used to normalize the NanoLuc luminescence values . In the first experimental setup with HEK293 cells , construct A had significantly higher luminescence compared to both B and C constructs ( Fig 5 ) . Again , in the second setup , with U-2 OS cells , the A construct demonstrated significantly higher luminescence than construct C . All comparisons between the control plasmid and construct luminescence levels were significant . The current canine reference genome ( CanFam3 . 1 ) shows a gap within NOG ( Fig 4 ) . We closed the gap by PCR and sequencing ( S8 Fig ) . The sequence overlapped the 5’ NOG sequence from Beagle [35] and corresponded with the variant with three copies of hexanucleotide insertion ( GenBank accession AB544074 . 1 ) . The closure of the gap in the reference genome permits the accurate positioning of the upstream deletion locus in relation with the coding sequence . The sequence corresponding to the 434 bp long gap in the reference is very similar to the corresponding human sequence ( S9 Fig ) . Alignment introduced six gaps ( 34 , 25 , 3 , 2 , 1 and 1 bp ) . The nucleotide-level identity was a remarkable 76% ( 330/434 ) suggesting conserved function . The NOG promoter ( ENSR00000096009 ) overlaps with the corresponding region in human and spans from 17:56592202 to 17:56594999 with the core promoter at 56592600–56594601 . Scanning the human core collection at the JASPAR2018 database [36] with the alignment of dog and human sequences at S9 Fig , uncovered 135 matrix IDs and altogether 1368 putative binding sites for them ( S8 Table ) . Bonferroni-adjusted P-values were calculated for all sites . Matrix IDs with adjusted P-value less than 0 . 05 for any site are shown in S9 Table . The corresponding transcription factors are histone 4 transcription factor ( HINFP ) , two E2F-related factors and three AP-2 family members .
Canine and human hip dysplasia represent one of the most complex and prevalent problems in veterinary and medical sciences . Our GWAS uncovered four novel protective and risk loci on chromosomes 1 and 9 . The loci on chromosome 9 differentiated the mild from the moderate-to-severe phenotypes . Alleles upstream of NOG displayed differential enhancer activity in vitro . Three additional candidate genes on chromosomes 1 and 9 were revealed: NOX3 , ARID1B and RNF43 . We identified putative regulatory variants of NOG that encodes for a well-known BMP inhibitor , noggin . Noggin is essential for the growth and patterning of the neural tube after neural induction [37 , 38] , but it is also required for embryonic chondrogenesis , osteogenesis and joint formation [38–40] . Joint formation in Nog knockout mice is defective and most joints are missing from the limbs [40] . In humans , NOG missense mutations segregate with proximal symphalangism and multiple synostosis syndrome , both of which are skeletal dysplasias resulting from decreased noggin activity [39 , 41] . Nog is also widely expressed in adult mouse joint cartilage and down-regulated in surgically induced arthritis [42] . Nog haploinsufficiency protected mice from arthritis induced by methylated bovine serum albumin [43] . Overexpression of murine noggin has been associated with impaired function of osteoblasts , resulting in osteopenia , fractures and decreased bone formation rate [38 , 44 , 45] . The affinity of noggin to different BMPs varies . Further , there are other BMP antagonists that can partially compensate for the lack of noggin ( e . g . chordin , follistatin , gremlin and sclerostin ) [37 , 40 , 46–48] . However , siRNA-mediated Nog knock-down led to increased BMP-mediated osteoblastic differentiation and extracellular matrix mineralization without compensatory induction of gremlin or chordin expression [49] . Our in vitro expression data suggests that the variant upstream of NOG has potential gene-regulatory consequences . It is possible that the regulation of noggin expression levels is suboptimal in hip joints of German Shepherds prone to develop moderate-to-severe hip dysplasia . Another study revealed a single-nucleotide variant affecting the expression of NOG 105 bp downstream of the transcription start site , when the researchers investigated targeted sequencing data of a GWAS locus for human cleft lip , with or without cleft palate [50] . We were not able to close the 434 bp gap upstream of NOG with the targeted resequencing data . The overall coverage was variable , and parts of the target region were not covered at all . This is a general caveat of using probe-enriched genomic DNA templates for sequencing . We finally used PCR and sequencing to close the gap , which enabled the accurate positioning of the upstream deletion locus in relation with the coding sequence . The close proximity with NOG and the high degree of conservation with the corresponding sequence in human NOG promoter suggest that the uncovered new genomic sequence might be involved in the regulation of NOG expression . Together with the discovery of functionally active variant alleles upstream of NOG ( Figs 4 and 5 ) , our results suggest more research should be targeted to the characterization of canine NOG and its regulation . The protective locus on chromosome 1 spans over a 1 . 1 Mb region and harbors two genes of interest: NOX3 and ARID1B . NOX3 belongs to the family of NADPH oxidases , which catalyse the formation of superoxides and other reactive oxygen species . NADPH oxidase enables the production of hydrogen peroxide ( H2O2 ) , which is ultimately used in a reaction cascade that participate in the initiation of articular cartilage degradation [51 , 52] . NOX3 is a non-phagocytic member of the NADPH oxidase family and it is mainly expressed in the inner ear and fetal tissues [53] . Thus , the role of NOX3 molecule in hip dysplasia remains uncertain , although as shown in S10 Table , an indirect link between NOX3 and TRIO , a protein encoded by another candidate gene for German Shepherd hip dysplasia has been reported in a study by Fels et al . ( 2014 ) [54] . The AT-rich interactive domain-containing protein 1B encoded by the second candidate gene ( ARID1B ) on chromosome 1 , functions as a transcriptional activator and repressor via chromatin remodeling [55] . Mutations in ARID1B cause Coffin-Siris syndrome ( CSS ) , which is a rare hereditary disorder affecting multiple body systems , for instance the nervous , cardiovascular , and skeletal systems [56 , 57] . As a consequence to this syndrome , ARID1B is associated with joint laxity ( 66% of the patients ) [56 , 57] . However , the dogs with hip dysplasia do not exhibit similar multisystemic symptoms as the CSS patients with causative ARID1B mutations . MAMDC2 is another potential candidate gene on chromosome 1 . It encodes a proteoglycan and has been associated with increased intraocular pressure [58] . Other putative candidate genes on chromosome 9 uncovered in the variant analysis ( Table 7 ) include MPO , RNF43 , RAD51C . Reactive oxygen species and MPO have been inferred to participate in the regulation of chronic inflammation [59–61] . Therefore , it was intriguing to discover a potentially deleterious missense variant of MPO ( Table 7 ) . The mutated amino acid , however , is not included in the predicted mature protein . Intronic variants close to splice regions in RNF43 are also potentially significant . RNF43 ubiquitin ligase [62] negatively regulates WNT signaling [63] . WNT signaling is implicated in osteoarthritis as reviewed in [64 , 65] , and a recent study also suggest it might be affected in CHD [66] . RAD51C is a well-known recombination factor [67] . Deciphering polygenic , multifactorial disorders requires large sample sizes . Although dogs have an unique genomic architecture [28 , 68–70] that facilitates association studies in smaller cohorts than in humans [28] , the lack of power is still a regular concern . Our GWAS was unexceptional in this respect . Even after we increased the sample size from 292 to 409 or 524 dogs , and consequently revealed two additional loci on chromosome 1 , none of the associations reached genome-wide significance . We observed strong LD in our data ( S1 Table ) , which was expected due to the genomic architecture of dogs . Therefore , Bonferroni correction threshold could be overly conservative for our data as explained in Methods , part “Genome-wide association analysis” . Also , the lack of power may be a consequence of the increasing variation among cases , when we included the mild hip dysplasia phenotypes in the second meta-analysis . We observed significant differences in the allele and genotype frequencies between mild cases ( C/C ) and the moderate-to-severe cases ( C/D , D/C or worse ) throughout the loci on chromosome 9 ( Table 5 ) , whereas mild cases did not differ from controls in these comparisons . Additionally , the fragment genotype frequencies related to NOG were similar for controls and mild cases but again differed significantly when these two groups were separately compared with the moderate-to-severe cases . These findings corroborate that the dogs with mild hip dysplasia are indeed at lower genetic risk for the disorder . It would be important to find out , if other genetic factors differentiating the dogs in these phenotype groups exist . In conclusion , using several genetic approaches we have discovered novel variants of a putative NOG enhancer that downregulate reporter gene expression in vitro . The variants are associated with healthy and mildly dysplastic hip joints in German Shepherds . Besides a larger replication study and investigation of the other candidate genes on chromosomes 1 and 9 , future research should focus on what kind of biological effects the variants have on the expression of noggin in the canine hips and on the development of hip dysplasia .
The Finnish Kennel Club ( FKC ) granted permission to use its data and CHD screening radiographs for our research . All radiographs have been scored by two specialized veterinarians , thus reducing inter-observer bias [7] . All hip score results are freely available from the FKC breeding database [71] . Our study cohort consisted altogether 531 German Shepherds ( 247 cases + 284 controls ) , born between 1993 and 2013 . Cases were dogs with an FCI score C or worse for both hips and controls were dogs with a score A for both hips . We discarded dogs with an FCI score B because their inclusion may lead in a confounded control phenotype . Five control dogs had to be excluded from the analyses due to ambiguous phenotypes . This left us with a total of 526 dogs ( 247 cases and 279 controls ) before quality control . However , one more control had to be removed during quality control due to an outlier genotype , after which we had 525 dogs left for the GWAS . At least one EDTA blood sample was collected from all the dogs between years 2006 and 2015 . The dogs were chosen for our study according to their hip scores and pedigrees , creating a balanced study population of working , mixed and show line dogs ( S10 Fig ) . Guidelines for research ethics and good scientific practices were followed . We hold an ethical license for collecting EDTA blood samples ( ESAVI/7482/04 . 10 . 07/2015 ) , from ELLA–Animal Experiment Board in Finland under The Regional State Administrative Agency for Southern Finland . The owners signed a form of consent and they were well informed of the project . The original EDTA-blood samples are stored at the Dog DNA bank at the University of Helsinki . DNA extraction from the EDTA-blood samples was carried out using Chemagic Magnetic Separation Module I ( MSMI ) with a standard protocol by Chemagen ( Chemagen Biopolymer-Technologie AG , Baeswieler , Germany ) , after which the samples were sent to Geneseek ( Lincoln , NE , US ) to be genotyped using the high density 173K canine SNP array from Illumina ( San Diego , CA , USA ) . Genotyping was executed in several batches as collection of the original EDTA-samples took place over several years . Batch effect was accounted for as a covariate in our meta-analyses . Our German Shepherd population was divided into five ( the original GWAS , S11 Fig ) or four ( meta-analyses , S12 Fig ) subpopulation clusters according to their genomic relationships . This was achieved by first calculating the appropriate number of clusters from a genomic relationship matrix with a package “mclust” [72] in R [73] , which uses covariance parametrization and selects appropriate clusters via Bayesian information criterion . A covariate vector was created according to the clustering data , so that each individual belongs to one of the clusters . This covariate was used in our model to account for any differences in disease association between the clusters . The same genotype data was used for both meta-analyses , but the number of included dogs in each analysis was determined by the stringency of the phenotypes . The quality control for this data was carried out in two steps . First , before merging the original data ( 292 dogs ) with the new genotypes ( 233 dogs ) , initial quality controls were executed separately on them with PLINK . The following thresholds were used for each data set: per sample call rate 0 . 90 , per SNP call rate 0 . 95 , minor allele frequency 0 . 05 , and p-value cut-off level < 0 . 00001 for the HWE check . Also , strand had to be flipped for 59980 SNPs in the original data set due to strand inconsistencies with the new genotype data . This was done with the --flip command in PLINK . Second , after merging of the data sets , the data was imported to R , where the QC was repeated with GenABEL for the whole data with the same QC thresholds . This left 88499 SNPs for the meta-analyses . After the meta-analyses , we checked the genotype call quality of the best SNPs to verify that the associations were not due to genotype-calling or other such errors . One SNP on chromosome 4 ( BICF2P491963 ) was observed to show false association in the first meta-analysis due to a batch-specific calling error . The error was not resolved by the use of batch covariates , and the SNP was therefore removed . Also , the genotyping batch of one dog was missing . This dog was therefore removed leaving 524 dogs for the meta-analyses . We performed a case-control GWAS to identify SNPs associated with canine hip dysplasia . The original association study included 160 controls and 132 cases . The GWAS was implemented in R with the package GenABEL [31] . The covariates were sex and the genomic cluster of the animal . In the meta-analyses we also used the genotyping batch as a covariate . We used FASTA [76] and QTSCORE [32] , in GenABEL to calculate the association test statistics . When used with a binary trait FASTA corresponds to the Cochran-Armitage trend test [77] . FASTA is an efficient tool for association analysis in family-based data sets . However , FASTA has the disadvantage of not being able to compute a genome-wide significance with permutation analysis , because the data structure of the test statistics is not exchangeable . This is due to incorporating the relationship matrix ϕ in the computation of the test statistics [76] . QTSCORE does not suffer from this , as the test statistics derive from the environmental residuals that are not correlated with each other . Thus , the data structure is exchangeable and permutation analysis can be used to calculate empirical experiment-wise genome-wide significance levels for the analyzed SNPs [32] . Bonferroni correction threshold for genome-wide significance was determined as ( P-value/Number of SNPs ) = 0 . 05/92315 = 5 . 42x10-7 for the original GWAS , and 0 . 05/88499 = 5 . 65x10-7 for the meta-analyses . However , Bonferroni correction is problematic in genetic association studies , because it expects independence between the comparisons , which does not hold for SNPs due to LD [78] . Consequently , when type I error is controlled with overly conservative Bonferroni adjustment , type II error rate might be inflated if the sample size is small , and some QTL with real effects may be ruled insignificant [78] . Therefore , we estimated the effective number of independent tests using simpleM [79] for use in permutation analysis for genome-wide significance as 24159 for the original GWAS and 26323 for the meta-analyses . We also used these values to calculate thresholds for significance that rely on more accurate estimates of independent tests: 0 . 05/24159 = 2 . 07x10-6 for the original GWAS and 0 . 05/26323 = 1 . 90x10-6 for the meta-analyses . We used the function “r2fast” [80] from the GenABEL-package in R to estimate the r2 values between the top SNPs from the genome-wide association analyses . For one SNP in the first meta-analysis ( BICF2P272135 ) , we re-calculated the r2 values with the RSQ-function in excel , because of a batch specific allele flip that affected the LD-estimation in R . To estimate the number of independently associated loci within the target regions on chromosomes 1 and 9 , we used a SNP clumping procedure . This was executed with the “clump . markers” function from the R-package cgmisc [81] . The threshold for forming the clumps were as follows . The physical distance cut-off for clumping was set to 7 . 5 Mb to cover all of the associated loci on both targeted chromosomes , so as not to create any clumps due to distance , but only due to association with the trait ( P-value threshold = 5 . 0x10-5–5 . 0x10-6 ) , and due to high enough correlation between the SNPs ( r2 threshold = 0 . 70 ) . A targeted sequencing of a 7 Mb region on canine chromosome 9 ( bases 30620001 to 37620000 from NC_006591 . 3 ) was executed by the DNA Sequencing and Genomics lab at the University of Helsinki . The study included 24 cases and 24 controls that were chosen by the combinations of their genotypes for the following markers: BICF2S23027935 , BICF2P742007 , BICF2G630834826 , BICF2G630837209 , BICF2G630837240 , BICF2G630837405 and BICF2P272135 ( See also Tables 1 and 2 ) . SNP genotype combinations for 24 controls and 23 cases were GAGAGCG and AGAGATC , respectively . In addition , one case had the combination AGARRYS . An indexed Illumina library was created for all 48 samples . Briefly , DNA was sheared using a Bioruptor NGS sonicator ( Diagenode , Denville , NJ , US ) and the obtained fragments were end-repaired , A-tailed and truncated Illumina Y-adapters ligated . In a PCR step ( 20 cycles ) full-length P5 and indexed P7 adapters were introduced using KAPA Hifi DNA Polymerase ( KAPA Biosystems , Wilmington , MA , US ) . Pools containing four samples each were made for sequence capture with custom SeqCapEZ probes ( Nimblegen/Roche , Madison , WI , US ) targeting the 7 Mb area from the genome . The sequence capture was performed according to the manufacturer’s protocols ( Nimblegen/Roche , Madison , WI , US ) . The captured fragments were amplified ( 20 cycles ) using Illumina adapters P5 and P7 as described above . The PCR products were purified , and size selected using AMPure XP beads ( Beckman Coulter Inc . , Brea , CA , US ) . The obtained final libraries were paired-end ( 300 bp + 300 bp ) sequenced on a MiSeq Sequencer ( Illumina , San Diego , CA , US ) . The adapter sequences were removed and the raw reads were filtered using PRINSEQ [82] . After quality control , the remaining 47272947 ( 94 . 6% ) reads were mapped to the reference sequence CanFam3 . 1 using Burrows-Wheeler Alignment tool [83] . The aligned reads were visualized in Tablet and Integrative Genomics Viewer [84 , 85] . We implemented a targeted re-sequencing analysis pipeline to screen for coding variants in comparison with CanFam3 . 1 reference genome . FASTX was used to perform base quality check of the raw reads and Burrows-Wheeler Aligner ( BWA ) version 0 . 5 . 9 [83] was used to map the reads to the reference genome . Picard tools ( http://broadinstitute . github . io/picard/ ) was used to sort and mark possible PCR duplicates . Re-alignment around indels and base quality score recalibration was done using GATK . The variant calling was carried out using the Genome Analysis Tool Kit ( GATK ) version 3 . 5 [86] and SAMtools version 1 . 2 [87 , 88] . The detected variants were annotated to Ensembl and NCBI gene annotation databases using ANNOVAR [89] . Using 258 controls ( hip scores A/A , including the sequenced 24 controls ) and 168 moderate-to-severe-cases ( hip scores C/D , D/C or worse , including the 24 sequenced cases ) , we determined the statistical association between the phenotype and SNP variants in the target area ( S4 Table ) . The Cochran-Mantel-Haenszel test variable M2 for the independence of variants and the phenotype could be determined for 217 SNPs ( S5 Table ) . The null distribution of maximum M2 from 10000 permutations had a mean value of 8 . 25 and with 95% confidence interval ranging between 4 . 06 and 15 . 01 . Using the null distribution as a reference , 28 of the 217 SNPs were statistically associated with the phenotype ( Bonferroni-corrected , adjusted p-value < 0 . 05 , N = 217 ) ( S5 Table , S4 Fig ) . We performed a PCR with a region of 400 bp encasing the deletion revealed in the targeted sequencing . We designed the primers for this with the NCBI Primer-BLAST tool [90] . The primer sequences are in the supporting information ( S11 Table ) . Basic and 5’-FAM-labeled primers were from Oligomer ( Helsinki , Finland ) . The annealing temperatures were calculated with Thermo Fisher Scientific Tm calculator for Phusion DNA polymerase [91] . The PCR was run with a T100 Thermal Cycler ( Bio-Rad , California , US ) with a standard 3-step protocol for Phusion reaction . Standard 1 . 2% and 2% agarose gels were used ( A9539; Sigma Aldrich , St . Louis , MO , US ) , with 1 x TBE buffer and ethidiumbromide staining . Sample and ladder volume were 5 μl in all lanes . We used GeneRuler 100 bp ( SM0242 ) and 100 bp Plus ( SM0321 ) , from Thermo Fischer Scientific ( Waltham , MA , US ) as the DNA ladders . The gel-imaging was performed with AlphaImager ( Alpha Innotech , Kasendorf , Germany ) . The PCR amplicon was validated with sequencing . PCR products from 18 dogs were ambiguous on gels and were sent for fragment analysis . Nine samples did not yield a product with either method and one sample remained ambiguous leaving us with 516 fragment genotypes . The DNA Sequencing and Genomics lab at the University of Helsinki carried out both the sequencing and the fragment analysis . They used capillary electrophoresis to analyze the fragments , with a GeneScan 500 ROX dye ( 4310361; Thermo Fisher Scientific , Waltham , MA , US ) size standard . Subsequently , we analyzed the data with Peak Scanner v1 . 0 ( Applied Biosystems , Foster City , CA , US ) . Logistic regression models with or without the NOG regulatory variants were computed in R [73] . The odds ratios corresponding with the GLM coefficients were calculated using R package ‘oddsratio’ [92] . AUC calculations and comparisons were done using R package ‘pROC’ [93] . A reference sequence was assembled using CSC computational hub based on the targeted sequencing reads from a case that did not exhibit the deletion upstream of NOG . The adapters were removed and quality of the fastq files was assessed using FastQC [94] . The de novo assembly was done using the Spades assembler [95] . Assembly was done for the following k-mer values ( 21 , 33 , 55 , 77 , 99 , 127 ) ; the Spades assembler then generates a combined assembly ( i . e . scaffolds ) based on the kmers used . The assembly QC for the scaffolds was done using ‘Quast’ [96] . Genomic DNA from Dog6 was amplified using primers CanNOG-F1 and CanNOG-R1 from Ishii et al . [35] . The PCR products were sequenced , low quality sequences were discarded and a consensus sequence was derived . The alignments between Dog6 , CanFam3 . 1 chr9 and GRCh38 chr17 were done using MAFFT [97] . The human core JASPAR2018 database [36] was queried with the alignment in S9 Fig using TFBSTools [98] . Bonferroni correction was used to adjust the P-values for each putative binding site for all the matrix ID’s . The matrix ID specific prediction was considered significant if the bonferroni-corrected P-value for any of its binding sites was less than 0 . 05 . According to the findings from the targeted sequencing we designed three different sequence variant constructs: A , B and C , where A is our German Shepherd reference sequence , and B and C are variants with deletion of eight or seven AGG-triplets . The construct sequences are shown in the supporting information ( S7 Table ) . The longest construct ( construct A ) was designed based on the Dog2 scaffolds generated from the resequencing data ( S7 Fig ) . The NOG enhancer sequence variants were cloned into the pNL3 . 1[Nluc/minP] NanoLuc luciferase vector ( Promega , Madison , WI , US ) . pGL4 . 54[luc2/TK] firefly luciferase was used as a constitutively expressed control plasmid . 24 h prior to transfection , 2 x 104 HEK293 or 8 x 103 U-2 OS cells were plated to 96 well plates in DMEM medium supplemented with 10% FBS and without antibiotics . The HEK293 cells were transfected with 50 ng of each plasmid DNA and 50 ug carrier DNA / well and the U-2 OS cells with 10 ng of each plasmid and 80 ug carrier DNA/well using Fugene HD transfection reagent ( Promega , Madison , WI , US ) . Luciferase activities were measured after 24 h using the Nano-Glo Promega Dual-Luciferase reporter assay system according to the manufacturer’s instructions . The NanoLuc luminescence values were normalized by division with the control firefly luminescence . The data for every setup ( three transfection experiments each with four technical replicates ) was analyzed in R using the Kruskal-Wallis rank sum test followed by Dunn’s test for multiple pairwise comparisons with Bonferroni adjustment for P-values . P-value < 0 . 05 was considered significant . | Hip dysplasia is a common orthopedic disorder in dogs and humans . It can pose a serious welfare problem with severe pain . The genetic background of this disorder remains inconclusive even after years of arduous research . We used the genotypes of 525 German Shepherds with carefully determined hip scores to identify genomic regions potentially harboring genetic risk factors for the disorder . We found four regions on chromosomes 1 and 9 exhibiting suggestive association with the disorder phenotypes . Further analysis of the identified loci on chromosome 9 by sequencing 48 dogs revealed deletions in a potential regulatory region of NOG - the gene encoding noggin , a known regulator of joint development in mice and in humans . Using a reporter assay , we demonstrated that the deletions decrease the enhancer activity of the regulatory region and could therefore affect the expression of NOG in hips . The deletions significantly differentiate the healthy and the mild phenotypes from the moderate-to-severe phenotypes . Therefore , our results suggest that the deletion protects against hip dysplasia . Future research should focus on how these regulatory variants affect the expression of noggin in canine hips , and what the roles of noggin and the other revealed loci are in canine hip dysplasia . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"dysplasia",
"pathology",
"and",
"laboratory",
"medicine",
"statistics",
"metaanalysis",
"variant",
"genotypes",
"vertebrates",
"dogs",
"animals",
"mammals",
"genetic",
"mapping",
"mathemati... | 2019 | Novel protective and risk loci in hip dysplasia in German Shepherds |
High-risk human papillomavirus ( HPV ) E6 proteins associate with the cellular ubiquitin ligase E6-Associated Protein ( E6AP ) , and then recruit both p53 and certain cellular PDZ proteins for ubiquitination and degradation by the proteasome . Low-risk HPV E6 proteins also associate with E6AP , yet fail to recruit p53 or PDZ proteins; their E6AP-dependent targets have so far been uncharacterized . We found a cellular PDZ protein called Na+/H+ Exchanger Regulatory Factor 1 ( NHERF1 ) is targeted for degradation by both high and low-risk HPV E6 proteins as well as E6 proteins from diverse non-primate mammalian species . NHERF1 was degraded by E6 in a manner dependent upon E6AP ubiquitin ligase activity but independent of PDZ interactions . A novel structural domain of E6 , independent of the p53 recognition domain , was necessary to associate with and degrade NHERF1 , and the NHERF1 EB domain was required for E6-mediated degradation . Degradation of NHERF1 by E6 activated canonical Wnt/β-catenin signaling , a key pathway that regulates cell growth and proliferation . Expression levels of NHERF1 increased with increasing cell confluency . This is the first study in which a cellular protein has been identified that is targeted for degradation by both high and low-risk HPV E6 as well as E6 proteins from diverse animal papillomaviruses . This suggests that NHERF1 plays a role in regulating squamous epithelial growth and further suggests that the interaction of E6 proteins with NHERF1 could be a common therapeutic target for multiple papillomavirus types .
Human papillomaviruses ( HPVs ) are small DNA tumor viruses that cause squamous epithelial papillomas in which the virus replicates . The papillomas are initially benign and the host is usually able to clear the underlying HPV infection over time . However , a subset of HPV infections may result in lesions that persist and grow to harmful size or that have a propensity to evolve into carcinomas [1] . The cancer-causing HPV types are called high-risk and the most commonly occurring high-risk types are HPV16 and HPV18 . Worldwide , high-risk HPVs are responsible for 5% of cancers , with cervical cancer being the most common [2] . HPV types that are not associated with malignancies are termed low-risk HPV; although low-risk for malignancies , the size and location of the benign papillomas can render these lesions medically serious [3] . Beyond HPVs , papillomaviruses have been isolated from mammalian species including rodents , primates , bats , cetaceans , and ungulates [4] , and are clustered into related genera based upon the divergence of the L1 capsid protein nucleotide sequence ( both high and low-risk mucosal HPV types discussed in this study belong to the primate Alpha genera ) [5] . Most non-human papillomaviruses encode E6 proteins that are similar in predicted fold to high-risk HPV16 E6 [6] . When diverse mammalian papillomaviruses are clustered based on their E6 sequence similarity , two main groups of papillomaviruses emerge: those that encode E6 proteins that bind to the Notch co-activator Mastermind Like 1 ( MAML1 ) , and those that bind to a cellular E3 ubiquitin ligase called E6-Associated Protein ( E6AP ) [7] . An E6 protein that preferentially binds MAML1 suppresses MAML1 transcriptional activation , while an E6 protein that preferentially binds E6AP stimulates E6AP E3 ubiquitin ligase activity to then target additional cellular proteins recruited by E6 to E6AP for ubiquitination and degradation by the proteasome [7] . The difference between the propensity of high and low-risk HPVs to cause cancer is secondary to differences between their respective E6 and E7 oncoproteins [8] . E6 and E7 from both high and low-risk HPVs bind cellular E3 ubiquitin ligases and hijack their ubiquitin ligase activity to perturb certain cellular proteins that are recruited by E6 or E7 [9] . Both high-risk and low-risk E7 proteins interact with ubiquitin ligases of the cullin and N-end rule families and target the degradation of additional cellular proteins recruited to E7 such as pocket-family proteins ( RB , RBL1 , and RBL2 ) and PTPN14 [10 , 11] . High and low-risk E7 proteins target certain cellular proteins in common ( such as the RBL2 p130 pocket protein ) [12] . However only high-risk HPV E7 types interact with and target the degradation of the retinoblastoma ( RB ) pocket protein [13 , 14] , which has implications for the carcinogenic properties of high-risk E7 . High and low-risk Alpha genera HPV E6 proteins interact with the cellular E3 ubiquitin ligase E6AP [15–17] , but only cellular proteins targeted for degradation by the high-risk E6 protein ( such as p53 ) are well established [18 , 19] . Another striking difference between high and low-risk E6 is the presence of a PDZ binding motif ( PBM ) at the extreme carboxy terminus of high-risk E6 proteins [20–22] . The high-risk PBM enables E6 to interact with a group of cellular proteins termed PDZ proteins , all of which contain PDZ ( PSD-90/Dlg1/ZO-1 ) homology domains [23] . The targeted degradation of cellular proteins that are recruited through interaction with the high-risk E6 PBM has been controversial , but the E6 PBM functionally promotes retention of the viral DNA plasmid within infected cells [24]; the E6 PBM function can be rescued by disruption of p53 function [25] . Although low-risk E6 proteins bind E6AP , they do not have a PBM at the carboxy-terminus [20] , do not interact with p53 [17] , and no cellular targets of the low-risk E6+E6AP complex have been described . Such cellular targets would be presumed to be of exceptional interest since they would be common to both high and low-risk E6 proteins , just as RBL2 is a common target of the high and low-risk HPV E7 protein . In this study , we identify the PDZ-adapter protein NHERF1 as degraded by both high and low-risk E6 proteins , in a manner dependent upon the ubiquitin ligase activity of E6AP and the proteasome . Other E6 proteins from diverse species where E6 could bind E6AP were also able to initiate NHERF1 degradation , indicating the conservation of this function . Interaction of NHERF1 with E6 required prior association of E6 with E6AP , and we identified a novel interaction domain within 16E6 that is required . Finally , the targeted degradation of NHERF1 by both low and high-risk E6 proteins resulted in the activation of canonical Wnt signaling , connecting the degradation of NHERF1 by E6 to the activation of an oncogenic signaling pathway .
NHERF1 was previously shown to be degraded by HPV16 E6 ( 16E6 ) ( but not by HPV18 E6 ( 18E6 ) or HPV11 E6 ( 11E6 ) ) through an interaction requiring the PBM of 16E6 [26] . In our proteomic studies of cellular proteins that associate with the 16E6 and 18E6 PBMs [27] , we did not identify NHERF1 , but in other experiments observed a reduction of NHERF1 protein levels by 16E6 , 18E6 , and 11E6 . In order to characterize the reduction of NHERF1 by these E6 proteins , we performed transient transfections into E6AP-null 8B9 cells reconstituted with either WT E6AP ( E6AP_WT ) or a mutant E6AP defective in ubiquitin ligase activity ( E6AP_Ub− ) . E6APs were co-transfected with plasmids encoding p53 , NHERF1 , and 16E6 , 16E6 deleted of the PBM ( 16E6ΔPBM ) , 11E6 , or 18E6 . Consistent with published literature , p53 was degraded by high-risk E6 proteins ( 16E6 and 18E6 ) independently of a PBM [25] and dependent upon E6AP ubiquitin ligase activity [28] ( Fig 1 ) . Expression of 11E6 together with E6AP_WT resulted in a lack of p53 degradation by low-risk E6 ( 11E6 ) , corroborating published findings [17 , 29] . However , NHERF1 protein levels were reduced by each observed E6 protein ( Fig 1A ) , in contrast to what has been previously published [26] . To ensure the reduction of NHERF1 by either high or low-risk E6 proteins was not due to an overexpression artifact , we performed an E6 titration experiment ( Fig 1B ) . Representative western blots from which the quantification in Fig 1B was derived are shown in S1 Fig . Three different E6 proteins were used: 16E6_WT , 16E6ΔPBM , and 11E6_WT . We used p53 as a control for 16E6-mediated degradation , as multiple studies have shown low-risk E6 proteins ( 11E6 ) do not degrade p53 [17 , 29] while high-risk 16E6 is able to degrade p53 independent of the presence of a PBM [25] . Observing the degradation of p53 in cells expressing variable amounts of E6 provided a guide for physiologically relevant E6 expression levels . NHERF1 and p53 protein levels were similarly reduced by both 16E6_WT and 16E6ΔPBM at the various E6 titrations ( Fig 1B ) . 11E6_WT was unable to initiate the degradation of p53 but targeted NHERF1 at levels similar to those required by 16E6 . Deletion of the 16E6 PBM did not impact the degradation of p53 or reduction of NHERF1 protein levels by 16E6 . To determine the physiologic relevance of the observed reduction of NHERF1 protein by E6 , we seeded keratinocytes either lacking or containing the HPV16 genome at equal confluencies and observed NHERF1 protein levels by western blot ( Fig 1C ) . Keratinocytes containing the HPV16 genome not only degraded p53 , as expected , but also reduced NHERF1 protein levels by 84% . We established that NHERF1 protein levels are reduced by E6 in a transient transfection system . To determine whether low levels of stable 16E6-expression could initiate the reduction of NHERF1 protein levels , we retrovirally transduced normal immortalized keratinocytes with either empty vector or 16E6_WT and observed NHERF1 protein levels . Initially , our results were variable . We hypothesized that keratinocyte confluency may affect NHERF1 protein levels . To test this possibility , we seeded vector-transduced and 16E6_WT-transduced keratinocytes at three different cell densities: 5x103 cells/cm2 ( very sub-confluent ) , 1 . 3x104 cells/cm2 ( sub-confluent ) , and 2 . 6x104 cells/cm2 ( mid-confluent ) . NHERF1 protein levels increased with an increase in cell density and 16E6_WT consistently reduced NHERF1 ( Fig 2A ) . In order to determine if changes in NHERF1 levels with confluency were secondary to changes in NHERF1 RNA levels , we performed qPCR on RNA extracted from keratinocytes retrovirally transduced with vector or 16E6_WT and plated as in Fig 2A . Interestingly , NHERF1 RNA levels did not differ between keratinocytes seeded at different densities expressing either empty vector or 16E6_WT ( Fig 2B ) . Because the ability of each E6 protein to reduce NHERF1 protein levels was also dependent upon the ubiquitin ligase activity of E6AP ( Fig 1A ) , we hypothesized that E6 reduction of NHERF1 levels would be secondary to proteasome activity . We seeded retrovirally transduced keratinocytes expressing either empty vector or 16E6_WT at similar confluency and treated with either DMSO , mitomycin C ( MMC ) to induce p53 [30] , or the proteasome inhibitor MG132 at differing concentrations for 8 hours . As expected , p53 levels increased in vector keratinocytes treated with MMC compared to untreated cells as well as in 16E6_WT cells exposed to increasing concentrations of MG132 [28] ( Fig 3 ) . NHERF1 protein levels increased significantly in a dose dependent manner upon treatment with MG132 in parallel to that seen with p53 . ( Fig 3 , lanes 3–8 ) . This indicated that NHERF1 is degraded through the proteasome by E6 in a manner dependent upon WT E6AP . The observation that NHERF1 was targeted by both high and low-risk HPV E6 proteins suggested that NHERF1 may also be a target of diverse non-primate E6 proteins . We examined the ability of E6 proteins from multiple different genera and different mammalian species to target NHERF1 for degradation ( Fig 4 ) . E6 proteins that preferentially bind MAML1 were unable to degrade NHERF1 ( Fig 4A and 4B ) . All of the tested Alpha ( primate ) , Dyodelta ( boar ) , and Dyopi ( porpoise ) genera E6 proteins that bind E6AP targeted NHERF1 . While E6AP-binding was necessary it was not sufficient , as E6 proteins from Omega ( polar bear , UmPV1 ) and Omikron ( cetaceans , PphPV1 and TtPV5 ) did not degrade NHERF1 ( Fig 4A ) . Interestingly , E6 proteins that bind E6AP but did not target NHERF1 degradation sequence-clustered separately from E6 proteins that did target NHERF1 degradation , suggesting evolutionary divergence of this function ( Fig 4B ) . Because the ability of E6 to degrade NHERF1 was not dependent upon the presence of a PBM ( Figs 1 and 4 ) , we attempted to identify which residue ( s ) of 16E6 were required to mediate degradation of NHERF1 . The crystal structure of 16E6 complexed with the E6-binding peptide from E6AP [31] was examined to identify amino acids that were at least 20% exposed , resulting in over eighty candidate residues ( S2 Fig ) . Candidate residues were individually mutated in the context of the 16E6 gene and the resulting point mutants were screened for their ability to degrade NHERF1 in the presence of E6AP_WT in transiently transfected C33A cells . To ensure our point mutants were not functionally defective ( i . e . could not fold properly or could not interact with E6AP ) , we also screened the mutants for ability to degrade p53 . A selection of mutants and the results of the screen are shown in Fig 5 . Four mutants stood out as selectively defective in their ability to degrade NHERF1 ( Fig 5A and 5B ) while still being able to degrade p53 ( Fig 5C ) : F69A , K72A , F69R and a double mutant: F69A/K72A . As evidenced in the crystal structure of 16E6 , the side chains of F69 and K72 ( Fig 6A and 6B ) are located along the connecting alpha-helix that links the amino-terminal and carboxy-terminal zinc-structured domains of 16E6 . The F69 and K72 side chains are aligned and adjacent on the connecting helix , which is on the opposite side of 16E6 from the p53 interaction surface [32] ( Fig 6C ) . We had identified the 16E6_F69A/K72A mutant in a transient transfection screen . To ensure the identified 16E6_F69A/K72A double mutant was selectively defective for degrading NHERF1 in the context of a stable cell line , keratinocytes retrovirally transduced with empty vector , 16E6_WT , 16E6ΔPBM , 16E6_F69A/K72A , or 11E6_WT were seeded at equal confluency and lysates prepared . Keratinocytes expressing 16E6_WT , 16E6ΔPBM , and 11E6_WT degraded NHERF1 ( Fig 7 , lanes 2 , 3 , 5 ) . However , keratinocytes expressing 16E6_F69A/K72A were unable to stimulate the degradation of NHERF1 ( Fig 7 , lane 4 ) , indicating a novel substrate interaction domain important for 16E6-mediated degradation of NHERF1 . Because the PBM of E6 proteins is not required to initiate the degradation of NHERF1 ( Figs 1 , 4 and 7 ) , we hypothesized that neither of the PDZ domains of NHERF1 would be required for 16E6 to initiate NHERF1 degradation . We truncated NHERF1 and deleted several characterized domains within the protein [33 , 34] ( Fig 8A ) . E6AP-null 8B9 cells were co-transfected with 16E6_WT , NHERF1 truncations , HA_GFP , and either E6AP_Ub− or E6AP_WT . NHERF1 protein levels were quantified , and then normalized to the internal transfection control ( HA_GFP ) . The various NHERF1 truncations displayed different expression levels . To account for these variations , levels of NHERF1 truncations in the presence of E6AP_Ub− were set to 100% and the expression level of the corresponding NHERF1 truncation in the presence of E6AP_WT was normalized accordingly ( Fig 8B and 8C , bar graphs ) . All NHERF1 truncations containing the EB domain were targeted for degradation by 16E6 in the presence of E6AP_WT ( highlighted in green in Fig 8A ) . Truncations of NHERF1 that lacked the EB domain were not targeted for degradation by 16E6 ( highlighted in red in Fig 8A ) . In addition , the NHERF1 PBM was not required for 16E6 mediated degradation ( Fig 8C , lanes 5 vs . 6 and 9 vs . 10 ) . We identified the NHERF1 EB domain as a requirement for 16E6 mediated degradation and the importance of 16E6 residues F69 and K72 . In order to examine the interactions between the 16E6+E6AP+NHERF1 complex , all three proteins were expressed in a yeast three-hybrid system so as to detect the heterotrimeric complex . We fused 16E6_WT and ubiquitin ligase dead E6AP ( E6AP_Ub− ) to the LexA DNA binding domain and co-expressed this fusion with either vector , 16E6_WT , or 16E6_F69A/K72A in yeast containing a LexA responsive LacZ reporter . These yeast were then mated to yeast expressing native p53 or Gal4 ( G4 ) transactivator fusions to NHERF1 121–358 ( containing the EB domain ) , NHERF1 121–297 ( deleted of the EB domain ) , 16E6_WT , or the tyrosine phosphatase PTPN3 ( a PDZ protein ) ( Fig 9 ) . The LexA_16E6 fusion co-expressed with p53 ( in the absence of E6AP ) resulted in very weak activation of the LacZ reporter ( spot 4B ) while co-expression with G4_PTPN3 resulted in strong transactivation ( spot 6B ) , but no interaction with NHERF1 ( spots 2B and 3B ) . We then co-expressed 16E6 and E6AP by using a LexA_E6AP_Ub− fusion together with native 16E6 . When LexA_E6AP_Ub− , untagged 16E6_WT , and p53 were co-expressed , a strong activation of the LacZ reporter was observed ( Fig 9 , spot 4D ) , illustrating that while p53 has a weak direct interaction with 16E6 , it interacts strongly with 16E6 bound to E6AP . This activation was also seen with 16E6_F69A/K72A in the presence of LexA_E6AP_Ub− and p53 ( Fig 9 , spot 4E ) , indicating the preserved ability of the 16E6_F69A/K72A double mutant to bind E6AP and recruit p53 . When LexA_E6AP_Ub− , 16E6_WT , and G4_NHERF1 121–358 ( contains the EB domain ) were co-expressed , we observed activation of the LacZ reporter , indicating the recruitment of NHERF1 to E6AP by 16E6_WT ( Fig 9 , spot 2D ) . Truncating the EB domain from the G4_NHERF1 ( G4_NHERF1 121–297 , spot 3D ) or the use of the 16E6_F69A/K72A double mutant ( spot 2E ) ablated the reporter transactivation , indicating the requirement of the EB domain and the importance of 16E6 residues F69 and K72 in the interaction of the 16E6+E6AP+NHERF1 complex . It has been shown that high-risk HPV E6 proteins augment the canonical Wnt/β-catenin signaling pathway [35–39] . Additionally , it has been shown that NHERF1 inhibits the canonical Wnt/β-catenin signaling pathway through multiple mechanisms . NHERF1 forms a complex with β-catenin [40] and can also bind to the intracellular PBM of certain isoforms of Frizzled [41] , a G-protein coupled receptor important in the activation of the canonical Wnt signaling pathway . Therefore , we hypothesized E6 degradation of NHERF1 would activate the Wnt/β-catenin signaling pathway in cells expressing E6 . To test this possibility , we utilized the TOP/FOP luciferase reporter assay . 16E6 , 16E6ΔPBM , 11E6 , and 18E6 all stimulated the activity of the Wnt/β-catenin pathway over vector-transfected cells ( Fig 10 ) . However , cells transfected with 16E6_F69A/K72A were unable to augment the canonical Wnt pathway over vector levels , indicating that the ability of E6 to degrade NHERF1 is required for E6 activation of the canonical Wnt/β-catenin signaling pathway . The earlier Accardi et al . study proposed that expression of 16E7 sensitized NHERF1 for degradation by the induction of NHERF1 phosphorylation [26] . Our experiments did not show either E7 induction of slow-migrating NHERF1 phosphorylated isoforms or an enhancement of E6-NHERF1 degradation upon co-expression of E7 ( S3 and S4 Figs ) .
E6 proteins from papillomaviruses can be separated into distinct groups: those that bind MAML1 and repress Notch signaling , and those that bind E6AP and hijack its ubiquitin ligase activity [7 , 42–44] . E6 proteins from papillomaviruses in the Alpha , DyoDelta , Dyopi , Omega , and Omikron genera all behave similarly in that they bind E6AP and activate its ubiquitin ligase activity [7] . Here , we describe the degradation of NHERF1 by E6 proteins from both high and low-risk HPVs , as well as from papillomaviruses from multiple divergent mammalian species . The ability of these E6s to degrade NHERF1 is dependent upon E6AP ( Fig 1 ) and the proteasome ( Fig 3 ) . In addition , we identify two amino acids in 16E6 ( F69 and K72 ) that are necessary for E6-mediated degradation of NHERF1 . These two residues are aligned , and adjacent in the outwardly oriented face of the E6 connecting alpha helix , suggesting a novel interaction domain ( Fig 6B and 6C ) . Among the E6 proteins that target NHERF1 for degradation there are residues homologous to 16E6 F69 , being either phenylalanine or leucine at that position . The residues homologous to 16E6 K72 are less well-conserved . Although many mutants were screened , there may be additional residues in 16E6 that contribute to the interaction . NHERF1 degradation by E6 requires the NHERF1 EB domain , but does not require the PBM at the extreme carboxy terminus of NHERF1 ( Fig 8B and 8C ) . The ability of E6 proteins to degrade NHERF1 augments the canonical Wnt/β-catenin signaling ( Fig 10 ) , an oncogenic pathway frequently active in cancer . NHERF1 is the product of the SLC9A3R1 gene . SLC9A3R1 mRNA is broadly expressed in epithelia , with the highest mRNA expression in kidney , gut , and esophagus . NHERF1 is not developmentally essential , although mice have considerably reduced lifespans [45] . NHERF1-null mice are prone to phosphate wasting , brittle bone structure , and hydrocephaly [45] due to the mislocalization of proteins with which NHERF1 normally associates [45–47] . NHERF1 contains two PDZ domains and an EB domain at the carboxy terminus through which it interacts with ezrin , radixin , and moesin to link itself , and proteins to which it is bound , to the actin cytoskeleton network [48] . While the functions of NHERF1 are varied due to its role as a scaffold , multiple studies indicate it regulates cell growth and differentiation , two key cellular functions that papillomaviruses disrupt in the process of viral infection . Whether NHERF1 is a tumor suppressor or an oncogene has been debated in the literature . There are numerous papers regarding NHERF1 human cancer phenotypes , but they are collectively inconsistent [49 , 50] . NHERF1-null mice do not have a direct cancer phenotype , but have lengthened intestines [51] , indicating a growth regulatory function of NHERF1 . The diminished life span of NHERF1 null mice could limit observation of cancer traits . However , a recent in vivo study provided strong genetic support for NHERF1 as a tumor suppressor . APCMin/+ mice bred as either heterozygote or knockout for NHERF1 experience considerably shorter survival than their NHERF1-expressing counterparts due to increased tumor burden , demonstrating a tumor suppressor phenotype for NHERF1 [51] . Additionally , these APCMin/+ mice lacking NHERF1 have greater activation of Wnt/β-catenin signaling , suggesting NHERF1 acts as a negative regulator of this oncogenic pathway . NHERF1-associated proteins that plausibly could regulate cell proliferation are numerous and include β-catenin [40] , Frizzled [41] , G-protein coupled receptors ( β-adrenergic type 2 , [52] ) , receptor tyrosine kinases ( PDGFR , [53] ) , phosphatases ( PTEN , [54] ) , transcriptional coactivators ( YAP1 , [55] ) , ion channels ( Kir1 . 1 and CFTR , [56] ) , phospholipase-C [57] , and actin anchoring proteins ( ezrin , radixin , and moesin , [48] ) . Several studies have indicated that HPV E6 proteins can activate canonical Wnt/β-catenin signaling [35–39] . Our work expands and builds upon the scope of these studies . The ability of E6 to degrade NHERF1 and activate Wnt signaling may aid in propagation of papillomaviruses by enhancing the stimulation of cellular proliferation and promoting cell survival . There are numerous cell growth regulatory avenues that E6 could manipulate by degrading NHERF1 and within this study we have explored one possibility: the canonical Wnt/β-catenin signaling pathway ( Fig 10 ) ; other possibilities will be the subject of future studies . The EB domain of NHERF1 is required for E6-mediated degradation in the presence of E6AP ( Fig 8B and 8C ) . This domain is responsible for linking NHERF1 to the actin cytoskeleton network via interaction with ERM proteins [48] . NHERF1 has a PBM at its extreme carboxy terminus and when the EB domain is not bound to ERM proteins , the NHERF1 PBM can self-associate with the NHERF1 PDZ2 domain , resulting in a closed NHERF1 conformation [33] . The head-to-tail closed NHERF1 confirmation is not required for E6-mediated degradation , as an NHERF1ΔPBM mutant was still targeted for degradation by E6 in the presence of E6AP_WT ( Fig 8C ) . Nor was the 16E6 PBM required for degradation of NHERF1 ( Figs 1 , 4 and 7 ) , contrary to a prior report [26] . In addition to the requirement of the NHERF1 EB domain , two E6 residues , F69 and K72 , are necessary for E6-mediated NHERF1 degradation ( Figs 5 and 7 ) . Crucially , the 16E6_F69A/K72A double mutant can still initiate the degradation of p53 , indicating it is still able to bind E6AP , recruit p53 to the complex , and trigger ubiquitination . The F69 and K72 residues are also required to form a tri-molecular complex between E6AP , E6 , and NHERF1 in yeast ( Fig 9 , spot 2E vs . 2D ) . Like the association of E6 with p53 , NHERF1 does not interact directly with E6 , but requires prior association of E6 with E6AP , indicating that NHERF1 requires an altered conformation of E6 that is secondary to E6 binding to E6AP [58] . As we were testing the ability of E6 proteins to degrade NHERF1 in stable keratinocyte cell lines , we discovered that NHERF1 protein levels are sensitive to cell confluency ( Fig 2A ) . As cell confluency increases , so does the concentration of NHERF1 protein . E6 decreased NHERF1 expression at all confluency states . Thus , misleading interpretations can occur if different samples within an experiment are harvested at non-equivalent confluency . The relationship between NHERF1 and cell confluency may have contributed to the lack of identification of NHERF1 as a degradation target of HPV18 E6 and low-risk E6 proteins in the past , as well as differences between our studies and a prior publication [26] . It is likely that the effect of NHERF1 sensitivity to cell confluency underlies conflicting findings between different laboratories regarding NHERF1 expression levels and cancer associated traits [49 , 50] . Future studies of NHERF1 must take into account and carefully control cell densities when performing experiments . Binding to E6AP is necessary for E6-induced degradation of NHERF1 , but it is not sufficient , as three tested E6 proteins that bind E6AP do not target NHERF1 for degradation: UmPV1 E6 ( polar bear ) , PphPV1 E6 ( porpoise ) , and TtPV5 E6 ( bottlenose dolphin ) ( Fig 4A ) . Interestingly , the three E6 proteins that do not degrade NHERF1 cluster together in phylogenetic relatedness ( Fig 4B ) . We utilized transfected human NHERF1 throughout our study , so it is possible that the inability of these three E6 proteins to target NHERF1 for degradation may be due to evolutionary divergence in the NHERF1 homologs . Future studies will explore if the lack of degradation of human NHERF1 by UmPV1 , PphPV1 , and TtPV5 is due to evolutionary divergence of the respective NHERF1 proteins compared to human NHERF1 . It would be of interest to determine if NHERF1 is a “universal” target of E6 proteins that act through association with E6AP . Discovery of NHERF1 as a novel target for not only high and low-risk mucosal and cutaneous HPV E6 as well as a wide range of E6 proteins across divergent host species indicates a significant and previously undescribed role for NHERF1 in papillomavirus biology . That NHERF1 is a conserved target of papillomavirus E6 proteins further elevates the importance of NHERF1 as a cell growth regulator . Finally , the identification of this highly conserved E6 degradation target may represent a novel avenue for therapeutic intervention against both low and high-risk HPV .
E6AP-null 8B9 cells ( a gift of Dr . Lawrence banks , ICGEB , Italy ) [59] and HPV-negative C33A cervical cancer cells ( ATCC ) were maintained and transfected using polyethylenimine ( PEI ) as previously described [58] . Normal immortalized keratinocytes ( NIKS , obtained from ATCC ) are spontaneously immortalized foreskin keratinocytes [60] that were cocultured with mitomycin C-treated 3T3 feeder cells in F medium as described previously [61] . NIKS were retrovirally transduced with replication-defective murine retroviruses based on pLXSN [62] for E7 expression or pBABEpuro for E6 expression [63] . NIKS were transfected with the HPV16 genome as previously described [25 , 61] . Retrovirally transduced NIKS cells and NIKS transfected with the HPV16 genome were counted and seeded at equal confluency in each experiment . Epitope tagged E6AP , GFP , E6 , and NHERF1 were all transiently expressed from the pcDNA3 plasmid . HA-tagged human NHERF1 originated from Vijaya Ramesh’s laboratory ( from Addgene , plasmid 11635 ) . 16E6 point mutants were created using QuikChange primer design ( Agilent Technologies ) . 16E6ΔPBM was created by mutating the PBM of 16E6 from ETQL* to EL* and E6AP_Ub− was created by mutating the active cysteine residue at position 843 to an alanine ( C843A ) . The E6AP constructs utilized express human E6AP isoform 3 . NHERF1 truncations were PCR generated and sequenced . 12 well plates of transfected mammalian cells were lysed in 0 . 5X IGEPAL as described previously [7] . Transduced NIKS were lysed in 1% SDS , 5mM EDTA , and 1 mM sodium vanadate and equilibrated for protein content ( Biorad assay kit ) . All lysates were resolved by SDS-PAGE electrophoresis and transferred to PVDF membranes . Antibodies: anti-HA ( Bethyl Laboratories , Inc . ) , anti-FLAG M2 ( Sigma ) , anti-p53 Ab-8 ( ThermoFisher Scientific ) , anti-16E6 6G6 ( a generous gift from Arbor Vita Corporation ) , anti-SLC9A3R1 ( Sigma ) , anti-GAPDH ( Cell Signaling Technology ) , and anti-MYC 9B11 ( Cell Signaling Technology ) . Retrovirally transduced NIKS were plated at different cell densities and harvested following a TRIzol RNA harvest protocol ( Invitrogen ) . cDNA was generated using random hexamers . Quantitative real-time PCR was performed on the cDNA using iQ SYBR Green Supermix ( BioRad #1708880 ) . The primers targeted the SLC9A3R1 gene ( BioRad Assay ID: qHsaCEP0050521 ) and the GAPDH gene ( BioRad Assay ID: qHsaCEP0041396 ) . Relative values were analyzed using the ΔΔCT method ( where CT is the threshold cycle ) and GAPDH as a control . C33A cells plated at 70% confluency were transiently transfected with DNA of the TOPFLASH or control FOPFLASH ( containing mutated TCF/β-catenin binding sites; 1 ug ) plasmid , Renilla luciferase ( 0 . 005 ug ) plasmid ( used to evaluate transfection efficiency ) , FLAG_E6AP_WT ( 0 . 35 ug ) plasmid , and the indicated E6 plasmids ( 0 . 3 ug ) . 18 hrs post-transfection , media was removed and Wnt3A conditioned media was added for 8 . 5 hours to stimulate the Wnt pathway . Luciferase levels were measured using the Dual-Luciferase Reporter Assay System ( Promega ) and a Cytation1 Plate Reader ( software version 3 . 04 . 17 ) . FOPFLASH luciferase readings were low , and were subtracted from the paired TOPFLASH readouts . 10% fetal bovine serum Wnt3A conditioned media was generated using L Wnt-3A murine fibroblasts ( ATCC , CRL-2647 ) as previously described [64] . Multiple protein sequence files were downloaded from the Papillomavirus Episteme [65] and aligned using the EMBL-EBI MUSCLE ( MUltiple Sequence Comparison by Log- Expectation ) program [66] . The phylogenetic tree was generated as a neighbour-joining tree without distance corrections within the MUSCLE program [66] . Modified LexA-based yeast three-hybrid assays were performed as previously described [58] . | Papillomaviruses cause benign squamous epithelial tumors through the action of virally encoded oncoproteins termed E6 and E7 , which are classified as either high or low-risk based upon the propensity of the tumor to evolve into cancer . E6 proteins from both high and low-risk HPVs interact with a cellular ubiquitin ligase called E6AP . High-risk E6 proteins hijack E6AP ubiquitin ligase activity to target p53 for degradation . Degradation targets of the low-risk E6 proteins in complex with E6AP have not been described . Here , we describe a protein called NHERF1 that is targeted for degradation by both high and low-risk E6 proteins , as well as E6 proteins from diverse animal species . Degradation of NHERF1 resulted in activation of an oncogenic cellular signaling pathway called Wnt . Identification of NHERF1 as a highly conserved E6 degradation target could inform therapies directed against both low-risk HPVs and cancer-inducing high-risk HPVs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [] | 2019 | E6 proteins from high-risk HPV, low-risk HPV, and animal papillomaviruses activate the Wnt/β-catenin pathway through E6AP-dependent degradation of NHERF1 |
In Taiwan , around 1 , 500 cases of dengue fever are reported annually and incidence has been increasing over time . A national web-based Notifiable Diseases Surveillance System ( NDSS ) has been in operation since 1997 to monitor incidence and trends and support case and outbreak management . We present the findings of an evaluation of the NDSS to ascertain the extent to which dengue fever surveillance objectives are being achieved . We extracted the NDSS data on all laboratory-confirmed dengue fever cases reported during 1 January 2010 to 31 December 2012 to assess and describe key system attributes based on the Centers for Disease Control and Prevention surveillance evaluation guidelines . The system’s structure and processes were delineated and operational staff interviewed using a semi-structured questionnaire . Crude and age-adjusted incidence rates were calculated and key demographic variables were summarised to describe reporting activity . Data completeness and validity were described across several variables . Of 5 , 072 laboratory-confirmed dengue fever cases reported during 2010–2012 , 4 , 740 ( 93% ) were reported during July to December . The system was judged to be simple due to its minimal reporting steps . Data collected on key variables were correctly formatted and usable in > 90% of cases , demonstrating good data completeness and validity . The information collected was considered relevant by users with high acceptability . Adherence to guidelines for 24-hour reporting was 99% . Of 720 cases ( 14% ) recorded as travel-related , 111 ( 15% ) had an onset >14 days after return , highlighting the potential for misclassification . Information on hospitalization was missing for 22% of cases . The calculated PVP was 43% . The NDSS for dengue fever surveillance is a robust , well maintained and acceptable system that supports the collection of complete and valid data needed to achieve the surveillance objectives . The simplicity of the system engenders compliance leading to timely and accurate reporting . Completeness of hospitalization information could be further improved to allow assessment of severity of illness . To minimize misclassification , an algorithm to accurately classify travel cases should be established .
Dengue fever is an acute mosquito-borne viral infection caused by the dengue virus ( DENV ) that imposes considerable morbidity globally , particularly in the tropics and subtropics [1–3] . DENV is transmitted by several mosquito species within the genus Aedes , principally A . aegypti although outbreaks have been also attributed to the less efficient vector A . albopictus . In non-endemic settings , imported cases are often reported and the presence of a viable vector in the environment can present a serious threat of autochthonous transmission . Although dengue fever is non-endemic in Taiwan , the vector is abundant and the rising incidence of dengue fever in Southeast Asia in recent years has led to an increase in indigenous cases in Taiwan [4] . Annually , around 1500 cases of dengue fever are reported and there has been an increase in the occurrence of annual epidemics since the 1990s and early 2000s [1] . Transmission occurs as intermittent epidemics , sometimes with intervals of several years , with the main focus of activity in southern Taiwan which is more conducive to dengue outbreaks because of its tropical climate and the presence of A . aegypti [5] . The Notifiable Diseases Surveillance System ( NDSS ) has been in operation in Taiwan since 1997 and provides a national web-based platform for reporting and monitoring notifiable diseases , including dengue fever . The objectives of dengue fever surveillance include monitoring trends in incidence and supporting case management , although it may also facilitate outbreak detection . Before 2009 , Taiwan Centers for Disease Control ( TCDC ) used only a clinical case definition for reporting dengue . Many endemic countries have relied heavily on clinical case definitions which may lead to overreporting of cases due to a wide spectrum of clinical manifestations and other circulating viruses [6] . TCDC’s current dengue case definition has been established since 2009 , requiring laboratory confirmation for reported cases , which has been shown in other systems to improve specificity [6 , 7] . Other changes made to the NDSS to optimize the surveillance objectives have included the ability to link complementary datasets , such as laboratory , vector surveillance , and geographic information system data to improve monitoring and early detection of transmission . The integration of multiple surveillance systems has been demonstrated to reduce dengue spread and associated morbidity [7–8] . Dengue fever surveillance through the NDSS was last evaluated by the TCDC in 2009 prior to the introduction of these changes , and showed good timeliness and a predictive value positive ( PVP ) of 86 . 2% ( personal communication , Yi-Chun Lo , TCDC ) . As part of quality improvement activities , regular evaluation of the system is needed to ensure that the stated objectives are being achieved . We aimed to assess and describe key attributes of dengue fever surveillance through the NDSS to ascertain the extent to which stated objectives are achieved and to make recommendations for improvements .
Dengue fever has been a notifiable disease in Taiwan since June 1988 . A suspected case is defined as fever ( >38°C ) with at least one of the following symptoms: retro-orbital pain , myalgia , arthralgia , rash , leukopenia , and haemorrhagic manifestations . A confirmed case is defined as positive DENV identification by virus isolation , nucleic acid testing , non-structural protein-1 ( NS-1 ) antigen testing , or seroconversion [9] . The Communicable Diseases Act in Taiwan makes it a statutory requirement that suspected or confirmed cases of dengue fever be reported via the NDSS to the relevant local health department and TCDC by the diagnosing clinician or laboratory , within 24 hours . The local public health department should initiate patient interviews within 24 hours of receipt of the report to support case finding , contact tracing , and environmental control measures if deemed necessary . In addition to notifications from physicians , the NDSS also captures data on cases detected through contact investigation , self-reporting , and active surveillance at quarantine stations . Public health professionals actively investigate persons in close contact with a laboratory-confirmed case for symptoms consistent with dengue fever and submit blood samples to TCDC for confirmatory testing . A self-reporting system allows patients with symptoms consistent with dengue fever to present to the local health department for DENV confirmatory testing with a financial incentive of 2 , 500 new Taiwanese dollars ( approximately USD $80 ) if test results are positive . Since 2006 , a parallel active surveillance system has been in operation to detect febrile inbound passengers by using remote-sensing infrared thermography at quarantine stations of all international airports and harbors . Passengers presenting with fever upon entry from an endemic country are tested for dengue on the spot using an NS-1 antigen test which is sent to TCDC for laboratory confirmation . Laboratory-confirmed cases are automatically reported into the NDSS from the laboratory system feed . Fig . 1 shows the data flow between components of the NDSS . The surveillance data are used to produce a suite of routine and ad-hoc outputs including annual reports , web-based tables and maps , and the Taiwan Epidemiology Bulletin which is disseminated to a range of stakeholders . The NDSS is operated and maintained by two full-time information specialists , with a back-up computer system located at a separate site . The annual maintenance cost including staff , hardware , and software is approximately two million New Taiwan Dollars and is funded by TCDC . As the NDSS is a universal platform for all notifiable diseases , we cannot determine the proportion of cost attributed to dengue alone . The evaluation design followed the updated guidelines for Evaluating Public Health Surveillance Systems published by the United States Centers for Disease Control and Prevention [10] . Using the stated objectives of the dengue fever surveillance system , appropriate attributes were selected based on the relevance to dengue fever surveillance and the ability to access corroborating information from other systems ( e . g . laboratory data ) . The selected attributes for the evaluation were simplicity , acceptability , data completeness , quality and validity , timeliness , and PVP . Sensitivity , specificity and representativeness were not assessed as time and resource constraints precluded the possibility of obtaining supplementary data from another system and they have been examined in previous evaluations . Data on all confirmed dengue fever cases reported through the NDSS between 1 January 2010 and 31 December 2012 were anonymized and extracted by TCDC staff . These were translated into English and collated onto a password protected Microsoft Excel database . Reporting activity was explored to describe the epidemiology of cases in Taiwan from 2009–2011 . Annual crude and age-specific incidence rates were calculated using age specific census mid-year estimates for the years 2009 to 2011 [11] . Categorical variables were summarised as counts and proportions and continuous variables presented using appropriate measures of central tendency and variation . Exact score mid-p values were ascertained from chi square test for categorical variables and chi square test for trend for continuous data , with a p value of ≤ 0 . 05 considered statistically significant . All statistical analyses were performed in Stata 12 . 1 ( StataCorp . College Station , USA ) . A 17 item semi-structured questionnaire was used to assess respondents’ knowledge , understanding and views on the current case definitions and objectives of NDSS as well as their views on the relevance , acceptability and ease of use of the system . The questionnaire was administered via face-to-face interviews with internal stakeholders considered key to the operation of the surveillance system . Interviewees were selected based on convenience ( availability and willingness to participate ) and expertise ( minimum of three years’ experience ) and interviewed by a TCDC staff member in Mandarin Chinese with the responses recorded in the questionnaire in English . Simplicity and acceptability were assessed using questions concerning compliance , ease of use , and number of steps in the system alongside users’ opinions on the appropriateness of variables collected and the current case definition . The responses from the interviewees were analysed and summarized in the following way: ( 1 ) the number and percentage of respondents providing a ‘Yes’ or ‘No’ response to questions with dichotomous responses; ( 2 ) the number and percentage of respondents providing a particular response to questions with scaled ( nominal ) options; and ( 3 ) questions with free text responses in a narrative style without any thematic coding . Completed questionnaires were transcribed into SelectSurvey ( SelectSurvey . Net; 2012 ) and exported to Microsoft Excel for analysis . Completeness was assessed by determining the percentage of case records with recorded data on each variable . Data quality was evaluated by assessing the completeness , validity , and chronological consistency of variables included in the World Health Organization ( WHO ) recommended minimum data set for dengue fever surveillance [12] . These variables consisted of case classification , unique identifier , patient name , age , sex , geographic information , date of onset , hospitalization , outcome , and travel history during the past two weeks . Validity was assessed by determining the percentage of case records with valid recorded data on each variable . The validity of travel categorization was assessed by ascertaining the number of cases recorded as travel-related on their patient record . Travel dates were then checked and cases categorized as travel-related using the system definition of “travel outside Taiwan within 3−14 days before illness onset” . Any patient that reported travel outside this window was deemed non-travel related . This was checked against what was recorded in the travel-related field . Chronological consistency was assessed by determining the percentage of case records in which there was consistency between reported date of onset and date of diagnosis . Guidelines were scare on what represents a good or acceptable level of data completeness or validity among individual fields in surveillance systems . Evaluation is often based on perceived importance of the field in question [13 , 14] . Using a threshold value of 90% as a satisfactory level of completeness/validity , we reported results as “satisfactory” or “unsatisfactory” if ≥ 90% or < 90% respectively of case records met the threshold value [15] . Timeliness was assessed by calculating two metrics; time that elapsed between onset of symptoms to diagnosis and from diagnosis to report . These were summarised as the number of days between these two time points with an appropriate measures of central tendency and variation . The percentage of cases not reported to NDSS within the recommended 24 hours was also calculated . PVP was assessed by calculating the percentage of all tested cases that were laboratory confirmed . Data reported to the NDSS are used for public health surveillance purposes . TCDC determined this study as non-research activity and therefore exempt from the review by the Institutional Review Board at the time it was undertaken ( 2012−2013 ) . All patient data were anonymized and verbal informed consent was obtained from all interviewees .
During 2010−2012 , 11 , 718 suspected dengue fever cases were reported to the NDSS; 5 , 072 were subsequently laboratory confirmed , with annual counts of confirmed cases decreasing over the period ( 1 , 897 in 2010 , 1 , 696 in 2011 and 1 , 479 in 2012 ) . The crude incidence decreased from 8 . 2/100 , 000 in 2010 to 6 . 4/100 , 000 in 2012 . The majority ( n = 4 , 740 , 93% ) of confirmed cases occurred during July to December ( Fig . 2 ) . Of the 5 , 072 laboratory-confirmed cases , the median age was 45 years ( range <1 to 98 years ) and 2 , 530 ( 49% ) were male ( Fig . 3 ) . During the period , 720 ( 14% ) of the 5 , 072 laboratory-confirmed cases were reported as travel-related and the proportion of travel-related cases did not change significantly annually ( p = 0 . 6 ) . Of the 720 travel-related cases , 111 ( 15% ) reported symptom onset more than 14 days after travel , which did not technically fit the definition of travel-related . Of travel-related cases , 57% were reported during July to October whereas the majority ( 91% ) of non-travel related cases were reported during September to December ( Fig . 4 and Fig . 5 ) . Four individuals participated in the face-to-face interviews; two each from the north and south of the country ( three TCDC staff and a member of the local health department ) . Their job roles were a public health nurse , an infectious disease doctor , an environmental prevention and control officer , and an information technology staff member . All respondents agreed that the current system was effective or very effective in meeting the objectives of dengue fever surveillance and felt that the variables collected were relevant to dengue fever specifically . All respondents agreed that the dengue surveillance system was clear and easy to use; useful , quick , and that the skills required to use the system existed in their teams . A mixture of uses of the system were reported including case reporting ( 75% ) , case monitoring ( 75% ) , data analysis ( 50% ) , and tracking outbreaks ( 50% ) . In addition , two respondents also reported using the system for advice on case management and to identify areas to target for vector surveillance . Free text comments from respondents revealed that some users felt the case definition was broad and in general , cases were more likely to be diagnosed based on clinicians’ judgment and experience . However , it was accepted that the severity of disease requires a sensitive case definition in order to capture cases . Table 1 summarized completeness of variables reported on confirmed dengue fever cases in the NDSS . The variable “hospitalized” was the only variable with an unsatisfactory level of completeness with 78 . 5% of records completed . All cases had both an onset date and diagnosis date recorded with no inappropriate values reported ( i . e . wrong format or nonsensical values ) . All recorded onset dates were chronologically consistent with the diagnosis date . All cases had a unique identifier reported . Overall , 99 . 9% ( 4810/4817 ) of cases with Taiwanese residency had a correctly structured unique identifier . The median length of time from symptom onset to presentation for clinical assessment/diagnosis was four days ( range 0–49 ) , and 4 , 556 ( 90% ) of cases were diagnosed within seven days of symptom onset . The elapse between the date of diagnosis to the report date was 0−17 days , with 5 , 009 ( 98 . 8% ) reported within the recommended 24 hours . Over the period , 11 , 796 suspected dengue fever cases were reported to the NDSS with 5 , 072 subsequently laboratory confirmed . This represents a calculated PVP of 43 . 0% ( Table 2 ) .
Our evaluation of dengue fever surveillance in Taiwan shows that the system is simple , acceptable to users and achieves its stated objectives . The high level of completeness of key data fields is encouraging and demonstrates the high quality of the surveillance system but is also an indirect measure of the engagement of stakeholders particularly those who report to the system . The only variable with low reporting was hospitalization information , but this is not a mandatory field and data entry still proceeds even if this is not recorded . Improving the completeness of this variable would be welcomed as it may allow more accurate estimates of severity of illness . The simplicity of this system and its processes engenders compliance which can facilitate the delivery of effective public health responses and improve outcomes at the individual and population levels [15] . In 2009 , changes were made to the NDSS to optimize the surveillance objectives which increased co-operation with other public health services such as laboratory reporting and entomological colleagues in order improve detection of early transmission; and a change in case definition to increase case specificity . These changes have been supported by users of the system and are reflected in the evaluation findings that show stakeholders are engaged and compliant . Engagement can be strengthened by incorporating other forms of surveillance and promote working towards a more integrated response [16–17] . We could not assess the outbreak detection capability of the system but we recommend that future evaluations include this to support the development of statistical algorithms to detect space-time clusters using all available linked data . The sensitivity of such as system may be improved by combining outbreak detection with virologic surveillance [18] . WHO recommend that effective prevention and control of epidemic dengue requires an integrated laboratory based surveillance system that can provide early warning of epidemic transmission [19] . Systems with standardised procedures for laboratory confirmation may improve the ability to distinguish between true cases and other causes of febrile illness . Our finding that under half ( 43 . 0% ) of reported cases tested had laboratory confirmed dengue fever represents a considerable decrease from the PVP of 86 . 2% observed during 2004–2008 ( personal communication , Yi-Chun Lo , TCDC ) . This may be due to the change in the confirmed case definition introduced in 2009 and the low threshold for reporting suspected cases by clinicians . Before 2009 , an individual with dengue-like symptoms who was epidemiologically linked to a confirmed case was classified as a confirmed case without the need for further laboratory confirmation . This may have led to an overestimation of the numerator and an artificially inflated estimate of PVP prior to 2009 . In addition , PVP may fluctuate during periods of high incidence and the evaluation of other dengue systems have reported that even with increasing PVP there is still a fairly high level of false-positive returns , probably due to sensitive case definitions and a broad spectrum of illness , especially during outbreaks [6] . The calculation and assessment of PVP is unlikely to be robust without incorporating estimates of system sensitivity , which we were unable to undertake in this study due to lack of access to a ‘gold-standard’ data source and so an accurate assessment of this is a key recommendation for further work . We found that a small number of travel cases were misclassified . To minimize this , we recommend the use of an algorithm to more accurately classify travel cases in order to improve accuracy and ensure adherence to WHO recommended standards for dengue fever [20] . Accurate estimates of imported and autochthonous cases are needed to allow comparisons of the burden of dengue in different geographical locations and time periods and support public health authorities to make informed decisions on resource allocations [21] . There is a possibility of selection bias arising from the convenience sample of stakeholders that were interviewed and this may preclude any generalisation to other key stakeholders . In addition the sample was small and not necessarily representative of the views of the majority of stakeholders . Nonetheless , the views expressed by the respondents are still useful and provided some insight into key aspects of the operation of the system . Our finding of a predominance of adult cases is consistent with findings from other parts of the region [22–25] . Age is known to be an important predictor of risk and outcome of dengue infection as adults have been found to be at greater risk of hospitalization and death [26] . A number of factors may underpin this observed age distribution including the demographic of travel-related cases and milder symptoms in children . Our study has identified a number of key recommendations including a further assessment of sensitivity and PVP , ability of all parts of the system to accurately capture and track outbreaks and efforts to improve data quality and completeness to provide a clearer picture of severity and actual burden of imported cases . These are consistent with the recommendations from a facilitated discussion involving members of the Dengue Prevention Board in the Asia-Pacific and America that include: collection of a minimum data set; additional studies to verify the sensitivity of the system; data sharing with laboratories; and goals for national surveillance systems to include early detection and prediction of dengue outbreaks [15] . This study had some limitations due to lack of access to data which meant that we were unable to undertake an assessment of estimates of system sensitivity . Also , the evaluation of simplicity and acceptability doesn’t cover cases’ perception and use of the reporting systems , only public health staff . Staff members were interviewed based on convenience due to availability , English language ability , and willingness . This could bias the sample and led to small numbers available for interview , leading to a less representative sample than is usually required . Issues with providing data to non-Taiwan CDC staff meant that the available dataset was only available for anonymous variables , and only case information; we were unable to investigate any linked data such as vector information or serological results . In conclusion , our study shows that comprehensive dengue fever surveillance is necessary for establishing a clear picture of the local burden and distribution of disease in populations , in order to effectively support public health response . The NDSS is a robust and valid integrated system which supports these efforts and assists with the reporting of a clear picture of the epidemiology of dengue fever in Taiwan . Integrated surveillance and outbreak preparedness remain central enabling factors that highly contribute to effective implementation of the global strategy to reduce the burden of dengue infection worldwide and further efforts to improve these should be on-going [19] . | In Taiwan , around 1 , 500 cases of dengue fever are reported annually . Surveillance and outbreak preparedness are important activities aimed at reducing the burden of dengue fever worldwide . A national web-based Notifiable Diseases Surveillance System ( NDSS ) for dengue fever has been established since 1997 in Taiwan to monitor trends , and support case and outbreak management . We evaluated this surveillance system based on the Centers for Disease Control and Prevention guidelines and operational staff interviews to ascertain the extent to which stated objectives are being achieved . The results indicate that the NDSS for dengue fever surveillance works well , is easy to use , and can be used to accurately identify cases . The findings have informed recommendations to improve surveillance and its use in outbreak response , including efforts to improve data completeness . Comprehensive surveillance exemplified by the NDSS for dengue fever in Taiwan is necessary for establishing a clear picture of the local burden and distribution of dengue fever to effectively support public health response . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Evaluation of the National Notifiable Diseases Surveillance System for Dengue Fever in Taiwan, 2010–2012 |
By means of a computer search for upstream promoter elements ( distal sequence element and proximal sequence element ) typical of small nuclear RNA genes , we have identified in the human genome a number of previously unrecognized , putative transcription units whose predicted products are novel noncoding RNAs with homology to protein-coding genes . By elucidating the function of one of them , we provide evidence for the existence of a sense/antisense-based gene-regulation network where part of the polymerase III transcriptome could control its polymerase II counterpart .
Recent advances in mammalian genome studies are bringing to light the occurrence of a widespread transcription of noncoding ( nc ) regions devoted to the regulation of the protein coding genome expression [1–4] . The mechanisms of action of these transcripts are various and of different natures , although all of them are devoted to the regulation of fundamental genetic pathways involved in the determination of the cell phenotype . The concomitant evolution of noncoding regulatory transcripts and proteins that target different RNA:RNA or RNA:DNA complexes emphasizes the importance of studying the regulatory processes mediated by nucleic acid interactions . It has been demonstrated that in both prokaryotes and eukaryotes , cis-acting RNA regulatory regions ( i . e . , 3′ UTRs forming secondary structures with regulatory features or containing sequences recognized by proteins involved in RNA stability modulation ) can be simultaneously regulated in trans by other noncoding RNAs ( i . e . , microRNAs ) or by protein complexes . The simultaneous occurrence of cis- and trans-regulatory elements brings to light the complexity of this network where the coexistence of different ncRNAs plays a key role in the control of other target gene expression [5] . In this context a prominent role is played by the enlarging family of microRNAs ( miRNAs ) that act at a posttranscriptional level by inhibiting the translation of protein-coding genes [6] . The known miRNAs , similar to protein-coding mRNAs , are synthesized as polyadenylated precursor molecules by the RNA polymerase ( pol ) II transcription machinery [7] . The vast majority of the tools used in molecular biology are based on transcript collections obtained by oligo ( dT ) reverse transcriptase ( RT ) -PCR , thus encompassing only polyadenylated pol II products . However , a wide contribution of nonpolyadenylated transcripts to the human transcriptome has been revealed [8] . The role of such transcripts in pol II transcriptome–expression regulation remains largely unexplored . Among the noncoding elements , one of the most investigated has been the Alu class of repetitive sequences that represents about one-tenth of the whole human genome . Although it is not yet possible to discern a peculiar role of Alu sequences , their short transcripts have been shown to be involved in several biological processes such as: RNA editing ( where Alus are preferential sites for A to I RNA editing thus having profound implications both in gene-expression regulation and in mammalian genome evolution ) [9] , alternative splicing ( internal exons that contain an Alu sequence are almost always alternatively spliced ) [10] , chromosomal recombination ( the recombination between Alu elements is at the base of genomic deletions associated with many human genetic disorders ) [11] , gene-expression regulation ( functioning as naturally occurring antisense RNAs ) [12] , cell-stress response ( such as heat-shock response and/or translation inhibition ) [13] , and as putative miRNAs targets [14] . The physiological role of Alus and all the other 7SL-derived transcripts needs to be studied in more detail . In particular , the fact that their transcription is RNA pol III–dependent brings to light a previously unexpected role in gene-expression regulation of this enzyme that deserves investigation . In this study , starting from the observation that pol III is specialized in transcription of ncRNA genes , we addressed the hypothesis that the human genome might contain pol III transcription units each specifically regulating one ( or more ) specific pol II genes , thus constituting functional “cogene”/gene pairs .
To test our hypothesis we focused on pol III type 3–extragenic promoters , which are located upstream of the transcribed region . We screened the human genome for regions containing the consensus sequences characteristic of pol III type 3 promoters: the proximal sequence element ( PSE ) and the distal sequence element ( DSE ) [15 , 16] . First , we tested the PSE sequences of three well-characterized pol III type 3–ncRNA genes ( U6 , H1 , and 7SK ) as query sequences for the search of similar ( if not equal ) elements in the human genome by using the BLAST algorithm ( http://www . ncbi . nlm . nih . gov/BLAST; under “Nucleotide” subsection select “short , nearly exact matches” option , then pull down “Homo sapiens” organism database ) ( see Materials and Methods for sequences used as query ) . While the search with U6 and 7SK PSE sequences did not identify a significant number of homologous regions scattered throughout the genome , the H1 PSE element shared a high homology with 60 novel putative PSE sequences . Among these we selected ( by a BLAST analysis ) those that contained a DSE sequence element within an arbitrarily defined distance of 1 , 000 basepairs upstream the PSE . In addition to the expected H1 , results evidenced 31 novel putative PSE/DSE-dependent promoters characterized by the concomitant occurrence of the PSE and the DSE sequences within that defined genomic distance . Moreover , a detailed sequence analysis showed that the vast majority of the distances between the PSE and a downstream TATA box or TATA-like element are within 18–22 basepairs , as expected for a canonical type 3 promoter [16] . Altogether these observations were taken as preliminary indication of a functional pol III type 3 structure of these novel promoters ( Table S1 ) . Since our search was based on pol III type 3 promoters , some additional features of these promoters needed to be considered: ( i ) the occurrence of a PSE consensus sequence does not identify per se a pol III type 3 promoter; that is , rather , the result of the simultaneous occurrence at an appropriate distance of the PSE and an A/T-rich ( TATA-like ) element . Indeed , it has been clearly shown that the occurrence of a PSE consensus that lacks a downstream A/T-rich element makes the promoter readable by RNA pol II , such as in the case of the U2 snRNA gene [17] . In this context , the transcription initiation region is not relevant for the choice of the RNA polymerase , at least in humans , although it seems to be of fundamental importance in Xenopus [18] . Therefore , the PSE/DSE-flanked , putative transcription units identified by our search might in theory be transcribed either by pol II or by pol III , depending on the occurrence of a functional A/T-rich region downstream of the PSE . ( ii ) Pol III transcription units are characterized by the presence of very simple and easily recognizable termination signals , consisting in a run of four or more consecutive T residues . We thus searched , within the collection of DSE/PSE-containing sequences , for the further occurrence of a TATA-like element downstream of the PSE and for the occurrence of a termination signal at a significant distance downstream of the hypothetical transcription start site , assumed to be located approximately 30 basepairs downstream of the TATA element . Such a search refinement revealed that most of the newly identified sequences have features compatible with a pol III type 3 promoter structure ( Table S2 ) . Our in silico search was based on the H1 PSE sequence , which was used as a query only allowing for mismatches in the first and last positions . The search thus likely identified only those promoters whose structure is very similar to that of H1 . This is supported by the fact that out of H1 no other previously known pol III type 3 promoters were found in our search . Given the divergence among the PSE sequences of U6 , 7SK , and H1 , it is to be expected that the use , as a query sequence , of a more degenerated PSE consensus , derived from the known pol III type 3 PSE sequences , would bring to light a considerably higher number of putative PSE-dependent transcription units in the human genome . To further characterize in silico the novel transcription units , we arbitrarily assumed as transcribed the region starting from the 30th nucleotide downstream of the first nucleotide of the predicted TATA box . In addition , a run of at least four T residues was considered as a pol III transcription–termination signal , although events of “read-through” are possible at T4 sequences depending on sequence context features [19 , 20] . While it has to be emphasized that the transcribed region of each element of this collection needs to be experimentally determined , we selected 32 putative novel transcripts to be subjected to additional analysis on the basis of their in silico characterization . To test if a common secondary structure could be a hallmark of the novel molecules , an in silico analysis of their secondary structure was performed by mfold algorithm ( http://www . bioinfo . rpi . edu/applications/mfold/rna/form1 . cgi ) [21] . Results showed that although hairpins with short stems ( 5–7 basepairs ) were frequent , no shared secondary structures were recurrent , indicating that a peculiar molecular organization is not the common hallmark of this set of noncoding molecules . Although their averaged free energy ( ΔG ) was extremely variable ( −42 . 7 ± 41 . 2 ) , a group of four transcripts ( 11A , 20A , 21A , and 29A ) showed a ΔG value significantly lower than all the others ( ΔG < −100 ) . A statistical analysis of such ΔG differences was performed evidencing that the differences between this group of transcripts ( 11A , 20A , 21A , and 29A ) , and the rest of the pool is highly significant ( Student's t-test , 33 degrees of freedom , α significance level = 0 . 1 corresponding to a p-value of 0 . 0001 ) , thus keeping in line with their physiologically functional–secondary structure organization ( Table S3 ) . In order to assess if the pool of transcription units was prevalently constituted by repeats such as retroposons , we analyzed their sequences by Repeat Masker algorithm ( http://repeatmasker . org ) evidencing that: ( i ) only two out of 32 ( 6 . 2% ) are short interspersed nucleotide elements ( 21A and 29A , which were marked as AluJb elements ) ; ( ii ) three of them ( 24A , 37A , and 38A ) are part of long interspersed nucleotide elements; ( iii ) two sequences ( 17A and 40A ) contained a mammalian interspersed repeat ( MIR ) , and ( iv ) three sequences ( 30A , 32A , and 44A ) contained different types of long terminal repeats ( Table S4 ) . Considering that Alus , long interspersed nucleotide elements , and MIRs constitute about 15% , 30% , and 1%–5% of the human genome , respectively , one would have expected a higher frequency of repeats in the pool of sequences . Altogether these observations provide evidence that the novel PSE-dependent transcripts are not associated to a specific class of repetitive sequences scattered throughout the human genome , but instead they constitute a novel heterogeneous set of type 3 promoter-driven elements . When these noncoding sequences were used to challenge the human genome database ( BLAST analysis ) , it was found that seven of them were internal to known or predicted protein-coding genes , four being in antisense and three in sense configuration . Most of the novel sequence elements not mapping in coding regions shared a high sequence homology ( ∼80% ) to a pol II transcript/expressed sequence tag that maps in a different locus ( Table S5 ) . Such homologies reached even higher values ( up to 90% ) , if only parts of the putative transcripts were considered . In fact , no expressed sequence tags entirely containing one of our transcription units were found , so that if a sense/antisense-based regulation would occur , it would likely be related to parts of the ncRNA sequences , while the other part could have structural properties that facilitate this regulatory action ( perhaps by binding specific structural proteins ) . Based on these observations , a novel control mechanism of gene expression could be postulated where pol III ( or pol III-like ) elements act as trans-locus antisense of their homologous protein-coding RNAs . In this model , the pol III cogenes in antisense configuration with respect to one ( or more ) specific target gene ( s ) could regulate their expression either by interfering with its mRNA maturation ( if the homologous region is internal to an intron ) or by inhibiting protein translation ( if the homology is associated to an exon ) . To test our hypothesis we selected one of the novel transcription units ( here referred to as 21A ) that maps in 8q24 . 13 . If aligned to the human genome , it shows several homology hits among which the most highly significant were associated to multiple intronic regions of centromeric protein F gene ( CENP-F; 1q32-q41 ) [22] , thus constituting its putative natural trans-chromosomal antisense ( Figure 1A–1C ) . Although , similarly to all of the 7SL/Alu-derived elements , 21A is expected to be primate-specific [23] , an evolutionary conservation analysis was performed aligning its sequence with the mouse-predicted CENP-F gene . No significant similarities were found indicating that in rodents a putative CENP-F antisense regulatory role , if any , would be associated to a different class of noncoding elements . Despite its high sequence similarity with other human Alus , 21A lacks the Alu-specific intragenic consensus elements needed to promote its pol III transcription such as the blocks A and B [24] . This observation further pointed to a 21A transcription driven by an extragenic type 3 pol III promoter . To check for 21A expression in cultured cells , we performed Northern blot analysis on total HeLa cell RNA and skin fibroblast RNA using a 21A dsDNA probe . Two positive bands were detected: one corresponding in size to the expected 21A transcript ( ∼300 nucleotides ) , and the other one corresponding to a high molecular mass transcript ( as expected for CENP-F mRNA ) ( Figure 2A ) . However , considering that the 21A double-strand cDNA probe would detect transcription of 21A-similar Alus from multiple loci , we also amplified a 21A-specific cDNA from total RNA samples , extracted from skin fibroblasts and four tumor cell lines ( 293T , LAN5 , HCT , and HeLa ) , by random hexamer-based RT-PCR in order to better identify a 21A-specific transcription product ( Figure 2B ) . The DNA band obtained was then purified and sequenced , evidencing that the amplification product was the expected 21A . In addition , to better assess 21A transcription , we fused its promoter region to a luciferase silencer hairpin and cotransfected this construct with a plasmid-expressing luciferase . Results showed a halved luciferase activity 48 h after transfection , thus demonstrating an efficient transcription directed by 21A promoter . In the same experiment , a set of five novel promoters from our collection was tested , demonstrating an active transcription of the hairpin promoted by four of them ( Figure 2C ) . These data support the conclusion that the majority of the novel putative transcription units is under the control of active extragenic PSE/TATA-containing promoters . The same experiment as above was repeated after 24 h of cell treatment with ML-60218 , a cell-permeable indazolo-sulfonamide compound that displays broad-spectrum inhibitory activity against pol III [25] . Results showed an efficient luciferase-silencing activity in the absence of the pol III inhibitor ( as evidenced by a decreased luciferase emission ) , while after treatment with ML-60218 the luciferase signal was increased ( Figure 2D ) . As a control , a similar experiment was performed by treating the cells for 12 h with a pol II inhibitor ( α-amanitin ) . In this case , no major transcription variations were observed either for the novel transcript units or for the well-known pol III-dependent H1 gene ( Figure 2E ) . To better assess pol III dependence of 21A transcription , we directly measured the endogenous 21A ( and 29A ) RNA amount in ML-60218 treated cells versus untreated control samples . Results evidenced a significant transcription downregulation of the two novel pol III units ( 50% and 30% inhibition in the case of 21A and 29A , respectively ) , thus keeping in line with the occurrence of a specific effect of ML-60218 on their promoters . To further assess the specificity of action of the two inhibitors , we analyzed by real-time RT-PCR the transcription activity of two pol III ( 5S rRNA and 7SK ) and two pol II-dependent genes ( c-Myc and glyceraldehyde 3 phosphate dehydrogenase [GAPDH] ) both in ML-60218 and in α-amanitin-treated cells . Results showed a significant inhibition of the pol III–dependent genes after ML-60218 treatment and a stable transcription level of the pol II–dependent genes in the same samples ( Figure 2F ) . On the contrary , pol III transcription activity was stable in α-amanitin–treated cells , while in the same samples the pol II–transcribed genes were downregulated , thus demonstrating the specificity of action of the two inhibitors in these experimental conditions ( Figure 2G ) . Altogether , these results provide evidence that the novel PSE/TATA-containing transcription units are transcribed by pol III . To test whether the 21A transcript acts as an antisense inhibitor of CENP-F expression , we measured by Western analysis CENP-F protein level in HeLa cells transiently transfected with four different 21A constructs carrying: ( i ) the whole 21A region containing both DSE and PSE elements ( p21A ) ; ( ii ) its upstream moiety , which contains the DSE and a MIR element , but not the CENP-F homology region ( p21A-1 ) ; ( iii ) the novel pol III type 3–transcription region ( which includes an Alu Jb module ) ( p21A-2 ) ; and ( iv ) an empty vector as mock control ( pMock ) . As shown in Figure 3 , starting at 24 h from transfection of the whole 21A region , inhibition of CENP-F accumulation ( followed by a rapid degradation ) was observed . Such inhibition was specifically associated to constructs expressing the 21A RNA ( p21A , p21A-2 ) , while the MIR element in the upstream moiety of the fragment ( p21A-1 construct ) was ineffective ( Figure 3A–3D ) . In this context , it has to be noted that a slight delay occurred in 21A-2 inhibitory action ( and in the expression of 21A-specific RNA ) , compared to what was observed with the complete 21A construct ( more rapid increase in 21A RNA expression and decrease of CENP-F protein levels ) , suggesting a positive transcriptional role of the DSE element . The actual occurrence of 21A transcription in transfected cells was analyzed by real-time quantitative RT-PCR . As expected , a very high amount of 21A transcript was detected in p21A and p21A-2–transfected cells ( 210- and 480-fold , respectively , at 48 h from transfection ) , while the 21A RNA content of samples transfected with pMock control plasmid and/or with a construct containing the promoter lacking the transcribed region ( p21A-1 construct ) were essentially stable , showing a very low basal level of 21A expression in untransfected HeLa cells ( Figure 3I–3N ) . All the PCR products were analyzed in their dissociation curve , showing a single characteristic pick ( at 78–79 °C ) in p21A/p21A2-transfected samples significantly reduced in pMOCK/p21A-1 . On the contrary , the cells transfected with the two control plasmids ( pMock/p21A-1 ) showed a dissociation pattern characteristic of a heterogeneous population of molecules ( Figure 3O ) . Again these results confirmed an active synthesis of the exogenous 21A ncRNA transcript in p21A/p21A-2–transfected samples that was strongly reduced at a very low endogenous–basal level in the samples lacking the transcript region ( pMOCK/p21A-1 ) . As a consequence of 21A very active transcription , the level of CENP-F mRNA ( as determined by real-time RT-PCR ) was significantly decreased in p21A/p21A-2–transfected cells , while no major CENP-F mRNA variations were observed in pMOCK/p21A-1–transfected cells ( Figure 3E–3H ) . Altogether these results demonstrate an inverse correlation between 21A transcription and CENP-F expression . Therefore , considering the high sequence homology between 21A transcript and three CENP-F hnRNA intronic portions , we suggest a mechanism of antisense inhibition of CENP-F mRNA maturation by the 21A transcript . Given the central role of CENP-F in mitosis , we tested the effect of ectopic 21A expression on cell proliferation . By measuring [3H]-thymidine incorporation , we observed a dramatic arrest of cell proliferation after 48 h in 21A-transfected cells . Again , the effect was specifically associated to the downstream 21A transcribed region ( p21A/p21A-2 constructs ) , while transfection of the MIR-containing upstream moiety ( p21A-1 construct ) did not alter cell proliferation ( Figure 4A ) . Although at the present state we cannot exclude a contribution to this effect by Alus from other loci , this experiment demonstrates an inverse correlation of 21A transcription and cell proliferation that is in accord with the inhibition of CENP-F synthesis demonstrated above . To further support the antisense role of 21A , we transfected HeLa cells with a construct expressing the transcript in antisense configuration ( here referred to as pAnti-21A ) , thus quenching the activity of the endogenous 21A molecules . Results showed an increased cell proliferation 24–48 h after transfection . Similar results were obtained when a 21A-specific small interfering RNA ( siRNA ) –expressing construct was transfected in HeLa cells , while the negative control sample ( cells transfected with an unrelated chicken-specific siRNA ) maintained a cell-proliferation rate similar to that of pMock-transfected cells ( Figure 4B ) . In both of the experiments an increased CENP-F expression was detected both at the protein and mRNA levels ( Figure 4E and 4F ) . As evidenced by real-time RT-PCR in the same experiment , a concomitant 21A-RNA decrease was observed after 24 h of transfection in anti/si21A treated cells , although a complete recovery of 21A RNA synthesis occurred after 48 h of transfection ( Figure 4G–4H ) . As shown in these experiments CENP-F modulation and 21A RNA decrease were analyzed only at 0 , 24 , and 48 h after transfection , rather than at 0 , 24 , 48 , and 72 h as in the previous experiments . In fact , at 72 h after transfection the CENP-F synthesis determinations would be strongly biased by an early cell culture overconfluence caused by the proliferation increase that follows 21A downregulation . These data suggest that the decreased amount of 21A transcript consequent to its siRNA-mediated silencing , as well as its suppression by antisense technology , specifically increases CENP-F synthesis , thus keeping in line with the proposed role of 21A as CENP-F regulatory cogene . In addition , it has to be considered that the increased proliferation rate observed here supports the idea of a widespread regulatory action of 21A that may control at the posttranscriptional level the expression of several target genes similarly to what has been proposed for miRNAs [26] . Considering that a 21A-driven cell-proliferation inhibition is expected to be primate specific ( Alu sequences were not found in other mammalian orders ) , we tested for its possible occurrence in mouse . We found that , after transfection of p21A , p21A-1 , p21A-2 , and pMock , the murine fibroblast NIH 3T3 cells did not show any proliferation decrease as assessed by [3H]-thymidine incorporation ( Figure 5 ) . The species-specificity of 21A action , together with its inability to cause a nonspecific cell reaction that leads to a proliferative blockade in mice , further strengthens a 21A-specific ( perhaps multilocus ) regulatory role . In fact , considering these data we rule out a nonspecific effect of 21A on cell proliferation , perhaps due to the activation of a more general biological process , such as the interferon response ( an antiviral cell reaction shared by all mammals ) , rather than a specific multilocus 21A regulatory action . As demonstrated by transfection experiments , 21A overexpression is inversely correlated to cell proliferation . In accordance with this finding , its basal expression level is very low in fully proliferating HeLa cells . To better investigate the inverse correlation between the endogenous 21A expression and cell proliferation , we analyzed by quantitative real-time RT-PCR the 21A expression levels in cell types characterized by different proliferation potentials . Results showed that in three immortalized , fully proliferating cell lines analyzed here ( HeLa as cervical adenocarcinoma; 293T as renal epithelial adenovirus transformed cells; LAN5 as neuroblastoma ) , the level of 21A transcription was very low if compared to the unproliferating/resting PBL ( peripheral blood lymphocyte ) cells , in which a 276-fold-increased–21A transcription was evidenced . In the same experiment , according to an inverse correlation between endogenous 21A transcription and the cell proliferation rate , the 21A RNA level in primary skin fibroblasts ( of which the proliferation rate is significantly lower than that of the tumor cell lines analyzed here ) showed a 23-fold increase compared to 393 T cells , and a very low expression level if compared to the resting/unproliferating PBL ( Figure 6 ) . Again the dissociation curve analysis of 21A amplification product showed in PBL a peak at 78–79 °C , characteristic of a single specific molecular species that resembled the one obtained in 21A/21A-2 transfected cells ( where the amount of 21A transcripts was strongly increased ) , although a slight shoulder , most likely due to a cross-amplification of other very similar transcripts , revealed a detectable endogenous Alu transcription background ( Figure 6 ) . Altogether these results evidence a very active 21A transcription in PBL/resting cells that further strengthens the idea of 21A as a novel key factor of cell-proliferation control . In order to check if the endogenous 21A overexpression in unproliferating cells was related to a widespread increased RNA polymerase III activity rather than a 21A-specific activation , we measured by real-time RT-PCR the 5S rRNA expression level in the same samples . The results showed no direct correlation between 5S rRNA expression and the cell-proliferation rate variations , evidencing that the 21A overexpression in resting cells was the consequence of a 21A-specific transcription activation rather than a wider , nonspecific increase of pol III activity ( Figure 6 ) . These data thus suggest the existence of an unexpectedly specific expression regulation of 21A promoter ( related to the cell proliferation state ) that needs to be investigated in detail . We propose that the noncoding fraction of the human genome includes a larger than expected number of ncRNA genes controlled by DSE and PSE promoter elements . Due to their promoter structure , a number of these genes is likely to be transcribed by pol III . We refer to them as cogenes since they could specifically coact with a protein-coding pol II gene . Given the very high sequence homology between pol III and pol II transcript pairs , and in light of the results we have obtained investigating the regulatory activity of the 21A transcription unit , we propose that a large part of these novel elements may act as antisense inhibitors of protein translation and/or mRNA maturation , although some of them ( those whose homology with the pol II target gene is in sense configuration ) could play a role in gene-expression regulation with different mechanisms . Altogether these findings provide evidence for the existence of an ncRNA gene set associated to PSE/DSE-containing promoters , whose products coact with a corresponding set of protein-coding targets . In conclusion , this study provides: ( i ) a collection of novel noncoding transcripts to be investigated for their potential regulatory action with respect to pol II target genes; ( ii ) a novel source of PSE-dependent promoters useful for the identification of common regulatory regions specific for this type of promoters; ( iii ) a novel class of molecules involved in the RNA-dependent gene expression regulation; and ( iv ) a novel transcript ( 21A ) , whose role in tumor cell proliferation deserves further investigation in the context of cancer studies .
All of the sequence searches and alignments were carried out by means of BLAST at the National Center for Biotechnology Information ( http://www . ncbi . nlm . nih . gov/BLAST ) . The sequences used to query were the following: H1 PSE , nCACCATAAAnGTGAAAn or nTTTCACnTTTATGGTGn; U6 PSE , CTTACCGTAACTTGAAAGT; 7SK PSE ( as reported in PMID: 2011518 ) , TTGACC-TAAGTG; DSE ( Octamer-Binding Transcription Factor 1 [Oct1] consensus sequence ) , ATTTGCAT or ATGCAAAT with or without a single base of mismatch . For transient transfections , HeLa cells ( grown in DMEM supplemented with 10% FCS ) , were grown in multiwell petri dishes 16 h before transfection . The expression ( 21A , 21A[1] , 21A[2] , 21A[3] ) constructs containing the regions of interest cloned in the pTopo vectors ( Invitrogen , http://www . invitrogen . com ) were introduced into the cells using the Fugene 6 transfection reagent ( Roche , http://www . roche . com/home . html ) according to the manufacturer's instructions . A plasmid expressing luciferase was used as a control of transfection efficiency ( to which all the results were normalized ) . Cells were harvested 24 , 48 , and 72 h after transfection , and firefly luciferase activity was measured by dual-luciferase reporter–assay system ( Promega , http://www . promega . com ) according to the manufacturer's protocol . To specifically inhibit RNA polymerase III and/or RNA polymerase II , a cell-permeable chlorobenzenesulfonamide ( ML-60218 ) ( Calbiochem , http://splash . emdbiosciences . com ) and/or α-amanitin ( Roche , http://www . roche . com ) were used at the concentration of 20 μM and 10 μg/ml , respectively , in the medium for 25 h ( ML-60218 ) and 12 h ( α-amanitin ) before the luciferase activity detection . To test the promoter activity of the novel transcription units , we prepared six plasmid constructs expressing a firefly luciferase-silencing hairpin ( Gregory Hannon , Cold Spring Harbor Laboratory , Cold Spring Harbor , New York , United States ) ; transcription was driven by the 11A , 14A , 21A , 29A , 38A , 51A promoters , respectively . The hairpin sequence ( targeting a firefly luciferase mRNA from a cotransfected expression plasmid [Promega] ) is: 5′-GGAUUCCAUUCAGCGGAGCCACCUGAUGAAGCUUGAUCGGGUCUCGCUGAGUUGGAAUCCAUU-3′ . Oligos used to subclone the novel pol III type III promoters within Not I/HinD III restriction sites ( uppercase letters ) were the following: 11AFprom Not I: 5′-atgcGCGGCCGCatttgcatgtcgctatgtg-3′ 11ARprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccacgcactcagctcgtg-3′ 14AFprom Not I: 5′-atgcGCGGCCGCaactgatgtatgattatatctt-3′ 14ARprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccattattatctcctttgttctgt-3′ 21AFprom Not I: 5′-atgcGCGGCCGCacagctgtagcagatgct-3′ 21ARprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccaccacacttggtcaactat-3′ 29AFprom Not I: 5′-atgcGCGGCCGCttctcacctaaaggagtc-3′ 29ARprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccttctaatcctcctaagatca-3′ 38AFprom Not I: 5′-atgcGCGGCCGCttcactaagatccagtgc-3′ 38Arprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccgattcatgaacacagaatatt3′ 51AFprom Not I: 5′-atgcGCGGCCGCgttgaacatttaactctgtat-3′ 51Arprom HinD III: 5′-gatcAAGCTTcatcaggtggctcccgctgaattggaatccctcatggcacttggagat-3′ In this analysis , the above constructs were cotransfected with a pGL3 plasmid-expressing ( Promega ) firefly luciferase as a target to be silenced and with a pRL plasmid expressing ( Promega ) renilla luciferase to which all the determinations were normalized . Cells were harvested 24 , 48 , and 72 h after transfection , and firefly/renilla luciferase activities were measured by dual-luciferase reporter–assay system ( Promega ) according to the manufacturer's protocol . The plasmid constructs p21A , p21A ( 1 ) , and p21A ( 2 ) were generated amplifying from a genomic DNA preparation of the regions of interest; the PCR products were then subcloned into the pNEB193 vector . The oligos used to generate p21A PCR fragments were the following: 21A forward: 5′-GGAAATCTTACCTTCCTGCC-3′ 21A reverse: 5′-TGGCTAGGTCATGTGACCAT-3′ 21A ( 1 ) forward: 5′-GGAAATCTTACCTTCCTGCC-3′ 21A ( 1 ) reverse: 5′-TTCATTCATTCATTCATTGATTCAC-3′ 21A ( 2 ) forward: 5′–CAGCTGCAGCAGATGCTAGCAGGGC-3′ 21A ( 2 ) reverse: 5′–TGGCTAGGTCATGTGACCATTC-3′ The plasmid construct pAnti-21A was generated amplifying the transcribed region from p21A plasmid using the following oligos: Anti-21A Terminator-containing forward: 5′-CTGAAAAAGTAGTCCCAGCACTTTG-3′ Anti-21A Bam HI-containing reverse: 5′-ATGCGGATCCGAGACAGGGTCTTGCTC-3′ Thus the transcribed region was generated in antisense configuration . The pAnti-21A promoter was obtained by amplifying p21A promoter with the following oligos: 21A Forward: 5′-GGAAATCTTACCTTCCTGCC-3′ p21A Bam HI-containing reverse: 5′-ATGCGGATCCGAGCCACCACACTTGGTC-3′ . The PCR products were digested with the restriction enzyme Bam HI , purified by gel electrophoresis , and ligated by T4 ligase ( Invitrogen ) . The insert obtained was then subcloned in pTOPO vector ( Invitrogen ) following the manufacturer's instructions . Prior to transfection all of the plasmids were sequenced by DNA sequencing kit ( Applied Biosystems , www . appliedbiosystems . com ) following the manufacturer's instructions . To isolate and sequence a partial 21A cDNA , we performed different RT-PCR reactions . Starting from about 5 μg of total RNA , cDNA was synthesized by using an oligo ( dT ) 12–18 primer or a random hexamers mix and a superscript first-strand synthesis system for RT-PCR ( Invitrogen ) . cDNAs were diluted 10–50 times , then subjected to PCR reactions . The oligos used to isolate 21A RT-PCR product were: oligo forward 21AF 5′-gctcacgtagtcccagcacttt-3′ and oligo reverse 21AR 5′-actatgttgcccaagctggtct-3′ . PCR products were separated on 1 . 5%–2% agarose gel . The DNA bands were cut , purified by the Millipore DNA gel extraction kit ( http://www . millipore . com ) , and sequenced . The RNA for 21A was measured by real-time quantitative RT-PCR using the PE ABI PRISM@ 7700 sequence detection system ( PerkinElmer , http://www . perkinelmer . com ) and Sybr Green ( Applied Biosytems ) . The sequences of 21A forward and reverse primers as designed by the Primer Express 1 . 5 software ( Applied Biosystems ) were 5′-GCTGAGGCAGGAGGATCACT-3′ and 5′-GCACTACCACACCCAGCTAATTTT-3′ . The sequences of CENP-F forward and reverse primers were 5′-CTGCAGAAAGAACTCTCTCAACTTC-3′ and 5′-TCAACAATTAAGTAGCTGGAACCA-3′ . For endogenous control , the expression of GAPDH gene was examined . The sequences for human GAPDH primers were 5′-GAAGGTGAAGGTCGGAGTC-3′ and 5′- GAAGATGGTGATGGGATTTC-3′ . The sequences for human 5S rRNA primers were 5′-TACGGCCATACCACCCTGAA-3′ and 5′-GCGGTCTCCCATCCAAGTAC-3′ . The sequences for human 7SK RNA primers were 5′-AGGACCGGTCTTCGGTCAA-3′ and 5′-TCATTTGGATGTGTCTGCAGTCT-3′ . The sequences for human c-Myc primers were 5′-CGTCTCCACACATCAGCATAA-3′ and 5′-GACACTGTCCAACTTGACCCTCTT-3′ . Relative transcript levels were determined from the relative standard curve constructed from stock cDNA dilutions and divided by the target quantity of the calibrator following manufacturer's instructions . The Anti-21A siRNA was synthesized against a region of the 21A transcript of no homology with CENP-F so that the silencing effect was specific for the pol III regulatory RNA and did not interfere with CENP-F RNA stability . The siRNA synthesis was carried out taking advantage of the siRNA construction kit ( Ambion , http://www . ambion . com ) according to the manufacturer's protocol . The sense/antisense oligos used were: 5′-aaGTGTGGTGGCTCACcctgtctc-3′ and 5′-aaGTGAGCCACCACACcctgtctc-3′ . We tested proliferation of HeLa cells transfected with 21A , 21A-1 , 21A-2 , 21A-3 , and Anti-21A constructs plating 5 × 105 cells per well in round-bottomed 96-well plates , incubated for 24 , 48 , and 72 h after transfection , and pulsed with [3H]-thymidine ( 1 . 0 μCi/10 μl/well ) ( Amersham Biosciences , http://www5 . amershambiosciences . com ) for the last 18 h . We harvested the cells and evaluated cell proliferation by counting the thymidine uptake . We calculated the averaged proliferation rate , measured as counts per minute , and standard deviation for the triplicate wells of each sample . Based on a single-step acid-phenol–guanidium method , total RNA was extracted using TRIzol reagent ( Invitrogen ) according to the manufacturer's protocol . Total RNAs , from HeLa cells , were electrophoresed through 1 . 5% agarose gels in the presence of formaldehyde and blotted onto Hybond N membranes ( Amersham ) . The blot was hybridized with an 85-bp–long probe contained in the region from nucleotide 1 , 194 to nucleotide 1 , 278 of the 21A reported sequence ( Table S2 ) , spanning a region internal to the transcript . The probe was obtained by PCR ( using the 21A plasmid construct as template ) using the following oligos: 21AF 5′- GCTCACGTAGTCCCAGCACTTT-3′ and 21AR 5′-AGACCAGCTTGGGCAACATAGT-3′ . Blot prehybridization was performed at 65 °C for 2 h in 333 mM NaH2PO4 ( pH 7 . 2 ) , 6 . 66% sodium dodecyl sulphate , and 250 mg/ml denatured salmon sperm DNA . Blot hybridization was performed at 65 °C for 18 h in the same solution containing 106 counts per minute/ml of denatured and labeled probes . After hybridization the blots were washed twice at 65 °C for 30 min in 0 . 2% sodium dodecyl sulphate , 2× SSPE and once at 65 °C for 30 min in 0 . 2% sodium dodecyl sulphate , and 0 . 2× SSPE . Membranes were exposed to autoradiographic films for 24–48 h and then developed . Equal amounts of proteins ( 10 μg/sample ) from each sample were loaded on standard 4%–12% NU-PAGE gradient gels ( Invitrogen ) . Blotting onto Protran nitrocellulose membranes ( Schleicher & Schuell , www . schleicher-schuell . com ) was performed in the X-Cell Sure Lock Electrophoresis Cell ( Invitrogen ) , according to the manufacturer's instructions . The membranes were saturated overnight in 3% nonfat milk in TTBS buffer ( 500 nM NaCl; 20 mM Tris/Cl [pH 7 . 5]; 0 . 05% Tween-20 ) and incubated for 4 h at room temperature with the human anti-mitosin/CENP-F ab90 ( ABCAM , http://abcam . com ) and/or anti-alpha tubulin ( OMIM 191110 ) ( Sigma-Aldrich , www . sigmaaldrich . com ) mouse monoclonal antibodies . The anti-mitosin antibody recognized a weak signal at a very high apparent molecular mass ( 350–400 kDa ) , while the anti-alpha tubulin showed a clear signal at 45 kDa . The immunoreactive band was revealed by an alkaline phosphate-conjugated affinity-purified monoclonal anti-rabbit–mouse IgG ( Sigma-Aldrich ) , and ( in the experiment indicated in Figure 1C ) the enzymatic chemiluminescence ( ECL ) detection system ( Amersham ) , or ( in the experiment indicated in Figure 1E ) the alkaline phosphatase substrate BCIP/NBT ( ICN Biomedicals , http://www . mpbio . com ) .
The National Center for Biotechnology Information ( http://www . ncbi . nlm . nih . gov ) accession numbers for the genes and gene products discussed in this paper are: U6 ( M14486 ) , 7SK ( 001445 ) , H1 ( 002312 ) , U2 ( 002716 ) , CENP-F ( 016343 ) , 5S rRNA ( V00589 ) , c-Myc ( 002467 ) , and GAPDH ( 002046 ) . | After the sequence of the human genome was determined , it was immediately recognized that a large part of the regulation of the gene expression occurring in the cells under physiological , as well as under pathological conditions , is carried out by RNA molecules that do not code for proteins ( the “noncoding portion” of the genome ) . Here , we focus on small RNA molecules transcribed by the RNA polymerase III and identify a novel set of approximately 30 noncoding ( nc ) RNA genes . We propose that these RNA transcripts play a key role in regulating the expression of specific protein-coding genes transcribed by the RNA polymerase II , thus constituting an unprecedented example of cogene/gene pairs . Furthermore , we provide evidence that the RNA polymerase III , in addition to the well-known task in the constitutive synthesis of small RNAs ( such as 5S rRNA and tRNAs ) , also plays a key role in the area of gene-expression control . A detailed investigation of the function of one of the novel ncRNA genes , called 21A , revealed that its transcript plays a role in the control of the proliferation of some tumor cells . The above findings significantly expand our understanding of the ncRNA universe and open the way to further studies aimed at the elucidation of the molecular pathways involving this novel class of regulatory RNAs . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"cell",
"biology",
"genetics",
"and",
"genomics",
"homo",
"(human)",
"molecular",
"biology"
] | 2007 | New Small Nuclear RNA Gene-Like Transcriptional Units as Sources of Regulatory Transcripts |
Promoters are structurally and functionally diverse gene regulatory regions . The presence or absence of sequence motifs and the spacing between the motifs defines the properties of promoters . Recent alternative promoter usage analyses in Drosophila melanogaster revealed that transposable elements significantly contribute to promote diversity . In this work , we analyzed in detail one of the transposable element insertions , named FBti0019985 , that has been co-opted to drive expression of CG18446 , a candidate stress response gene . We analyzed strains from different natural populations and we found that besides FBti0019985 , there are another eight independent transposable elements inserted in the proximal promoter region of CG18446 . All nine insertions are solo-LTRs that belong to the roo family . We analyzed the sequence of the nine roo insertions and we investigated whether the different insertions were functionally equivalent by performing 5’-RACE , gene expression , and cold-stress survival experiments . We found that different insertions have different molecular and functional consequences . The exact position where the transposable elements are inserted matters , as they all showed highly conserved sequences but only two of the analyzed insertions provided alternative transcription start sites , and only the FBti0019985 insertion consistently affects CG18446 expression . The phenotypic consequences of the different insertions also vary: only FBti0019985 was associated with cold-stress tolerance . Interestingly , the only previous report of transposable elements inserting repeatedly and independently in a promoter region in D . melanogaster , were also located upstream of a stress response gene . Our results suggest that functional validation of individual structural variants is needed to resolve the complexity of insertion clusters .
Promoters are crucial regions for the transcriptional regulation of gene expression . Recent computational and experimental advances in functional genomics techniques have allowed defining the promoter architecture to an unprecedented level . Several core promoter motifs such as the Initiator ( Inr ) and the Downstream core Promoter Element ( DPE ) have been described , and it is likely that many others remain to be discovered . The presence or absence of the core promoter motifs influences enhancer-promoter communication and thus gene regulation [1] . Promoter regions also harbour transcription factor binding motifs , which are another important component in the regulation of gene expression [2] . Besides cis-regulatory elements that influence the temporal and spatial expression patterns of genes , proximal promoters often contain alternative transcription start sites ( TSSs ) [1 , 3] . Rather than being “biological noise” from imprecise binding of the transcription initiation machinery , genome-wide analyses of TSSs usage showed that alternative TSSs play an important role in the diversification of gene expression patterns [4–8] . Transposable elements ( TEs ) , long proposed to play an important role in gene regulation [9 , 10] , have recently been found to provide at least 1 , 300 alternative TSSs in the Drosophila melanogaster genome [8] . TEs can also add Transcription Factor Binding Sites ( TFBSs ) to the promoter of genes as has been recently shown in Drosophila and humans [11–13] . As a result of adding particular sequence elements , many TEs confer their intrinsic regulatory properties to nearby genes demonstrating that they distribute cis-regulatory modules [8] . Finally , TEs inserted in promoter regions can also influence gene expression by disrupting the promoter architecture . This is the case , for example , of naturally occurring P-element insertions in the promoter of heat shock protein ( hsp ) genes [14] . One of the TEs identified as providing an alternative TSS by Batut et al ( 2013 ) [8] , named FBti0019985 , was previously reported in a screening designed to identify putatively adaptive TE insertions in D . melanogaster [15] . However , this particular TE was not further studied because its population frequency could not be accurately determined [15] . FBti0019985 is a roo solo-LTR inserted in the 5’-UTR of CG18446 gene , which is nested in the first intron of crossbronx ( cbx ) ( Fig 1 ) . TEs from the roo family have long been proposed to affect the expression of nearby genes by adding and distributing cis-regulatory regions [16–19] . Specifically , roo LTRs contain several TFBSs and the Inr sequence characteristic of core promoters [8 , 20] . Interestingly , CG18446 has been identified as a candidate gene for cold resistance: it is upregulated in fly strains that have been selected for increased cold resistance compared with control strains that were not subjected to cold-stress [21] . Cold resistance is an ecologically and evolutionarily relevant trait because it influences the ability of the species to adapt to different climatic conditions and thus , their geographical distribution [22 , 23] . There is good evidence suggesting that D . melanogaster adapts to cold environments and a growing list of candidate genes involved in this thermotolerance phenotype is being identified [21 , 24–28] . However , the molecular variants responsible for the adaptive cold-stress resistance phenotype remain elusive [29] . In this work , we further analyzed the presence/absence of FBti0019985 in four natural populations of D . melanogaster . We found that besides FBti0019985 , eight other roo elements have inserted in a 368 bp region around CG18446 transcript start site . These roo elements differ in the insertion site and in their orientation . On the other hand , all elements have the same size and show high sequence conservation: all cis-regulatory elements previously described in roo LTRs are highly conserved [8 , 30] . We further investigated whether these different insertions were functionally equivalent by performing 5’-RACE , gene expression , and phenotypic analyses . Our results showed that the functional consequences of the different roo insertions depend on the particular position where the element is inserted . Among the nine different roo solo-LTR insertions , only FBti0019985 is consistently associated with increased viability in nonstress and cold-stress conditions across genetic backgrounds .
We first aimed at estimating the frequency of FBti0019985 in non-African natural D . melanogaster populations . Thus , we checked using PCR whether this insertion was present , polymorphic , or absent in 28 strains from a natural population collected in North Carolina ( North America , DGRP strains [31 , 32] ) and in 15 strains from a natural population collected in Bari ( Italy , Europe [33] ) ( Table 1 ) . We obtained PCR results for 39 of the 43 strains tested: nine strains produced PCR bands consistent with FBti0019985 being present , five strains appeared as heterozygous , 13 strains showed unexpected band patterns , and 12 strains appeared as absent ( Table 1 ) ( see Material and Methods ) . To verify these results , we sequenced 32 of the 39 strains including all the strains that showed some evidence of presence ( Table 1 ) . Only four of the nine strains classified as present , according to the PCR results , had the FBti0019985 insertion . For the rest of this work , we considered the position where FBti0019985 is inserted as the "reference position" . The other five present strains , the five heterozygous strains , and 12 of the 13 strains that gave unexpected PCR bands contained different roo solo-LTR insertions ( Table 1 ) . Overall , besides FBti0019985 , we found eight other 428 bp roo solo-LTRs inserted in eight different positions ( Fig 2 ) . Three roo insertions are located downstream of the reference position: roo+7 , roo+175 , and roo+278 ( Fig 2 ) . Two of the four strains carrying roo+7 have a duplication of the 95 bp region located immediately upstream of the insertion ( Table 1 ) . roo+175 element is inserted in the 5’-UTR region , and roo+278 is inserted in the first exon of CG18446 gene . Both roo+175 and roo+278 have a conserved Inr motif . If transcription starts in these insertions , flies carrying roo+175 would have a 100 bp shorter 5'-UTR , and flies carrying roo+278 would have a 35 amino acids shorter CG18446 protein . The other five roo insertions are located upstream of the reference position: roo-19 , roo-28 , roo-44 , roo-68 , roo-90 ( Fig 2 ) . Four of them , roo-19 , roo-28 , roo-44 , and roo-68 , are inserted in reverse orientation . We used Tlex-2 software to further analyze the frequency of the nine roo insertions in 21 additional DGRP strains , in 26 strains from a Swedish natural population , and in 42 strains from a population collected in the ancestral range of the species , Zambia ( Fig 2 and S1 Table ) ( see Material and Methods ) [34] . Overall , we found that 67 strains , out of the 128 strains analyzed , contained one of the nine roo solo-LTR insertions . The two most common roo insertion in out-of-Africa populations are roo-90 and FBti0019985 present in 13% and 10% of the strains tested , respectively ( Fig 2 ) . Besides , some insertions are only present in the North Carolina natural population while others are specific to the Italian natural population ( Fig 2 ) . Only three of the nine insertions described in North Carolina and Italian populations are present in the Swedish population . However , we did not perform de novo discovery of TEs in this population . Thus , it could be that other private insertions are present in the Swedish population . Finally , all the nine insertions were present in the African population although most of them were present at very low frequencies ( Fig 2 ) . In summary , we have found that besides the FBti0019985 insertion annotated in the reference genome , eight other 428 bp roo solo-LTRs are inserted nearby CG18446 TSS in natural populations of D . melanogaster ( Fig 2 ) [35] . Each one of the strains analyzed contains a single solo-LTR roo insertion and most of the analyzed strains contain one of the nine solo-LTR roo insertions . We identified the Target Site Duplications ( TSD ) of the nine different roo insertions using data from the 26 present strains sequenced in this work ( Table 1 ) . We could identify the TSD for all roo insertions except for roo+278 . We found that six of the eight TSDs identified are five nucleotides long as has been previously described for this family [36] ( Fig 2 ) . However , the TSD sequences did not match the proposed TSD consensus sequence [34 , 36 , 37] . We thus used all the available roo TSD sequences to build a new consensus ( S1 Fig ) . The different roo solo-LTR insertions had different TSDs suggesting that they are independent insertions ( Fig 2 ) . Furthermore , all the roo elements located in a given insertion site have the same exact TSD and are inserted in the same orientation suggesting that each one of them is a unique insertion event ( Fig 2 ) . To test whether these nine insertion events were the result of a burst of transposition , we constructed a phylogenetic tree . We included the nine roo insertions sequenced in this work and 115 other roo insertions present in the D . melanogaster genome ( S2 Fig and S1 Text ) . We found that not all the newly described roo insertions clustered together suggesting that they did not insert at the same time ( S2 Fig and S1 Text ) . All the TEs identified in CG18446 proximal promoter region belong to the roo family . Thus , we also investigated whether roo elements annotated in the reference genome are preferentially inserted into gene proximal promoter regions as has been previously described for other TE families [38 , 39] . We analyzed the 138 insertions belonging to the roo family annotated in the D . melanogaster reference genome ( v5 ) . We found 21 roo insertions located in the 1 kb region upstream of a gene or overlapping the 5’-end of a gene . Thus , only 15 . 2% of the roo elements in the D . melanogaster genome are located in gene promoters and/or 5’-UTRs . In summary , TSD analyses of the nine insertions characterized in this work suggested that they are independent insertions , and confirmed the length but not the sequence previously reported as the TSD consensus for this family . Our results are not consistent with the nine roo insertions being the result of a single burst of transposition . Finally , our analyses also suggested that roo elements do not preferentially insert in 5’ gene regions . We analyzed multiple sequence alignments of all the roo insertions located nearby CG18446 . We identified TFBSs using the JASPAR database ( see Material and Methods ) . We also specifically looked for conservation of the regulatory regions previously described in the roo family [8 , 30] , and for conserved core promoter motifs [1] ( Fig 3A and S2A Table ) . Overall , there was very little diversity among the nine solo-LTRs ( S3A Fig ) . The five TFBSs and the Inr sequence previously identified in the consensus sequence of roo LTRs are conserved in all the roo copies located in the proximal promoter of CG18446 [8] . Additionally , we found another four TFBSs that are also highly conserved in all the copies ( Fig 3A and S3A Fig ) . The nine transcription factors are involved in developmental processes . Additionally , Deaf1 and Nub are also involved in immune response [40 , 41] . Finally , three previously identified Matrix Associated Regions ( MARs ) in LTRs from the roo family are also highly conserved in the nine insertions ( Fig 3A and S3B Fig ) [30] . These results suggest that these roo solo-LTR insertions are introducing the same cis-regulatory regions in the CG18446 proximal promoter region . Still , the functional effect of these insertions might be different because they are located in different positions and have different orientations ( Fig 2 ) . We analyzed the proximal promoter region of CG18446 in the 30 strains sequenced in this work . We could not identify the TATA box suggesting that CG18446 has a DPE promoter [1] . We identified eight TFBSs in the proximal promoter of CG18446 ( Fig 3B and S2C Table ) . These eight TFBSs are highly conserved in all the strains analyzed ( S3C Fig ) . The different roo insertions characterized in this work do not disrupt any of the identified core promoter motifs or TFBSs ( Fig 3B ) . However , they do affect the spacing between the different regulatory motifs , which might affect the protein-protein interaction at the CG18446 promoter and thus the expression level of this gene ( Fig 3B ) [14] . Besides affecting the spacing of transcription factor binding site , another mechanism by which roo insertions could be affecting CG18446 expression is by recruiting piRNAs that would lead to heterochromatin formation [42 , 43] . We mapped piRNA reads from three different available libraries to a 1 . 4 kb region including FBti0019985 ( Fig 4A ) ( see Material and Methods ) [44–46] . We found that most of the piRNAs mapping to the insertion were sense reads , suggesting that FBti0019985 is not acting as a target for heterochromatin assembly [42] . We also looked for evidence of HP1a binding to FBti0019985 using modENCODE data ( see Material and Methods ) [47] . HP1a is a structural chromosomal protein that mediates both gene expression and gene silencing [48] . We did find evidence of HP1a reads binding to FBti0019985 ( Fig 4B ) . Thus , by recruiting HP1a , FBti0019985 could be affecting the expression of CG18446 . The same results were obtained for the other eight roo solo-LTR insertions: most of the piRNAs mapping to the insertions were sense reads and we found evidence of HP1a binding to all of them ( S3 Table ) . Overall , our results are suggestive but not conclusive of HP1a binding to the nine roo insertions described in this work . To further investigate the possible functional consequences of the roo insertions , we focused on the five insertions present at higher population frequencies in out-of-Africa populations: FBti0019985 , roo+7 , roo-44 , roo-90 , and roo-68 ( Fig 2 ) . We investigated whether roo insertions could be providing an alternative TSS to CG18446 . Batut et al ( 2013 ) [8] reported that the TSS of CG18446 is located inside FBti0019985 . However , this finding was obtained using RAMPAGE and was not further validated using 5’-RACE . For this reason , we performed a 5’-RACE with the RAL-810 strain that carries FBti0019985 and with the RAL-783 strain that does not carry any of the nine roo solo-LTR insertions . As expected , we found that the TSS of CG18446 is inside the TE: the first 50 bp of the 276 bp 5’-UTR correspond to FBti0019985 ( Fig 5 ) . Additionally , flies with the insertion have also a shorter transcript , with a 201 bp 5’-UTR , that does not start in FBti0019985 ( Fig 5 ) . Most of the sequenced transcripts start in the FBti0019985 insertion ( 14 out of 20 transcripts analyzed ) . Flies without the FBti0019985 insertion only have the 201 bp 5’-UTR transcript ( Fig 5 ) . We then checked whether roo+7 , located only 7 bp downstream of FBti0019985 , roo-90 , which is the most distal insertion , and roo-44 , which is inserted in reversed orientation , also provide an alternative TSS to CG18446 . We found that roo+7 affects the TSS of CG18446 ( Fig 5 ) . Indeed , the TSS in roo+7 is in the same nucleotide position as in FBti0019985 . Thus , CG18446 transcript in flies with roo+7 is 7 bp shorter compared with the transcript in flies with FBti0019985 . Similarly to FBti0019985 , most of the sequenced transcripts started in the roo+7 insertion ( 18 out of 22 transcripts analyzed ) . On the other hand , we did not find evidence of a TSS inside roo-90 , which might indicate that the distance of the TE to the nearby gene affects its ability to provide an alternative TSS ( Fig 5 ) . Finally , we analyzed two different strains carrying the roo-44 insertion in the same position and we could not find evidence for a transcript with the TSS in roo-44 ( Fig 5 ) . Overall , we found that only FBti0019985 and roo+7 insertions modify the length of CG18446 transcript . These two roo insertions are located a few nucleotides from the gene and both are inserted in 5’ to 3’ orientation . We further analyzed whether different roo insertions were associated with changes in CG18446 expression in embryos , where this gene is highly expressed [49] . For FBti0019985 , we analyzed the expression of CG18446 in flies with four different genetic backgrounds . In three of the four backgrounds , FBti0019985 is associated with upregulation of CG18446 ( Fig 6A ) . This result is significant in two genetic backgrounds , RAL-810 and IV68 , and marginally significant in a third background , RAL-639 ( t-test p-value = 0 . 045 , p-value = 0 . 005 and p-value = 0 . 062 , respectively ) ( Fig 6A ) . On the other hand , only in one of the three genetic backgrounds analyzed for roo+7 , the insertion is associated with downregulation of this gene ( t-test p-value = 0 . 015 for RAL-405 ) ( Fig 6B ) . We also checked the expression of CG18446 in flies with two roo solo-LTR insertions that do not provide an alternative TSS to this gene: roo-90 and roo-44 . We found that roo-90 is only associated with CG18446 upregulation in one of the three backgrounds analyzed ( p-value = 0 . 001 , for RAL-21 ) ( Fig 6C ) . Two different strains with the roo-44 solo-LTR insertion did not show differences in the level of expression of CG18446 compared with strains without the insertion ( p-values > 0 . 05 in both cases ) ( Fig 6D ) . Overall , we found that FBti0019985 is associated with CG18446 upregulation in three of the four backgrounds analyzed ( Fig 6A ) . In the majority of strains , roo+7 , roo-90 , and roo-44 are not associated with changes in CG18446 expression level ( Fig 6B–6D ) . However , we can not discard that the presence of these insertions is associated with changes in the expression of CG18446 in other developmental stages and/or in tissues not analyzed in this work . We have shown that FBti0019985 affects the transcript length and it is associated with upregulation of CG18446 in most of the genetic backgrounds analyzed ( Figs 5 and 6A ) . Because CG18446 has been previously identified as a cold-stress candidate gene , we tested whether flies with and without FBti0019985 differed in their sensitivity to cold-stress [21] . We first compared RAL-810 , which carries FBti0019985 , with RAL-783 , which does not carry any of the nine roo insertions ( Fig 7A ) . We performed three biological replicates . ANOVA analyses showed that the experimental condition ( nonstress or cold-stress ) and the insertion genotype ( presence or absence of FBti0019985 ) were significant ( Table 2 ) . Flies with FBti0019985 had a higher viability than flies without this insertion in both nonstress and cold-stress conditions . Furthermore , the interaction between these two factors was also significant suggesting that the effect of the insertion is larger in cold-stress conditions ( Fig 7A and Table 2 ) . We repeated the experiment using flies with different genetic backgrounds: RAL-802 that carries FBti0019985 and RAL-908 that does not carry this insertion ( Fig 7B ) . ANOVA analyses showed that the experimental condition and the insertion genotype are significant while the interaction between these two factors was not significant ( Table 2 ) . RAL-802 flies had a higher egg-to-adult viability in nonstress and in cold-stress conditions compared with flies without FBti0019985 . Finally , we tested whether flies from a different population , IV68 carrying FBti0019985 and IV22 without this particular insertion both collected in Italy , also showed significantly increased viability in nonstress and in cold-stress conditions ( Fig 7C and Table 2 ) . We found that IV68 flies had a higher viability than flies without the FBti0019985 insertion in both nonstress and cold-stress conditions ( Table 2 ) . Overall , we found consistent results , across genetic backgrounds from two different natural populations , suggesting that flies with the FBti0019985 insertion are associated with increased viability compared to flies without this insertion in nonstress and in cold-stress conditions . In all cases , the effect of the presence of the insertion was either medium or large ( Table 2 ) . In one of the genetic backgrounds , the effect was larger under cold-stress conditions ( Fig 7A ) while no interaction between experimental condition and insertion genotype was found in the other two backgrounds ( Fig 7B and 7C ) . We further checked whether another four roo solo-LTR insertions described in this work are associated with cold-stress phenotypes . For each insertion , we compared the egg-to-adult viability of flies with two different genetic backgrounds with the egg-to-adult viability of RAL-783 that does not carry any of these insertions ( Fig 8 ) . In all cases , we performed ANOVA analyses to check whether the experimental conditions , insertion genotype , and/or the interaction between these two factors were significant ( Table 2 ) . We found that the experimental condition had a significant effect on egg-to-adult viability in all the strains tested ( Table 2 ) . On the other hand , the effect of the insertion was only significant in some of the genetic backgrounds ( Table 2 ) . Among strains that carry the roo+7 insertion , the insertion genotype had an effect only in one of the two backgrounds tested ( Fig 8A and 8B and Table 2 ) . RAL-405 flies with roo+7 insertion showed decreased viability ( Fig 8A and Table 2 ) . The presence/ absence of roo-90 did not have a significant effect on egg-to-adult viability ( Fig 8C and 8D and Table 2 ) . For roo-44 , while the insertion genotype had a significant effect on the two backgrounds tested , results were not consistent . In one background , the presence of the insertion is associated with increased viability under cold-stress conditions and the interaction between the treatment and the insertion genotype is significant ( Fig 8E and Table 2 ) , while in the other background the presence of roo-44 is associated with decreased viability ( Fig 8F and Table 2 ) . Finally , the presence of roo-68 significantly affected viability in only one of the two backgrounds tested: RAL-716 flies carrying roo-68 showed decreased viability ( Fig 8H and Table 2 ) . Overall our results suggested that the presence of roo+7 , roo-90 , roo-44 , and roo-68 solo-LTR insertions reported in this work was not consistently associated with cold-stress phenotypes ( Fig 8 ) . These other insertions could have no phenotypic effect or could be involved in phenotypes not analyzed in this work . We looked for evidence of positive selection in the 2 kb region flanking the FBti0019985 insertion . We analyzed the number of segregating sites ( S ) in this region and estimated Tajima´s D , iHS , nSL , H12 and XP-EHH ( see Material and Methods ) . We found reduced diversity in the strains with FBti0019985: the number of segregating sites in this region is significantly smaller than the number of segregating sites found in 2 kb regions of chromosome 2R , where the FBti0019985 insertion is located ( p-value = 0 . 015 ) ( S4 Table ) . We also found that Tajima’s D was significantly negative in the 2 kb region where FBti0019985 is inserted , as expected if this region is under positive selection ( p-value = 0 . 009 ) ( S4 Fig and S4 Table ) . Finally , we also found significant values of iHS and H12 in the region flanking the FBti0019985 insertion ( p-value = 0 . 048 and p-value = 0 . 023 , respectively ) ( S5 Fig and S4 Table ) . We also looked for evidence of selection taking into account not only the strains in which FBti0019985 is inserted , but all the strains that contain one of the nine roo insertions described in this work . In this case , only iHS showed a marginally significant value ( p-value = 0 . 049 ) ( S6 Fig ) . Overall , our results suggest that the strains carrying FBti0019985 might be evolving under positive selection while the evidence for positive selection taking into account all the strains with one of the nine roo solo-LTRs , was only marginally significant .
Besides FBti0019985 , we have discovered eight other roo solo-LTR elements inserted in the 368 bp region nearby the TSS of the cold-stress response gene CG18446 ( Fig 2 ) [21] . Each strain contained a single roo insertion and the population frequency of the different individual insertions varies from 1% to 17% ( Fig 2 ) . Full-length elements from the roo family are 8 . 7 kb long . Such long insertions in the proximal promoter of CG18446 located in the first intron of cbx , might be deleterious , which could explain why all the identified insertions were solo-LTR elements . In D . melanogaster , repeated insertions of TEs have only been described in the proximal promoters of a particular gene class: hsp genes [50] . The susceptibility of hsp genes to TE insertions was attributed to their peculiar chromatin architecture: constitutively decondensed chromatin and nucleosome-free regions [51 , 52] . However , promoter regions of non-hsp genes with similar chromatin architecture are not targets for TE insertions suggesting that chromatin accessibility is not sufficient to explain the susceptibility of hsp genes to TE insertions [50] . From a functional point of view , the presence of TEs in the promoter regions of hsp genes has been suggested to allow a rapid gene expression response to unpredictable temperature changes [50] . Similarly , the presence of roo insertions in the promoter of CG18446 could also be enhancing the ability of this gene to respond to environmental challenges , although only one of the nine roo insertions was associated with cold-stress tolerance ( see below ) . Interestingly , almost 100% of the insertions described in heat-shock genes are P-element insertions , and all the insertions described here are roo elements . P-elements preferentially insert in the 5' end of genes where they recognize a structural motif rather than a sequence motif [38 , 39] . While 81% of P-elements insert in 5’ gene regions , our results showed that only 15 . 2% of the roo elements annotated in the reference genome are inserted in 5’ gene regions . Thus , with the data currently available , roo insertions do not seem to preferentially insert into 5’ gene regions although analyses of de novo insertions should shed more light on this issue . Our results showed that the different roo elements inserted in the proximal promoter of CG18446 differ in their molecular and functional effects ( Table 3 ) . We found that the two insertions that are more closely located to CG18446 , FBti0019985 and roo+7 , provided an alternative TSS to this gene ( Fig 5 and Table 3 ) . However , only FBti0019985 is associated with upregulation of CG18446 expression ( Fig 6 and Table 3 ) . Besides providing an alternative TSS , the effect of the FBti0019985 insertion on CG18446 expression could be due to the addition of new regulatory regions ( Fig 3A ) , to the disruption of the spacing of pre-existing ones ( Fig 3B ) , and/or to the recruitment of HP1a protein that could also lead to changes in the expression of CG18446 ( Fig 4B ) . Finally , we cannot discard that polymorphisms other than the presence/absence of the FBti0019985 insertion also affect the expression of CG18446 . We found that the FBti0019985 insertion , which is associated with increased CG18446 expression , is consistently associated with increased viability in nonstress and in cold-stress conditions ( Fig 7 and Table 3 ) . Although we cannot exclude that other variants linked to FBti0019985 contribute to the increased viability phenotypes , we argue that it is unlikely that the association between the FBti0019985 insertion and increased viability in three different genetic backgrounds from two different natural populations would occur spuriously [53] . These results also suggest that CG18446 is likely to play a role in cold tolerance as was previously suggested based on cold-stress selection experiments in which this gene was found to be overexpressed [21] . However , FBti0019985 is present in only 10% of the out-of-Africa natural strains analyzed in this work . Our screening was focused on three out-of-Africa populations , thus we cannot discard that FBti0019985 is present at higher frequencies in other populations . Alternatively , it is also possible that the relatively low frequency of FBti0019985 is due to negative fitness effects of this insertion on other phenotypes . Cold-stress resistance has been associated with decreased starvation resistance [54 , 55] and reduced fecundity [56 , 57] . Therefore , the benefit of flies carrying FBti0019985 in cold-stress conditions might be a cost , for example , when food resources are scarce . While FBti0019985 has a consistent cold-stress tolerance phenotype , four other roo insertions also located on the proximal promoter of CG18446 did not ( Fig 8 and Table 3 ) . The insertion that is present at higher frequencies in out-of-Africa populations is roo-90 ( Fig 2 ) . However , this insertion is not associated with changes of expression of CG18446 in embryos ( Fig 6 ) and was not found to be associated with cold-stress tolerance phenotypes ( Fig 8C and 8D and Table 3 ) . It could be that this insertion has no phenotypic effect . Alternatively , roo-90 could be affecting a phenotype other than cold tolerance . A recent update in FlyBase revealed that CG18446 is also an ethanol-regulated gene that could contribute to ethanol sensitivity or tolerance [58] . Another possibility is that roo-90 affects cbx . As the other roo insertion described in this work and CG18446 gene , roo-90 is inserted in the first intron of cbx which has been functionally classified as a defense response to bacterium and spermatogenesis gene [59] ( Fig 1 ) . Elucidating whether roo-90 has an adaptive effect is beyond the scope of this paper . Overall , we did not find evidence of positive selection at the DNA level in the region where the nine roo solo-LTR elements are inserted . We did find evidence of reduced diversity in this region when only the strains containing FBti0019985 were considered ( S4–S6 Figs and S4 Table ) . Further analyses with a bigger dataset of strains is needed in order to determine whether this region shows signals of positive selection at the DNA level . In summary , our results showed that different TE insertions in the same gene promoter region might have different molecular and functional consequences . Thus , the description of complex regions , as the one reported in this work , should be followed by functional analysis of the structural variants if we want to elucidate which ones are functionally relevant .
We used inbred strains from the Drosophila Genetic Reference Panel ( DGRP [31 , 32] ) and isofemale strains from an Italian population collected in Castellana Grotte ( Bari , Italy [33] ) to perform the molecular and phenotypic assays . We used a PCR approach to check for presence/ absence of FBti0019985 in 28 strains from the North Carolina population and in 15 strains from Italy . The primers used were FBti0019985_FL ( 5’-GGCATCATAAAACCGTTGAACAC-3’ ) , FBti0019985_L ( 5’-AGTCCCTTAGTGGGAGACCACAG-3’ ) and FBti0019985_R ( 5’-CGTAGGATCAGTGGGTGAAAATG-3’ ) ( Fig 1 ) . Primers FBti0019985_L and FBti0019985_R are expected to give a 616 bp band when the TE is present . Primers FBti0019985_FL and FBti0019985_R are expected to give a 638 bp band when the TE is absent and a 1066 bp band when the TE is present . All PCR bands giving evidence of presence and some of the PCR bands giving evidence of absence were cloned using TOPO TA Cloning Kit for Sequencing ( Invitrogen ) following the manufacturer’s instructions and Sanger-sequenced using M13 forward and/or M13 reverse primers to verify the results . Sequences have been deposited in GenBank under accession numbers KU672690-KU672720 . We estimated the frequencies of the nine roo solo-LTR insertions described in this work using T-lex2 software [34] . Because T-lex2 works only for annotated TEs , we constructed eight new reference sequences including each one of the newly described roo solo-LTR insertions . The new reference sequences included 500 bp at each side of the TE and the TSD of each insertion . We run T-lex2 in strains from three different populations: 50 strains from North Carolina ( DGRP [31 , 32] ) , 27 strains from a population collected in Stockholm , Sweden [33] , and 67 strains from a population collected in Siavonga , Zambia [60] . As a control , we also run T-lex2 in the strains for which we have PCR results ( S1 Table ) . We obtained results for 21 out of 50 DGRP strains , 26 out of 27 Swedish strains and 42 out of 67 Zambian strains . In some of the strains , T-lex2 detects more than one insertion per strain . However , PCR analyses of these strains revealed that only one insertion was present . These results suggest that T-lex2 cannot accurately estimate the frequency of insertion when they are closely located to each other . We thus discarded T-lex2 results indicating the presence of more than one insertion per strain . Other factors such as the quality of the reads and the coverage of the different strains could also be affecting T-lex2 results . Target site motifs were constructed in WebLogo ( http://weblogo . berkeley . edu ) using six TSDs sequences obtained in this work and 41 TSDs sequences predicted with T-lex2 software [34] . For each roo solo-LTR insertion , we constructed a consensus sequence taking into account the 26 strains sequenced in this work using Sequencher 5 . 0 software . We aligned the nine roo insertion consensus sequences with 115 of the 137 other roo insertions present in the D . melanogaster genome using the multiple sequence aligner program MAFFT [61] . The quality sequence of the other 22 roo insertions was too low to include them in the alignment . A maximum likelihood tree was inferred using RAxML Version 8 [62] under the general time-reversible nucleotide model and a gamma distribution of evolutionary rates . We use the ETE toolkit Python framework for the analysis and visualization of trees [63] . We looked for conservation of the Transcription Factor Binding Sites ( TFBSs ) previously described in the roo family [8] in all the roo solo-LTRs characterized in this work . First , we downloaded from FlyBase version r6 . 06 ( http://flybase . org ) the fasta file of FBti0019985 sequence ( genome region 2R: 9 , 871 , 090–9 , 871 , 523 ) . We also searched for TFBSs in the roo insertions and in the CG18446 promoter regions using all the available JASPAR CORE Insecta matrices ( http://jaspar . genereg . net ) . Only those sites predicted with a relative score higher than 0 . 995 were considered . We identified four new TFBS in FBti0019985 sequence: Deaf1 , ara , mir , and caup . We then look for conservation of the identified motifs in all the roo solo-LTR sequences described in this work . For some strains , we used the information available in http://popdrowser . uab . cat [64] . We used three piRNA libraries [44–46] to map piRNA reads to a 1 . 4 kb region including FBti0019985 and to all the roo insertions described in this work following the methodology described in Ullastres et al ( 2015 ) [33] . Briefly , we used BWA-MEM package version 0 . 7 . 5 a-r405 [65] to align the reads and then we used SamTools and BamTools [66] to index and filter by sense/antisense reads . The total read density was obtained using R ( Rstudio v0 . 98 . 507 ) [67] . We used modENCODE ChIP-Seq data [47] to map HP1a reads to a 1 . 4 kb region including FBti0019985 and to all the roo insertions described in this work following the methodology described in Ullastres et al ( 2015 ) [33] . We aligned the reads using BWA-MEM package version 0 . 7 . 5 a-r405 [65] . The total read density was obtained using R ( Rstudio v0 . 98 . 507 ) [67] . 5-to-7 day-old flies were placed in a fly cage with egg-laying medium ( 2% agar with apple juice and a piece of fresh yeast ) during 4 hours . Then , adult flies were separated and embryos were collected following the suspension method described in Schou ( 2013 ) [68] . Embryo dechorionation was done by bleach ( 50% ) immersion . Total RNA was extracted using TRIzol Plus RNA Purification Kit ( Ambion ) . RNA was then treated on-column with DNase I ( Thermo ) during purification , and then treated once more after purification . 5’-RACE was performed with FirstChoice RLM-RACE Kit and using Small-scale reaction RNA processing with RNA samples of RAL-783 ( roo- ) , RAL-810 ( FBti0019985 ) , RAL-405 ( roo+7 ) , RAL-21 ( roo-90 ) , RAL-383 ( roo-44 ) and RAL-195 ( roo-44 ) . The gene specific outer primer was 5’-GACACTCTTCGGTTGGTGGA-3’ and the gene specific inner primer was 5’-ACAACTGTTCTGTAGGATCGC-3’ . The control primer was 5’-TAGTCCGCAGAGAAACGTCG-3’ . Inner PCR products were then cloned and Sanger-sequenced as mentioned above . Sequences have been deposited in GenBank under accession numbers KU672721-KU672722 . Embryo collection and RNA extraction was performed as described before . Reverse transcription was carried out using 500 ng of total RNA using Transcriptor First Strand cDNA Synthesis Kit ( Roche ) . The cDNA was then used in a 1/50 dilution for qRT-PCR with SYBR green master-mix ( Bio-Rad ) on an iQ5 Thermal cycler . CG18446 expression was measured using specific primers ( 5’-GAGCAGTTGGAATCGGGTTTTAC-3’ and 5’-GTATGAATCGCAGTCCAGCCATA-3’ ) spanning 99 bp cDNA in the exon 1/exon 2 junction of CG18446 . The primer pair efficiency was 99 , 1% ( r2 larger than 0 . 99 ) . CG18446 expression was normalized with Act5C expression levels ( 5’-GCGCCCTTACTCTTTCACCA-3’ and 5’-ATGTCACGGACGATTTCACG-3’ ) . Embryo collection was performed as mentioned above . Embryos were put into 50 ml fresh food vials . When embryos were 4–8 hour-old , they were kept at 1 C for 14 hours and then they were kept at room temperature ( 22–25 C ) . Simultaneously , control vials were always kept at room temperature ( 22–25 C ) and never exposed to cold-stress . A total of 8–20 vials were analyzed per experiment . The same number of embryos per vial , 30 or 50 , were used for all the replicates of a given experiment . Percentage viability was calculated based on the number of emerged flies to the total number of embryos placed in each vial . Statistical significance was calculated performing two-way ANOVA using SPSS v21 . We combined all the data into a full model: experimental condition ( stress and nonstress ) , insertion genotype ( presence/absence of the insertion ) and interaction between these two factors . For those experiments in which more than one replicate was performed , the replicate effect was also taken into account . Because our dependent variable was a proportion , we used the arcsine transformation of the data before performing statistical analysis . We tested whether the data was normally distributed using Kolmogorov-Smirnov test . When the data was not normally distributed after the arcsine transformation , we applied the rank transformation . When the statistical test was significant , we estimated partial eta-squared values as a measure of the effect size ( 0 . 01 small effect , 0 . 06 medium effect , and 0 . 14 large effect ) . We estimated the number of segregating sites ( S ) , Tajima´s D , iHS , nSL and XP-EHH in the 2 kb region flanking the FBti0019985 insertion ( chromosome 2R: 5758000–5760000 ) in 10 DGRP strains containing this insertion , in the 23 DGRP strains containing one of the roo insertions described in this work , and in the 15 strains that do not contain any insertion in the promoter region of CG18446 . Note that the coordinates of FBti0019985 in the r5 of the D . melanogaster genome used by the DGRP project to generate the vcf files are 2R: 5 , 758 , 595–5 , 759 , 028 . S and Tajima´s D are standard mesures of neutrality . iHS and nSL tests identify hard sweeps although they have some power to detect soft sweeps as well [69 , 70] . H12 tests for positive selection on new variation and standing genetic variation within a population , that is , it searches both for soft and hard sweeps in a population [71] . Finally , XP-EHH is a statistical test of positive selection in one population that uses between populations comparisons to increase power in regions near fixation in the selected population [72] . We have used vcftools to calculate the number of segregating sites , and Tajima´s D using parameters –maf 1/ ( 2n ) , where n is the sample size , and –remove-indels . We have obtained iHS , nSL , and XP-EHH using the selscan software with default parameters [73] . Finally , we have calculated H12 with ad hoc scripts . The four latter statistics require phased data . Thus , chromosome 2R of the 205 DGRP strains were phased together using ShapeIt [74] . To calculate the significance for the number of segregating sites , we resampled at random the same number of strains from the 205 DGRP strains available and calculated the distribution of segregating sites in the same 2 kb region . To calculate the significance of Tajima´s D , iHS , nSL and XP-EHH , we have used the empirical distributions of these statistics obtained from chromosome 2R . | The presence of several transposable element insertions in the promoter region of a Drosophila melanogaster gene has only been described in heat shock protein genes . In this work , we have discovered and characterized in detail several naturally occurring independent transposable element insertions in the promoter region of a cold-stress response gene in the fruitfly Drosophila melanogaster . The nine transposable element insertions described are clustered in a small 368 bp region and all belong to the same family of transposable elements: the roo family . Each individual insertion is present at relatively low population frequencies , ranging from 1% to 17% . However , the majority of strains analyzed contain one of these nine roo insertions suggesting that this region might be evolving under positive selection . Although the sequence of these insertions is highly similar , their molecular and functional consequences are different . Only one of them , FBti0019985 , is associated with increased viability in nonstress and in cold-stress conditions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"sequencing",
"techniques",
"invertebrates",
"gene",
"regulation",
"animals",
"invertebrate",
"genomics",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"genetic",
"elements",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
... | 2016 | Multiple Independent Retroelement Insertions in the Promoter of a Stress Response Gene Have Variable Molecular and Functional Effects in Drosophila |
4-anilino quinazolines have been identified as inhibitors of HCV replication . The target of this class of compounds was proposed to be the viral protein NS5A , although unequivocal proof has never been presented . A 4-anilino quinazoline moiety is often found in kinase inhibitors , leading us to formulate the hypothesis that the anti-HCV activity displayed by these compounds might be due to inhibition of a cellular kinase . Type III phosphatidylinositol 4-kinase α ( PI4KIIIα ) has recently been identified as a host factor for HCV replication . We therefore evaluated AL-9 , a compound prototypical of the 4-anilino quinazoline class , on selected phosphatidylinositol kinases . AL-9 inhibited purified PI4KIIIα and , to a lesser extent , PI4KIIIβ . In Huh7 . 5 cells , PI4KIIIα is responsible for the phosphatidylinositol-4 phosphate ( PI4P ) pool present in the plasma membrane . Accordingly , we observed a gradual decrease of PI4P in the plasma membrane upon incubation with AL-9 , indicating that this agent inhibits PI4KIIIα also in living cells . Conversely , AL-9 did not affect the level of PI4P in the Golgi membrane , suggesting that the PI4KIIIβ isoform was not significantly inhibited under our experimental conditions . Incubation of cells expressing HCV proteins with AL-9 induced abnormally large clusters of NS5A , a phenomenon previously observed upon silencing PI4KIIIα by RNA interference . In light of our findings , we propose that the antiviral effect of 4-anilino quinazoline compounds is mediated by the inhibition of PI4KIIIα and the consequent depletion of PI4P required for the HCV membranous web . In addition , we noted that HCV has a profound effect on cellular PI4P distribution , causing significant enrichment of PI4P in the HCV-membranous web and a concomitant depletion of PI4P in the plasma membrane . This observation implies that HCV – by recruiting PI4KIIIα in the RNA replication complex – hijacks PI4P metabolism , ultimately resulting in a markedly altered subcellular distribution of the PI4KIIIα product .
Hepatitis C virus ( HCV ) is an enveloped , single-stranded RNA virus classified as member of the Hepacivirus genus within the Flaviviridae family . The 9 . 6 kb positive-sense RNA genome contains a single open reading frame encoding a polyprotein of about 3 , 000 amino acids , flanked by highly structured 5′ and 3′ untranslated ( UTR ) regions . Following its release into the cytoplasm of the host cell , viral RNA is translated via an internal ribosome entry site ( IRES ) , giving rise to a single polypeptide that is cleaved into 10 different mature protein products: Core , gpE1 , gpE2 , p7 , NS2 , NS3 , NS4A , NS4B , NS5A , and NS5B . HCV RNA replication takes place in the cytoplasm , in association with a virus-induced intracellular membrane structure termed “membranous web” , onto which NS proteins assemble to form the so-called RNA replication complexes . It is estimated that 3% of the world's population are chronically infected by the hepatitis C virus ( HCV ) . Most infections become chronic and over time evolve into chronic hepatitis . The most unwanted complication of chronic hepatitis is cirrhosis , a massive liver fibrosis , which can lead to liver failure and hepatocellular carcinoma . Since the discovery of hepatitis C virus ( HCV ) in the late 1980's much progress has been made in the understanding of the viral life cycle of HCV . Nonetheless , to date no vaccines are available and the current standard of care , involving lengthy treatment with a combination of ribavirin and pegylated interferon-α ( peg-IFN-α ) , eradicates the infection in half of treated patients . A large effort has been made in the past two decades in order to develop novel anti-HCV therapies with greater efficacy . Two oral direct-acting antiviral agents ( DAA ) targeting the HCV NS3/4 protease , boceprevir and telaprevir , have recently reached the market and more are being developed [1] . While the initial efforts to the discovery of DAA focused almost exclusively on the best characterized HCV enzymes required for viral replication – the NS3/4A protease and the NS5B polymerase – in the past few years the NS5A viral protein has been attracting more and more attention as a target for drug development [1] , [2] . NS5A possesses no known enzymatic activity . It is a multifunctional non-structural protein important for viral replication [3]–[6] as well as viral assembly [7]–[9] . It is a phosphoprotein consisting of three domains [10] . Domain I is highly conserved and forms a dimeric structure [11] , [12] , whereas domains II and III are believed to adopt a “natively unfolded” conformation [13] , [14] . In recent years , several anti-HCV compounds identified using cell-based replicon screens were indicated to target NS5A based on the analysis of the mutations associated with emergence of resistance in the replicon system [15]–[17] . The most studied series of these “NS5A inhibitors” is represented by BMS-790052 , an agent that is leading the field , having demonstrated potent antiviral activity in clinical studies [18] . Compounds in this class are characterized by a complex , dimeric or pseudo-dimeric structure and a high molecular weight , when compared with conventional “drug-like” small molecules [17] , [19] . Resistance mutations against these compounds emerge readily in domain I of NS5A [20] , with the most recurrent of these changes corresponding to variant of tyrosine at position 93 [20] . Although direct interaction with purified NS5A has not been demonstrated , compelling reverse genetic experiments [20] as well as molecular models [15] , [21] strongly support the notion that NS5A is the direct target of these compounds . A less characterized series of compounds , belonging to a different chemical class , was also initially indicated to target NS5A on the basis of the mutation pattern observed in resistant replicons [21] . The common structural element of this latter class of inhibitors is a 4-anilino quinazoline core . A representative member of this class of compounds is A-831/AZD-2836 , an experimental antiviral agent that entered clinical trials but was later discontinued due to the lack of adequate exposure [17] . For these agents , the mutations reported to be associated with resistance were found to be different from those expected for the NS5A inhibitor described above , pointing to a different mechanism of action: a few mutations were found at the C-terminal end of NS5A domain I ( E212D , L199F and T200P ) , whereas most mutations occurred in NS5A domains II and III ( P299L , S370P , V388D , V362A , S390G and S370P ) . Additional mutations were also found in NS4B ( S258T ) and NS5B ( S76A ) [17] , [21] , [22] . Reverse genetics studies in which these mutations were reintroduced in the replicon , however , did not recapitulate the resistant phenotype observed in the original cellular clones [17] , leaving thus the possibility open that these compounds act through a different viral or cellular target . Interestingly , many kinase inhibitors , including some approved antitumoral drugs ( gefitinib , lapatinib , erlotinib ) are 4-anilino quinazoline derivatives [23]–[25] . Altogether , these considerations led us to investigate whether the anti-HCV activity displayed by these compounds might be due to inhibition of a cellular kinase . Recently , several small-interfering RNA ( siRNA ) screening campaigns have identified type III phosphatidylinositol 4-kinases ( PI4K ) as crucial host factors for HCV replication . In particular , PI4KIIIα was found to be required for HCV RNA replication in a cell line- and genotype-independent manner , whereas the requirement for the β isoform was observed to be less dramatic and limited to Con-1 ( genotype 1b ) replicons [26]–[29] . It was shown that the catalytic activity of PI4KIIIα is required to rescue HCV replication in cells with a stable knock-down of PI4KIIIα . In addition , it has been proposed that NS5A stimulates PI4KIIIα activity by direct interaction via domain I [30]–[32] . All these observations taken together made us consider the phosphatidylinositol 4-kinases a potential alternative candidate target for 4-anilino quinazoline inhibitors of HCV replication . In this paper , we present evidence that AL-9 , a member of this class of compounds previously reported to target NS5A , inhibits PI4P formation by direct inhibition of phosphatidylinositol 4-kinase IIIα ( PI4KIIIα ) . In addition , we provide evidence that pharmacological inhibition of PI4KIIIα with AL-9 results in altered subcellular distribution of NS5A similar to that observed after RNAi knock-down of the PI4KIIIα mRNA , strongly supporting a mechanism of HCV inhibition mediated by the inhibition of PI4KIIIα . Moreover , we show that HCV subverts components of the phosphatidylinositol-4 phosphate ( PI4P ) pathway to function in favor of its own life cycle , thereby enriching the PI4P concentration in the membranous web while depleting the plasma membrane PI4P pool .
AL-9 is a member of 4-anilino quinazoline-containing HCV replication inhibitors described previously ( [21]; Figure 1 ) . In order to confirm its anti-HCV activity , we tested the effect of this compound on HCV replication in Huh7 . 5 cells stably expressing genotype 1b or 2a subgenomic replicons ( Con1-SR and JFH-A4 , respectively ) . The EC50 values , calculated by measuring viral RNA after incubation with AL-9 for three days , are reported in Table 1 . Replicon EC50 values for AL-9 were found to be 0 . 29 µM and 0 . 75 µM for genotype 1b and 2a , respectively . In order to prove that AL-9 inhibits HCV replication not only in the context of a HCV subgenomic replicon , but also in the context of the complete viral life-cycle , we determined the inhibitory activity using the J6/JFH-1 HCV virus . In this case , the EC50 value was found to be 1 . 2 µM , a figure comparable with the result obtained with genotype 2a subgenomic replicon . CC50 values are shown for Con1-SR , JFH-A4 and Huh7 . 5 cells , respectively . In summary , AL-9 inhibits HCV across different genotypes with activity in the sub-micromolar to low micromolar range in the absence of significant cytotoxic effects . In the following experiment , we investigated whether AL-9 inhibits the purified type III phosphatidylinositol 4-kinases PI4KIIIα and PI4KIIIβ ( Figure 2 ) . Both enzymes were inhibited by AL-9 with a five-fold preference for PI4KIIIα ( IC50 of 0 . 57 µM and 3 . 08 µM , respectively ) . This result demonstrates that AL-9 inhibits type III PI4 kinases in vitro at concentrations similar to those required for its anti-HCV activity , displaying a moderate selectivity for the α over the β isoform . We also tested the activity of AL-9 on two class I PI3-kinases ( p110α and p110β ) . While PI3-kinase p110α was inhibited with an IC50 of 1 . 1 µM , the potency of AL-9 for PI3-kinase p110β was significantly lower ( 40% inhibition @10 µM , data not shown ) . Our hypothesis is that AL-9 inhibits HCV replication via inhibition of PI4KIIIα . Thus , we wanted to assess whether AL-9 also inhibited PI4KIIIα in living cells . To this aim , we needed to set up an assay that allowed us to monitor the activity of this kinase in intact cells . PI4KIIIα is primarily localized to the ER , whereas PI4KIIIβ is localized to the Golgi membranes [33] . It was shown that PI4KIIIβ contributes to the synthesis of PI4P at the Golgi membranes [34] , [35] . Subcellular localization of the enzymes , however , does not always coincide with their function . Thus , PI4KIIIα , considered to be an ER-resident enzyme , has previously been shown to be critical for the generation and maintenance of the plasma membrane PI4P pool during phospholipase C activation and Ca2 signaling in HEK-293 or Cos-7 cells [35] , [36] as well as in resting Cos-7 cells [37] . Whether PI4KIIIα is responsible for the maintenance of the plasma membrane PI4P pool under normal cell culture conditions in hepatoma cells is currently not known . Hammond et al [37] have developed immunocytochemical techniques that enable selective staining of the PI4P pool present in the plasma membrane ( plasma membrane staining protocol ) or in the intracellular membranes ( Golgi staining protocol ) , respectively . We used this technique , in combination with RNA gene silencing or pharmacological inhibition , to decipher which of the type III enzymes participates in the synthesis of the Golgi- or plasma membrane PI4P-pools in Huh7 . 5 hepatoma cells . To address which type III PI4 kinase is responsible for the synthesis of the different cellular PI4P pools , Huh7 . 5 cells were treated with siRNAs targeting PI4KIIIα , PI4KIIIβ or an unrelated siRNA ( mock-siRNA ) as described in the Materials and Methods section . Immunoblots assays show specific knockdown of PI4KIIIα or PI4KIIIβ by their corresponding siRNAs ( Figure 3C ) . Three days after siRNA treatment , PI4P was revealed either by the plasma membrane staining protocol ( Figure 3A , upper panel ) or by the Golgi membrane staining protocol ( Figure 3A , lower panel ) . In cells treated with the unrelated siRNA ( mock-siRNA ) , PI4P was detected both in the plasma membrane and in intracellular membranes . Intracellular PI4P was localized primarily in the Golgi membranes , as judged by the colocalization with the Golgi marker giantin . Silencing of PI4KIIIα resulted in a significant decrease of the PI4P level in the plasma membrane . Concomitantly with the decrease in the plasma membrane PI4P levels , we consistently observed a pronounced increase of PI4P level in the Golgi membrane following PI4KIIIα knockdown . In the case of PI4KIIIβ knockdown , we observed a ∼30% decrease of Golgi membrane PI4P level , whereas the PI4P levels of the plasma membrane remained substantially unaffected ( Figure 3B ) . These results are in line with the previously reported role for PI4KIIIα in maintaining the PI4P plasma membrane pool [35]–[37] and confirm the importance of PI4KIIIβ for the synthesis of at least part of the Golgi membrane PI4P [34] , [35] . We also observed that decreased expression of PI4KIIIα resulted in an unexpected increase in the level of the Golgi membrane pool ( Figure 3A ) , suggesting a complex level of cross-talk between the cellular type III PI4 kinases in maintaining the physiological PI4P levels at the Golgi membrane , at least in our experimental model . In order to confirm and extend the results described above , we utilized a known pharmacological inhibitor of the type III PI4 kinases . PIK93 was previously exploited to distinguish between the roles of the two PI4KIII isoforms [38] , [39] . In particular , a concentration of 0 . 5 µM PIK93 is expected to affect only PI4KIIIβ , whereas 30 µM PIK93 should inhibit both PI4KIIIβ and PI4KIIIα . Thus , Huh7 . 5 cells were treated with 0 . 5 µM or 30 µM PIK93 or with DMSO as control . After two hours of incubation , PI4P was revealed either by the plasma membrane staining protocol ( Figure 4A , upper panel ) or by the Golgi staining protocol ( Figure 4A , lower panel ) . PI4P levels associated with the Golgi membranes decreased by ∼25% after incubation with 0 . 5 µM PIK93 ( Figure 4B ) . This is in line with PI4KIIIβ contributing to the production of PI4P present in the Golgi membranes ( PI4KIIα , another contributor of Golgi-localized PI4P is not inhibited by PIK93 [38] , [39] ) . Increasing PIK93 concentration to 30 µM further increased the inhibition of the intracellular membrane PI4P pool , to ∼65% ( Figure 4B ) . This could be due to a more complete inhibition of PI4KIIIβ; however , based on this experiment , we cannot rule out a contribution of PI4KIIIα activity to the maintenance of the Golgi membrane PI4P pool . In contrast to what observed in the Golgi-associated membranes , the plasma membrane PI4P level was not significantly affected upon incubation with 0 . 5 µM PIK93 , but decreased by nearly 50% after incubation with 30 µM of PIK93 ( Figure 4B ) . Combined with the RNAi experiments described above , these results support the notion that , in Huh7 . 5 cells , PI4KIIIα is involved in the maintenance of the plasma membrane PI4P pool , whereas PI4KIIIβ is at least partly responsible for the maintenance of the Golgi membrane PI4P pool . We then evaluated the PI4K inhibitory activity of AL-9 in Huh7 . 5 cells using the same methodology . Briefly , Huh7 . 5 cells were incubated either with DMSO or with increasing concentration of AL-9 ( 1 , 2 , 4 or 8 µM ) for two hours ( Figure 5A ) . Treatment with AL-9 gradually reduced the amount of PI4P in the plasma membrane ( Figure 5B ) . Conversely , the concentration of PI4P in the Golgi-associated membranes remained substantially unaltered up to the highest AL-9 concentration used ( Figure 5B ) . In all , the results described above suggest that AL-9 inhibits PI4KIIIα also in living cells , while not appreciably affecting the activity of PI4KIIIβ . This is in line with the selectivity for PI4KIIIα over PI4KIIIβ observed in the biochemical assays . Viral infection induces modification of intracellular membrane structures [40] and , for some RNA viruses including HCV , it has been shown that these induced membranous structures are highly enriched for PI4P [32] , [41] . Before testing the activity of AL-9 in HCV-infected cells , we wanted to know what the impact of HCV on cellular membrane structures was , with special regard to the subcellular membrane distribution of PI4P . Naïve Huh7 . 5 cells or cells actively replicating the genotype 2a or 1b HCV subgenomic replicon were investigated for their PI4P concentration in internal membranes or plasma membranes , respectively ( Figure 6A ) . As previously shown , cells expressing the HCV replicon form a membranous web that is highly enriched for PI4P ( Figure 6A , lower panel ) . The level of PI4P in these virus-specific membrane structures is markedly higher in JFH-A4 cells , containing the very efficient genotype 2a JFH-1 replicon , compared to the Con1-SR cells , which are based on the genotype 1b Con1 replicon , possibly mirroring the different RNA replication efficiency . It is well established that the kinase responsible for the production of the PI4P pool present in these structures is PI4KIIIα . In the current model , PI4KIIIα interacts with the viral protein NS5A , leading to up-regulation of the kinase activity and accumulation of PI4P in the virus-specific membranous web [30]–[32] . Conversely , the results shown in Figures 3–5 suggest that – in absence of viral replication – a major function of PI4KIIIα is the synthesis of the PI4P pool in the plasma membrane . We therefore asked ourselves whether the presence of HCV could not only influence distribution and enrichment of PI4P in internal membranes , but also alter the PI4P plasma membrane pool . In Figure 6A , we show that , concomitantly with the increase of PI4P in the internal membranes ( lower panel ) , HCV replication promotes a marked decrease of PI4P concentration in the plasma membrane ( upper panel ) . Relative quantification of the PI4P levels in the different experimental conditions is shown in Figure 6B . This experiment demonstrates that the presence of HCV causes a dramatic change of PI4P localization in cellular membranes , whereby the increase of PI4P concentration in the virus-specific membranous structures appears to be accompanied by a depletion of the PI4P pool normally present in the plasma membrane . We next investigated whether HCV-associated changes in PI4P distribution could be reverted upon cure of the HCV replicon by specific inhibitors . We treated JFH-A4 cells for two weeks either with the HCV RdRP inhibitor HCV-796 or with the HCV NS3/4A protease inhibitor MK-5172 and followed PI4P localization in internal membranes and in the plasma membrane ( Figure 7 ) . Independent of the type of inhibitor used , the result shows that the HCV-induced PI4P-enriched membranous web in JFH-A4 cells disappeared upon suppression of HCV replication and that the intracellular PI4P localization returned to the Golgi-localization as observed in the naïve Huh7 . 5 cells ( left column ) . In parallel , the plasma membrane concentration of PI4P increases to the levels observed in naïve cells ( middle column ) . NS5A staining ( right column ) as well as real-time RT-PCR ( not shown ) indicated that the prolonged treatment with HCV-inhibitor led to complete and stable suppression of viral protein expression and undetectable level of HCV RNA . Thus , removal of HCV RNA brings PI4P synthesis and distribution back to a level comparable to naïve Huh7 . 5 cells . It is worth of note , however , that the previous presence of HCV replicons in the cured cells induced some irreversible morphological changes of unknown nature , such as a smaller cell size . We have shown that PI4KIIIα is inhibited by AL-9 in naïve Huh7 . 5 cells . As discussed above , in HCV-replicating cells , the kinase activity of PI4KIIIα is up-regulated by a direct protein-protein interaction with the viral protein NS5A [30] , [32] . In the following experiment ( Figure 8 ) , we explored whether AL-9 is able to inhibit PI4KIIIα also in this context . JFH-A4 cells were incubated with increasing concentration of AL-9 for 4 hours and PI4P concentration in the HCV membranous web was followed by immunostaining ( Golgi staining protocol ) . Treatment of cells with AL-9 lead to clear inhibition of PI4P accumulation in the HCV membranous web . Incubation with 8 µM AL-9 depleted as much as 70% of the PI4P present in the intracellular membranes of replicon-containing cells . This result confirms that AL-9 inhibits PI4KIIIα independent of its membranous localization and suggests that this inhibition could be responsible for the observed antiviral effect . Since AL-9 has anti-HCV activity in the concentration range used here , longer incubation of HCV replicons with AL-9 results in inhibition of HCV RNA- and protein-synthesis . As a consequence , the PI4P-enriched HCV membranous web would disintegrate . In this case loss of PI4P in the internal membranes could be not a direct consequence of PI4KIIIα inhibition , but a consequence of disintegration of the HCV membranous web . In order to rule out this possibility , we checked localization of NS5A , a presumed marker for HCV replication sites , after 4 hours of incubation with AL-9 . Localization of NS5A does not change , suggesting that AL-9 does not significantly change the structure of the HCV membranous web upon 4 hours of treatment . Moreover , incubating the same replicon cells for 4 hours with HCV-796 , an HCV polymerase inhibitor , did not lead to appreciable depletion of the membranous web PI4P pool indicating that the loss of PI4P in the HCV-induced intracellular membranes is the direct consequence of inhibition of PI4KIIIα , and not the consequence of inhibition of HCV replication . Additional evidence is provided in the experiment below , in which expression of the HCV polyprotein , and consequently formation of a membranous web , was driven by cDNA plasmid rather than by autonomously replicating HCV RNA . It was previously shown that knock-down of PI4KIIIα expression by RNAi resulted in the production of large NS5A clusters . This was achieved in an experimental setting where the HCV polyprotein was expressed from DNA constructs , thus avoiding potential confounding effects due to inhibition of HCV RNA replication [27] , [32] . We wanted to assess whether pharmacological inhibition of PI4KIIIα kinase activity would lead to similar effects on NS5A subcellular localization . Thus we followed the effect of AL-9 on NS5A localization after transient DNA transfection in Huh7-Lunet/T7 cells with a plasmid expressing genotype 2a nonstructural proteins NS3-NS5B under the control of a T7 promoter [42] . Cells were treated either with DMSO ( upper panels ) or with 8 µM AL-9 ( lower panels ) for 2 , 8 or 16 hours and localization of NS5A as well as PI4P were followed by indirect fluorescence microscopy ( Figure 9 ) . Cells successfully transfected with the HCV polyprotein expressed NS5A and induced the PI4P-enriched membranous web . After 8 hours of treatment with AL-9 , changes in NS5A localization in form of larger clusters become visible . At the same time , PI4P concentration in the membranous web started to decrease . After 16 hours of incubation with AL-9 , NS5A was concentrated almost exclusively in large clusters . At this time-point , PI4P in the internal membranes had completely vanished . In summary , this experiment shows that , in cells expressing the HCV polyprotein from cDNA , prolonged treatment with AL-9 results in a redistribution of NS5A into large clustered structures with high resemblance to the structures previously observed after silencing of the PI4KIIIα gene by RNAi [27] , [32] . Concomitantly , we observed a depletion of the PI4P pool present in the HCV-induced membranous structures . These results indicate that the catalytic activity of PI4KIIIα is directly or indirectly required for the proper localization of HCV NS5A protein into the membranous web . Furthermore , the experiment just described lands additional support to the notion that the antiviral effect of AL-9 is mediated by the inhibition of PI4KIIIα .
In the present paper , we show that a compound belonging to the class of 4-anilino quinazoline inhibitors of HCV replication is an inhibitor of PI4KIIIα , a cellular lipid kinase required for viral replication . PI4KIIIα belongs to the family of type III phosphatidylinositol 4-kinases , enzymes that catalyze the conversion of phosphatidylinositol to phosphatidylinositol 4-phosphate ( PI4P ) . PI4P is the most abundant monophosphorylated inositol phospholipid in mammalian cells and the importance of this phospholipid is just started to be unraveled [43] . In addition to playing important roles in intracellular signaling and membrane trafficking , phosphatidylinositol lipids and their metabolizing enzymes are also exploited by many different viruses in order to transform cellular membranes in structures supporting their replication [40] , [44] , [45] . PI4KIIIβ was shown to be a host factor required for enterovirus replication [41] , whereas several reports have demonstrated that PI4KIIIα is crucial for HCV replication [26]–[29] . Owing to the importance of this pathway , the need for specific inhibitors of PI4III kinases is increasing . Only recently , some enviroxime-like compounds with antiviral activity against enterovirus have been demonstrated to target PI4KIIIβ . One of these agents is a very specific inhibitor of the β-isoform of the type III PI4-kinases [46] . So far , no such compound exists for the PI4KIII-α isoform . A commonly used inhibitor for type III phosphatidylinositol 4-kinases is PIK93 , which has originally been designed to inhibit class I PI3-kinases [38] . This compound allows differential inhibition of PI4KIIIβ alone or PI4KIIIα and PI4KIIIβ together depending on the concentration used . In this paper , we show that a 4-anilino quinazoline derivative , termed AL-9 ( Figure 1 and Figure S1 ) , is able to inhibit PI4KIIIα in a test tube as well as in living cells . AL-9 inhibited purified PI4KIIIα , with a moderate ( ∼5-fold ) selectivity over the β isoform ( Figure 2 ) . In cell culture , we observed that treatment with AL-9 efficiently inhibits the maintenance of the plasma membrane PI4P pool in Huh7 . 5 cells while not significantly affecting the Golgi membrane pool at the highest concentration used ( Figure 5 ) . This finding is in line with the moderate selectivity observed in the biochemical assay . Thus , AL-9 represents a lead candidate for the development of more potent and more specific inhibitors of PI4KIIIα . Anti-HCV compounds of the 4-anilino quinazoline class were previously assumed to exert their antiviral effect via inhibition of the viral protein NS5A . This conclusion rested on analysis of the mutations found in the HCV replicon in association with resistance to these agents [21] . Mutations generated against 4-anilino quinazolines were localized mainly in NS5A , in triplets that occurred all in NS5A or appeared concomitant with changed in NS4B or NS5B [17] , [22] ( see also Introduction ) . Reverse genetic experiments , in which these mutations were reintroduced in the replicon ( single , double and triple combinations ) , however , did not support a role for these mutations in conferring resistance to 4-anilino quinazolines [17] . In order to assess whether the reported mutations conferred any level of resistance to AL-9 , we independently performed reverse genetics studies in which selected mutations triplets , reported to be associated with the higher level of resistance , were reintroduced in a genotype 1b replicon with the same genetic background as the one reported in the original resistance study ( Figure S2 ) . These mutation triplets are: FAG: L199F+V362A+S390G ( NS5A ) , DLD: E212D+P299L+V388D ( NS5A ) , and PPA: T200P+S370P ( NS5A ) +S76A ( NS5B ) . We observed that the replicon containing the first triplet lost the ability to replicate at significant level . For replicons containing the latter two combinations of mutations , RNA replication could be measured , although at a lower level compared to the parental construct ( 35% and 20% , respectively ) . These replicons , however , remained equally sensitive to AL-9 as the parental replicon ( Figure S2 ) , opening the question as to which really is the target of this compound class . We are currently trying to select HCV replicons resistant to AL-9 . So far we were unable to identify mutations that confer resistance to AL-9 . Our new data on AL-9 suggest that inhibition of HCV replication by 4-anilino quinazoline compounds is a consequence of PI4KIIIα inhibition . Our conclusion rests on a number of experimental findings . First of all , we showed that AL-9 is an inhibitor of purified type III PI4 kinases . Furthermore , we clearly demonstrated that AL-9 inhibits PI4KIIIα both in naïve Huh7 . 5 cells ( Figure 5 , discussed above ) as well in cells harboring actively replicating HCV RNA ( Figure 8 ) . In cells where HCV replication occurs , PI4KIIIα interacts physically with HCV NS5A . This interaction , in turn , leads to the stimulation of PI4P synthesis at the HCV replication sites [32] . Treatment of replicon-harboring cells with AL-9 leads to efficient suppression of the PI4P pool at the HCV replication sites and does so independently of inhibition of HCV replication . This indicates that – although the enzymatic activity of PI4KIIIα is modulated by the interaction with the HCV protein NS5A – it remains sensitive to the action of the 4-anilino quinazoline inhibitor . We also investigated whether the dramatic changes observed in PI4P membrane levels by treatment with AL-9 could be associated with alteration in the subcellular distribution of type III PI4 kinases . To this aim , we analyzed the subcellular distribution of the type III PI4 kinases in Huh7 . 5 or Luc-A4 cells following incubation with AL-9 ( Figure S3 ) . We observed no major effect of AL-9 on the localization of either PI4KIIIα or PI4KIIIβ , in line with the notion that the observed effects are primarily due to the inhibition of the kinase activity rather than to an altered protein subcellular distribution . In cells that express the HCV polyprotein from a trans-gene , knock-down of PI4KIIIα by RNAi was previously shown to cause a dramatic change in NS5A subcellular distribution , from a pattern consistent with localization in the membranous web replication complexes to abnormally large cytoplasmic clusters [27] , [30] , [32] . In Figure 9 , we show that AL-9 treatment of cells ectopically expressing the HCV nonstructural proteins results in a time-dependent depletion of PI4P and a concomitant change of NS5A localization to the large-clustered structures discussed above , reinforcing the notion that the anti-HCV effect of AL-9 and related compounds are likely to be mediated by the inhibition of PI4KIIIα . We also found that PI3K p110α is inhibited by AL-9 in vitro at concentration similar to those needed to inhibit type III PI4-kinases . However , no Class I PI3-kinase has been shown to influence HCV replication thus inhibition of HCV replication by AL-9 is not due to inhibition of Class I PI3-kinases . So far , the only PI3-kinase that resulted as positive hit for HCV replication inhibition in siRNA screens is PI3-kinase C2 gamma [28] . Future work will have to address whether AL-9 inhibits PI3KC2G in addition to Type III PI4-kinases . During the characterization of AL-9 we focused our attention on various aspects of PI4P metabolism in Huh7 . 5 cells with and without replicating HCV . We observed a typical Golgi localization of PI4P in intracellular membranes of naïve Huh7 . 5 cells and confirmed a role for PI4KIIIβ in maintaining at least part of this pool . In order to get the complete picture we also investigated the PI4P pool present in the plasma membrane . In yeast , Stt4p , the ortholog to the mammalian PI4KIIIα , is localized at the plasma membrane and it is the major contributor for the synthesis of the plasma membrane-localized PI4P [43] . In mammalian cells , the role of PI4KIIIα for the maintenance of the plasma membrane PI4P pool has been demonstrated in HEK-293 and Cos-7 cells [35]–[37] . Here we demonstrate that liver-derived Huh7 . 5 cells are endowed with a rich PI4P pool in the plasma membrane and that the enzyme responsible for its maintenance is PI4KIIIα . In HCV-replicating cells , the subcellular PI4P distribution is profoundly altered . As already reported previously , the presence of HCV causes the induction of a membranous web highly enriched for PI4KIIIα-syntesized PI4P . In accordance , several reports demonstrate that NS5A recruits PI4KIIIα to the membranous web by direct protein-protein interaction , thereby stimulating its enzymatic activity [30]–[32] . Concomitantly with the induction of highly PI4P-enriched internal membranes , we observe a marked decrease of PI4P in the plasma membrane . One possible explanation could be that – by hijacking PIKIIIα – HCV might be able to enrich PI4P in the virus-induced membranous web not only by directly activating the enzymatic activity of PI4KIIIα recruited into the HCV RNA replication compartment , but also by preventing transport of the PI4KIIIα-synthesized PI4P from the synthesis site to the plasma membrane . How PI4KIIIα , localized at the ER , synthesizes the PI4P pool present in the plasma membrane it is still an enigma . This topological discrepancy can partially be resolved assuming that PI4KIIIα-dependent PI4P production occurs on ER-PM contact sites , that is , sites of close apposition between ER and PM . In yeast it has been demonstrated that a complex interplay between different proteins regulate the PI4P metabolism at the plasma membrane [47] . Among these proteins are Osh , the yeast ortholog of the human OSBP and the ER membrane VAP proteins Scs2 and Scs22 , the yeast orthologs of human VAP proteins . Interestingly , h-VAP-33 and OSBP have been shown to be important for HCV replication [48]–[50] . It may be possible that recruitment of PI4KIIIα to the HCV membranous web through NS5A prevents interaction of PI4KIIIα with its cellular protein partners required to direct PI4P to the plasma membrane . Upon withdrawal of HCV from the cells ( Figure 7 ) PI4KIIIα is again free for interaction with the adequate partners . A possible role of PI4KIIIα in PI4P trafficking between the plasma- and intracellular membranes is suggested by our finding that RNAi silencing of this PI4 kinase results in decreased concentration of PI4P in the plasma membrane with a concomitant increase in the level of PI4P in the endomembranes ( Figure 3 ) . Such a function of PI4KIIIα would have to be independent of the kinase activity , since pharmacological inhibition ( with PIK93 or AL-9 ) does not recapitulate this phenomenon observed by knocking down the protein expression . In summary , the presence of HCV may change PI4P metabolism not only by activating the catalytic activity of PI4KIIIα by NS5A but also by modulating the PI4P distribution between different membrane compartments . The net result is an enrichment of the PI4P pool in the HCV-induced membranous web with a concomitant depletion of the plasma membrane PI4P pool . Concluding , in this paper we demonstrate that a class of HCV inhibitors originally proposed to target NS5A does in fact target the host factor PI4KIIIα . Compounds targeting host factors may have the general advantage of imposing a higher genetic barrier to the development of resistance . AL-9 , a member of this class of compounds , inhibits PI4KIIIα and to our knowledge , it is the first compound with a clear preference for PI4KIIIα over PI4KIIIβ . For this reason , AL-9 offers a good candidate as lead compound for the development of more potent and specific pharmacological inhibitors of PI4KIIIα to be used both as important research tools as well as leads for initial drug discovery .
The HCV RNA polymerase inhibitor HCV-796 and the PI kinase inhibitor PIK93 were a gift from Arrow Pharmaceuticals . The HCV protease inhibitor MK-5172 was purchased from Selleck Chemicals . Nucleic acids were manipulated according to standard protocols . Plasmid FBac-His-CD-PI4KA was constructed as follows: the catalytic domain of PI4KIIIα was amplified by PCR using the oligonucleotides 5′-CACTGCGGATCCATAATGGGGATGATGCAGTGTGTGATTG-3′ ( sense ) , 5′-CCTGCGAATTCTCAGTAGGGGATGTCATTC-3′ ( antisense ) and the plasmid pEF1A-PIK4CA untagged ( a kind gift from G . Randall , Department of Microbiology , University of Chicago ) as template . The resulting PCR fragment was subcloned into the vector pCR-Blunt II-Topo ( Invitrogen ) and finally cloned into the BamH1–XhoI cloning sites of the plasmid vector pFastBac THT-B . The resulting protein expressed from this plasmid contains an N-terminal hexa-histidine tag and starts at PI4KIIIα amino acid G873 ( reference sequence NM_058004 ) . pTM-NS3-5B expression vector expressing the HCV genotype 2a nonstructural proteins under the control of the T7 promoter was a generous gift from V . Lohmann ( Department of Molecular Virology , University of Heidelberg ) [42] . Synthesis of compound AL-9 is described in Supporting Information . The human hepatoma-derived cell line Huh7 . 5 [51] were grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum , 100 U/ml penicillin , 100 µg/ml Streptomycin and 2 mM L-glutamine; G418 ( 0 . 8 mg/ml ) was added to cell lines containing the HCV replicon . Stable cell lines expressing HCV genotype 1b or 2a subgenomic replicons were generated by electroporation of in vitro-transcribed RNA into Huh7 . 5 cells [52] and following selection with G418 ( 0 . 8 mg/ml ) for three weeks . Con1-SR: Huh7 . 5 cells replicating the Con1 subgenomic replicon with the adaptive mutations E1202G in NS3 and S2204R in NS5A . JFH-A4: Huh7 . 5 cells replicating the JFH-1 subgenomic replicon together with the luciferase reporter gene constructed as described previously [53] . JFH-A4 cells were cured from the HCV replicon by two weeks of treatment with the protease inhibitor MK-5172 ( 0 . 2 µM ) or the HCV RNA polymerase inhibitor HCV-796 ( 2 µM ) , respectively . Huh7-Lunet/T7 cells were a kind gift from V . Lohmann ( Department of Molecular Virology , University of Heidelberg , Germany ) . For replication assays , JFH-A4 or Con1-SR cells were plated at the density of 3×104 or 6×104 cells/well , respectively , in 24-well dishes the day before the experiment . Cells were treated with AL-9 resulting in a final concentration of 1% DMSO in the cell medium . After three days of treatment , RNA was extracted using the RNeasy Mini Kit ( Qiagen ) and HCV RNA was quantified by real time PCR using the following oligonucleotide and probe set designed for the HCV IRES as described previously [52]: sense ( 5′-GCGAAAGGCCTTGTGGTACT-3′ ) , antisense ( 5′-CACGGTCTACGAGACCTCCC-3′ ) , and probe ( 5′-CCTGATAGGGTGCTTGCGAGTGCC-3′ , 5′ 6-carboxyfluorescein [FAM]/3′ 6-carboxytetramethylrhodamine [TAMRA] ) . GAPDH mRNA was used as internal control for data normalization . Production of infectious virus was performed as follows: J6/JFH-1 chimeric RNA ( 1-846 ( J6CF ) /847-3034 ( JFH1 ) was electroporated into Huh7 . 5 cells using the protocol described previously [52] . Briefly , 2×106 cells were electroporated with 10 µg of RNA in a final volume of 200 µl and 4×106 cells were plated in a T-75 flask . Three days post electroporation , medium was harvested and stored at −20°C in small aliquots . Calculation of EC50 of AL-9 using the infectious HCV virus was performed as follows: Huh7 . 5 cells were plated at 4×104 cells/well in 24-well plates the day before infection . Infection was started by addition of 10 µl of cell medium containing infectious virus ( see above ) at an MOI of 50 in a final volume of 400 µl . After 6 hours of incubation , medium was removed and replaced with 400 µl of fresh medium containing serial dilutions of AL-9 . RNA was collected after 72 hours of incubation and quantified by real time PCR . Cell cytotoxicity ( CC50 ) of AL-9 was calculated using the cell viability assay CellTiter-Blue ( Promega ) . Huh7 . 5 , JFH-A4 , or Con1-SR cells ( 5×103 cells/well in 96-well dishes ) were plated the day before treatment . AL-9 was added and cell viability was measured after four days of treatment . Recombinant baculovirus was generated with the plasmid FBac-His-CD-PI4KA using the Bac-to-Bac system following the instructions of the manufacturer ( Invitrogen ) . For protein expression , Sf9 cells were infected with recombinant baculovirus at a density of 2×106 Sf9 cells/ml for 3 days at 20°C . To prepare cell extract ( 1 . 5×108 cells ) , cells were incubated in hypotonic buffer ( 10 mM HEPES ( pH 7 . 5 ) , 10 mM NaCl , 1 mM Tris ( 2-carboxyethyl ) phosphine ( TCEP ) and EDTA-free protease inhibitor cocktail ( Complete , Roche ) for 30 min in ice and mechanically broken by 20 strokes of a Dounce homogenizer . After homogenizing , cells were incubated in lysis buffer ( 50 mM HEPES ( pH 7 . 5 ) , 500 mM NaCl , 10% glycerol , 1% Triton-X100 , 1 mM TCEP and EDTA-free protease inhibitor cocktail ( Complete , Roche ) for further 30 min in ice and cell extract was cleared by centrifugation for 45 min at 20 . 000 g . The cleared supernatant was incubated in batch with Ni-Sepharose High Performance ( GE Healthcare ) for 2 hours at 4°C with continuous shaking . The resin was first washed with 10 resin-volumes of wash buffer ( 50 mM HEPES ( pH 7 . 5 ) , 10% glycerol , 0 . 4% Triton X-100 , 150 mM NaCl and 20 mM imidazol ) followed by elution with wash buffer containing 250 mM imidazole . Active fraction ( 0 . 5 ml ) were dialyzed against 50 mM HEPES ( pH 7 . 5 ) , 150 mM NaCl , 1 mM DTT , 0 . 4% Triton X-100 and 10% glycerol and stored at −80°C in small aliquots . PI4K kinase activity was assayed with the ADP-Glo Kinase Assay ( Promega ) , according to the manufacturer's instructions . Briefly , 0 . 5 µl of PI4KIIIα-CD or 0 . 05 µl PI4KIIIβ ( 32 ng , Invitrogen ) were preincubated with DMSO or AL-9 in reaction buffer ( 20 mM Tris ( pH 7 . 5 ) , 5 mM MgCl2 , 2 mM DTT , 0 . 5 mM EGTA , 0 . 4% Triton X-100 ) for 10 min at room temperature in a final volume of 8 µl . The reaction was started by addition of 2 µl of ATP and PI∶PS Lipid Kinase Substrate ( Invitrogen ) to give a final concentration of 100 µM and 150 µM , respectively . After 1 hour of incubation at room temperature the reaction was stopped and further processed as described by the manufacturer . In parallel the reaction was performed without PI∶PS substrate in order to detect contaminating ATPase activity present in the protein fractions . This activity was subtracted from the measured kinase activity . Kinase activity of PI3Kα ( p110α/p85α ) and PI3Kβ ( p110β/p85α ) was assayed as above using 5 ng or 20 ng , respectively ( Millipore ) . Reaction buffer was changed to 50 mM HEPES pH 7 . 5 , 10 mM MgCl2 and 1 mM DTT . Cells were plated one day before the experiment in 24-well plates ( 5×104 cells/well for Huh7 . 5 and JFH-A4 cells , 7×104 cells/well for Con1-SR and 1×105 cells/well for cured JFH-A4 cells ) . Cells were either untreated or treated with compounds for the time as indicated in the figure legend . PI4P staining of the plasma membrane or internal membranes was performed exactly as described previously [37] . Primary antibodies used were: anti-PI4P IgM ( Cat . No . Z-P004 , 1∶300 , Echelon ) , anti-Giantin antibody ( Cat . No . PRB-114C-200 1∶1000 , Covance ) , affinity-purified rabbit anti-NS5A antibody ( 1∶2000 ) [54] , anti-PI4KIIIα kinase ( Cat . No . 4902 , 1∶50 , Cell Signaling ) , anti-PI4KIIIβ kinase ( Cat . No , 611817 , 1∶500 , BD Transduction ) . Secondary antibodies used were goat anti-mouse IgM Alexa Fluor 488 ( Cat . No . A-21042 , 1∶600 , Invitrogen ) and goat anti-rabbit Alexa Fluor 568 ( Cat . No . A-11011 , 1∶600 , Invitrogen ) . For type III PI4K kinases or NS5A staining , all incubations were performed at room temperature . Cells were washed once with PBS and fixed with 300 µl of 4% PFA for 15 min . Cells were washed three times with PBS and permeabilized with 500 µl of 0 . 1% Triton X-100 ( or 0 . 5% for PI4KIIIα kinase staining ) in PBS for 10 min . Unspecific binding was blocked by incubation with 3% BSA in PBS ( for PI4KIIIα staining no blocking was performed ) . After incubation with the primary antibody in blocking buffer , cell were washed with PBS and subsequently incubated with goat secondary antibodies conjugated to Alexa-Fluor 568 , or Alexa-Fluor 488 at a dilution of 1∶600 . Nuclei were stained with Hoechst dye 33342 ( Sigma; 1∶4000 ) . Slides were then mounted with 5 µl ProLong Gold Antifade ( Invitrogen ) and analyzed by using an inverted Leica TCS SP5 scanning laser confocal microscope . Digital images were taken using LAS AF software ( Leica ) and processed using Volocity software ( Perkin Elmer ) . Quantification of fluorescence intensity was determined from multiple images using Volocity . Relative changes in fluorescence intensity mean values where obtained from four randomly picked fields for each condition ( 150∼300 cells ) . For plasma membrane staining , total PI4P fluorescence intensity obtained in each condition was normalized to the number of cells present in each field . For the quantification of relative PI4P levels in internal membranes , PI4P fluorescence intensity was normalized using the fluorescence intensity of the Golgi marker giantin . Quantitative immunofluorescence data are presented as means ± the standard error of the mean ( SEM ) . For the calculation of statistical significance , a two-tailed , unpaired t-test was performed . 3×104 Huh7 . 5 cells/well were seeded in 24-well plates on microscope cover glasses and transfected with 50 nM of siRNAs in serum-free Opti-MEM ( Invitrogen ) using Lipofectamine RNAiMAX ( Invitrogen ) , according to the manufacturer's protocol . For western blot analysis , the transfection reaction was proportionally scaled up to 6-well plates . In order to maximize the silencing efficiency , 24 hours after the first transfection , the cells were subjected to a second round of siRNA transfection . siRNA sequences were the following ( 5′→3′ sense strand ) : mock siRNA , 5′-GUAUGACCGACUACGCGUA[dT][dT]-3′ ( custom , Sigma-Aldrich ) ; PI4KIIIα siRNA , 5′-CCGCCAUGUUCUCAGAUAA[dT][dT]-3′ ( custom , Sigma-Aldrich ) ; and PI4KIIIβ siRNA , 5′-GCACUGUGCCCAACUAUGA[dT][dT]-3′ ( Silencer Validated siRNA s10543; Ambion ) . Three days after the initial transfection , cells were stained for PI4P as described previously [37] , or subjected to western blot analysis . For immunoblot analysis of protein expression , cells were harvested with TEN buffer ( 10 mM Tris/HCl pH 8 . 0 , 1 mM EDTA , 100 mM NaCl ) , washed once with PBS and lysed with 2X protein sample buffer ( 125 mM Tris-HCl pH 6 . 8 , 10 mM EDTA , 0 , 003 gr bromophenol blue , 20% glycerol , 4% SDS and 10% β-mercaptoethanol; 200 µl ) . The samples were then sonicated , heated at 95°C and loaded onto 7 . 5% polyacrylamide-SDS page ( Criterion , Biorad ) . After electrophoresis proteins were transferred to a nitrocellulose membrane and unspecific binding was blocked by PBS supplemented with 0 . 5% Tween ( PBS-T ) and 5% milk . Membranes were then incubated overnight at 4°C with primary antibodies ( anti-PI4KIIIα , cat no . 4902 , 1∶250 Cell Signaling , anti-PI4KIIIβ , cat . No . 611817 , 1∶3000 BD Transduction Laboratories , mouse anti-β-actin , cat . No . A1978 , 1∶5000 , Sigma ) . HRP-conjugated secondary antibodies ( donkey anti-rabbit , Cat . No . 9341 and sheep anti-mouse , Cat . No . 9311 , GE Healthcare ) were incubated for 1 hour at room temperature and detection was performed using SuperSignal-Femto chemiluminescent substrate ( Pierce-Thermo Scientific ) . 1 . 5×106 Huh7-Lunet/T7 cells/100 mm dish were transfected with 20 µg pTM-NS3-5B using the transfection reagent Lipofectamine 2000 ( Invitrogen ) . Six hours after transfection , cells were seeded in 24-well plates on microscope cover glasses for indirect immunofluorescence . After 5 hours , cells were treated either with DMSO or with 8 µM AL-9 for 2 , 8 or 16 hours and co-staining of NS5A and PI4P was performed using the Golgi staining protocol , as described previously [37] . | It is estimated that 3% of the world's population are chronically infected by the hepatitis C virus ( HCV ) . Most infections become chronic and eventually evolve into cirrhosis and hepatocellular carcinoma . Host factors are interesting targets for anti-HCV therapies due to their inherent high genetic barrier to resistance . Recently , phosphatidylinositol 4-kinase α ( PI4KIIIα ) has been identified as a crucial host factor for HCV replication . Many different pathogens , including HCV , subvert components of the phosphatidylinositol-4 phosphate ( PI4P ) pathway to function in favor of their own life cycle . In this paper , we show that HCV dramatically alters cellular PI4P metabolism and distribution , resulting in the enrichment of PI4P in the membranous web required for viral replication with a concomitant decrease of PI4P in the plasma-membrane . Moreover , we demonstrate that 4-anilino quinazolines , antiviral agents previously believed to target HCV NS5A , do in fact inhibit PI4P formation by inhibition of PI4KIIIα . This compound class is a promising lead for the development of a novel antiviral therapy based on PI4KIIIα inhibition . Specific PI4KIIIα inhibitors would also be important research tools required for a deeper understanding of the functions and regulation of PI4P . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology",
"biology",
"microbiology"
] | 2012 | Metabolism of Phosphatidylinositol 4-Kinase IIIα-Dependent PI4P Is Subverted by HCV and Is Targeted by a 4-Anilino Quinazoline with Antiviral Activity |
In acute Plasmodium falciparum ( P . falciparum ) malaria , the pro- and anti-inflammatory immune pathways must be delicately balanced so that the parasitemia is controlled without inducing immunopathology . An important mechanism to fine-tune T cell responses in the periphery is the induction of coinhibitory receptors such as CTLA4 and PD1 . However , their role in acute infections such as P . falciparum malaria remains poorly understood . To test whether coinhibitory receptors modulate CD4+ T cell functions in malaria , blood samples were obtained from patients with acute P . falciparum malaria treated in Germany . Flow cytometric analysis showed a more frequent expression of CTLA4 and PD1 on CD4+ T cells of malaria patients than of healthy control subjects . In vitro stimulation with P . falciparum-infected red blood cells revealed a distinct population of PD1+CTLA4+CD4+ T cells that simultaneously produced IFNγ and IL10 . This antigen-specific cytokine production was enhanced by blocking PD1/PDL1 and CTLA4 . PD1+CTLA4+CD4+ T cells were further isolated based on surface expression of PD1 and their inhibitory function investigated in-vitro . Isolated PD1+CTLA4+CD4+ T cells suppressed the proliferation of the total CD4+ population in response to anti-CD3/28 and plasmodial antigens in a cell-extrinsic manner . The response to other specific antigens was not suppressed . Thus , acute P . falciparum malaria induces P . falciparum-specific PD1+CTLA4+CD4+ Teffector cells that coproduce IFNγ and IL10 , and inhibit other CD4+ T cells . Transient induction of regulatory Teffector cells may be an important mechanism that controls T cell responses and might prevent severe inflammation in patients with malaria and potentially other acute infections .
Malaria remains one of the leading health burdens worldwide with about 600 000 deaths per year . Most of these deaths are attributable to the species Plasmodium falciparum ( P . falciparum ) [1] . Primary infection with P . falciparum initially induces a strong Th1-type CD4+ T cell response . While a strong proinflammatory Th1 response can contribute to control of parasitemia and protection to subsequent infections [2 , 3] , it can also be pathological as it activates the endothelium and thereby promotes sequestration of parasitized red blood cells in the microvessels of vital organs such as the brain [4 , 5] . This impedes parasite clearance by the spleen and enhances severe manifestations of malaria such as cerebral malaria [6–8] . Therefore , a tight coordination of the immune response is needed to ensure the optimal outcome for the patient . Strong proinflammatory responses activate counteracting pathways such as the induction of regulatory T cell ( Treg ) populations and the production of anti-inflammatory cytokines , which both are crucial for preventing immunopathology in malaria and other parasitic diseases [9–12] . Another key mechanism that regulates potentially immunopathological T cell responses in the periphery is the induction of coinhibitory receptors such as cytotoxic T-Lymphocyte attenuator 4 ( CTLA4 ) and programmed death 1 ( PD1 ) on T cells . The importance of PD1 and CTLA4 in T cell regulation has largely been studied in various chronic viral diseases , including HIV , hepatitis B , and hepatitis C . Such chronic viral diseases induce sustained PD1 and CTLA4 expression on activated T cells . This is associated with T cell exhaustion and reduced effector functions of the cells . Blockade of these receptors can partially rescue the T cell responses to these viruses , thereby reducing the viral burden [13–17] . Several recent studies , however , suggest that coinhibitory receptors inhibit T effector functions not only in chronic viral but also in acute infections [18–22] . In support of this , we and other groups have shown that in experimental murine malaria , conventional T cells strongly express the coinhibitory receptors CTLA4 and PD1 [18 , 23] . Blocking coinhibitory receptors improves parasite control [20 , 24 , 25] but also leads to more severe manifestations of disease in several models of experimental malaria [18 , 20 , 23] . Recent studies reported that CTLA4 and PD1 are also upregulated on the T cells of patients with acute malaria or children that are regularly exposed to P . falciparum in endemic areas [24 , 26–28] . However , it remains unclear how the expression of these coinhibitory receptors influences the immune response to acute P . falciparum infection in humans . We evaluated the CD4+ T cell response in patients with acute imported malaria in Hamburg , Germany to investigate whether the induction of coinhibitory receptors downregulates the T cell response in acute P . falciparum malaria and to further elucidate involved regulatory pathways . High numbers of CD4+ T cells in the peripheral blood of malaria patients expressed CTLA4 and PD1 . These PD1+CTLA4+CD4+ T cells showed two distinct functions . Firstly , they proliferated and produced IFNγ and IL10 in response to P . falciparum . Blockade of CTLA4 and PD1 markedly increased P . falciparum-specific cytokine production . Secondly , these T cells also suppressed the proliferation of other CD4+ T cells in response to polyclonal and P . falciparum-specific stimulation . Thus , PD1+CTLA4+CD4+ T cells may be a new population of regulatory Teffector ( Teff ) cells that arises from the CD4+ Teff population during acute infection to downregulate proinflammatory and potentially immunopathological responses .
The present study explored the role of coinhibitory receptors as an immune regulatory pathway in acute P . falciparum malaria . We initially hypothesized that the induction of coinhibitory receptors , primarily CTLA4 and PD1 , is an important and adaptable mechanism by which CD4+ T cell responses in malaria are downregulated . Indeed , we found that the majority of P . falciparum-specific CD4+ T cells expressed both CTLA4 and PD1 and that blocking both receptors enhanced their P . falciparum-specific effector responses . Surprisingly , subsequent experiments showed that the PD1+CTLA4+CD4+ Teff cells in malaria patients had an acquired extrinsic regulatory function by which they suppressed the P . falciparum-specific and polyclonal proliferation of other T cells . These PD1+CTLA4+CD4+ Teff cells therefore constitute a distinct regulatory T cell population which is induced during acute malaria . Our collection of blood from adult patients with imported P . falciparum malaria in Hamburg , Germany , enabled us to conduct complex functional experiments in the setting of acute infection and disease that would not be possible in holoendemic areas , where adults rarely develop acute febrile malaria and only small amounts of blood can be collected from sick children . Our patient cohort consisted largely of two groups , namely , Caucasian patients who had malaria for the first time , and patients of West African descent who reside in Germany and whose last treatment for malaria was at least 5 years before the index infection ( S1 Table ) . Notably , we did not observe a difference between the two main groups of our patient cohort in terms of their T cell phenotypes and regulatory function ( S12 Fig ) . Recently , several studies have observed that patients with acute malaria , and children in endemic regions , who are exposed to P . falciparum , express higher levels of CTLA4 and PD1 on their T cells than uninfected or unexposed control subjects [24 , 26–28] . This increased expression has been proposed to be a sign of T cell exhaustion . Indeed , when coinhibitory receptors are blocked , T cells from patients with Plasmodium vivax malaria produce higher cytokine levels in response to parasite antigens [28] . However , whether these receptors induce a similar downregulation of the T cell response to P . falciparum malaria had not been addressed so far . Our study showed that during acute P . falciparum malaria , high numbers of CD4+ T cells expressed CTLA4 and PD1 and mostly coexpressed both receptors . Although both , CTLA4 and PD1 , are rapidly upregulated after T cell receptor-mediated activation , their expression can also be upregulated by alternative cytokine-based mechanisms [34 , 35] . While clearly the vast majority of P . falciparum-specific T cells in our study were CTLA4+ and PD1+ , it was not possible to determine if all PD1+CTLA4+CD4+ T cells were specific for P . falciparum-antigens or whether bystander activation contributed to the high numbers of CTLA4+ and PD1+ T cells . Importantly , patients with severe cerebral malaria showed a higher frequency of CTLA4+CD4+ T cells than patients with uncomplicated malaria . This analysis was restricted by the low number of patients with cerebral malaria ( n = 3 ) included in our study and needs to be further validated with higher numbers of patients with severe malaria . But our data support results from an earlier study with children in Ghana , in which a higher percentage of CTLA4+CD4+ T cells was observed in children with severe malaria , predominantly severe anemia , than in children with uncomplicated malaria [36] . These observations suggest that a dysregulation of the CD4+ T cell function is associated with an increased risk for severe complications of malaria . By contrast , we found that the degree of parasitemia at time of diagnosis did not correlate with the frequency with which CD4+ T cells expressed CTLA4 or PD1 . This is different from observations in chronic and acute viral diseases where it has been shown consistently that viral load correlates with PD1 and CTLA4 expression on T cells [14 , 37–39] . It should be noted that peripheral parasitemia does not reflect the numbers of sequestered parasites in the microcirculation and is therefore a poor predictor of complete parasite biomass [40] . Besides , peak parasitemia and highest frequencies of CTLA4+ and PD1+CD4+ T cells might occur at later time points than time of diagnosis ( Fig 1D ) . Thus , a correlation between CTLA4 and PD1 expression and complete parasite biomass or peak parasitemia cannot fully be excluded . Interestingly , a study by Schlotmann et al . detected a correlation between CTLA4 expression and peak parasitemia ( but not parasitemia at time of diagnosis ) in adult patients with acute malaria [26] . However , it is also conceivable that the lack of correlation between parasitemia and CTLA4 and PD1 expression in our study reflects the fact that the parasite-induced upregulation of coinhibitor expression is readily saturated by antigen-load and cannot be further increased by higher parasitemia . Other factors such as the duration of antigen exposure or cytokine levels might additionally influence the CTLA4 and PD1 expression in malaria . To date , several studies have reported that T cell responses to plasmodial and other antigens are reduced during acute malaria and recover after treatment and convalescence [41–45] . Several mechanisms underlying this phenomenon have been proposed , including impairment of antigen-presenting cell function , the production of anti-inflammatory cytokines , and the induction of apoptosis [45 , 46] . Similarly , we found in our study that only 40% of our patients showed measurable T cell proliferation and 65% produced cytokines ( predominantly IFNγ and IL10 ) by intracellular staining in response to P . falciparum , respectively . We propose that the strong induction of CTLA4 and PD1 contributes to the reduced effector function of P . falciparum-specific T cells during acute malaria . Blocking both , CTLA4 and PDL1 , enhanced the T cell proliferation in 4 of 11 malaria patients . The blockade showed a stronger effect on cytokine production and increased the malaria-specific IFNγ and IL10 production of all patients who exhibited a cytokine response to P . falciparum and demarcated a positive IFNγ response in one patient who was priorly unresponsive . Our results corroborate observations by us and other groups in rodent malaria models that administration of antibodies against CTLA4 and/or PDL1 in vivo increases the proinflammatory T cell responses in murine malaria [18 , 23 , 24 , 47] . Moreover , a recent study showed that blocking coinhibitory receptors , including PDL1 and CTLA4 , enhanced cytokine responses to P . vivax antigens in vitro [28] . Of note , we observed a small but significant increase of the frequencies of PDL2+ macrophages , T cells and B cells . Further enhancement of the T cell effector function might be achieved with additional blockade of the PD1/PDL2 axis [48] , which has not been examined in malaria so far . Taken together , our data support the notion that the observed dysfunction of P . falciparum-specific T cells during acute malaria can be attributed , at least in part , to the upregulation of CTLA4 and PD1 . PD1+CTLA4+CD4+ T cells were the main source of P . falciparum-specific cytokines and predominantly produced IFNγ and IL10 . Most coproduced both of these cytokines . Both , IFNγ and IL10 , play critical roles in the immune homeostasis in acute malaria . IFNγ is needed to activate phagocytosis and control parasitemia , while IL10 is crucial for counterbalancing the inflammatory response [49–52] . While IL10 can be produced by various cell populations , including B cells , monocytes , NK cells , Th2-type T cells , and Type 1 Tregs ( Tr1 cells ) , several studies suggest that IFNγ+IL10+ Th1 cells are a particularly important source of IL10 in human and murine parasitic infections [9 , 50 , 53] . Importantly , IFNγ+IL10+ Th1 cells have been shown recently to prevent immunopathology in rodent models of toxoplasmosis and malaria [10 , 49] . Similarly , studies in endemic areas identified IFNγ+IL10+ Th1 cells in children regularly exposed to or infected with P . falciparum [51 , 53 , 54] . Moreover , low frequencies of these cells were associated with the development of severe malaria in children in the Gambia [11] . Importantly , we also noticed a trend towards a decreased IL10+ to IFNγ+ T cell ratio in our patients with severe cerebral malaria compared to patients with uncomplicated malaria but our sample size of patients with severe cerebral malaria was low ( n = 3 ) . Specifically , there was a lower proportion of IL10 single-positive Th1 CD4+ T cells . This is different to the prior study in Gambia , where a polycloncal stimulus with PMA/Ionomycin was used . High ratios of TNFα and INFγ levels to IL10 or TGFβ levels have also been observed in other studies in patients with severe malaria [55 , 56] . Taken together , our data showed that IFNγ+IL10+ CD4+ T cells are present during the first-ever symptomatic infection with P . falciparum in naïve individuals as well as in individuals who were last exposed > 5 years ago . Repeated exposure is therefore not required for the induction of IFNγ+IL10+ CD4+ T cells . Notably , we also observed that both the IFNγ+IL10+ cells and even a substantial percentage of the IL10 single-positive cells expressed Tbet , suggesting that they derive from Th1 lineage . This is consistent with in vitro studies employing the complement factor CD46 that showed Th1 effector cells acquire the ability to produce IL10 and convert from IFNγ+ into IFNγ+IL10+ cells during ongoing stimulation before terminally transitioning to IL10 single-positive Th1 CD4+ T cells . The latter cells exert a suppressive function and have a Tr1-like phenotype [57] . It is highly likely that the strong antigen exposure and the cytokine milieu during acute malaria activates the IFNγ to IL10 switch , which serves as a negative feedback loop to prevent overwhelming inflammation and tissue damage . This concept is further supported by the observation by us and others that IFNγ to IL10 ratios might be altered in patients with severe malaria . Interestingly , isolated PD1+CTLA4+CD4+ T cells acted like antigen-specific Tr1-like cells: they suppressed the P . falciparum-specific and anti-CD3/28-induced proliferation of the CD4+ T cell population from malaria patients . However , they did not suppress the responses to other antigens such as tetanus toxoid . This suggests strongly that PD1+CTLA4+ T cells are P . falciparum-specific and require prior activation to exert their cell-extrinsic suppressive function . This activation can be mediated by a polyclonal stimulans such as anti-CD3/28 or by their specific P . falciparum antigen . The PD1+CTLA4+ T cells in our malaria patients were clearly distinct from natural Tregs as they did not express Foxp3 or CD25 , nor were they CD49b+LAG3+ ( S13 Fig ) , which are recently identified markers for Tr1 cells [58] . However , the expression of both receptors , CTLA4 and PD1 , has repeatedly been associated with a regulatory phenotype . In particular , the CTLA4 expression of our PD1+CTLA4+CD4+ T cells was significantly higher than in Foxp3+ natural Tregs . This is interesting because although CTLA4 was originally discovered as a cell-intrinsic inhibitor , there is some evidence that conventional CTLA+ T cells can play a cell-extrinsic regulatory function but current observations are restricted to murine experiments [59–61] . PD1 expression on the other hand associates with IL10-producing Tr1 or Tr1-like Tregs [62 , 63] . Indeed , we found that the PD1+CTLA4+CD4+ Teff cells in our malaria patients produced the regulatory cytokine IL10 in response to P . falciparum-antigens . But unlike the inhibitory function of classical Tr1 cells , the suppressive effect of the PD1+CTLA4+CD4+ Teff cells was not abrogated by blocking IL10 . In fact , even the combined blockade of CTLA4 , PDL1 , TGFβ , and IL10 did not alter the suppressive function of the PD1+CTLA4+CD4+ Teff cells . Expression of coinhibitory receptors and the cytokines TGFβ and IL10 are important tools to maintain immune homeostasis [18 , 23 , 24 , 47 , 49 , 64] . But the observed suppressive effect of PD1+CTLA4+CD4+ Teff cells is independent of these regulatory mechanisms . Transwell experiments revealed that the suppression was dependent on T-T cell contact , similar to what is seen in nTregs . It is highly possible that the suppressive effect of the PD1+CTLA4+CD4+ Teff cells is mediated by several mechanisms and not one single mechanism can be identified , as has been shown in nTregs . Possible mechanisms include release of granzyme B and perforin , disrupting the metabolic state of T cells via the production of the ectoenzymes CD39 and CD73 or consumption or downregulation of IL2 [65] . The fact that the PD1+ CTLA4+ CD4+ Teff cells expressed only low levels of CD25 argues against consumption of IL2 as a possible mechanism underlying their suppressive activity . But further studies that determine how PD1+ CTLA4+CD4+ T cells suppress other T cells are clearly warranted . Mouse models of malaria are probably a powerful tool to further investigate the suppressive mechanisms . To our knowledge , this is the first time a population of antigen-specific Teff cells with a cell-extrinsic suppressor function has been reported in an acute infection in humans . Our observations partially support previous findings by Häringer et al . , who identified a population of IFNγ+IL10+ effector-like T cells with regulatory function in the blood of healthy volunteers [62] . Similar to our study , these cells express high levels of PD1 and CTLA4 and low levels of CD127 and respond to persisting antigens such as Candida and CMV but not to vaccine antigens . However , the suppressive activity of these IFNγ+IL10+ T cells mainly depends on IL10 . Several other studies have described populations of allergen-specific PD1+CTLA4+IL10+ T cells in the peripheral blood but their suppressive function in vitro depends on multiple factors , including CTLA4 , PD1 , and IL10 [33 , 63] . Interestingly , a study by Che et al . showed that naïve T cells failed to expand in vitro in the presence of HIV-primed T cells , and that this suppressive effect was also dependent on T-T cell contact [66] . While peripheral control of inflammation has generally been attributed to adaptive Tr1 cells , which are thought to be a distinct lineage , our observations raise the intriguing hypothesis that during an acute infection , regulatory function is transiently acquired by Teff cells to control the T cell response by three mechanisms: a ) autoregulation of the Teff response through upregulation of CTLA4 and PD1 which inhibit cytokine production b ) cytokine switch from IFNγ to IL10 production and c ) cell-extrinsic inhibition of the T cell proliferation . This possibility is potentially exciting given the increasing focus on immune-modulating therapies in the last few years that has led to the licensing of antibodies against coinhibitory receptors and the first trials on the transfer of T cell populations [67 , 68] . Modulating Teff cells and their regulatory pathways could offer new and exciting approaches for preventing immune-mediated pathologies in severe infections or for modulating vaccine responses in the future . Further studies that elucidate the factors involved in the induction of PD1+ CTLA4+ T cells , their effector and regulatory functions , and their immunological memory are strongly warranted . Studies that explore the relationship between the regulatory pathways and the outcome of P . falciparum infection in humans in endemic areas would also be very valuable . In conclusion , our results shed further light on the complex immune regulation in acute malaria . We conclude that PD1+CTLA4+ Teff cells play a crucial role in the immune regulation: they arise from Th1 effector cells and acquire a suppressive phenotype , by expressing coinhibitory receptors , switching to IL10 production and by cell-extrinsic suppression of other T cells . Such a negative feedback loop is probably not unique for P . falciparum malaria and could be a common autoregulatory mechanism that serves to control overwhelming Th1 responses and prevent immunopathology in acute infections .
In total , 25 patients with acute malaria were enrolled during their inpatient treatment at the University Hospital Eppendorf in Hamburg between the first and sixth day of the treatment . Patients were enrolled consecutively between March 2013 and September 2014 . The clinical characteristics of the enrolled patients are summarized in S1 Table . The majority of these patients ( 14 of 25 ) were of West African origin and had been living in Germany for ≥5 years ( 5–20 years ) . 8 of the remaining patients were Germans who were infected during vacation or short-term work assignments in a malaria endemic country . Two were German expatriates , living in Western Africa and one patient from Gambia was visiting Germany on vacation . Patients of West African origin had suffered from repeated malaria episodes as children but had not been diagnosed with malaria > 5 years . The 25 enrolled patients were on average 42 ( range 18–70 ) years old . All had microscopically detectable parasitemia and five had severe malaria , according to the definitions of the WHO [69] . Of the five patients with severe malaria , three suffered from cerebral malaria and two were diagnosed with hyperparasit emia . Patients with cerebral malaria were treated with Artesunate iv , followed by atovaquone-proguanil . All other patients were treated with artesiminin-combination therapy ( dihydroartemisin-piperaquin ) or atovaquone-proguanil p . o . . Between 5 and 50 ml of heparinized blood was collected from each malaria patient and used for flow cytometric analysis and functional assays . Due to limitations on number of cells , only a subset of functional assays and/or flow cytometric staining could be carried out for each blood sample . In addition to the enrolled patients , anonymous blood samples ( 1 ml each ) from 15 adult patients with an active P . falciparum infection were obtained from the clinical diagnostic lab of the Bernhard Nocht Institute of Tropical Medicine . Baseline information such as age , parasitemia and presentation of symptoms were available for these patients . The latter samples were only subjected to ex vivo flow cytometric T cell analysis ( Fig 1B , 1C and 1E ) . For a small number of patients , additional follow-up samples were received during treatment via the diagnostic lab and used for ex vivo flow cytometric T cell phenotyping . The 19 control subjects were healthy adult staff members of the Bernhard Nocht Institute of Tropical Medicine and the University Medical Centre Hamburg Eppendorf . P . falciparum infection was determined microscopically by experienced lab technicians at the Bernhard Nocht Institute of Tropical Medicine . Thick and thin blood smears were stained with 4% Giemsa and examined under oil immersion ( original magnification × 100 ) . Ex vivo staining for immunophenotypic analysis was conducted on fresh whole blood samples from malaria patients and healthy volunteers . For malaria patients , samples were included on the day of diagnosis ( day 1 ) prior to treatment whenever available . Alternatively , the first blood sample obtained from that specific patient was included for flow cytometric analysis . Thus , 100 μl of fresh whole blood was incubated for 30 min at 4°C with the following antibodies: CD4 BV510 ( clone OKT4 ) , CD8 AF700 ( RPA-T8 ) , CD14 AF 700 ( HCD14 ) , CD19 BV510 ( HIB19 ) , CD11c PECy7 ( Bu15 ) , CD45RA FITC ( HI100 ) , CCR7 PECy7 ( G043H7 ) , PD1 PerCP Cy5 . 5 ( EH12 . 2H7 ) , BTLA Biotin ( MIH26 ) with Streptavidin-BV 421 , CD25 PE Cy7 ( BC96 ) , CD127 BV 421 ( A019D5 ) , PD-L1 APC ( 29E . 2A3 ) , PD-L2PE ( 24F . 10C12 ) , CD80 BV 421 ( 2D10 ) , and CD86 AF488 ( IT2 . 2 ) ( all from BioLegend ) . Thereafter , the blood samples were lysed and fixed using a Lysis/Fixation buffer ( BioLegend ) , and washed . Subsequent intracellular staining for CTLA4 PE ( L3D10 ) , Foxp3 AF647 ( 259D ) , Tbet APC ( 4B19 ) , Granzyme B AF647 ( GB11 ) , CD3 APC Cy7 ( SK7 ) ( all from BioLegend ) , and Ki67 AF 488 ( MK167 ) ( eBioscience ) was conducted using the eBioscience Foxp3/Transcription factor staining buffer set according to the manufacturer’s protocol . The flow cytometric data were collected on a LSR II Cytometer ( 4 laser , Becton Dickinson ) and analyzed using FlowJo software ( Treestar ) . Compensations were set using single stained/unstained beads . At least 100 000 cells were collected in the lymphocyte gate . Cells were first gated on forward/sideward scatter to exclude cell debris and gates were set for lymphocytes and monocytes . T cells were then gated on the basis of the lineage markers CD3 , CD4 and CD8 . Monocytes/macrophages were defined as CD14+CD3- cells . B cells were defined as CD19+CD3- cells . Fluorescence minus one ( FMO ) controls were used to define subsequent gates for coinhibitory receptors and ligands and regulatory , memory , and activation markers . FMO is a staining control that includes all antibodies used in a flow cytometric assay except for one fluorochrome of interest ( termed FMO ) to control for the contribution of spectral overlap to the background when using multiple fluorochromes [70 , 71] . In vitro P . falciparum antigen-specific CD4+ T cell proliferation and cytokine expression were measured using P . falciparum-infected red blood cells ( iRBC ) as the antigen . For this , P . falciparum cultures ( strain 3D7 ) were maintained in 0+ erythrocytes in RPMI-1640 medium supplemented with 0 . 5% AlbuMAX II ( Invitrogen ) at 37°C according to standard methods [72] . Schizont and late trophozoite stage parasites were isolated magnetically using the Miltenyi LD column . The percentage of infected erythrocytes after magnetic isolation was typically 80–90% . The iRBC were stored in PBS at -70°C until use in cell culture at two iRBC per T cell/PBMC . Uninfected red blood cells ( uRBC ) , namely , the 0+ erythrocytes from the same donors that were used for the parasite cultures , served as negative controls at two uRBC per T cell/PBMC . CMV pp65 protein ( Miltenyi , 1 μl/ml ) and tetanus toxoid ( 10 μg/ml ) served as alternative antigens . For polyclonal stimulation in proliferation assays , 1 μg/ml soluble anti-CD3 ( clone UCHT1 , eBioscience ) and 1 μg/ml anti-CD28 ( clone CD28 . 2 . , eBioscience ) were used . Alternatively , anti-CD3/28-coated beads ( Dynal T cell expander ) were used at a ratio of five cells/bead . PHA ( Sigma-Aldrich at 5 μg/ml ) or PMA/Ionomycin ( 50ng/500ng ) served as the positive controls for intracellular cytokine staining . PBMC were isolated from heparinized whole blood within 2 hours of collection by standard density gradient centrifugation with Ficoll Paque ( GE Healthcare ) using the SepMate System ( Stemcell technologies ) according to the manufacturer’s instructions . The isolated PBMC were then directly subjected to antigen/mitogen stimulation assays that measured cell proliferation and cytokine production by intracellular staining or underwent further T cell isolation steps . For cell cultures , fresh PBMC were resuspended in complete RPMI ( cRPMI ) , consisting of RPMI 1640 ( PAA laboratories ) supplemented with 10% pooled human AB serum ( Sigma-Aldrich ) , 4 mM L-glutamine , 25 mM HEPES , and 80 mg/ml gentamicin ( both PAA laboratories ) . For isolation of T cell subsets , freshly isolated PBMC were washed and resuspended in MACS buffer ( PBS with 0 . 5% BSA and 2 mmol/l EDTA ) . The PBMC were then negatively selected for CD4+ T cells using the CD4+ T cell isolation kit from Miltenyi according to the manufacturer’s protocol . The isolated CD4+ T cell preparations were routinely >90% pure . An aliquot of each CD4+ T cell preparation was set aside for cell culture . The remaining CD4+ T cells were resuspended in PBS supplemented with 1% human AB serum and surface-stained with anti-CD4 APC ( RPA-T4 ) , anti-CD25 PECy7 ( BC96 ) , anti-CD127 AF 488 ( A019D5 ) , anti-PD1 PerCP Cy5 . 5 ( EH12 . 2H7 ) , and anti-CTLA4 PE ( L3D10 ) . Gates were set based on FMO controls . Natural Tregs were first excluded on the basis of high expression of CD25 and low expression of CD127 . The remaining CD4+ T cells were then sorted into PD1+ CD4+ T cells and PD1neg CD4+ T cells . The PD1+ T cell fractions were 50–95% pure for PD1 and CTLA4 after conducting intracellular staining for CTLA4 ( S10 Fig ) . The CD4- cells that were acquired in the magnetic-bead sorting step were irradiated with 3000 rad and served as antigen-presenting cells . After isolation , the cells were resuspended in cRPMI and used for cell culture . Freshly isolated PBMC or T cells were cultured in round-bottom 96-well plates directly after isolation . To assess antigen-specific responsiveness of PBMC , the cells were cultured at 1×106/ml for 120 hours with 2 iRBC or uRBC/cell , after which 80 μl of supernatant was removed for cytokine detection assays . To assess whether blocking the ligation of PDL1 and CTLA4 would enhance the proliferation and cytokine production of these PBMC , the cells were coincubated with 10 μg/ml anti-PDL1 ( MIH1 , eBioscience ) and 10 μg/ml anti-CTLA4 ( Ipilimumab ) . Isotype controls for human and mouse IgG1 ( BioLegend ET901 and eBioscience P3 . 6 . 2 . 8 . 1 ) were used at 10μg/ml . To assess the ability of PD1+CTLA4+CD4+ T cells to suppress other CD4+ T cells , 1 . 25×105 CD4+ T cells/ml were stimulated with anti-CD3/28 beads ( 5 cells/bead ratio ) . PD1+CTLA4+CD4+ T cells were added to the cultures at different CD4+/ PD1+ ratios , namely , 1:0; 1:1; and 1:0 . 5 . For antigen-specific suppression assays , CD4+ T cells were cultured at 5×105 T cells/ml with iRBC , uRBC , CMV , or tetanus toxoid with or without PD1+CTLA4+CD4+ T cells in the presence of a final concentration of 5×105/ml irradiated CD4- cells , which served as antigen-presenting cells . After 120 hours of culture , 80 μl of supernatant was removed for cytokine detection assays . Lymphocyte proliferation in the PBMC and suppression assays was determined by pulsing the cultures after 120 hours with 1 μCi/ml 3H-Thymidine for 18 hours , harvesting the cells onto cellulose filters , and measuring radiolabeled thymidine incorporation by scintillation counting . The results were expressed as counts per minute ( cpm ) . A stimulation index >2 ( mean cpm of stimulated cells/mean cpm of unstimulated cells ) was considered to be a positive proliferative response . To determine the suppressive mechanism , the suppression cultures were co-cultured with anti-PDL1 ( MIH1 , eBioscience ) , anti-CTLA4 ( Ipilimumab ) , anti-TGFβ ( 1D11 . 16 . 8 , eBioscience ) , and anti-IL10R ( 3F9 , BioLegend ) ( all at 10 μg/ml ) . Transwell experiments were conducted using Nunc transwell inserts . Cell culture conditions were set up in triplicates when sufficient cell numbers were available . However , limitations in cell numbers meant that not all samples could be subjected to all experimental conditions . Freshly isolated PBMC were labeled with 15 mM CFSE ( Molecular Probes ) and cultured with iRBC , uRBC or medium ( negative controls ) , or anti-CD3/28 ( positive control ) , harvested after 72 hours , surface-stained for CD4 , CD8 , CD3 and PD1 , and intracellularly stained for CTLA4 using the eBioscience Foxp3/Transcription factor staining buffer . Dead cells were excluded using the UV live/dead stain ( Life Technologies ) . To detect intracellular cytokines , PBMC were cultured at 106 cells/ml and stimulated with iRBC , uRBC , PHA , anti-CD3/28 , CMV or without antigen for 18 hours . Brefeldin A and Monensin ( 1 μl/ml each , both from BioLegend ) were added after 6 hours of culture . For control cultures with PMA/Ionomycin , PBMC were stimulated with 50ng PMA and 500ng Ionomycin for a total duration of 12h . The cells were then washed and surface-stained with the following antibodies: CD4 BV 510 , CD8 AF 700 , PD1 PerCP Cy5 . 5 , and CD127 BV 421 . To detect dead cells , the cells were also stained with UV live/dead ( Life Technologies ) . The cells were then fixed and permeabilized using the eBioscience Foxp3 /Transcription factor staining buffer set and intracellularly stained using the following antibodies: IFNγ FITC ( 4S . B3 ) , IL10PE ( JES3-9D7 ) , IL4BV 421 ( MP4-25D2 ) , TNFα PECy7 ( MAB11 ) , IL13PE ( JES10-5A2 ) , CTLA4 PE ( L3D10 ) , Foxp3 AF 647 ( 259D ) , CD3 APC Cy7 ( SK7 ) and Tbet APC ( 4B19 ) . Samples were acquired on a LSR II and analyzed using FlowJo software . Gates were set on the basis of FMO controls . Cultural supernatants were harvested after 120 hours and frozen at -20°C . The cytokines IFNγ and IL10 were measured using the eBioscience ELISA ready-SET-go kit according to the manufacturer’s recommendations . The detection limits were 10 pg/ml for IL10 and 20 pg/ml for IFNγ . For statistical analysis , samples with optical density readings below the limit of the standard curve of the assay were assigned a value half that of the detection level . Two groups were compared using the student’s t-test . For comparison of several parameters between two groups , t-tests with Holm-Sidak correction for multiple comparisons were used ( Graph Pad Prism 7 ) . Three or more groups were compared by ANOVA with the Bonferroni post hoc test . For data that was not normally distributed , the Wilcoxon matched pairs test or Friedman test with Dunn´s multiple comparison test was used . Correlation analysis was conducted by calculating Pearson’s product-moment correlation coefficient . IFNγ/IL10 cytokine combinations displayed in pie charts were compared between the two groups by performing a partial permutation test in SPICE [73] . Ethical approval was obtained from the Ethikkommission Hamburg . Written informed consent was obtained from all enrolled participants prior to inclusion in the study . | In acute infections like malaria , our immune systems must achieve a careful balance between inflammatory and anti-inflammatory responses to successfully fight the infection without causing harm to the host . In this study , we examined the CD4+ T cell response and CD4+ T cell regulation in patients with acute malaria . Important mechanisms to control CD4+ T cell activity include specific regulatory T cell populations which suppress other T cells or the expression of so-called coinhibitory receptors which inhibit the inflammatory response of the expressing cells . We showed that in malaria patients , high numbers of T cells expressed the coinhibitory receptors CTLA4 and PD1 . Surprisingly , despite the high expression of coinhibitory receptors , malaria-specific effector function was predominantly found in the PD1+CTLA4+CD4+ T cell population . A blockade of the receptors enhanced the effector response . Even more surprising , we found that although PD1+ CTLA4+CD4+ T cells contained the majority of malaria-specific T cells , they showed , at the same time , cell-extrinsic suppressor activity and actively downregulated T cell proliferation . Thus , our observations describe a new population of “regulatory” T effector cells which are induced during acute malaria . This will contribute to our understanding of the complex immune pathways activated during acute malaria . | [
"Abstract",
"Introduction",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"blood",
"cells",
"flow",
"cytometry",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"immune",
"cells",
"immune",
"physiology",
"cytokines",
"plasmodium",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"pa... | 2016 | Acute Malaria Induces PD1+CTLA4+ Effector T Cells with Cell-Extrinsic Suppressor Function |
Previously , we identified an adolescent idiopathic scoliosis susceptibility locus near human ladybird homeobox 1 ( LBX1 ) and FLJ41350 by a genome-wide association study . Here , we characterized the associated non-coding variant and investigated the function of these genes . A chromosome conformation capture assay revealed that the genome region with the most significantly associated single nucleotide polymorphism ( rs11190870 ) physically interacted with the promoter region of LBX1-FLJ41350 . The promoter in the direction of LBX1 , combined with a 590-bp region including rs11190870 , had higher transcriptional activity with the risk allele than that with the non-risk allele in HEK 293T cells . The ubiquitous overexpression of human LBX1 or either of the zebrafish lbx genes ( lbx1a , lbx1b , and lbx2 ) , but not FLJ41350 , in zebrafish embryos caused body curvature followed by death prior to vertebral column formation . Such body axis deformation was not observed in transcription activator-like effector nucleases mediated knockout zebrafish of lbx1b or lbx2 . Mosaic expression of lbx1b driven by the GATA2 minimal promoter and the lbx1b enhancer in zebrafish significantly alleviated the embryonic lethal phenotype to allow observation of the later onset of the spinal curvature with or without vertebral malformation . Deformation of the embryonic body axis by lbx1b overexpression was associated with defects in convergent extension , which is a component of the main axis-elongation machinery in gastrulating embryos . In embryos overexpressing lbx1b , wnt5b , a ligand of the non-canonical Wnt/planar cell polarity ( PCP ) pathway , was significantly downregulated . Injection of mRNA for wnt5b or RhoA , a key downstream effector of Wnt/PCP signaling , rescued the defective convergent extension phenotype and attenuated the lbx1b-induced curvature of the body axis . Thus , our study presents a novel pathological feature of LBX1 and its zebrafish homologs in body axis deformation at various stages of embryonic and subsequent growth in zebrafish .
Scoliosis is defined as lateral curvature of the spine with a Cobb angle greater than 10 degrees [1] . It is categorized into congenital , idiopathic , and secondary scoliosis [2] . Congenital scoliosis ( CS ) is caused by embryonic vertebral malformation that results in deviation of the normal spinal alignment [3] . Idiopathic scoliosis ( IS ) is a twisting condition of the spine characterized by rotation of the vertebrae without their malformation and is further categorized into infantile , juvenile , and adolescent type by age of onset . Among these forms , adolescent IS ( AIS ) accounts for 80% of all human scoliosis and develops in 2–4% of children aged between 10 and 16 years across all racial groups [1 , 4] . Secondary scoliosis is attributed to a wide variety of causes such as cerebral palsy , paralysis , Duchenne muscular dystrophy , Marfan syndrome , and Ehlers-Danlos syndrome [5–7] . In contrast , the precise disease mechanisms of both IS and CS are understood poorly [8] . Axial skeletal development occurs through a sequential and coordinated series of events regulated by various growth/differentiation factors [2 , 9] . During gastrulation , the vertebrate embryo elongates along the anterior-posterior axis through a process called convergent extension [10] . The notochord is then formed ventral to the neural tube as the transient embryonic backbone prior to vertebral bone formation in vertebrates [11] . Following somite segmentation in the paraxial mesoderm , which is formed in a well-defined order along the head to tail axis , the sclerotome derived from the ventral part of the somite eventually gives rise to the vertebrae , the annulus fibrosus of the intervertebral discs , and the rib cage [2] . Any anomalies in these processes are considered to result in the development of both CS and IS . The role of hereditary or genetic factors especially in the development of AIS has been widely accepted [8] . AIS is a complex polygenic disease influenced by more than one allele at different loci [12] . Indeed , genome-wide association studies identified several novel susceptibility loci including ladybird homeobox 1 ( LBX1 ) , G protein-coupled receptor 126 , zinc finger protein basonuclin 2 , and paired box 1 ( PAX1 ) [13–16] . Among them , a single nucleotide polymorphism ( SNP ) , rs11190870 in the 3′-flanking region of LBX1 , has been replicated consistently in independent studies using Chinese [17–19] and Caucasian populations [20] . Human LBX1 was first identified as a gene with homology to the ladybird late ( lbl ) gene in Drosophila [21] . The ladybird protein is a member of the homeobox transcription factor family with an engrailed repressor domain [22] . In vertebrates , Lbx genes are expressed in the dorsal spinal cord and hindbrain [23] , a subpopulation of cardiac neural crest cells [24] , muscle precursor cells , and satellite cells of regenerating adult skeletal muscle [25 , 26] . Ectopic expression of LBX1 in chicken somites and limbs activates myogenic markers such as myogenin and myod , owing to the expansion of the myoblastic cell population [27] . Previous in vivo studies using Lbx1 knockout mice and lbx gene knockdown morphants in zebrafish or Xenopus did not reveal phenotypes associated with scoliosis [25 , 28–30] . To our knowledge , the pathological features of Lbx1 and lbx genes in body axis deformation have not been explored . Scoliosis has long been considered to be exclusive to bipedal vertebrates [31] . It has been proposed that the unique human upright posture alters spinal conditions toward the eventual development of scoliosis [32] . Naturally occurring scoliosis is quite rare in quadrupedal vertebrates such as rats and mice [31] . The lack of good animal models in vivo has been a major challenge for studying the etiology of scoliosis . Previously , the experimental animal model available for scoliosis research was the young melatonin-deficient chicken , which develops a three-dimensional spinal deformity consisting of lateral curvature after pinealectomy [33] . Recently , it has become clear that several types of fish including zebrafish are suitable for exploring human scoliosis [14 , 34–38] . AIS-like scoliosis develops in loss-of-function mutants of protein tyrosine kinase 7 ( ptk7 ) [37] and kinesin family member 6 in zebrafish [34] . Sharma et al . reported that the PAX1 enhancer locus in humans is associated with susceptibility to IS in females and its enhancer activity is disrupted by IS-associated SNPs [14] . Loss of collagen type VIII alpha 1 function also reportedly causes CS-like vertebral malformations [38] . In this study , we characterized the most significantly associated SNP , rs11190870 , using chromosome conformation capture ( 3C ) , electrophoretic mobility shift assays ( EMSAs ) , and dual luciferase assays , and then examined the effects of the misregulated expression of LBX1 on axial skeletal development using zebrafish as an animal model by both gain-of-function and loss-of-function approaches . We demonstrate that the elevated expression of human LBX1 or zebrafish lbx1 homologs in zebrafish causes axial developmental defects including defective convergent extension movement and body curvature , which could be attributed to the impairment of non-canonical Wnt/planar cell polarity ( PCP ) signaling . Some zebrafish transiently overexpressing lbx1b survived to develop mild body axis deformation including spinal curvature during larval or juvenile stage . Taken together , our study demonstrated the pathological contribution of lbx genes to body axis deformation in zebrafish .
Human LBX1 and FLJ41350 are located approximately 0 . 6 kb apart in a head-to-head arrangement on human chromosome 10 , and rs11190870 lies 7 . 5 kb downstream of LBX1 ( Fig 1A ) . FLJ41350 is a hypothetical gene that is found only in the human genome , and its function is uncharacterized [15] . We characterized FLJ41350 through exon connection and 5′-rapid amplification of cDNA ends . We confirmed that FLJ41350 is composed of 3 exons , with the predicted translational start site located at exon 1 followed by an open reading frame of 120 amino acids with no known motif ( S1 Fig ) . No orthologs of FLJ41350 are found in any other species except humans . To investigate the functional impact of rs11190870 , we performed EMSAs and found that some nuclear proteins bound specifically to the genome sequences around rs11190870 with higher affinity to the risk allele than the non-risk allele ( Fig 1B ) . We also analyzed the physical interaction between the genome sequence surrounding rs11190870 and its nearby genome regions using the 3C assay [39] with A172 human glioblastoma cells ( A172 cells ) ( Fig 1C ) . The 3C assay is a powerful technique for analyzing chromatin organization to reveal the physical interaction between two distal genomic elements [40] . Digestion of cross-linked chromatin with a restriction enzyme and subsequent intra-molecular ligation produces novel junctions between restriction fragments in proximity in the nucleus , which can be detected by PCR . We confirmed that the specific band with primers F4 and R1 was of the expected length ( Fig 1C ) and corresponded to each primer region by sequencing . This result indicates that the F4 and R1 primer regions are adjacent to each other and that the genome sequence surrounding rs11190870 physically interacts with the promoter region of LBX1-FLJ41350 . We then cloned approximately 1 kb of the LBX1 promoter region ( -917 to +153 ) and evaluated its promoter activity by luciferase assay . In A172 cells , the region had relatively high promoter activity in the direction of LBX1 , but not in that of FLJ41350 ( Fig 1D ) . Moreover , the LBX1 promoter , combined with a 590-bp sequence around rs11190870 that is highly conserved across species , had higher transcriptional activity with the risk allele than with the non-risk allele in HEK 293T cells ( Fig 1E ) . These results suggest that rs11190870 confers AIS susceptibility by upregulating LBX1 transcription . To investigate the effect of the elevated expression of LBX1 on body axis formation , we performed a series of gain-of-function experiments using zebrafish . We overexpressed zebrafish lbx1a , lbx1b , lbx2 , and their mutated genes without the homeodomain or the engrailed homology domain by mRNA injection ( Fig 2A and 2B ) . By 48 hours post-fertilization ( hpf ) , the larvae developed body curvature by the ubiquitous overexpression of any one of these lbx genes , but not by that of the mutated genes without the functional domains ( Fig 2B ) . The incidence of body curvature was highest with lbx1b overexpression and increased in a dose-dependent manner ( Fig 2B and 2C ) . Notably , some lbx1b-overexpressing larvae exhibited notochord deformity and a displaced dorsal melanophore stripe ( S2 Fig ) . In addition , a reduction or complete deletion of the forebrain and eyes was observed in many larvae ( S2 Fig ) . Injection of human LBX1 mRNA caused body curvature in embryos , but injection of human FLJ41350 mRNA failed to induce any obvious phenotype related to body axis morphology ( Fig 2B ) . We also examined the loss-of-function effect on axial development in transcription activator-like effector nuclease ( TALEN ) -mediated knockout zebrafish ( S3 and S4 Figs ) . Unlike overexpression of lbx genes , lbx1b-/- and lbx2-/- mutant larvae displayed a straight trunk comparable to wild-type larvae ( Fig 2D ) , suggesting the involvement of gain-of-function but not loss-of-function of lbx1b in the body curvature phenotype that might be related to scoliosis susceptibility . To confirm defective axial development in the established line with uniform expression of lbx1 in an inducible manner , we employed a Gal4/UAS-based bidirectional expression system for the stable overexpression of lbx1b in zebrafish [41] . The F0 driver transgenic carriers were crossed with the F0 responder transgenic carriers to produce Tg ( hsp:Gal-VP;EGFP:UAS:lbx1b ) F1 progeny with different copies of the Tol2 insertion . Responding to heat shock , embryos with both driver and responder transgenes expressed EGFP and lbx1b driven by the E1b promoter ( Fig 3A ) . The positive correlation between the levels of EGFP and lbx1b expression was examined in Tg ( hsp:Gal-VP;EGFP:UAS:mcherry ) ( S5 Fig ) . Overexpression of lbx1b in Tg ( hsp:Gal-VP;EGFP:UAS:lbx1b ) after heat shock was confirmed by western blotting ( Fig 3B ) . By 48 hpf , body curvature became evident in Tg ( hsp:Gal-VP;EGFP:UAS:lbx1b ) embryos exposed to heat shock at 4 hpf ( Fig 3C and 3D ) . All larvae with lbx1b overexpression died within 7 days post-fertilization ( dpf ) . The severity of body curvature was related to the fluorescence intensity of EGFP in a dose-dependent manner ( Fig 3C and 3E ) . Taken together , we conclude that overexpression of lbx genes , especially lbx1b , induces body curvature in zebrafish embryos . To elucidate the mechanism by which lbx1 overexpression causes embryonic body curvature , we traced back to the pregastrulation stages . Convergent extension movement during gastrulation ( 5 . 25–10 . 33 hpf ) shapes the body axis , narrowing all germ layers in the mediolateral direction and extending them along the anterioposterior axis ( Figs 4A and 5A ) . Embryos overexpressing lbx1b in the gastrulation stage showed mediolateral elongation of somites ( Fig 4B and 4C ) , suggesting some perturbations occur in the formation of the body axis due to abnormal convergent extension . Embryos exposed to heat shock at 4 hpf exhibited a more profound convergent extension defect and more severe body curvature than those at 12 hpf ( Fig 4B–4E ) , demonstrating a positive correlation between the extent of defective convergent extension with the severity of body curvature and the presence of a critical time window for lbx1b overexpression . In situ hybridization for the characteristic markers for the ectoderm or mesoderm revealed a marked delay of convergent movement in embryos overexpressing lbx1b ( Fig 5B and 5C ) . Compared with sibling controls , lbx1b-overexpressing embryos showed a wider neural ectoderm border ( dlx3b ) , broader paraxial mesoderm ( papc ) , and mediolateral elongation ( uncx4 ) of somites ( Fig 5B ) . We also found a significant delay of extension movement ( Fig 5D and 5E ) , which elongated the embryo from head to tail . By contrast , the expression pattern of a dorsal marker , chordin ( chd ) , and a ventral marker , ventral homeobox ( vox ) , in early gastrula was not significantly altered in lbx1b-overexpressing embryos ( S6 Fig ) , indicating that dorsoventral patterning was not affected in these embryos . These results indicate that defective convergent extension resulting from elevated lbx1b expression during gastrulation provokes impaired body axis formation . In vertebrates , non-canonical Wnt/PCP signals are mainly involved in the regulation of convergent extension [10] . Loss of function of wnt5b or wnt11 , the ligand for the non-canonical Wnt/PCP signaling pathway , leads to severely defective convergent extension movement in zebrafish [42 , 43] . Our in situ hybridization study revealed that wnt5b was downregulated in gastrulation embryos upon lbx1b overexpression ( Fig 6A and 6B and S7 Fig ) . In contrast , no significant change was observed in wnt11 expression ( Fig 6A and 6B and S7 Fig ) . We also confirmed by quantitative RT-PCR that wnt5b expression was significantly downregulated at the gastrulation stage in lbx1b-overexpressing embryos ( Fig 6C ) . We then performed an in vivo luciferase assay in zebrafish embryos to analyze the in vivo effects of lbx1b overexpression on the transcriptional activity of two potential promoter regions of wnt5b . Two transcripts encoding 363 ( MN1309037 ) or 380 amino acids were found on the Ensembl website . We tested the sequences upstream of wnt5b including these promoters ( P1 and P2 ) . In 90% epiboly embryos , the P2 promoter had much stronger transcriptional activity ( about 50-fold induction ) than the P1 promoter . Co-injection of lbx1b mRNA repressed the transcriptional activity of P2 by 66 . 7% ( Fig 6D ) . Thus , lbx1b overexpression during gastrulation downregulated the expression of wnt5b largely through repression of the P2 promoter . These results suggest that defective convergent extension caused by the overexpression of lbx1b in embryos could be attributed to impairment of non-canonical Wnt/PCP signaling . To evaluate further the effect of misregulation of non-canonical Wnt/PCP signaling in defective convergent extension caused by lbx1b overexpression , we performed a rescue experiment by overexpressing wnt5b , a ligand of the Wnt/PCP pathway . We optimized the amount of wnt5b mRNA injection to avoid defects caused by its overexpression in embryos . Defective migration of dlx3b-positive cells in embryos injected with lbx1b mRNA was rescued by co-injection of lbx1b and wnt5b mRNA ( Fig 7A and 7B ) . Wnt5b mRNA injection mostly rescued the body curvature phenotype in Tg ( hsp:Gal-VP;EGFP:UAS:lbx1b ) with heat shock at 4hpf ( Fig 7C and 7D ) . We further examined whether defects caused by lbx1b overexpression can be rescued by overexpressing RhoA or Rac1 small GTPases , both of which are downstream effectors of the Wnt/PCP pathway . RhoA rescued both the defective convergent extension and body curvature phenotype ( Fig 7A and 7B ) , whereas Rac1 failed to rescue the convergent extension defects and body curvature ( S8 Fig ) . Interestingly , RhoA overexpression was not effective in larvae with heat shock at 12 hpf ( S9 Fig ) . These results demonstrate that impairment of non-canonical Wnt/PCP signaling , especially the wnt5b/RhoA pathway , caused by lbx1b overexpression , contributes to defective convergent extension and curvature of the body axis . To investigate the effects of lbx1b overexpression on endogenous expression domains during axial development , we forced lbx1b expression under the control of the previously characterized lbx1b enhancer [44] and the GATA2 minimal promoter by microinjecting a GATA2-1b:lbx1b plasmid ( Fig 8A ) . We confirmed that reporter expression driven by the regulatory elements in Tg ( GATA2-1b:EGFP ) generally recapitulated the endogenous expression of lbx1b , lbx1a , or lbx2 at different developmental stages ( S10 Fig ) . Similarly to embryos injected with mRNAs , many 48 hpf embryos injected with GATA2-1b:lbx1b developed severe body curvature ( S11 Fig ) , abnormalities in somite morphology ( Fig 8B ) , notochord deformity ( S12A Fig ) , and a displaced dorsal melanophore stripe ( Fig 8C ) . Some of the larvae with a displaced dorsal melanophore stripe had no apparent notochord deformity ( S12B Fig ) . The majority of Tg ( GATA2-1b:lbx1b ) F1 embryos presented with a severe malformation and died within 24 hpf ( S13 Fig ) . Some were alive at 48 hpf , developing serious axial body curvature , but died within 72 hpf ( S13 Fig ) . Unlike the F1 generation of Tg ( GATA2-1b:lbx1b ) , which is embryonic lethal , some founder Tg ( GATA2-1b:lbx1b ) with almost a straight trunk could survive to adulthood , thus allowing our observation of the later developmental stages in this model . We monitored embryos with a mild notochord deformity induced by injection of GATA2-1b:lbx1b ( S12A Fig ) ( n = 41 ) until adulthood , together with wild-type siblings as controls ( n = 45 ) . Thirteen Tg ( GATA2-1b:lbx1b ) and two control zebrafish died within 21 days . The deformed notochord ( S12A Fig red arrow ) gradually ossified to form a spine , leading to vertebral malformations ( n = 27 , p < 0 . 01 ) such as hemivertebrae ( Fig 8D , white arrow ) and block vertebra ( Fig 8D , yellow arrow ) at the location of the notochord deformity . Eventually , these zebrafish showed scoliosis with vertebral malformations mimicking CS ( Fig 8E and S1 video ) . No apparent spinal deformity was identified in the control ( Fig 8B–8E ) . Thus , local notochord deformity in founder Tg ( GATA2-1b:lbx1b ) develops into CS-like scoliosis with vertebral malformations . To investigate the possibility of AIS-mimicking scoliosis in Tg ( GATA2-1b:lbx1b ) during the period corresponding to human adolescence ( Fig 8F ) , we kept transgenic larvae that had a displaced dorsal melanophore stripe without an apparent notochord deformity ( n = 45 ) , together with their wild-type siblings ( n = 60 ) . Eight Tg ( GATA2-1b:lbx1b ) and two control zebrafish died within 30 days . In 19 of the 37 surviving Tg ( GATA2-1b:lbx1b ) ( p < 0 . 01 ) , significant scoliosis , with rotation of the longitudinal body axis but without visible vertebral malformations , was observed by 55 dpf and then developed progressively until 90 dpf ( Fig 8F , 8G and S2 video ) . Additionally , there was a significant female bias ( 16/19 ) for the prevalence of scoliosis ( p < 0 . 01 ) . No apparent spinal deformity was identified in the control group ( 0/58 ) . These results indicate that mild body axis deformation resulting from the increased expression of lbx1b could cause irregular trunk development such as notochord deformity and a displaced dorsal melanophore stripe , further leading to the later development of CS- or AIS-like scoliosis .
We demonstrate here that the most significantly associated SNP , rs11190870 [15] could confer AIS susceptibility by activating LBX1 transcription . Our gain-of-function approaches using the zebrafish model revealed that the elevated expression of human LBX1 or any of the zebrafish genes lbx1a , lbx1b , and lbx2 causes body axis deformation at various stages of embryonic and subsequent growth in zebrafish . Embryonic body curvature prior to vertebral column formation is associated with defective convergent extension involving the downregulation of wnt5b during gastrulation to disrupt axial development . Defective convergent extension and embryonic body curvature phenotypes were mostly rescued by the overexpression of wnt5b and RhoA , key molecules in the Wnt/PCP signaling pathway . An embryonic lethal phenotype could be alleviated by chimeric expression of lbx1b under the control of the GATA2 minimal promoter and the lbx1b enhancer in larvae , thus allowing observation of the later onset of the spinal curvature with or without vertebral malformation in zebrafish . Thus , as a step towards better understanding of the genetic pathophysiology of scoliosis , our study provide a new evidence for a pathological role of LBX1 and its zebrafish homologs in body axis deformation . The most significant SNP associated with AIS ( rs11190870 ) is located in the intergenic region [15] . The nearest gens are LBX1 and FLJ41350 , which are 7 . 5 kb upstream and 8 . 1 kb downstream of rs11190870 , respectively . Using 3C assays , we found that the genome sequence surrounding rs11190870 physically interacts with the LBX1 and FLJ41350 promoters . In luciferase assays , significantly higher promoter activity was detected in the direction toward LBX1 , but not toward FLJ41350 . EMSAs revealed that some nuclear proteins bound specifically to the genome sequences around rs11190870 with higher affinity to the risk allele . Given that risk variants could disrupt or create a binding site for a transcription factor , any change of LBX1 expression driven by the variants , including downregulation , upregulation , and alteration in temporospatial distribution , would be possible . Expression quantitative trait loci ( eQTL ) data are available only for peripheral blood cells , which showed no association between the LBX1 expression level and the rs11190870 genotype ( Human genetic variation database . ( http://www . genome . med . kyoto-u . ac . jp/SnpDB/index . html ) . However , further studies on eQTL are hampered by a lack of information on which types of tissues or cells are relevant to AIS pathogenesis . So far , phenotypes associated with CS and AIS have not been reported in Lbx1 null mice and lbx gene knockdown morphants in zebrafish or Xenopus [25 , 28–30] . Our lbx1b or lbx2 single knockout zebrafish mutants generated by targeting the first exon using Platinum TALENs [45] also did not exhibit embryonic axial body curvature or scoliosis . The database of Zebrafish Mutation Project also shows that normal development is observed in lbx1a or lbx2 nonsense mutants ( https://www . sanger . ac . uk/sanger/Zebrafish_Zmpsearch/lbx1 ) , although information on the associated phenotype of double or triple knockout zebrafish is not available . Previous studies demonstrated a dominant-negative effect by the removal of the engrailed domain from Xenopus Lbx1 that normally functions as a repressor [22 , 46] . Injection of lbx1aΔeh , lbx1bΔeh , or lbx2Δeh mRNA did not cause any body curvature as shown in our study . Thus , the current data do not support the possibility that loss-of-function of LBX1 is involved in susceptibility to scoliosis . In contrast , we found a significant increase of promoter activity in the presence of the genomic region with rs11190870 found in the risk allele . Considering that rs11190870 could confer AIS susceptibility by activating LBX1 transcription , it would be reasonable to assume that upregulation of human LBX1 may contribute to some aspects of the pathogenic mechanism in scoliosis . The ladybird protein is a member of the homeobox transcription factor family with an engrailed repressor domain at the N-terminus [22] . Overexpression of LBX1 and any one of lbx1a , lbx1b , or lbx2 caused defective convergent extension movements that led to curvature of the body axis . Upon overexpression of the lbx genes without the engrailed repressor domain , body curvature was not observed in the embryos , suggesting that Lbx genes negatively regulate their target genes as repressors . Indeed , our in vivo luciferase assays revealed that lbx1b significantly represses the promoter activity of wnt5b . Hence , lbx1b downregulates wnt5b expression during gastrulation at the transcriptional level , thereby causing defective convergent extension followed by deformation of the body axis . Both canonical and non-canonical Wnt signaling pathways are involved in convergent extension movements during gastrulation . A shortened-curled tail was reported in a Wnt-5 mutant ( ppt−/− ) with defective convergent extension [47] . AIS- and CS-like scoliosis are also observed in zebrafish mutants of ptk7 , which regulates both canonical and non-canonical Wnt signaling activity [37 , 48] . The same group identified a novel sequence variant within a single IS patient that disrupted PTK7 . In this study , we found that the elevated expression of lbx1 in zebrafish evokes wnt5b downregulation , suggesting that aberrant Wnt/PCP signaling causes defective convergent extension in our experimental model . Interestingly enough , our approach investigating the etiology of scoliosis from the opposite direction also led to the hypothesis that a dysregulated Wnt signaling pathway is involved in both CS and IS pathogenesis . Non-canonical Wnt/PCP signaling is involved in a variety of events independently of β-catenin [49] . During axis formation in vertebrates , the Wnt/PCP pathway regulates cell polarity and cell motility by modulating the activity of Rho family small GTPases . Especially , RhoA-ROCK signaling mainly acts downstream of wnt5 and wnt11 in zebrafish embryos [50] . Co-injection of mRNA for wnt5b or RhoA mRNA with lbx1b mRNA rescued defective convergent extension leading to embryonic body curvature . These findings strongly support our hypothesis that misregulation of Wnt/PCP signaling induced by lbx1b overexpression is responsible for defective convergent extension followed by body axis deformation . To date , CS and IS have been considered not to be etiologically relevant , but it has been reported previously that a family history of IS was observed in 17 . 3% of 237 families with CS [51] . Another study of 31 CS cases also reported that three ( 10% ) had first-degree relatives with IS [52] . The overlapping familial aggregates of CS and IS suggested the possibility of a common cause for these clinically distinct diseases . The uniform overexpression of lbx1b either ubiquitously or in the endogenous expression domain results in severe defective convergent extension leading to morphological defects in both mesoderm and ectoderm patterning followed by early death prior to notochord mineralization to form the spine . In contrast , mosaic expression of lbx1b under the control of the lbx1b enhancer in larvae alleviated the embryonic lethal phenotype with body curvature and thereby allowed the later onset of scoliosis with or without vertebral malformation in zebrafish . The F1 embryos generated by the AIS- and CS-like mosaic transgenic founders presented with severe body axis deformation including convergent extension defects and body curvature . Thus , our observations that the elevated expression of lbx1b causes both AIS- and CS-like scoliosis may provide a new perspective for the shared genetic basis of AIS and CS . Some of the founder Tg ( GATA2-1b:lbx1b ) with a displaced dorsal melanophore stripe without apparent notochord deformity developed scoliosis with rotation around the longitudinal axis of the body , but without visible vertebral malformations . These results suggest that subtle deformities in the early body axis may be later accentuated during the growth spurt . In fact , AIS patients appear to be quite normal until adolescence . It is reasonable to postulate that an early event such as defective axial development resulting from the upregulation of LBX1 may be too mild to be detected in potential AIS patients until the growth spurt . In a late-onset polygenic disease such as AIS , even such subtle abnormalities may be sufficient to accumulate growth irregularities and greatly aggravate biomechanical instability during adolescence in association with additional genetic or environmental factors . However , at present , considering that the GATA2 minimal promoter and an lbx1b enhancer could drive lbx1b expression in neural tissue later in development [44] , we cannot exclude the possibility that lbx1b expression after convergent extension causes idiopathic scoliosis . Thus , we need to determine carefully the mechanism by which lbx1b causes the AIS-like phenotype in the mosaic transgenic founders . Polygenic diseases including AIS are triggered by the combination of a number of susceptibility genes whose individual contribution may be relatively small . It is also considered that these diseases could occur when a threshold of quantitatively-varying risk or liability influenced genetically and environmentally is exceeded [53] . Unlike a monogenic disease caused by a mutation in one gene , it appears that the cumulative effects combined with additional factors for a relatively long time lead to the onset of clinical manifestations of AIS , even though the contribution of each individual gene is small . Our study provides a new evidence for the possible involvement of LBX1-induced mild defects during embryonic axial development in AIS susceptibility . As the faithful recapitulation of the late-onset polygenic disease in the animal model has not been generally established yet , our current experimental approaches are still fraught with limitations . Further studies are necessary for establishment of a genetic animal model recapitulating the expression of LBX1 in an analogous way to that in AIS patients .
All of the animal experimental procedures used in this study were approved by the Animal Care Committee of the Institute for Frontier Medical Sciences , Kyoto University and conformed to institutional guidelines for the study of vertebrates Rhabdomyosarcoma cells and A172 human glioblastoma cells ( A172 cells ) from ATCC and HeLa cells were obtained from the Japanese Collection of Research Bioresources Cell Bank ( Osaka , Japan ) . The cells were maintained at 37°C under 5% CO2 in Dulbecco’s modified Eagle’s medium-high glucose supplemented with penicillin ( 50 U/mL ) , streptomycin ( 50 g/mL ) , and 10% fetal bovine serum . The cells were crosslinked with 37% formaldehyde solution at a final concentration of 1% in a 37°C dry incubator for 10 min , followed by an additional incubation at 4°C for 2 h . The crosslinked protein-chromatin material was purified by 8 M urea ultracentrifugation and digested with Sau3AI as described previously . A 2-g aliquot of chromatin was diluted in a ligation buffer and ligated with T4 DNA ligase ( Fermentas ) for 4 h . After reversing the crosslinks , the ligated DNA was amplified by PCR with various combinations of primers using GoTaq Hot Start Master Mix ( Promega ) . We prepared nuclear extracts from rhabdomyosarcoma and A172 cells as described previously [54] . We prepared probes for the risk ( R ) and non-risk ( N ) alleles of rs11190870 by annealing 17-bp complementary oligonucleotides and labeling with digoxigenin ( DIG ) -11-ddUTP ( Roche ) . For competition experiments , nuclear extracts were pre-incubated with excess unlabeled probes . We detected DNA-protein complexes using a DIG gel shift kit according to the manufacturer’s instructions ( Roche ) . We amplified the LBX1 promoter fragment ( -917 to +153 ) in both directions by PCR and cloned them into the pGL4 . 10 promoter-less luciferase reporter vector ( Promega ) . The constructs were co-transfected with the pGL4 . 73 Renilla luciferase vector ( hRluc/SV40 ) as an internal control . Transfection of each construct was performed using TransIT-LT1 ( Mirus Bio LLC ) . HEK 293T cells were maintained at 37°C under 5% CO2 in Dulbecco’s modified Eagle’s medium-high glucose supplemented with 10% fetal bovine serum . Transfection was performed with Lipofectamine LTX and PLUS reagent ( Life Technologies ) . After 24 h of transfection , the cells were harvested and luciferase activity was measured using a Pick&gene dual luciferase detection kit ( Toyo B-Net Co . ) . The RIKEN Wako ( RW ) strain and AB strain were obtained from the Zebrafish National BioResource Center of Japan ( http://www . shigen . nig . ac . jp/zebra/ ) and Kondoh ERATO Laboratory , respectively . Adult fish were maintained under a 14 h light–10 h dark cycle at 28°C . Embryos were kept at 28°C and staged by hpf or dpf [55] . The RW strain was subjected to micro-injection and whole-mount in situ hybridization . The AB strain was used for the preparation of total RNA . The established line Tg ( UAS:EGFP ) [41] was generously provided by Dr . Koichi Kawakami ( National Institute of Genetics ) . Specific primers for zebrafish lbx1a ( NM_001025532 ) , lbx1b ( NM_001163312 ) , and lbx2 ( NM_001007134 ) , and human LBX1 ( NM_006562 . 4 ) , FLJ41350 ( NR_029380 ) , and RhoA ( NM_001664 . 2 ) were designed based on the nucleotide sequences from GenBank ( S1 Table ) . The cDNAs were amplified by PCR from a cDNA library and cloned into the pCS2 ( + ) vector . Deletion constructs of the engrailed domain and homeodomain of lbx1 were generated by inverse PCR . Capped mRNAs were synthesized using an SP6 RNA polymerase in vitro transcription kit ( Life Technologies ) and purified using a MEGAclear Kit ( Life Technologies ) according to the manufacturer’s instructions . A mixture containing 50/100/150 pg mRNA for lbx1a , lbx1b , lbx2 , LBX1 , and FLJ41350 , 15 pg mRNA for RhoA , and 40 pg mRNA for wnt5b and RAC1 was injected into the cytoplasm of one-cell-stage embryos . Highly active Platinum TALENs were constructed using two-step Golden Gate assembly method as described previously with a slight modification [45] . DNA-binding modules were assembled with the two-step Golden Gate method using the Platinum Gate TALEN Kit ( Addgene , Kit #1000000043 ) . pCS2-based vectors were used as destination vectors . The target sequence was 5’-TAAACCCCCTGGACCACcttccaccacccgcgAGCTCCAACAA GCCCTTA-3’ , where uppercase and lowercase letters indicate lbx1b TALEN recognition sequence and spacer sequence , respectively . We found polymorphism of RW WT in the left TALE-binding sequence; TAAACCCCCTGGACCAC and TGAATCCCCTGGACCAC , both of which are silent mutations . The target sequence was 5’-TTGCAGTCCAGCGGCGAG gagaggcggcggggtCCCTTGGACCAACTCCCA-3’ , where uppercase and lowercase letters indicate lbx2 TALEN recognition sequence and spacer sequence , respectively . Genomic DNA was extracted from the caudal fins of lbx1b TALENs mRNA-injected zebrafish . For sequencing , PCR products were amplified from the genomic DNA and phosphorylated by T4PNK ( TAKARA BIO INC . ) , and then subcloned into EcoRV site of pBluescript II SK ( + ) vector . We identified F0 fishes carrying multiple mutations in the target site , and then generated and screened a F1 fish with a nonsense mutation by crossing the F0 and wild type fishes . lbx1b+/- and lbx2+/- mutant were generated by crossing F1 and wild type fishes . lbx1b-/- and lbx2-/- were further generated by intercrossing lbx1b+/- and lbx2+/- , respectively . The Tol2 transposon/transposase system [41 , 56–59] was employed for the establishment of transgenic zebrafish . The coding sequence of lbx1b was cloned into pME-MCS to generate pME-lbx1b . The complementary sequence of the E1b promoter , EGFP , and polyA were cloned into p5E-UAS-E1b to generate p5E-polyA-EGFP-E1b-UAS-E1b . The driver construct ( Fig 3A ) was generated by recombining p5E-hsp70I , pME-Gal4VP16 , p3E-polyA , and pDestTol2CG2 with Gateway LR Clonase II Enzyme mix ( Life Technologies ) . Similarly , the responder construct ( Fig 3A ) was generated by recombining p5E-polyA-EGFP-E1b-UAS-E1b , pME-lbx1b , p3E-polyA , and pDestTol2CG2 . Capped mRNA of medaka Tol2 transposase was prepared by in vitro translation as described above . A mixture containing 50 pg transposase mRNA and 40 pg Tol2 transgenic plasmid was injected into the cytoplasm of one-cell-stage embryos . F1 fish were acquired by outcrossing EGFP-positive F0 with RW fish , and screened by cardiac fluorescence . F2 lines were then generated by outcrossing F1 and RW fish . All kept F2 lines yielded about 50% EGFP-positive progeny when mated to RW fish , which suggested there was a single Tol2 insertion site . The lbx1b enhancer located from +1316 to +2383 bp downstream of the lbx1b transcription start site [44] was cloned into pME-MCS . The enhancer activity in vivo was confirmed using zebrafish enhancer detection ( ZED ) [60] . To generate p5E-lbx1b enhancer-GATA2 , the 2 . 3-kb BamHI fragment from the ZED-lbx1b enhancer , was cloned into the BamHI site of p5E-MCS . The constructs GATA2-1b:lbx1b or GATA2-1b:MCS were generated by recombining p5E-lbx1b enhance-GATA2 , pME-lbx1b or pME-MCS , p3E-polyA , and pDestTol2CG2 . Embryos at 4 hpf or 12 hpf in E3 buffer were placed on block incubator and heated up to 38°C gradually , and then maintained at 38°C for 30 min . After heat shock treatment , they were gradually cooled to 28 . 5°C . The heat shock treatment causes neither an anomaly nor a decrease in viability . Larvae were homogenized with a Dounce tissue grinder and lysed with lysis buffer ( 50 mM Tris-HCl [pH 7 . 5] , 150 mM NaCl , 1% NP-40 , 0 . 1% SDS ) containing protease inhibitor cocktail ( Roche ) . Lysates were mixed with 5× Laemmli sampling buffer containing 100 mM DTT and boiled at 95°C for 3 min . Proteins were separated by SDS-PAGE and transferred onto PVDF membranes ( Merck Millipore ) . After blocking with BLOCKING ONE ( Nacalai Tesque ) , the membranes were incubated with primary antibodies in phosphate-buffered saline containing 0 . 1% Tween 20 and 10% BLOCKING ONE , followed by incubation with horseradish peroxidase-conjugated secondary antibodies . Signals were detected with SuperSignal West Pico Chemiluminescent Substrate ( Thermo Scientific ) and images were captured by ImageQuanta LAS 4000 ( GE Healthcare Bio-Sciences ) . Whole-mount in situ hybridization was performed as described previously [61] . Sense and antisense riboprobes for hgg1 , dlx3b , ntl , papc , uncx4 , wnt5b , wnt11 [42] , chd and vox [48] were generated by in vitro translation using a digoxigenin ( DIG ) RNA labeling kit with T7 or T3 RNA polymerase ( Roche ) . Hybridization signals were detected with an alkaline phosphatase-conjugated anti-DIG antibody ( Roche ) according to the manufacturer’s instructions . For quantification , the image colors of in situ hybridization were inverted , and Area , Integrated Density , and Mean Gray Value were measured by ImageJ . The corrected Gray Value = Integrated Density − ( Area of the selected embryos × Mean Gray Value of background readings ) . Total RNA was extracted from zebrafish embryos injected with lbx1b mRNA using an RNeasy Plus Mini kit ( QIAGEN ) . Two hundred nanograms of total RNA were used to synthesize cDNA with a PrimeScript RT reagent Kit ( Takara Bio ) . Quantitative RT-PCR was performed using SYBR Premix Ex Taq II ( Takara Bio ) on a StepOne instrument ( Life Technologies ) . Relative mRNA expression was normalized to ef-1α and calculated using the 2−ΔΔCT method . Specific primers for quantitative RT—PCR are listed in S1 Table . A 569-bp insulator of chicken β-globin ( BGI ) and firefly luciferase ( luc ) were amplified from the ZED vector and pGL3-basic vector ( Promega ) , respectively . The resultant amplification products were cloned into pME-MCS to construct the promoter-less pME-BGI-luc plasmid . Two fragments of approximately 2 kb upstream of each transcription start site of wnt5b were amplified from zebrafish genome DNA with the primers listed in S1 Table . These were cloned into pME-BGI-luc , pME-BGI-P1-luc plasmid ( P1 ) , and pME-BGI-P2-luc ( P2 ) by the In-Fusion technique ( Clontech ) . A mixture containing fluorescein isothiocyanate ( FITC ) -dextran ( SIGMA ) and luciferase plasmids with or without lbx1b mRNA was injected into one-cell-stage embryos . FITC fluorescence intensity was quantified using a fluorescence microscope ( Leica MZ 16 FA ) and ImageJ software . Embryos were then lysed individually and luciferase activity was measured as described previously [62] . The measured activity was normalized by the FITC fluorescence intensity of an individual embryo . Bones in fixed larvae were stained with alizarin red ( Wako ) . Vertebral bone morphology of adult zebrafish was analyzed by micro-computed tomography scans with inspeXio SMX-90CT ( SHIMADZU ) . Three-dimensional reconstruction and videos were generated with ImageJ software . Embryos were examined and scored for relevant phenotypes . Statistical analysis ( SPSS 16 . 0 ) was performed by chi-square analysis for enumeration data and independent-samples t test or one-way ANOVA for measurement data to calculate p values under various conditions . Spearman’s correlation between relative fluorescence intensity and body curvature severity was calculated . A linear regression equation was calculated with SPSS . | Scoliosis is the most common type of spinal deformity with a lateral spinal curvature of at least 10 degrees , affecting 2–4% of the population . Scoliosis caused by a primary problem related to the spine itself is classified into congenital scoliosis ( CS ) and idiopathic scoliosis ( IS ) . Among these , adolescent idiopathic scoliosis ( AIS ) , the most common form of scoliosis , is known as a common polygenic disease . Severe curving of the spine in scoliosis leads to profound psychological and social impacts , but etiology-based therapies have not been established since the precise pathological mechanisms of both IS and CS remain undefined . Previously , we identified an AIS susceptibility locus near human ladybird homeobox 1 ( LBX1 ) by a genome-wide association study . Here , we report the functional characterization of the most significantly associated single nucleotide polymorphism ( SNP ) , rs11190870 and LBX1 as well as its zebrafish homologues . Overexpression of LBX1 and zebrafish lbx genes caused lateral body curvature in association with the impairment of non-canonical Wnt/planar cell polarity signaling . Thus , our study presents a novel pathological feature of LBX1 in body axis deformation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"classical",
"mechanics",
"vertebrates",
"animals",
"somites",
"animal",
"models",
"osteichthyes",
"developmental",
"biology",
"model",
"organisms",
"organism",
"development",
"damage",
"mechanics",
"molecular",
"biology",
"techniques",
"embryos",
"research",
"and",
"anal... | 2016 | Functional Investigation of a Non-coding Variant Associated with Adolescent Idiopathic Scoliosis in Zebrafish: Elevated Expression of the Ladybird Homeobox Gene Causes Body Axis Deformation |
Viruses and bacteria are critical components of the human microbiome and play important roles in health and disease . Most previous work has relied on studying bacteria and viruses independently , thereby reducing them to two separate communities . Such approaches are unable to capture how these microbial communities interact , such as through processes that maintain community robustness or allow phage-host populations to co-evolve . We implemented a network-based analytical approach to describe phage-bacteria network diversity throughout the human body . We built these community networks using a machine learning algorithm to predict which phages could infect which bacteria in a given microbiome . Our algorithm was applied to paired viral and bacterial metagenomic sequence sets from three previously published human cohorts . We organized the predicted interactions into networks that allowed us to evaluate phage-bacteria connectedness across the human body . We observed evidence that gut and skin network structures were person-specific and not conserved among cohabitating family members . High-fat diets appeared to be associated with less connected networks . Network structure differed between skin sites , with those exposed to the external environment being less connected and likely more susceptible to network degradation by microbial extinction events . This study quantified and contrasted the diversity of virome-microbiome networks across the human body and illustrated how environmental factors may influence phage-bacteria interactive dynamics . This work provides a baseline for future studies to better understand system perturbations , such as disease states , through ecological networks .
Viruses and bacteria are critical components of the human microbiome and play important roles in health and disease . Bacterial communities have been associated with disease states , including a range of skin conditions [1] , acute and chronic wound healing conditions [2 , 3] , and gastrointestinal diseases , such as inflammatory bowel disease [4 , 5] , Clostridium difficile infections [6] and colorectal cancer [7 , 8] . Altered human viromes ( virus communities consisting primarily of bacteriophages ) also have been associated with diseases and perturbations , including inflammatory bowel disease [5 , 9] , periodontal disease [10] , spread of antibiotic resistance [11] , and others [12–17] . Viruses act in concert with their microbial hosts as a single ecological community [18] . Viruses influence their living microbial host communities through processes including lysis , host gene expression modulation [19] , influence on evolutionary processes such as horizontal gene transfer [20] or antagonistic co-evolution [21] , and alteration of ecosystem processes and elemental stoichiometry [22] . Previous human microbiome work has focused on bacterial and viral communities , but have reduced them to two separate communities by studying them independently [5 , 9 , 10 , 12–17] . This approach fails to capture the complex dynamics of interacting bacteria and phage communities , which frequently share genetic information and work together to maintain ecosystem structure ( e . g . kill-the-winner dynamics that prevent domination by a single bacterium ) . Removal of bacteria or phages can disrupt or even collapse those ecosystems [18 , 23–32] . Integrating these datasets as relationship-based networks allow us to capture this complex interaction information . Studying such bacteria-phage interactions through community-wide networks built from inferred relationships begins to provide us with insights into the drivers of human microbiome diversity across body sites and enable the study of human microbiome network dynamics overall . In this study , we characterized human-associated bacterial and phage communities by their inferred relationships using three published paired virus and bacteria-dominated whole community metagenomic datasets [13 , 14 , 33 , 34] . We leveraged machine learning and graph theory techniques to establish and explore the human bacteria-phage network diversity therein . This approach built upon previous large-scale phage-bacteria network analyses by inferring interactions from metagenomic datasets , rather than culture-dependent data [28] , which is limited in the scale of possible experiments and analyses . We implemented an adapted metagenomic interaction inference model that made some improvements upon previous phage-host interaction prediction models . Previous approaches have utilized a variety of techniques , such as linear models that were used to predict bacteria-phage co-occurrence using taxonomic assignments [35] , and nucleotide similarity models that were applied to both whole virus genomes [36] and clusters of whole and partial virus genomes [37] . Our approach uniquely included protein interaction data and was validated based on experimentally determined positive and negative interactions ( i . e . who does and does not infect whom ) . We built on previous modeling work as a means to our ends , and focused on the biological insights we could gain instead of building a superior model and presenting our work as a toolkit . We therefore did not focus on extensive benchmarking against other existing models [36 , 37–40] . Through this approach we were able to provide an initial basic understanding of the network dynamics associated with phage and bacterial communities on and in the human body . By building and utilizing a microbiome network , we found that different people , body sites , and anatomical locations not only support distinct microbiome membership and diversity [13 , 14 , 33 , 34 , 41–43] , but also support ecological communities with distinct communication structures and robustness to network degradation by extinction events . Through an improved understanding of network structures across the human body , we aim to empower future studies to investigate how these communities dynamics are influenced by disease states and the overall impact they may have on human health .
We studied the differences in virus-bacteria interaction networks across healthy human bodies by leveraging previously published shotgun sequence datasets of purified viral metagenomes ( viromes ) paired with bacteria-dominated whole community metagenomes . Our study contained three datasets that explored the impact of diet on the healthy human gut virome [14] , the impact of anatomical location on the healthy human skin virome [13] , and the viromes of monozygotic twins and their mothers [33 , 34] . We selected these datasets because their virome samples were subjected to virus-like particle ( VLP ) purification , which removed contaminating DNA from human cells , bacteria , etc . To this end , the publishing authors employed combinations of filtration , chloroform/DNase treatment , and cesium chloride gradients to eliminate organismal DNA ( e . g . bacteria , human , fungi , etc ) and thereby allow for direct assessment of both the extracellular and fully-assembled intracellular virome ( S1A and S1B Fig ) [14 , 34] . Each research group reported quality control measures to ensure the purity of the virome sequence datasets , using both computational and molecular techniques ( e . g . 16S rRNA gene qPCR ) ( S1 Table ) . These reports confirmed that the virome libraries consisted of highly purified virus genomic DNA . The bacterial and viral sequences from these studies were quality filtered and assembled into contigs ( i . e . genomic fragments ) . We further grouped the related bacterial and phage contigs into operationally defined units based on their k-mer frequencies and co-abundance patterns , similar to previous reports ( S2 and S3 Figs ) [37] . This was done both for dimensionality reduction and to prevent inflation of node counts by using contigs which are expected to represent multiple fragments from the same genomes . This was also done to create genome analogs that we could use in our classification model which was built using genome sequences . We referred to these operationally defined groups of related contigs as operational genomic units ( OGUs ) . Each OGU represented a genomically similar sub-population of either bacteria or phages . Contig lengths within clusters ranged between 103 and 105 . 5 bp ( S2 and S3 Figs ) . The original publications reported that the whole metagenomic shotgun sequence samples , which primarily consisted of bacteria , had an average viral relative abundance of 0 . 4% ( S1 Table ) [13 , 14 , 33 , 34] . We confirmed these reports by finding that only 2% ( 6/280 OGUs ) of bacterial OGUs had significantly strong nucleotide similarity to phage reference genomes ( e-value < 10−25 ) [13 , 14 , 33 , 34] . Additionally , no OGUs were confidently identified as lytic or temperate phage OGUs in the bacterial dataset using the Virsorter algorithm [44] . We also supplemented the previous virome fraction quality control measures ( S1 Table ) to find that , in light of the rigorous purification and quality control during sample collection and preparation , 77% ( 228/298 operational genomic units ) still had some nucleotide similarity to a given bacterial reference genome ( e-value < 10−25 ) . It is important to note that interpreting such alignment is complicated by the fact that most reference bacterial genomes also contain prophages ( i . e . phages integrated into bacterial genomes ) , meaning we do not know to what extent the alignments were the result of bacterial contaminants in the virome fraction and what were true integrated prophages . As most phages in these communities have been shown to be temperate ( i . e . they integrate into bacterial genomes ) using methods that included nucleotide alignments of phages to bacterial reference genomes , we expected that a large fraction of those phages were temperate and therefore shared elements with bacterial reference genomes—a trend previously reported [14] . To ensure the purity of our sample sets , we supplemented the quality control measures by filtering out all OGUs that could be potential bacterial contaminants , as described previously [37] . This resulted in the removal of 143 OGUs that exhibited nucleotide similarity to bacterial genomes but no identifiable known phage elements . We were also able to identify two OGUs as representing complete , high confidence phages using the stringent Virsorter phage identification algorithm ( class 1 confidence group ) [44] . We predicted which phage OGUs infected which bacterial OGUs using a random forest model trained on experimentally validated infectious relationships from six previous publications [36 , 45–49] . Only bacteria and bacteriophages were used in the model . The training set contained 43 diverse bacterial species and 30 diverse phage strains , including both broad and specific ranges of infection ( S4A and S4B Fig , S2 Table ) . While it is true that there are more known phages that infect bacteria , we were limited by the information confirming which phages do not infect certain bacteria and attempted to keep the numbers of positive and negative interactions similar . Phages with linear and circular genomes , as well as ssDNA and dsDNA genomes , were included in the analysis . Because we used DNA sequencing studies , RNA phages were not considered ( S4C and S4D Fig ) . This training set included both positive relationships ( i . e . a phage infects a bacterium ) and negative relationships ( i . e . a phage does not infect a bacterium ) . This allowed us to validate the false positive and false negative rates associated with our candidate models , thereby building upon previous work that only considered positive relationships [36] . It is also worth noting that while a positive interaction is strong evidence that the interaction exists , we must also be conscious that negative interactions are only weak evidence for a lack of interaction because the finding could be the result of our inability to reproduce conditions in which those interactions occur . Altogether we decided to maintain a balanced dataset at the cost of under-sampling the available positive interaction information because the use of such a severely unbalanced dataset often results in over-fit and uninformative model training . However , as an additional validation measure , we used the extensive additional positive interactions as a secondary dataset to confirm that we could identify infectious interactions from a more diverse bacterial and phage dataset . Using this approach , we confirmed that 382 additional phage reference genomes , representing a diverse range of phages , were matched to at least one reference bacterial host genome of the species that they were expected to infect ( S5 Fig ) . Because the model was built on full genomes and used on OGUs , we also assessed whether our model was resilient to incomplete reference genomes . We found that the use of our model on random contigs representing as little as 50% length of the original reference phage and bacterial genomes resulted in minimal reduction in the ability of the model to identify relationships ( S6 Fig ) . Four phage and bacterial genomic features were used in our random forest model to predict infectious relationships between bacteria and phages: 1 ) genome nucleotide similarities , 2 ) gene amino acid sequence similarities , 3 ) bacterial Clustered Regularly Interspaced Short Palindromic Repeat ( CRISPR ) spacer sequences that target phages , and 4 ) similarity of protein families associated with experimentally identified protein-protein interactions [50] . These features were calculated using the training set described above . While the nucleotide and amino acid similarity metrics were expected to identify prophage signatures , the protein family interaction and CRISPR signatures were expected to aid in identifying lytic phages in addition to temperate phages . We chose to utilize these metrics that directly compare nucleotide sequences between sample phages and bacteria , instead of relying on alignment to reference genomes or known marker genes , because we were extrapolating our model to highly diverse communities which we expect to diverge significantly from the available reference genomes . The resulting random forest model was assessed using nested cross validation , and the median area under its receiver operating characteristic ( ROC ) curve was 0 . 788 , the median model sensitivity was 0 . 905 , and median specificity was 0 . 538 ( Fig 1A ) . This balance of confident true positives at the cost of fewer true negatives was ideal for this type of dataset which consisted of primarily positive connections ( S7 Fig ) . Nested cross validation of the model demonstrated that the sensitivity and specificity of the model could vary but the overall model performance ( AUC ) remained more consistent ( S8 Fig ) . This suggested that our model would perform with a similar overall accuracy despite changes in sensitivity/specificity trade-offs . The most important predictor in the model was amino acid similarity between genes , followed by nucleotide similarity of whole genomes ( Fig 1B ) . Protein family interactions were moderately important to the model , and CRISPRs were largely uninformative , due to the minimal amount of identifiable CRISPRs in the dataset and their redundancy with the nucleotide similarity methods ( Fig 1B ) . Approximately one third of the training set relationships yielded no score and therefore were unable to be assigned an interaction prediction ( Fig 1C ) . We used our random forest model to classify the relationships between bacteria and phage operational genomic units , which were then used to build the interactive network . The master network , analogous to the universal microbiome network concept previously described [51] , contained the three studies as sub-networks , which themselves each contained sub-networks for each sample ( S9 Fig ) . Metadata including study , sample ID , disease , and OGU abundance within the community were stored in the master network for parsing in downstream analyses ( S9 Fig ) . The phage and bacteria of the master network demonstrated both narrow and broad ranges of infectious interactions ( S10 Fig ) . Bacterial and phage relative abundance was recorded in each sample for each OGU and the weight of the edge connecting those OGUs was calculated as a function of those relative abundance values . The separate extraction of the phage and bacterial libraries ensured a more accurate measurement of the microbial communities , as has been outlined previously [52 , 53] . The master network was highly connected and contained 38 , 337 infectious relationships among 435 nodes , representing 155 phages and 280 bacteria . Although the network was highly connected , not all relationships were present in all samples . Relationships were weighted by the relative abundances of their associated bacteria and phages . Like the master network , the skin network exhibited a diameter of 4 ( measure of graph size; the greatest number of traversed vertices required between two vertices ) and included 433 ( 154 phages , 279 bacteria , 99 . 5% total ) and 38 , 099 ( 99 . 4% ) of the master network nodes and edges , respectively ( Fig 1E and 1F ) . Additionally , the subnetworks demonstrated narrow ranges of eccentricity across their nodes ( S11 Fig ) . Graph node eccentricity , a measurement to supplement diameter , is the shortest distance of each node to the furthest other node within the graph . The phages and bacteria in the diet and twin sample sets were more sparsely related , with the diet study consisting of 80 ( 32 phages , 48 bacteria ) nodes and 1 , 290 relationships , and the twin study containing 130 ( 29 phages , 101 bacteria ) nodes and 2 , 457 relationships ( Fig 1E and 1F ) . As a validation measure , we identified five ( 1 . 7% ) examples of phage OGUs which contained similar genomic elements to the four previously described , broadly infectious phages isolated from Lake Michigan ( tblastx; e-value < 10−25 ) [54] . Diet is a major environmental factor that influences resource availability and gut microbiome composition and diversity , including bacteria and phages [14 , 55 , 56] . Previous work in isolated culture-based systems has suggested that changes in nutrient availability are associated with altered phage-bacteria network structures [25] , although this has yet to be tested in humans . We therefore hypothesized that a change in diet would also be associated with a change in virome-microbiome network structure in the human gut . We evaluated the diet-associated differences in gut virome-microbiome network structure by quantifying how central each sample’s network was on average . We accomplished this by utilizing two common weighted centrality metrics: degree centrality and closeness centrality . Degree centrality , the simplest centrality metric , was defined as the number of connections each phage made with each bacterium . We supplemented measurements of degree centrality with measurements of closeness centrality . Closeness centrality is a metric of how close each phage or bacterium is to all of the other phages and bacteria in the network . A higher closeness centrality suggests that the effects of genetic information or altered abundance would be more impactful to all other microbes in the system . Because these are weighted metrics , the values are functions of both connectivity as well as community composition . A network with higher average closeness centrality also indicates an overall greater degree of connections , which suggests a greater resilience against network degradation by extinction events [25 , 57] . This is because more highly connected networks are less likely to degrade into multiple smaller networks when bacteria or phages are randomly removed [25 , 57] . We used this information to calculate the average connectedness per sample , which was corrected for the maximum potential degree of connectedness . Unfortunately our dataset was insufficiently powered to make strong conclusions toward this hypothesis , but this is an interesting observation that warrants further investigation . This observation also serves to illustrate the types of questions we can answer with more comprehensive microbiome sampling and integrative analyses . Using our small sample set , we observed that the gut microbiome network structures associated with high-fat diets appeared less connected than those of low-fat diets , although a greater sample size will be required to more properly evaluate this trend ( Fig 2A and 2B ) . Five subjects were available for use , all of which had matching bacteria and virome datasets and samples from 8-10 days following the initiation of their diets . High-fat diets appeared to exhibit reduced degree centrality ( Fig 2A ) , suggesting bacteria in high-fat environments were targeted by fewer phages and that phage tropism was more restricted . High-fat diets also appeared to exhibit decreased closeness centrality ( Fig 2B ) , indicating that bacteria and phages were more distant from other bacteria and phages in the community . This would make genetic transfer and altered abundance of a given phage or bacterium less capable of impacting other bacteria and phages within the network . In addition to diet , we observed a possible trend that obesity influenced network structure . This was done using the three mother samples available from the twin sample set , all of which had matching bacteria and phage samples and confirmed BMI information . The obesity-associated network appeared to have a higher degree centrality ( Fig 2C ) , but less closeness centrality than the healthy-associated networks ( Fig 2D ) . These results suggested that the obesity-associated networks may be less connected . This again comes with the caveat that this is only an opportunistic observation using an existing sample set with too few samples to make more substantial claims . We included this observation as a point of interest , given the data was available . Skin and gut community membership and diversity are highly personal , with people remaining more similar to themselves than to other people over time [13 , 58 , 59] . We therefore hypothesized that this personal conservation extended to microbiome network structure . We addressed this hypothesis by calculating the degree of dissimilarity between each subject’s network , based on phage and bacteria abundance and centrality . We quantified phage and bacteria centrality within each sample graph using the weighted eigenvector centrality metric . This metric defines central phages as those that are highly abundant ( AO as defined in the Methods ) and infect many distinct bacteria which themselves are abundant and infected by many other phages . Similarly , bacterial centrality was defined as those bacteria that were both abundant and connected to numerous phages that were themselves connected to many bacteria . We then calculated the similarity of community networks using the weighted eigenvector centrality of all nodes between all samples . Samples with similar network structures were interpreted as having similar capacities for network robustness and transmitting genetic material . We used this network dissimilarity metric to test whether microbiome network structures were more similar within people than between people over time . We found that gut microbiome network structures clustered by person ( ANOSIM p-value = 0 . 008 , R = 1 , Fig 3A ) . Network dissimilarity within each person over the 8-10 day sampling period was less than the average dissimilarity between that person and others , although this difference was not statistically significant ( p-value = 0 . 125 , Fig 3B ) . Four of the five available subjects were used because one of the subjects was not sampled at the initial time point . The lack of statistical confidence was likely due to the small sample size of this dataset . Although there was evidence for gut network conservation among individuals , we found no evidence for conservation of gut network structures within families . The gut network structures were not more similar within families ( twins and their mothers; intrafamily ) compared to other families ( other twins and mothers; inter-family ) ( p-value = 0 . 547 , Fig 3C ) . In addition to the gut , skin microbiome network structure was conserved within individuals ( p-value < 0 . 001 , Fig 3D ) . This distribution was similar when separated by anatomical sites . Most sites were statistically significantly more conserved within individuals ( S12 Fig ) . As an additional validation measure , we evaluated the tolerance of these findings to inaccuracies in the underlying network . As described above , our model is not perfect and there is likely to be noise from false positive and false negative predictions . We found that additional random noise , both by creating a fully connected graph or randomly reducing the number of edges to 60% of the original , changed the statistical significance values ( p-values ) of our findings but not by enough to change whether they were statistically significant ( p-value < 0 . 05 ) . Therefore the comparisons between groups are resilient to potential noise resulting from model false positive and false negative predictions ( S13 Fig ) . Extensive work has illustrated differences in diversity and composition of the healthy human skin microbiome between anatomical sites , including bacteria , virus , and fungal communities [13 , 42 , 58] . These communities vary by degree of skin moisture , oil , and environmental exposure; features which were defined in the original publication [13] . As viruses are known to influence microbial diversity and community composition , we hypothesized that these differences would still be evident after integrating the bacterial and viral datasets and evaluating their microbe-virus network structure between anatomical sites . To test this , we evaluated the changes in network structure between anatomical sites within the skin dataset . The anatomical sites and their features ( e . g . moisture & occlusion ) were defined in the previous publication through consultation with dermatologists and reference to previous literature [13] . The average centrality of each sample was quantified using the weighted eigenvector centrality metric . Intermittently moist skin sites ( dynamic sites that fluctuate between being moist and dry ) were significantly more connected than the moist and sebaceous environments ( p-value < 0 . 001 , Fig 4A ) . Also , skin sites that were occluded from the environment were less connected than those that were constantly exposed to the environment or only intermittently occluded ( p-value < 0 . 001 , Fig 4B ) . We also confirmed that addition of noise to the underlying network , as described above , altered the values of statistical significance ( p-values ) but not by enough to change whether they were statistically significant ( S14 Fig ) . To supplement this analysis , we compared the network signatures using the centrality dissimilarity approach described above . The dissimilarity between samples was a function of shared relationships , degree of centrality , and bacteria/phage abundance . When using this supplementary approach , we found that network structures significantly clustered by moisture , sebaceous , and intermittently moist status ( Fig 4C and 4E ) . Occluded sites were significantly different from exposed and intermittently occluded sites , but there was no difference between exposed and intermittently occluded sites ( Fig 4D and 4F ) . These findings provide further support that skin microbiome network structure differs significantly between skin sites .
Foundational work has provided a baseline understanding of the human microbiome by characterizing bacterial and viral diversity across the human body [13 , 14 , 41–43 , 60] . Here we integrated the bacterial and viral sequence sets to offer an initial understanding of how phage-bacteria networks differ throughout the human body , so as to provide a baseline for future studies of how and why microbiome networks differ in disease states . We implemented a network-based analytical model to evaluate the basic properties of the human microbiome through bacteria and phage relationships , instead of membership or diversity alone . This approach enabled the application of network theory to provide a new perspective while analyzing bacterial and viral communities simultaneously . We utilized metrics of connectivity to model the extent to which communities of bacteria and phages interact through mechanisms such as horizontal gene transfer , modulated bacterial gene expression , and alterations in abundance . Just as gut microbiome and virome composition and diversity are conserved in individuals [13 , 41 , 42 , 59] , gut and skin microbiome network structures were conserved within individuals over time . Gut network structure was not conserved among family members . These findings suggested that the community properties inferred from microbiome interaction network structures , such as robustness ( meaning a more highly connected network is more “robust” to network degradation because a randomly removed bacteria or phage node is less likely to divide or disintegrate [25 , 57] the overall network ) , the potential for horizontal gene transfer between members , and co-evolution of populations , were person-specific . These properties may be impacted by personal factors ranging from the body’s immune system to external environmental conditions , such as climate and diet . We observed evidence supporting the ability of environmental conditions to shape gut and skin microbiome interaction network structure by observing that diet and skin location were associated with altered network structures . We observed evidence that diet was sufficient to alter gut microbiome network connectivity , although this needs to be interpreted cautiously as a case observation , due to the small sample size . Although the available sample size was small , our findings provide some preliminary evidence that high-fat diets are less connected than low-fat diets and that high-fat diets may therefore lead to less robust communities with a decreased ability for microbes to directly influence one another . We supported this finding with the observation that obesity may have been associated with decreased network connectivity . Together these findings suggest the food we eat may not only impact which microbes colonize our guts , but may also impact their interactions with infecting phages . Further work will be required to characterize these relationships with a larger cohort . In addition to diet , the skin environment also influenced the microbiome interaction network structure . Network structure differed between environmentally exposed and occluded skin sites . The sites under greater environmental fluctuation and exposure ( the exposed and intermittently exposed sites ) were more connected and therefore were predicted to have a higher resilience against network degradation when random nodes are removed from the network . Likewise , intermittently moist sites demonstrated higher connectedness than the moist and sebaceous sites . These findings agree with previous work that has shown that bacterial community networks differ by skin environment types [51] . Together these data suggested that body sites under greater degrees of fluctuation harbored more highly connected microbiomes that are potentially more robust to network disruption by extinction events . This points to a link between microbiome and environmental robustness toward network homeostasis and warrants further investigation . While these findings take us an important step closer to understanding the microbiome through interspecies relationships , there are caveats and considerations to our findings . First , as with most classification models , the infection classification model developed and applied is only as good as its training set—in this case , the collection of experimentally-verified positive and negative infection data . Large-scale experimental screens for phage and bacteria infectious interactions that report high-confidence negative interactions ( i . e . , no infection ) are desperately needed , as they would provide more robust model training and improved model performance . Furthermore , just as we have improved on previous modeling efforts , we expect that new and creative scoring metrics will improve future performance . Other creative and high performing models are currently being developed and the applications of these models to community network creation will continue to move this field forward [38–40] . Second , although our analyses utilized the best datasets currently available for our study , this work was done retrospectively and relied on existing data up to seven years old . These archived datasets were limited by the technology and costs of the time . For example , the diet and twin studies , relied on multiple displacement amplification ( MDA ) in their library preparations–an approach used to overcome the large nucleic acids requirements typical of older sequencing library generation protocols . It is now known that MDA results in biases in microbial community composition [61] , as well as toward ssDNA viral genomes [62 , 63] , thus rendering the resulting microbial and viral metagenomes largely non-quantitative . Future work that employs larger sequence datasets and that avoids the use of bias-inducing amplification steps will build on and validate our findings , as well as inform the design and interpretation of further studies . Although our models demonstrated satisfactory accuracy and overall performance , it was important to interpret our findings under the realization that our model was not perfect . This caveat is not new to the microbiome field , with a notable example being the use of 16S rRNA sequencing using the V4 variable region [53] . Use of the V4 variable region excluded detection of major skin bacterial members , meaning that the findings were not able to completely describe the underlying biological environment . Despite this caveat , skin microbiome studies provided valuable biological insights by focusing on the community differences between groups ( e . g . disease and healthy ) which were both analyzed the same way . Similarly , here we focused our conclusions on the differences between the groups which were all treated the same , so that we can minimize our dependence on a perfect predictive model . We also provided explicit evidence that the introduction of noise equally to the compared groups did not significantly impact our findings . Third , the networks in this study were built using operational genomic units ( OGUs ) , which represented groups of highly similar bacteria or phage genomes or clustered genome fragments . Similar clustering definition and validation methods , both computational and experimental , have been implemented in other metagenomic sequencing studies , as well [37 , 64–66] . These approaches could offer yet another level of sophistication to our network-based analyses . While this operationally defined clustering approach allows us to study whole community networks , our ability to make conclusions about interactions among specific phage or bacterial species or populations is inherently limited , compared to more focused , culture-based studies such as the work by Malki et al [54] . Future work must address this limitation , e . g . , through improved binning methods and deeper metagenomic shotgun sequencing , but most importantly through an improved conceptual framing of what defines ecologically and evolutionarily cohesive units for both phage and bacteria [67] . Defining operational genomic units and their taxonomic underpinnings ( e . g . , whether OGU clusters represent genera or species ) is an active area of work critical to the utility of this approach . As a first step , phylogenomic analyses have been performed to cluster cyanophage isolate genomes into informative groups using shared gene content , average nucleotide identity of shared genes , and pairwise differences between genomes [68] . Such population-genetic assessment of phage evolution , coupled with the ecological implications of genome heterogeneity , will inform how to define nodes in future iterations of the ecological network developed here . Even though we are hesitant to speculate on phage host ranges at low taxonomic levels in our dataset , the data does agree with previous reports of instances of broad phage host range [54 , 69] . Additionally , visualization of our dataset interactions using the heat map approach previously used in other host range studies , suggests a trend toward modular and nested tropism , but we do not have the strain-level resolution that powered those previous experimental studies . Finally , it is important to note that our model was built using available full genomes with known interactions , while the experimental datasets resulted in OGUs created from metagenomic shotgun sequence sets , as described above . While this is an informative approach given available data , it is not ideal . We envision future work focusing on training models using metagenomic shotgun sample sets from “mock communities” of bacteria and phages with experimentally validated infectious relationships . This would also be more informative than relying on simulated metagenomic sample sets , whose use would result in models built on simulations and more assumptions instead of empirical data . Together this way the training set can be subjected to the same pre-processing , contig assembly , and OGU binning processes as the experimental data . Furthermore , exciting advances in long read sequencing platforms such as the Oxford Nanopore MinIon system will provide more accurate genomic scaffolds than de novo assembled contigs , allowing for more accurate training and predictions of our models . As discussed above , it is because our current model is susceptible to this noise that we focus our conclusions on comparisons between experimental groups that were both treated the same . This is also why it was important for us to evaluate the susceptibility of our results to noise caused by the less-than-perfect prediction model . Together our work takes an initial step towards defining bacteria-virus interaction profiles as a characteristic of human-associated microbial communities . This approach revealed the impacts that different human environments ( e . g . , the skin and gut ) can have on microbiome connectivity . By focusing on relationships between bacterial and viral communities , they are studied as the interacting cohorts they are , rather than as independent entities . While our developed bacteria-phage interaction framework is a novel conceptual advance , the microbiome also consists of archaea and small eukaryotes , including fungi and Demodex mites [1 , 70]—all of which can interact with human immune cells and other non-microbial community members [71] . Future work will build from our approach and include these additional community members and their diverse interactions and relationships ( e . g . , beyond phage-bacteria ) . This will result in a more robust network and a more holistic understanding of the evolutionary and ecological processes that drive the assembly and function of the human-associated microbiome .
A reproducible version of this manuscript written in R markdown and all of the code used to obtain and process the sequencing data is available at the following GitHub repository: https://github . com/SchlossLab/Hannigan_ConjunctisViribus_ploscompbio_2018 . Raw sequencing data and associated metadata were acquired from the NCBI sequence read archive ( SRA ) . Supplementary metadata were acquired from the same SRA repositories and their associated manuscripts . The gut virome diet study ( SRA: SRP002424 ) , twin virome studies ( SRA: SRP002523; SRP000319 ) , and skin virome study ( SRA: SRP049645 ) were downloaded as . sra files . For clarity , the sample sizes used for each study subset were described with the data in the results section . Sequencing files were converted to fastq format using the fastq-dump tool of the NCBI SRA Toolkit ( v2 . 2 . 0 ) . Sequences were quality trimmed using the Fastx toolkit ( v0 . 0 . 14 ) to exclude bases with quality scores below 33 and shorter than 75 bp [72] . Paired end reads were filtered to exclude sequences missing their corresponding pair using the get_trimmed_pairs . py script available in the source code . Contigs were assembled using the Megahit assembly program ( v1 . 0 . 6 ) [73] . A minimum contig length of 1 kb was used . Iterative k-mer stepping began at a minimum length of 21 and progressed by 20 until 101 . All other default parameters were used . Contig simulations were performed by randomly extracting a string of genomic nucleotides that represented a defined percent length of that genome . This was accomplished using our RandomContigGenerator . pl , which was published in the associated GitHub repository . Contigs were concatenated into two master files prior to alignment , one for bacterial contigs and one for phage contigs . Sample sequences were aligned to phage or bacterial contigs using the Bowtie2 global aligner ( v2 . 2 . 1 ) [74] . We defined a mismatch threshold of 1 bp and seed length of 25 bp . Sequence abundance was calculated from the Bowtie2 output using the calculate_abundance_from_sam . pl script available in the source code . Contigs often represent large fragments of genomes . In order to reduce redundancy and the resulting artificially inflated genomic richness within our dataset , it was important to bin contigs into operational units based on their similarity . This approach is conceptually similar to the clustering of related 16S rRNA sequences into operational taxonomic units ( OTUs ) , although here we are clustering contigs into operational genomic units ( OGUs ) [60] . Contigs were clustered using the CONCOCT algorithm ( v0 . 4 . 0 ) [75] . Because of our large dataset and limits in computational efficiency , we randomly subsampled the dataset to include 25% of all samples , and used these to inform contig abundance within the CONCOCT algorithm . CONCOCT was used with a maximum of 500 clusters , a k-mer length of four , a length threshold of 1 kb , 25 iterations , and exclusion of the total coverage variable . OGU abundance ( AO ) was obtained as the sum of the abundance of each contig ( Aj ) associated with that OGU . The abundance values were length corrected such that: A O = 10 7 ∑ j = 1 k A j ∑ j = 1 k L j Where L is the length of each contig j within the OGU . To confirm a lack of phage sequences in the bacterial OGU dataset , we performed blast nucleotide alignment of the bacterial OGU representative sequences using an e-value < 10−25 , which was stricter than the 10−10 threshold used in the random forest model below , against all of the phage reference genomes available in the EMBL database . We used a stricter threshold because we know there are genomic similarities between bacteria and phage OGUs from the interactive model , but we were interested in contigs with high enough similarity to references that they may indeed be from phages . We also performed the converse analysis of aligning phage OGU representative sequences to EMBL bacterial reference genomes . We ran both the phage and bacteria OGU representative sequences through the Virsorter program ( 1 . 0 . 3 ) to identify phages ( all default parameters were used ) , using only those in the high confidence identification category “class 1” [44] . Finally , we filtered out phage OGUs that had bacterial elements as described above , but also lacked known phage elements by using the tblastx algorithm and a maximum e-value of 10−25 . Open reading frames ( ORFs ) were identified using the Prodigal program ( V2 . 6 . 2 ) with the meta mode parameter and default settings [76] . The classification model for predicting interactions was built using experimentally validated bacteria-phage infections or validated lack of infections from six studies [36 , 45–49] . No further reference databases were used in our alignment procedures . Associated reference genomes were downloaded from the European Bioinformatics Institute ( see details in source code ) . The model was created based on the four metrics listed below . The four scores were used as parameters in a random forest model to classify bacteria and bacteriophage pairs as either having infectious interactions or not . The classification model was built using the Caret R package ( v6 . 0 . 73 ) [77] . The model was trained using five-fold cross validation with ten repeats , and the median model performance was evaluated by training the model on 80% of the dataset and testing performance on the remaining 20% . Pairs without scores were classified as not interacting . The model was optimized using the ROC value . The resulting model performance was plotted using the plotROC R package . The performance of our model for identifying diverse infectious relationships between bacteria and phages , beyond those that were included in the model creation step , were validated using additional bacterial and phage reference genomes , which could be linked by the records of which phage strains were isolated on which bacteria under laboratory conditions . Viral and bacterial reference genomes were downloaded from the GenBank repository on February 19 , 2018 using the viral location ftp://ftp . ncbi . nih . gov/refseq/release/viral/ and the bacterial location ftp://ftp . ncbi . nih . gov/refseq/release/bacteria/ . This resulted in the use of 539 complete phages reference genomes ( with identified hosts ) and 3 , 469 bacterial reference genomes . We used the same prediction model to predict which phages were infecting which hosts , so as to confirm that the model was capable of identifying interactions in a more diverse dataset . Bacteria interactions were identified at the species level . The random contig iteration analysis was performed using a subset of bacterial reference genomes , for computational performance reasons . Only single representative genomes for each species were used . The bacteria and phage operational genomic units ( OGUs ) were scored using the same approach as outlined above . The infectious pairings between bacteria and phage OGUs were classified using the random forest model described above . The predicted infectious pairings and all associated metadata were used to populate a graph database using Neo4j graph database software ( v2 . 3 . 1 ) [81] . This network was used for downstream community analysis . Tolerance to false negative and false positive noise within the networks was assessed by randomly removing a defined fraction of network edges before re-running the downstream analysis work flows . This was accomplished using functionality within the igraph R package ( v1 . 0 . 1 ) [82] . We quantified the centrality of graph vertices using three different metrics , each of which provided different information graph structure . When calculating these values , let G ( V , E ) be an undirected , unweighted graph with |V| = n nodes and |E| = m edges . Also , let A be its corresponding adjacency matrix with entries aij = 1 if nodes Vi and Vj are connected via an edge , and aij = 0 otherwise . Briefly , the closeness centrality of node Vi is calculated taking the inverse of the average length of the shortest paths ( d ) between nodes Vi and all the other nodes Vj . Mathematically , the closeness centrality of node Vi is given as: C C ( V i ) = ( ∑ j = 1 n d ( V i , V j ) ) - 1 The distance between nodes ( d ) was calculated as the shortest number of edges required to be traversed to move from one node to another . Intuitively , the degree centrality of node Vi is defined as the number of edges that are incident to that node: C D ( V i ) = ∑ j = 1 n a i j where aij is the ijth entry in the adjacency matrix A . The eigenvector centrality of node Vi is defined as the ith value in the first eigenvector of the associated adjacency matrix A . Conceptually , this function results in a centrality value that reflects the connections of the vertex , as well as the centrality of its neighboring vertices . The centralization metric was used to assess the average centrality of each sample graph G . Centralization was calculated by taking the sum of each vertex Vi’s centrality from the graph maximum centrality Cw , such that: C ( G ) = ∑ i = 1 n C w - c ( V i ) T The values were corrected for uneven graph sizes by dividing the centralization score by the maximum theoretical centralization ( T ) for a graph with the same number of vertices . Degree and closeness centrality were calculated using the associated functions within the igraph R package ( v1 . 0 . 1 ) [82] . We assessed similarity between graphs by evaluating the shared centrality of their vertices , as has been done previously . More specifically , we calculated the dissimilarity between graphs Gi and Gj using the Bray-Curtis dissimilarity metric and eigenvector centrality values such that: B ( G i , G j ) = 1 - 2 C i j C i + C j Where Cij is the sum of the lesser centrality values for those vertices shared between graphs , and Ci and Cj are the total number of vertices found in each graph . This allows us to calculate the dissimilarity between graphs based on the shared centrality values between the two graphs . Differences in intrapersonal and interpersonal network structure diversity , based on multivariate data , were calculated using an analysis of similarity ( ANOSIM ) . Statistical significance of univariate Eigenvector centrality differences were calculated using a paired Wilcoxon test . Statistical significance of differences in univariate eigenvector centrality measurements of skin virome-microbiome networks were calculated using a pairwise Wilcoxon test , corrected for multiple hypothesis tests using the Holm correction method . Multivariate eigenvector centrality was measured as the mean differences between cluster centroids , with statistical significance measured using an ANOVA and post hoc Tukey test . | The human microbiome , the collection of microbial communities that colonize the human body , is a crucial component to health and disease . Two major components of the human microbiome are the bacterial and viral communities . These communities have primarily been studied separately using metrics of community composition and diversity . These approaches have failed to capture the complex dynamics of interacting bacteria and phage communities , which frequently share genetic information and work together to maintain ecosystem homestatsis ( e . g . kill-the-winner dynamics ) . Removal of bacteria or phage can disrupt or even collapse those ecosystems . Relationship-based network approaches allow us to capture this interaction information . Using this network-based approach with three independent human cohorts , we were able to present an initial understanding of how phage-bacteria networks differ throughout the human body , so as to provide a baseline for future studies of how and why microbiome networks differ in disease states . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"microbiome",
"bacteriophages",
"microbiology",
"diet",
"viruses",
"nutrition",
"network",
"analysis",
"bacteria",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"viral",
"genomics",
"c... | 2018 | Biogeography and environmental conditions shape bacteriophage-bacteria networks across the human microbiome |
Following respiratory syncytial virus infection of adult CB6F1 hybrid mice , a predictable CD8+ T cell epitope hierarchy is established with a strongly dominant response to a Kd-restricted peptide ( SYIGSINNI ) from the M2 protein . The response to KdM282-90 is ∼5-fold higher than the response to a subdominant epitope from the M protein ( NAITNAKII , DbM187-195 ) . After infection of neonatal mice , a distinctly different epitope hierarchy emerges with codominant responses to KdM282-90 and DbM187-195 . Adoptive transfer of naïve CD8+ T cells from adults into congenic neonates prior to infection indicates that intrinsic CD8+ T cell factors contribute to age-related differences in hierarchy . Epitope-specific precursor frequency differs between adults and neonates and influences , but does not predict the hierarchy following infection . Additionally , dominance of KdM282-90 –specific cells does not correlate with TdT activity . Epitope-specific Vβ repertoire usage is more restricted and functional avidity is lower in neonatal mice . The neonatal pattern of codominance changes after infection at 10 days of age , and rapidly shifts to the adult pattern of extreme KdM282- 90 -dominance . Thus , the functional properties of T cells are selectively modified by developmental factors in an epitope-specific and age-dependent manner .
Infants are uniquely affected by respiratory syncytial virus ( RSV ) , the leading viral pathogen of the lower respiratory tract in this age group worldwide [1] . In adults RSV is primarily an under-recognized cause of upper respiratory tract illness [2] . RSV causes yearly winter epidemics with most children becoming infected during their first RSV season . Ninety percent of infants are infected by 2 years of age , with the incidence of severe disease peaking between 6 weeks and 6 months [3] . RSV also causes significant morbidity in children and is the number one cause of hospitalization in those under the age of 12 months [4] . Young infants experience increased vulnerability to infectious agents , particularly viral pathogens , suggesting that their T cell-mediated immune responses are different from those in adults [5] , [6] . Viruses such as RSV and influenza that cause acute infections in immunocompetent adults often result in protracted illnesses in neonates [2] , [7] . These infections ultimately resolve , suggesting delayed viral clearance rather than a persistent inability to clear the offending pathogen . CD8+ T cells play an important role in immunity against viruses through cytokine production and the killing of infected target cells [8] , [9] Activation of naïve T lymphocytes and the expression of effector activity by activated CD8+ T cells requires engagement of the T cell receptor ( TCR ) by viral epitopes presented on MHC class I complexes displayed by antigen-presenting cells ( APCs ) [10] . Although viral pathogens encode thousands of potentially immunogenic determinants , CD8+ T cell responses are usually targeted against a few viral epitopes . These targeted epitopes generally conform to a hierarchy , with one or two dominant epitopes and several subdominant epitopes [11] . Depending on the nature of the stimulus , however , engagement of the TCR and accessory signaling can result in a variety of outcomes for responding T cells that range from full activation and differentiation through to aborted activation and anergy . The differences in T cell-mediated immunity between neonates and adults are not well understood . The limited number of human studies performed to date suggest that the neonatal T-cell compartment may be immature [12] and that neonatal CD8+ T cells may require additional stimuli [13] . Murine models have provided further insights into the differences between neonatal and adult CD8+ T cell populations . Studies have shown that neonatal mice have 1-2 log10 fewer T cells than adult mice [14] , [15] . Furthermore , neonatal TCR chains have a lower frequency of N-nucleotide additions due to a deficiency of terminal deoxynucleotidyl transferase ( TdT ) until approximately 4 days after birth [16] . Thus , neonatal TCRβ CDR3 loops are shorter by an average of one amino acid [17] , [18] . However neonatal mice develop a diverse T cell repertoire in the thymus and periphery that approximates the same order of magnitude compared to adults [14] . In addition , despite differences that would suggest a deficient cytolytic response , a more robust “adult-like”CD8+ T cell response can be generated in certain circumstances , including DNA vaccination and other vaccinations that involve Toll-like receptor ( TLR ) ligands [19] , [20] , [21] . Here , we elucidate key differences between adult and neonatal CD8+ T cell responses using a murine model of RSV infection . Responses in adult hybrid CB6F1 mice [22] were compared to responses in neonatal CB6F1 mice with respect to the hierarchy , function , phenotype and clonotypic composition of epitope-specific CD8+ T cell populations . Of the two primary CD8+ T cell populations that typically respond to RSV infection , we find that the KdM282-90-dominant CD8+ T cell epitope hierarchy found in adult CB6F1 mice is absent in neonates; instead , the KdM282-90 and DbM187-195 epitope-specific CD8+ T cell responses are numerically codominant . These patterns are maintained throughout primary infection , and sustained during the memory phase . Neonatal CD8+ T cells appear highly functional following stimulation with saturating concentrations of peptide , but have lower functional avidities compared to the corresponding adult cells . The neonatal and adult epitope-specific TCRβ repertoires also differ , but the observed differences in epitope hierarchy are not due to limited TdT activity . We performed the first reported naïve precursor frequency analysis in neonates and found a lower frequency of KdM282-90-specific cells , which does not predict the epitope hierarchy established following infection . Rather , a distinct transition time point governs the switch to an adult-like response pattern . Infection prior to day of life 10 results in a codominant response , but infection at or after day of life 10 results in a KdM282-90-skewed , adult-like response . Critically , after transfer of naïve adult CD8+ T cells into congenic neonates , the transferred cells respond with the expected adult-like hierarchy while the neonatal host cells display the typical codominant neonatal pattern . These data indicate that factors intrinsic to CD8+ T cell maturation and development can determine the qualitative nature of the effector response and represent the first example of a differential age-dependent epitope-specific CD8+ T cell response hierarchy to the same viral pathogen . Collectively , these findings may have important implications for understanding the pathogenesis of infectious diseases and immunity in neonates .
Viral titer kinetics were performed after infection of both neonatal ( 7 days old ) and adult ( 6-8 weeks old ) CB6F1 mice with 2×106 plaque forming units ( PFU ) of RSV A2 under isoflurane anesthesia . Titers in the lung were measured at days 1-8 post-infection . Adult mice experienced an eclipse phase early in infection , followed by strong replication in the lung between days 2 and 4 post-infection . Viral load peaked above 5 logs , and dropped rapidly after day 6 post-infection ( Figure 1A ) . Neonatal mice infected with the same dose of virus did not clear the virus as effectively in the early phase , and had higher titers at days 1 and 2 post-infection compared to adults . Replication in neonatal mice was more modest , and peaked at about one log lower compared to replication in adults , after which a gradual tapering of viral titers was observed ( Figure 1A ) . To see if viral titers depended on the challenge dose , both adult and neonatal mice were infected with 1×107 PFU , and viral titers were measured at day 4 ( peak viral titer ) . Adults infected with a higher dose had a one log increase in viral titer compared to those infected with a lower dose ( Figure 1B ) . However , neonatal mice infected with a high-titer dose did not achieve a viral titer above 4 logs ( Figure 1B ) . Thus , both low dose and high dose infections demonstrated that neonatal CB6F1 mice are less permissive for replication than adults of the same strain . Next , we analyzed CD8+ T cell responses specific for two major viral epitopes in both neonatal and adult CB6F1 mice infected with 2×106 PFU of RSV . We have previously shown that the KdM282-90 epitope is immunodominant when compared to the DbM187-195 epitope in infected adult CB6F1 mice [22] . Adult RSV infected mice showed a strongly dominant KdM282-90 response in the lung throughout the course of primary infection ( between days 5 and 14 , Figure 2A ) . Mice infected at 7 days of age exhibited lower overall CD8+ T cell responses compared to adults . Surprisingly , in contrast to adult mice , neonates showed a similar ( codominant ) response to both the KdM282-90 and the DbM187-195 epitopes at all time points following infection ( Figure 2B ) . To further evaluate epitope dominance , the KdM282-90 response of each individual mouse was divided by the DbM187-195 response to yield a response ratio of the two epitopes . The KdM282-90-dominant response in adult mice resulted in ratios of 4-6 after day 7 post infection . Codominant neonatal responses resulted in ratios between 0 . 5 and 1 . 5 throughout the course of infection ( Figure 2C ) . Differences in the response to acute infection were maintained during the memory phase . Analysis of epitope-specific memory CD8+ T cells 72 days after infection showed a KdM282-90-skewed memory response in mice infected as adults ( average response ratio = 2 . 72 ) , and a codominant response in mice that were infected neonatally ( average response ratio = 1 . 08 ) ( Figure 2D ) . Intracellular cytokine staining ( ICS ) was performed to evaluate the function of epitope-specific CD8+ T cells from neonatal and adult mice at day 7 after RSV-infection , which represents the peak of the CD8+ T cell response . Tetramer responses were evaluated in parallel with samples that were stimulated with 1×10−6M of M282-90 , M187-195 or flu peptide controls as described in the Materials and Methods . The percentages of tetramer-binding , epitope-specific cells from infected adults and neonates were compared to the percentages of cells that produced cytokine following specific peptide stimulation . Cells were evaluated for the production of IFN-γ , TNF-α and IL-2 . However , almost all CD8+ T cells produced either IFN-γ alone , or IFN-γ and TNF-α together upon specific stimulation in both neonates and adults . As we have described previously , the adult response to the KdM282-90 epitope was less functional than the response to the DbM187-195 epitope with regards to cytokine production ( i . e . a lower percentage of cells are capable of producing cytokine ) [22] ( Figure 3A ) . In contrast , neonatal cells were highly functional upon peptide stimulation , and almost all cells specific for both the KdM282-90 and DbM187-195 produced IFN-γ , or IFN-γ and TNF-α following peptide stimulation ( Figure 3A ) . The response of CD8+ T cells to stimulation with saturating concentrations of peptide may not reflect either the avidity of the responding populations or the functional response to physiological concentrations of peptide in vivo . We therefore assessed the functional avidities of epitope-specific CD8+ T cell populations from both the lungs and spleens of adult and neonatal mice at day 7 post-infection by measuring IFN-γ production following stimulation with peptide concentrations ranging between 1×10−6 and 1×10−12 M . The percentages of cells producing IFN-γ upon stimulation with each dose of peptide were normalized to the percentages of cells producing cytokine at a saturating peptide concentration ( 10−6M ) for each epitope , and log transformation and nonlinear fits were performed using Graphpad PRISM . The functional avidity of each epitope-specific CD8+T cell population was then measured as a half maximal response ( concentration of peptide necessary for 50% of the cells to produce IFN-γ ) . In the lung , adult DbM187-195-specific cells responded to lower concentrations of cognate peptide ( LogEC50 = -9 . 6 ) than KdM282-90-specific cells ( LogEC50 = -8 . 8 ) , indicating that the subdominant DbM187-195 response has better functional avidity at the site of infection ( Figure 3B ) . Neonatal responses in the lung demonstrated a similar difference of about one log between the codominant DbM187-195 and KdM282-90 responses ( LogEC50 of -9 . 4 and -8 . 5 , respectively ) , and were overall slightly lower ( <0 . 4 logs ) compared to the corresponding functional avidities observed in adults ( Figure 3B ) . A different picture emerged in the spleen , where adult DbM187-195 and KdM282-90-specific CD8+ T cells each responded to considerably lower peptide concentrations than the corresponding cells from the adult lung ( LogEC50 of -10 . 5 and -10 . 2 , respectively ) , while neonatal DbM187-195 and KdM282-90-specific cells were slightly less responsive to peptide ( LogEC50 of -9 . 6 and -8 . 7 , respectively ) ( Figure 3C ) . The resulting difference between CD8+ T cell responsiveness in adults and neonates was much larger than that observed in the lungs . In both adult and neonatal mice , and in both the lung and the spleen , DbM187-195-specific CD8+ T cells exhibited higher functional avidities than KdM282-90-specific CD8+ T cells . RSV has several described immunomodulatory strategies , most attributed to the G protein , which may impact CD8+ T cell epitope hierarchy . To evaluate whether the differences in epitope dominance hierarchy between neonatal and adult mice was RSV-specific , the M and M2 antigens from RSV were expressed in the form of a fusion protein from a replication-incompetent rAd5 vector ( see Materials and Methods ) . In this way , both epitopes were equally available and processed from the same protein . Adults and neonates were infected intranasally with 5×107 FFU , and KdM282-90 and DbM187-195 CD8+ T cell responses were measured at day 7 after infection . Consistent with the patterns observed in RSV infection , adult mice displayed a strongly KdM282-90-skewed response to rAd-MM2 ( Figure 4A ) , with a response ratio of 6 ( Figure 4B ) . Neonates generated a codominant CD8+ T cell response to rAd-MM2 infection with a response ratio of approximately 1 ( Figures 4A and B , p<0 . 0001 ) . Similar results were observed using a Semliki Forest virus vector to deliver the MM2 fusion protein ( data not shown ) . Thus , the observed differences in epitope hierarchy do not result from immunomanipulation by RSV , but are an intrinsic feature of the age-associated CD8+ T cell response . Next , we evaluated the naïve repertoire of the CD8+T cell compartment in adult and neonatal mice by staining splenic lymphocytes with antibodies to CD3 and CD8 , and a panel of TCR Vβ specific antibodies . Overall , we observed a similar T cell receptor Vβ ( TRBV ) expression profile in adults and neonates , with some small differences in minor populations ( Figure 5A ) . TRBV13-2/13-3 ( IMGT nomenclature , www . imgt . org ) ( TCR Vβ 8 . 2/8 . 1 ) was the most dominant variable chain expressed in both adult and neonatal naïve CB6F1 mice , followed by TRBV13-1 ( TCR Vβ 8 . 3 ) and TRBV19 ( TCR Vβ 6 ) . We then conducted a similar evaluation of epitope-specific CD8+ T cells by costaining with tetramer and the panel of Vβ-specific antibodies at day 7 after RSV infection . The KdM282-90 response was comprised almost exclusively of cells using TRBV13-2/13-3 in both infected adults and neonates . The response to DbM187-195 was also fairly restricted , with predominant usage of TRBV17 ( TCR Vβ 9 ) in both the adult and neonatal CD8+ T cell populations . However , adult mice showed some usage of TRBV13-2/13-3 , which was not observed in the neonatal DbM187-195-specific response ( Figure 5B ) . Thus , by this crude analysis , TRBV usage within RSV-specific CD8+ T cell populations was largely similar in neonates and adults . To extend our characterization of the RSV-specific CD8+ T cell repertoire , we sequenced TCRs from single tetramer-positive cells sorted by flow cytometry from two infected adult and neonatal mice [23] . Overall , the patterns of TRBV usage observed with antibody staining were mirrored in the single cell sequencing results , with the KdM282-90 response consisting primarily of cells using TRBV13-2 , and the DbM187-195-specific response comprising mainly TRBV17-positive cells ( Figure 6A ) . Again , the major difference between adults and neonates pertained to the DbM187-195 response , which contained a minority of clonotypes with more diverse TRBV usage in adults ( Figure 6A ) . Single-cell TRBV sequences were also analyzed at the CDR3 level The KdM282-90-sorted cells showed considerable diversity across CDR3 amino acid sequences in both adults and neonates , with only a couple of common sequences between the groups ( Supplemental Figure S1 ) . The DbM187-195-sorted cells also showed relatively few common CDR3β sequences between adults and neonates ( Supplemental Figure S2 ) ; in addition , there was considerably less diversity within the neonatal response , with more than 50% of the sequences obtained having the same CDR3β sequence ( CASSDWGGAEQFF ) . In further analyses , we generated consensus sequences for the CDR3β regions ( http://weblogo . berkeley . edu/ ) . Despite considerable diversity , the consensus CDR3β sequences for the responses to the KdM282-90 epitope were similar between adults and neonates and featured the highly predominant usage of central glycine residues ( Figure 6B ) . Similarly , CDR3β motifs between adult and neonatal DbM187-195-specific cells were largely similar . Our past work has shown that the adult DbM187-195 response contains a relatively conserved DWG motif [24] , which was even more prominent in the neonatal consensus sequence due the relatively restricted use of a DWG containing CDR3β ( Figure 6B ) . Terminal deoxynucleotidyl Transferase ( TdT ) , the enzyme responsible for all non-template nucleotide additions within the CDR3 region , is not expressed until 4-7 days after birth in neonatal mice . Consistent with lower TdT activity , neonatal epitope-specific CD8+ T cell populations had slightly shorter CDR3β amino acid lengths compared to those of adults ( KdM282-90: 12 . 1 vs . 13 . 1 , DbM187-195: 13 . 2 vs . 13 . 6 respectively; data not shown ) . To determine if the epitope dominance disparity between adults and neonates is due to low or limited TdT expression in early life , we infected adult wild-type and TdT knockout ( TdT-/- ) CB6F1 mice . Overall , epitope-specific responses were lower in TdT-/- mice than they were in wild-type mice ( Figure 7A ) . The KdM282-90/DbM187-195 response ratio in TdT-/- mice was significantly higher than in wild-type mice , indicating a more KdM282-90-skewed response ( Figure 7B , p = 0 . 0015 ) . Furthermore , CDR3β sequences obtained from TdT-/- mice were on average 1 and 2 amino acids shorter for the DbM187-195-specific and KdM282-90-specific CD8+ T cell responses , respectively ( data not shown ) . These data indicate that a lack of TdT activity is not responsible for the codominant epitope hierarchy observed following RSV infection of neonatal mice . TCR Vβ usage by tetramer-positive CD8+ T cells from infected wild-type and TdT-/- adult mice was similar for both the KdM282-90-specific and DbM187-195-specific responses ( Figure 7C ) . To determine if the differences in response hierarchy between adult and neonatal mice was due to a difference in epitope-specific CD8+ T cell precursor frequencies , we enumerated naive precursors using a double tetramer enrichment protocol as described in the Materials and Methods . For adult mice , spleens and macroscopic lymph nodes were harvested; for neonatal mice only spleens were harvested , and 8-12 spleens were pooled for each sample . Cells from naïve adult and neonatal CB6F1 mice were processed at the same time as samples from OT-I/RAG1-/- mice ( negative controls ) , and samples from immune/memory CB6F1 mice that were infected as adults at least one month previously . Data were generated with two sets of tetramers for each epitope: one set was purchased from Beckman Coulter , and the second set was produced in-house as described previously [25] . Although the data were consistent within each set of tetramers , some differences were observed between sets . Representative raw data are shown in Supplemental Figures S3 and S4 . Data generated using the in-house tetramers appeared more comparable to published data using the double tetramer method , with positive cells staining equally with each tetramer and forming a clean diagonal; for this reason the in-house tetramers were used for the data presented in Figure 8 . Despite the KdM282-90-skewed response observed following RSV infection , naïve adult CB6F1 mice had similar numbers of precursors specific for the KdM282-90 and DbM187-195 epitopes ( average of 419 vs . 370 cells per mouse , respectively; p = 0 . 54; Figure 8A ) , with a resulting median ratio of 1 . 27 ( Figure 8B ) . In contrast , naïve neonatal CB6F1 had significantly fewer precursors specific for KdM282-90 compared to DbM187-195 ( average of 2 . 2 vs . 5 . 2 cells per neonate , respectively; p = 0 . 0003; Figure 8A ) . The median ratio of KdM282-90/DbM187-195-specific CD8+ T cells in naïve neonates was 0 . 46 ( Figure 8B ) . As expected , OT-I/RAG1-/- had no precursors specific for either KdM282-90 or DbM187-195 . Further , as expected based on our staining results for memory CD8+ T cell responses , mice infected at least one month prior to analysis had significantly higher numbers of precursors specific for KdM282-90 ( average of 8114 cells/mouse ) compared to DbM187-195 ( average of 2811 cells/mouse; Figure 8C ) , with a median response ratio of 3 . 01 ( Figure 8D ) ; these data mirror the tetramer straining results obtained with memory populations in immune mice ( Figure 2D ) . For comparison , data obtained using the commercial tetramers to enrich cells from the same groups of mice are presented in Supplemental Figure S5 . Next , we asked when the epitope hierarchy shifted by infecting mice between the ages of 3 days of life and 13 days of life and measuring CD8+ T cell responses at day 7 post-infection . Mice infected before day of life 10 exhibited codominant CD8+ T cell responses to KdM282-90 and DbM187-195 . Starting most notably in mice infected at 10 days of life , however , the response became increasingly KdM282-90-dominant . This pattern continued through mice infected at day of life 13 into adulthood ( Figure 9A ) and is clearly visualized as an upward trend on the response ratio graph beginning with mice infected at day of life 10 ( Figure 9B ) . Changes in epitope hierarchy in mice infected at days of life 3 through 13 were not associated with viral titer increases ( Figure 9C ) . To assess the contribution of intrinsic CD8+ T cell factors to the establishment of epitope hierarchy following infection , naïve adult CD8+ T cells were adoptively transferred intra-peritoneal into naïve congenic neonatal mice 2–3 days prior to infection with RSV . Spleen and lymph node CD8+ T cells were purified by untouched isolation from two adults for transfer into each neonate; the final number of transferred CD8+ T cells was between 7×106 and 1 . 5×107 per neonate . The CD8+ T cell responses of adult ( transferred ) and neonatal ( endogenous ) cells were evaluated at day 7 post-infection ( Supplemental Figure S6 ) . The endogenous CD8+ T cell response in neonatal mice in the absence of adoptive transfer was codominant for KdM282-90 and DbM187-195 , with an average response ratio of 1 . 3 ( Figure 10A and B ) . In mice that received naïve adult CD8+ T cells , the endogenous response was similarly codominant with an average response ratio of 1 . 2 . However , the adult transferred cells within the same animals generated a strongly KdM282-90 skewed response with an average ratio of 15 . 2; p = 0 . 0031; Figure 10A and B ) similar to that observed following infection of adult mice . These data indicate that intrinsic CD8+ T cell factors play an important role in the establishment of epitope hierarchy following infection .
Respiratory syncytial virus is a significant cause of morbidity and mortality in infancy , with most severe disease occurring before 6 months of age . Currently , prophylactic treatment with the humanized monoclonal antibody Synagis is the only licensed intervention for RSV and is used to prevent hospitalization caused by severe RSV disease . However , passive antibody treatment is expensive , only partially effective , and available only to those at highest risk . There is a great need for a preventive vaccine , or effective therapeutic treatment strategies . To be of most benefit , an RSV vaccine would need to be given early in life , and overcome many obstacles associated with the generation of effective immunity during infancy [26] . Induction of effective immunity to RSV is challenging due to the failure of natural immunity to protect against reinfection . Vaccine safety is also a concern because of the history of vaccine-enhanced disease following natural infection after a formalin-inactivated vaccine given in the early 1960s [27] , [28] . These challenges highlight the critical significance of understanding neonatal adaptive immune responses during both infection and vaccination . Several studies have demonstrated the importance of the age of infection with regard to the immune response to RSV [29] , [30] , [31] , [32] , [33] , [34] , [35] , [36] . Studies in neonatal mice have focused on the influence that neonatal infection has on the immune response to subsequent reinfection with RSV . It has been established that when neonatal BALB/c mice are infected prior to 7 days of age , they experience enhanced disease following reinfection as adults . This disease enhancement has been associated with increased airway hyperresponsiveness , increased airway inflammatory cell recruitment and immunopathology , mucus hyperproduction , eosinophil recruitment , and enhanced Th2 cytokine production [30] , [31] , [32] , [33] , [34] , [35] . Additionally , infecting neonates with recombinant RSV expressing IFN-γ was found to abrogate disease enhancement , while RSV expressing IL-4 resulted in further enhancement of Th2 responses and eosinophilia [37] . Importantly , in all studies , disease enhancement is not observed following reinfection of mice initially infected as adults . These studies in BALB/c mice provide an example of how early infection can modify and shape subsequent responses to reinfection or other airway insults . These studies also recapitulate some facets of human infection , where severe disease early in life has been linked to the development of childhood wheezing [38] and vaccination during infancy caused enhanced disease following natural infection [28] . As in humans , the CD8+ T cell response in RSV-infected mice is known to play a critical role in viral clearance . Using CB6F1 hybrid mice , the expression of both d- and b- MHC haplotypes creates a more complex phenotype and increases the number of T cell responses that can be measured . In BALB/c mice , the KdM282-90 response is so dominant that other responses contribute very little to the outcome of infection . In CB6F1 mice , a reproducible epitope hierarchy is established following RSV infection . The KdM282-90 response initially described in the BALB/c parent strain dominates , with an approximately 5-fold lower response to the DbM187-195 epitope initially described as dominant in the C57BL/6 parent strain [22] , [39] . In striking contrast , mice infected as neonates were found to have codominant responses to KdM282-90 and DbM187-195 throughout the course of primary infection , which were maintained in the memory phase . It is generally accepted that neonatal CD8+ T cell responses are lower than adult responses , and this has been described for the KdM282-90 epitope following neonatal infection of BALB/c [34] . While we also measure lower KdM282-90 responses in CB6F1 neonates , the response to the DbM187-195 epitope is higher in infected neonates than adults . This observation demonstrates that the ability of neonates to generate CD8+ T cell responses is epitope-dependent and in some cases , may be superior to adult T cell responses . We found that the epitope hierarchy makes a radical shift when mice are infected between 3 and 13 days of age , with the emergence of KdM282-90 dominance starting after infection at day of life 9 . The dominance pattern skews dramatically for mice infected between days 9 and 13 and is associated with a lowering of the response to the DbM187-195 epitope and a significant increase in the response to KdM282-90 . It is likely that dampening of the DbM187-195 response following emergence of the dominant KdM282-90 response reflects immunodomination by this epitope , which we have described in detail in adult CB6F1mice [40] . The relatively low KdM282-90 response generated in younger neonates may not reach a magnitude or functional activity sufficient for immunodomination of DbM187-195 . Functionally , neonatal CD8+ T cells are adept at producing effector cytokines following stimulation with saturating concentrations of cognate peptide . They were as proficient , if not more , in the production of IFN-γ and TNF-α as compared to CD8+ T cell responses generated in the adult . This suggestion of “adult-like” function may be deceptive , however , as peptide titrations showed clearly lower functional avidities for CD8+ T cell responses generated in the neonate , which may play a role in their responsiveness in vivo . This difference was particularly apparent in the spleen , where approximately one log more peptide was required to reach a half maximal response in neonatal CD8+ T cells . In both adults and neonates , DbM187-195-specific cells were found to have higher functional avidities than KdM282-90-specific cells , a property that appears to be independent of epitope hierarchy . Factors known to influence CD8+ T cell epitope dominance fall into three main categories . The first is antigen processing and presentation and includes factors involved in peptide liberation , transport , and class I binding affinity . The second involves characteristics inherent to the CD8+ T cell response such as the T cell repertoire and precursor frequency , and the ability of cells to respond to stimulation by activation and proliferation . Finally , regulation of CD8+ T cell responses , either by other CD8+ T cells ( immunodomination ) or by regulatory T cells can play a role in the establishment of epitope hierarchy . Adoptive transfer experiments in which adult naïve CD8+ T cells were transferred into neonates gave us the ability to study the response of both neonatal and adult cells within the same infected host . The results of these experiments heavily suggested that intrinsic CD8+ T cell factors help dictate the dominance patterns we observed following RSV infection . We investigated several CD8+ T cell factors that may be involved , starting with characterization of the TCR Vβ repertoire in naïve and infected mice . Despite the dramatic difference in epitope response hierarchy , neonatal responses were surprisingly similar to responses seen in adults in many ways , particularly when responses were studied at the level of TCR Vβ protein expression . Further analysis by single-cell clonotyping was necessary to identify differences at the CDR3β amino acid level , and showed overall less diversity in the neonatal response , particularly for the DbM187-195 epitope . Despite some sequence and diversity differences , however , the general motifs within KdM282-90 and DbM187-195 CDR3βs were similar between adults and neonates . The impact that these relatively subtle repertoire differences between adults and neonates exert on epitope hierarchy is a subject of further investigation . A lack of TdT activity in early life has been found to be responsible for shaping the murine neonatal repertoire [41] , [42] , and neonatal CD8+ T cell responses to infection can consist of shorter CDR3 sequences than those of adults [43] . We hypothesized that TdT may play a role in epitope hierarchy differences between adults and neonates , and infected adult TdT-/- CB6F1 mice to address this possibility . RSV-infected TdT-/- animals had an overall lower CD8+ T cell response , but the relative dominance of the KdM282-90-specific response was greater than in wild-type mice , indicating that TdT deficiency does not favor the DbM187-195 response or account for codominance in infected neonates . Additionally , the TCR Vβ repertoire within both the KdM282-90 and the DbM187-195 response was similar between wild-type and TdT-/- animals . Precursor frequency is another CD8+ T cell factor that has been found to correlate with epitope hierarchy post-infection [44] , [45] , [46] . Here , we describe the first reported enumeration of precursor frequencies in neonatal mice . Unlike in adults , lymph nodes cannot be acquired from naïve neonatal mice . The availability of only spleen tissue and the relative lymphopenia of neonates necessitated pooling of at least 8 neonates to work with sufficient cell numbers . Technical constraints in both adults and neonates imply that naïve precursor frequencies are an underestimate in each case , but relative comparisons between epitopes offers reproducible and meaningful results . We analyzed each cell population independently using the same tetramer conjugated to two different fluorochromes simultaneously . Consistent and expected results were seen for each set with regard to the negative controls , and the ratio of KdM282-90/DbM187-195 cells seen in immune mice . While neonates had a more DbM187-195-skewed population of precursors than adults , precursor frequency did not predict the final epitope hierarchy post-infection . This interpretation was consistent with data generated with either in-house or commercial tetramers . Adult CB6F1 have as many , if not more DbM187-195-specific precursors , yet generate a severely KdM282-90-skewed response . Similarly , the presence of more precursors specific for DbM187-195 than for KdM282-90 in the neonate does not lead to dominance of the DbM187-195 response . La Gruta et al . , have reported that precursor frequency is unrelated to the immunodominance hierarchy observed in adult mice following infection with influenza A virus , and suggest that subdominance is a consequence of inefficient cell recruitment and clonal expansion [47] . It is likely that immunodominance in the RSV model is influenced by a combination of precursor frequencies and differing abilities of naïve CD8+ T cells to be recruited and proliferate . Figure2A clearly illustrates a large net proliferative advantage for KdM282-90-specific cells over DbM187-195-specific cells between days 5 and 7 post-infection in adult mice . The basis for this difference in net frequency of tetramer-positive adult T cells in the lung is currently being explored along with other factors including antigen processing and presentation , CD8+ T cell regulation , and the influence of MHC I-peptide epitope structure and TCR affinity on the functional T cell hierarchy . In summary , we have described differences in epitope dominance between adult and neonatal CD8+ T cell responses , with associated differences in TCR diversity , functional avidity , and precursor frequency . We show that there are intrinsic properties of adult T cells that result in distinct functional responses in an epitope-dependent manner . The factors that account for the dramatic shift in T cell function between days 9 and 10 of life are the subject of ongoing investigation . This phenomenon is unlikely to be RSV-specific , and we believe that epitope dominance disparities due to age are likely to exist in other mouse models and in humans . A better understanding of the differences between how adult and neonatal responses are generated is of critical importance . The well-known plasticity of the neonatal response can be powerfully manipulated , but immune response patterns in neonates cannot necessarily be predicted by responses in adults . Therefore care must be taken to ensure that optimal effectiveness and safety of vaccine-induced immunity is achieved in neonates .
All mice used in this study and analysis were maintained according to the guidelines of the NIH Guide to the Care and Use of Laboratory Animals and the approval of the Animal Care and Use Committee of the Vaccine Research Center ( VRC ) , National Institute of Allergy and Infectious Diseases at the National Institute of Health . All mice were housed in a facility fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International ( AAALAC ) . All procedures were conducted in strict accordance with all relevant federal and National Institutes of Health guidelines and regulations . BALB/c ( female ) and C57BL/6 ( male ) breeders were purchased from Jackson Labs and pups were obtained by time-mating . Adult ( 8–10 weeks old ) female CB6F1/J mice ( Jackson Labs , Bar Harbor , ME or bred in-house ) were used . TdT-/- mice on BALB/c and C57BL/6 backgrounds were generously provided by Jonathan Yewdell ( NIAID , NIH ) and Ann Feeney ( Scripps ) , respectively . CB6F1 TdT-/- mice were bred in-house . Thy1 . 1+ C57BL/6 mice were obtained from Jackson ( B6 . PL-Thy1a/CyJ ) , Thy1 . 1+ BALB/c mice were a generous gift from Jonathan Yewdell , and Thy1 . 1+ CB6F1 mice were bred in house . OT-I/RAG-1-/- mice were purchased from Taconic and bred in-house . All mice were housed in our animal care facility at NIAID under specific , pathogen-free conditions , and maintained on standard rodent chow and water supplied ad libitum . All studies were reviewed and approved by the NIH Animal Care and Use Committee . Mice were anesthetized using isoflurane ( 3% ) prior to intranasal inoculation with 2×106 PFU live RSV in 10% EMEM ( 100 µl for adults , 25–35 µl for neonates ) . Neonatal mice were infected at day 7 of life unless stated otherwise . All mice were euthanized by lethal injection with pentobarbital ( 250 mg/kg ) . Mice were sacrificed and lung tissue was removed and quick-frozen in 10% EMEM . Thawed tissues were kept chilled while individual samples were ground using a GentleMACS machine ( Miltenyi , Germany ) on program Lung 02 . Samples were centrifuged , and dilutions of clarified supernatant were inoculated on 80% confluent HEp-2 cell monolayers in triplicate and overlaid with 0 . 75% methyl cellulose in 10% EMEM . After incubation for 4 days at 37°C , the monolayers were fixed with 10% buffered formalin and stained with hematoxylin and eosin . Plaques were counted and expressed as log10 PFU/gram of tissue . The limit of detection was 1 . 8 log10 PFU/gram of tissue . A replication-defective adenovirus expressing a fusion protein of M and M2 was constructed using sequences for RSV-A2 ( M , NCBI accession number AAB86660; M2 , NCBI sequence number AAB86677 ) , which were codon optimized for expression in humans and mice using the GeneOptimizer technology from GeneArt . The rAd5-MM2 vector was produced by GenVec ( Gaithersburg , MD ) using complementing mammalian cells ( 293-ORF6 ) as described previously [48] , and 5×107 focus forming units ( FFU ) were given intranasally to adults and neonates under isoflurane anesthesia . RSV M282-90 ( SYIGSINNI ) and RSV M187-195 ( NAITNAKII ) peptides were derived from the RSV M2 and M proteins , respectively . The H2-Kd-binding influenza virus A/Puerto Rico/8/34 nucleoprotein NP147-155 ( TYQRTRALV ) peptide , and H2-Db-binding influenza A/Puerto Rico/8/34 NP366-374 ( ASNENMETM ) peptide were used as negative controls . All peptides were synthesized by Anaspec , Inc . ( San Jose , CA ) , and confirmed to be >95% pure by analytical high-performance liquid chromatography at the NIAID peptide core facility ( Bethesda , MD ) . Mice were sacrificed and lung and/or spleen tissues were harvested at the indicated times post-infection . Lung and spleen tissues were disrupted by tissue dissociation using a GentleMACS machine ( Miltenyi ) . Lymphocytes were purified using Fico- LITE at room temperature , washed , then resuspended in 10% RPMI . For ICS , lymphocytes were incubated at 37°C for 5 hours with 1 µM of the appropriate peptide , 1 µg/ml of co-stimulatory antibodies against CD28 and CD49d , and 1 µg/ml of monensin . After incubation , cells were surface stained with fluorochrome-conjugated antibodies against CD3 ( 145-2C11 ) , CD4 ( GK1 . 5 ) , and CD8 ( 2 . 43 ) then fixed and permeabilized using an intracellular cytokine staining kit according to the manufacturer's instructions ( BD , San Diego , CA ) . Intracellular stains were performed with labeled antibodies to IFN-γ ( XMG1 . 2 ) , IL-2 ( JES6-5H4 ) and TNF-α ( MP6-XT22 ) for 20 min at 4°C . For tetramer analysis , cells were surface stained with KdM282-90 or DbM187-195 tetramer ( Beckman Coulter , San Diego , CA ) together with labeled antibodies specific for CD3 , CD4 , and CD8 . Samples for TCR Vβ usage analysis also included an antibody from the TCR Vβ screening panel ( BD ) . After staining , cells were washed and analyzed by flow cytometry . Samples were collected on an LSR-II flow cytometer ( BD , San Jose , CA ) and data were analyzed using FlowJo version 8 . 8 . 5 ( Tree Star , San Carlos , CA ) . For ICS analysis , Boolean gating was performed after single gating for each cytokine , and background from flu peptide-stimulated control samples was subtracted in Pestle ( software provided by Mario Roederer , Bethesda , MD ) prior to graphing . To assess functional avidity , the same ICS procedure was performed following a 5 hour , 37° incubation of cells from infected lungs or spleens with 1 µg/ml of co-stimulatory antibodies against CD28 and CD49d , 1μg/ml of monensin and serial dilutions of the indicated peptide . Single cell sequencing was performed using multiplex nested RT-PCR of the TCRβ chain from single sorted cells and data were analyzed as described previously [23] . Weblogos were generated on http://weblogo . berkeley . edu/following input of all sequences normalized to the same length by insertion of Xs into the center of shorter CDR3βs . Single cell suspensions were isolated from the spleen and macroscopic lymph nodes ( inguinal , axillary , brachial , cervical , and mesenteric ) of individual mice ( adults ) or from the spleen only of neonatal mice ( pooled spleens from 8–12 mice/sample ) by manual disruption between frosted glass slides . Precursor frequencies were evaluated using an approach similar to that described for assessing naïve CD4+ T cell populations [49] , [50] , with other described modifications to enhance the detection of epitope-specific CD8+ T cells with less background [44] , [45] . Briefly , cells were incubated in MACS buffer ( Miltenyi ) containing 0 . 1% sodium azide , Fc block ( 2 . 4G2 , BD ) , and a labeled antibody to CD8α ( 53-6 . 7 , BD ) . Tetramers of the same specificity labeled separately with phycoerythrin ( PE ) and allophycocyanin ( APC ) were added for one hour at room temperature . Cells were washed and incubated with 50 µL each of anti-PE and anti-APC beads for 30 minutes at 4°C before washing and passing through a magnetized MS column ( Miltenyi ) as directed for positive enrichment of tetramer-specific T cells . Enriched cells were resuspended in 100 µL of MACS buffer and stained with additional surface antibodies against CD3 , CD4 , CD44 ( IM-7 , e-Bioscience ) and B220 ( RA3-6B2 , e-Bioscience ) , CD11b ( M1/70 . 15 , e-Bioscience ) , CD11c ( N418 , e-Bioscience ) , and F4/80 ( BM8 , Caltag-Invitrogen ) antibodies were included in a dump gate that also excluded dead ( ViViD+ ) cells . Samples were acquired in their entirety on an LSR II flow cytometer and analyzed using FlowJo software to determine the number of tetramer double-positive cells acquired from each adult mouse or pool of neonatal mice . All data are presented as number of cells obtained per mouse . Samples from naïve CB6F1 adults , naïve CB6F1 neonates , immune/memory CB6F1 adults ( positive controls ) , and OT-I/RAG1-/- mice ( negative controls ) were run with two sets of tetramers for each epitope: one set made in-house , and the other set obtained from Beckman Coulter . Cells were harvested from the spleens and macroscopic lymph nodes of naïve adult CB6F1 mice by manual disruption between frosted glass slides . Lymphocytes were harvested using Ficoll-LITE , then CD8+ T cells were purified by untouched isolation using a CD8+ T cell isolation kit ( Miltenyi ) . Purity was assessed to be 88-95% by flow cytometry following each sort . Between 7×106 and 1 . 5×107 purified CD8+ T cells ( the amount isolated from two adults for each neonate ) were transferred intra-peritoneally into naïve , congenic neonates at days 4-5 of life , and infection with 2×106 PFU of RSV was performed at day 7 of life . Lungs and spleens were harvested 7 days post infection and cells were isolated and stained as described previously with the addition of antibodies to Thy1 . 1 ( HIS51 , e-Bioscience ) and Thy1 . 2 ( 53-2 . 1 , e-Bioscience ) to discriminate the host response from the response of adoptively transferred cells . Statistical analyses between two groups were conducted using a two-tailed students t test . Comparisons between multiple groups were performed in GraphPad Prism using a 1 way or 2 way ANOVA followed by Bonferroni's post-tests for multiple comparisons between all groups . Supplemental Figures S1 and S2 show CDR3β amino acid sequences derived from single cell sequencing of KdM282-90 or DbM187-195 tetramer-sorted cells , respectively from infected adults and neonates . Supplemental Figures S3 and S4 show raw flow cytometry data from double tetramer sorts using tetramers made by in-house ( Supplemental Figure S3 ) or commercially obtained from Beckman Coulter ( Supplemental Figure S4 ) . Supplemental Figure S5 shows the results of precursor frequency experiments conducted using Beckman Coulter tetramer sets . Supplemental Figure S6 shows the experimental outline , and gating strategy for naïve CD8+ T cell adoptive transfer experiments . | RSV causes yearly winter epidemics of respiratory disease with peak hospitalization rates at 2 . 5 months of age . Clearance of virus-infected cells depends on CD8 T-cells , and defining mechanisms of CD8 T-cell regulation is essential for understanding RSV disease pathogenesis and guiding therapeutic interventions . CD8 T-cells recognize a virus-infected cell by detecting peptides cleaved from viral proteins that are presented in host cell MHC molecules . The strength of CD8 T-cell response to processed peptide epitopes from the virus commonly assumes a predictable response hierarchy . In adult hybrid mice that have both H-2b and H-2d MHC alleles , the majority of the CD8 T-cell response is targeted at a peptide from the M2 protein and presented by the Kd MHC molecule , KdM282-90 , and a smaller subset of CD8 T-cells respond to a peptide from the M protein , DbM187-195 . Interestingly , when infecting neonatal hybrid mice the dominance pattern is not seen and the DbM187-195 response is equal to or greater than the KdM282-90 response . We show that the adult dominance pattern emerges at 10 days of age and that T cells are modified by developmental factors in an epitope-specific and age-dependent manner . These observations may influence future vaccine design . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"immune",
"cells",
"respiratory",
"infections",
"immunity",
"to",
"infections",
"immunology",
"microbiology",
"pulmonology",
"adaptive",
"immunity",
"immune",
"defense",
"immunoregulation",
"lower",
"respiratory",
"tract",
"infections",
"asthma",
"animal",
"mo... | 2011 | Neonatal CD8 T-cell Hierarchy Is Distinct from Adults and Is Influenced by Intrinsic T cell Properties in Respiratory Syncytial Virus Infected Mice |
At least six histone H1 variants exist in somatic mammalian cells that bind to the linker DNA and stabilize the nucleosome particle contributing to higher order chromatin compaction . In addition , H1 seems to be actively involved in the regulation of gene expression . However , it is not well known whether the different variants have distinct roles or if they regulate specific promoters . We have explored this by inducible shRNA-mediated knock-down of each of the H1 variants in a human breast cancer cell line . Rapid inhibition of each H1 variant was not compensated for by changes of expression of other variants . Microarray experiments have shown a different subset of genes to be altered in each H1 knock-down . Interestingly , H1 . 2 depletion caused specific effects such as a cell cycle G1-phase arrest , the repressed expression of a number of cell cycle genes , and decreased global nucleosome spacing . On its side , H1 . 4 depletion caused cell death in T47D cells , providing the first evidence of the essential role of an H1 variant for survival in a human cell type . Thus , specific phenotypes are observed in breast cancer cells depleted of individual histone H1 variants , supporting the theory that distinct roles exist for the linker histone variants .
Eukaryotic DNA is packaged into chromatin through its association with histone proteins . Chromatin is composed of nucleosomes . The nucleosome core particle consists of 146 base pair units wrapped around a histone octamer consisting of two copies each of the core histone proteins H2A , H2B , H3 and H4 . The linker histone H1 sits at the base of the nucleosome near the DNA entry and exit sites and is involved in the folding and stabilization of the 30 nm chromatin fiber [1] , [2] . The amount of H1 per nucleosome is very variable , and the paradigm of one H1 per nucleosome is more the exception than the rule [3] . Histone H1 is a lysine-rich protein with a short basic N-terminal tail , a highly conserved central globular domain and a long positively-charged C-terminal tail . These tails are post-translationally modified , mostly by phosphorylation , but also by acetylation and methylation [4] , [5] . CDK-dependent phosphorylation of H1 occurs progressively throughout the cell cycle , with a maximum during mitosis [6] . Histone H1 in vertebrates is a family of closely related , single-gene encoded proteins , showing much less evolutionary conservation than core histones . In mammals , five somatic subtypes ( from H1 . 1 to H1 . 5 ) , a terminally differentiated expressed isoform ( H1 . 0 ) , two tissue-specific variants ( H1 testis and H1 oocyte ) and a recently described , poorly characterized H1x variant have been identified [7]–[10] . Histone H1 participates in nucleosome positioning or spacing and formation of the higher-order chromatin structure . H1-containing chromatin is more resistant to nuclease digestion and shows strong inhibition of nucleosome sliding [11] . Consequently , H1 is seen as a structural component related to chromatin compaction and inaccessibility to transcription factors or RNA polymerase . Nonetheless , it has been suggested that histone H1 plays a more dynamic and gene-specific role , participating in activation or repression of gene expression . Previous studies on the effect of H1 depletion on global gene expression have reported changes in the expression of small groups of genes , instead of it affecting the vast majority of cellular genes [12]–[16] . Overexpression experiments have also contributed to challenge the concept of H1 as a general repressor of chromatin activity . In Xenopus laevis embryos , over-expression of the somatic H1 variant repressed oocyte- but not somatic-type 5S rRNA genes or other Pol III transcripts [17] . Overexpression of H1 . 0 and H1 . 2 isoforms increase both the basal and hormone-induced levels of the mouse mammary tumor virus ( MMTV ) reporter gene transcript [18] . Gene-specific effects of H1 might result from interactions with specific regulatory factors or DNA-binding proteins . This might also be the origin of reported specific functions for some H1 variants . H1 . 5 cooperates with the transcription factor MSX1 for inhibition of specific target promoters and myogenesis [19] . H1 . 4 is involved in a heterochromatinization process . H1 . 4 at Lys26 is deacetylated by SirT1 and subsequently methylated , facilitating recruitment of Polycomb complexes and HP1 , whereas simultaneous phosphorylation of Ser27 blocks HP1 binding [20]–[23] . Recently , a novel H1 . 2 complex acting as a repressor of p53-mediated transcription has been isolated [24] . In addition , H1 . 2 has been involved in apoptosis induced by DNA double-strand breaks , acting as a cytochrome c-releasing factor that appears in the cytoplasm after X-ray irradiation [25] . A strong decrease in histone H1 has drastic consequences on the life span of some organisms [26]–[28] . However , single or double H1 variant knock-out mice have no apparent phenotype due to the compensatory up-regulation of other subtypes [29] . These reports have favored the thinking that H1 variants are redundant , lacking specific functions in chromatin organization or gene expression control . Knocking out additional subtypes cannot be compensated for fully by up-regulation of the remaining subtypes [30] . Nonetheless , triple null H1 . 2-H1 . 3-H1 . 5 mouse embryonic stem ( ES ) cells with a 50% global reduction in total H1 can be obtained [12] . The electrostatic effect of this reduction is apparently compensated for by a reduction in nucleosome repeat length ( NRL ) , in line with a previously reported strong linear relationship between NRL and the H1/nucleosome ratio ( reviewed in [3] ) , together with other mechanisms such as changes in core histone modifications . In this triple KO ES cell line , a very limited number of genes change their expression . These include imprinted genes regulated by DNA methylation , in which specific CpG regions are downmethylated in the absence of H1 [12] . We have previously described a complex role of the linker histone in the chromatin remodeling events taking place at the MMTV promoter in response to progestin . In vitro , histone H1 enhances transactivation in the simultaneous presence of the progesterone receptor ( PR ) and Nuclear Factor 1 ( NF1 ) . The first step following initial binding of PR to the exposed HREs is phosphorylation of histone H1 . Then , PR recruits an ATP-dependant chromatin remodeling complex , which changes the structure of the nucleosomal core particles making its DNA more accessible for NF1 and additional PR binding . In a last step , H1 leaves the promoter to enable efficient transcription initiation [31] , [32] . In this work , no attempt was made to define the role of individual H1 variants . In order to further characterize in vivo the role of histone H1 variants and , in particular , to ascertain whether there is any functional specificity , we developed an inducible shRNA expressing system for the depletion of individual H1 variants in the human breast cancer cell line T47D . We achieved reliable specific inhibition for each variant , avoiding compensation effects and demonstrated significant differences among H1 variants , with regard to cell cycle progression and gene expression . On the one hand , depletion of H1 . 4 leads to cell death in T47D . This is the first time that it has been noted that one variant is essential for survival in a human cell type . On the other , we describe the involvement of H1 . 2 in cell cycle progression . Its inhibition caused a G1 arrest , defects in chromatin structure and changes in expression of specific genes linked to cell cycle .
H1 knock-down cell lines were established from T47D-MTVL cells ( carrying one stably integrated copy of luciferase reporter gene driven by the MMTV promoter; [33] . These cell lines were grown in RPMI 1640 medium , supplemented with 10% FBS , 2 mM L-glutamine , 100 U/ml penicillin , and 100 µg/ml streptomycin . Doxycycline ( Sigma ) was added at 2 . 5 µg/ml when indicated . Along a 6-day treatment with Dox , cells were passaged at day 3 . When indicated , serum-containing media was replaced with serum-free media at day 4 for growth arrest . Images of H1 KD cell lines grown in normal conditions were taken using an inverted Leica DMR microscope . Other cell lines used were grown as follows: MCF-7 cell line was grown at 37°C with 5% CO2 in MEM medium containing 10% fetal bovine serum , 1% penicillin/streptomycin , 1% non-essential amino acids , 1% sodium pyruvate and 1% glutamine . MCF10A was grown at 37°C with 5% CO2 in F-12 MEM medium containing 10% fetal bovine serum , 1% penicillin/streptomycin , 10 µl of insulin , 10−8 M cholera toxin , 0 . 5 µg/ml of hydrocortisone and 10 ng/ml of EGF . HEK 293T was grown at 37°C with 5% CO2 in Dulbecco's modified Eagle medium ( DMEM ) containing 10% fetal bovine serum , 1% glutamine and 1% penicillin/streptomycin . HeLa was grown at 37°C with 5% CO2 in DMEM medium containing 10% fetal bovine serum and 1% penicillin/streptomycin . Plasmids for the lentivirus vector-mediated drug-inducible RNA interference system ( pLVTHM , ptTR-KRAB-Red , pCMC-R8 . 91 and pMD . G ) were provided by D . Trono ( University of Geneva ) [34] . The 64-mer oligonucleotides for histone H1 shRNA cloning into MluI/ClaI-digested pLVTHM were designed , annealed and phosphorylated as communicated by D . Trono ( http://tronolab . epfl . ch/ ) . Oligonucleotides have the following general structure: 5′-CGCGTCCCC-N19 TTCAAGAGA-rcN19-TTTTTGGAAAT-3′ and 5′-AGGGG-N19-AAGTTCTCT-rcN19-AAAAACCTTTAGC-3′ , being N19 the specific target sequence for each H1 variant and rcN19 its reverse complementary sequence . The corresponding 19-mer gene-specific target sequences for interference were designed manually following standard rules ( AAN19 , GC % 30–70 ) . Target sequences ( N19 ) are CGCTGACTCGCAGATCAAG for H1 . 0 , AGAGCGTAGCGGAGTTTCT for H1 . 2 , CTGCCAAGAGTCCAGCTAA for H1 . 3 , GAAGAGCGCCAAGAAGACC for H1 . 4 and GGCAACTAAGAAGGCTGCC for H1 . 5 . For the production of viral particles containing the HIV-derived vectors , 2 . 5×106 HEK 293T cells ( Clontech ) were transfected with plasmids ptTR-KRAB-Red , pLVTHM-shH1 . n or pEV833-HA-H1 . n ( 10 µg ) , pCMV-R8 . 91 ( 6 . 5 µg ) and pMD . G ( 3 . 5 µg ) in 10 cm dishes using calcium phosphate . Medium was collected every 24 hours for 2 days and centrifuged 1h30min at 26 , 000 rpm at 4°C in a sucrose gradient to concentrate viruses . Pellet containing viral particles was dissolved in medium and used for cell infection . Cells were infected using the spinoculation method , i . e . plates were centrifuged at 1 , 200× g for 2 h at 25°C . Initially , a cell line expressing the Dox-responsive KRAB repressor and RedFP ( ptTR-KRAB-Red ) was generated . Then , this cell line was infected with viruses for expression of the different H1 variants shRNAs ( pLVTHM ) . The inducible knocked-down cell lines were sorted in a FACSvantageSE ( Becton Dickinson ) for RedFP-positive and GFP-positive fluorescence after 3 days of Dox treatment . Then cells were amplified in the absence of Dox until an experiment was performed . Human histone H1 variants were PCR-amplified from genomic DNA and cloned into pCDNA4-HA vector provided by D . Reinberg's group ( NYU Medical School ) . The complete H1-HA cassette was cloned into the lentiviral expression vector pEV833 provided by E . Verdin ( Gladstone Institutes ) upstream an IRES-GFP cassette . Virus production and infections with pEV833-derived lentivirus were performed as described above . shRNA- resistant H1 . 2 and H1 . 4 were created by site-directed mutagenesis with QuikChange Mutagenesis Kit ( Stratagene ) . The introduced changes ( in lower case: AGAaCGgtcCGGcGTTagT for H1 . 2 , and GAAatctGCgAAGAAGACC for H1 . 4 ) did not alter the amino acidic sequence . The N-terminal domain of human H1 variants H1 . 1 to H1 . 5 was used to design peptides for production of subtype-specific antibodies ( Figure S1 ) . Peptides were synthesized and conjugated to a carrier protein , Keyhole Limpet Hemocyanin ( KLH ) , in the Proteomics Facility of the Pompeu Fabra University . Each of two rabbits was injected with 200 µg of peptide for each H1 variant , at two weeks intervals , for a total of 4 immunizations . Sera were collected and analyzed a week after last immunization . Antibodies were purified in a HiTrap Protein A HP column ( Amersham Biosciences ) according to manufacturer instructions and some are available through Abcam Ltd . ( Cambridge , UK ) : H1 . 1 ( ab17584 ) , H1 . 3 ( ab24174 ) and H1 . 5 ( ab24175 ) . A second generation of antibodies ( including H1 . 2 ab17677 and H1 . 5 ab18208 ) was produced in collaboration with Abcam starting from two-branched peptides 2× ( specific peptide ) -Lys-Cys as reported elsewhere [35] . Histone H1 was purified by 5% perchloric acid lysis for 1 hour at 4°C . Soluble acid proteins were precipitated with 30% trichloroacetic acid over night at 4°C , washed twice with 0 . 5 ml of acetone and reconstituted in water . Whole cell extracts were obtained lysing cells in Tris-HCl ( pH 7 . 4 ) 25 mM , EDTA 1 mM , EGTA 1 mM and SDS 1% plus protease and phosphatase inhibitors . Protein concentration was determined in both cases by Micro BCA protein assay ( Pierce ) . Lysates or purified proteins were subjected to 12% SDS-PAGE , transferred to a nitrocellulose membrane , blocked with 5% non-fat milk for 1 hour , incubated with primary antibodies over night at 4°C and secondary antibodies conjugated to peroxidase for 1 hour at room temperature . Bands were visualized by chemiluminiscence using ECL system ( Amersham ) . Quantifications were performed with Image Gauge ( FujiFilm ) software . The anti H1 . 0 ( ab11079 ) , H1 phospho-T146 ( ab3596 ) , HA-tag ( ab9110 ) and CDC2 ( ab18 ) antibodies were from Abcam; anti tubulin was from Sigma; anti CDK2 ( m2 ) , CDK4 ( c22 ) , CCND1 , p27 ( c19 ) , p16 ( h156 ) and p21 ( c19 ) antibodies were from Santa Cruz; phospho-Rb ( Ser 608 ) was from Cell Signalling . Immunofluorescence and chromatin immunoprecipitation ( ChIP ) assays with the HA-tag antibody were performed as described previously [36] . CDKs were immunoprecpitated from total cell extracts and its kinase activity measured on Rb or H1 protein substrates as described elsewhere [37] , [38] . Recombinant human H1 subtypes were purchased from Alexis Biochemicals ( produced in E . coli bacteria ) , or provided by N . Happel ( University of Göttingen ) ( produced in yeast ) . Recombinant Rb fragment 792–925 was kindly provided by O . Bachs ( IDIBABS-UB ) . Cells were washed with cold PBS 1× , fixed in 70% ethanol and stained with Analysis solution: 3% Ribonuclease A ( Sigma ) 10 mg/ml , 3% solution A ( 38 mM sodium citrate , 500 µg/ml propidium iodide ) in PBS 1× . Samples were analyzed using a FACS Calibur machine ( Becton Dickinson ) , CellQuest analysis software and ModFit program . Pellet of cells growing in rich medium and treated or not for 6 days with Dox was dissolved in buffer A ( 10 mM Tris-HCl pH 7 . 4 , 10 mM NaCl , 3 mM MgCl2 , 0 . 3 M sucrose and 0 . 2 mM PMSF ) plus 0 . 2% of NP40 and incubated for 10 min at 4°C . Nuclei were obtained after centrifugation and digested with 2 . 5 units of Mnase ( Worthington ) per 4 million of nuclei for 25 min at room temperature in buffer A plus 10 mM CaCl2 . DNA was purified through a Qiagen column and run on 1 . 2% agarose gel . Total RNA was extracted using the RNAsy Kit ( Qiagen ) . The quality of the RNA was analyzed using the Agilent Bioanalyzer 2100 and the RNA 6000 LabChip Kit ( Agilent ) with the Eukaryote Total RNA Nano Assay . cDNA was generated from 100 ng of RNA using Superscript First Strand Synthesis System ( Invitrogen ) . Gene products were analyzed by qPCR using SYBR Green Mix ( Roche ) and specific oligonucleotides in a Roche 480 Lightcycler . Each value was corrected by human GAPDH and expressed as relative units . Each experiment was performed in triplicate . Gene-specific oligonucleotide sequences are available on request . The RT Profiler PCR Array ( APHS-020 ) from SuperArray Biosciences Corporation is a 96-well plate containing primers for 84 cell cycle related genes , plus 5 housekeeping genes and 3 RNA and PCR quality controls . The system also includes an instrument-specific master mix and an optimized first strand synthesis kit . First strand synthesis and qPCR was performed following instructions from manufacturer and run in a Roche 480 Lightcycler . Procedures for microarray hybridization and data analysis are described elsewhere [36] , [39] and detailed in the Text S1 . Microarray data is available at GEO with accession numbers GSE11294 and GSE12299 ( http://www . ncbi . nlm . nih . gov/geo/index . cgi ) .
In order to explore the specificity of the different H1 subtypes in gene expression control in breast cancer cells , we generated T47D-derived stable cell lines with reduced expression of each of the H1 variants specifically ( H1 . 0 to H1 . 5 ) . H1 . 1 was left out as it is not expressed in T47D cells ( Figure S1 ) . We used an inducible shRNA expression system based on a TeT-On strategy [34] . shRNAs targeting divergent gene regions were designed to specifically inhibit the H1 . 0 , H1 . 2 , H1 . 3 , H1 . 4 and H1 . 5 variants of the linker histone and cloned in the lentivirus vector pLVTHM . Knocked-down cell lines were generated first by infecting the regulator vector that expresses a RedFP marker and a repressor TetR-KRAB fusion protein , then the vector codifying for the shRNA and reporter GFP . Upon doxycycline ( Dox ) addition , derepression of GFP and shRNA occurs and the targeted gene is knocked down . Cells simultaneously expressing RedFP and GFP upon Dox addition were selected by cell sorting . A control cell line was generated by infection with the empty pLVTHM vector . Depletion of the different H1 variants was analyzed by Western blot with specific antibodies ( See Materials and Methods and Figure S1 ) . H1 . 4 was detected with a H1 phospho-T146 antibody characterized in Figure S1 . Western blot analysis of H1-extracted sample from each knocked-down cell line revealed a consistent and specific reduction of every H1 variant ( Figure 1A ) . No significant compensation effects by other H1 subtypes were observed at protein level , but a H1 . 0 increase after H1 . 4 depletion . Efficient depletion was achieved after six days of treatment with Dox ( Figure 1B ) . Relative quantification of H1 . 2 inhibition by Western blot showed a 8-fold decrease in protein level ( Figure 1C ) . Depletion of other isoforms oscillated between 70 and 99% ( Figure 1A ) . Electrospray mass spectrometry of H1 preparations from all Dox-treated cell lines also showed specific reduction of each H1 subtype identified by its differential molecular weight ( data not shown ) . The knock-down effect was also measured by reverse transcription and real-time PCR ( RT-qPCR ) analysis of H1 variants gene expression before and after Dox treatment for all the cell lines generated ( Figure 1D ) . This showed specific reduction of the targeted gene ( between 60% and 90% depending on the variant ) and some increase in H1 . 0 mRNA in both H1 . 2 and H1 . 4 knocked-down cell lines . Significant reduction of H1 variants transcripts due to shRNA expression occurred 2 days after Dox addition ( Figure 1E ) , earlier than the apparent reduction of H1 proteins . Upon Dox treatment , we observed differences in growth rate among H1 variant knock-downs , in particular , H1 . 2 and H1 . 4 depleted cells failed to reach confluency ( Figure 2A ) and exhibited a slower growth rate . To quantify this we mixed , at a 1∶1 ratio , each of the shRNA expressing cell lines ( RedFP and GFP-positive upon Dox treatment ) with parental T47D cells ( RedFP and GFP-negative ) , and we monitored by FACS the proportion between the two populations over time in culture in the presence of Dox ( Figure 2B ) . Slow progression of H1 . 4 knocked-down cells was seen at six days after Dox addition and of H1 . 2 depleted cells at day 9 . At day 12 , H1 . 5 depleted cells were also less abundant than the parental ones . In addition , at day 6 , H1 . 4 cells had changed morphology towards a necrotic phenotype ( Figure 2A and Figure S2 ) . To understand the causes of these proliferation defects , the cell cycle profile of the different cell lines was examined by FACS analysis of propidium iodide-stained cells six days after Dox addition and compared to untreated cells . The knocked-down cell lines for H1 . 0 and H1 . 3 did not show any significant difference in the cell cycle profile in comparison to uninduced cells or to the control cell line . On the other hand , both H1 . 2 and H1 . 4 depleted cells exhibited a dramatic reduction in the S phase cell population ( Figure 2C and Table S1 ) . H1 . 5 depletion caused a slight decrease of S phase . Interestingly , in the H1 . 2 knock-down , an increase in the G1 peak over controls and a concomitant decrease in G2 peak were observed . That is , depletion of H1 . 2 induced G1 arrest . We cannot discard that complete depletion of H1 . 0 , H1 . 3 or H1 . 5 could have different effects . In the H1 . 4 knock-down , a small increase of G2 and almost no change in G1 was observed . Although the cell cycle pattern was not very dramatic , the increase in the subG1 peak indicates a high level of mortality in the H1 . 4 depleted cells ( Figure 2D ) , in accordance with the cell growth experiments . Experiments performed to measure apoptotic cells ( such as formation of apoptotic DNA ladder , anexin V binding or TUNEL assays ) have failed to find differences between H1 . 4-depleted and control cells , indicating that cell death is not mediated by apoptosis ( data not shown ) . Based on these results , we conclude that H1 . 4 knocked-down cells show a cell death phenotype , probably by necrosis , although more accurate analysis needs to be done . The deleterious effect of H1 . 4 depletion in T47D cells was rescued by transient transfection of a plasmid expressing recombinant shRNA-resistant H1 . 4 ( Figure S2 ) , discarding off-target effects . To further characterize the G1 arrest phenotype of H1 . 2-depleted cells , we studied the cell cycle progression after recovering from a serum-starvation block in G1 . H1 . 2 knock-down ( KD ) cells treated or not with Dox for 6 days were serum-starved for the last two days . After serum addition , the cell cycle profile was determined at 24 , 30 and 48 h ( Figure 2E and Table S1 ) . Dox untreated cells progressed along the cycle , with a maximum of cells in S phase at 24 h , in G2/M at 30 h , and in G1 at 48 h . Dox-treated cells remained arrested in G1-phase , with little increase in cells in S-phase at 30 h after serum addition . This indicated that the G1 block caused by H1 . 2 depletion is strong and very few cells are able to escape and progress through the cell cycle . Hydroxyurea ( HU ) and methyl methanosulfonate ( MMS ) are damaging agents that block cycling cells in S and S/G2 phase , respectively . We investigated how these drugs affect the cell cycle profile of ±Dox-treated H1 . 2 and H1 . 4 KD cells . While the cell cycle profile of H1 . 4 KD in response to HU or MMS was practically indistinguishable in the presence or absence of Dox ( data not shown ) , H1 . 2 KD cells treated with Dox remained accumulated in G1 whether or not HU or MMS was present ( Figure 2F ) . Taken together , our results indicated that H1 . 2 depletion causes a strong G1 arrest phenotype and that H1 . 4 KD cell cycle is not greatly affected in the few surviving cells . In order to demonstrate that the cell cycle arrest was specifically due to the depletion of the H1 . 2 variant , we stably integrated into the H1 . 2 KD cell line a lentiviral vector for the expression of an shRNA-resistant H1 . 2-encoding gene C-terminally fused to the HA peptide ( rH1 . 2-HA ) . The levels of rH1 . 2-HA protein were similar to the endogenous histone variant as shown by Western blot analysis ( Figure 3A ) . Cell cycle analysis demonstrated that the G1 arrest observed in the H1 . 2 depleted cells was reversed when the rH1 . 2-HA protein was present ( Figure 3B ) . Moreover , to exclude that the phenotype was not due to a reduction in the total H1 content , we performed the experiment stably expressing ectopically each of the other HA-tagged H1 variants in the inducible H1 . 2 KD cell line . Incorporation of recombinant HA-tagged H1 variants into chromatin was tested by several means , including Western blot of H1 extracted from chromatin , immunofluorescence and chromatin immunoprecipitation ( Figure S3 ) . Expression of H1 . 4-HA and H1 . 5-HA was comparable to rH1 . 2-HA , H1 . 0-HA level was sensibly lower , and H1 . 3-HA approximately 10-fold inferior ( Figure 3A ) . As Figure 3C shows , none of the other variants was able to reverse the G1 arrest caused by the depletion of H1 . 2 . Taken together , this indicates that H1 . 2 variant is specifically required for normal cell cycle progression and its depletion causes cells to accumulate in G1 phase . Complementation results with H1 . 3-HA and , probably , H1 . 0-HA are not definitive due to its lower expression levels . In order to investigate whether the effect of H1 . 2 and H1 . 4 deletion on cell proliferation was specific to the breast cancer cell line T47D used or if it is a more extended phenotype , we introduced the inducible shRNA expression system in a different breast adenocarcinoma cell line ( MCF7 ) , a non-tumoral breast epithelial cell line ( MCF10A ) , a cervical adenocarcinoma cell line ( HeLa ) and an embryonic kidney cell line ( 293T ) . Efficient H1 . 2 and H1 . 4 depletion in response to Dox treatment was obtained for all cell lines , ranging from 63 to 95% ( Figure 4A and 4B ) . In vivo observation of cells by microscope after 6 days of Dox treatment indicated that H1 . 2 depletion caused a clear defect in MCF7 cell growth , but not on HeLa , 293T or MCF10A ( Figure 4C ) . Interestingly , H1 . 2 depletion was more pronounced in MCF10A than in MCF-7 . In order to better describe the cell proliferation defect , H1 . 2 KD cells were cocultured 1∶1 with parental cells , treated with Dox and growth of the two populations was monitored by FACS along time ( Figure 4D ) . Growth of MCF7 H1 . 2 KD , but not MCF7 control or MCF10A H1 . 2 KD , was slower than parental cells , and comparable to the results shown for T47D H1 . 2 KD shown in Figure 2B . Consequently , the effect of H1 . 2 depletion on cell proliferation was not restricted to T47D , but observed also in a different breast cancer cell line , MCF7 . However , analysis of the cell cycle profile has failed to show G1 arrest . Instead a clear increase in the subG1 peak was observed in MCF7 cells ( Figure 4E ) . It is worth noting that , unlike T47D , the MCF7 cell line is p53-positive and , consequently , prone to apoptotic death in response to adverse stimuli . This could explain the differences between the two breast cancer cell lines in response to H1 . 2 depletion . On the other hand , H1 . 4 depletion was also obtained in the four cell lines tested , but only in MCF10A did it affect cell proliferation , as occurred in T47D ( Figure 4B , 4C and 4D ) . Noteworthy , H1 . 4 depletion was stronger in HeLa cells , but no effect on cell proliferation was observed . We cannot rule out that higher depletion efficiencies ( e . g . in MCF7 H1 . 4 KD ) would produce further effects . Nonetheless , our results strongly suggest that H1 . 2 and H1 . 4 are involved in processes related to normal cell cycle progression in a cell type-dependent fashion . Reduced H1 content ( down to 50% ) in triple-H1 null mouse embryonic stem ( ES ) cells leads to reduced nucleosome spacing ( nucleosome repeat length ) , as measured after micrococcal nuclease ( MNase ) digestion of nuclear bulk chromatin [12] . We investigated nucleosome spacing of chromatin from the different H1 variant T47D knock-downs by MNase digestion of isolated nuclei ( Figure 5A ) . H1 . 2 depletion , but not depletion of other H1 variants , caused a striking reduction in the spacing between nucleosomes . The calculated nucleosome repeat length ( NRL ) was decreased from approximately 184 . 7 to 173 . 5 upon H1 . 2 depletion ( Figure 5B , obtained from five independent experiments ) . This was surprising as H1 . 2 only represents approximately 23% of the total H1 content in T47D cells ( Figure S1 ) . As H1 . 2 depletion reached 90–95% and compensatory increases in the expression of other H1 variants were not detected , we estimate that total H1 content was reduced by approximately 20% . This was confirmed by Coomassie staining of H1 preparations from H1 . 2 KD cells treated or not with Dox ( Figure S1 ) . Stable expression of HA-tagged , shRNA-resistant H1 . 2 in the H1 . 2 KD cell line reverted the reduction in nucleosome spacing caused by endogenous H1 . 2 depletion ( Figure 5C ) . In triple-H1 null ES cells , nucleosome spacing reduction was accompanied by global changes in different histone post-translational modifications [12] . Histone modifications , such as trimethylation of K27 or trimethylation of K9 in core histone H3 , remain unchanged in our H1 . 2 depleted cells ( data not shown ) . H1 . 2 depletion alters global chromatin structure as seen by the reduction in nucleosome spacing . This could have some effect on nuclear processes such as DNA replication or gene expression and , ultimately , cell proliferation . We studied the effect of H1 . 2 depletion on global gene expression . For this , we used a custom-made breast cancer microarray platform containing 826 cDNA clones ( see Text S1 ) to compare H1 . 2 KD cell line treated for 6 days with Dox with untreated cells and with control cells ±Dox . In order to avoid comparing cells arrested in the cell cycle ( H1 . 2 KD +Dox ) with cells normally cycling ( H1 . 2 KD −Dox ) , all cells were equally arrested in G1 by serum deprivation over the last two days of ±Dox treatment prior to RNA extraction ( Figure 2E ) . Depletion of H1 . 2 caused an alteration in the expression of a limited number of genes . Most of altered genes were repressed in the absence of the histone: 11 genes were up-regulated and 59 genes were down-regulated with a fold-change higher than 1 . 4 ( q≤0 . 1 ) , with the H1 . 2-encoding gene ( hist1h1c ) being the strongest down-regulated cDNA due to the shRNA specific action ( Figure 6 ) . These changes were not observed in control cells , discarding an effect of Dox on gene expression . This indicates that H1 . 2 , rather than being a general repressor , is acting direct or indirectly as positive regulator of gene expression . Among the repressed genes , many of them are involved in cell cycle control . Gene ontology identified 20 out of the 59 repressed genes as being related to cell cycle , and 10 to cell division . This proportion is higher than would be expected , considering the abundance of these gene categories in the customized array . Alteration in the expression of one or a subset of cell cycle genes could cause the G1 arrest . The analysis of the consequences of H1 depletion on global gene expression was extended to the other H1 variants using genome-wide Illumina microarrays containing more than 22 , 000 probes ( see Text S1 ) . All KD cell lines were treated or not with Dox for 6 days , in duplicate , and arrested in G1 by 48 h of serum-starvation prior to RNA extraction . A summary of the number of genes up- or down-regulated with a fold-change ≥1 . 4 ( false discovery rate q≤0 . 1 ) in each H1 variant KD cell line upon Dox treatment in comparison to untreated cells is presented in Figure 7 and Tables S2 and S3 . The proportion of altered genes , up- or down-regulated , ranged between 1 and 2-% , depending on the H1 variant under study . Globally , approximately 6% of the genes are altered by depletion of some of the variants . H1 depletion caused 1 . 5 times more down- than up-regulation of genes . In particular , the ratio down- to up-regulated ranged between 1 for H1 . 5 and 2 . 7 for H1 . 2 . As reported above , H1 . 2 depletion mostly leads to gene down-regulation . A total of 533 genes were up-regulated by depletion of some H1 variant . Of these , 443 ( 83% ) were unique to a single variant . Seventy genes were up-regulated by two H1 variants , and 20 by three or four variants . The number of genes down-regulated by depletion of some variant was 812 . Of these , 615 ( 76% ) were uniquely regulated , 111 genes were regulated by two variants , and 86 by more than two variants . Interestingly , H1 . 2 , H1 . 3 and H1 . 4 were the variants that produced a higher proportion of uniquely-regulated genes , indicating that these are the variants with a more specific role in gene expression control . On the other hand , the majority of genes down-regulated in H1 . 0 and H1 . 5 KD were also affected by depletion of other H1 variants . In conclusion , depletion of the different H1 variants caused an alteration in the expression of a limited number of genes , different in each H1 variant KD . Most of the genes are affected by a single H1 variant , supporting the theory that H1 isoforms play specific roles in gene expression . Nonetheless , a proportion of genes are altered by more than one H1 variant , suggesting that redundant roles for H1 variants may also exist . To further explore the relationship between H1 . 2 depletion and the cell cycle gene expression profile , we tested expression of 84 cell cycle-related genes by RT-real time PCR using the SuperArray ( Bioscience Corporation ) platform ( Figure 8 ) . Fourteen genes were significantly repressed ( more than 2-fold ) when H1 . 2 was depleted after Dox addition for 6 days in serum-starved cells ( q≤0 . 5 ) . This number was increased to 24 repressed genes when considering ≥1 . 5-fold and q≤0 . 1 , and included 10 genes already identified in the customized microarray experiments: BIRC5 , CDKN3 , CDC20 , CCNB2 , CDC2 , CCNB1 , MAD2L1 , CKS2 , BCL2 and RBBP8 . Among the newly identified genes were CDK2 , PCNA , MCM2 , MCM3 , MCM5 , CDKN2A ( p16INK4A ) and CDKN2B ( p15INK4B ) . Only two genes were activated ≥2-fold: RBL2 and CCNG2 . In the same experiment , cells kept ±Dox ( 6 days ) and serum-starved ( last two days ) were further incubated with 10% serum for an additional 12 hours period and RNA was extracted . Differences in expression of the repressed genes were even higher , confirming that expression of this subset of genes was affected as a consequence of H1 . 2 depletion ( Figure 8 ) . Expression of some of the identified genes was further analyzed by conventional RT-qPCR with specific oligonucleotides . Interestingly , we detected induction of the cell cycle inhibitor CDKN1A ( p21Cip1 ) , which had not been observed using the previous methods . Some of the genes were repressed ( CDC2 , CDKN3 , RFC3 ) or activated ( CDKN1A ) as early as 2 days after Dox addition . Other genes were not repressed until day 6 ( CDC20 ) ( Figure S4 ) . Next , we tested whether inhibition of genes in H1 . 2 depleted cells could be reverted by stable reintroduction of recombinant H1 . 2 ( Figure 9A ) . Inhibition of CDC2 , CCNB2 , CDC20 , MAD2L1 and RFC3 was reverted by restoring H1 . 2 levels . Induction of CDKN1A ( p21Cip1 ) was only partially reverted by reintroduction of H1 . 2 . Expression of cell cycle-related genes may vary during cycle progression . We investigated the effect of H1 . 2 depletion on the expression of these genes over time after release from a serum-starvation induced G1-arrest block . In H1 . 2-depleted cells ( Dox-treated ) , the response of the genes to serum was , in most of the cases , delayed and reduced ( Figure S4 ) . In order to test whether repression of this cell cycle gene subset was specific to H1 . 2 depletion , we tested gene expression in control , H1 . 0 , H1 . 2 and H1 . 4 knock-down cells ±Dox , in parallel , with the SuperArray platform ( Table S4 ) . Twenty-seven genes were repressed ( ≥1 . 5-fold ) in comparison to control cells ±Dox when H1 . 4 was depleted , with approximately 50% coincidence with genes repressed upon H1 . 2 inhibition . This included CDKN3 , MAD2L1 and CCNB2 . Only one gene was slightly activated . On the other hand , H1 . 0 depletion caused activation of approximately 20 genes ( ≥1 . 5-fold ) and repression of only 2 genes . Our results indicate that depletion of different H1 variants affect expression of cell cycle-related genes diversely . In order to better characterize the G1 arrest caused by H1 . 2 depletion , we analyzed accumulation or phosphorylation of several cell cycle regulators in ±Dox-treated H1 . 2 KD and control cells by Western blot with specific antibodies ( Figure 9B ) . Phosphoryation of retinoblastoma protein ( Rb ) is a key gatekeeper of G1 to S transition through the restriction point of the cell cycle [40] , [41] . Rb phosphorylation was highly reduced in Dox-treated H1 . 2 KD cells , but not in control cells , as shown in two different experiments . Higher phoRb in H1 . 2 KD ( −Dox ) compared to parental T47D observed in experiment 1 was not reproducible . Rb phosphorylation is required for transcription of S-phase genes , e . g . CDC2 , PCNA and cyclins E/A . Some of them are consistently down-regulated in H1 . 2 KD , such as CDC2 and PCNA . Accordingly , the level of CDC2 protein was reduced in H1 . 2 KD cells +Dox ( Figure 9B ) . Rb is consecutively phosphorylated in the G1 phase by the CDK4/6-Cyclin D1 and CDK2-Cyclin E pairs . We did not detect changes in the accumulation of Cyclin D1 or CDK4 proteins ( Figure 9B ) . Instead , CDK2 levels were reduced in H1 . 2 KD , as well as the CDK2 -associated kinase activity immunoprecipitated from these cells ( Figure 9B and 9C ) . CDK2 gene expression was also reduced upon H1 . 2 depletion ( Figure 8 and Table S4 ) . As with CDC2 , CDK2 down-regulation in H1 . 2 KD cells could also help to explain the G1 arrest . In the DNA damage response , p53-dependent induction of expression of the CDK inhibitor p21Cip1 is involved in the defect in Rb phosphorylation [42] . In our experiment , the levels of the CDK inhibitors p21Cip1 and p27Kip1 did not change ( Figure 9B ) , although CDKN1A gene expression was increased in H1 . 2-depleted cells ( Figure 9A ) . Considering that T47D cells are p53-defective , we cannot rule out that a different player affected ultimately by H1 . 2 depletion , such as inhibitors of the INK4 family , is impacting on Rb phosphorylation . One of these inhibitors , p16INK4A was not detected in our cells ( Figure 9B ) . We can conclude that G1 arrest in H1 . 2 KD cells is caused by deregulation of cell cycle factors upstream the phosphorylation of Rb , but its exact causes still remain elusive to us as does , more importantly , the way in which H1 . 2 governs regulation of these factors .
We have investigated the consequences of sudden depletion of a particular H1 subtype . We found that H1 . 2 and H1 . 4 have an essential role in normal growth of a human breast cancer cell line in culture . On the other hand , 70–90% depletion of H1 . 0 , H1 . 3 or H1 . 5 subtypes was compatible with normal cell growth . We cannot discard that further depletion of these isoforms could have additional consequences . To our knowledge , no previously reported features of H1 subtypes anticipated these results . H1 . 0 is highly abundant in terminally differentiated cells , while somatic H1 variants ( H1 . 1-H1 . 5 ) are present in dividing cells with differing abundance depending on cell type . Interestingly , no human cell line lacking expression of H1 . 2 or H1 . 4 has been found to date , supporting the important role of these two isoforms [45] , [46] . These reports show that all cell types tested express H1 . 2 and H1 . 4 at a high level , each accounting for at least 20% of the cellular H1 content . In contrast , cell types can be found where some of the other H1 subtypes are not detectable or at very low levels ( ≤5% of total H1 ) . This is in agreement with our estimations of the H1 content in the five cell lines used in this report . Analysis of the H1 variants pattern in T47D cells combining gel electrophoresis and immunoblotting ( Figure S1 and data not shown ) indicated that the relative contribution of subtypes to the total H1 content was approximately as follows: 9% for H1 . 0 , 23% for H1 . 2 , 13% for H1 . 3 , 24% for H1 . 4 and 31% for H1 . 5 . On the other hand , the amount of H1 . 5 was low in HeLa and 293T , H1 . 3 was undetectable in HeLa , and H1 . 0 almost inexistent in MCF10A ( Figure 4 ) . Histone H1 . 4 depletion had a strong impact on cell survival of T47D cells . This cell line is functionally defective in p53 and caspase-3 , and consequently these cells do not undergo apoptosis in response to stress or damage . Indeed , we have been unable to detect signs of apoptotic death upon H1 . 4 depletion . Six days after induction of H1 . 4 KD , few cells remained attached to the culture flask , and these cells presented a necrotic phenotype . When the cell cycle profile was analyzed , very few cycling cells were detected and FACS analysis identified the presence of cells in the subG1 window . Gene expression profiles showed the alteration in expression of a considerable number of cell cycle related genes . Although our data suggests that H1 . 4 is essential for T47D cell growth , at this point it cannot be discarded a more complicated explanation as a consequence of indirect effects of H1 . 4 depletion , such as the resulting increased expression of H1 . 0 . Depletion of H1 . 2 in T47D cells caused G1 arrest and , consequently , slow proliferation . This effect is specific to H1 . 2 depletion , as it did not occur in KD cells for the other variants and it was complemented solely by the reintroduction of recombinant , shRNA-resistant H1 . 2 . The effects of H1 . 2 or H1 . 4 depletion are not unique to the T47D cell line , but are also observed in MCF7 and MCF10A cells , respectively . Noteworthy , H1 . 2 depletion did not cause G1 arrest in MCF7 but rather an apoptotic phenotype . Differences in genetic background between cell lines may explain different outcomes for the same stress . For instance , unlike T47D , MCF7 is p53-positive and prone to apoptosis . Cell lines from other origins did not exhibit a cell cycle phenotype in response to H1 . 2 KD . This is perplexing as H1 . 2 accounts for almost 50% of the total H1 content in HeLa ( Figure 4A and [46] ) , and is also highly abundant in 293T . This suggests that the effect of H1 depletion is cell type specific and unlinked to the relative amount of each variant in the cell , nor correlates with the efficiency of depletion achieved by RNA interference . Tumor cells undergo uncontrolled proliferation , arising from the ability to bypass the quiescent state characteristic of most normal cells from adult tissues [41] . H1 . 2 depletion blocks the proliferative potential of breast cancer cells and therefore unraveling the underlying mechanism could provide insights into how normal cells become tumorigenic . Triple null H1-2-H1 . 3-h1 . 5 mouse embryonic stem ( ES ) cells present a 15 base pair reduction in the bulk chromatin NRL [12] . H1 . 2 depletion of breast cancer cells altered chromatin organization , measured as nucleosome spacing after controlled MNase digestion of nuclei . NRL was decreased by about 11 bp upon H1 . 2 depletion , and the effect was specific to this subtype and complemented by recombinant H1 . 2 expression . This was unexpected as a reduction of NRL has been reported only after a significant ( 50% or higher ) decrease in H1 content [12] . We estimated that H1 . 2 accounts for only about 20–25% of the total H1 content in T47D cells and depletion by shRNA was not complete . In our experiments , depletion of other H1 subtypes accounting for a similar or higher percentage of total H1 ( H1 . 4 and H1 . 5 ) had no effect on chromatin structure , indicating that H1 . 2 may play a specific role in chromatin organization or nucleosome spacing . Triple KO mouse ES cells proliferate at the same rate as WT ES cells and also exhibit a decrease in the levels of some core histone modifications as well as in CpG methylation within the regulatory regions of certain genes [12] . So far , we have not detected changes in global histone modifications nor in DNA methylation upon H1 . 2 depletion ( data not shown ) . Using a microarray platform to measure methylation of CpG loci in 807 genes ( including cell cycle control , apoptosis , chromosome X-linked , and imprinted genes ) , we have not observed significant changes in H1 . 2 KD cells treated or not with Dox for 6 days ( M . Sancho , A . Fernández , M . Beato , M . Esteller , A . Jordan; unpublished results ) . Epigenetic changes to compensate for H1 depletion in ES cells may require longer times and may allow cell survival under selective pressure , whereas rapid induction of H1 . 2 shRNA does not allow for compensation effects to appear before cell progression is compromised . We have not yet been able to establish a causative relation between the reduced nucleosome spacing and the cell cycle arrest in G1 phase observed upon H1 . 2 depletion in T47D . Alterations of chromatin organization could affect progression of the replication fork , or could expose to damage regions of DNA leading to a damage response . It has been recently reported that H1-depleted cells ( triple KO ES cells ) are hyper resistant to DNA damage and present hypersensitive checkpoints ( G2/M arrest induced at lower doses ) , explained by increased signaling generated at each DNA break [47] . If so , upon H1 . 2 depletion we would expect S or G2/M arrest in response to damaging agents such as HU or MMS , respectively . But we have clearly shown that G1-arrest is dominant with regard to arrest caused by such compounds . H1 depletion could also have an effect on the formation and stability of mitotic chromosomes . In the absence of H1 , chromosomes assembled and replicated in Xenopus egg extracts failed to compact properly , leading to segregation anomalies [48] . In triple KO mouse ES cells , no abnormalities in mitosis and normal division time were detected [12] . Nonetheless , the decrease in H1 leads to a general increase in telomere length [47] . It remains to be explored if selective H1 . 2 depletion leads to chromosomal abnormalities . Chromatin missorganization could also cause changes in the expression of sensitive promoters that may include genes relevant for cell proliferation . Alternatively , H1 . 2 may be a specific regulator of some particular promoters . It has been reported before that H1 . 5 interacts with transcription factor MSX1 and both are recruited to specific promoters to control myogenesis [19] . On the other hand , H1 . 2 has been found to be part of a complex that represses p53-mediated transcription in HeLa cells [24] . With this idea in mind , we have explored what changes in gene expression profiles occur upon depletion of different H1 subtypes . Less than 10% of the genes present in a customized microarray ( and almost 2% on a genome-wide array ) were altered in its expression upon H1 . 2 knock-down and most of them were repressed . This behavior has been reported before [14]–[16] and is indicative of H1 acting as an active regulator of gene expression , not solely a repressor due to its chromatin compaction properties . Repressed genes included a relevant proportion of cell cycle-related genes . We reject the possibility that this is due to different proportion of cells in the different cell cycle phases upon exposure to the shRNA , as all cell populations were equivalently arrested in G1 due to serum-starvation for the gene expression experiments . Rather we found an effect of H1 . 2 deprivation on a subset of cell cycle related genes , some of which are in line with the G1 arrest phenotype and may be the link to the cell proliferation defect . Replicative or premature , stress-induced senescence is characterized by maintenance of cells in a growth-arrest state ( comparable to G1 arrest ) , by forming heterochromatin at proliferation-promoting gene loci ( E2F target genes ) . It has recently been suggested that alteration in histone H1 metabolism may be involved in a cellular senescence-inducing mechanism [49] . Senescent fibroblasts show reduced levels of chromatin-bound endogenous H1 . That report did not clarify whether histone H1 loss plays a causative role in senescence induction or is only an effect of the process , as RNA interference of H1 was unsuccessful . Our data suggests that specific H1 subtype loss from chromatin may cause by itself this phenotype in T47D cells , indicating that H1 loss may be a messenger between senescence-inducing signals and the molecular response leading to growth arrest . G1 progression and entry into S phase depends on the activity of two cell cycle kinase-complexes , CDK4/6-cyclin D and CDK2-cyclin E , that work in concert to relieve inhibition of a dynamic transcription complex that contains retinoblastoma ( Rb ) protein and E2F . Phosphorylation of Rb by CDK4/6 and CDK2 dissociates the Rb-HDAC repressor complex , permitting transcription of key S-phase-promoting genes , including some required for DNA replication [40] . Downstream genes controlled by E2F , such as CDC2 and PCNA are repressed in the absence of H1 . 2 . CDKs activity is regulated by members of the INK4 ( p15 , p16 , p18 , p19 ) or KIP/CIP ( p27 , p21 ) families of cyclin dependent kinase inhibitors induced by different agents or growth conditions promoting cell cycle arrest , differentiation or apoptosis . In H1 . 2 knock-down cells ( +Dox ) , Rb remains unphosphorylated when compared with −Dox cells , indicating that the cause of G1 arrest is upstream Rb phosphorylation by CDKs . p16INK4A , p27Kip1 and p53-inducible p21Cip1 could be candidates to explain low Rb phosphorylation , but we have not appreciated any difference in the level of either inhibitor in H1 . 2 KD cells , despite the fact that the CDKN1A ( p21Cip1 ) gene is induced upon H1 . 2 depletion . Involvement of other inhibitors of the INK4 family needs to be tested . Cyclin D1 and CDK4 levels also remain unaltered . On the contrary , CDK2 and CDC2 levels are reduced in H1 . 2 KD cells . The CDC2 ( CDK1 ) -Cyclin E pair has been involved recently in G1/S transition and could substitute CDK2 , and so it could also account for Rb phosphorylation [50] , [51] . Reduced levels of these two proteins may account for cell cycle arrest by itself , or in combination with additional yet unrecognized factors . Among the repressed genes in H1 . 2 KD cells , there are genes involved in initiation of replication such as those encoding for CDC6 , CDC45L , PCNA and several MCM proteins . Down-regulation of these factors could prevent entry into S phase and , consequently , cells would remain in G1 . Another set of repressed genes encode for proteins involved in G2/M transition , such as CDC2 , and mitosis progression , such as MAD2L1 , CDC20 and CDC23 , cyclin B , aurora kinase B . Their repression does not explain the G1 arrest , but may be part of a gene network co-regulated by H1 . 2 depletion . It is difficult to unequivocally assign a gene or group of genes as direct targets of the H1 . 2 depletion or as responsible of the G1 arrest phenotype . The first would require investigation into the participation of this particular histone subtype in the promoter activation process of candidate genes by chromatin immunoprecipitation , but H1 subtype specific antibodies generated have been useless to date for this application . The second would require testing the effect of interfering with each cell cycle-related candidate gene . Probably , simultaneous , incomplete silencing of several genes would be needed to recapitulate the G1 arrest observed here [52] . Our data and data published elsewhere are indicative that depletion of cell cycle regulatory proteins such as CDC2 , CDK2 , CDC20 , MCMs or PCNA is deleterious for cell cycle progression [53]–[55] . Global gene expression profiling combining inducible H1 variant depletion and genome-wide microarrays has allowed us to define the subsets of genes regulated directly or indirectly by each variant . The overall number of regulated genes and the proportion of genes up- or down-regulated differ between H1 isoforms . Depletion of the different H1 variants caused an alteration in the expression of a limited number of genes , different in each H1 variant KD . Most of the genes ( 79% ) are affected by a single H1 variant , supporting the theory that H1 isoforms play specific roles in gene expression . Nonetheless , a proportion of genes are altered by more than one H1 variant , suggesting that redundant roles for H1 variants in gene expression regulation may also exist . 13% of genes were altered in two variant KDs , and only 1% of genes were affected ( down-regulated ) in each of the five KDs . This is in agreement with a recent report showing that overexpression of mouse H10 and H1c ( human H1 . 0 and H1 . 2 orthologs ) altered a limited number of variant-specific genes , as well as a smaller number of commonly regulated genes [16] . In that case , too , genes were either up- or down-regulated upon H1 overproduction , with a stronger predominance of up-regulated genes in the H1c-overexpressing cells and a major number of down-regulated genes in the H10 cell line . Similarly , we have obtained a higher ratio down- to up-regulated genes in the H1 . 2 KD ( ratio DW:UP = 2 . 7 ) than in the H1 . 0 KD ( ratio = 1 . 9 ) . This supports the theory that H1 . 0 is a stronger repressor of transcription and H1 . 2 may play a role as transcription activator or facilitator . Nonetheless , following with this argument , H1 . 4 and H1 . 5 would be even more repressive ( ratio = 1 . 4 and 1 , respectively ) . This , actually , fits with previous description of histone H1 variant distribution and compaction capability . It has been proposed that mouse H1 . 0 , H1 . 4 and H1 . 5 promote formation of compact chromatin and are enriched in more compact chromatin , while H1 . 2 condenses chromatin less effectively and is enriched in less condensed chromatin regions [8] , [56] . Our analysis has also shown that H1 . 3 , and especially H1 . 2 and H1 . 4 , are the most specific variants in controlling gene expression , with a higher number of uniquely regulated genes . This might be related to the essentiality of H1 . 2 and H1 . 4 in breast cancer cells reported here , and with the fact that no cell types lacking any of the two variants have been found to date . These two isoforms may play more specific roles , while the other variants may be more redundant . This is still speculation and to proof it would need to cross the different H1 variant knock-downs with all the vectors for ectopic expression of HA-tagged recombinant isoforms . Considering that histone H1 is a global chromatin component , our and previous reports studying global gene expression changes in response to H1 depletion [12]–[15] , show that H1 affects expression of a surprisingly reduced number of genes . Here , approximately 6% of the genes present in the 22K-gene microarray show altered expression ( >1 . 4-fold ) upon inhibition of some of the five H1 variants tested . Nevertheless , not all the genes present in the microarray platform are expressed in this breast cancer cell line growing under the conditions used , i . e . G1 arrest caused by serum starvation . Only about 36% of the genes were expressed above background ( data not shown ) . This increases the proportion of genes being affected by depletion of some H1 variant in up to 16% of expressed genes ( 6 . 6% up- and 10% down-regulated by H1 depletion ) . This is already a considerable number of affected genes , larger than the 0 . 56% of expressed genes affected ( >2-fold ) in mouse triple H1 knock-out embryonic fibroblasts published elsewhere [12] . If we also consider a >2-fold variation in gene expression for the analysis of data , 2% of expressed genes are still affected upon depletion of some H1 variant . The striking difference between the two studies —with all due caution with regard to differences in the data analysis procedure applied— might be due to the inducible nature of the knock-down strategy used here . Along a small number of cell divisions , cells are encountered with a reduced level of a particular isoform . In the triple mutant , cells have been selected for survival and have adapted to the new situation , allowing for the resetting of a hypothetical basal gene expression profile required for the maintenance and progression of cells . In conclusion , our approach of inducible depletion of individual H1 variants using RNA interference technology has revealed that H1 subtypes are not fully redundant but may have differentiated functions in some of their nuclear roles , including nucleosome spacing and gene expression control , impacting on important cellular events such as proliferation . | Eukaryotic DNA is packaged into chromatin through its association with histone proteins . The linker histone H1 sits at the base of the nucleosome near the DNA entry and exit sites to stabilize two full turns of DNA . In particular , histone H1 participates in nucleosome spacing and formation of the higher-order chromatin structure . In addition , H1 seems to be actively involved in the regulation of gene expression . Histone H1 in mammals is a family of closely related , single-gene encoded proteins , including five somatic subtypes ( from H1 . 1 to H1 . 5 ) and a terminally differentiated expressed isoform ( H1 . 0 ) . It is not well known whether the different variants have distinct roles or if they regulate specific promoters . We have explored this by inducible knock-down of each of the H1 variants in breast cancer cells . A different subset of genes is altered in each H1 knock-down , and depletion has different effects on cell survival . Interestingly , H1 . 2 and H1 . 4 depletion specifically caused arrest of cell proliferation . Concomitant with this , H1 . 2 depletion caused decreased global nucleosome spacing and repressed expression of a number of cell cycle genes . Thus , specific phenotypes are observed in breast cancer cells depleted of individual histone H1 variants . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/chromatin",
"structure",
"molecular",
"biology/transcription",
"initiation",
"and",
"activation"
] | 2008 | Depletion of Human Histone H1 Variants Uncovers Specific Roles in Gene Expression and Cell Growth |
A central challenge in host-pathogen systems biology is the elucidation of general , systems-level principles that distinguish host-pathogen interactions from within-host interactions . Current analyses of host-pathogen and within-host protein-protein interaction networks are largely limited by their resolution , treating proteins as nodes and interactions as edges . Here , we construct a domain-resolved map of human-virus and within-human protein-protein interaction networks by annotating protein interactions with high-coverage , high-accuracy , domain-centric interaction mechanisms: ( 1 ) domain-domain interactions , in which a domain in one protein binds to a domain in a second protein , and ( 2 ) domain-motif interactions , in which a domain in one protein binds to a short , linear peptide motif in a second protein . Analysis of these domain-resolved networks reveals , for the first time , significant mechanistic differences between virus-human and within-human interactions at the resolution of single domains . While human proteins tend to compete with each other for domain binding sites by means of sequence similarity , viral proteins tend to compete with human proteins for domain binding sites in the absence of sequence similarity . Independent of their previously established preference for targeting human protein hubs , viral proteins also preferentially target human proteins containing linear motif-binding domains . Compared to human proteins , viral proteins participate in more domain-motif interactions , target more unique linear motif-binding domains per residue , and contain more unique linear motifs per residue . Together , these results suggest that viruses surmount genome size constraints by convergently evolving multiple short linear motifs in order to effectively mimic , hijack , and manipulate complex host processes for their survival . Our domain-resolved analyses reveal unique signatures of pleiotropy , economy , and convergent evolution in viral-host interactions that are otherwise hidden in the traditional binary network , highlighting the power and necessity of high-resolution approaches in host-pathogen systems biology .
Protein-protein interactions ( PPIs ) can be broadly classified into two fundamentally different classes: those within the same species , such as within-host PPIs , and those between different species , such as host-pathogen PPIs . Are there general , systems-level principles that distinguish host-pathogen PPIs ( exogenous interactions ) from within-host PPIs ( endogenous interactions ) ? Surprisingly , little is known about the existence and nature of such global principles , in part because they are not amenable to investigation by traditional methods , which examine specific host-pathogen PPIs individually . The most well-studied host-pathogen interaction systems are host-virus interactions , and the combined results of decades of detailed studies on specific host-virus interactions suggest that such global principles may exist . Endogenous interactions among host proteins are expected to be cooperative: proteins encoded within the same genome interact with one another to carry out biological function in a coordinated and synergistic fashion . On the contrary , exogenous interactions between viral proteins and host proteins are expected to be largely antagonistic: viruses physically manipulate host cell machinery to perpetuate their genomes at the host's expense . In addition to hijacking host macromolecular complexes to make new viral products , viruses are known to modulate the host response to infection in order to escape detection and prevent the host from interfering with viral replication [1]–[3] . Many viral proteins directly compete with host proteins for binding sites [4] , and some even modify host proteins chemically , e . g . marking them for degradation by the host's own machinery [5]–[7] . Despite providing such detailed information about the molecular mechanisms and consequences of specific exogenous interactions , traditional virology studies are highly focused and thus are often unable to draw general conclusions about the mechanisms governing exogenous interactions even among closely related viruses . As a result , despite these detailed studies on specific host-virus interaction systems , the overarching principles that distinguish host-virus interactions from within-host interactions have not yet been elucidated . A systems biology approach is therefore essential in order to obtain a global perspective on host-pathogen interactions . Recent advances in high-throughput experimental and computational biology have enabled the reconstruction and analysis of large-scale host-pathogen PPI networks [8]–[18] . These systematic studies have successfully revealed global patterns in host-pathogen systems that are otherwise inaccessible by the traditional reductionist approach , which studies host-pathogen PPIs one at a time . For example , global analyses have revealed that viral proteins have repeatedly evolved to target host proteins central to the host PPI network ( e . g . hubs with many physical interaction partners ) [8] , [11] , [19] . In addition to targeting common host pathways regulating viral infection and replication in general [11] , [20] , different classes of viruses also target host pathways uniquely involved in class-specific mechanisms of infection and replication [20] . Despite these advances , current host-pathogen systems biology approaches are highly abstract and coarse-grained , treating proteins as nodes and PPIs as edges; therefore , the insights generated by these analyses are strongly limited in spatial and mechanistic resolution . A high-resolution approach is needed to uncover more general rules governing host-pathogen PPI networks [21] . One approach to increase resolution in PPI networks has been to construct three-dimensional ( 3D ) structural models to protein interactions [22]–[25] . We recently applied this technique to build an atomic-resolution map of human-virus and within-human PPI networks by constructing 3D structural models of exogenous and endogenous PPIs using homology modeling [4] . A direct comparison between the resulting human-virus and within-human structural interaction networks revealed systematic and significant differences between exogenous and endogenous interactions that are otherwise hidden in the binary PPI networks . For example , we found that viral proteins preferentially bind to and mimic existing endogenous interfaces on their human target proteins , rather than creating entirely new interfaces . In addition , interface mimicry tends to be achieved without structural similarity in the human-virus PPI network as compared to the within-human PPI network . Finally , endogenous interfaces mimicked by virus proteins tend to evolve quickly , and mediate many endogenous interactions that are transient and regulatory in function , as compared to generic endogenous interfaces [4] . Although 3D structure information can be used to interrogate host-pathogen interaction networks at atomic resolution in a reasonably unbiased manner , coverage in these analyses is limited by the number of high-quality 3D homology models that can be built for endogenous and exogenous interactions [26] . In this work , we probe high-resolution principles governing exogenous and endogenous PPI networks using a domain-resolved approach that annotates proteins with known domains , and PPIs with known domain-centric interaction mechanisms ( domain-domain interactions and domain-motif interactions; Figure 1 ) . This domain-resolved network is of higher accuracy than the binary PPI network , and of higher coverage than the 3D structural interaction network . Although domain-based studies of host-pathogen PPIs have been previously reported for specific pathogens [27] , [28] , a systematic , quantitative comparison between exogenous and endogenous PPI networks at the level of domains has never been attempted before . Domain-motif interactions have been previously reported to be important in host-pathogen interactions [29] , but their prevalence in the global host-pathogen interaction network remains unknown relative to the within-host network [29] . Our global , domain-resolved map of human-virus and within-human PPI networks enables , for the first time , the discovery of novel systematic and statistically significant differences between exogenous and endogenous PPIs in terms of domain interaction usage . While two human proteins competing to bind the same domain tend to have global sequence similarity , viral proteins competing with human proteins do not . Viral proteins preferentially target human proteins containing linear motif-binding domains independent of their degree in the endogenous network . In addition , viral proteins use linear motifs to mediate protein-protein interactions more often than human proteins do . Finally , viral proteins contain a higher density of linear motifs than generic human proteins . Collectively , these observations suggest that the exogenous network is very different from the endogenous network in terms of domain interaction usage . While the endogenous network evolves largely by gene duplication followed by divergence , the exogenous network is dominated by convergent evolution of domain-motif and domain-domain interactions . Compared to human proteins , viral proteins tend to convergently evolve and pack multiple linear motifs mediating many biophysical interactions that are functionally diverse in order to manipulate complex host processes . Together , these results strongly support the utility of a domain-resolved approach for interrogating host-pathogen interaction networks , and in particular for determining the general principles that distinguish exogenous and endogenous interactions .
We constructed high-resolution human-virus ( exogenous ) and within-human ( endogenous ) protein-protein interaction ( PPI ) networks by annotating proteins and PPIs with known domain information . We considered two major categories of domain-centric interaction mechanisms: domain-domain interactions ( DDIs ) , in which a globular domain from one protein binds to a globular domain from a second protein ( Figure 1A ) , and domain-motif interactions ( DMIs ) , in which a linear motif-binding ( LMB ) domain from one protein binds to a short , linear peptide motif in a second protein ( Figure 1B ) . Some PPIs can be annotated with both DDI and DMI mechanisms . The endogenous portion of our network contains 39 , 329 PPIs among 9 , 870 human proteins , of which 48 . 7% can be assigned to at least one of the two domain-centric mechanisms ( Figure 2 ) . There are 8 , 277 DDIs mediated by 1 , 164 unique domain types forming 3 , 209 unique types of interacting domain pairs . In addition , there are 16 , 785 DMIs mediated by 554 unique LMB domain types ( Figure 2 ) . The exogenous portion of our network contains 1 , 670 interactions between 267 viral proteins and their 954 human protein targets . The viral proteins represent 17 viral families and 7 Baltimore classes ( Table S1 ) [30] . 30 . 5% of all exogenous interactions can be assigned at least one domain-centric interaction mechanism , which can be further divided into the following five cases ( Figure 2 ) . ( i ) 30 exogenous DDIs involve a human domain homolog present in a viral protein , presumably due to horizontal gene transfer ( Table S2 ) . ( ii ) 110 exogenous DDIs involve human domains interacting with virus-specific domains . ( iii ) 443 exogenous DMIs are mediated by an LMB domain-containing human protein binding to a viral protein with the corresponding linear motif . ( iv ) 11 exogenous DMIs are mediated by a viral protein containing a human-like LMB domain binding to a human protein with the corresponding linear motif ( e . g . the SH2 and SH3 domain-containing Src protein from Avian sarcoma virus , and the kinase domain-containing BGLF4 protein from Epstein-Barr virus ) . ( v ) Exogenous DMIs mediated by virus-specific LMB domains are only just starting to be characterized [31] , and are not represented in our network . Annotating exogenous and endogenous PPI networks with domain-centric interaction mechanisms yields networks with increased resolution compared to binary networks , while maintaining high coverage . We annotated proteins with complete taxonomic and Pfam domain information [32] , [33] , and PPIs with interacting domain information [34]–[36] . These annotated PPIs are of higher quality than generic PPIs , as measured by the overlap with a gold standard set of PPIs reported by at least two independent publications ( “confirmed interactions” ) . Specifically , endogenous interactions annotated with domain-centric mechanisms are 52% more likely to be confirmed than non-annotated endogenous interactions ( Figure 3 ) , and exogenous interactions annotated with domain-centric mechanisms are 28% more likely to be confirmed than non-annotated exogenous interactions ( Figure 3 ) . Hence , in addition to providing mechanistic insights , annotation of endogenous and exogenous interactions with domain interaction information raises our confidence in the accuracy of the underlying interactions . In our previous work based on 3D structural models of exogenous and endogenous interactions , we demonstrated that viral proteins frequently bind to human target proteins at sites of existing endogenous interfaces ( “interface mimicry” ) [4] . Moreover , compared to overlap among endogenous interfaces , exogenous-endogenous interface overlap was much less likely to involve global structural similarity between the two proteins targeting the same interface [4] . Here , we reexamined this result in the context of our domain-resolved human-virus PPI network . In the absence of 3D structural information , it is not possible to determine if two proteins bind to the same interface on a third protein . However , in our domain-resolved human-virus PPI network , it is possible to determine if two proteins bind to the same domain in the third protein ( Figure 4A–C ) , which is a prerequisite for interface mimicry . A similar approach has been previously used in the yeast 3D structural interaction network to distinguish between singlish-interface hub proteins , which mediate mutually exclusive PPIs , and multi-interface hub proteins , which mediate multiple simultaneous PPIs [22] . Among DDIs in the endogenous network , of the 3 , 493 cases where two human proteins bind to the same domain of a third human protein , 72% are mediated by domains sharing significant sequence similarity ( Figure 4D ) . In contrast , among DDIs in the combined exogenous-endogenous network , of the 46 cases where a viral protein and a human protein bind to the same domain of another human protein , only 24% are mediated by domains sharing significant sequence similarity ( Figure 4D ) . The results from these domain-resolved analyses are consistent with our previous findings using 3D structural networks: viral proteins are significantly less likely than human proteins to bind to the same domain of a human target protein by means of global sequence similarity to an endogenous binding partner ( Fisher's exact test , two-tailed P<10−10; Figure 4D ) . Viruses have been known to use linear peptide motifs to target endogenous LMB domains [21] , [29]; however , it is unknown how prevalent this mechanism of interaction is . Here , we quantified how frequently viral proteins target host proteins using a domain-motif interaction mechanism . We examined the domain composition of human proteins targeted by viruses , and compared it with the domain composition of generic human proteins in the network . We found that human proteins targeted by viruses are significantly enriched for LMB domains relative to generic human proteins ( Fold enrichment = 1 . 36; Fisher's exact test , two-tailed P<10−15; Figure 5 ) . With the exception of Orthomyxoviruses , the direction of this trend holds for exogenous interactions from all major viral families in the network , and cannot be attributed to a specific type of virus ( Figure 5 ) . In contrast , the difference in enrichment for non-LMB domains between human proteins targeted by viruses and generic human proteins is only marginally significant ( Fold enrichment = 0 . 96; P = 0 . 012; Figure 5 ) , suggesting that the observed enrichment for LMB domains among human proteins targeted by viruses is not a simple result of superior domain annotation among these proteins . Previous work has revealed a tendency for viral proteins to target host protein hubs [11] , [27] , [28] . Because LMB domains recognize small peptide motifs which may occur in many proteins , we expect LMB domain-containing human proteins to participate in more endogenous interactions than proteins without LMB domains , and hence be more hub-like . Indeed , the average LMB domain-containing human protein in our network participates in 10 . 5 endogenous interactions , while the average LMB domain-free protein participates in only 6 . 4 endogenous interactions . As a result , our finding that viruses tend to target LMB domain-containing proteins may be confounded by the viral preference for targeting hub proteins . We examined the effects of endogenous degree on the relationship between a human protein containing an LMB domain and the likelihood of that protein being a viral target . We stratified human proteins according to endogenous degree and then compared the probability of being a viral target among proteins with and without LMB domains ( Figure 6 ) . Consistent with previous findings that viruses target host protein hubs , we observe that the probability of being a viral target increases with increasing endogenous degree , and that this trend holds for both LMB domain-containing proteins and LMB domain-free proteins ( Figure 6 ) . More importantly , for a fixed endogenous degree , LMB domain-containing human proteins are more likely to be targeted by viruses than human proteins without LMB domains ( Figure 6 ) . This finding suggests that viruses preferentially target LMB domain-containing human proteins independent of their higher average degree . To quantify the statistical significance of this assertion , we measured concordance between ( i ) having an LMB domain and ( ii ) being a viral target , among pairs of human proteins with the same degree . We picked a pair of proteins with the same degree in which one had an LMB domain while the other did not , and considered the pair concordant if the LMB domain-containing protein was a viral target whereas the LMB domain-free protein was not , and discordant if the LMB domain-containing protein was not a viral target whereas the LMB domain-free protein was . We observed a strong preference for concordant protein pairs over discordant protein pairs ( 58% concordant versus 42% discordant ) , favoring a degree-independent association between LMB domain-containing proteins and viral targets . The degree-independent association between a human protein containing an LMB domain and being a viral target is statistically significant ( one-tailed P = 0 . 006; Figure 6 ) , as calculated by a degree-preserving random permutation of LMB domain and viral target annotations among sets of human proteins . The results of the previous section establish that viruses tend to preferentially target human proteins containing LMB domains by comparing properties of human proteins targeted by viruses against all other human proteins . Next , we determined if the viral preference for targeting LMB domain-containing proteins also holds at the level of PPIs , when comparing the fraction of domain-motif interactions ( DMIs ) between viral proteins and human proteins . We observed that viral proteins have higher fraction of DMIs out of total number of PPIs per protein than human proteins ( permutation test , two-tailed P = 0 . 047; Figure 7A ) . To ensure this trend was not due to superior annotation in either the endogenous or exogenous dataset , we repeated the analyses on confirmed interactions and observed the same trend ( P = 0 . 018; Figure 7B ) . This result suggests that although the endogenous network contains more proteins and PPIs and has a higher fraction of domain annotation than the exogenous network ( Figure 2 ) , viral proteins are more likely on average than human proteins to interact using a domain-motif interface ( Figure 7 ) . We next examined whether viral preference for targeting LMB domain-containing proteins is reflected in elevated linear motif occurrence in viral proteins as compared to human proteins . We determined density of linear motifs and LMB domains targeted per protein , rather than directly comparing the total number of linear motifs and LMB domains targeted per protein , to account for the large difference in protein size between viral and human proteins: within our network , the median viral protein length ( 306 residues ) is 34% smaller than the median human protein length ( 464 residues ) . We first calculated the density of unique LMB domains targeted per residue for viral proteins and human proteins . We found that viral proteins target a greater variety of unique LMB domains per residue than human proteins ( permutation test , two-tailed P = 0 . 012; Figure 8A ) . This calculation directly compares the properties of experimentally determined endogenous and exogenous PPIs , and may be confounded by methodological differences in mapping endogenous versus exogenous interactions: only 22% of endogenous interactions are reported by small-scale experiments ( reporting fewer than 100 interactions ) , whereas as many as 73% of exogenous interactions are reported by small-scale experiments . To ensure that the aforementioned trend observed in our network cannot be explained by this difference in methodology , we repeated the analyses on a host-virus PPI network built from the previously published “HI-2005” and “VirHost” interactome datasets , which were generated using the same methodology [37] , [38] , and observed the same trend ( P = 0 . 049; Figure 8A ) . This analysis supports our earlier conclusion that viral proteins interact with a greater variety of distinct LMB domains on a per residue basis than human proteins . Our observation that viral proteins target more LMB domains per residue than human proteins may still be confounded by subtle differences in experimental procedures for mapping endogenous versus exogenous interactomes . To control for such differences , we calculated the density of linear motif types per residue for each viral and human protein , regardless of whether the motifs were used to mediate known interactions . This measure is interactome-independent , and thus is free of any procedural biases in experimental interactome maps . Consistent with our previous findings , we found that viral proteins have significantly more unique linear motif types per residue than human proteins ( P<0 . 001; Figure 8B ) . These results indicate that in addition to preferentially targeting LMB domain-containing proteins ( Figures 5 and 6 ) , viral proteins are more likely to target a greater variety of unique LMB domains per residue than human proteins ( Figure 7 ) , and have a higher density of unique linear motifs than human proteins ( Figure 8 ) .
We constructed a domain-resolved map of host-virus and within-host protein-protein interaction ( PPI ) networks to probe general , systems-level principles that distinguish host-pathogen , exogenous PPIs from within-host , endogenous PPIs . Annotation of proteins and interactions with known domain information yields a domain-resolved network with higher resolution and quality than the binary PPI network , and higher coverage than the 3D structural interaction network . Classification of endogenous and exogenous PPIs into domain-domain interactions ( DDIs ) and domain-motif interactions ( DMIs ) reveals global differences in domain interaction patterns between host-pathogen and within-host networks that are otherwise hidden in traditional binary PPI networks . While our domain-centric annotations reduce the rate of false positives in PPI networks , additional potential limitations include false negatives , incomplete annotation , and methodological biases . In this work , we have minimized the effects of such incompleteness and biases by carefully controlling for them when performing systematic comparisons between exogenous and endogenous networks . A potentially more significant limitation is investigator bias: most host-pathogen studies are conducted on clinically significant human pathogens , such as HIV . Despite this investigator bias , our exogenous network represents a wide variety of viral families ( Table S1 ) . We emphasize that our comparisons and contrasts between exogenous and endogenous PPIs are carried out within our domain-resolved interaction networks , and therefore our conclusions should be minimally confounded by systematic biases inherent in a domain-resolved approach . Our analyses reveal systematic , mechanistic differences between exogenous and endogenous interactions . The most pronounced of these differences is the tendency for viruses to mimic human interactions by means of convergent evolution . We find that viral proteins and human proteins tend to target the same domain of another human protein without any shared sequence similarity , extending the results of our previous work using 3D structural interaction networks [4] . In addition , we demonstrate for the first time that viral proteins are more likely than human proteins to mediate interactions using short linear motifs , which can easily arise by convergent evolution due to their small size and minimal genomic constraints . These observations support the hypothesis that viral proteins tend to convergently evolve mechanisms to mimic existing endogenous binding interfaces . In addition , viral proteins are more economical and functionally more pleiotropic than human proteins in that viral proteins target more LMB domain-containing proteins , and also target more unique LMB domains per residue . Furthermore , we found that viral proteins contain more unique linear motif types per residue . Given the knowledge that linear motifs in disordered regions tend to be conserved and are more likely to be the target of binding by LMB domains [39] , we further investigated whether or not viral proteins are more disordered than human proteins . Indeed , we find that viral proteins are enriched for disorder-promoting residues [40] relative to human proteins ( Student's t-test , two-tailed P<0 . 0001 ) . Additionally , considering only motifs in “disordered regions” ( a region ±10 residues around a motif , containing >60% disorder-promoting residues [40] ) , we observe that viral proteins continue to have higher density of unique linear motif types per residue than human proteins do ( permutation test , two-tailed P<0 . 05 ) . Our results demonstrate that viral proteins and virus-host PPIs are in general very different from host proteins and within-host PPIs: viral proteins are small , complex , multifunctional polypeptides which can mediate multiple host-virus interactions , typically using the highly economical and highly pleiotropic method of domain-motif interactions largely through convergent evolution . These signatures of pleiotropy , economy , and convergent evolution in the virus-host PPI network are a direct consequence of the intense selection pressure on viruses to establish and maintain , with very limited genomic resources at their disposal , extensive and effective physical interactions with the host necessary for their survival . These global trends are applicable in general to viral proteins and exogenous interactions , and do not reflect a bias in a specific viral type , nor in a specific methodology for determining PPIs . Our results suggest that annotating viral proteins with domain-centric interaction mechanisms , especially by scanning viral protein sequences for linear motifs , can provide a novel approach to identifying host protein interaction partners for study . It may also be possible to use this domain-centric annotation approach to identify therapeutic treatments based on competition for motif binding sites . Thus , our study highlights the importance of a high-resolution , domain-resolved approach to host-pathogen network biology for revealing general mechanistic principles governing host-pathogen interactions .
We collected reports of endogenous ( human-human ) protein-protein interactions ( PPIs ) from the IntAct database , and reports of exogenous ( human-virus ) PPIs from the IntAct and VirusMINT databases [41] , [42] . We discarded PPIs with missing protein sequence information in UniProt [33] . Exogenous PPIs were further filtered to exclude ( i ) virus species that do not normally target mammalian hosts , and ( ii ) deltaviruses , which ( as subviral satellites ) cannot infect a host without co-infection by another virus . The viral proteins represent 17 viral families and all Baltimore classes [30] . We assigned Pfam domains to the human and viral proteins in our networks using the Pfam batch search utility , subject to an E-value cutoff of 10−2 [32] . To avoid misclassifying proviral fragments in the human proteome as native human domains , we removed human proteins from the analysis if they were annotated as viral fragments or polyproteins in Uniprot . Using protein sequence and domain information , we then assigned putative interaction mechanisms to endogenous and exogenous PPIs in our dataset . We classified PPIs as domain-domain interactions ( DDIs ) if a domain in the first protein was known or predicted to interact with a domain in the second protein . Pairs of putative interacting domains were assembled from the DOMINE database [36] , which integrates results from a variety of DDI curation and prediction studies . In addition , we classified PPIs as domain-motif interactions ( DMIs ) if one of the proteins contained a putative linear motif-binding ( LMB ) domain and the second protein contained a linear motif recognized by that LMB domain . We utilized predicted domain-motif associations from Neduva et al . [35] and manually curated domain-motif associations from the database of Eukaryotic Linear Motifs ( ELM ) [34] . Motifs in these datasets take the form of regular expressions which can be searched directly against an amino acid sequence using standard pattern matching tools . To assess the quality of the endogenous and exogenous networks , we compared them individually against a gold standard set of endogenous and exogenous PPIs; these subsets of interactions were constructed by querying for interactions that were reported by at least two independent publications . We measured concordance between having an LMB domain and being a viral target by picking pairs of human proteins with the same degree in which one had an LMB domain while the other did not . The protein pair was concordant if the LMB domain-containing protein was a viral target and the LMB domain-free protein was not a viral target . Conversely , the protein pair was discordant if the LMB domain-containing protein was not a viral target and the LMB domain-free protein was a viral target . All other protein pairs were considered to be non-informative . To evaluate the statistical significance of this test , we completed 1 , 000 repetitions of random permutation of the LMB domain and viral target annotations among sets of human proteins with the same endogenous degree and repeated our procedure . For permutation-based comparisons of virus and human proteins , we first compute the mean of each group and then evaluate the difference between these means . To evaluate if such a difference is likely to arise at random , we repeatedly permute the “virus” and “human” protein labels and then calculate the difference in the means of the newly randomized groups . Over a large number of trials ( e . g . 1000 ) , the fraction of permutations in which the random difference is at least as large as the observed difference approximates the probability of observing such a difference at random ( p-value ) , and serves as a measure of the statistical significance of the observed measurement . | The goal of host-pathogen systems biology is to examine the complex interactions between species , such as those between a virus and its host . Analysis of protein-protein interaction ( PPI ) networks can identify general principles that distinguish between within-species and between-species interactions . However , PPI data are limited by their low resolution , and cannot provide detailed information about the physical mechanisms underlying interactions between proteins . Using protein domain-based annotation methods , we have constructed an integrated human-virus PPI network which better highlights the mechanistic differences between human-human and human-virus PPIs . Our findings suggest that viral proteins use unique strategies to interact with human proteins , a finding with significant implications for pathogen research . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2013 | Signatures of Pleiotropy, Economy and Convergent Evolution in a Domain-Resolved Map of Human–Virus Protein–Protein Interaction Networks |
Alzheimer disease ( AD ) , Frontotemporal lobar degeneration ( FTD ) , Amyotrophic lateral sclerosis ( ALS ) and Parkinson disease ( PD ) have a certain degree of clinical , pathological and molecular overlap . Previous studies indicate that causative mutations in AD and FTD/ALS genes can be found in clinical familial AD . We examined the presence of causative and low frequency coding variants in the AD , FTD , ALS and PD Mendelian genes , in over 450 families with clinical history of AD and over 11 , 710 sporadic cases and cognitive normal participants from North America . Known pathogenic mutations were found in 1 . 05% of the sporadic cases , in 0 . 69% of the cognitively normal participants and in 4 . 22% of the families . A trend towards enrichment , albeit non-significant , was observed for most AD , FTD and PD genes . Only PSEN1 and PINK1 showed consistent association with AD cases when we used ExAC as the control population . These results suggest that current study designs may contain heterogeneity and contamination of the control population , and that current statistical methods for the discovery of novel genes with real pathogenic variants in complex late onset diseases may be inadequate or underpowered to identify genes carrying pathogenic mutations .
Neurodegenerative diseases like Alzheimer Disease ( AD ) , Frontotemporal dementia ( FTD ) , Parkinson disease ( PD ) and Amyotrophic Lateral Sclerosis ( ALS ) share clinical and pathologic features . Dementia is characteristic of AD and FTD , but may also present in PD and ALS [1] . In all these diseases we can observe two types of manifestations , either a rare and following Mendelian inheritance , or a more common seemingly non-familial representation [2] . All these diseases share the pathologic hallmark of presenting protein aggregates in different areas of the central nervous system . The rare familial forms have been key to our understanding of each disease’s pathology; these are well characterized by a dominant pathologic protein aggregate in a specific location within the nervous system caused by the impairment of a specific set of genes [2] . However , a clinical , and pathological crossover has been observed in the idiopathic forms [3] allowing for the assumption that different types of dementia may have overlapping genetic causes . AD is the most common neurodegenerative disease affecting over 5 . 5 million Americans [4] . AD’s main pathologic hallmark is the extracellular deposit of β-amyloid plaques , followed by the presence of intracellular aggregates of neurofibrillary tangles of phosphorylated tau protein ( MAPT ) . Pathogenic variants in APP , PSEN1 and PSEN2 were thought to exclusively cause early onset familial AD ( EOFAD ) [5] , but the screening of some late onset families and even sporadic cases , revealed that the contribution of familial genetic variants in APP , PSEN1 and PSEN2 in idiopathic cases cannot be neglected [6–8] . Intracellular aggregates of hyperphosphorylated tau ( encoded by MAPT ) is the second neuropathological hallmark of AD . Tau aggregates also characterize a subgroup of FTD cases ( FTD-Tau ) in which genetic linkage to MAPT was found [9] . The discovery of the gene involved in tau deposits stimulated a series of studies seeking association between polymorphisms within MAPT and AD . Preliminary studies investigating the genetic relationship between MAPT and AD produced contradictory results [10 , 11] although the latest reports provide more compelling evidence that common variants in the MAPT region confer risk for AD , particularly in cases who are non-carriers of the APOE ε4 allele [11–15] . Other familial FTD cases that are negative for tau protein aggregates present ubiquitin inmunoreactivity ( FTD-U ) , which was initially associated with genetic variants in GRN [16] . Together with MAPT , GRN genetic variants cause up to 50% of familial FTD cases , but GRN genetic variants have also been regarded as risk factors for AD [17–20] . Other familial genes later associated with FTD-U subtypes are C9ORF72 , TARDBP , VCP , CHMP2B , UBQLN2 and FUS [21] . A hexanucelotide repeat expansion in C9ORF72 was identified as major risk factor for FTD [22 , 23] although its pathogenicity towards ubiquitin reactive deposits is not well understood . C9ORF72 is also the most common risk factor for familial ALS [24 , 25] and multiple studies have reported clinical AD cases with the full C9ORF72 expansion [26–29] , extending the possible genetic continuum from AD to FTD and ALS . Shared genetic variants between FTD and ALS have been found in TARDBP , VCP , UBQLN2 and FUS [30–35] but their association with AD has not yet been reported . Other than the shared variants with FTD , the majority of familial ALS is caused by mutations in SOD1 [36 , 37] . Recently , mutations in OPTN , and PFN1 have been reported in ALS kindreds [37] . No associations have yet been reported between ALS genes and AD pathology , despite some attempts to demonstrate an association between common SOD1 variants and AD under the paradigm of a common oxidative stress pathway [38] . Finally , although AD is primarily characterized by cognitive deficits and PD by motor impairment , a clinical and pathological cross-over has been identified in several instances by the presence of dementia in PD patients [39] and motor symptoms in AD patients [40] . Pathologically , tau aggregates can be present to different degrees in sporadic PD [41] and more than 50% of people with AD show α-synuclein aggregates [41] . Lewy bodies , composed of α-synuclein aggregates , the pathological hallmark of PD , are attributed to pathogenic variants in the SNCA gene; although genetic variants in LRRK2 , PARK2 , PARK7 and PINK1 have also been linked to familial PD [42] . Combined meta-analysis of AD and PD GWAS revealed the lack of variants that increase the risk of developing both diseases [43]; although later on , a genetic overlap between AD and PD at the MAPT locus was detected [44] and recent studies have detected pathogenic PARK2 mutations in sporadic early onset AD cases [15] . A genetic overlap among all these neurodegenerative diseases cannot be ignored , and may certainly be underestimated since most of the previous studies performed either two by two gene/disease analysis , which does not cover the full spectrum of genes , or GWAS , which does not cover genetic variants across the frequency spectrum . In a previous work we reported an enrichment of known pathogenic and novel rare variants in APP , PSEN1 , PSEN2 , GRN and MAPT in LOAD families [45] . In this work , we expanded our analysis to a thorough examination of all genes known to cause Mendelian forms of AD , FTD , ALS , and PD . We evaluate the presence of rare and nonsynonymous genetic variants in these 30 genes not only in a large independent familial dataset ( 467 families ) , but also in two large sporadic cohorts , 851 in-house sporadic AD cases and controls and in the sporadic ADSP ( Alzheimer Disease Sequencing Project ) dataset ( https://www . niagads . org/adsp/content/home , accessed August 2016 ) .
Mutations in Mendelian AD genes are known to be autosomal dominant with complete penetrance; but we found that the variants identified in this study did not always present complete penetrance or segregate perfectly with disease status ( Table 2 ) . First , the pathogenic variant PSEN1 p . ( Ala79Val ) [46] was detected in a 67 yrs control from a large family with several affected members and an average AAO around the seventies ( Fam #1 ) . Genotyping of the mutation in up to 19 members of the family indicated incomplete penetrance ( Fig 1A ) , given the presence of phenocopies and some non-affected carriers , possible presymptomatic cases . In addition , this variant was found in seven cases from the ADSP dataset with an average AAO of 68 yrs . The pathogenic variant PSEN1 p . ( Leu85Pro ) [47] was detected in a family of four affected and two non-demented individuals ( Fam #2 ) , but only one of the affected individuals was a carrier of the genetic variant . The pathogenic variant PSEN1 p . ( Gly206Ala ) [48] was found in one case of an AAO 55 yrs , self-reported Caribbean origin ( Fam #3 ) . This variant was originally reported in 194 families of Caribbean Hispanic origin [48 , 49] . We also detected this variant in four cases from the sporadic ADSP dataset , one of European American origin ( AAO = 63 ) and three with Hispanic ethnicity ( mean AAO = 70 ) . Variant PSEN2 p . ( Asn141Ile ) , also pathogenic [51] , was detected in one family ( Fam #5 ) and in one sporadic case ( AAO = 53 ) of the KANL cohort . However , after further examination of the clinical history of this participant we detected a reported family history of dementia ( Supplementary results ) . Variant PSEN2 p . ( Met174Val ) [52] was detected in two families: Fam #6 was composed of six affected individuals and one diagnosed as non-demented . Only one affected and the cognitively-normal family members were carriers of the variant ( Fig 1B ) . Similarly , Fam #7 has five affected members and three cognitively normal members , in which only one cognitively normal member ( the marry-in ) was a carrier of the reported pathogenic variant . This variant was also observed in one control of the KANL sporadic dataset ( ALA = 72 ) and in two cases ( mean AAO = 82 ) and four cognitively normal ( 64 , 63 , 84 and 85 yrs ALA ) individuals from the sporadic ADSP dataset . Together these results suggest that previously reported pathogenic variants in Mendelian AD genes may present imperfect segregation , given the existence of phenocopies ( Fam #1 , Fam #6 ) ; and/or incomplete penetrance due to the presence of older cognitive normal carriers that may carry additional modifying factors . Also , their finding within the sporadic cohorts calls for a reexamination of the non-affected carriers as possible presymptomatic individuals . Known pathogenic variants in these genes were also observed exclusively within the “sporadic” datasets . Two affected individuals of the sporadic KANL cohort with age at onset ( AAO ) in their 50s were carriers of a known pathogenic APP variant , p . ( Ile716Val ) [53]; and variant APP p . ( Val717Phe ) [54] was found in one affected individual ( AAO = 70 ) of the ADSP cohort . The PSEN1 variant p . ( Leu226Arg ) [55] was detected in one affected participant ( AAO = 51 ) of the KANL cohort; variants PSEN1 p . ( His214Tyr ) [56] and p . ( Val412Ile ) [57] were found in two cases ( 85 yrs and 84 yrs AAO ) and variant p . ( Ala409Thr ) [58] was found in one cognitively normal participant ( ALA = 89 ) of the ADSP cohort . PSEN2 variant p . ( Ala85Val ) [59] was detected in one cognitively normal participant ( ALA = 89 ) and variant p . ( Leu238Pro ) [60] was found in two cases ( 80 and 63 yrs AAO ) of the ADSP cohort . Mutations in FTD genes are also known to segregate in a dominant pattern . Among the seven FTD genes examined , we observed four pathogenic variants in GRN , three pathogenic variants in MAPT , two variants in TARDBP and one pathogenic variant in VCP . We also found the repeat expansion C9ORF72 in several unrelated cases ( 0 . 85% ) and in 10 individuals with family history . The GRN variant p . ( Arg110* ) [61] was present in three siblings and one cousin of a family with a history of reported AD ( Fam #8 , Fig 1C ) , but only three of the affected members were carriers of the genetic variant . This variant was also detected in one affected participant ( AAO = 74 ) of the ADSP cohort . The GRN variant p . ( Thr382fs ) [62] was detected in a female AAO 60 yr with an AD diagnosis from the sporadic dataset . After examination of her clinical history it was discovered that she had two siblings and one cousin diagnosed with dementia ( Fam #9 ) . Genotyping of this variant in six family members revealed that the genetic variant was present in all members affected by dementia , and one young cognitively normal participant ( ALA = 65 ) . Pathology reports available indicated that the index individual had AD pathology and Pick bodies with a frontotemporal lobar atrophy pattern consistent with Pick’s disease; and the two siblings were later pathologically diagnosed as FTD and as non-AD dementia . Two other variants in GRN p . ( Arg493* ) [63] and p . ( Cys521Tyr ) [64] were detected in the sporadic ADSP cohort: GRN p . ( Arg43* ) was found in 4 affected participants ( average AAO = 73 ) and p . ( Cys521Tyr ) was found in 2 affected participants ( average AAO = 82 . 5 ) . The incomplete penetrance observed for GRN within the families can be the result of phenocopies ( Fam #8 ) or presymptomatic cases ( Fam #9 ) . The MAPT variant p . ( Gly389Arg ) [65] was found in one cognitively normal participant ( ALA = 91 ) from the ADSP cohort . MAPT variant p . ( Arg406Trp ) [66] was detected in two participants of the sporadic dataset , both cognitively normal in their 60s with no symptoms of presymptomatic AD ( not by imaging or CSF Aβ levels ) . MAPT variant p . ( Gln424Lys ) ( personal communication in 2005 by Brice to AD&FTDMDB Curator ) was detected in one family ( Fam #10 ) in which three of the five affected members were carriers of the variant . Variants in TARDBP were exclusively found in the ADSP cohort . Variant p . ( Asn267Ser ) [67] was detected in two affected members ( Average AAO = 77 ) and variant p . ( Asn390Ser ) [68] was detected in one affected member with AAO 74 yrs . The VCP variant p . ( Arg155His ) [69] was detected in Fam #11 but segregation could not be performed since DNA was only available for one of the affected members . Mutations in PD genes present different patterns of segregation; SNCA and LRKK2 are known to cause dominantly inherited PD , whereas PARK2 , PARK7 and PINK1 are known to cause early onset PD with a recessive inheritance mode [70] . We detected 10 different known pathogenic PD variants in LRKK2 , PARK2 and PINK1 , in 70 different carriers , all of whom were heterozygous for the variant . One LRRK2 variant , p . ( Gly2019Ser ) [71] , was only present in two families ( Fam #12 , Fam #13 ) . Fam #12 includes five members , four diagnosed with AD ( two carriers ) and one cognitively normal ( a carrier heterozygous for p . ( Gly2019Ser ) ) . Fam #13 is composed of two affected individuals ( one carrier ) and two cognitively normal ( no carriers ) . Mutations in the LRRK2 gene are the most common genetic cause of PD; but this gene is also known to have pleomorphic pathology [72] and the penetrance of the variant p . ( Gly2019Ser ) is known to vary in different populations and ages [73] , so , the incomplete penetrance observed here is not surprising . Seven variants were detected in PARK2: p . ( Gln34fs ) [74] was found in one case ( AAO = 79 ) from Fam #14 although the other case was a non-carrier , in two affected participants ( AAO of 31 and 82 yrs ) of the KANL sporadic dataset , and in three cases ( average AAO = 83 ) and three cognitive normal participants ( average ALA = 86 ) from ADSP . PARK2 variant p . ( Pro113fs ) [75] was present in four of 10 cases from three families ( Fam #15 , Fam #16 , Fam #17 ) and in two cases ( AAO 73 and 83 yrs ) and two cognitively normal ( ALA 64 and 87 ) participants of the ADSP cohort . In our study , variant PARK2 p . ( Met192Leu ) [76] was the most common of the PD mutations . All four members from Fam #18 ( three cases and one control ) were carriers of the genetic variant , as was the case from Fam #19 ( AAO = 55 ) . Within the KANL this variant was found in one affected ( AAO = 64 ) and one cognitively normal ( ALA = 68 ) member . Up to 13 affected carriers ( Average AAO = 88±7 ) and 11 cognitively normal carriers ( average ALA = 84±5 ) were detected in the ADSP cohort . Four other variants in PARK2 were exclusively found in the sporadic cohorts . Variant p . ( Thr240Met ) [77] was detected in one affected ( AAO = 77 ) and one cognitively normal ( ALA = 64 ) member of the KANL sporadic cohort , and in three affected ( average AAO = 82±10 ) and two cognitively normal ( average ALA = 88 ) members of the ADSP sporadic cohort . PARK2 variants p . ( Leu238Pro ) [78] and p . ( Gly430Asp ) [76] were found in the same cognitively normal individual of the ADSP cohort ( ALA = 68 ) , whereas PARK p . ( Arg366Trp ) [79] was found in two non-affected carriers ( average ALA = 65 ) from the KANL sporadic cohort . Finally , two variants were detected in PINK1 . Variant p . ( Arg464His ) was found in one affected ( AAO = 86 ) participant of the ADSP cohort . Variant p . ( Arg492* ) [80] was present in only one case ( AAO = 68 ) and one cognitively normal ( ALA = 73 ) member from Fam #20 , and in one affected ( AAO = 65 ) and three non-affected ( ALA 73 , 86 , 88 ) individuals from the ADSP cohort . Mutations in PARK2 and PINK1 are known to cause early onset PD ( AAO range 12–58 yr ) with a recessive pattern of inheritance . Segregation could not be determined in our study due to lack of familial stages . With the exception of the case in Fam #20 ( AAO of 55 ) , the individuals reported here all had an AAO > 60 , suggesting that these PARK2 and PINK1 variants would be modifiers rather than causative of AD . For a full description of all known pathogenic variants detected , see Supplementary Results . Once we confirmed that known pathogenic mutations in the AD , FTD and PD genes can be found in sporadic as well as in late-onset familial samples , we wanted to determine if an overall increase of low frequency non-synonymous coding variants can be found in these genes in AD cases compared to cognitively normal participants . We performed this analysis in our unrelated dataset and in the sporadic ADSP dataset . The cases in the unrelated KANL dataset present a larger enrichment of rare pathogenic variants compared to the cases of the sporadic ADSP dataset ( Table 3 ) . In both datasets , the enrichment increases when we focus on very rare non-synonymous variants ( Table 3 ) . None of these enrichments was significant after multiple test correction , but we observe some suggestive and nominally significant results . Within the unrelated KANL cohort we found CHMP2B ( OR = 2 . 24 , P = 0 . 06 ) and VPS35 ( OR = 6 . 77 , P = 0 . 05 ) to have suggestive significance-values , and EIF4G1 ( OR = 1 . 71 , P = 0 . 01 ) to be nominally significant for rare variants , those with minor allele frequency below 1% and a predicted high or moderate effect on the final protein ( MAF≤1% HM ) . Also , the global effect of AD genes ( APP , PSEN1 and PSEN2 ) was nominally significant ( OR = 2 . 31 , P = 0 . 028 ) if we considered only very rare variants , those with just one allele count in the population and a predicted high or moderate effect on the final protein ( AC1 HM ) . The enriched effect of CHMP2B remained significant when we considered only APOE ε4 carriers; some of the AD genes became nominally significant ( e . g . APP and PSEN2 ) and the global effect of AD genes resulted in nominal significance for rare variants and significant after multiple test correction ( Total AD , OR = 14 . 76 , P = 1 . 83×10−4 ) when we considered very rare variants ( S3 Table ) . EIF4G1 remained nominally significant ( OR = 2 . 01 , P = 0 . 02 ) for the set of APOE ε4 non-carriers; FBXO7 became nominally significant ( OR = 3 . 09 , P = 0 . 03 ) and VPS35 gained suggestive significance ( OR = 7 . 05 , P = 0 . 06 ) within the set of rare variants ( MAF≤1% HM ) . In addition , the global effect of very rare variants ( AC1 HM ) from FTD genes was also nominally significantly associated ( OR = 1 . 84 , P = 0 . 04 ) with APOE ε4 non-carriers ( S4 Table ) . Within the sporadic ADSP cohort , none of the enriched genes presented significant p-values ( Table 3 ) . However , PSEN1 resulted in nominally significant association with the APOE ε4 non-carrier for both sets , rare ( OR = 2 . 14 , P = 0 . 05 ) and very rare variants ( OR = 2 . 76 , P = 0 . 02 ) ( S3 Table ) . Because the lack of significant associations could be due to the sample size , we decided to perform gene-based analyses comparing all the unrelated KANL cases ( n = 672 ) or the sporadic ADSP cases ( N = 5 , 679 ) with the non-Finnish European ( NFE ) ExAC population as a cognitive normal dataset ( n = 33 , 000 ) . The coverage of the tested genes in ExAC and in our population is comparable . PSEN1 appeared as nominally significantly enriched in these two cohort cases for the set of rare variants ( MAF≤1% HM ) ; and it was statistically significant for sporadic ADSP cases within the very rare set of variants ( OR = 3 . 134 , P = 3 . 34×10-5 ) . Interestingly , PINK1 appeared as significantly associated with the unrelated cases examined regardless of the set of variants tested . EIF4G1 also remained significantly associated with unrelated KANL cases within the rare variants subset ( OR = 1 . 69 , P = 8 . 96×10−4 ) and for sporadic ADSP cases within the very rare set of variants ( S5 Table ) . No statistical association was found for the overall AD , FTD and PD genes in either the KALN or ADSP dataset ( S5 Table ) .
We would like to conclude by reinforcing that despite the limitations of this study the results obtained are sound and in the expected direction . First , this research is based on clinical AD participants and cognitive normals , so we cannot rule out the presence of presymptomatic cases , comorbidities , phenotypic changes with disease evolution , or even misdiagnoses , as we have already stated . This limits the power of this study to establish causality of the detected pathogenic variants and it could as well be limiting our statistical power and enrichment . Second , sporadic and familial AD are categorically different entities , despite the fact that we believe the genetic architecture and molecular load are largely similar [91] . However , the complex and heterogeneous presentation of AD makes it difficult to recruit large cohorts with good phenotypic characterization . Therefore , despite our effort to increase our sample size by creating an unrelated cohort , we are still underpowered to detect significant association for genes that we know are directly involved in the pathogenicity of the disease . Nonetheless , these findings highlight the need to join efforts to gather large sample sizes . Finally , we are also aware that the comparison against ExAC data has some methodological flaws . We even acknowledge the possible presence of pre-symptomatic cases within the Exac database . That is why we only consider those results as illustrative of what could be the genetic load when a larger dataset of cognitive normal participants is available .
Samples from the Washington University School of Medicine ( WUSM ) site included in this study were recruited by either the Charles F . and Joanne Knight Alzheimer's Disease Research Center ( Knight-ADRC ) at the Washington University School of Medicine in Saint Louis or the National Institute on Aging Genetics Initiative for Late-Onset Alzheimer’s Disease ( NIA-LOAD ) . From this point onwards , we will refer to these samples as KANL ( Knight-ADRC-NIA-LOAD ) . This study was approved by each recruiting center’s Institutional Review Board . Research was carried out in accordance with the approved protocol . Written informed consent was obtained from participants and their family members by the Clinical and Genetics Core of the Knight ADRC . The approval number for the Knight ADRC Genetics Core family studies is 201104178 . All the cases received a diagnosis of dementia of the Alzheimer's type ( DAT ) , using criteria equivalent to the National Institute of Neurological and Communication Disorders and Stroke-Alzheimer's Disease and Related Disorders Association for probable AD [96 , 97] . Cognitively normal participants received the same assessment as the cases , and were deemed non-demented . Written consent was obtained from all participants . The Alzheimer’s Disease Sequencing Project ( ADSP ) is a collaborative work of five independent groups across the USA that aims to identify new genomic variants contributing to increased risk for LOAD . ( https://www . niagads . org/adsp/content/home ) . During the discovery phase , they generated WGS data from members of multiplex AD families and whole exome sequence WES data collected in a large case-control cohort . These data are available to qualified researchers through the database of Genotypes and Phenotypes ( https://www . ncbi . nlm . nih . gov/gap Study Accession: phs000572 . v7 . p4 ) . Samples coming from the KANL site were sequenced using either whole-exome sequencing ( WES , 83 . 53% ) or whole-genome sequencing ( WGS , 16 . 46% ) . Exome libraries were prepared using Agilent’s SureSelect Human All Exon kits V3 and V5 or Roche VCRome . Both , WES and WGS samples were sequenced on a HiSeq2000 with paired ends reads , with a mean depth of coverage of 50x to 150x for WES and 30x for WGS . We performed joint analysis and quality control ( QC ) for all samples coming from the KANL site as well as for the ADSP familial study-design downloaded from dbGAP . Whether we started from BAM files or SRA files , all were converted to fastq files . Alignment was conducted against GRCh37 . p13 genome reference . Variant calling was performed separately for WES and WGS following GATK’s 3 . 6 Best Practices ( https://www . broadinstitute . org/gatk/ ) and restricted to Agilent’s V5 kit plus a 100 bp of padding added to each capture target end . WGS data was filtered to remove low complexity regions , and regions with excessive depth . Only those variants and indels that fell within the above 99 . 9% confidence threshold were considered for analysis; additional variant filters included allele-balance ( AB = 0 . 3–0 . 7 ) , quality depth ( QD ≥5 for indels and QD≥2 for SNPs ) , and missingness ( geno = 0 . 05 ) . Variants out of Hardy Weinberg equilibrium ( P<1x10-6 ) or with differential missingness between cases and controls , WES and WGS or different sequencing platforms were removed from analysis . In addition , individuals with more than 10% of missing variants and whose genotype data indicated a sex discordant from the clinical database were removed from dataset . Finally , individual and familial relatedness was confirmed using PLINK1 . 9 ( https://www . cog-genomics . org/plink2/ibd ) and an existing GWAS dataset for these individuals . Functional impact and population frequencies of variants were annotated with SnpEff [100] . At this point we separated familial from sporadic KANL datasets and generated the unrelated dataset . For the ADSP case-control dataset we downloaded plink file after alignment , variant calling and QC had been performed to which we performed additional QC . Briefly , we checked for allele-balance ( AB = 0 . 3–0 . 7 ) and differential missingness between cases and controls . We used the entire dataset for discovery of pathogenic variants but we later restricted our analysis to individuals with self-reported non-hispanic white ethnicity that we corroborated with PCAs . We focused our analysis on genes and variants reported as pathogenic and causing disease in a Mendelian pattern in AD , FTD , PD or ALS . For AD and FTD we restricted our analysis to those genes listed in the AD&FTD mutation database ( http://www . molgen . vib-ua . be/ADMutations/ , accessed November , 2016 ) ; particularly , we focused on APP , PSEN1 and PSEN2 for AD and CHMP2B , FUS , GRN , MAPT , TARDBP , TBK1 and VCP for FTD . For PD we started off with those genes and variants listed in the PD mutation database ( http://www . molgen . vib-ua . be/PDMutDB/ , accessed November , 2016 ) , namely , LRRK2 , PARK2 , PARK7 , PINK1 and SNCA; we also included UCHL1 , ATP13A2 , GIGYF2 , HTRA2 , PLA2G6 , FBXO7 , VPS35 , EIF4G1 and DNAJC16 for being reported as causative of Mendelian PD in several occasions [103 , 104] . To our knowledge , there is no ALS mutation database so we restricted our analysis to those genes consistently reported in the literature as causative of familial ALS , i . e . SOD1 , OPTN , UBQLN2 and PFN1 [37]; other familial ALS genes like FUS , VCP , UBQLN2 and SQSTM1 are included as FTD causing . | In this study we provide further genetic evidence of the clinical , pathological and molecular overlap between neurodegenerative diseases . We screened the known Mendelian genes in Alzheimer disease ( AD ) , Frontotemporal Dementia ( FTD ) , Parkinson disease ( PD ) and Amyotrophic Lateral Sclerosis ( ALS ) disease in over 13 , 292 individuals . We report the presence of known pathogenic mutations in AD , FTD and PD genes in both sporadic and familial late onset AD cases and even cognitively normal individuals , pointing to contamination of control cohorts with asymptomatic cases . We also observed an enrichment , although not statistically significant , in most of AD , FTD and PD genes , reinforcing the idea of pathologic crossover among these diseases . However , the fact that genes carrying known pathogenic mutation do not show significant association calls for a reevaluation of current statistical methods . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"neurodegenerative",
"diseases",
"pathogens",
"amyotrophic",
"lateral",
"sclerosis",
"alzheimer's",
"disease",
"motor",
"neuron",
"diseases",
"dementia",
"mental",
"health",
"and",
"p... | 2017 | Analysis of neurodegenerative Mendelian genes in clinically diagnosed Alzheimer Disease |
The Global Programme to Eliminate Lymphatic Filariasis ( GPELF ) , launched in 2000 , has the target of eliminating the disease as a public health problem by the year 2020 . The strategy adopted is mass drug administration ( MDA ) to all eligible individuals in endemic communities and the implementation of measures to reduce the morbidity of those suffering from chronic disease . Success has been recorded in many rural endemic communities in which elimination efforts have centered . However , implementation has been challenging in several urban African cities . The large cities of West Africa , exemplified in Nigeria in Kano are challenging for LF elimination program because reaching 65% therapeutic coverage during MDA is difficult . There is therefore a need to define a strategy which could complement MDA . Thus , in Kano State , Nigeria , while LF MDA had reached 33 of the 44 Local Government Areas ( LGAs ) there remained eleven ‘urban’ LGAs which had not been covered by MDA . Given the challenges of achieving at least 65% coverage during MDA implementation over several years in order to achieve elimination , it may be challenging to eliminate LF in such settings . In order to plan the LF control activities , this study was undertaken to confirm the LF infection prevalence in the human and mosquito populations in three urban LGAs . The prevalence of circulating filarial antigen ( CFA ) of Wuchereria bancrofti was assessed by an immuno-chromatography test ( ICT ) in 981 people in three urban LGAs of Kano state , Nigeria . Mosquitoes were collected over a period of 4 months from May to August 2015 using exit traps , gravid traps and pyrethrum knock-down spray sheet collections ( PSC ) in different households . A proportion of mosquitoes were analyzed for W . bancrofti , using dissection , loop-mediated isothermal amplification ( LAMP ) assay and conventional polymerase chain reaction ( PCR ) . The results showed that none of the 981 subjects ( constituted of <21% of children 5–10 years old ) tested had detectable levels of CFA in their blood . Entomological results showed that An . gambiae s . l . had W . bancrofti DNA detectable in pools in Kano; W . bancrofti DNA was detected in between 0 . 96% and 6 . 78% and to a lesser extent in Culex mosquitoes where DNA was detected at rates of between 0 . 19% and 0 . 64% . DNA analysis showed that An . coluzzii constituted 9 . 9% of the collected mosquitoes and the remaining 90 . 1% of the mosquitoes were Culex mosquitoes . Despite detection of W . bancrofti DNA within mosquito specimens collected in three Kano urban LGAs , we were not able to find a subject with detectable level of CFA . Together with other evidence suggesting that LF transmission in urban areas in West Africa may not be of significant importance , the Federal Ministry of Health advised that two rounds of MDA be undertaken in the urban areas of Kano . It is recommended that the prevalence of W . bancrofti infection in the human and mosquito populations be re-assessed after a couple of years .
Lymphatic filariasis ( LF ) is a disease caused by Nematode worms that live in the lymphatic vessels of humans . There are three species of filarial parasites which infect humans—Wuchereria bancrofti , Brugia malayi and Brugia timori . In Africa , LF is caused by W . bancrofti and is transmitted by Anopheles mosquitoes in rural settings and Anopheles and culicine mosquitoes in urban areas [1 , 2] . LF is a significant health problem in many developing countries with about 1 . 4 billion people known to live in endemic areas , with one quarter potentially infected [3] . LF is also the second leading cause of permanent disability and undermines the social and economic welfare of affected people and communities [4] . The World Health Assembly passed a resolution in 1997 to eliminate LF as a public health problem by the year 2020 . The World Health Organization has since 2000 launched a Global Programme to Eliminate Lymphatic Filariasis ( GPELF ) [5] . Annual mass treatment with single-dose diethylcarbamazine ( DEC ) or ivermectin ( IVM ) in combination with albendazole ( ALB ) for 4–6 years is the principal tool of the elimination strategy . This has resulted in dramatic progress in reducing the prevalence of LF in many areas around the world through MDA [6] . The problem of urban LF remains a matter of debate , yet it is recognized as being a key challenge to the ongoing global efforts to eliminate LF as a public health problem [7] . This is , in part , due to the challenges of translating MDA procedures that were developed for rural areas to the urban environment , with its distinctive patterns of human organization and behaviour , and in part due to the proliferation of potential urban vectors in poorly planned urban settlements [2] . In West Africa where Anopheles mosquitoes are the main vectors of LF and Culex plays little , if any , role in LF transmission , the status of urban transmission remains ill- defined as studies of LF transmission in cities in Sierra Leone and Guinea found no evidence of transmission [8 , 9] . Given the rapid expansion of urban centres in West Africa over recent decades , driven by population increases , insecurity and rural-urban migration the policy for treatment of these populations needs to be defined based on the evidence that transmission is ongoing and is likely to be sustained . In the State of Kano , Nigeria , LF MDA started in 2010 and as of 2014 , had reached 33 of the 44 Local Government Areas ( LGAs ) known to be LF endemic . Under the plans for 2015 LF MDA was to be extended to the remaining 11 LGAs including , six ‘urban’ LGAs . In 2010 , baseline ICT prevalence surveys conducted in these six urban LGAs revealed high ICT prevalence rates between 2 and 22% . Given such high pre-control endemicity levels added to the challenges of MDA implementation in urban settings , it may be useful to confirm the prevalence rates before embarking on MDA . The main goal of this study was to verify previous survey results carried out in Kano urban settings which suggested that LF is endemic . Entomological surveys were then conducted to determine the intensity of ongoing transmission in large cities in the light of recent findings that LF transmission is not sustainable in the capitals of Sierra Leone [8] and Guinea [9] despite the presence of antigen positives .
Ethical Clearance ( MOH/OFF/797/Ti/15 ) was given by the Kano State Ministry of Health Ethical Committee . Additionally , all surveys were conducted in accordance with the study protocol approved by the Institutional Ethics Review Board of the Liverpool School of Tropical Medicine ( 1189RS ) . All levels of leadership in the three Local Government Areas used for this study were met and their approval of the work was received . Written informed consent was obtained from all study participants above the age of 18 years , as well as parental consent from children below the age of 18 years . The children below 18 years provided oral consent for the study . Selection of study communities was based on a previous mapping by the Federal Ministry of Health , which indicated that urban LGAs were endemic for LF in Kano State . The mapping survey was undertaken in 2010 , following WHO recommendations [10] . Communities were selected in LGAs , based on key informant identification of communities most likely to be endemic for filariasis . Following community sensitization , 50–100 consenting volunteers were invited to a designated place where they were examined for W . bancrofti antigenemia using the ICT cards . The results of the mapping survey are presented in Table 1 . The current study was carried out in three of the urban LGAs included in the mapping survey ( Ungogo , Fagge and Nasarawa ) . The LGAs and study communities were randomly selected among those having the lowest and the highest antigenemia rates recorded during the mapping survey . However , in Nasarawa LGA Goji , which was peri-urban , was replaced by Gama ( an urban community ) . Jaba in Fagge LGA and Dankukuru in Ungogo LGA were also urban communities . In these LGAs neither LF MDA nor MDA for onchocerciasis based on ivermectin has been previously conducted . Kano State is located in Northern Nigeria , on latitude 110 30′ N—110 50′ N and longitude 80 30′ E—80 50′ E . The state is bound by four other states , all of which are endemic for LF- Katsina State to the North-west , Jigawa State to the North-east , Kaduna State to the South-west and Bauchi State on the South-eastern border . The status of Kano State as an LF endemic state has previously been established as a 1 . 6% prevalence was found in some villages [11] . A serological survey was carried out following the WHO guidelines [12] in May 2015 . Individuals living in the selected communities were sensitized and informed by community health workers and health officers of the urban LGAs about the survey . The sensitization was done in churches and mosques and other socio-economic groups . Additionally , a town crier was used in the selected communities to inform community members about the survey and individuals ≥5 years willing to take part in the survey were then invited to a designated community location ( schools and urban health center ) where finger-prick blood samples were taken . We used the immune-chromatography card test ( ICT ) ( Alere , NOW , ICT filariasis kits; Binax , Portland , USA ) for the detection of circulating filarial antigen in finger-prick blood samples taken during the day . All batches of ICT cards used were tested for quality control using serum from LF positive individuals , provided by the Noguchi Memorial Institute for Medical Research ( NMIMR ) . All test results were read after 10 minutes according to the manufacturer’s instructions . Mosquitoes were collected over a 4-month period from May to August 2015 , using exit traps , gravid traps and pyrethrum knock-down spray sheet collections ( PSC ) in different households . The use of these various trapping methods was to augment the number of samples collected and analyzed for the study . The estimated sample size was 1000 vector mosquitoes ( 100 pools of 20 mosquitoes per pool per site ) required in order to estimate an infection rate of 1% with a power of 0 . 80 [13] . Mosquito collections were done in the same communities where the parasitological surveys were undertaken . According to the information provided by the district health officers of the three urban communities , mosquito collection points were also selected based on the observation of risk factors such as presence of mosquito breeding sites , increasing potential exposure of the population to mosquito bites . Firstly , exit ( window ) traps , targeting host-seeking adults were installed in fifteen different randomly selected households , five in each LGA . Householders were asked to collect mosquitoes for 10-days , each morning , during the mosquito collection period—for which they received a small remuneration . This yielded a total of 600 sampling days ( 3 LGAs x 5 traps x 10 days x 4 months ) . Both exit traps and PSC were performed in different households on the same day . However , when mosquito catches were consistently low in any household , such households were replaced by others in the neighborhood . Pyrethrum spray catches ( PSC ) targeting resting adult mosquitoes were carried out inside the sleeping room of three randomly selected households ( different from those used for the exit trap ) in each site . Each month the collection period lasted for 7 days , therein yielding a total sampling effort of 252 sampling days ( 3 LGAs x 4 months x 7 days per month x 3 households per sampling day ) . A different set of three houses was sampled on each of the seven sampling days per month . The PSC method consisted of spraying an insecticide ( pyrethroid ) in selected rooms . Mosquitoes were collected on a white sheet covering the floor space of the sleeping room 10 minutes after spraying the rooms . Gravid trap collections were performed outside five randomly selected collection points in each LGA . Each month the collection period lasted 7 days therein yielding a total sampling effort of 420 days ( 3 LGAs x 4 months x 7 days per month x 5 traps / site ) . Gravid traps are highly efficient for sampling Culex species . The trap attracts females with an oviposition attractant medium contained in a pan below the trap [14] . Previous studies in West African cities have shown that the mosquito fauna is dominated by Culex species [8 , 9 , 15] , as a result of the polluted breeding environments . While Culex is considered as a non-vector species in West Africa [16] , it has been shown that parasite DNA can be identified in both vector and non-vector species after ingestion of parasite positive blood [17 , 18] . As such , identifying W . bancrofti parasite or DNA in any mosquito species is enough to reveal the presence of an infected individual . The use of Culex gravid traps was therefore to augment the chances of identifying infected mosquitoes in the study areas . Each month , during the entomology survey mosquitoes collected by each trapping method were taken to the nearest suitable field laboratory for further processing . Mosquitoes collected were identified to the genus and species level using morphological identification keys [19 , 20] . Following identification , the specimens were placed individually in coded vials and kept dry in tins containing silica gel for future processing . The mosquitoes were then sent to the NTD Reference laboratory of the Noguchi Memorial Institute for Medical Research , for molecular identification and processing for W . bancrofti infection . For the species identification of the members of the An . gambiae complex , a maximum of 100 mosquitoes were processed from each community . Where the number was less than 100 , all mosquitoes were processed . Genomic DNA was extracted from the legs of each mosquito , using the boiling preparation method; the legs were crushed in 100ml of distilled water and boiled at 95°C for 10 minutes . The supernatant was subsequently used as template for the PCR . Mosquitoes in the An . gambiae complex were then identified to species and molecular forms using the PCR method of Fanello et al . [21] . Mosquitoes were pooled according to community , trapping method and abdominal status . The pool range was 1–20 . DNA was extracted from each pool using the Qiagen DNEasy tissue kit . The identification of W . bancrofti DNA was performed using the LAMP method [22] and the conventional PCR method [23] , for confirmation . The LAMP method uses four primers that are simultaneously used to initiate DNA synthesis from the original unamplified DNA to generate a stem-loop DNA for subsequent LAMP cycling , and is thus faster , more sensitive and less susceptible to inhibition , compared to the conventional PCR method . We did not assess the infectivity rate of W . bancrofti stage specific L3 parasites . Only the presence of W . bancrofti DNA in mosquitoes was assessed , hence we cannot conclude on the status of transmission . All assays included a known positive control ( DNA extracted from a mosquito pool seeded with W . bancrofti microfilariae ) , and a negative control . All W . bancrofti positive pools were confirmed , by repeating the assay a second time . If a positive pool turn negative after repeat , the sample is re-run a third time for confirmation . If it is still negative , it is considered negative . The numbers of Anopheles and Culex mosquitoes were collated and calculated as percentages of the totals for site trapping method and month of collection . The pool screening software ( Version 2 . 0 ) was used to estimate the maximum likelihood estimates ( MLE ) of the prevalence of W . bancrofti infection in the mosquitoes [24 , 25] .
During this survey , 305 , 304 and 372 individuals were examined for circulating filarial antigen ( CFA ) from Fagge , Ungogo and Nasarawa , respectively , giving a total of 981 individuals ( Table 2 ) . None of the subjects had detectable circulating W . bancrofti parasite antigen . A total of 19 , 690 mosquitoes consisting of 2 , 600 An . gambiae , 17 , 082 Culex mosquitoes and 8 mosquitoes belonging to other species were collected ( Table 3 ) . The number of Anopheles mosquitoes collected was low in the dry season months of May , June and July , but peaked at the beginning of the rainy season in August . A proportion of the mosquitoes were dissected to determine infection with W . bancrofti . In all , 8 , 809 mosquitoes were dissected- 1324 An . gambiae and 7 , 485 Culex mosquitoes . All mosquitoes dissected to examine for larvae of W . bancrofti were negative suggesting that transmission was not taking place . Molecular identification of W . bancrofti identification revealed the presence of DNA in both An . gambiae and Culex mosquitoes . 602 pools of mosquitoes ( pool ranges of 1–20 ) were analyzed out of which 62 pools were positive for W . bancrofti DNA . Overall , higher prevalences were observed in Anopheles compared to Culex ( Table 4 ) . The prevalence of W . bancrofti DNA in Fagge was 0 . 69% in the An . gambiae and 0 . 64% in Culex spp . In Nasarawa , the prevalence of DNA in samples was 6 . 78% and 0 . 48% in An . gambiae and Culex , respectively , while in Ungogo it was 4 . 55% and 0 . 19% in An . gambiae and Culex , respectively . The data was also analyzed based on the trapping method used ( Table 5 ) . It can be observed that the use of the PSC and the gravid traps resulted in higher infection rates , compared to the used of the exit traps . Molecular analysis of samples of the An . gambiae s . l . showed that most members of the species complex in the three study communities were An . coluzzii , previously known as the M form of An . gambiae s . s . Only one An . gambiae s . s . and one An . arabiensis were identified ( Table 6 ) .
A cross-sectional ICT antigen detection survey carried out in May 2015 in the three urban LGAs of Kano state revealed that none of the 981 individuals aged ≥5 years had detectable levels of CFA . This contradicts the mapping results undertaken by the NTD program from the Federal Ministry of Health in the study area in 2010 ( 5 years earlier ) which reported prevalence of circulating filarial antigen varying between 2% and 12% in these LGAs . It will be important to know what factors have led to the reduction in LF prevalence from the mapping records or if the ICT tests available were recording false positives . It is known that vector control interventions have the potential to impact LF elimination activities [26 , 27] and the level of vector control interventions implemented between previous mapping and the current study could provide further evidence for LF elimination using vector control in Africa . In Kano state free distribution of insecticide treated nets was reported , which showed ITN ownership increasing more than 10-fold , from 6% to 71% [28] . It is further possible that the insecurity in northern Nigeria could have contributed significantly to the increase in the number of people migrating from rural areas to cities such as Kano . 3 . 3 million persons were displaced due of Boko Haram attacks from Borno , Yobe and Adamawa States to Kano in the last 5 years [29] . As a result it may be possible that the LF prevalence in urban Kano in 2010 may be a result of rural to urban migration of antigen positive individuals that are likely to have acquired the infection elsewhere . The two main mosquito species observed in the study sites were Culex quinquefasciatus and An . gambiae s . l . The higher proportion of An . coluzzii ( formally the M form of An . gambiae ) could be explained by the fact that these prefer ephemeral breeding sites , found more in urban and arid areas compared to An . gambiae s . s [30] . The study demonstrated the presence of An . gambiae s . l ( principally An . coluzzi ) and Culex quinquefasciatus specimens containing DNA of W . bancrofti in urban settings of Kano state . The prevalence of W . bancrofti DNA in Culex species from Fagge , Nasarawa and Ungogo LGAs were low ( ≤0 . 65% , 0 . 48% and 0 . 19% respectively ) compared to An . gambiae in which 0 . 96% , 6 . 78% and 4 . 55% DNA positivity were recorded respectively . Despite these levels of W . bancrofti DNA prevalence , we are unable to confirm transmission by An . gambiae and Culex quinquefasciatus as vectors of W . bancrofti in Kano because the detection of parasite DNA in mosquitoes does not necessarily imply that LF transmission is ongoing in a given setting . Studies have shown that mosquitoes that fed on people with very low levels of MF sometimes ingest MF but rarely produced infective larvae [27] . In Fagge LGA , a low DNA detection level ( 0 . 69% ) was recorded after processing 586 specimens of An . gambiae s . l . To confirm on-going LF transmission in Ungogo LGA where a high infection rate of W . bancrofti has been reported within An . gambiae species , the use of newer generation diagnostic tools , such as Wb123 [31] , might prove useful in the identification of current and active transmission in the untreated population . Similar to other investigations elsewhere [9 , 32] , this study does point to the fact that molecular xenomonitoring methods are superior to parasitological surveys in humans . Nonetheless , defining the elimination thresholds using molecular xenomonitoring requires defining at which prevalence estimate , in mosquitoes , transmission can be said to be interrupted . The observed prevalence of parasite DNA in the mosquitoes fall within the cut-off points of 0 . 25% , 0 . 5% and 1% suggested for Culex areas [33–35] , and 0 . 65% and 1 . 0% for Anopheles areas [13 , 36] . However , these cut-offs are only suggestive . The challenge therefore is to clearly define the elimination thresholds in vectors considering , vector control activities , vector species and abundance , biting rates etc . A limitation to this survey was the inability to attain the required sample sizes , and as such , the wide confidence intervals , especially in the Anopheles collections indicate the need for much larger sample sizes . This could be due to the period of mosquito collection , as the mosquitoes were collected mostly during the dry season . As such the timing of xenomonitoring surveys , barring programmatic difficulties , is very important to make meaningful assessments . While Anopheles species are considered the major vectors of W . bancrofti in West Africa , earlier studies in rural settings of Kano State revealed L3 W . bancrofti infection in Culex quinquefasciatus [11] . Other studies elsewhere in Nigeria also revealed high infection and infectivity rates in Culex mosquitoes [37 , 38] . The identification of DNA of the parasite in both mosquito genera goes to lend support to these earlier findings . Further , An . coluzzii is believed to be a more efficient vector of LF compared to An . gambiae s . s [39 , 40] , and the predominant nature of the former could imply a more efficient transmission of LF in the study locations . However , the promotion of environmental sanitation , coupled with vector control measures would go a long way in reducing LF infection and transmission . It is note-worthy that while the parasitological survey did not detect any infection , entomological surveys targeted at mosquito breeding sites identified infections , indicating these as being more sensitive in detecting any ongoing transmission . The parasitological surveys would also have identified infections in the community , if these were targeted at high-risk individuals such as those living in the vicinity of those breeding sites . Future assessments should therefore aim at identifying such individuals and stratifying the study communities so as to improve the outcome of parasitological surveys . The higher local prevalence of infection observed in the mosquitoes , compared to lower aggregated values may reflect the spatial heterogeneity and clustering of LF prevalence and transmission [41 , 42] . Analyzing disaggregated data to the trap level ( especially if the coordinates of trapping sites were taken ) might have been useful in identifying focal areas of transmission , considering that mosquitoes have limited flight ranges . This study also reveals the utility of analyzing mosquitoes that have recently fed in determining mosquito infection prevalence [13] . As can be seen from Table 4 , fed and gravid mosquitoes generally had higher infection levels compared to unfed mosquitoes . While some pools of unfed mosquitoes were found positive , parity dissection were not done in this study . This is because the mosquitoes sent to the NMIMR were stored dry on silica gel , and parity dissections are best done on fresh samples . Thus , the positive unfed mosquitoes may represent parous mosquitoes . Filaria DNA can be detected in mosquitoes that have processed W . bancrofti infected blood [43] and it is therefore important that molecular xenomonitoring studies aimed at analyzing unfed mosquitoes , limit these analyses to parous mosquitoes in order to minimize cost . Further the analysis based on the trapping methods ( Table 5 ) reveal that trapping of indoor resting mosquitoes using the PSC and gravid mosquitoes may be better tools in xenomonitoring than exit traps . Assessing W . bancrofti DNA within mosquito specimens is an indicator of the presence of infected humans and possible ongoing transmission [41 , 44] . However , confirming the transmission of LF requires the identification of infective stage larvae either through molecular methods or through dissection . While our results revealed widespread presence of W . bancrofti DNA in the mosquito population , we were unable to find a single positive mosquito given the number of mosquitoes dissected . Even though PCR is far more sensitive than dissection [13] , a limitation of this study is that the mosquitoes dissected for W . bancrofti larvae were not also subjected to the more sensitive molecular analysis . Such mosquitoes were teased , examined and discarded , and only un-dissected mosquitoes were shipped to the NMIMR for analyses . Additionally , assessing W . bancrofti DNA within mosquitoes is not a good indicator to confirm the presence of L3 larvae of LF parasite . But due to financial and logistic challenges related to assessment of the infectivity–L3 –rate ( because mosquito specimens should be stored with RNA processed through real time PCR technique ) we were not able to evaluate the infectivity rate . Molecular xenomonitoring , however , proved more effective in this study and has also been shown to be a promising tool in post-MDA surveillance [32 , 33] . In conclusion , despite the fact that infection of W . bancrofti DNA within Anopheles gambiae specimens was above 1% in some areas , the decline in antigen prevalence from the 2010 survey to the current survey led to the conclusion that LF transmission is not sustainable in the three urban LGAs of Kano state . While the mosquito biting rate could not be determined in this study , it is believed that the low numbers of An . gambiae collected is not sufficient to sustain transmission , as studies suggest that 2700 to over 100 , 000 infective bites are required for a new infection to be established [45] . The increased use and coverage of insecticide treated nets [28] , and other activities such as mosquito-proofing houses with screens and ceilings [46] , will lead to a reduction in the indoor densities of mosquitoes . These in turn will further decrease the vector-human contact in these urban areas , and ultimately the inefficient transmission of the parasite by mosquitoes . Based on these we could not recommend MDA . However , the Federal Ministry of Health advised that two rounds of MDA should be undertaken . The first round of MDA was undertaken in 2016 . This study demonstrated that entomological surveys are probably more sensitive indicators of the presence of infection . However , we could not demonstrate evidence for active transmission as we did not assess the infectivity rate for technical reasons . While no individual was found positive for antigen in the sample population surveyed , it is possible that antigen positive and potentially microfilaremic adults could potentially be contributing to transmission , especially since positive mosquitoes were detected . The use of the wb123 antibody assays may provide further evidence to the presence or absence of active transmission . Re-assessing the prevalence of W . bancrofti infection ( both in the human and mosquito population ) after a couple of years is recommended . | Mass drug administration ( MDA ) for the control of elephantiasis in the state of Kano in Nigeria , started in the year 2010 . It was estimated that by 2015 , the MDA programme will be extended to 11 remaining urban Local Government Areas ( LGAs ) . However , MDA in urban areas faces specific challenges , the most prominent being the need to achieve coverage rates of 65% and above . As such MDA alone may not be sufficient to achieve the required programme impacts of reducing LF transmission to levels below which transmission cannot be sustained , and additional interventions may be required . This study set out to confirm the LF infection prevalence in the human and mosquito populations in three urban LGAs in Kano . Individuals were tested for signs of the disease , and mosquito samples were collected and also tested for the worms that cause the disease . The study revealed that of 981 people tested , none had circulating filarial antigen in the blood . However , the mosquitoes collected revealed the presence of the disease-causing worms , but the level of infection was low . The infection in the mosquitoes was also detected in two different types of mosquitoes . Based on the outcomes of this study , and evidence from other West African cities on the transmission of LF , the Federal Ministry of Health recommended that two rounds of MDA be undertaken in urban areas of Kano . A further reassessment after a couple of years is warranted . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"geographical",
"locations",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"filariasis",
"neglected",
"tropical",
"diseases",
"molecular",
"biology",
"techniques",
"insect",
"vectors",
"lymphatic",
"f... | 2017 | Is mass drug administration against lymphatic filariasis required in urban settings? The experience in Kano, Nigeria |
The glycosylation of cell surface proteins plays a crucial role in a multitude of biological processes , such as cell adhesion and recognition . To understand the process of protein glycosylation , the reaction mechanisms of the participating enzymes need to be known . However , the reaction mechanism of retaining glycosyltransferases has not yet been sufficiently explained . Here we investigated the catalytic mechanism of human isoform 2 of the retaining glycosyltransferase polypeptide UDP-GalNAc transferase by coupling two different QM/MM-based approaches , namely a potential energy surface scan in two distance difference dimensions and a minimum energy reaction path optimisation using the Nudged Elastic Band method . Potential energy scan studies often suffer from inadequate sampling of reactive processes due to a predefined scan coordinate system . At the same time , path optimisation methods enable the sampling of a virtually unlimited number of dimensions , but their results cannot be unambiguously interpreted without knowledge of the potential energy surface . By combining these methods , we have been able to eliminate the most significant sources of potential errors inherent to each of these approaches . The structural model is based on the crystal structure of human isoform 2 . In the QM/MM method , the QM region consists of 275 atoms , the remaining 5776 atoms were in the MM region . We found that ppGalNAcT2 catalyzes a same-face nucleophilic substitution with internal return ( SNi ) . The optimized transition state for the reaction is 13 . 8 kcal/mol higher in energy than the reactant while the energy of the product complex is 6 . 7 kcal/mol lower . During the process of nucleophilic attack , a proton is synchronously transferred to the leaving phosphate . The presence of a short-lived metastable oxocarbenium intermediate is likely , as indicated by the reaction energy profiles obtained using high-level density functionals .
Protein glycosylation is known to play a pivotal role in many aspects of protein biochemistry , and there have been many examples where carbohydrate structures ( glycans ) carry out a significant biological function . [1–3] Glycans exist in a vast array of diverse structures built up from just a few small basic fragments . This can therefore be directly compared to the protein world , constructed purely from simple amino acids . However , in striking contrast to proteins , the structures of glycans are not encoded in any specific form analogous to the genome . [1] The so-called glycocode is just implicitly present in the regulation of hundreds of different highly specialized enzymes , glycosidases and glycosyltransferases , forming the glycosylation cascade . For this reason , understanding the reactivity of glycosyltransferases is essential to being able to decode the glycocode . Glycosyltransferases can be divided into two main groups based on whether they invert or retain the stereochemical configuration on the anomeric carbon . The reaction mechanism of inverting glycosyltransferases is well understood and both experiments and molecular modeling support a direct displacement SN2-like mechanism with a protein amino acid functioning as a catalytic base . However , the same level of understanding has not yet been reached for members of the retaining group . A lot of scientific attention has been recently focused on this issue in an attempt to determine the reaction mechanism of retaining glycosyltransferases , with mixed results . [4 , 5] Throughout the group of retaining glycosyltransferases , two main mechanisms were suggested to explain the reaction . The first of them is the double-displacement mechanism , where the reaction is thought to proceed via two consecutive configuration-inverting nucleophilic substitutions , first forming a covalent enzyme-carbohydrate intermediate and then transferring the carbohydrate onto the acceptor molecule . In this mechanism , a suitably positioned amino acid residue functioning as the catalytic base is required and two enzymes , namely α-1 , 3-galactosyltransferase [6] ( a3GalT ) and blood-group A and B α-1 , 3-galactosyltransferase [7] were proposed to proceed with this mechanism . Theoretical calculations on truncated QM models [8] predicted this mechanism to be energetically possible . Later QM/MM calculations [9–11] also supported this mechanism . However , there are many retaining glycosyltransferases that lack any residues that could serve as a nucleophile for the formation of the covalent intermediate . Therefore , the other possible reaction mechanism , the “internal return-like” , also called the SNi-like mechanism , has been suggested for these enzymes . This mechanism does not require a nucleophilic residue to be present . In this case , the reaction can proceed either as a concerted mechanism via an oxocarbenium ion pair transition state , or as a stepwise mechanism via a metastable intermediate that is subsequently captured by the acceptor nucleophile . [5] Compared with the double displacement mechanism , SNi substitution also seems to match the available kinetic isotope effect data . [12] The SNi-like mechanism was proposed for lipopolysaccharide α-1 , 4-galactosyltransferase C ( LgtC ) [13] and supported by theoretical studies . [4 , 14] Recent experimental evidence for the retaining glycosyltransferase , trehalose-6-phosphate synthase ( OtsA ) [12 , 15] is consistent with the SNi mechanism and also supports the theory that the hydrogen bond between the phosphate group and the acceptor hydroxyl plays a role in stabilizing the transition state suggested by calculations . [14] However , the existence of a short-lived intermediate remains an open question . Recently , several QM/MM theoretical studies [10 , 16 , 17] have been carried out in an attempt to shed some light on this problem . All three studies used a hybrid QM/MM model of the entire enzyme and came to the same general conclusion that the SNi-like mechanism is the most probable one . Unfortunately , due to the substantial approximations used in these studies , many unanswered questions about the validity of their results remain . The study on LgtC by Gómez et al . [10] used a static approach of QM/MM geometry optimisations constrained to points along a single predefined reaction coordinate , describing the difference in the lengths of the dissociating and newly forming C-O bond . Obviously , this completely neglects the second transfer process taking place at the same time—the transfer of a proton from the acceptor hydroxyl moiety onto a base represented by the leaving phosphate group . This , together with the low resolution of the scanned coordinate , led to a sudden jump of the proton upon crossing the main reaction barrier , indicated by a sharp spike in potential energy . In the end , the resulting energy profile does not describe a minimum energy path on a single potential energy surface , but a combination of two unconnected path segments corresponding to the endpoint locations of the proton . When our manuscript was being prepared for publication , Gómez et al . published another study [17] very similar to the LgtC one , focused on the ppGalNAcT2 glycosyltransferase . Although the conclusions presented there are again in agreement with theoretical expectations and available experimental findings , the ppGalNAcT2 study shares many of the methodological problems of the LgtC one . The authors have scanned potential energy along a single predefined reaction coordinate , using a very modest quantum-chemical description of the active site , namely the Becke-Perdew pure density functional together with a small basis ( SVP ) and a small quantum region ( 80 atoms ) . Just the basis set itself casts serious doubts on the usability of their results , as the authors themselves show that the related error in the potential energy barrier is at least 5 kcal mol−1 ( compared to TZVP basis ) , that is , about one third of the estimated barrier height . The influence of the simple density functional additionally seems to be of roughly the same magnitude . This can be related to the overall negative charge of the used QM region , as anionic systems are notoriously difficult to describe using pure density functionals due to a large self-interaction error . However , the most important shortcoming of the study in question is the fact that the authors were unable to find the transition state of the reaction , precluding any validation of the proposed reaction path . In contrast , the study on OtsA by Ardèvol and Rovira [16] took a more rigorous approach , sampling both the nucleophilic substitution and proton transfer processes by means of two independent collective variables ( CV ) . Their results are based on QM/MM Car-Parrinello molecular dynamics , using the metadynamics method to improve CV sampling and calculate free energy profiles , and support a single displacement with a two-step mechanism . [16] However , enhanced sampling methods like metadynamics provide correct free energy data only after the system reaches the regime of free diffusion along the reaction path . Unfortunately , computational resource constraints limit the achievable simulation lenght so severely that the free diffusion is essentially never reached . This leads to extremely noisy energy profiles , making unambiguous interpretation of the results obtained very difficult and their agreement with the expected reaction mechanism largely coincidental . In this work , we aim to describe the reaction mechanism of a retaining glycosyltransferase as thoroughly as possible , combining the results from two different approaches in order to leverage the advantages of both while avoiding their usual shortcomings . Multidimensional energy scans are able to provide an overall view of the potential energy surface ( PES ) , but often suffer from unsampled degrees of freedom leading to discontinuities that can pass undetected . On the other hand , minimum energy reaction path ( MERP ) optimization methods enable the sampling of a virtually unlimited number of dimensions , and thus guarantee that a single contiguous path will be obtained . However , there is no indication whether a given minimum energy path is the most probable and physically sound one . It can thus easily happen that a given MERP is deemed to be correct , even though an alternative path with a lower barrier exists in a different region of the PES . Such a situation is obviously impossible to detect without global information about the shape of the PES . By applying both approaches together and cross-checking the results , possible errors and artifacts can be easily identified . If the optimised MERP path is geometrically and energetically consistent with the PES , the possibility of discontinuities in the PES can be ruled out with confidence . At the same time , the shape of the PES can rule out the existence of alternative reaction pathways , validating the MERP . Unfortunately , although the idea of a combined approach is straightforward and its advantages are obvious , such a method is still not being ordinarily used to study enzymatic reactivity . Instead , studies based only on a single method with all its weaknesses are still very common . We chose polypeptide UDP-GalNAc transferase , human isoform 2 [18] ( ppGalNAcT2 ) as the subject of the study . This enzyme catalyses the first step in O-linked ( mucin-type ) protein glycosylation by transferring an N-acetylgalactosaminyl ( GalNAc ) group onto the serine or threonine hydroxyl moieties of an acceptor protein ( Fig . 1 ) . This glycosyltransferase exists in a large variety of isoforms exhibiting different spatial and temporal expression patterns and substrate specificities . [19] Increased activity of ppGalNAcT2 has been linked to the metastatic ability of various types of carcinoma , suggesting that targeted inhibition of a certain isoform could open the way towards selective anti-cancer drugs . [20] Detailed knowledge of the reaction mechanism and especially the transition state structure could then be used to design a potent inhibitor . Using the combined approach outlined before , we were able to obtain a reliable description of the reaction mechanism of ppGalNAcT2 , including a fully optimised structure of the main transition state .
The initial model was prepared from the X-ray structures of human isoform 2 [21] ( PDB: 2FFU ) and isoform 10 [22] ( PDB: 2D7I ) , where the former includes a short acceptor peptide EA2 and the UDP part of the donor molecule , and the latter includes a hydrolyzed UDP-GalNAc . In both cases , the protein consists of the main catalytic domain exhibiting the common GT-A fold and a C-terminal ricin-like lectin domain in a trefoil fold . [23] Both domains are connected by a flexible linker , and as such the lectin domain does not visibly interact with the catalytic domain . Because it is also experimentally known to not be essential for catalytic activity [24] , the lectin domain was cut off at conserved [25] proline 435 and not included in further studies . The native enzyme structure contains a manganese ion in the active site , coordinating the diphosphate fragment of UDP . However , manganese usually occurs in complexes in a high-spin state possessing 5 unpaired electrons . [26] Because this fact would entail a spin-unrestricted treatment of the active site , leading to an almost twofold increase in computational cost and possible convergence problems , we opted for replacing it with magnesium . Such a change has often been used in studies of similar enzymes to allow for spin-restricted calculations , based on tests by Kóňa and Tvaroška . [27] Although the general applicability of this replacement is uncertain , experimental data on the ppGalNAc transferase isoform 1 clearly show that 89% of its activity ( measured as kcat using deglycosylated ovine submaxillary mucin as a poly-acceptor substrate ) is retained when magnesium is used instead of manganese . [28] Computational results presented later in this work confirmed the applicability of this replacement . The system was described by a QM/MM model , where the QM zone consisted of 275 atoms treated by density functional theory at the OPBE-D3/TZP level . Initial geometry optimisation of the model led to a dissociated carbocationic state . The reactant and product structures were subsequently obtained by pulling the anomeric carbon towards the respective oxygen atom using a restraint and then fully optimising the geometry after releasing the restraint . The structures of the reactant and product complex after optimisation ( Fig . 2 ) can be described by the parameters shown in Table 1 . Based on this data , the initial 2D energy scan was done by scanning the C1-OA distance from 3 . 00 Å to 1 . 50 Å and the O1-H distance from 1 . 80 Å to 1 . 05 Å , both in steps of 0 . 15 Å . The resulting potential energy surface map is shown in Fig . 3 . It shows a large discontinuity in the location of the apparent barrier , caused by a dissociation of the GalNAc-phosphate C1-O1 bond ( S3 and S4 Figs . ) . This implies that the calculated surface is , in fact , an artificial combination of two separate fragments of the respective 2D potential surfaces for the bound and dissociated state of UDP-GalNAc . The 2D PES region that would normally connect these two fragments is completely missing due to the inadequate sampling of the aforementioned bond dissociation process . This situation precludes any further utilisation of the results of this scan . Attempts to correct for this problem by running a three-dimensional scan ( with the C1-O1 glycosidic bond length added as a third coordinate ) purely in the expected transition state region failed to locate a saddle point . This leads us to conclude that even the apparent position of the reaction barrier is incorrect because of the inadequate description of the reactive processes by the chosen scan coordinates . Unfortunately , running a three-dimensional scan spanning the whole area from reactant to product is not feasible , due to the required number of scan points needed to achieve satisfactory resolution . For this reason we opted for a different set of two scan coordinates . Coupled formation and dissociation of bonds can be described efficiently using distance difference coordinates . These are well known in the field of molecular dynamics , but they are not supported by common QM/MM software packages . After implementing them into the ADF program , we carried out another 2D energy scan , varying the nucleophilic substitution coordinate from 1 . 60 Å to −1 . 80 Å in steps of 0 . 20 Å , and the proton transfer coordinate from 0 . 80 Å to −0 . 30 Å in steps of 0 . 10 Å . This resulted in a smooth surface with no identifiable discontinuities , depicted in Fig . 4 . However , several isolated data points exhibited energy significantly different from their neighbors , caused by the relatively loose geometry convergence criteria applied in order to keep the computational cost manageable . To create a clearly understandable visualisation with physically relevant contour lines , these points were removed prior to visualisation when their energy differed by more than 2 kcal mol−1 from the average of four directly adjacent points ( for interior points ) or two adjacent points along the boundary ( for points on surface boundary ) . In total , 12 such points were removed for visualisation , amounting to only 5% of the total point count ( S5 Fig . ) . From the obtained PES map , we can predict a single important transition state between 16 and 18 kcal mol−1 above reactant energy , representing the nucleophilic attack well after dissociation of the GalNAc-phosphate bond . The extent of the saddle region delimited by the 16 and 18 kcal mol−1 contour lines is particularly noteworthy , as it is a clear sign of the relatively low curvature of the PES around the expected transition state . This low curvature makes direct identification of the TS candidates from the potential energy map difficult . We assume that this was the reason why attempts to optimize the transition state structure from only the scan results have been unsuccessful . It is also clear from the potential energy map that the proton transfer cannot serve as the initiating step of the reaction , because that would lead the system into the energetically unfavorable region in the top left corner of the map . Instead , the proton is spontaneously transferred during relaxation into the product minimum , as indicated by the low-energy region in the bottom left corner and the absence of a separate proton transfer barrier in the same region . The first phase of the reaction consists of the dissociation of the GalNAc-phosphate bond , corresponding to an increase in energy around d ( C1-OA ) −d ( C1-O1 ) = 1 . 00 Å . No clear barrier can be identified for this process , as it merely appears to form a shoulder of the main reaction barrier . NEB path optimisation from the initial approximation generated by restraint-based coordinate driving converged successfully in 100 path optimisation steps ( S8 Fig . ) . Projection of the initial and final paths into the distance-difference 2D map are shown in Fig . 4 . It is apparent that the overall path shape did not change during path relaxation . The potential energy of the individual images is depicted in Fig . 4 by the color of the path points and is clearly in reasonable agreement with the surrounding potential surface . Both facts ( the consistency of the path location and the image energies ) provide important evidence that the results obtained using both methods are not influenced by errors stemming from incorrect description of the reaction by the selected 2D scan coordinates or an unphysical initial path approximation . The overall energy profile along the NEB path shown in Fig . 5 A again exhibits the same main features found previously in the PES . A single very large barrier is present , rising to a maximum relative energy of 14 . 1 kcal mol−1 at image 20 , followed by a steep yet smooth decline to the product minimum 6 . 7 kcal mol−1 below the reactant energy . The predicted barrier height of approx . 14 kcal mol−1 is in very good agreement with the phenomenological free energy barrier of approx 17 kcal mol−1 , that can be calculated using transition state theory from the experimentally determined kcat value of 3 . 70 s−1 . [21] Additionally , the SNi mechanism observed in our study is also supported by experimental kinetic isotope effect data . [12] To get a clearer picture of all the key processes taking place during the reaction , we can analyse the evolution of key bond lengths presented in Fig . 5B . The first phase consists of a dissociation of the C1-O1 glycosidic bond , covered by images 2–5 . The distance of the attacking nucleophile does not change appreciably during this event , showing that the nucleophile is not directly involved in initiating it . On the other hand , the hydrogen bond between threonine hydrogen and phosphate oxygen shortens visibly by about 0 . 2 Å , as this bond is made stronger by the increased negative charge on the oxygen atom after the heterolytic cleavage of the C1-O1 bond . The following path segments up to image 20 describe a phase of significant spatial rearrangement of the reacting species with no changes to their bonding pattern . The length of the ( now dissociated ) C1-O1 bond increases as the phosphate leaving group relaxes to a less strained position than the one at the start of the reaction . This gradual separation of the oxocarbenium ion—leaving group pair is probably the main reason for the gradual rise in energy , creating the nearly flat top of the barrier . At the same time , the nucleophile hydroxyl moiety is slowly inserted between the anomeric carbon and phosphate , as shown by the decreasing C1-OA distance . After crossing the saddle point region at image 20 , the energy starts to decrease and from image 23 to 25 , key changes in bonding take place: A new C1-OA glycosidic bond forms between the acceptor and GalNAc , as indicated by the C1-OA distance decreasing from 2 . 0 to 1 . 5 Å . The proton is transferred to the phosphate , while maintaining an exceptionally strong hydrogen bond to threonine with a bond length of only 1 . 34 Å . The phosphate moves back 0 . 3 Å closer to the GalNAc , probably thanks to the decreasing repulsion with the nucleophile oxygen . In the last five images , the energy decreases as the system releases the conformational strain and the bond lengths relax to their equilibrium values . To assess the impact that the approximations taken in this study might have on the validity of the calculated reaction energy profile , we calculated single-point energies on the optimized image geometries using several different methods . Replacing the magnesium ion with the natural manganese ion in high-spin configuration together with a spin-unrestricted calculation does not alter the energies significantly and the shape of the energy profile is almost entirely retained ( Fig . 5A ) . The differences are mainly noticeable in the region of the saddle point and product minimum , where the overall barrier height is lowered by 0 . 8 kcal mol−1 to about 13 . 3 kcal mol−1 . The sign and magnitude of this energy difference agrees well with the experimentally observed difference [28] in reactivity for magnesium and manganese , although considering the accuracy limits of the computational methods employed , this could be just a coincidental agreement . Similarly , single-point energies for the profile were recomputed with a larger basis set to check for possible basis incompleteness issues . The profile calculated using the QZ4P basis [29] is qualitatively unaltered , exhibiting only a slightly increased barrier height , by 1 . 8 kcal mol−1 . Because a proper description of the reaction is wholly dependent on the performance of chosen density functional , energies were recomputed using several functionals to find possible artifacts . Although any density functional can exhibit its own share of problems , it is much less probable that a given artifact would be present in the data calculated using several different density functionals . Additionally , more complex density functionals are inherently more accurate because they are based on fewer approximations in various energy terms . For example , hybrid density functionals suffer much less from artificial locality and self-interaction error than GGA functionals , thanks to the use of the exact electron exchange term in hybrid functionals . Further improvement in accuracy is available by also including the kinetic energy density term , forming the so-called group of meta-hybrid density functionals that represent essentially the best QM methods usable for systems with hundreds of atoms . The results obtained using the hybrid PBE0 functional [30 , 31] agreed well with the original OPBE ones , but a shallow minimum corresponding to a metastable oxocarbenium intermediate was now visible . The same conclusion could be drawn from a very similar profile calculated by the state-of-the-art meta-hybrid M06-2X density functional [32 , 33] . This observation is similar to the findings of the OtsA study , which predicted a single displacement but a two-step reaction . [16] On the other hand , the commonly used Becke-Perdew GGA density functional [34 , 35] completely failed to provide a physically sound energy profile ( S7 Fig . ) . Unfortunately , as the minimum and thus also the preceding barrier is only present in the energy profiles calculated by hybrid density functionals , geometry optimisation of the respective stationary points would be extremely computationally demanding for a QM region consisting of 275 atoms . This is in contrast with the three confirmed stationary points that were successfully optimised using the much faster OPBE functional . For this reason , we selected image 6 as a representative structure of the first transition state and image 9 as the intermediate . Both structures are presented in Fig . 6 . Additionally , even though a minimum is present , it is only 0 . 7 kcal mol−1 deep , explaining the extremely short lifetime of the intermediate . Finally , it has to be noted that the results are based purely on potential energy data while the real physical process is controlled by free energy . The depth of the minimum may thus be significantly affected by the entropic effects included in free energy . The structure of the main transition state was refined by optimising the structure of image 20 from the NEB path along the first eigenvector of an approximate numerical Hessian . After reaching convergence , the transition state depicted in Fig . 6 was obtained . Structural changes during transition state optimisation were very small , with changes in the key distances being on the order of 0 . 01 Å . The final geometry of the second transition state thus has the same major features as NEB image 20 . The proton is still attached to the acceptor oxygen , but at the same time it participates in a very strong hydrogen bond with the leaving group . The carbohydrate ring is in an envelope conformation with a partial half-chair character ( S9 Fig . ) , and a new C1-OA glycosidic bond is being formed . Its length in the optimised transition state is 2 . 33 Å , almost exactly matching the length of the dissociating C1-O1 glycosidic bond in the estimated structure of the first transition state . This observation supports the previously proposed concept [5 , 12] of the two transition states involving each glycosidic bond being very similar , almost “mirror images” of each other . To verify the correctness of the obtained transition state , a full QM/MM numerical Hessian was subsequently calculated . It contains exactly one negative eigenvalue in both the non-mass-weighted and mass-weighted ( normal mode ) coordinate systems , confirming that the structure corresponds to a first-order saddle point . The calculated imaginary frequency of this normal vibration mode is 96i cm−1 , in line with the previously observed low curvature of the PES . Visual inspection of the normal mode motion confirmed that it represents the nucleophilic substitution process . We have identified several interactions that probably play a key role in facilitating the reaction . These can be divided into three main groups: interactions with structural role ( enforcing a proper relative positioning of the substrates ) , those stabilizing the positive charge on the GalNAc oxocarbenium ion and finally interactions stabilizing the negative charge on the diphosphate moiety of the UDP leaving group . Among the structural interactions there are several hydrogen bonds coordinating the GalNAc moiety: a bond between the amidic backbone hydrogen of Gly309 and the N-acetyl carbonyl oxygen , two hydrogen bonds between Glu334 and the O4 and O6 hydrogens of GalNAc and a hydrogen bond between Arg208 and the O4 GalNAc oxygen . All of those interactions keep the GalNAc moiety rotated around the glycosidic bond towards the diphosphate group , exposing the glycosidic bond and C1 atom to a nucleophilic attack by the acceptor . The positive charge that develops on C1 after bond dissociation interacts in a charge-dipole manner with the carbonyl group of Ala307 , a member of a loop covering the β-face of GalNAc . Apart from the electrostatic effects of the metal cation and hydrogen bonding with the water molecules coordinated to this ion , the leaving group is additionally forming a strong hydrogen bond with Tyr367 , a hydrogen bond with Arg362 and another , relatively weak hydrogen bond with the amidic hydrogen of the acceptor threonine . There is also an important intramolecular hydrogen bond between a phosphate oxygen and the amidic hydrogen of GalNAc that further contributes to keeping the saccharide moiety suitably rotated . However , probably the most interesting of all the active site residues is Trp331 . It plays a double role: Forms a CH-pi interaction with the C6 hydrogen atoms of GalNAc and at the same time donates a hydrogen bond to the phosphate oxygen participating in the original glycosidic bond . This hydrogen bond is quite weak in reactant ( having a length of 2 . 69 Å ) , but grows much stronger after dissociation of the glycosidic bond , achieving its minimum length of 1 . 76 Å around the transition state and then getting weaker again after the nucleophilic capture ( 2 . 02 Å in product ) . All of the enzyme residues participating in these interactions are highly conserved and were experimentally identified as being crucial for preserving reactivity ( Table 2 ) . In this study , we have shown that human isoform 2 of the polypeptide UDP-GalNAc transferase catalyses a same-face nucleophilic substitution with internal return ( SNi ) . The optimized transition state for the reaction is 13 . 8 kcal mol−1 higher in energy than the reactant , while the energy of the product complex is 6 . 7 kcal mol−1 lower . This corresponds to a dissociated oxocarbenium state just before its nucleophilic capture by the acceptor threonine oxygen . During the process of nucleophilic attack , a proton is synchronously transferred to the leaving phosphate . By coupling two different QM/MM-based approaches for investigating the reaction mechanism , namely a PES scan in two distance-difference dimensions and a MERP optimisation using the NEB method , we were able to rule out the most significant sources of potential errors . We can therefore conclude that the observations based on theoretical modeling are in good agreement with available experimental evidence [12] , including the recent X-ray structures based on modified substrates . [36] The reaction starts with a dissociation of the C1-O1 glycosidic bond of the donor UDP-GalNAc . The barrier of this step is lower than 10 kcal mol−1 and is thus hidden under the barrier of the rate-determining step . However , the presence of a short-lived metastable oxocarbenium ion is likely , because the corresponding energy minimum is visible in energy profiles obtained using higher-level density functionals . On the other hand , the minimum is only ca . 1 kcal mol−1 deep , and such a subtle energy difference is at the limit of the accuracy provided by applicable theoretical modeling approaches . Additionally , the stability of the intermediate can be affected by other entropic and environmental phenomena not considered here . We have shown that distance-difference coordinates provide an exceedingly useful tool for the description of common reactive processes , and are certainly useful for the wider scientific community interested in mapping reaction potential energy surfaces . Additionally , the Nudged Elastic Band method for MERP optimisation is suitable for a rapid exploration of reaction pathways , and it is immune to the coordinate sampling problems common in static energy mapping . However , knowledge of the shape of the PES is necessary to ensure that a physically relevant path is selected for optimisation .
The X-ray structures of ppGalNAcT isoforms 2 ( PDB ID: 2FFU ) and 10 ( PDB ID: 2D7I ) were superimposed using Accelrys Discovery Studio Visualizer 3 . 1 to minimize RMS distance between corresponding C-alpha atoms of the catalytic domain ( S1 Fig . ) . The final RMSD was 0 . 88 Å for C-alpha atoms and 1 . 33 Å for all protein atoms . Visual inspection of the active site showed near perfect overlap of the UDP molecules and neighboring side chains , allowing the coordinates of GalNAc to be directly transferred from 2D7I into 2FFU . Afterwards , hydrogen atoms were added to the structure using the pdb2adf tool from the ADF [37] suite . First , the required fragment file for UDP-GalNAc was generated by the antechamber tool from AmberTools 1 . 4 [38] with partial atomic charges calculated by the AM1-BCC method [39] to give a total charge of −2 . The protonation and oxidation states of relevant protein residues were assigned automatically by pdb2adf based on their chemical environment and visually checked for correctness . The protonation state of histidine 359 was manually overridden to HID as the automatically generated one ( HIE ) was obviously nonsensical . This residue is a ligand of the metal cofactor and therefore needs to have the N-ε atom unprotonated . Furthermore , water molecules present in the active site were manually rotated to create a network of hydrogen bonds where possible . The QM region was defined to include the essential parts of the substrates and residues experimentally known to be crucial for reactivity . The UDP donor was represented by methyl diphosphate , as the ribose and uracil parts of the molecule are quite far from the reactive site . The acceptor threonine 7 of the EA2 peptide was included together with its direct neighbors , threonine 6 ( excluding its amine group ) and proline 8 ( excluding its carboxyl group ) . Finally , 12 highly conserved residues interacting with the substrates were included ( Table 2 ) as well as six water molecules that were well defined in the crystal structure ( B-factors below 20 ) . Three of these water molecules are located close to the metal ion with one of them directly serving as a ligand and the other two forming a hydrogen bond network between the first water molecule and neighboring active site residues . The other three water molecules are located approximately in the plane of the pyranose ring or slightly towards its beta-face ( the face opposite to the glycosidic bond ) . These molecules form a hydrogen bond network , acting as hydrogen bond donors to the GalNAc O5 oxygen and the acceptor threonine oxygen . However , they are separated from the leaving diphosphate group by the carbohydrate and threonine moieties and thus cannot directly participate in the reaction as catalytic acids or bases or proton transfer mediators . The resulting system contains 252 real ( non-capping ) atoms in the QM region and 6051 atoms in total . All QM/MM calculations were carried out using the Amsterdam Density Functional package [37 , 40 , 41] in versions 2012 . 01d ( used only for the 2D scans and M06-2X single point calculations ) and 2013 . 01 . The NEWQMMM implementation of molecular mechanics in ADF was employed to manage the MM part of the system , described by the AMBER ff94 forcefield [42] combined with GLYCAM06 parameters [43] on the GalNAc group . The AddRemove QM/MM coupling scheme [44] was used and charges on the QM atoms were updated in every geometry iteration from the MDC decomposition [45] up to the dipole level . Hydrogen capping atoms were added by the AddRemove scheme to saturate link bonds crossing the QM/MM boundary , bringing the overall QM atom count to 275 . The QM part was described by density functional theory at the generalized gradient approximation level using the OPBE functional ( a combination of the OPTX optimized exchange functional by Handy and Cohen [46] and Perdew-Burke-Ernzerhof [47 , 48] correlation functional ) . This functional has been shown to be the best GGA functional for describing nucleophilic substitution reactions . [49] In tests , it provided results qualitatively similar to the M06-2X meta-hybrid density functional [32 , 33] . The description of weak interactions was augmented by the DFT-D3 empirical dispersion correction [50 , 51] in “zero-damping” form . All calculations were carried out using the all-electron Slater-type TZP basis [29] with the charge fitting set distributed with ADF . Two types of numerical quadrature grids were employed to evaluate the electrostatic and exchange-correlation potential . A Becke grid integration scheme [52 , 53] was used for all calculations using ADF 2013 . 01 , with resolution given by the “Normal” preset . A Voronoi cell based integration method [54] was used in the energy scans and M06-2X single point calculations , because the newer Becke scheme is not available in ADF 2012 . 01d . The number of integration points for this method was automatically determined by ADF to meet predefined accuracy level 4 for PES scans or 6 for M06-2X calculations . To reduce random integration noise in the gradients and prevent it from spoiling the Hessian estimates , a smoothing method based on conservation of the Voronoi cells and integration points across the geometry steps was applied [53] . Because the QM/MM implementation in ADF treats MM as a perturbation to the QM system , full convergence of molecular mechanics is required in every QM geometry step . This was ensured by optimising the MM system by the scaled conjugate gradient method [55] until all MM gradient vector components decreased below 0 . 01 kcal mol−1 Å−1 . SCF optimisation of the QM region was stopped when the maximum element of the commutator of the last two Fock matrices decreased below 10−5 au . Two-dimensional potential energy scans were carried out using two sets of scan coordinates , always starting from the optimized reactant structure . In the first scan , two distance coordinates were used: d ( C1-OA ) and d ( O1-H ) . The second scan used two distance differences instead to provide a better description of the respective processes: d ( C1-OA ) −d ( C1-O1 ) describing the nucleophilic substitution and d ( O1-H ) − ( OA-H ) describing the proton transfer . All scan coordinates were implemented using restraints . Support for distance difference restraints was implemented into ADF and subsequently added to the mainline distribution . In both scans , first the respective nucleophilic substitution coordinate was scanned from the reactant value to the ( approximate ) product value with the second coordinate frozen . Afterwards , the second coordinate was scanned to the product value , generating the second scan dimension . The optimisation of each point started with the geometry , charges and MO coefficients of the preceding point in a given scanline and proceeded using a quasi-Newton optimizer with Broyden-Fletcher-Goldfarb-Shanno Hessian updates until the maximum gradient component decreased below 0 . 01 Hartree Å−1 . The minimum energy reaction path was described using the Nudged Elastic Band ( NEB ) approach with improved tangent estimates [56 , 57] . The algorithm was based on the ASE [58] Python toolkit coupled to ADF for MM optimisation and gradient evaluation . Cartesian coordinates of all 252 real QM atoms were used to describe the configuration of each NEB image , leading to an optimisation of the reaction path in the full 756-dimensional space without any a priori assumptions regarding the reaction coordinate . Just as in the case of PES scans , the positions of all MM atoms for each image were fully optimised before every gradient calculation . The path was discretized into 30 images , where the reactant and product endpoints were kept fixed and the rest was optimized simultaneously using the FIRE algorithm [59] . This algorithm was selected because quasi-Newton algorithms do not work well with NEB , both due to the high dimensionality ( 28 images × 756 coordinates each = 21 , 168-dimensional optimisation space ) and especially because the NEB Hessian matrix is not symmetric [60] and therefore can not describe the locally quadratic surface assumed by quasi-Newton algorithms . The FIRE algorithm is the most sophisticated algorithm implemented in ASE that is able to deal with this problem . All internal optimizer parameters were kept at their default values . We observed a significant instability of the pure NEB path in regions where the PES is relatively flat ( especially in the reactant and product basins ) , leading to the formation of “kinks” in the path and subsequent divergent lengthening of the affected path segments . A partial perpendicular force term [56] was added to keep the path stable and smooth . The fraction of perpendicular force included was determined by f ( Φ ) = 1 2 1 + cos π cos Φ where ϕ denotes the angle between adjacent path segments . Although this means that the result is no longer a rigorously correct minimum potential energy path , but just an approximation of it , the difference is minimal , and it is only visible in regions that are not particularly interesting in terms of the reaction mechanism . The force constant k for NEB springs was set to 5 eV Å−1 . Path optimisation was stopped when the maximum element of total NEB gradient decreased below 0 . 0025 Hartree Å−1 . Initial approximation of the minimum energy reaction path was created by manually picking 5 points ( pairs of distance difference values ) from the potential energy surface and applying spline interpolation to obtain 21 distance difference pairs uniformly spaced along the curve . Structures were then generated by successive optimisation starting from the reactant structure with the two distance difference coordinates restrained to the values corresponding to a given point . All these structures were directly used as NEB images . Because the NEB approach outlined above forces the images to be equidistant with respect to their 756-D Euclidean distance , additional images were added by linear interpolation between the obtained structures until all path segments were shorter than 0 . 25 Å . To keep the implementation simple , only the positions of QM atoms were interpolated; the positions of MM atoms and the QM charges were directly copied from the nearest parent structure . The final transition state structure was obtained by optimising image 20 of the converged NEB path using a quasi-Newton optimizer with Bofill’s Hessian update scheme [61] and a gradient convergence criterion of 0 . 001 Hartree Å−1 . Full numerical Hessian from a preceding calculation on image 23 was used as the initial Hessian for the optimisation . After reaching convergence , a full numerical Hessian of the total QM/MM energy was calculated for the optimized structure using symmetric central two-point finite differentiation of gradients with a step of 0 . 01 Å . Cremer-Pople conformational parameters [62] for the carbohydrate ring were calculated using the cp . py script by Hill and Reilly [63] . | Cell surface proteins are covered by a diverse array of glycan structures , important for mutual cell recognition and communication . These glycans are complex branched molecules assembled from monosaccharide units by a sophisticated cascade of enzymes from the group of glycosyltransferases . Disruptions in the synthesis of glycans are linked to various diseases with the most prominent example being cancer . To understand or control the process of glycosylation , the reaction mechanisms of the participating enzymes need to be known . Here we investigate the catalytic mechanism of human glycosyltransferase ppGalNAcT2 using the tools of computational chemistry . By modelling the crucial parts of the enzyme using a quantum mechanics-based description , we are able to trace the whole reaction path leading from the reactant state to the product state . Our results provide a reliable description of the motion of all important atoms during the reaction and they are fully consistent with available experimental data . The insights obtained in this study can be further used to design a potent inhibitor molecule , usable as a potential drug for diseases involving increased activity of the enzyme . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [] | 2015 | Stepwise Catalytic Mechanism via Short-Lived Intermediate Inferred from Combined QM/MM MERP and PES Calculations on Retaining Glycosyltransferase ppGalNAcT2 |
Plants have developed sophisticated systems to monitor and rapidly acclimate to environmental fluctuations . Light is an essential source of environmental information throughout the plant’s life cycle . The model plant Arabidopsis thaliana possesses five phytochromes ( phyA-phyE ) with important roles in germination , seedling establishment , shade avoidance , and flowering . However , our understanding of the phytochrome signaling network is incomplete , and little is known about the individual roles of phytochromes and how they function cooperatively to mediate light responses . Here , we used a bottom-up approach to study the phytochrome network . We added each of the five phytochromes to a phytochrome-less background to study their individual roles and then added the phytochromes by pairs to study their interactions . By analyzing the 16 resulting genotypes , we revealed unique roles for each phytochrome and identified novel phytochrome interactions that regulate germination and the onset of flowering . Furthermore , we found that ambient temperature has both phytochrome-dependent and -independent effects , suggesting that multiple pathways integrate temperature and light signaling . Surprisingly , none of the phytochromes alone conferred a photoperiodic response . Although phyE and phyB were the strongest repressors of flowering , both phyB and phyC were needed to confer a flowering response to photoperiod . Thus , a specific combination of phytochromes is required to detect changes in photoperiod , whereas single phytochromes are sufficient to respond to light quality , indicating how phytochromes signal different light cues .
Plant photoreceptor signaling networks are sensitive to a large dynamic range of light inputs . Plant light signaling systems are sensitive enough to induce germination in response to extremely short exposures of light , as encountered during soil tillage , and very low light intensities , as experienced under soil litter , and yet able to detect subtle variations in light quality under full sunlight or changes in photoperiod . These abilities depend partly on the existence of multiple photoreceptor families with differential spectral properties and on the sub-functionalization of photoreceptor family members , which resulted in the emergence of photoreceptors with distinct properties and the capacity to interact to modulate light sensitivity [1 , 2] . In the model plant Arabidopsis ( Arabidopsis thaliana ) , thirteen different sensory photoreceptors have been characterized to date , with absorption spectra ranging from UV-B ( 280 nm ) to far-red light ( 730 nm ) ( FR ) [3] . The phytochromes are a family of red light ( R ) and FR photoreceptors consisting of five members ( phyA-phyE ) [4] , and plant development under natural conditions depends on their ability to toggle between the Pr ( inactive ) form , which absorbs R , and the Pfr ( active ) form , which absorbs FR . The phytochromes are synthesized in the Pr form and photoconverted to the Pfr form after absorption of R . FR can then convert Pfr back to Pr . Thus , the proportion of phytochromes in the Pfr form is a function of the R to FR ratio . As plants absorb R for photosynthesis , but reflect FR , a decrease in the R/FR ratio indicates the presence of neighboring vegetation , which at some point may compete for light resources [5] . Shade-intolerant plants respond to low R/FR ratios by elongating their stems and petioles , to outcompete their neighbors , and accelerating flowering , to ensure their reproductive success . These responses are known collectively as the shade avoidance syndrome ( SAS ) [1] . Phytochromes regulate plant development throughout the life cycle , from germination to flowering . For instance , phytochromes promote germination by stimulating gibberellin ( GA ) synthesis and sensitivity [6 , 7] , promote deetiolation during early seedling development , inhibit hypocotyl and stem elongation by altering auxin levels [8] , entrain the circadian clock , and regulate flowering time [2] . Phytochromes exist as homodimers or heterodimers , which are translocated to the nucleus upon photoconversion to the Pfr form [9] . Upon light exposure during seedling establishment , phytochromes alter the expression of thousands of genes [1 , 2 , 10] . They induce these complex responses by interacting with members of a family of basic-helix-loop-helix ( bHLH ) transcription factors , the PHYTOCHROME INTERACTING FACTORS ( PIF ) . The PIFs are repressors of germination and deetiolation and are degraded by proteasomes upon interaction with phytochromes [1 , 2 , 7] . The roles of phytochromes were mostly studied in the model systems Arabidopsis and rice ( Oryza sativa ) , using photoreceptor mutants [11–17] . Given their similar spectral properties , it is not possible to dissect the roles of each phytochrome in plant development without using genetic tools . Studies have mainly analyzed single phytochrome mutants , and few have examined higher order mutant combinations [2] . Although these studies have been fundamental in establishing which phytochromes contribute to which developmental responses , they failed to pinpoint which particular phytochrome activates which developmental program or pathway . Further , cross-talk has been reported to exist between phytochrome signaling pathways [15 , 18 , 19]; however , it is unclear if the identified interacting phytochromes are sufficient for the interaction or if other phytochromes are also required . Increased complexity may also be expected , because heterodimers form among phytochrome family members [20 , 21] . In studies in which phytochromes were overexpressed , other phytochromes were present in the genetic background , raising questions as to whether the signal was generated by single or multiple phytochromes [22–25] . To resolve these issues , we used a bottom-up approach , similar to that employed by Coen and Meyerowitz to generate their ABC model of plant flower development [26] . They used a mutant background devoid of the three kinds of flower development genes ( A , B , and C ) to which they “added” each gene separately and then by pairs to dissect the roles of each gene and to decipher the mutual interactions that led to flower development . Using a similar approach , we obtained the whole set of quadruple and triple phytochrome mutants in the same genetic background ( i . e . , Columbia ) . In this way , we “added” each phytochrome alone and each possible pair of phytochromes to a phytochrome-less background to study the effects of each individual phytochrome in isolation and to examine all pair wise interactions , both direct and indirect . Our work reveals the distinct roles of each phytochrome and identifies how the phytochromes interact with each other , thereby revealing novel properties of the phytochrome signaling system that are necessary for regulating the photoperiodic response .
After obtaining the whole set of phytochrome quadruple and triple mutants in the Columbia background , we evaluated their germination under different light conditions . Under continuous white light ( WL ) , phyA , phyB , or phyD alone was sufficient to induce germination , phyE was a poor inducer of germination , and phyC did not induce germination . The quintuple phytochrome mutant failed to germinate under WL ( S1 Fig ) . Under R and FR treatments , only phyB produced a R/FR reversible response , while phyA produced a response to continuous FR and to a single FR pulse ( the so-called very low fluence response , VLFR ) ( Fig 1 ) , as expected [18] . As under WL , phyC did not induce germination at all under continuous FR or FR pulses and phyE only rarely did so . The simplest explanation for the low activity of phyC is the inability of the PHYC apoprotein to accumulate in the absence of phyB [20 , 27] ( S2 Fig ) . On the other hand , phyD promoted germination in response to continuous R but not to single R pulses ( Fig 1 ) . The requirement of continuous R by phyD is consistent with previous observations in phyD overexpressing lines which also required continuous R to inhibit hypocotyl elongation [24] . Triple mutant combinations allowed us to study phytochrome interactions . Despite its negligible role as an inducer of germination when present alone , phyE acted synergistically with phyC and with phyD to induce germination after exposure to continuous R or to a single pulse of R ( Fig 1 ) . Further , phyE acted synergistically with phyB at low R:FR ratios ( i . e . , during FR treatments ) , consistent with previous reports [6 , 19] . phyC , phyD , and phyE interacted synergistically with phyA under continuous FR , a treatment that specifically activates phyA [18] . When we evaluated the GA sensitivity of the various mutants , a different picture emerged . phyB and phyA were the most important positive regulators of GA sensitivity ( S1 Table , S3A and S3C Fig ) and were both antagonized by phyC , phyD , and phyE ( S3A Fig ) . These results indicate that phyC , phyD , and phyE not only promote germination by acting synergistically ( Fig 1 ) , but also antagonize germination by reducing the effects of phyA and phyB on GA sensitivity ( S3A and S3C Fig ) . This “gas and brake” behavior could be important for regulating GA signaling homeostasis once seeds germinate , as GA and light antagonize each other during seedling emergence [1] . As mentioned above , accumulation of PHYC apoprotein depends on phyB ( S2 Fig ) [20 , 27] . We also found increased PHYA levels in the presence of phyC , phyD or phyE in light-grown seedlings ( S2 Fig ) . Therefore , the synergistic interactions in the promotion of germination may be explained , to some extent , by increased photoreceptor levels . Conversely , the antagonistic interactions that decrease GA sensitivity suggest more specific roles for phyC , phyD , and phyE . To evaluate how each phytochrome pathway influences flowering and is influenced by temperature , we measured flowering time for all genotypes under long-day ( LD ) conditions , at temperatures of 18 to 24°C ( Fig 2A and S4A Fig ) . The quintuple phytochrome mutant was poorly responsive within this range of temperatures ( slope: -0 , 365±0 , 217 leaves/°C ) . Surprisingly , phyE was found to be the strongest repressor of flowering and its effect was stronger at lower temperatures ( slope: -2 , 580±0 , 429 leaves/°C Fig 2A ) . A similar effect was observed for phyB ( slope: -2 , 121±0 , 389 leaves/°C Fig 2A ) . The temperature-dependent effects of phyE and phyB underscore the interaction between the light and temperature signaling pathways ( Fig 2A ) . Conversely , phyD was a weak flowering repressor under all conditions tested , showing that phyD and phyE have distinct roles , with the former being more effective in promoting germination and the latter more effective in influencing flowering ( Figs 1 and 2 ) . phyA behaved as a weak repressor of flowering under LD at all temperatures tested ( Fig 2A and S4A Fig ) . However , when combined with other phytochromes , phyA emerged as a strong antagonist , mainly by antagonizing phyE signaling . The presence of phyA eliminated most of the temperature responsiveness of plants bearing phyE alone , but not phyB , suggesting that signaling downstream phyE might be different , at least in part , from that downstream of phyB , and that some temperature effects might also be specific to phyE . phyA also antagonized the relatively weak role of phyD as a flowering repressor . These data suggest that the well-known role of phyA in promoting flowering [28] results , at least in part , from its antagonism of phyE and phyD signaling ( Fig 2 ) . This phyA effect is unlikely to be due to lower PHYD or PHYE levels in the presence of phyA ( S2 Fig ) . phyC had negligible effects when present alone . Similar to phyA , phyC antagonized the action of phyE , but contrary to phyA , also antagonized phyB at the lower temperatures ( Fig 2A and S4A Fig ) . Taken together , these results emphasize the importance of both positive and negative phytochrome interactions in achieving a WT response to both photoperiod and temperature . Finally , when only phyA and phyC were present , plants flowered slightly earlier than the quintuple phytochrome mutant and plants bearing only phyA or phyC , suggesting a novel interaction between phyA and phyC leading to flowering promotion ( Fig 2A and S4A Fig ) . The flowering behavior of plants bearing phyE and phyA revealed that phytochrome signaling was strongly influenced by ambient temperature ( Fig 2A and S4A Fig ) . In addition , the absence of phytochromes in the quintuple phytochrome mutant significantly reduced the sensitivity to temperature in LD conditions ( Fig 2A ) and the photoperiodic response was absent at 24°C ( Fig 2B and S4B Fig ) . Surprisingly , in short days ( SD ) the quintuple phytochrome mutant flowered much later at 18°C compared to 24°C , showing that temperature also regulates flowering in a phytochrome-independent manner ( Fig 2B and S4B Fig ) . Nevertheless , individual phytochromes also contributed to flowering repression in SD at 18°C . phyE was again the most efficient repressor , followed by phyA and phyB . phyD showed only weak effects on its own , whereas phyC effects were negligible . As in LD , phyC acted antagonistically to phyE , but contrary to LD , in SD at 18°C phyC acted synergistically with phyB and phyD to repress flowering . Noteworthy , the quintuple phytochrome mutant responded to photoperiod only at 18°C , but not at 24°C ( Fig 2B , see S5 Fig for the photoperiod effect ) . These results underscore the importance of interactions between temperature and phytochrome signaling in the control of flowering , but also show that there is at least one temperature responsive pathway that is phytochrome independent ( Fig 2B and S4A Fig ) . To evaluate the photoperiodic response of the mutants , we compared the flowering time of plants grown in LD and SD conditions at either 18 or 24°C ( Fig 2B , S5 Fig ) . At 24°C , none of the phytochromes conferred a photoperiodic response when present alone . Even phyB , which has roles in photoperiodic responses that have been extensively studied in single mutant analyses [29] , failed to confer a photoperiodic response under the conditions tested . Genotypes bearing phyA showed a weak photoperiodic response; phyA behaved as a weak flowering repressor under SD conditions ( Fig 2B ) . Interestingly , only the combination of phyB and phyC produced a strong photoperiodic response at 24°C ( Fig 2B , S5 Fig ) . phyC and phyB form heterodimers and phyC requires phyB in Arabidopsis and rice [21 , 30] . However , our results show that both phyB and phyC are required to confer a photoperiodic response , suggesting that the phyB/phyC heterodimer may have a specific and important role . Further , this specificity seems to be essential for the photoperiodic response , but not for the hypocotyl response to R ( Fig 3 ) , since phyB was sufficient to restore a WT response to R on its own , whereas other phytochromes promoted only subtle phenotypic changes in response to R ( Fig 3 ) . Consistent with a role for the phyB-phyC pair in photoperiodism , this phytochrome pair was the most effective in inhibiting hypocotyl elongation under SD and LD , but not in response to blue light ( S6 Fig ) . These results underscore the role of phyC in the photoperiodic response , which changes from a flowering promoter under LD conditions ( by antagonizing phyB and phyE; Fig 2A ) to a flowering repressor under SD conditions ( by acting in combination with phyB; Fig 2B ) . The differential effects of each phytochrome could be due to differences in the intrinsic properties of each photoreceptor or in the mRNA expression levels , translatability or distribution patterns . The intrinsic properties include the differential capacity to accumulate at the protein level , to heterodimerize or to signal to downstream factors , and the photochemical properties of each phytochrome . To rule out the effects of mRNA expression and distribution patterns , we generated transgenic lines in which each phytochrome fused to the hemaglutinin ( HA ) tag was driven by the 35S constitutive promoter in the quintuple phytochrome mutant background . At least eight independent transgenic lines for each phytochrome were obtained without previous selection for the phenotype other than herbicide resistance . Individual lines were evaluated for the flowering and germination phenotypes ( S7 Fig ) . The use of the same epitope tag allowed us a direct comparison of the levels of PHY apoprotein accumulation in each transgenic line ( S8 Fig ) . Only one out of fourteen phyC lines expressed detectable levels , which is consistent with the phyC dependence on phyB ( S2 Fig ) [21] . PHYD apoprotein was expressed at somewhat lower levels ( 50% in average ) , and PHYE to even lower levels ( 12% ) , when compared to PHYB expressing lines , which is consistent with previous reports on the overexpression of these photoreceptors [24] ( S8 Fig ) . We used several independent lines , that were also positive for PHY accumulation , to compare the effectiveness of each photoreceptor in the regulation of germination and flowering , in the absence of other phytochromes ( Fig 4 ) . Several 35S:phyD-HA lines restored germination to levels of above 50% , but only weakly delayed flowering , to timing similar to that of the quadruple phyA phyB phyC phyE mutant , except for a unique late flowering line ( 1 out of 8 independent lines ) . Conversely , 35S:phyE-HA lines did not germinate better than the phyA phyB phyC phyD line , except for a single line that expressed phyE to very high levels ( 1 out of 12 independent lines , S8 Fig ) , but several of these lines significantly delayed flowering , despite their relative lower levels of apoprotein accumulation ( Fig 4 and S2 Table ) . These results confirm that the roles of phyD and phyE differ due to the nature of the photoreceptors themselves rather than to differences in the expression patterns of their mRNAs . Lines expressing 35S:phyA-HA did not restore full phyA activity and had only weak effects on flowering time and germination ( S7 Fig ) . phyA is known to accumulate to high levels in etiolated seedlings , about 8-fold more than phyB [31] . In our transgenic lines , phyA-HA accumulated to levels not much higher than phyB-HA in etiolated seedlings ( S8 Fig ) . Hence , these results could be due to differences in expression level compared to that driven by the native phyA promoter [32] . Finally , expression of 35S:phyC-HA did not restore germination at all and did not delay flowering significantly on its own , but did delay flowering under SD conditions when transformed into plants bearing only phyB ( Fig 5 ) . Further , 35S:phyC-HA also antagonized phyE activity , consistent with previous results ( Figs 2 and 5 ) . To account for differences in T-DNA insertion sites , we performed a similar flowering experiment with F1 lines that were each the product of a cross between the 35S:phyC-HA lines ( in the quintuple phytochrome mutant background ) and lines bearing only phyB ( i . e . , phyA phyC phyD phyE quadruple mutants ) . Despite being heterozygous for the PHYB locus ( PHYB/phyB ) and hemizygous for the 35S:phyC-HA insertion , these F1 lines flowered significantly later than the 35S:phyC-HA homozygous lines ( in the quintuple mutant background ) and the phyA phyC phyD phyE quadruple mutants homozygous for PHYB ( S9 Fig ) , further confirming the mutual requirement of phyB and phyC for regulating the photoperiodic response . Phytochrome heterodimers are known to exist for the phyB/phyC , phyB/phyD , phyB/phyE and phyC/phyD pairs [20 , 21] . Phytochrome fusions to fluorescent proteins were instrumental in studies about the dynamics of phytochrome nuclear localization . However , these studies did not distinguish between phytochrome homodimers and heterodimers [22 , 33–35] . On the other hand , the interactions among phytochromes evidenced above could be either direct or indirect . To test whether direct interactions were possible and if heterodimers differ in intracellular localization patterns , we examined all possible phytochrome pairs by bimolecular fluorescence complementation ( BiFC ) analysis ( S10 Fig ) . We fused the C-terminus of each phytochrome to either the N-terminal of the Enhanced Yellow Fluorescence Protein ( nEYFP ) or the C-terminal ( cEYFP ) . When two molecules of phytochrome interact , the two EYFP halves are close enough to reconstitute the fluorescence activity . In these assays , the complexes mature with time and the equilibria may be displaced , therefore they can not be taken as a measure of binding affinity . On the other hand , we tested the expression of each phytochrome in the Nicotiana benthamiana transient system and they were not expressed at similar levels ( S8C Fig ) . Therefore , our assays must be interpreted in a qualitative rather than quantitative manner . We co-expressed each pair of constructs in Nicotiana benthamiana leaves and kept the plants in the dark for two days before observing the EYFP fluorescence ( S10 Fig ) . phyC was the only phytochrome that did not yield detectable fluorescence when paired with itself , which is in accordance with reported data that phyC does not form homodimers [21] . By contrast , we detected phyE-phyE interactions . Coupled with the finding that phyE is biologically active in the absence of other phytochromes ( Fig 2 ) , this result strongly suggests that phyE forms homodimers . Furthermore , self-interactions were observed for phyA , phyB , and phyD , consistent with these phytochromes forming homodimers [21 , 22] . phyA , phyB , phyD and phyE interacted with each other ( S10 Fig ) . The phyA/phyD and the phyA/phyC signal was relatively weaker , but still above background , indicating that these heterodimers may be possible . The interactions of phyC with phyB and with phyD , the interactions of phyD with phyE and with phyB and the interactions of phyE with phyB are consistent with the reported heterodimers [21] , whereas the interaction of phyC with phyE , and phyA with phyB and with phyE indicates that the existence of phyC/phyE , phyA/phyB and phyA/phyE heterodimers may also be possible . Interestingly , fluorescence was mostly localized to the nuclei when phyB was combined with phyC and to some extent when phyD was combined with phyC ( S10 Fig ) . The existence of phyB/phyC heterodimers was recently described [21 , 30] , but its intracellular localization pattern was never observed . The phyB/phyC heterodimers could have been rapidly transported into the nuclei in response to the light emitted during confocal microscopy . To avoid this , we collected the leaves under a green safe light and fixed the leaves in the dark before confocal microscopy . Again , an important fraction of phyB/phyC fluorescence remained in the nuclei ( Fig 6A ) , lower panel , green nuclei ) , while in the same conditions most of the phyB/phyB fluorescence was cytoplasmic ( Fig 6A , upper panel , green cytoplasm and blue nuclei marked by ECFP fused to the SV40 NLS ) . This result suggests that phyC alters the nuclei/cytoplasm partitioning of phyB in Nicotiana benthamiana . To test this possibility , we compared the rate of reaccumulation of phyB/phyB homodimers and phyB/phyC heterodimers in the cytoplasm in R and FR treated WL-grown plants . After infiltration , plants were grown for 12 h under WL , given a FR pulse and then grown for 36 h in the dark . After the dark period , plants were treated with R for 3 h and analyzed by confocal microscopy ( Fig 6B and 6C , R control , time zero ) . Both phyB/phyB homodimers and phyB/phyC heterodimers were localized to both cytoplasm and nuclei . However , after treatments with FR , a substantial amount of phyB/phyC heterodimers remained in the nuclei ( green nuclei ) , whereas most of the phyB/phyB homodimers were cytoplasmic ( green cytoplasm and blue nuclei , Fig 6B and 6C ) . To further test these localization patterns with a BiFC independent assay , we evaluated the localization of phyC and phyB when coexpressed with each other ( Fig 6D and 6E ) . For this purpose we used GFP and Cerulean fusions of both phyB and phyC . After agroinfiltration with these constructs , plants were grown for 12 h under WL , given a FR pulse and then grown for 36 h in the dark . After the dark period , leaves were collected under a safe green light , fixed and analyzed by confocal microscopy . Once again we observed that phyB was mostly cytoplasmic when expressed alone , whereas it was mostly nuclear when coexpressed with phyC . Conversely , the phyC pattern was mostly nuclear when coexpressed with phyB . Contrary to BiFC data ( S10 Fig ) , phyC was detected when expressed alone ( Fig 6D ) , suggesting that it was stabilized by association with endogenous tobacco phytochromes . Interestingly , phyC was also more nuclear localized than phyB . These results suggest that phyC forces phyB to localize to the nucleus , after periods of darkness and in a light-quality independent manner . However , in Arabidopsis nuclei , phyB remains nuclear under prolonged periods of darkness , low quality or low irradiance , but changes its pattern of localization in nuclear bodies , from large to small nuclear bodies [36 , 37] . To address the behavior of phyB in the presence of phyC in Arabidopsis , we generated phyB-GFP lines either in the quintuple phytochrome mutant background or in a background having only phyC , by crossing phyB-GFP phyA phyB phyC phyD phyE lines with either the quintuple phyA phyB phyC phyD phyE mutant or the phyA phyB phyD phyE quadruple mutant bearing only phyC . The F1 lines were grown in SD conditions in which we observed effects of phyC and phyB on hypocotyl elongation ( S6B Fig ) and flowering ( Fig 2B ) . We observed the pattern of phyB localization at two time points , during the last hour of a SD and during the last hour of a long night ( Fig 6F and 6G ) . phyB-GFP was localized to large nuclear bodies after the light period regardless the presence of phyC . However , after 15 h in the dark , phyB-GFP remained in large nuclear bodies in the presence of phyC , whereas in the absence of phyC , phyB-GFP showed a diffused pattern within the nuclei ( Fig 6F and 6G ) . It was recently shown that phytochrome nuclear bodies are required to inhibit hypocotyl elongation during a prolonged dark period [37] . In the light of these findings , our results strongly suggest that phyC is important to maintain phyB in these active nuclear bodies . Besides the pattern of nuclear localization , phyC could affect total nuclear phyB . We measured phyB and phyC levels in nuclear extracts of plants bearing only phyB , only phyC or both ( S11A and S11B Fig ) during the last hour of a SD or the last hour of the long night period . As expected , phyC did not accumulate in the absence of phyB ( S11B Fig ) , but phyB nuclear levels were higher in the presence of phyC ( S11A Fig ) , suggesting that phyB/phyC heterodimers may be more stable within the nuclei than phyB homodimers . To further test if phyC could increase the activity of phyB at the end of the night phase , we measured the expression of genes that respond to either light quality or dawn cues during the dark to light transition ( S11C–S11F Fig ) , a condition where phyC promotes de accumulation of phyB in large nuclear bodies ( Fig 6F and 6G ) . The mRNA levels of PIF3-LIKE 1 ( PIL1 ) and ARABIDOPSIS THALIANA HOMEOBOX PROTEIN 2 ( ATHB2 ) were repressed by the photoreceptors phyB and phyC to a lower level than phyB alone , even during the last hour of the dark period ( S11C and S11D Fig ) [37] . The morning-expressed clock genes NIGHT LIGHT-INDUCIBLE AND CLOCK-REGULATED1 ( LNK1 ) and CIRCADIAN CLOCK ASSOCIATED 1 ( CCA1 ) [38 , 39] showed the strongest response to light when both phyB and phyC were present ( S11E and S11F Fig ) . Interestingly , the phase of CCA1 expression was delayed in plants bearing phyB and phyC compared to plants bearing phyB alone . This could be a mechanism underlying the sensitivity of plants bearing phyB and phyC to photoperiod . Taken together , these gene expression studies show that the coordinated action of phyB and phyC is observed when the nuclear phyB/phyC heterodimers are expected to be relatively more abundant in large nuclear bodies , at the dark to light transition after a long night period .
When present alone , phyC is barely active , but causes a slight decrease in GA sensitivity during germination ( S3B Fig ) , a slight inhibition of hypocotyl elongation under R , blue and white-light ( Fig 3; S6 and S12A Figs ) , and an increase in hook opening ( S12B Fig ) , consistent with previous reports [14 , 40 , 42] . This low residual activity of phyC is also consistent with the lack of detectable phyC homodimerization [21] ( S10 Fig ) and the low accumulation of phyC in the absence of phyB ( S2 Fig ) [20 , 27] . However , Triticum aestivum ( wheat ) phyC homodimerizes in Arabidopsis and elicits photomorphogenic responses [43] , and forcing homodimerization of phyC triggers photomorphogenic responses in Arabidopsis [22] . Hence , it is possible that the residual effects of phyC may be due to very low levels of phyC homodimers . It is unclear whether phyE is active independently of the other phytochromes and whether it forms homodimers . Clack et al . did not detect phyE homodimers [21] , but more recently , expressed phyE-GFP in transgenic plants was shown to form homodimers [22] . However , native phyE was not tested in a background devoid of all other phytochromes . Our results suggest that native phyE forms homodimers , as phyE is biologically active in the absence of other phytochromes ( Figs 1 , 2 and 4; S6 and S7 Figs ) and it interacts with itself in a BiFC assay ( S10 Fig ) . Nevertheless , it cannot be ruled out that the lack of homodimerization of phyE reported previously may be due to natural phyE variants , or to the use of different accessions [21 , 22] . Interestingly , we found that phyE repressed flowering , even to a greater extent than did phyB ( Fig 2 ) . This repression was highly dependent on ambient temperature under LD conditions , accounting for the temperature-dependent flowering of phyB mutants [13 , 41] . Thus , phyE might compensate for the lack of phyB at low temperatures . Conversely , the individual effects of phyE on germination and hypocotyl elongation were subtle ( Figs 1 , 3 and 4; S1 , S3 and S7B Figs; S2 Table ) , but interactions with other phytochromes emerged ( see below , Fig 7 ) . In contrast to phyE , phyD was more efficient to promote germination ( Figs 1 and 4; S1 and S7B Figs; S2 Table ) , but only weakly repressed flowering ( Figs 2 and 4; S7A Fig; S2 Table ) . Hence , phyD and phyE , which were previously believed to work mostly in a redundant fashion [6 , 11 , 12 , 14 , 15 , 22 , 24 , 40 , 41] , have distinct roles . At this point it is unclear how this may occur . It might imply different capacity of phyD and phyE to interact with downstream signaling components . On the other hand , even if our data does not support that general changes in phytochrome stability may explain the differences between phyD and phyE , we cannot rule out that specific mechanisms to regulate the level of each phytochrome in a tissue specific manner may exist and could account for the opposite efficiencies with which phyD and phyE promote germination and repress flowering . After phyB , phyA was the most important promoter of germination and GA sensitivity , and these effects were more evident under FR ( Fig 1; S3 Fig ) . These findings are consistent with the established roles of phyA [18] . However , rather than behaving as a flowering promoter [28] , phyA , when present in isolation , turned out to be a weak flowering repressor , strongly suggesting that phyA promotes flowering in an indirect manner , by modulating the signaling pathway transduced by phyE and , to a lesser extent , phyD ( Figs 2 and 7 ) . Further , phyA was the only phytochrome to confer some degree of photoperiod sensitivity on its own , showing a stronger repressive role in SD conditions ( Fig 2B ) . As expected , phyB was the phytochrome that had the greatest stimulatory effect on germination and GA sensitivity ( Fig 1; S1 , S3 and S7B Figs ) and the greatest inhibitory effect on hypocotyl elongation ( Fig 3 ) . Despite the widely accepted role for phyB in the photoperiodic flowering response [29] , phyB did not confer photoperiod sensitivity on its own , showing that phyB is necessary but not sufficient for the photoperiodic response ( Fig 2B ) . In our analysis of binary interactions , we detected both positive and negative interactions ( Fig 7 ) . Whereas phyC , phyD , and phyE acted synergistically to promote germination in response to light ( Fig 1; S1 Fig ) , they antagonized both phyA and phyB in terms of GA sensitivity ( Fig 2; S1 Table ) . As GA promotes germination but antagonizes phytochromes in the control of hypocotyl elongation [1] , this mechanism may allow seedlings to regain phytochrome control of hypocotyl elongation ( seedling emergence ) after germination . Interestingly , phyC , phyD , and phyE also acted synergistically with phyA to inhibit hypocotyl elongation under blue light ( S6A Fig ) . This synergistic effect , similar to that observed for FR-promoted germination ( Fig 1 and S3 Fig ) , may be due to higher protein levels of phyA in the presence of any of the three phytochromes , phyC , phyD or phyE ( Fig 7 and S2 Fig ) . Our results also highlight the importance of the antagonistic effects of phytochromes in achieving proper flowering time . Plants bearing only phyB or phyE flowered significantly later than did WT plants under LD conditions , underscoring the importance of phyA and phyC antagonism mainly on phyE and phyB , respectively ( Figs 2 and 7 ) . The interactions among phytochromes shown here could be due , at least in part , to the formation of heterodimers . Heterodimers of phyB/phyC , phyB/phyD , phyB/phyE , and phyD/phyE , and , to a lesser extent , phyD/phyC were previously shown to exist [20 , 21] . However , heterodimers containing phyA and phyC/phyE heterodimers were not found . In this study , we detected interactions between phyA and all phytochromes ( including a very weak signal for phyA/phyC ) and also phyC/phyD and phyC/phyE interactions by using BiFC ( S10 Fig ) . These results must be taken with caution as agroinfiltration in tobacco leaves may lead to high expression levels ( S8D Fig ) . Nevertheless , these results raise the possibility that the pair wise interactions observed may be due , at least in part , to direct protein-protein interactions among the phytochrome members and suggest that the phytochrome signaling network is more complex than previously thought . The role of phyB in the photoperiodic response has been extensively reported in diverse species [29 , 44 , 45] . However , although phyB can repress flowering on its own , it requires the presence of phyC to confer photoperiod responsiveness ( Fig 2 and S5 Fig ) . On the other hand , phyC requires phyB , consistent with previous reports in both rice and Arabidopsis [15 , 30 , 40] . The strict requirement for the activity of both phyB and phyC could indicate that components downstream of the photoreceptors interact or that the activity of phyB/phyC heterodimers differs from that of the individual photoreceptors . Our data favor the second possibility . First , the existence of phyB/phyC heterodimers was reported previously in both Arabidopsis and rice [21 , 30] and we have shown here that phyB and phyC interact in vivo ( Fig 6 and S10 Fig ) . Second , we showed that the phyB/phyC heterodimers are consistently localized to the nucleus in transient assays in Nicotiana benthamiana , even after prolonged dark periods or under low red to far-red light ratios , suggesting an emerging property of the phyB-phyC system ( Fig 6 and S10 Fig ) . Third , neither phyB nor phyC conferred even a subtle photoperiodic response on their own; this is unlikely if phyB and phyC coaction is due to the interaction of downstream components of phyB and phyC . Interestingly , phyB repressed hypocotyl elongation on its own in response to R ( Fig 3 ) and repressed flowering under both SD and LD conditions ( Fig 2 ) , but only in the presence of phyC was phyB able to restore photoperiodism . Recent reports strengthen the idea that phyB in large nuclear bodies is active and necessary to trigger downstream processes during the night period [35–37] . We show here that phyC promotes the localization of phyB to large nuclear bodies after a long night period ( Fig 6F and 6G ) , suggesting that phyB/phyC heterodimers are active for longer periods of darkness . Hence , the extended activity phyB/phyC heterodimers in the night and its availability early in the dark to light transition might be important to repress flowering and hypocotyl elongation specifically under SD conditions . Two interesting previous observations may be explained in light of our results . It was reported that PHYB antisense lines with a ~75% reduction in phyB display longer hypocotyls , as do phyB mutants , but normal flowering time [46] . Similarly , the phyB-28 allele , which lacks most of the HKRD domain , has long hypocotyls but an almost normal flowering time [47] . These observations can be explained if , under low phyB levels , phyB is found mostly as a heterodimer with phyC and hence retains normal photoperiodic responses with respect to flowering , while impairing normal hypocotyl responses that are mostly due to phyB/phyB homodimers . Similarly , the effects of the phyB-28 allele could be explained if this mutation either affected phyB levels ( which may not be the case ) or if phyB/phyC heterodimers were more abundant than phyB/phyB homodimers . We show here that phyB alone is sufficient to confer full hypocotyl and germination responses to R and to repress flowering , which is also consistent with phyB alone being sufficient to confer a response to light quality ( response to R/FR ratios ) [42] . However , it has not been easy to dissect the role of phyB in the flowering response to light quality from its role in photoperiodic flowering [28 , 29 , 48] . Our data support the notion that phyB/phyB homodimers are involved in the responses to light quality , whereas the phyB/phyC heterodimers are involved in the photoperiodic response . The role of phyC in photoperiodism may be widely conserved . In population studies , strong phyC alleles were found to be more abundant at higher latitudes [49] , which could indicate that these alleles have an increased sensitivity to photoperiod . In wheat , Brachypodium distachyon , and Hordeum vulgare ( barley ) , phyC promotes flowering more effectively in LD conditions [43 , 50 , 51] . These results highlight the importance of phyC in photoperiodic responses in diverse habitats and species and are consistent with our finding that phyC is essential for the photoperiodic response . It would be interesting to study if phyC also promotes flowering under LD conditions by antagonizing phyB in wheat and related grasses and to establish the possible involvement of phyB/phyC heterodimers in the photoperiodic response of these species . An interesting aspect of phytochromes is that their effects are altered by temperature ( Fig 2A ) . An absence of phytochromes resulted in very low temperature sensitivity under LD conditions . However , phyE and phyB repressed flowering more efficiently as the temperature decreased , indicating that cross-talk exists between the phytochrome and temperature signaling pathways ( Fig 2A ) . The specific effect of phyA on temperature-dependent phyE signaling , but not on phyB signaling , strongly suggests that there are differences in the signaling pathway downstream phyB and phyE ( Fig 2A ) . How phyE regulates flowering is still unknown , but these results raise the possibility that a CONSTANS ( CO ) -independent mechanism that differs from the phyB-mediated effect on CO stability may function downstream of phyE [29] . We think of two possible mechanisms to explain how phyB/phyC heterodimers might contribute to photoperiod detection . One possibility is that the effects of phyB on CO stability [29] may be due indeed to phyB/phyC heterodimers . A second possibility is supported by the role of phyB in regulating the phase of the circadian clock [52] and also supported by our gene expression data ( S11 Fig ) : phyB/phyC heterodimers might affect the phase of clock and flowering time genes and hence , photoperiod detection . Flowering was reported to be insensitive to the photoperiod in the absence of phytochromes [10] , and we found similar results when plants were grown at 24°C . However , we also found that photoperiod responsiveness was restored at low temperatures ( Fig 2B ) . Together , these results suggest the existence of phytochrome-dependent and -independent mechanisms that regulate the flowering response to temperature , consistent with previous genetic evidence [53] . PIF4 [54] and two transcription factors that form heterodimers , SHORT VEGETATIVE PHASE ( SVP ) and FLOWERING LOCUS M ( FLM/MAF1 ) , regulate flowering in response to ambient temperature [55 , 56] . Further experimentation is needed to determine if the PIF4 and SVP/FLM pathways correspond to phytochrome-dependent and -independent pathways .
phyA-211 , phyB-9 , phyC-2 , phyD-201 , and phyE-201 alleles are in the Columbia background [4 , 15 , 42] . Segregating populations were genotyped as previously described [42] to identify triple and quadruple mutants . For experiments with seedlings , sterilized seeds were suspended in 100 μM GA4+7 ( Duchefa Biochemie , Haarlem , The Netherlands ) , stratified for 3 days at 4°C , and then pipetted onto plates of Murashige Skoog Salts media and 0 . 8% Plant Agar ( Duchefa Biochemie ) . Light treatments were performed in dedicated growth chambers ( Model I30BLL , Percival Scientific , Perry , IA , U . S . A . ) . For red and far-red light , light-emitting diodes were used . For the hypocotyl measurement assays , 15 seeds were plated per replicate , and the average height of the 10 tallest seedlings was recorded per replicate . In the germination assays , sterilized seeds were directly plated on MS salts plates ( 0 . 8% agar ) and given a post-imbibition saturating 5-min FR pulse to revert seed phytochrome to the Pr form . Then , the seeds were stratified in darkness for 3 days at 4°C . After stratification , seeds were incubated at 23°C for 6 days under the indicated light regimes , before counting the germinated seeds ( i . e . , radicle emergence ) . Each pool of seeds used in the germination assays was collected from plants grown side by side under the same conditions . This process was repeated several times and pools of seeds grown under the same conditions , but at different times , were collected . The phytochrome cDNAs were obtained by retrotranscription from Col-0 RNA , and cloned into the pCHF5 plasmid fused to the C-terminus HA tag ( S3 Table ) . Single locus insertion lines from the T3 generation were selected for each experiment . ( See also Supporting Information . ) Agrobacterium tumefaciens ( GV3301 ) containing the pCardo1-C-nEYFP or pCardo1-C-cEYFP vector ( harboring each phytochrome tagged with nEYFP or cEYFP , respectively ) , the pBIN19-35S-P19 vector ( containing the P19 suppressor of silencing ) , and the pCardo2 . 1-ECFP-NLS vector ( containing the nuclear marker ECFP-NLS ) were co-infiltrated into the leaves of Nicotiana benthamiana plants grown under LD at 23°C essentially as described [57] with some modifications . After infiltration , plants were grown in the same LD conditions for 12 h , up to the end of the photoperiod , and then treated with FR pulses , R pulses or darkness , as indicated in each figure legend . The confocal images were taken using a Zeiss LSM 710 microscope . EYFP was excited at 514 nm and observed at 520–539 nm , whereas ECFP was excited at 458 nm and observed at 466–480 nm . 600 mg of tissue were frozen in liquid nitrogen and gently grinded in a mortar . Then , the nuclei extraction was performed as an simplified version of [58] without the Percoll gradient . The nuclei was pelleted and then washed twice to enrich in the nuclei fraction . Equal volumes of each nuclei preparation were used for immunoblots , and phytochromes quantitated relative to H3 . The sequence of genes used in this study can be found in the GenBank/EMBL or the Arabidopsis Genome Initiative databases under the following accession numbers: AT1G09570 ( PHYA ) , AT2G18790 ( PHYB ) , AT5G35840 ( PHYC ) , AT4G16250 ( PHYD ) , AT4G18130 ( PHYE ) , AT4G16780 ( ATHB2 ) , AT2G46970 ( PIL1 ) , AT2G46830 ( CCA1 ) . AT5G64170 ( LNK1 ) . | As sessile organisms , plants respond to and integrate environmental information . An intriguing aspect is how plants integrate this information . We studied the interactions among members of the phytochrome family of photoreceptors , which detect the changes in light quality that occur upon shading by other plants , as well as the duration of daylength , which indicates seasonal changes . We conducted these studies in Arabidopsis , which bears five phytochromes ( phyA-phyE ) . We show that the individual roles of each phytochrome ( in the absence of others ) are different but , more importantly , that their combinations give different properties to the system . phyE , for instance , regulates flowering in a temperature-dependent manner , indicating that phyE signaling is a point of integration between light and temperature cues . phyC is poorly active on its own , but it is essential for the phytochrome-dependent photoperiodic flowering . In long-days ( mimicking late spring ) phyC promotes flowering indirectly , by inhibiting phyB and phyE signaling , which are themselves repressors of flowering . Conversely , under the short days of winter phyC becomes a flowering repressor because phyB requires phyC only for this specific response . Therefore , phyC is essential for the detection of photoperiod by phytochromes and suggest a conserved role for phyC in the photoperiodism of angiosperms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"anatomy",
"plant",
"growth",
"and",
"development",
"plant",
"embryo",
"anatomy",
"brassica",
"social",
"sciences",
"plant",
"physiology",
"neuroscience",
"developmental",
"biology",
"plant",
"science",
"model",
"organisms",
"plants",
"flowering",
"plants",
"... | 2016 | Bottom-up Assembly of the Phytochrome Network |
Despite the global threat caused by arthropod-borne viruses , there is not an efficient method for screening vector populations to detect novel viral sequences . Current viral detection and surveillance methods based on culture can be costly and time consuming and are predicated on prior knowledge of the etiologic agent , as they rely on specific oligonucleotide primers or antibodies . Therefore , these techniques may be unsuitable for situations when the causative agent of an outbreak is unknown . In this study we explored the use of high-throughput pyrosequencing for surveillance of arthropod-borne RNA viruses . Dengue virus , a member of the positive strand RNA Flavivirus family that is transmitted by several members of the Aedes genus of mosquitoes , was used as a model . Aedes aegypti mosquitoes experimentally infected with dengue virus type 1 ( DENV-1 ) were pooled with noninfected mosquitoes to simulate samples derived from ongoing arbovirus surveillance programs . Using random-primed methods , total RNA was reverse-transcribed and resulting cDNA subjected to 454 pyrosequencing . In two types of samples , one with 5 adult mosquitoes infected with DENV-1- and the other with 1 DENV-1 infected mosquito and 4 noninfected mosquitoes , we identified DENV-1 DNA sequences . DENV-1 sequences were not detected in an uninfected control pool of 5 adult mosquitoes . We calculated the proportion of the Ae . aegypti metagenome contributed by each infecting Dengue virus genome ( pIP ) , which ranged from 2 . 75×10−8 to 1 . 08×10−7 . DENV-1 RNA was sufficiently concentrated in the mosquito that its detection was feasible using current high-throughput sequencing instrumentation . We also identified some of the components of the mosquito microflora on the basis of the sequence of expressed RNA . This included members of the bacterial genera Pirellula and Asaia , various fungi , and a potentially uncharacterized mycovirus .
Dengue virus types 1–4 are emerging members of the genus Flavivirus in the Flaviviridae family , which consists of a group of related enveloped viruses with positive-stranded RNA genomes [1] . The four types can be distinguished through serologic assays , though each is capable of causing a spectrum of disease ranging from a mild or unapparent viral syndrome to severe and often deadly manifestations of hemorrhagic disease known as Dengue hemorrhagic fever ( DHF ) and Dengue shock syndrome ( DSS ) . DHF is characterized by a sudden onset of fever with hemorrhagic complications such as petechiae and/or gastrointestinal hemorrhage and can be followed by shock and low blood pressure , hallmarks of DSS [2] . There is no approved specific therapeutic for dengue infection; treatment is limited to supportive care . The viruses are transmitted throughout the tropics and subtropical areas by Aedes aegypti , a diurnal mosquito , and in South East Asia by a related species , Ae . albopictus [3] , [4] . The estimated global burden of disease caused by dengue is 100 million cases per year with 250 , 000 to 300 , 000 cases of DHF annually , for which the case fatality rate is 5% [2] . Although dengue is arguably one of the more significant arthropod-borne ( arbo- ) viruses in terms of the morbidity and mortality it causes [5] , it is not the only arbovirus that causes a significant threat to humans; West Nile virus , Japanese encephalitis virus , yellow fever virus , and others are also of major global concern [1] . Despite the ever-present global threat caused by arboviruses , there is not yet a single detection system that is capable of detecting all arboviruses simultaneously [6] . In general , traditional viral detection and surveillance methods can be costly and time consuming and generally require prior knowledge of the etiologic agent , as they rely on virus-specific primers or antibodies . Therefore , these techniques are unsuitable for situations when the causative agent of an outbreak is entirely novel or is an unknown sequence variant . The pitfalls of using specific PCR targets were vividly demonstrated when it was found that a new variant of a genital Chlamydia trachomatis strain had escaped detection for several years because it had acquired a deletion in the region of a virulence plasmid that was a target for a commonly used real-time PCR assay [7] . Recently , several groundbreaking studies have been published that used de novo high-throughput bead-based pyrosequencing of DNA [8] to provide putative identification of viral disease agents [9] , [10] , [11] . These studies all use a metagenomic approach [12]- sequencing a sample containing total nucleic acids- with no separation of host nucleic acids from those of the infecting or commensal microorganisms . In one such study , pyrosequencing was used to conduct a survey of microorganisms in honeybee colonies associated with colony collapse disorder ( CCD ) in comparison to colonies that were free of CCD [11] . In another study , organs were transplanted from the same donor to several recipient patients , all who died weeks later of a febrile illness . A variety of assays aimed at identifying the etiologic agent were uninformative . The authors of that study employed 454 pyrosequencing of the patients' samples and were able to identify , amidst a host background of 103 , 632 total reads , 14 reads corresponding to a novel , deadly arenavirus [9] . In a third recent study , a novel strain of Ebola virus was identified in Uganda by using pyrosequencing to examine patients' samples [10] . These three examples highlight the utility of pyrosequencing for relatively unbiased detection of an etiologic agent that is either unsuspected or novel , situations in which more traditional methods might not be appropriate . While this pioneering work highlights the promise of metagenomic sequencing based virus detection there has not been a study that measures the relationship of the number of DNA sequences detected to the number of virus particles present ( although a recent study that utilized metagenomic sequencing of citrus phloem cells estimated the limit of detection for Candidatus liberibacter asiaticus , the bacterium thought to be the causative agent of Huanglongbing citrus disease , to be one bacterial cell per every 52 phloem cells [13] ) . In the current study we explore the usefulness of pyrosequencing in the context of arboviral infections . Ae . aegypti mosquitoes experimentally infected with DENV-1 were pooled with noninfected mosquitoes to simulate samples that might be derived from real-world arbovirus surveillance programs . Using random-primed methods , total RNA was reverse-transcribed and the resulting cDNA subjected to high-throughput pyrosequencing . We successfully detected DENV-1 sequences from samples containing infected mosquitoes . Sequences matching DENV-1 were absent from a control sample of noninfected mosquito RNA . Based on the results from this survey we have produced a model of the type of sequencing capacity necessary to detect novel variants of dengue virus in screens of wild-caught Ae . aegypti .
Aedes aegypti ( Rockefeller strain ) were reared using standard methodology . Briefly , eggs were flooded , allowed to hatch overnight and resultant larvae were provided ground fish food and reared in a walk-in incubator maintained at 26°C . Pupae were removed and adults allowed to emerge into 3 . 8-liter cardboard cages with netting over the open end . Adult mosquitoes were transported to a Biological Safety Level-3 containment laboratory where they were inoculated intrathoracically with 0 . 3 uL of a DENV-1 ( HAW strain ) virus suspension containing 106 . 2 PFU/mL of inoculum ( 102 . 7 PFU/mosquito ) . Inoculated mosquitoes were maintained in an incubator at 26°C for 7 days . Mosquitoes were then killed by freezing at −20°C for 5 min and then triturated either individually or with the addition of uninfected Ae . aegypti in 0 . 6 mL of diluent ( 10% heat-inactivated fetal bovine serum in Medium 199 with Earle's salts [Invitrogen , Inc . , Carlsbad , CA] and antibiotics ) . A 0 . 1-mL aliquot of the mosquito suspension was removed and added to 0 . 9 mL of diluent . This was frozen at −70°C until testing by plaque assay on Vero cells to determine the number of viral particles present in the sample . 1 . 5 mL of TRIzol-LS ( Invitrogen , Inc . , Carlsbad , CA ) was then added to each tube and the contents of each vial were divided into two cryovials to create duplicate samples . TRIzol-LS extraction of total RNA from homogenized mosquitoes was performed according to the manufacturer's instructions . The final RNA pellet was resuspended in 50 uL of RNase-free water and stored at −70°C until use . PCR-ESI/MS was performed as described previously . Briefly , RT-PCR was performed using an 8-primer pair pan-Flavivirus assay on the Ibis T-5000 [14] . The assay targets conserved genomic regions of the known pathogenic mosquito- and tick-borne members of the genus Flavivirus . After PCR , the reactions were desalted , and the purified DNA products were individually sprayed into a Bruker Daltonics microToF ( Billerica , MA ) mass spectrometer . Proprietary signal-processing software was used to deconvolute raw data from the mass per charge . This molecular mass was then assigned to the amplicon's empirical molecular mass and correlating base composition , which was matched with those in the system's database . For every RT-PCR reaction well , the signal amplitude of the internal positive control ( which is spiked into every reaction at a known concentration of 100 copies per reaction ) and the sample were compared and interpreted to give quantitative results . Mosquito suspensions were tested for infectious virus by plaque assay on MK-2 ( monkey kidney ) cells . Briefly , serial 10-fold dilutions of the mosquito suspension were made in diluent and 0 . 1 mL added to each well of a 6-well plastic tissue culture plate . An agarose overlay was added 1 hour later and a second overlay , containing neutral red , added 5 days later . Plaques were enumerated the following day . Methods in a 454 Application Note for sequencing Influenza Virus RNA [15] were modified slightly to be used for detecting flavivirus RNA amidst insect RNA at concentrations relevant to experimental inoculation . The most substantial modification to the procedure was the starting material; rather than starting from purified viral nucleic acid , total RNA ( that of mosquito , viral , and potential commensal organisms ) was used as input material . Enrichment for viral RNA via a biotin-labeled oligonucleotide specific for viral sequences was not performed . Additionally , other small differences in our procedure as compared to that of the Application Note consisted of using random heptamer primers for reverse transcription , other slightly modified oligo sequences , and omission of the second round of RNA Clean and AMPure ( Agencourt Biosciences Corporation , Beverly , MA ) bead purifications . Briefly , total mosquito RNA was fragmented at 82°C for 2 minutes in fragmentation buffer as described in the Application Note , purified with one round of RNAClean , and then RT-PCR was performed using a random heptamer primer ( 5′-phosphate-N7-3′ ) . RT-PCR conditions , RNA removal , reaction neutralization , and single-stranded cDNA recovery were performed as per the Application Note . In preparation for FLX sequencing , adaptors with overhangs were made by annealing together pairs of complimentary oligos . Adaptor A was made by annealing the following 2 oligos together: mod sscDNA OligoA′ 5′-NNN NNN CTG ATG GCG CGA GGG AGG-dideoxyC-3′ and sscDNA OligoA 5′-GCC TCC CTC GCG CCA TCA G-3′ , and Adaptor B was made by annealing together the following 2 oligos: sscDNA OligoB 5′-biotin-GCC TTG CCA GCC CGC TCA GNN NNN N-phosphate-3′ and sscDNA OligoB′ 5′-phosphate-CTG AGC GGG CTG GCA AGG-dideoxyC-3′ . Annealing and directional ligation of the adaptors onto cDNA was performed essentially as described in the Application Note . The final cDNA library was subjected to only one round of RNAClean . An amplification PCR was performed prior to emulsion PCR using the following oligos: Amplification Primer A 5′-GCC TCC CTC GCG CCA-3′ and Amplification Primer B 5′ GCC TTG CCA GCC CGC-3′ , Advantage 2 polymerase mix ( Clontech , Mountain View , CA ) , and the PCR conditions described in the Application Note . All oligos were purchased from Invitrogen ( Carlsbad , CA ) . Emulsion PCR was performed using 454 emPCR kits II and III ( Roche , Indianapolis , IN ) . After enrichment , the resulting emPCR II- and III-derived beads were pooled 1∶1 for the sequencing run . A 9 ul ( ∼600 ng ) aliquot of the sample in question ( consisting of total RNA extracted from a pool of 5 male Ae . aegypti experimentally inoculated with DENV-1 ) was taken and subjected to eukaryotic ribosomal RNA depletion using the RiboMinus Kit ( Invitrogen ) followed by ethanol precipitation with glycogen ( Applied Biosystems/Ambion , Austin , TX ) as a carrier for lower molecular weight RNA species . After the ethanol-precipitated , rRNA-depleted , RNA was air-dried and resuspended in DEPC-treated water , it was fragmented , reverse-transcribed and sequenced as described in the metagenomic sequencing protocol ( above ) . For sake of comparison , a second aliquot ( 3 ul ) of the same total RNA prep was also fragmented , reverse-transcribed , sequenced , and analyzed in parallel , with the exception that this second aliquot was not subjected to rRNA depletion . For the rRNA depletion experiments , BLASTn was used to find hits to Ae . aegypti and Ae . albopictus rRNA genes as well as DENV-1 genome ( GenBank Accession DVU88536; BLAST cutoff used was bit score greater than 100 ) . Transcript analysis and remapping analyses were performed using CLC Genomics Workbench v3 . 7 ( CLC Inc , Aarhus , Denmark ) . The Ae . aegypti whole genome shotgun data [Genbank: AEGE00000000] was downloaded in December 2009 . CLC Reference assembler was run with Insertion cost = 3 , Deletion cost = 2 and Mismatch cost = 2 . Contigs with BLAST match scores greater than 100 to the Ae . albopictus rRNA operon [Genbank: MQSRAGN] were flagged as probably containing Ae . aegypti rRNA operons . The AaegL1 . 2 collection of 18 , 760 known transcripts was downloaded from VectorBase ( www . vectorbase . org ) . CLC RNA-seq was run with Minimum length fraction of 0 . 9 and Minimum exon coverage fraction of 0 . 2 . GS Reference Mapper software ( v2 . 0 . 01 . 14; Roche/454 ) was used to produce reference-guided assemblies of each of the mosquito datasets with respect to the DENV-1 genome ( GenBank; DVU88536 ) and those results were viewed in Geneious software ( Biomatters Ltd . , Auckland , New Zealand ) . Contigs and individual reads were compared to sequences in NCBI nucleotide ( nt ) , refseq_RNA , environmental sequences ( env_nt ) , and whole genome sequence ( WGS ) databases using megaBLAST and protein ( nr ) database using BLASTx . The best BLAST hit was chosen and reported based on e value and sequence identity . Individual reads for each library were also uploaded to the MG-RAST server at http://metagenomics . nmpdr . org and are publicly accessible , listed under the identifiers 4441794 . 3 ( N2187 ) , 4441733 . 3 ( N2173 ) , 4441793 . 3 ( N2175 ) , 4450608 . 3 ( N2473 ) , and 4450615 . 3 ( N2474 ) . For the bacterial sequences identified , phylogenetic profiles were examined and most numerous hits were reported based on the RDP or SEED datasets . Minimum alignment length used was 50 and maximum e-value was set at 0 . 01 .
In order to assess 454 pyrosequencing as a platform for detection of arboviruses in their insect vectors , adult female Ae . aegypti were inoculated with 0 . 3 uL of a suspension containing 106 . 2 plaque-forming units ( PFU ) /mL ( a total of 102 . 7 PFU per mosquito ) . These mosquitoes were held at 27°C for 7–12 days and then mixed with noninfected control mosquitoes to simulate samples that might result from real-world arbovirus surveillance programs . The pooled mosquitoes were triturated and total RNA was extracted using TRIzol and reverse-transcribed using random primers . 454 GS FLX libraries were constructed from the resulting total cDNA ( see materials and methods ) . Library N2173 was made from five DENV-1-infected mosquitoes , library N2187 was made from 5 noninfected control mosquitoes , and library N2175 was made from five pooled mosquitoes , only one of which was infected with DENV-1 . These three FLX libraries were sequenced individually using the 454 GS FLX sequencer and each run resulted in roughly 200 , 000 to 350 , 000 reads per library post 454 quality filtering- roughly 30–60% less reads than would be expected from a whole genome sequencing run on the FLX instrument . The average read length was 208 bases long , noticeably shorter than the average read length generally produced from GS FLX sequencing of genomic DNA ( ∼250 bases ) , likely a result of the heat fragmentation step used to create appropriate sized RNA fragments , in place of the usual nebulization step involved in sequencing genomic DNA . In the case of these mosquito cDNA libraries , the average size of DNA fragments before binding adaptors was 150 bases , while for standard FLX libraries made from DNA fragmented via nebulization with Nitrogen gas , the average DNA fragment size is generally around 200 bases . Post sequencing , the resulting reads were assembled de novo into contigs using the 454 Newbler assembler with default parameters ( minimum overlap length of 40 and minimum overlap identity of 90 ) . A relatively small number ( less than 4% ) of the overall reads assembled into large contigs ( those greater than 500 nucleotides in length ) . As no physical means were used to purify or separate by origin the various RNA species present in each sample , it was expected that mosquito ribosomal RNA would make up a large percentage of the total . In fact , upon analysis of the resulting reads , Ae . aegypti ribosomal RNA was indeed found to constitute a significant percentage of the resulting reads , making up as much as 92% of the reads in the case of the N2173 library ( Table 1 ) . ( For a detailed analysis of which de novo assembled contigs from each dataset map to mosquito ribosomal RNA , see Tables S5 , S6 , S7 , S8 . ) However , despite the large contribution of rRNA , DENV-1-specific reads were still detected . The individual reads from each library were aligned against the NCBI nonredundant nucleotide and protein databases using BLASTn or BLASTx respectively [16] and additionally , reads from each of the 3 datasets were also assembled using the DENV-1 genome as a reference and results of the latter are displayed in Figure 1B . In the case of the library constructed from 5 infected mosquitoes ( N2173 ) , there were 227 reads ( 0 . 08% of the total number of reads ) that mapped to the DENV-1 genome ( Table 2 ) . Consistent with the relative numbers of infected mosquitoes there were proportionately 5 times fewer DENV-1 hits in the dataset from the library made from 1 infected mosquito pooled with 4 noninfected mosquitoes ( N2175 ) , indicating that the relative percentage of virus-specific hits per sequencing run directly correlates with the proportion of infected to noninfected mosquitoes and titer of virus present in a sample of pooled insects . The DENV-1 matches for these two datasets were relatively well spaced over the entire Dengue genome , suggesting that there was no significant inherent bias toward one end of the genome over the other in efficiency of reverse-transcription or adaptor ligation ( Figure 1 ) . The large contigs produced by de novo assembly of the libraries were queried against the NCBI nucleotide ( nt ) , refseq_RNA , environmental sequences ( env_nt ) , and whole genome sequence ( WGS ) databases using megaBLAST or the protein ( nr ) database using BLASTx . For N2173 , the library made from 5 infected mosquitoes , 7 of 23 large contigs were found to be DENV-1-specific ( Figure 1A ) . The total sequence encompassed by those 7 large contigs was 8517 bases ( 212 reads ) , representing 79 . 3% coverage of the 10 . 7 kB viral genome in just the large contigs alone . By comparison , for N2175 , the library created from 1 infected out of 5 total mosquitoes , only 2 of the 23 large contigs ( made up of 21 reads ) were DENV-1-specific , and for N2187 , the library made from 5 noninfected mosquitoes , none of the 38 large contigs were a match to the DENV-1 genome . Not surprisingly , reference-guided assembly of the reads resulted in contigs with greater coverage of the Dengue genome and included more DENV-1 reads than were included in the de novo assembled large contigs . Average depth of coverage of the DENV-1 genome scaled with the proportion of mosquitoes infected with DENV-1 , being 4 . 9× for N2173 ( 5 infected ) and 1× for N2175 ( 1 of 5 infected; Table 2 ) . As such a large proportion of each dataset was found to consist of ( mainly host ) ribosomal RNA sequence , we sought to decrease the proportion of eukaryotic ribosomal RNA sequences and thereby improve the sensitivity of detection of DENV-1 . For this experiment , starting material consisted of total RNA extracted from a pool of 5 male Ae . aegypti experimentally inoculated with DENV-1 . An aliquot of this RNA sample was treated using a commercially available kit that consists of biotinylated bead-bound probes designed to hybridize and selectively deplete eukaryotic rRNA ( mouse and human ) . Following this treatment , the rRNA-depleted sample ( N2473 ) was reverse-transcribed and sequenced as described in the methods . In parallel , a second aliquot of this same RNA sample ( N2474 ) was also reverse-transcribed and sequenced , but without being subjected to rRNA depletion . The sequence quality and output of each of the two sequencing runs was overall similar , producing 391 , 103 and 336 , 905 reads respectively . As expected , in the N2474 dataset , there was an overall decrease in the total number and proportion of reads corresponding to Aedes rRNA sequences ( proportion decreased from 61 . 4% to 54 . 5% ) and an increase in the depth of coverage of the DENV-1 genome ( Table 2 ) . Each of the five datasets were aligned to the DENV-1 genome and results are shown in Figure 1B . This reference-guided assembly of the N2473 dataset demonstrates 4 . 5× average depth of coverage of the DENV-1 genome as compared to 8 . 3× average depth of coverage by the N2474 dataset . Notably , although rRNA depletion resulted in a higher average depth of coverage , the coverage was skewed toward the 5′ end of the DENV-1 genome . From the sequencing data above , we estimated the proportion of reads attributable to a single infecting DENV-1 pathogen in each experiment , a number we called “pIP” . This was done by dividing the proportion of DENV-1 reads found in each experiment by an estimate of the number of infecting viruses ( Table S9 ) . The number of DENV-1 genomes per mosquito was assessed both by testing an aliquot of each sample using PCR electrospray ionization mass spectrometry ( PCR-ESI/MS ) and the Ibis T5000 biosensor or using titration on MK2 cells . Based on the IBIS data the pIP ranged from as low as 2 . 75×10−8 for N2173 to 1 . 08×10−7 for the depleted N2474 experiment . The empirically derived pIP value can be useful for planning future experiments that screen mosquitoes for DENV-1 like viruses by shotgun sequencing ( see discussion ) . For comparison the 100 most abundantly expressed transcripts in each female mosquito sample were ranked by RPKM value ( the number of reads which map per kilobase of exon model per million mapped reads , as in [17]; see Tables S1 , S2 , S3 , S4 ) . Overall , each sample shared about 50% similarity in the top ranked expression list . Six highly expressed transcripts that were in the top ten for all three libraries ( Table 3 ) were AAEL017647-RA and AAEL017654-RA , both of which are ribosomal RNA; AAEL017811-RA , RNase MRP; AAEL017868-RA , eukaryotic type signal recognition particle; and AAEL001702-RA and AAEL017571-RA both of which have no description in VectorBase . Relatively few genes known to be associated with a function were differentially expressed in the N2173 sample versus the other samples . An exception was a putative salivary gland mucin [GenBank: DQ440007 . 1] [18] , which was found to be up-regulated and may represent a host defense response to Dengue virus infection . Other exceptions include down-regulated transcripts corresponding to trypsin 3a1 and up-regulated transcripts corresponding to tRNAGly . The Dengue-infected mosquito RNA contained a higher percentage of Ae . aegypti rRNA transcripts and correspondingly lower percentage of transcripts from other genes . Overproduction of rRNA in response to viral infection has been previously reported in silk worms infected with cytoplasmic polyhedrosis virus [19] and human liver-derived cells infected with Hepatitis C virus [20] and may represent a physiological response to viral stress on the mosquito cells or may be the result of viral takeover of the host protein making machinery , although the true biological implications of this finding are not clear at this time . The top expressed transcript in both the infected samples was AAEL017413-RA , a transcript for which there is no description in VectorBase , but upon BLAST analysis , bears significant similarity to large subunit ribosomal RNA . The shotgun sequencing of cDNA derived from triturated insect total RNA is in essence a metagenomic experiment , and so it was expected that the results of this sequencing could include identification of various bacterial or other types of symbiotic microorganisms living on or in the mosquitoes . In order to more fully characterize the diversity of organisms present in these mosquito samples , the datasets were analyzed using MG-RAST , a publicly available web service with a pipeline that analyzes user-uploaded metagenomic datasets and provides automated metabolic and phylogenetic reconstructions according to various parameters that can be adjusted by the user [21] . The reads were compared against the RDP dataset of ribosomal RNA ( rRNA ) sequences [22] and thresholds were set for a minimum alignment length of 50 and maximum E-value of 0 . 01 . Based on this automated phylogenetic analysis of the reads , bacterial ribosomal RNA was found to constitute roughly 3 to 6 . 5% of the total number of classifiable reads for each library's dataset ( without treatment to reduce the contribution of eukaryotic rRNA ) . Specific bacterial taxa with the highest number of hits were the two genera Pirellula and Asaia . Additionally , unclassified bacteria , including unclassified Clostridiales and unclassified Enterobacteriaceae , were found to constitute a relatively large percentage of the total bacterial rRNA contribution ( Results are summarized in Table S10 ) . Based on the alignment of reads against the RDP database , as well as BLAST analysis of the large contigs , as much as 0 . 12% of reads from a given mosquito pool had their best hit to rRNA from members of the genus Asaia . This may likely indicate that all of the Ae . aegypti mosquitoes used in this study were infected or colonized with Asaia sp . , regardless of their DENV-1-infected status . However , as there is currently no fully sequenced genome from a member of the genus Asaia available in GenBank for comparison , it is conceivable that the total number of Asaia-specific reads in our dataset could be underestimated . In addition to the various bacterial rRNA sequences , fungal rRNA was found to constitute between 0 . 6% and 2 . 3% of the total reads in each library's dataset ( using the SEED algorithm with a minimal alignment length of 50 and maximum e-value of 0 . 01 ) . With these parameters , the top hit was to Saccharomyces cerevisiae . De novo assembly resulted in large contigs that by BLAST had their best hit to a rRNA of various fungal origins ( including the genera Penicillium and Aspergillus ) and in all five samples an RNA-dependent RNA polymerase encoded by a mycovirus was identified by SEED analysis of reads using MG-RAST and in 2/3 cases examined , by BLAST analysis of large contigs as well .
The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy , Department of Defense , United States Army , nor the U . S . Government . Some of the authors are military service members or employees of the U . S . Government . This work was prepared as part of their official duties . Title 17 U . S . C . §105 provides that ‘Copyright protection under this title is not available for any work of the United States Government . ’ Title 17 U . S . C . §101 defines a U . S . Government work as a work prepared by a military service member or employee of the U . S . Government as part of that person's official duties . | Traditional methods for virus detection often rely on specific attributes , such as DNA sequences , of the viruses and therefore they not only require a priori knowledge of the agent in question , but they also are generally very specific in nature , capable of detecting viruses only from within a specific family , for example . Nextgen sequencing shows much promise for detection/diagnostic applications because of its ever-increasing throughput , decreasing cost , and unbiased nature . We investigated the applicability of 454 pyrosequencing for viral surveillance of insect populations , using Aedes aegypti mosquitoes experimentally inoculated with Dengue virus type 1 ( DENV-1 ) and calculated what proportion of the total nucleic acid from crushed mosquitoes was contributed by the virus . We concluded that 454 pyrosequencing is capable of detecting even very small amounts of a known virus from within a pool of infected and noninfected mosquitoes , but for the amount of sequencing reads required to detect the virus , this technique may currently be too cost-prohibitive for use in large-scale surveillance efforts . Interesting byproducts of our study included a glimpse into what symbiotic organisms Ae . aegypti may harbor , as well as what genes may be differentially expressed in a DENV-1-infected versus noninfected mosquito . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression",
"virology/emerging",
"viral",
"diseases",
"genetics",
"and",
"genomics/gene",
"expression"
] | 2010 | Arbovirus Detection in Insect Vectors by Rapid, High-Throughput Pyrosequencing |
Trypanosoma cruzi is the obligate intracellular parasite that causes Chagas disease . The pathogenesis of this disease is a multifactorial complex process that involves a large number of molecules and particles , including the extracellular vesicles . The presence of EVs of T . cruzi was first described in 1979 and , since then , research regarding these particles has been increasing . Some of the functions described for these EVs include the increase in heart parasitism and the immunomodulation and evasion of the host immune response . Also , EVs may be involved in parasite adhesion to host cells and host cell invasion . EVs ( exosomes ) of the Pan4 strain of T . cruzi were isolated by differential centrifugation , and measured and quantified by TEM , NTA and DLS . The effect of EVs in increasing the parasitization of Vero cells was evaluated and the ED50 was calculated . Changes in cell permeability induced by EVs were evaluated in Vero and HL-1 cardiomyocyte cells using cell viability techniques such as trypan blue and MTT assays , and by confocal microscopy . The intracellular mobilization of Ca2+ and the disruption of the actin cytoskeleton induced by EVs over Vero cells were followed-up in time using confocal microscopy . To evaluate the effect of EVs over the cell cycle , cell cycle analyses using flow cytometry and Western blotting of the phosphorylated and non-phosphorylated protein of Retinoblastoma were performed . The incubation of cells with EVs of trypomastigotes of the Pan4 strain of T . cruzi induce a number of changes in the host cells that include a change in cell permeability and higher intracellular levels of Ca2+ that can alter the dynamics of the actin cytoskeleton and arrest the cell cycle at G0/G1 prior to the DNA synthesis necessary to complete mitosis . These changes aid the invasion of host cells and augment the percentage of cell parasitization .
Trypanosoma cruzi is an intracellular protozoan parasite that causes Chagas disease or American trypanosomiasis . An estimated 8 million people are infected with this parasite worldwide , with some 300 , 000 new cases and 15 , 000 deaths annually [1] . T . cruzi has a life cycle that includes mammals and blood-sucking bugs ( Hemiptera , Reduviidae ) as hosts . Humans can be infected through the insects faeces , by vertical ( congenital ) transmission , transmission by blood transfusions , organ transplants , or oral contamination via tainted fluids and foods [2] . Chagas disease displays symptomatic and pathological variations among infected individuals [3] but is characterized by an acute as well as a chronic stage . During the chronic stage , approximately 30% of the patients develop significant complications , which may include megasyndromes of the gastrointestinal tract ( such as megacolon or megaesophagus ) , neurological complications , and cardiomyopathy [4–7] . The pathogenesis of Chagas disease is a multifactorial process . The molecular invasion mechanisms by T . cruzi trypomastigotes ( T ) and the associated regulatory pathways have been intensely investigated for many years [8] . A large number of molecules have been involved and are described as part of the secretome of T . cruzi [9] . Some of them are included in extracellular vesicles ( EVs ) . EVs are small membrane-bound vesicles classified based on their size , biogenesis , and composition [10] in: a ) exosome-like EVs ( 20–100 nm ) , b ) ectosome-like EVs ( 100–1000 nm ) and c ) apoptotic blebs ( >1000 nm ) [9 , 11] . The presence of EVs of T . cruzi was first described in 1979 by da Silveira et al . , who demonstrated the secretion of plasma-membrane vesicles by T . cruzi epimastigote forms [12] . These vesicles were later detected by Gonçalves et al . ( 1991 ) in host-cell-derived trypomastigotes [13] . Since then , numerous publications concerning EVs have appeared , demonstrating their role in cell-to-cell communication , pathogenesis , evasion of the immune response and diagnosis . The cargo of EVs of T . cruzi contain proteins involved in host-parasite interactions , signalling , trafficking , and membrane fusion , transporters , oxidation-reduction , etc . [9] . Small RNAs derived from tRNAs and rRNAs have also been reported [14] . Some of the functions described for these EVs include the increase in heart parasitism and the immunomodulation of the host response [15]; the evasion of innate immunity [16]; and the induction of the release of EVs by the host blood cells that are involved in inhibiting complement-mediated lysis [17] . Also , EVs may be involved in parasite adhesion to host cells and host-cell invasion [15 , 17–20] . Recently , EVs have proved useful in evaluating disease severity as well as vaccine and drug effectiveness against chagasic cardiomyopathy [21] . However , little is known about the capacity of EVs to modulate the host-cell conditions . In this sense , the present work seeks to elucidate certain effects exerted by T . cruzi EVs over the parasite establishment inside the host cell .
Vero ( ECACC 84113001 ) and 3T3 cells ( CRL 1658 ) were cultured in Nunc cell-culture flasks of 75 cm2 surface area ( Thermo Fischer Scientific , USA ) in Modified Eagle’s Medium ( MEM ) ( Sigma , USA ) supplemented with 10% foetal bovine serum ( Gibco , USA ) previously inactivated at 56°C for 30 min ( IFBS ) plus antibiotics ( penicillin 100 U/mL , streptomycin 100 μg/mL ) . The cultures were maintained at 37°C , in a moist atmosphere enriched with 5% CO2 . HL-1 cardiac muscle-cell line was grown as described above , using Claycomb medium supplemented with 10% IFBS , norepinephrine 0 . 1 mM , L-glutamine 2 mM and antibiotics ( penicillin 100 U/mL , streptomycin 100 μg/mL ) . The cell cultures were routinely monitored for Mycoplasma by PCR . Vero cells were initially infected with purified metacyclic trypomastigotes of the Pan4 ( Tc Ia + Tc Id ) strain of T . cruzi obtained in vitro , according to de Pablos et al . ( 2011 ) [22] . After 120 h of the intracellular development of the parasite , tissue-culture cell-derived trypomastigotes ( TcT ) were harvested by centrifugation . Parasites were collected routinely every 120 h from the infected cell monolayer . The culture medium was centrifuged at 3 , 000 xg for 5 min and the pellet with the parasites was washed in PBS four times . To obtain the EVs from the TcT , we followed the procedure described previously by de Pablos et al . ( 2016 ) [18] , with some modifications . Parasites were incubated for 5 h at 37°C in RPMI medium ( Sigma , USA ) buffered with 25 mM HEPES at 7 . 2 and supplemented with 10% exosome-free IFBS . Afterwards , parasites were removed by centrifugation at 3 , 500 xg for 15 min and the supernatant was collected and centrifuged at 17 , 000 xg for 30 min at 4°C . This supernatant was filtered through a 0 . 22 μm pore filter ( Sartorius , Germany ) and ultracentrifuged at 100 , 000 xg for 16–18 h to obtain the EVs ( mostly exosomes ) . All the steps were performed in an ultracentrifuge Avanti J-301 ( Beckman Coulter , USA ) with a JA-30 . 50 Ti rotor . The resulting pellet was washed three times in PBS in an ultracentrifuge Sorwal WX80 ( Thermo Fisher Scientific , USA ) with F50L-24 x 1 . 5 fixed-angle rotor and resuspended in 100 μL PBS . The isolation procedure was evaluated by Transmission Electron Microscopy ( TEM ) , Nanoparticle Tracking Analysis ( NTA ) and Dynamic Light Scattering ( DLS ) . The proteins from the EVs were quantified using the Micro-BCA protein assay kit ( Thermo Fischer Scientific , USA ) , using bovine-serum albumin as standard . Viability of the TcT after shedding of EVs was evaluated using the trypan blue exclusion test . After 5 h , no significant death was detected and over 99% of the parasites were viable . To demonstrate the specificity of the effects of the EVs from the TcT of T . cruzi and to evaluate the effect of the EVs of trypomastigotes of the Pan4 strain over the infection of cells with T . cruzi from a different DTU and another intracellular microorganism , we performed the isolation of EVs from Crithidia mellificae and the 3T3 cell line ( a fibroblast cell line ) and evaluated the effect of these EVs over the parasitization percentages of cells infected with T . cruzi Pan4 . We also employed T . cruzi 4162 strain ( Tc IV ) and tachyzoites of Toxoplasma gondii RH for the infection of cells previously incubated with EVs of T . cruzi Pan4 . For the isolation of EVs from choanomastigotes of Crithidia mellificae , 1x107 parasites were incubated for 24 h at 28°C in LIT medium . Nunc cell-culture flasks of 75 cm2 surface area ( Thermo Fischer Scientific , USA ) with confluent monolayers of 3T3 cells were washed 3 times with MEM without IFBS and the cells were incubated for 24 h at 37°C with MEM ( Sigma , USA ) . After 24 h of incubation , the culture media were collected and centrifuged at 3 , 500 xg for 15 min and the obtained supernatants were handled the same way as for the isolation of EVs from TcT . DLS and the quantification of the protein concentration of these samples using the Micro-BCA protein assay kit ( Thermo Fischer Scientific , USA ) were performed as described above . The purification of metacyclic trypomastigotes of T . cruzi 4162 strain was also performed according to de Pablos et al . ( 2011 ) [22] and TcT were obtained after the infection of Vero cells as described . Tachyzoites of Toxoplasma gondii RH strain were maintained in our laboratory by serial passage , in semiconfluent monolayers of Vero cells and cultured in the same conditions as T . cruzi . The egressed parasites were harvested , centrifuged at 5 , 000 xg for 10 min , washed three times in PBS and added to the cell cultures in a ratio 5:1 ( parasites:cell ) . To confirm the presence of EVs in our samples , we resuspended an aliquot of the pellet in 0 . 1 M Tris HCl ( pH 7 . 2 ) and 5 μL each sample were adsorbed directly onto Formvar/carbon-coated grids and stained with 2% ( vol/vol ) uranyl acetate , for the direct observation in a TEM , LIBRA 120 PLUS Carl Zeiss microscope . The diameter of the EVs was measured by ImageJ 1 . 41 software . Distribution , size , and concentration of T . cruzi EVs from trypomastigotes was determined by measuring the rate of Brownian motion according to the particle size , using a Nanosight NS300 ( Malvern Instruments , UK ) . This system was equipped with a sCMOS camera and a blue 488 nm laser beam . Samples were diluted 1/100 just before the analysis , in low-binding Eppendorf tubes with PBS and the measurements were performed at 25°C . For data acquisition and information processing , we used the NTA software 3 . 2 Dev Build 3 . 2 . 16 . The particle movement was analysed by NTA software with the camera level at 16 , slider shutter at 1200 , and slider gain at 146 . To confirm the results obtained by NTA , we also performed Dynamic Light Scattering ( DLS ) of the EVs of trypomastigotes , choanomastigotes and cells using a Zetasizer nano ZN90 ( Malvern Instruments , UK ) . Samples were prepared the same way as described for the NTA and the measurements were also performed at 25°C . For data acquisition and information processing , the Zetasizer Ver . 7 . 11 software was employed . The presence of some molecules without orthologues in other organism and involved in the invasion process of T . cruzi was evaluated in EVs by Western blotting . Briefly , 300 μg of EVs isolated from TcT of the Pan4 strain were resolved by SDS-PAGE , transferred to a nitrocellulose membrane and blocked overnight with 5% non-fat milk in PBS-0 . 1% Tween 20 . Primary antibodies anti-cruzipain ( 1:3 , 000 ) , anti-TS ( mAb 39 ) ( 1:1 , 000 ) , and anti-MASPs ( signal peptide , SP ) ( 1:1 , 000 ) [16] were incubated overnight at 4°C . The membranes were washed with PBS-0 . 1% Tween 20 and incubated for 1 h with goat anti-mouse IgGs conjugated with peroxidase ( 1:1 , 000 ) ( Dako Agilent Pathology Solutions , USA ) in the case of TS and MASPs and goat anti-rabbit IgGs conjugated with peroxidase ( 1:2 , 000 ) in the case of cruzipain ( Dako Agilent Pathology Solutions , USA ) . The reaction was visualized using Clarity ECL Western substrate ( BioRad , Spain ) in a ChemiDoc Imaging system ( BioRad , Spain ) . Cultures of 5x104 Vero cells were grown in MEM supplemented with 10% IFBS over round 13-mm coverslips ( Marienfeld , Germany ) , in Nunc 24-well plates ( Thermo Fischer Scientific , USA ) for 24 h at 37°C and 5% CO2 . After this time , coverslips with cells were washed 3 times in MEM and incubated for 2 h with 0 . 1 , 0 . 25 , 0 . 5 , 1 and 2 . 5 μg/mL EVs in MEM . After the incubation , cells were infected with T . cruzi trypomastigotes of the Pan4 strain , at a parasite:host cell ratio of 5:1 . After 4 h , parasites were removed and the cells were washed and maintained in culture for 24 h . The cultures were fixed with methanol and stained with Giemsa . Parasitization percentages and parasitization indexes ( number of amastigotes per cell ) were calculated after counting at least 400 cells . The invasion assays were also performed using T . cruzi EVs submitted to thermal and chemical treatments . For the thermal treatments , EVs were incubated in a water bath at 50°C , 70°C , and 90°C for 30 min . For the chemical treatments , EVs were incubated with the proteolytic enzymes trypsin ( 0 . 5 mg/mL ) and proteinase K ( 0 . 5 mg/mL ) for 1 h at 37°C and with sodium periodate ( 10 mg/mL ) for 20 h at room temperature , in the dark , to reduce the glycoconjugates surrounding the EVs . After the treatments , EVs were washed twice in PBS by ultracentrifugation at 100 , 000 xg for 1 h , incubated with the Vero cell cultures for 2 h . The protocol for cell infection was followed as described above . The specificity of the effects of the EVs isolated from trypomastigotes of T . cruzi Pan4 and the effect of EVs from TcT of the Pan4 strain over the invasion of another T . cruzi strain and intracellular parasite were evaluated . For these experiments , cultures of 5x104 Vero cells were grown the same way as described for the invasion assays using the Pan4 strain . The cells were incubated for 2 h at 37°C with 0 . 38 μg/mL EVs from Crithidia mellificae or 3T3 cells . After this time , the cells were infected with T . cruzi trypomastigotes of T . cruzi Pan4 ( parasite:host cell ratio of 5:1 ) and after 4 h of interaction , the parasites were removed . The cells were washed and maintained in culture for 24 h , when they were fixed with methanol and stained with Giemsa . Parasitization percentages and indexes were calculated . The effect of EVs of the Pan4 strain of T . cruzi over the infection of cells with trypomastigotes of T . cruzi 4162 strain ( Tc IV ) and tachyzoites of T . gondii RH were performed . In this case , cells were incubated for 2 h with EVs of T . cruzi Pan4 and then infected with trypomastigotes of T . cruzi 4162 strain or tachyzoites of T . gondii RH in a parasite:cell ratio of 5:1 . After 4 h , the parasites were removed , the cells were washed and maintained in culture for 24 h . Parasitization percentages and indexes were calculated after the evaluation of the cells by Giemsa stain . To assess the potential capacity of EVs to permeabilize cells , we cultured 5 x 104 Vero cells in 12-well plates as described above . The potential changes in permeability during or after the EVs-cell interaction was evaluated using the Aspergillus giganteus ribotoxin α-sarcin , a ~17 kDa protein that inhibits protein biosynthesis when the cells are previously permeabilized [23–25] . Briefly , after 24 h of culture , cells were washed 3 times with MEM and incubated with 0 . 38 μg/mL EVs in MEM for 2 h . Cells were washed once and 20 μM of α-sarcin ( Sigma , USA ) was added for 4 h . After this time , cells were washed 4 times and subsequently incubated with MEM supplemented with 10% IFCS . In a parallel assay , the EVs and α-sarcin were added simultaneously to the cell culture . Viability of the cells was determined using the trypan blue exclusion test as well as the MTT viability assay ( Sigma , USA ) . Cell viability was also determined after the incubation of the cell cultures with EVs , α-sarcin and cells without any treatment as negative controls . The HL-1 cell line was derived from atrial cardiomyocytes is a cell line that maintains a series of cardiac characteristics such as morphological , biochemical , and electrophysiological properties in vitro [26] . On round 13-mm coverslips , 5x104 cells were grown in Claycomb medium supplemented with 10% IFBS , norepinephrine 0 . 1 mM , L-glutamine 2 mM and antibiotics . After 24 h of culture , cells were washed and incubated with 0 . 38 μg/mL EVs in MEM for 2 h . Afterwards , coverslips were washed 3 times and fixed with a solution of 2% paraformaldehyde and 1% glutaraldehyde for 1 h , washed 3 times in PBS and blocked with a solution containing 1% BSA and 0 . 3 M glycine in PBS , for 1 h . Cells were washed 3 times and incubated with an anti-β2-adrenergic receptor primary antibody ( 1:500 ) ( Thermo Fisher Scientific , USA ) for 1 h . Afterwards , cells were washed 3 times and incubated in the dark , with a goat anti-rabbit IgG antibody conjugated with Alexa Fluor 647 ( 1:500 ) ( Thermo Fisher Scientific , USA ) for 1 h , at 37°C . Finally , samples were washed 4 times , mounted in Vectashield mounting medium with DAPI ( Vector Laboratories , USA ) and imaged with a Leica DM5500B inverted microscope ( Leica Microsystems , Germany ) . HL-1 cells cultured and fixed as described above but treated with a solution of NP-40 in 10 mM citric acid ( pH 6 ) were employed as positive control of permeabilization of the assay . Vero cells were grown overnight in MEM without phenol red plus IFBS and in MEM without phenol red plus IFBS and 2 . 5 μM EDTA , on μ-slide ibidi multichamber dishes . Cells were washed 3 times in MEM without phenol red and incubated at 37°C for 20 min with Fluo4-AM ( Thermo-Fisher , USA ) in 1 ) MEM without phenol red , 2 ) MEM without phenol red plus 2 . 5 μM EDTA and 3 ) a culture medium similar to MEM but without Ca2+ and Mg2+ . Fluo-4 is an indicator that exhibits greater fluorescence upon binding intracellular free Ca2+ . It presents an AM grouping ( acetoxymethyl ester ) that , when internalized , is cleaved by intracellular esterases and released to bind to cytoplasmic calcium [20] . After 20 min of incubation of the cells with Fluo-4 AM , EVs of TcT of T . cruzi Pan4 were added to the cells and followed-up in time until 25 min of interaction . Basal controls of fluorescence in cells before the application of the stimulus with EVs were included . Images were taken every 5 min with a confocal microscope Nikon A1 ( Nikon Instruments , The Netherlands ) equipped with 10x , 20x multi-immersion , 40x oil , 60x oil , and 60x water objectives and a system of cell incubation at 37°C with enriched atmosphere with 5% CO2 . The Fluo4 probe was excited at 494–506 nm and light emission was detected at 516 nm . The fluorescence intensity was analysed and normalized with reference to the basal fluorescence using NIS Elements Software ( Nikon Instruments , The Netherlands ) . The analysis of images was performed using Fiji software ( Fiji is Just Image J ) . Controls of cells incubated with A23187 ( a calcium ionophore ) and 3-isobutyl-1-methylxantine ( IBMX ) ( a non-specific inhibitor of cAMP and cGMP phosphodiesterase that induces calcium release from intracellular stores ) were included . Cultures of 5x104 Vero cells were seeded over round 13-mm coverslips ( as described above ) and allowed to attach to the coverslips overnight . The cells were washed 3 times with MEM and incubated in MEM during different times with 0 . 38 μg/mL EVs in MEM . After this incubation step , coverslips were washed 3 times and fixed with cold acetone ( Scharlab , Spain ) for 15 min at -20°C . After the fixation step , coverslips were washed 3 times with PBS and permeabilized in a solution of 0 . 1% Triton X-100 ( Sigma , USA ) for 10 min . The cells were washed 3 times and blocked with 1% BSA and 0 . 3 M glycine in PBS for 1 h in PBS . After this time , the cells were washed and incubated with 5 μg/mL of a vimentin polyclonal antibody ( 1:300 ) ( Thermo Fischer Scientific , USA ) for 1 h , washed 4 times and incubated in the dark , with a goat anti-rabbit IgG antibody conjugated with Alexa Fluor 647 ( 1:500 ) ( Thermo Fisher , USA ) for 1 h , at 37°C . The coverslips were washed 4 times and stained with a solution of phalloidin , fluorescein isothiocyanate labelled ( Sigma , USA ) for 30 min . Samples were finally mounted in Vectashield mounting medium with DAPI ( Vector Laboratories , USA ) and imaged in a Leica DM5500B inverted microscope ( Leica Microsystems , Germany ) . Images were captured 15 min , 30 min , 120 min , and 24 h after EVs-cell contact . Cells not treated with EVs and cells incubated with the final supernatant from the EVs purification medium were employed as controls . Vero cells ( 1x105 ) were synchronized according to the method described previously by Osuna et al . ( 1984 ) [27] . Cells were seeded in 6-well plates with a culture medium with 25 mM thymidine for 12 h , when the medium was replaced with MEM + 10% IFCS . Afterwards , cells were washed with MEM and 1 h later , EVs were added directly to each well . One hour after this EV-cell contact , cells were washed and maintained for 2 and 8 h . Afterwards , the culture medium of the corresponding wells was removed , the cells were washed with PBS , fixed with 70% cold ethanol and incubated with a solution ( 0 . 2 M Na2HPO4 , 0 . 1 M citric acid , pH 7 . 8 ) for 15 min at 37°C . Cells were then centrifuged , washed with PBS and resuspended in 250 mL of a solution of propidium iodide ( 40 mg/mL ) and RNAse ( 100 mg/mL ) for 30 min at 37°C in the dark , according to Carrasco et al . ( 2014 ) [28] . Finally , the samples were analysed in a FACS Calibur ( BD Biosciences , San Jose , CA , USA ) flow cytometer . The results were analysed with FlowJo software ( v 7 . 6 . 5 , Tree Star , Inc . ) . Phosphorylation of the protein Rb after the incubation of cells with EVs was evaluated by immunoblotting . Briefly , 1x105 cells were grown in 6-well plates for 24 h . Cells were washed with MEM and incubated with EVs for 5 , 10 , 30 and 60 min . After this session , the cells were washed with PBS and lysed in RIPA lysis buffer with a protease inhibitor cocktail ( Roche , Switzerland ) for 15 min . Cells were harvested with a cell scraper and centrifuged at 14 , 000 xg for 10 min at 4°C . Supernatants were transferred to new Eppendorf tubes and stored at -20°C . The protein from cell lysates was quantified using the Bradford reagent ( Sigma , USA ) and 90 μg of protein from cell lysates were resolved by SDS-PAGE , transferred to a nitrocellulose membrane and blocked overnight with 5% non-fat milk in PBS-0 . 1% Tween 20 . Rb ( 1:2 , 000 ) and phospho-Rb ( 1:1 , 000 ) ( Cytoskeleton , USA ) primary antibodies ( Sigma , USA ) were incubated overnight at 4°C . Tubulin antibody ( 1:5 , 000 ) ( Cytoskeleton , USA ) was used as the loading control . The membranes were washed with PBS-0 . 1% Tween 20 and incubated for 1 h with goat anti-mouse IgGs conjugated with peroxidase ( 1:1 , 000 ) ( Dako Agilent Pathology Solutions , USA ) in the case of Rb , goat anti-rabbit IgGs conjugated with peroxidase ( 1:2 , 000 ) in the case of phospho-Rb ( Dako Agilent Pathology Solutions , USA ) , and rabbit anti-sheep IgGs conjugated with peroxidase ( 1:5 , 000 ) ( Dako Agilent Pathology Solutions , USA ) in the case of tubulin . The reaction was also visualized using Clarity ECL Western substrate ( BioRad , Spain ) in a ChemiDoc Imaging system ( BioRad , Spain ) . Quantification values represent the means of two or more independent experiments , each performed in triplicate . Means and standard deviations of the EVs size ( NTA ) , percentage of infected cells ( invasion assays ) and percentage of live cells ( permeabilization assays ) were calculated . One-factor ANOVAs were performed to detect significant differences between cells treated with EVs and control cells in the case of permeabilization and cell-cycle assays . Multiple post hoc comparisons were performed using the Tukey-Kramer test on GraphPad Prism 5 Software ( USA ) . Values with p<0 . 0001 were considered statistically significant ( *** ) .
After the incubation of 1x107 trypomastigotes of the Pan4 strain for 5 h at 37°C in the culture medium for the release of EVs , 12 μg of total protein of EVs were obtained . The isolation of the EVs by the ultracentrifugation protocol described was evaluated by TEM , NTA and DLS and the presence of surface molecules of T . cruzi in these samples was confirmed by Western blotting ( S1 Fig ) . Most of the particles visualized by negative staining under TEM were of 30–100 nm size ( S1A Fig ) . Analyses by NTA revealed a majority of EVs with a size of 70 . 7 ± 7 . 3 nm ( S1C Fig ) and a concentration of approximately 5x1010 ± 3 . 9x109 particles/mL . In the DLS analyses , two populations of EVs with different sizes were observed for TcTs , choanomastigotes and 3T3 cells ( S1D , S1E and S1F Fig ) ; EVs of T . cruzi showed a population of 23 . 05 ± 6 . 96 nm and a population of 55 . 74 ± 13 . 97 nm ( S1D Fig ) . Western blotting confirmed the presence of cruzipain , trans-sialidase and MASPs ( SP ) in the EVs of T . cruzi Pan4 ( S1B Fig ) . To evaluate the effect of EVs of TcT from the Pan4 strain in host-cell invasion , we tested different doses measured in total μg/mL of protein . After 2 h of incubation of the cells with the EVs , the infection with T was performed in a ratio 5:1 ( T:cell ) for 4 h . Counts were performed 24 h later , as described in the Methods section . The results indicate that the parasitism increased the most when the cells were treated with 0 . 5 μg/mL of EVs . These values in percentage of parasitization ( 88 . 88 ± 3 . 73 ) significantly differed with respect to the other doses analysed ( S2A Fig ) . From these results , the Effective Dose 50 ( ED50 ) for subsequent trials was set as 0 . 38 μg/mL . Regarding the effect of EVs on cells over time ( the increase in cell parasitization ) , these were treated with 0 . 38 μg/mL of total protein of EVs for 2 h . After the cells were washed three times in medium without serum , they were infected at different time points after the treatment with EVs at a T:cell ratio 5:1 , as described above . The results ( S2B Fig ) show that the difference between the percentage of parasitization of the cells treated with the EVs vs . the percentage of parasitization of non-treated control cells were statistically significant up to 8 h after the treatment . The effects were more evident at 2 and 4 h , when the increase in the percentages was higher compared to untreated cultures . This effect was not appreciated when the cells were infected 24 h after the treatment with the EVs . The parasitization indexes ( number of parasites per cell ) calculated also differed . For example , in the case of the cells incubated with EVs for 2 h was 2 . 78 ± 0 . 55 , an index that was twice the parasitization index of the control infected cells without the previous treatment ( 1 . 33 ± 0 . 18 ) . A series of experiments were performed to study whether the thermal treatment , the treatment with proteolytic enzymes or the reduction of the glycoconjugates surrounding the EVs alters their ability to induce higher levels of parasitization in the host cells with which they interact . Results are shown in S2C and S2D Fig . The thermal treatment of the EVs at 50°C , 70°C and 90°C annuled the action of the EVs on the cells , so the increase in the percentage of parasitization was not observed . The enzymatic treatment with the two proteases employed and the treatment with sodium periodate ( for the reduction in the content of carbohydrates of the EVs ) also inhibited the capacity of increasing the cell parasitization by trypomastigotes of T . cruzi in the cultures . Finally , Fig 1A , 1B and 1C show the effect of the incubation of cells with EVs of C . mellificae and 3T3 cells prior to the infection with T . cruzi Pan 4 and the increase in the infection of cells when these are incubated with EVs isolated from trypomastigotes of the Pan4 strain and then infected with TcT of the 4162 strain or tachyzoites of T . gondii RH . In Fig 1A is possible to observe that the incubation of cells with EVs from another source different than T . cruzi did not boosted an increase in the percentage of infected cells as it happens when the cells are in contact with EVs of T . cruzi Pan4 prior to the infection . When the cells were incubated with EVs of C . mellificae and 3T3 cells , the percentage of infected cells obtained were 29 . 3 ± 5 . 0 and 29 . 8 ± 4 . 3 , respectively . These results did not differ significantly from the results obtained in the case of the cells infected only with TcT of T . cruzi Pan4 without the previous treatment with the EVs ( 36 . 5 ± 3 . 8 ) . In these experiments , the parasitization indexes obtained were 1 . 31 ± 0 . 08 in the cells incubated with EVs of C . mellificae , 1 . 79 ± 0 . 28 in the cells incubated with EVs of the 3T3 cell line and 1 . 54 ± 0 . 31 in the cells infected with TcT without the prior incubation with EVs , but it rose to 2 . 60 ± 0 . 14 when the cells were previously treated with EVs of T . cruzi Pan4 . This Figure also shows that this increase in the parasitization percentage and index of cells is independent of the strain of T . cruzi employed to infect the cells , as we proved that the infection with TcT of a strain that is classified as Tc IV almost doubled the percentage of infected cells not incubated previously with the EVs ( 53 . 5 ± 6 . 0 vs 32 ± 2 . 6; parasitization indexes: 2 . 10 ± 0 . 14 vs 1 . 39 ± 0 . 21 , respectively ) . In the case of the cells treated with EVs of T . cruzi Pan4 prior to the infection with tachyzoites of T . gondii RH , in Fig 1B is possible to observe a slight , non-significant increase in the percentage of infected cells in those incubated with EVs of T . cruzi prior to the infection with tachyzoites , when compared to the cells infected with the parasite without the previous incubation with the EVs of T . cruzi ( 56 . 50 ± 4 . 80 vs . 46 . 50 ± 4 . 93 ) . A toxin of Aspergillus giganteus , α-sarcin , constitutes a ribotoxin with a molecular weight of 16 . 8 kDa that acts at the ribosomal level , inhibiting the protein synthesis . This toxin has been used to determine the potential permeabilization induced in the cells on the entry of certain viruses [29] . The cells were treated in two ways: i ) they were simultaneously incubated with EVs of T . cruzi Pan4 and α-sarcin or ii ) they were incubated with EVs of T . cruzi Pan4 for 2 h and then with α-sarcin for 4 h . After 24 h of the treatment , the trypan blue exclusion assay was performed . Vero cells treated with EVs and α-sarcin registered mortality percentages of 76 . 10 ± 6 . 81 ( simultaneous incubation ) and 82 . 20 ± 10 . 17 ( separated incubation of EVs and the toxin ) . The control cells incubated only with the toxin or with the EVs showed mortality percentages of 16 . 90 ± 4 . 10 and 17 . 23 ± 7 . 73 , respectively , while the percentage of viability of the untreated control culture cells was 15 . 56 ± 7 . 67 . These results were confirmed using the MTT cell-viability assay , as shown in Fig 2A . In this case , the percentage of viability of the cells treated with EVs and the toxin was 37 . 36 ± 15 . 39 , while the same percentages in the cells treated only with the toxin or the EVs alone were 90 . 10 ± 7 . 03 and 95 . 30 ± 3 . 30 , respectively . The percentage of viability in the untreated control cell cultures was 99 , 0 ± 0 . 1 . Fig 2B shows the appearance of the different cell monolayers preincubated with EVs and treated with α-sarcin and the different control cells at 24 h of the treatment . The beta-2 adrenergic receptor ( β2 adrenoreceptor ) , also known as ADRB2 , is a cell membrane-spanning beta-adrenergic receptor that binds epinephrine or adrenaline , whose signalling , via a downstream L-type calcium channel interaction , mediates physiological responses such as smooth-muscle relaxation and bronchodilation [30] . Using an antibody that recognizes an epitope between the amino acids 340–413 ( which corresponds to the intracytoplasmic domain of the receptor ) , we demonstrated that the EVs of T . cruzi can alter and permeabilize the cell membrane , exposing the epitope to the antibody , as shown in Fig 2C . In the case of the untreated cells , this effect was not detected . An image similar to that of the cells treated with the EVs resulted in the cultures that , prior to the incubation with the antibody , were permeabilized with a solution containing NP-40 . The time-course measurements of the intracellular calcium levels of Vero cells treated with the EVs of T . cruzi Pan4 are shown in Fig 3 . Results show that , when the cells were incubated in a culture medium with Ca2+ and Mg2+ , the fluorescence levels increased up to 3 . 83 ± 0 . 62 times the initial values as soon as 10 minutes of interaction . Moreover , when the interaction was performed in a culture medium depleted of Ca2+ and Mg2+ , there was also a progressive increase in the fluorescence levels . For example , at 10 min of the interaction , there was a 1 . 40 ± 0 . 39-fold increase in the fluorescence levels of the cells , which could correspond to a mobilization of the ions from their intracellular deposits to the cytoplasm . In the case of the cells incubated in MEM with EDTA , a calcium chelator , there was also a 1 . 69 ± 0 . 01 increase in the fluorescence levels at 10 minutes of incubation , when the fluorescence intensity started to decrease . At this point , it is possible to observe a different pattern of distribution of the fluorescence , with the appearance of a more granulated cytoplasm . The calcium ionophore A23187 was used as the control for the assays of the cells incubated in the medium containing Ca2+ and Mg2+ and prompted a 64 . 80-fold rise in the fluorescence levels at 25 min . The cAMP phosphodiesterase inhibitor , 3-isobutyl-1-methylxanthine ( IBMX ) was used as the control for the induction of ion output from the intracellular Ca2+ deposits to the cytoplasm and prompted a 2 . 37-fold rise in the fluorescence levels at 10 min . The analysis of the effect of the EVs of T . cruzi Pan4 over the actin cytoskeleton is shown in Fig 4A , 4B and 4C . The disorganization of the actin filaments was visible in the cells showing greater globular actin ( GA ) from 15 min up to 120 min after the treatment with EVs ( Fig 4A , 4B and 4C ) . On the other hand , vimentin appeared to withdraw from the areas in the cytoplasm where the actin is disorganized and concentrated in the parts of the cytoplasm where GA is less patent . The morphology of the treated cells appeared to be altered and filopodia ( F ) were visible , giving a dendritic aspect to the cell . These effects were reversible 24 h after the treatment; both the cytoskeleton images and the morphology of the cells incubated with EVs were similar to that of the control cells not treated with EVs . The influence of the EVs of T . cruzi Pan4 on Vero cells cycle was analysed with Vero cells previously synchronized in the S phase and treated as described in Methods . The changes in the cell cycle were analysed by flow cytometry 2 and 8 h after the addition of the EVs . Fig 5A shows the percentage of cells in the different phases of the cell cycle . At 8 h after the addition of the EVs , the percentage of cells increased at phases G0/G1 and decreased at phase S . Fig 5B and 5C show the results of the levels of the protein of retinoblastoma ( pRb ) in its phosphorylated and non-phosphorilated states , in the cells treated with EVs of T . cruzi Pan4 and the untreated control cells . The protein expression increased for the phosphorylated pRb from the first minutes of the interaction of the EVs with the cells , reaching the maximum phosphorylated state at 15 min of the interaction and declining to values similar to those of the control cells at 60 min of treatment .
The cell-invasion process of T . cruzi has been widely studied . Numerous mechanisms are known to be involved in preparing the cell to induce the endocytosis of trypomastigotes into non-phagocytic cells [31–36] and among the natural agents that induce massive entry and cellular infection are the EVs secreted by trypomastigotes . EVs from trypomastigotes of the Pan4 strain were first isolated in 2016 , when de Pablos et al . confirmed that the C-terminal region of MASPs proteins is present in the EVs secreted by the trypomastigotes derived from tissue-culture cells [18] . Following the methodology described above using differential centrifugation and then Nanoparticle Tracking analysis , we obtained an homogeneous population of EVs under our experimental conditions . Also , the yield was higher than that of other authors using others strains or forms of the parasite . Trocoli Torrecilhas et al . ( 2009 ) employed 5 μg of protein from 1x105 trypomastigotes of the Y strain [15] . Garcia Silva et al . ( 2014 ) have reported a protein yield of the vesicular fraction of 1 . 2 μg per 1x1010 epimastigotes from the DM28c strain [14] . In our case , we found 12 μg of protein of EVs after the incubation of 1x107 trypomastigotes of the Pan4 strain for 5 h at 37°C in the culture medium for the release of EVs . To track the time course of the effect of EVs in increasing cell parasitism , we performed infections at different times after incubating the cells with EVs and we found that the cells are still susceptible to increased parasitism at least for 8 h after the treatment ( p <0 . 0001 ) . However , 24 h after the incubation with the EVs , the percentage of infected cells in the treated and non-treated cultures were similar ( S2B Fig ) . On the other hand , we observed that increasing amounts of EVs can increase the percentage of infected cells with sigmoidal kinetics , reaching a maximum of 89% infected cells ( 0 . 5 μg/mL ) and ED50 of 0 . 38 μg/mL , a dose that was employed in subsequent experiments . In 2014 , Garcia Silva et al . observed that the treatment with small amounts of EVs ( 160 ng ) can prevent cells from appearing oversaturated in a tRNAGlu-derived 5′ halves visualization analysis by FISH [14] . In the same study , the authors determined that 30 min after the treatment of cells with EVs , a diffuse fluorescent cytoplasmic pattern appeared , which becomes granular after 2 h of treatment . Regarding the interaction and posterior infection of the cells after the treatment with the EVs , Cestari et al . ( 2012 ) pre-incubated Vero cells for 30 min before adding the trypomastigotes of the Sylvio X10/6 , DTU I strain [17] . This reflects that the incubation conditions with respect to the amount of EVs used and the incubation time vary among the different research groups , and it should be taken into account that the conditions selected by a researcher do not necessarily correspond to the conditions of a natural infection [14] . It has been demonstrated in vitro ( in non-phagocytic cells and monocytes ) , as well as in vivo , that EVs of T . cruzi increase the number of infected cells [15 , 17 , 37–38] . Under our experimental conditions , cells pretreated with EVs of T . cruzi Pan4 registered infection percentages of over 3 . 5-fold higher than for cells infected without the prior incubation with the EVs . The parasitization index ( number of parasite per cell ) was also two-fold that of the parasitization index of the control infected cells without prior treatment . In 2009 , Trocolli Torrecilhas et al . reported that T . cruzi trypomastigotes invade 5-fold more susceptible cells when these were preincubated with purified parasite EVs [15] . Cestari et al . ( 2012 ) also demonstrated that THP-1 derived plasma membrane vesicles ( ectosomes ) could simultaneously induce an increase in Vero cell invasion [17] . This invasion was dose dependent , non-specific for parasite strain or eukaryotic cell line , and dependent on the parasite infective stage . We also proved that the increase in parasitization is specific to T . cruzi trypomastigotes but non-specific for the parasite strain and that the incubation of cells with EVs from another trypanosomatid species or those from eukaryotic cells didn´t increased the percentages of parasitization . The interaction EVs-cell and the latter activity of EVs appears to depend on their binding through lectins to the plasma membrane and on a presumably enzymatic protein activity , given that the thermal and chemical treatments of EVs with trypsin , proteinase K , and sodium periodate drastically reduced parasitism of the cells with which they interacted . During the adhesion and invasion process of T to the host cells , a number of glycosylated molecules are expressed on the surface of the parasite . Examples include mucins , trans-sialidases , MASPs and the gp85 family of proteins [35] . Glycosylated proteins have also been detected in EVs by proteomic analyses [9 , 39] some with important activities and biological significance , such as trans-sialidases ( TS/SAPA , TC85 , gp82 , gp90 , CRP ) [40] , cruzipain [41] , gp63 , MASPs [42] , and other types of mucins [43] . As these proteins are located on the surface of the EVs , they can bind specifically to the proteins in the plasma membrane and this would explain why the reduction of the carbohydrates of the EVs after the sodium periodate treatment can affect the binding of EVs to the surface of the cells . On the other hand , some of these glycoproteins have enzymatic activities essential for the interaction of trypomastigotes during the invasion process and maybe in the interaction of EVs with the plasma membrane . It has been reported that the infection with some type of viruses lead to a permeabilization mechanism where the plasma membrane allows the entry of some high molecular weight molecules such as α-sarcin ( 16 . 8 KDa ) [44–46] , toxin that lacks a membrane receptor , unlike other toxins that affect the protein synthesis and are internalized via endocytosis [45 , 47] . The same effect was observed incubating cells with EVs of trypomastigotes of T . cruzi , as our results demonstrate that they can permeabilize Vero and HL-1 cell lines . Cell counts with trypan blue 24 h after the preincubation of cells with EVs and subsequent incubation with α-sarcin registered as much as 76 . 10% mortality . The control cells incubated only with α-sarcin , only with EVs , and without either showed percentages of cell death of 16 . 90% , 17 . 23% , and 15 . 56% , respectively . The percentages of mortality ( 100 - % of viability in Fig 2 ) of the cells using MTT were 62 . 64% in the case of the cells previously incubated with EVs and 9 . 90% in cells incubated only with the toxin . In 1990 , Castanys et al . have reported that infective metacyclic forms of T . cruzi secrete a glycoprotein involved in cell permeabilization that enabled the entrance of molecules such as α-sarcin [24] . Permeabilization was also evaluated in HL-1 cardiac muscle cells with confocal microscopy , using an antibody directed to an epitope of the β2-adrenergic receptor located in the intracytoplasmic region of the receptor anchoring . In this experiment , fluorescence was detected in the cytoplasm of cells previously incubated with EVs and in cells not treated with EVs but permeabilized with the detergent NP-40 . This permeabilization may be the result of changes in the cell membrane that allow the direct entry of the antibodies into the cytoplasm or by a transient disorganization of the membrane , capable of exposing the antigens present in inside the cell , with consequent exposure to the immune system . The presence of autoantibodies against β-adrenergic receptors in the serum of chagasic patients has been reported [48–51] , although authors have related such emergence to the recognition of exposed parts in the membrane due to cross reactivity to ribosomal acidic proteins P0 of the parasite [52] . A study related to the recognition of epitopes of the β-adrenergic receptors inserted into the inner side of the membrane by autoantibodies would be necessary to confirm the possible hypothesis that the permeabilization of the cardiac cells by the parasitic EVs lead to the exposure of these receptors to the immune system and then elicit the production of autoantibodies . The recognition and the invasion processes of the trypomastigotes to the host cells involve molecules over the surface of both the trypomastigote and the host cell . A ligand-receptor recognition occurs and generates in both a series of events that raises the intracellular Ca2+ levels due to the mobilization of the ions from the endoplasmic reticulum and the mitochondria [53–57] . This boost in calcium levels is also responsible for higher cAMP levels [58] , facilitating the release of EVs and their later fusion with the plasma membrane [59] . The treatment of Vero cells with EVs of T . cruzi Pan4 raised cytoplasmic Ca2+ levels and to determine the kinetics and origin of these Ca2+ ions , we treated cells with EVs and studied the result every 5 min under confocal microscopy . From 5 min of the treatment , fluorescence intensified in both culture media ( with and without calcium ) . This implies that the contact of Vero cells with EVs raises cytoplasmic Ca2+ levels that could come both from the intracellular deposits of calcium and the extracellular medium . The fluorescence pattern detected resembles the one in control cells treated with the xanthine IBMX for the Ca2+ mobilization from the intracellular deposits [60] or when the cells were incubated with the ionophore A23187 [61] , a compound that allows Ca2+ to enter cells from the culture medium . It has been demonstrated in different types of eukaryotic cells that the intracellular levels of calcium induce an asymmetric distribution of phospholipids in the plasma membrane by the activation of the enzymes scramblase and floppase . Then , phosphatidylserine and phosphatidylethanolamine are exposed in the outer side of the membrane and contribute to the activation of Ca2+-dependent proteases , followed by the release of EVs [62] . The exposure of anionic phospholipids to EVs strengthens the fusogenic properties of these vesicles , which could be a prerequisite for the release of the EV content . This could mean that the higher intracellular calcium levels and the changes in the distribution of phospholipids could explain the permeabilization induced in the host cell after the treatment with the EVs of T . cruzi , allowing the entrance of a toxin of ~17 kDa such as α-sarcin . Moreover , increases in intracellular Ca2+ could trigger a greater release of EVs from the cells exposed to the parasite [3 , 17 , 63] , as more calcium prompts a strong response of EV release in other cell lines [64–65] . Calcium ions also contribute to the reorganization of the cytoskeleton through the activation of cytoplasmic proteins such as calpain and gelsolin . These proteins cut the actin cytoskeleton protein network , allowing membrane budding and removing capping proteins at the end of the actin filaments [66–67] . A disruption in actin filaments and vimentin at the time of the invasion of cells with trypomastigotes has been demonstrated [68–69] using drugs like cytochalasin B and latrunculin , which affect the cytoskeletal structure and functions and , therefore , the entrance of the parasite in non-phagocytic cells [32–33 , 70] . It has been mentioned that the increase in intracellular Ca2+ leads to a rapid and transient reorganization of host-cell microfilaments , including the disassembly of the actin cytoskeleton , which is important for the entry of T into the host cells [71–73] . Studying the gene-expression changes caused by microvesicles of T . cruzi epimastigotes of the DM28c strain in mammalian host cells , Garcia Silva et al . ( 2014 ) observed an induction of a broad response , including the modification of the host-cell cytoskeleton and the extracellular matrix [74] . In fact , the regulation of actin cytoskeleton is one of the pathways identified as being affected by EV treatment in the profile of transcriptome changes [74] . Noting increased fluorescence in cells incubated with EVs in the presence of Fluo-4AM , we suspected that these changes in the Ca2+ mobilization induced by EVs could directly affect the actin cytoskeleton , as happens when the parasite begins to invade the host cell . Our results showed a clear disruption of host-cell actin from 15 min after the incubation with EVs , an effect that remains at 120 min but not 24 h after the treatment with EVs . Ferreira et al . ( 2006 ) indicated that different strains of Mt of T . cruzi can invade host cells through both actin cytoskeleton-dependent and independent routes , by engaging different surface molecules for attachment while triggering different signal-transduction pathways [34] . For example , host-cell invasion by the strain CL Mt , mediated mainly by the surface molecule gp82 , is associated with F-actin disassembly whereas the G strain is gp35/50-mediated invasion by strain G depends on target-cell actin cytoskeleton [34 , 75] . The analysis of the cell cycle events revealed how at 8 h after the addition of the EVs the percentage of cells in each of the cycle phases significantly differed when compared to control values , showing an arrest of the cell cycle in the G0/G1 phases . Previous observations regarding the in vitro life cycle of T . cruzi in cultured cells demonstrated a low cell-division rate among cells infected with the parasite [76–78] . In this regard , Ca2+ may be responsible for the cell cycle changes , as they act as second messengers in the control of the cell cycle . Thus , Ca2+/calmodulin activate the complex CDK4/cyclin D1 , which regulates the protein of Retinoblastoma ( pRb1 ) , the main inhibitor of the DNA synthesis [79] . From our results , it is evident that the phosphorylation of the protein Rb takes place from the first few min of the EVs/cell interaction . Here , phosphorylated pRb increased rapidly , while in the cells without the treatment with EVs no such change was detected ( Fig 5B and 5C ) . However , at 60 min of treatment with EVs , these increases in phosphorylation returned to normal levels . This apparently arrested synthesis , preventing the cells from entering phase S of the cycle . Moreover , a series of “calcium sensors” present in the cell cytoplasm , such as the stromal interaction molecule 1 ( STIM1 ) , is involved in the progression of mitosis . Cells lacking this protein may arrest the cells in phases G0/G1 , as occurred in our experiments . This implies that this protein is required for the progression of the cells in the phase of DNA synthesis or phase S [80] . The arrest of cells in the G0/G1 phases exerted by EVs of T . cruzi Pan4 was possibly caused by increased expression of cyclin-dependent kinase inhibitor p21 and the subsequent decrease of phosphorylated protein ( pRb ) [81] . The higher intracellular calcium levels were also involved in cell-cycle events . In fact , the indirect role of high levels of calcium in cell arrest in these phases has been examined by Wu et al . ( 2006 ) [82] , who employed capsaicin and blocked the cell cycle in the previous phase of DNA synthesis . This effect was reversed with BAPTA , an intracellular Ca2+ chelator . Together with the rises of intracellular calcium levels , and because of these high levels , these researchers have recently questioned the role of actin networks nucleated by the complex Arp2/3 in the signalling events necessary for the progression of the cell cycle in non-transformed cells [82–83] and demonstrated that Arp2/3 is not able to act as a sensor for the start of the phase S in the cell cycle per se , such as the actin filaments . Previous studies have shown that the use of cytochalasin B at very low doses detains the cell cycle in phases G0/G1 as in our experiments with EVs while the inhibitors that act in the polymerization of actin stopped the cell cycle before the cytokinesis [84–86] . In conclusion , it has been shown that the incubation of cells with EVs of TcT of T . cruzi Pan4 strain induce a number of changes in the host cells that include 1 ) a change in cell permeability , and 2 ) higher intracellular levels of Ca2+ that can alter the dynamics of the actin cytoskeleton and arrest the cell cycle at G0/G1 prior to the DNA synthesis necessary to complete mitosis . In the end , these changes induced by the EVs aid their invasion of host cells , augment the percentage of cell parasitization , and possibly cause some characteristic manifestations of Chagas disease . | Extracellular vesicles ( EVs ) are a diverse group of nanoparticles involved in intercellular communication under physiological and pathological conditions . Trypanosoma cruzi , the protozoan that causes Chagas disease , releases EVs that facilitate parasite invasion of the host cell , immunomodulate the host response , and help the parasite to evade this response . However , little is known about how the host cell is altered . In this work , we confirm that EVs of tissue-culture cell-derived trypomastigotes of the Pan4 strain increase cell parasitism . We also demonstrate that EVs affect cell permeability in Vero cells and cardiomyocytes and raise intracellular Ca2+ levels , altering the actin filaments and arresting the cell cycle at the G0/G1 phases . This work seeks to elucidate the way in which EVs influence certain aspects of the cell physiology that favour the establishment of this parasite inside the host cell . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"vero",
"cells",
"medicine",
"and",
"health",
"sciences",
"viral",
"transmission",
"and",
"infection",
"parasitic",
"cell",
"cycles",
"biological",
"cultures",
"microbiology",
"parasitic",
"diseases",
"protozoan",
"life",
"cycles",
"parasitic",
"protozoans",
"parasitolo... | 2019 | Extracellular vesicles of Trypanosoma cruzi tissue-culture cell-derived trypomastigotes: Induction of physiological changes in non-parasitized culture cells |
Ubiquitylation targets proteins for proteasome-mediated degradation and plays important roles in many biological processes including apoptosis . However , non-proteolytic functions of ubiquitylation are also known . In Drosophila , the inhibitor of apoptosis protein 1 ( DIAP1 ) is known to ubiquitylate the initiator caspase DRONC in vitro . Because DRONC protein accumulates in diap1 mutant cells that are kept alive by caspase inhibition ( “undead” cells ) , it is thought that DIAP1-mediated ubiquitylation causes proteasomal degradation of DRONC , protecting cells from apoptosis . However , contrary to this model , we show here that DIAP1-mediated ubiquitylation does not trigger proteasomal degradation of full-length DRONC , but serves a non-proteolytic function . Our data suggest that DIAP1-mediated ubiquitylation blocks processing and activation of DRONC . Interestingly , while full-length DRONC is not subject to DIAP1-induced degradation , once it is processed and activated it has reduced protein stability . Finally , we show that DRONC protein accumulates in “undead” cells due to increased transcription of dronc in these cells . These data refine current models of caspase regulation by IAPs .
Ubiquitylation describes the covalent attachment of ubiquitin , a 76 amino acid polypeptide , to proteins which occurs as a multi-step process ( reviewed in [1] , [2] ) . E1-activating enzymes activate ubiquitin and transfer it to E2-conjugating enzymes . E3-ubiquitin ligases mediate the conjugation of ubiquitin from the E2 to the target protein . Repeated ubiquitylation cycles lead to the formation of polyubiquitin chains attached on target proteins . Polyubiquitylated proteins are delivered to the 26S proteasome for degradation . However , non-proteolytic roles of ubiquitylation have also been described ( reviewed in [3] , [4] ) . From E1 to E3 , there is increasing complexity . For example , the Drosophila genome encodes only one E1 enzyme , termed UBA1 , which is required for all ubiquitin-dependent reactions in the cell [5] . In contrast , there are hundreds of E3-ubiquitin ligases which are needed to confer substrate specificity . Programmed cell death or apoptosis is an essential physiological process for normal development and maintenance of tissue homeostasis in both vertebrates and invertebrates ( reviewed in [6] ) . A highly specialized class of proteases , termed caspases , are central components of the apoptotic pathway ( reviewed in [7] ) . The full-length form ( zymogen ) of caspases is catalytically inactive and consists of a prodomain , a large and a small subunit . Activation of caspases occurs through dimerization and proteolytic cleavage , separating the large and small subunits . Based on the length of the prodomain , caspases are divided into initiator ( also known as apical or upstream ) and effector ( also known as executioner or downstream ) caspases [7] . The long prodomains of initiator caspases harbor regulatory motifs such as the caspase activation and recruitment domain ( CARD ) in CASPASE-9 . Through homotypic CARD/CARD interactions with the adapter protein APAF-1 , CASPASE-9 is recruited into the apoptosome , a large multi-subunit complex , where it dimerizes and auto-processes leading to its activation [8] , [9] . Activated CASPASE-9 cleaves and activates effector caspases ( CASPASE-3 , -6 , and –7 ) , which are characterized by short prodomains . Effector caspases execute the cell death process by cleaving a large number of cellular proteins [10] . In Drosophila , the initiator caspase DRONC and the effector caspases DrICE and DCP-1 are essential for apoptosis [11]–[18] . Like human CASPASE-9 , DRONC carries a CARD motif in its prodomain [19] . Consistently , DRONC interacts with ARK , the APAF-1 ortholog in Drosophila ( also known as DARK , HAC-1 or D-APAF-1 ) [20]–[22] for recruitment into an apoptosome-like complex which is required for DRONC activation [20] , [23]–[31] . After recruitment into the ARK apoptosome , DRONC cleaves and activates the effector caspases DrICE and DCP-1 [25] , [31]–[34] . Caspases are subject to negative regulation by inhibitor of apoptosis proteins ( IAPs ) ( reviewed in [35] , [36] ) . For example , DRONC is negatively regulated by Drosophila IAP1 ( DIAP1 ) [37] , [38] . diap1 mutations cause a dramatic cell death phenotype , in which nearly every mutant cell is apoptotic , suggesting an essential genetic role of diap1 for cellular survival [39]–[41] . DIAP1 is characterized by two tandem repeats known as the Baculovirus IAP Repeat ( BIR ) , and one C-terminally located RING domain [42] . The BIR domains are required for binding to caspases [37] , [38] , [43] . The RING domain provides DIAP1 with E3-ubiquitin ligase activity , required for ubiquitylation of target proteins [35] , [36] . Importantly , the BIR domains can bind to caspases independently of the RING domain [37] , [43] . Usually , IAPs bind to and inhibit activated , i . e . processed caspases , including CASPASE-3 , CASPASE-7 and CASPASE-9 as well as the Drosophila caspases DrICE and DCP-1 ( reviewed in [35] , [36] ) . However , a notable exception to this rule is DRONC . DIAP1 binds to the prodomain of full-length DRONC [37] , [38] , [43] . This unusual behavior suggests an important mechanism for the control of DRONC activation . Indeed , it has been shown that the RING domain of DIAP1 ubiquitylates full-length DRONC in vitro [38] , [44] . It has also been proposed that DIAP1 ubiquitylates auto-processed DRONC [33] . These ubiquitylation events are critical for the control of apoptosis , as homozygous diap1 mutants which lack a functional RING domain ( diap1ΔRING ) are highly apoptotic [41] . Because the BIR domains are intact in diap1ΔRING mutants , binding of DIAP1 to DRONC is not sufficient for inhibition of DRONC under physiological conditions , and ubiquitylation is the critical event for DRONC inhibition . Although the importance of DIAP1-mediated ubiquitylation of DRONC is well established , it is still unclear how this ubiquitylation event leads to inactivation of DRONC and of caspases in general . Because DRONC protein accumulates in diap1 mutant cells that are kept alive by expression of the effector caspase inhibitor P35 , generating so-called ‘undead’ cells , it has been proposed that DIAP1-mediated ubiquitylation triggers proteasomal degradation of full-length DRONC in living cells , thus protecting them from apoptosis [33] , [38] , [45] , [46] . However , degradation of full-length DRONC in living cells has never been observed and non-degradative models have also been proposed [44] . Furthermore , ubiquitylation of mammalian CASPASE-3 and CASPASE-7 has been demonstrated in vitro [47]–[49] . However , evidence for proteasome-dependent degradation of these caspases in vivo , i . e . in the context of a living animal , is lacking . In fact , a non-degradative mechanism has been demonstrated for the effector caspase DrICE in Drosophila [50] . Here , we further characterize the role of ubiquitylation for the control of DRONC activation . Consistent with a previous report [44] , we find that ubiquitylation of DRONC by DIAP1 is critical for inhibition of DRONC's pro-apoptotic activity . Using loss and gain of diap1 function , we provide genetic evidence that DIAP1-mediated ubiquitylation of full-length DRONC regulates this initiator caspase through a non-degradative mechanism . We find that the conjugation of ubiquitin suppresses DRONC processing and activation . Interestingly , once DRONC is processed and activated , it has reduced protein stability . Finally , we show that dronc transcripts accumulate in P35-expressing ‘undead’ cells , accounting for increased DRONC protein levels in these cells . These data refine the current model of caspase regulation by IAPs .
It has previously been shown that complete loss of ubiquitylation due to mutations of the E1 enzyme Uba1 causes apoptosis in eye imaginal discs as detected by an antibody that recognizes cleaved , i . e . activated , CASPASE-3 ( CAS3* ) [5] , [51] , [52] ( see also Figure 1A ) . Because ubiquitylation of DRONC does not occur in Uba1 mutants , we hypothesized that inappropriate activation of DRONC accounts for the apoptotic phenotype of Uba1 mutants . To test this possibility , we targeted dronc by RNA interference ( RNAi ) in Uba1 mutant cells in eye imaginal discs using the MARCM system and labeled for apoptosis using CAS3* antibody . In this system , Uba1 mutant cells expressing dronc RNAi are positively marked by GFP . Consistent with our hypothesis , knock-down of dronc strongly reduces apoptosis in Uba1 mutant clones ( Figure 1B ) . Furthermore , we tested clones doubly mutant for Uba1 and ark , the Drosophila ortholog of APAF-1 that is required for DRONC activation ( see Introduction ) . Apoptosis induced in Uba1 mutant clones is strongly suppressed if ark function is removed ( Figure S1 ) . These observations suggest that the apoptotic phenotype in Uba1 clones is caused by inappropriate activation of DRONC , presumably due to lack of ubiquitylation . However , the protein levels of DIAP1 are increased in Uba1 mutant clones [5] , [52] . There are two possibilities to explain the apoptotic phenotype in Uba1 mutants despite increased DIAP1 levels . First , the DIAP1 levels may not be sufficiently increased to inhibit DRONC . Alternatively , binding of DIAP1 to DRONC alone may not be sufficient for inhibition of DRONC; instead , ubiquitylation by DIAP1 is required to block DRONC activation , as previously suggested [44] . To distinguish between these two possibilities , we strongly overexpressed diap1 in Uba1 mutant clones in eye discs using the MARCM system and imaged for apoptosis by CAS3* labeling . Surprisingly , despite massive expression of diap1 ( >20 fold over wild-type levels; Figure 1C′″ ) , apoptosis still proceeds in Uba1 mutant clones ( Figure 1C′ ) , even though expression of the same transgene can block strong apoptotic phenotypes in several apoptotic paradigms ( Figure S2 ) . Apparently , overexpression of DIAP1 is not sufficient to inhibit DRONC and to protect Uba1 mutant cells from apoptosis . Because DIAP1 is the key regulator of DRONC and because DRONC is required for the apoptotic phenotype of Uba1 mutant cells , as evidenced by knock-down of dronc ( Figure 1B ) , our data provide genetic evidence that binding of DIAP1 is not sufficient for DRONC inhibition in Uba1 mutant cells . Consistent with this view , it has previously been shown that DIAP1 does ubiquitylate full-length DRONC in vitro [33] , [38] , [44] . We tested whether DIAP1 can also ubiquitylate DRONC in vivo . Because the available DRONC antibodies failed to immunoprecipitate endogenous DRONC , we transfected DRONC-V5 along with DIAP1+ or DIAP1ΔRING mutants ( CΔ6 , lacking the last six C-terminal residues , and F437A changing a critical Phe residue in the RING domain to Ala [53] ) and His-tagged Ubiquitin into Drosophila S2 cells . Ubiquitylated proteins were affinity purified under denaturing conditions using Ni columns . The eluates were subsequently examined by immunoblotting with anti-V5 antibodies to detect ubiquitylated forms of DRONC . Under these conditions , DIAP1+ readily ubiquitylates full-length DRONC in S2 cells ( Figure 2 ) , whereas the RING mutants DIAP1CΔ6 and DIAP1F437A were significantly impaired in their ability to ubiquitylate DRONC ( Figure 2 ) . These results indicate that DIAP1 ubiquitylates full-length DRONC in a RING-dependent manner in cultured cells . Because DIAP1 readily ubiquitylates DRONC , it has been postulated that DIAP1-mediated ubiquitylation leads to proteasomal degradation of DRONC [33] , [38] , [45] . However , this has never been rigorously tested in vivo . Therefore , we examined , whether overexpression of diap1 in wild-type animals can influence DRONC protein levels in vivo . To this end , we generated clones overexpressing diap1 ( marked by absence of GFP ) in eye discs , and analyzed the protein abundance of DRONC . Interestingly , despite high expression of diap1 ( Figure 3A′″ ) , the levels of DRONC remained unchanged and were not influenced by DIAP1 ( Figure 3A′ ) . The anti-DRONC antibody used in this assay is specific for DRONC ( Figure S3 ) . Importantly , the diap1 transgene used produces a functional DIAP1 protein that is able to inhibit apoptosis in several paradigms ( Figure S2 ) . Therefore , these data suggest that overexpressed DIAP1 does not target DRONC for degradation in living cells . Because of the surprising observation that overexpressed DIAP1 does not cause degradation of DRONC , we tested whether removal of DIAP1 changes DRONC protein levels . Expression of the IAP antagonist reaper ( rpr ) induces DIAP1 degradation and apoptosis [54]–[58] . Therefore , we examined whether RPR-induced degradation of DIAP1 changes DRONC protein levels . If DIAP1 targets DRONC for degradation , we would expect that DRONC protein levels would accumulate in response to rpr expression . Expression of rpr in eye imaginal discs posterior to the morphogenetic furrow ( MF ) using the GMR promoter ( GMR-rpr ) is well suited for this analysis . The MF is a dynamic structure that initiates at the posterior edge of the eye disc and moves towards the anterior during 3rd instar larval stage [59] , [60] ( Figure 4A , arrow ) . Expression of rpr by GMR is induced in all cells posterior to the MF [61] ( red in Figure 4A ) . Therefore , GMR-rpr eye discs provide a continuum of all developmental stages in which cells close to the MF have only recently induced rpr expression , while cells towards the posterior edge of the disc have been exposed to rpr progressively longer . Therefore , if DRONC accumulates during any of these stages , we should be able to detect it . In wild-type eye discs , DRONC protein is homogenously distributed throughout the disc . Only in the MF , higher levels of DRONC are detectable ( arrowhead in Figure 4B″ ) . This high expression of DRONC in the MF serves as an orientation mark . DIAP1 protein levels are low anterior to the MF , but increase in the MF ( arrowhead ) and posterior to it in wild-type discs ( Figure 4B′ ) . In GMR-rpr eye discs , overall DIAP1 levels are reduced in the rpr-expressing domain posterior to the MF ( Figure 4C′ ) , but particularly strongly reduced in the CAS3*-positive area ( Figure 4C′ , D′ , arrow ) consistent with previous reports [54]–[58] . However , accumulation of DRONC is not observed ( Figure 4C″ , D″ ) . In contrast , it appears that DRONC levels are also reduced . They are still high in the MF ( Figure 4C″ , arrowhead ) , but drop immediately thereafter . We also related DRONC levels to caspase activation . In the MF , where CAS3* activity is not detectable , DRONC is still high ( Figure 4D′ , D″; arrowhead ) , but in the CAS3*-positive area , DRONC levels are reduced ( Figure 4D′ , D″; arrow ) . These data indicate that loss of DIAP1 does not cause accumulation of DRONC protein implying that DIAP1 does not induce degradation of DRONC . In contrast , it appears that DIAP1 stabilizes DRONC at least under these conditions ( see Discussion ) . Finally , we analyzed DRONC protein levels in diap1ΔRING mutants which cannot ubiquitylate DRONC [44] . It has previously been shown that clones of the RING mutant diap122-8s accumulate DRONC protein [45] , [46] implying that ubiquitylation by the RING domain of DIAP1 causes degradation of DRONC . We repeated these experiments and indeed confirmed that DRONC levels are increased in diap122-8s mutant clones ( Figure S4 ) . Thus , these results appear inconsistent with the data presented in Figure 3 and Figure 4 in which manipulating DIAP1 levels did not provide evidence for DIAP1-mediated degradataion of DRONC . However , one caveat with the diap122-8s experiment was the use of the caspase inhibitor P35 which kept diap122-8s mutant cells in an ‘undead’ condition [45] . It has been pointed out that the ‘undead’ state may change the properties of the affected cells ( reviewed by [62] ) and in fact abnormal induction of transcription in ‘undead’ cells has been reported [45] , [63]–[66] . Thus , to explain the conflicting results between the diap122-8s data [45] and our data shown here , we hypothesized that p35-expressing ‘undead’ diap122-8s clones induce dronc transcription , leading to accumulation of DRONC protein . To test this hypothesis , we used a transcriptional lacZ reporter containing 1 . 33 kb of regulatory genomic sequences upstream of the transcriptional start site of the dronc gene fused to lacZ ( dronc1 . 33-lacZ ) [67] , [68] . Compared to controls ( Figure 5A , 5A′ ) and consistent with the hypothesis , dronc1 . 33-lacZ reporter activity is increased in p35-expressing ‘undead’ diap122-8s cells in wing imaginal discs and matches the increased DRONC protein pattern ( Figure 5B′-5B′″ ) . We also produced ‘undead’ cells in eye imaginal discs by co-expression of the IAP-antagonist hid and the caspase inhibitor p35 in the dorsal half of the eye disc using a dorsal eye- ( DE- ) GAL4 driver ( Figure 5C ) . Similar to wing discs , dronc reporter activity is increased in ‘undead’ cells in the dorsal half of the eye ( Figure 5D ) . Expression of p35 alone does not trigger transcription of dronc ( Figure 5E ) suggesting it is not the mere presence of P35 which causes dronc transcription , but the ‘undead’ nature of the affected cells . These observations may explain why DRONC protein accumulates in ‘undead’ diap122-8s mutant cells , but they still do not rule out the possibility that DRONC protein accumulates in diap122-8s mutants due to lack of ubiquitylation and thus degradation . To clarify this issue we examined dronc1 . 33-lacZ and DRONC levels in diap122-8s mutant clones without simultaneous p35 expression . Without the inhibition of apoptosis by P35 , diap122-8s clones die rapidly . Nevertheless , we were able to recover wing discs which contained small diap122-8s mutant clones . In these clones , neither dronc1 . 33-lacZ nor DRONC levels are detectably increased ( Figure 5F ) . Notably , these clones are located in the wing pouch in which we observed accumulation of dronc reporter activity and DRONC protein under ‘undead’ conditions ( Figure 5B″ ) . Thus , the ‘undead’ condition of p35-expressing diap122-8s mutant cells causes accumulation of DRONC protein due to induction of dronc transcription , explaining the observations of Ryoo et al . ( 2004 ) [45] . In the absence of p35 expression , transcription of dronc and accumulation of DRONC protein are not observed , providing additional evidence that ubiquitylation of DRONC by the RING domain of DIAP1 does not trigger degradation of DRONC . Our in vivo analysis implies that DIAP1-mediated ubiquitylation does not trigger proteasomal degradation of DRONC . To identify the role of ubiquitylation for regulation of DRONC activity , we analyzed the fate of DRONC protein in RING mutants of diap1 . Of note , these mutants retain their ability to bind to DRONC , because DRONC binding is not mediated by the RING domain , but by the BIR2 domain [37] , [38] , [43] . The RING mutant diap133-1s is especially suitable for this analysis because a premature stop codon results in deletion of the entire RING domain but leaves the BIR domains intact [44] ( Figure 6A ) , thus abrogating its E3 activity , but not caspase binding . Importantly , diap133-1s is characterized by a strong apoptotic phenotype , suggesting inappropriate caspase activation [41] , [45] . Indeed , we showed previously that diap1ΔRING mutant phenotypes are dependent upon DRONC , because dronc mutants suppress diap1ΔRING phenotypes [11] . Therefore , ubiquitylation of DRONC by DIAP1 is critical to maintain cell survival . We examined the cause of the diap133-1s apoptotic phenotype . First , as a control , we determined whether the diap133-1s gene produces a stable protein in vivo . We chose to analyze stage 6–9 embryos , because normal developmental cell death starts at stage 11 [69]; therefore , stage 6–9 diap133-1s mutant embryos allow analysis of DIAP1 in the absence of upstream apoptotic signals . In immunoblots of embryonic extracts obtained from a cross of heterozygous diap133-1s males and females , the DIAP133-1s protein is easily distinguished from wild-type DIAP1+ protein due to its faster electrophoretic mobility ( Figure 6A , top panel ) . The presence of the DIAP133-1s protein suggests that the apoptotic phenotype in diap133-1s mutant embryos is not caused by instability of the mutant protein . Interestingly , the protein levels of DIAP1+ and RING-deleted DIAP133-1s are similar ( Figure 6A , top panel ) suggesting that loss of the RING domain does not influence the protein stability of DIAP1 in the absence of pro-apoptotic signals . Next , we analyzed DRONC protein in extracts from diap133-1s mutant embryos . Consistent with the data in Figure 4 and Figure 5 , we do not detect a significant increase in the protein levels of DRONC in these embryos ( Figure 6A , middle panel ) . However , a significant amount of DRONC is present in a processed form in extracts of stage 6–9 diap133-1s mutant embryos which is absent in control extracts from wild-type embryos ( Figure 6A , middle panel ) . Therefore , DRONC processing , and thus activation , occurs in RING-depleted diap133-1s mutant embryos despite the fact that the BIR domains of DIAP1 are unaffected . The processed form of DRONC in this mutant of MW ∼36 kDa corresponds to the major apoptotic form of DRONC composed of the large subunit and the prodomain of DRONC [70] . This finding , and the one presented in Figure 1 , confirms that binding of DIAP1 to DRONC is not sufficient to prevent processing and activation of DRONC . Instead , the RING domain is required to control DRONC processing . Because the RING domain contains an E3-ubiquitin ligase activity [45] , [55]–[58] and because ubiquitylation of full-length DRONC does not trigger proteasomal degradation ( Figure 3 , Figure 4 , and Figure 5 ) , we conclude that ubiquitylation of DRONC by the RING domain of DIAP1 is necessary to inhibit DRONC processing and thus activation . To further characterize the role of ubiquitylation in the regulation of DRONC processing , we performed an immunoblot analysis with extracts from wild-type and Uba1 mosaic imaginal discs , which , under our experimental conditions , are about half mutant for Uba1 and half wild-type . Immunoblot analysis demonstrated that a significant amount of DRONC is processed in Uba1 mosaic discs ( Figure 6B ) . Thus , these data further support the notion that ubiquitylation of full-length DRONC is necessary for inhibition of DRONC processing .
Based on biochemical studies in vitro and overexpression studies in cultured cells , many of cancerous origin , it was initially proposed that binding of IAPs to caspases through their BIR domains is sufficient to inhibit caspases [71]–[80] . However , when ubiquitylation of caspases by IAPs was described [38] , [44] , [47] , [48] , it was unclear what role ubiquitylation would play for control of caspase activity , especially since for none of them , ubiquitin-mediated degradation has been observed ( see below ) . Because the RING domain is also required for auto-ubiquitylation of DIAP1 [54]–[58] , mutations of the RING domain would be expected to increase DIAP1 protein levels and protect cells from apoptosis . However , the opposite phenotype , massive apoptosis , was observed [41] . Nevertheless , despite the biochemical studies showing that the BIR domains of DIAP1 are sufficient for interaction with DRONC [37] , [38] , [43] , one could argue that DIAP1ΔRING mutants have lost the ability to interact with DRONC in vivo . While we cannot exclude this possibility due to the inability of our antibodies to immunoprecipitate endogenous proteins , another experiment supports the notion that ubiquitylation is necessary for DRONC inhibition: when wild-type diap1 was strongly overexpressed in an ubiquitylation-deficient Uba1 mutant background , DRONC-dependent apoptosis was not inhibited ( Figure 1C ) , suggesting that binding of DIAP1 is not sufficient for inhibition of DRONC . Instead , ubiquitylation is critical for inhibition of DRONC activity . The current model holds that DIAP1-mediated ubiquitylation leads to proteasomal degradation of full-length DRONC in living cells [33] , [38] , [45] . However , our data do not support this model in vivo . This model is based on biochemical ubiquitylation studies without in vivo validation and does not take into account that ubiquitylation often serves non-proteolytic functions [1] , [3] , [4] . Both overexpression and loss of diap1 does not cause a detectable alteration of the protein levels of DRONC ( Figure 3 , Figure 4 , Figure 5 ) , arguing against a degradation model . The only example where DRONC accumulation has been observed is in P35-expressing ‘undead’ diap1ΔRING mutant cells [45] , [46] , and we showed here that the ‘undead’ nature of these cells causes transcriptional induction of dronc ( Figure 5 ) . Together , these observations argue against a degradation model of full-length DRONC and favor a non-traditional ( non-proteolytic ) role of ubiquitylation regarding control of DRONC activity . Similarly , DIAP1-mediated ubiquitylation of the effector caspase DrICE inactivates this effector caspase through a non-degradative mechanism [50] . Interestingly , in GMR-rpr eye imaginal discs , DRONC protein levels appear to be reduced in apoptotic cells compared to living cells ( Figure 4C″ , 4D″ ) . Due to the apoptotic activity of REAPER , DRONC is present in its processed and activated form . Reduced protein stability of DRONC has also been reported after incorporation into the ARK apoptosome where it is processed and activated [46] . Combined , these observations suggest that while DIAP1-mediated ubiquitylation of full-length DRONC does not trigger its degradation , processed and activated DRONC has reduced protein stability and may indeed be degraded . It is unclear whether degradation of activated DRONC is mediated by DIAP1 , as proposed previously [33] . In GMR-rpr eye imaginal discs , reduced DRONC levels correlate with a reduction of DIAP1 protein ( Figure 4C′ , 4D′ ) . This correlation indicates that DIAP1 may actually stabilize DRONC protein , at least in part . Alternatively , because DRONC and DIAP1 form a complex [37] , REAPER-induced degradation of DIAP1 may result in co-degradation of complexed DRONC in the process . Further studies are needed to determine the cause of decreased DRONC stability in apoptotic cells . In addition to Drosophila DRONC , mammalian CASPASE-3 and CASPASE-7 have been reported to be ubiquitylated in vitro [47] , [48] . However , proteasome-mediated degradation of these caspases in vivo has not been reported . Although a decrease of CASPASE-3 levels has been noted upon overexpression of XIAP , this was shown for an artificial CASPASE-3 mutant , in which the order of the subunits was reversed and the Cys residue in the active site changed to Ser [47] . This catalytically inactive CASPASE-3 mutant is not proteolytically processed [47] . Therefore , physiological in vivo evidence for IAP-mediated degradation of mammalian caspases is lacking . Moreover , the phenotype of a RING-deleted XIAP mutant mouse is consistent with our data [49] . The XIAPΔRING mutant , which was generated by a knock-in approach in the endogenous XIAP gene , is characterized by increased caspase activity [49] . Intriguingly , the protein levels of CASPASE-3 , CASPASE-7 and CASPASE-9 did not significantly change in the XIAPΔRING mutant despite the fact that ubiquitylation of CASPASE-3 was reduced . However , processing of these caspases was easily detectable in XIAPΔRING mutants [49] . These data are very similar to the ones presented here for diap133-1s ( Figure 6 ) , and together strongly suggest that non-proteolytic ubiquitylation controls caspase processing and activity in both vertebrates and invertebrates . Non-proteolytic roles of ubiquitylation have been described in recent years and are involved in many processes including signal transduction , endocytosis , DNA repair , and histone activity ( reviewed in [1] , [3] , [4] ) . Two types of ubiquitylation lead to non-proteolytic functions . Monoubiquitylation is involved in endocytosis , DNA repair and histone activity . In fact , mammalian CASPASE-3 and CASPASE-7 have been found to be monoubiquitylated in vitro [48] . However , it is unclear whether DRONC is monoubiquitylated by DIAP1 . The presence of high molecular-weight ubiquitin conjugates in vitro ( Figure 2 ) suggests that DRONC may be polyubiquitylated , at least under the experimental conditions [38] , [44] . Polyubiquitylation serves both proteolytic and non-proteolytic functions depending on the Lysine ( K ) residue used for polyubiquitin chain formation . In general , if polyubiquitylation occurs via K48 , the target protein is subject to proteasome-mediated degradation . If it occurs on a different Lys residue , such as K63 , non-proteolytic functions are induced [1] , [3] , [4] . The best studied examples of both K48- and K63-linked polyubiquitylation are in the NF-κB pathway ( reviewed in [3] , [81] ) . While K48-polyubiquitylation leads to proteasomal degradation , K63-linked polyubiquitin chains act as scaffolds to assemble protein complexes required for NF-κB activation [3] , [81] . It is unknown what type of polyubiquitin chain is attached to DRONC , but it is possible that it is not K48-linked . Interestingly , in this context it has been shown that auto-ubiquitylation of DIAP1 preferentially involves K63-linked polyubiquitin chains [82] . By analogy , DIAP1 may ubiquitylate DRONC through formation of K63-linked polyubiquitin chains . This will be an interesting avenue to explore in future experiments . Conjugated monoubiquitin and polyubiquitin chains can serve as docking sites for factors containing ubiquitin-binding domains ( UBD ) [2] , [4] , [83] . The UBD-containing factors interpret the ubiquitylation status of the target protein , and trigger the appropriate response . For example , K48-linked polyubiquitin chains are recognized by Rad23 and Drk2 which deliver them to the proteasome [2] . Other forms of ubiquitin conjugates are recognized by different UBD-containing factors which control the activity of the ubiquitylated protein . Therefore , it is possible that an as yet unknown UBD-containing protein binds to ubiquitylated DRONC and controls its activity . For example , such an interaction could block the recruitment of ubiquitylated DRONC into the ARK apoptosome . Another possibility is that ubiquitylation could block dimerization of DRONC , which is required for activation of DRONC [34] . ‘Undead’ cells can be obtained by expression of the effector caspase inhibitor P35 [84] . In these cells , apoptosis has been induced , but cannot be executed due to effector caspase inhibition . Nevertheless , the initiator caspase DRONC is active in ‘undead’ cells and can promote non-apoptotic processes [51] . Recent work has suggested that ‘undead’ cells may alter their cellular behavior . For example , ‘undead’ cells change their size and shape , and have some migratory abilities to invade neighboring tissue [62] . They are also able to promote proliferation of neighboring cells causing hyperplastic overgrowth [15] , [45] , [63]–[66] ( reviewed by [85] , [86] ) . Altered transcription of the TGF-ß/BMP member decapentaplegic ( dpp ) , the Wnt-homolog wingless ( wg ) , and the p53 ortholog dp53 has also been reported in ‘undead’ cells [45] , [64]–[66] . Intriguingly , while dpp and wg are usually not expressed in the same cells [87] , ‘undead’ cells co-express them ectopically , clearly indicating an altered transcriptional program . As part of this altered transcriptional program , we showed that ‘undead’ cells also stimulate transcription of the initiator caspase dronc ( Figure 5 ) . Interestingly , p35 expression in normal cells does not induce dronc transcription suggesting that it is not the mere presence of P35 that induces dronc transcription , but instead the ‘undead’ condition of the affected cells . This transcriptional induction of dronc provides an explanation why DRONC protein levels are increased in ‘undead’ cells . It may also help to explain another observation regarding ‘undead’ cells . It has been demonstrated that although these cells are unable to die , they maintain the apoptotic machinery indefinitely [62] , [88] . Therefore , as part of this maintenance program , ‘undead’ cells stimulate dronc transcription . This is likely not restricted for dronc . Martin et al . ( 2009 ) [62] also showed that DrICE protein levels remain high in ‘undead’ cells which may also be caused by increased drICE transcription . It is also possible that the induction of dp53 by ‘undead’ cells [66] is part of this maintenance program , because we have shown that Dp53 induces expression of hid and rpr [89] and a positive feedback loop between dp53 , hid and dronc exists in ‘undead’ cells [66] . This may all occur at a transcriptional level . The mechanism by which ‘undead’ cells stimulate expression of dpp , wg , dp53 , dronc and potentially drICE are currently unknown and are avenues for future research .
Fly crosses were conducted using standard procedures at 25°C . The following mutants and transgenes were used: Uba1D6 [5]; arkG8 [26]; diap122-8s and diap133-1s [44]; vps25N55 [90]; droncI29 [11]; UAS-droncIR ( dronc inverted repeats ) [91]; GMR-rpr [92]; dronc1 . 33-lacZ [67] , [68] , ubx-FLP [93] , nub-GAL4 [94] , DE- ( dorsal eye- ) GAL4 [95] , and UAS-hid [96] . nub-FLP is nub-GAL4 UAS-FLP . UAS-p35 and UAS-FLP were obtained from Bloomington . Uba1D6 is a temperature sensitive allele which at 25°C is a hypomorphic allele , but at 30°C is a null allele [5] . In the experiments in Figure 1 , Figure 6B , and Figure S1 , Uba1D6 and Uba1D6 arkG8 mosaic larvae were incubated at 25°C; 12 hours before dissection the temperature was shifted to 30°C . This treatment allows recovery of Uba1D6 null mutant clones , which otherwise are cell lethal . Mutant clones were induced in eye-antennal imaginal discs using the FLP/FRT mitotic recombination system [97] using ey-FLP [98] . In this case , clones are marked by loss of GFP . Expression of diap1 and dronc RNAi in Uba1D6 clones ( Figure 1 ) was induced from UAS-diap1 or UAS-droncIR transgenes using the MARCM system [99] . Here , clones are positively marked by GFP . For induction of diap1-expressing clones in Figure 3 , the UAS-diap1 transgene was crossed to hs-FLP; tub<GFP<GAL4 ( < = FRT ) . Clones are marked by the absence of GFP . MARCM clones and diap1-overexpressing clones were induced in first instar larvae by heat-shock for one hour in a 37°C water bath . Expression of UAS-p35 in diap122-8s mosaic discs was accomplished by nub-GAL4 . Eye-antennal imaginal discs from third instar larvae were dissected using standard protocols and labeled with antibodies raised against the following antigens: DIAP1 ( a kind gift of Hermann Steller and Hyung Don Ryoo ) ; cleaved CASPASE-3 ( CAS3* ) ( Cell Signaling Technology ) and anti-ß-GAL ( Promega ) . The DRONC antibody used for immunofluorescence was raised against the C-terminus of DRONC in guinea pigs [44] . This antibody is specific for DRONC ( Figure S3 ) . Cy3- and Cy-5 fluorescently-conjugated secondary antibodies were obtained from Jackson ImmunoResearch . In each experiment , multiple clones in 10–20 eye and wing imaginal discs were analyzed , unless otherwise noted . Images were captured using an Olympus Optical FV500 confocal microscope . Schneider S2 cells were co-transfected with pMT-DRONC C>A V5 , pAcDIAP1 ( wt or CΔ6 , F437A , respectively , described in [50] ) and pAc His-HA-Ub at equal ratios . Expression of DRONC was induced overnight with 350 µM CuSO4 . Cells were lysed under denaturing conditions and ubiquitylated proteins were purified using Ni2+-NTA agarose beads ( QIAGEN ) . Immunoblot analysis was performed with α-V5 ( Serotec ) and α-DIAP1 antibodies [37] , [43] . For the immunoblots in Figure 6A , embryos were collected , decorionated and snap frozen in liquid nitrogen . Embryos were sonicated in Laemmli SDS loading buffer while being frozen . The equivalent of 30 lysed embryos was loaded per lane . Immunoblots were done using standard procedures . The anti-DRONC antibody used in Figure 6A is a peptide antibody raised against the large subunit of DRONC . The anti-DRONC antibody used in Figure 6B was raised against the C-terminus of DRONC in guinea pigs . | The Drosophila inhibitor of apoptosis 1 ( DIAP1 ) readily promotes ubiquitylation of the CASPASE-9–like initiator caspase DRONC in vitro and in vivo . Because DRONC protein accumulates in diap1 mutant cells that are kept alive by effector caspase inhibition—producing so-called “undead” cells—it has been proposed that DIAP1-mediated ubiquitylation would target full-length DRONC for proteasomal degradation , ensuring survival of normal cells . However , this has never been tested rigorously in vivo . By examining loss and gain of diap1 function , we show that DIAP1-mediated ubiquitylation does not trigger degradation of full-length DRONC . Our analysis demonstrates that DIAP1-mediated ubiquitylation controls DRONC processing and activation in a non-proteolytic manner . Interestingly , once DRONC is processed and activated , it has reduced protein stability . We also demonstrate that “undead” cells induce transcription of dronc , explaining increased protein levels of DRONC in these cells . This study re-defines the mechanism by which IAP-mediated ubiquitylation regulates caspase activity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [
"cell",
"death",
"developmental",
"biology",
"cell",
"biology",
"biology",
"molecular",
"cell",
"biology"
] | 2011 | Drosophila IAP1-Mediated Ubiquitylation Controls Activation of the Initiator Caspase DRONC Independent of Protein Degradation |
West Nile virus ( WNV ) is an arbovirus maintained in nature in a bird-mosquito enzootic cycle which can also infect other vertebrates including humans . WNV is now endemic in the United States ( U . S . ) , causing yearly outbreaks that have resulted in an estimated total of 4–5 million human infections . Over 41 , 700 cases of West Nile disease , including 18 , 810 neuroinvasive cases and 1 , 765 deaths , were reported to the CDC between 1999 and 2014 . In 2012 , the second largest West Nile outbreak in the U . S . was reported , which caused 5 , 674 cases and 286 deaths . WNV continues to evolve , and three major WNV lineage I genotypes ( NY99 , WN02 , and SW/WN03 ) have been described in the U . S . since introduction of the virus in 1999 . We report here the WNV sequences obtained from 19 human samples acquired during the 2012 U . S . outbreak and our examination of the evolutionary dynamics in WNV isolates sequenced from 1999–2012 . Maximum-likelihood and Bayesian methods were used to perform the phylogenetic analyses . Selection pressure analyses were performed with the HyPhy package using the Datamonkey web-server . Using different codon-based and branch-site selection models , we detected a number of codons subjected to positive pressure in WNV genes . Thirteen of the 19 completely sequenced isolates from 10 U . S . states were genetically similar , sharing up to 55 nucleotide mutations and 4 amino acid substitutions when compared with the prototype isolate WN-NY99 . Overall , these analyses showed that following a brief contraction in 2008–2009 , WNV genetic divergence in the U . S . continued to increase in 2012 , and that closely related variants were found across a broad geographic range of the U . S . , coincident with the second-largest WNV outbreak in U . S . history .
West Nile virus ( WNV ) emerged in the United States in 1999 and has become endemic , having caused annual outbreaks each subsequent year . WNV is a Flavivirus maintained in nature in an enzootic cycle between birds and mosquitoes . Other vertebrate hosts may be infected and develop disease , as occurs with humans and horses , which are considered dead-end hosts since they do not develop sufficient viremia to re-infect mosquitoes [1 , 2] . Transmission may also occur between humans via blood transfusion and transplantation of organs from infected individuals [3 , 4] . Since 2003 , donated blood has been routinely screened for WNV by nucleic acid testing ( NAT ) , and thousands of transmissions have been prevented [5] . Approximately 80% of humans infected with WNV develop no symptoms . Symptoms of WNV infections may vary from fever , rash and flu-like symptoms to severe neurological disease , which develops in less than 1% of cases and can result in death [6–8] . According to the U . S . Centers for Disease Control and Prevention ( CDC ) , WNV poses an ongoing public health threat , having infected millions of people and caused 1 , 765 deaths in the U . S . through the end of 2014 [9] . WNV is the most widely geographically distributed Flavivirus in the world , present on every continent except Antarctica . WNV infection had been observed in Africa , Asia , Australia/Oceania , and southern Europe prior to 1999 . In 1999 , the first cases of WNV in the Americas were observed in the U . S . in New York City , and the virus has since spread westward across the 48 contiguous states and Canada , and southward into Mexico , the Caribbean islands , Central America and South America , where it has caused human disease as far south as Argentina [10–12] . In the U . S . , WNV causes annual outbreaks of varying size and severity . Peaks of WNV activity have been observed in 2002–2003 , 2006 and 2012 . Reduced WNV activity was observed from 2008–2011 compared to 2002–2007 [9] . Following this period of relatively low activity , a large outbreak of WNV disease occurred in the 48 contiguous states in 2012 with 5 , 674 reported cases including 2 , 873 neuroinvasive cases and 286 deaths , the largest numbers reported to the ArboNET for any year since 2003 . [9] . WNV disease cases peaked in late August 2012 , with 5 , 199 ( 92% ) cases having onset of illness during July—September . The incidence of WNV neuroinvasive disease increased in 2012 to 0 . 92 per 100 , 000 . More than half of the neuroinvasive disease cases in 2012 were reported from four states: Texas ( n = 844 ) , California ( n = 297 ) , Illinois ( n = 187 ) , and Louisiana ( = 155 ) [9 , 14] . There are an estimated 30–70 non-neuroinvasive disease cases for every reported case of WNV neuroinvasive disease [6 , 8 , 13] . Therefore , an estimated 86 , 000–200 , 000 non-neuroinvasive disease cases might have occurred in 2012 but only 2 , 801 were diagnosed and reported . [14] . The reason for the increased incidence of WNV disease in 2012 is unknown and may involve multiple environmental and ecological factors as well as selection and dissemination of genetically best-fitted viruses . The spread of WNV in the Americas has offered a unique opportunity to observe evolution and genetic adaptation occurring in an arbovirus introduced to a new environment . The prototype strain from the 1999 New York outbreaks became known as the NY99 genotype , and is believed to share a common genetic origin with a 1998 Israeli isolate IS-98 [11 , 15] . In 2002 , a new WNV genotype , WN02 , appeared and was characterized by one amino acid substitution , E-V159A , and 13 conserved nucleotide mutations [16 , 17] . WN02 was found to be more efficiently transmitted by New World mosquitoes than NY99 , and eventually completely replaced NY99 [18] . This genetic shift coincided in time with large U . S . outbreaks in 2002–2003 and may have contributed to WNV’s spread across North America . Even with the genetic changes observed as WNV spread through North America , genetic variability of human isolates remained relatively low , increasing from 0 . 18% in 2002 to 0 . 37% in 2005 [19] . A second new genotype termed SW/WN03 , defined by two additional fixed amino acid substitutions , NS4A-A85T and NS5-K314R , was first observed in isolates collected in 2003 . WN02 and SW/WN03 genotypes displaced the ancestor NY99 genotype in the U . S . [20] . High WNV activity in the U . S . continued through 2006 and 2007 , and during this period , further genetic diversification of WNV strains was observed . A new well-defined viral cluster occurring within genotype SW/WN03 , termed MW/WN06 , was observed in strains collected from blood donors in the Midwestern and Northwestern U . S . in 2006 and 2007 [21] . The number of genetic mutations appearing in U . S . WNV strains continued to increase over this period , but the number of conserved mutations decreased slightly . Some nucleotide mutations which were previously believed to have been fixed in WNV isolates occurring after 2003 appeared to revert to the NY99 sequence , but other mutations associated with the WN02 genotype remained fixed [21] . The increased virulence of the WN02 genotype in mosquitoes is believed to have facilitated westward spread in 2002–2003 with a dramatic increase in infections , causing the largest WNV outbreak ever recognized worldwide and the largest viral encephalitis outbreak ever recognized in North America . This spread highlighted the need to monitor mutations occurring in the WNV genome and the genetic relationships of viral isolates causing disease in the U . S . over time [10–12 , 17 , 21] . Here we report results obtained from sequencing and phylogenetic analysis of 19 human WNV isolates from 13 U . S . states: Arizona ( AZ ) , California ( CA ) , Georgia ( GA ) , Illinois ( IL ) , Louisiana ( LA ) , Nebraska ( NE ) , New Mexico ( NM ) , North Dakota ( ND ) , Mississippi ( MS ) , Ohio ( OH ) , South Dakota ( SD ) , Texas ( TX ) , and Wyoming ( WY ) , from blood donations collected during the 2012 epidemic season . Thirteen of the 19 completely sequenced isolates from 10 U . S . states ( ND , SD , WY , TX , MS , GA , NM , OH , NE , IL ) were genetically similar , sharing up to 55 nucleotide mutations and 4 amino acid substitutions when compared with WN-NY99 ( GenBank accession number AF196835 ) . Phylogenetically , these 13 isolates clustered together with previously published 2012 isolates from TX [22 , 23] and some 2012 isolates from CO published in GenBank suggesting that this genetic variant was widely geographically distributed in 2012 . Isolates from AZ and CA were different from these genetic variants and phylogenetically clustered within local clades .
All human specimens used in this study were obtained from blood donors who signed the blood center’s Institutional Review Board ( IRB ) approved informed consent . These specimens were anonymized ( unlinked ) before shipment . Use of these unlinked specimens has been approved as exempt by the U . S . Food and Drug Administration ( FDA ) IRB ( Human Subjects Research—Exempt RIHSC Protocol #127B ) . The study included 19 isolates obtained after cultivation of residual blood specimens from blood donors who tested reactive for WNV RNA by FDA-approved commercial nucleic acid test assays used to screen blood donations . These 19 samples were representative of 13 states of the U . S . : AZ , CA , GA , IL , LA , NE , NM , ND , MS , OH , SD , TX , and WY ( Table 1 ) . A single passage in Vero cells ( ATCC # CCL-81 ) was performed for virus isolation from each specimen as previously described by Grinev et al . [19]; cell culture supernatants were harvested within 7 days and used for viral RNA extraction by the QIAamp Viral Mini RNA extraction kit ( Qiagen , Valencia , CA ) according to the manufacturer’s protocol . Reverse transcription reactions , PCR amplification and purification of amplicons were performed as described earlier [19] . Amplicons covering an entire WNV genome of each studied isolate were subjected to Sanger sequencing using the amplification primers and additional internal sequencing primers . Sequencing reactions were performed as described before [19] . Sequencing data were assembled and analyzed using the Vector NTI Advance 11 . 5 software package ( Invitrogen ) . Nucleotide ( nt ) and deduced amino acid ( aa ) sequences from studied isolates were aligned using the Align X program and compared to the genomic sequence of the parental WNV isolate WN-NY99 ( AF196835 ) . Nucleotide sequences reported in this paper were deposited into the GenBank database and accession numbers are shown in Table 1 ( KM012170—KM012188 ) . For Maximum-likelihood phylogeny we used MEGA 6 [24] . The Maximum-likelihood method employing the General Time Reversible ( GTR ) + Γ + I model was used to produce phylogenetic trees . This model was determined using the selection tool available in MEGA 6 . The parental strain WN-NY99 ( AF196835 ) was used to root the trees . The 19 newly sequenced WNV strains from this study ( Table 1 ) were aligned with 851 complete or near complete North American WNV sequences available in GenBank , as of September 2015 , using MEGA 6 . The dataset used in this study is composed of a total of 870 WNV ORF sequences from strains derived from the 1999–2012 epidemic seasons ( 1999 , n = 13; 2000 , n = 15; 2001 , n = 84; 2002 , n = 129; 2003 , n = 176; 2004 , n = 60; 2005 , n = 60; 2006 , n = 55; 2007 , n = 49; 2008 , n = 65; 2009 , n = 36; 2010 , n = 20; 2011 , n = 31; 2012 , n = 77 ) , shown in S1 Table . A selection analysis of ORFs of WNV strains isolated in 2012 ( n = 77 , S1 Table ) was performed using the Datamonkey web-server ( www . datamonkey . org ) . In addition to the Single-likelihood ancestor counting ( SLAC ) , Internal Fixed effects likelihood ( IFEL ) , Fixed effects likelihood ( FEL ) , Random Effects likelihood ( REL ) , Mixed Effects Model of Evolution ( MEME ) , Fast , Unconstrained Bayesian Approximation for inferring selection ( FUBAR ) methods , and Evolutionary fingerprint , we also employed the Conant-Stadler Property Informed Models of Evolution ( PRIME ) method . We have used the PRIME method to study site-specific aa properties ( e . g . chemical composition , charge , polarity ) which are being conserved or altered by the evolutionary process . Because of Datamonkey server restrictions , the REL method was only used to evaluate 74 sequences , which was the largest dataset that could be successfully analyzed ( KJ501432 , KJ501434 and KJ501437 were excluded randomly ) . A Bayesian skyline plot ( BSP ) was used to estimate the viral effective population size through time . Evolutionary rates for the WNV ORF sequences ( n = 870 ) were calculated using the Bayesian Markov-chain Monte Carlo ( MCMC ) approach employed by BEAST ver . 1 . 8 . 1 [25] and the BEAGLE library [26] . The dataset was analyzed using the TN93+Γ4 substitution model and the non-parametric Bayesian Skyline plot model , under relaxed uncorrelated lognormal ( UCLN ) molecular clocks as described elsewhere [21] . Four independent MCMC chains were run on a Tesla K20 computing processor until convergence to the stationary distribution was achieved ( ~500–600 million states with sampling frequency of 50 , 000 ) . Posterior distributions were examined in Tracer v1 . 6 [27] to ensure adequate mixing and convergence . All chains were combined in LogCombiner with a burn-in value set to 30% of generations . The maximum clade credibility tree ( MCC ) and BSP ( after resampling to 100 , 000 ) were generated . The MCC tree was visualized using FigTree v1 . 4 . 2 [27] .
Complete genomic sequences from 19 studied isolates from the 2012 epidemic were compared to the prototype strain WN-NY99 ( AF196835 ) . Most mutations ( ~89% ) were silent transitions ( U↔C , A↔G ) . The total number of nt mutations ranged from 54 to 83 . Shared nucleotide mutations identified in the studied WNV isolates are shown in Table 2 . All 19 WNV 2012 isolates from this study shared 7 nt mutations ( T1442C , C2466T , A4146G , C4803T , C6426T , C6996T and A10851G ) . Four mutations were shared by 18 of the 19 isolates: T7938C and T8811C ( excepting BSL140 ) ; T7015C ( excepting BSL178 ) ; and C9352T ( excepting BSL85 ) . In addition , seventeen isolates except BSL53 and BSL178 shared transition C6138T . Thirteen of the 19 completely sequenced isolates from 10 U . S . states ( ND , SD , WY , TX , MS , GA , NM , OH , NE , IL ) shared more than 50 nucleotide mutations when compared with prototype strain WN-NY99 ( Table 2 and S2 Table ) . Among 2012 WNV isolates from this study , the number of deduced aa substitutions ranged from 4 to 13 when compared to WN-NY99 , most of which are conservative changes . The transition T1442C is the non-silent mutation leading to the aa substitution E-V449A ( V159A , in the Envelope protein numeration ) . This substitution is common for all WNV isolates collected in the U . S . since 2003 , and therefore fixed in all strains of the WN02 and SW/WN03 genotypes . In addition to the aa substitution E-V449A , six WNV isolates shared NS2A-V1201I and 12 isolates shared the substitution NS2A-R1331K . Thirteen isolates reported here shared the substitution NS4B-I2513M ( Table 3 and S3 Table ) . Analysis of the nt variation in the ORFs of the North American WNV dataset ( n = 870 , S1 Table ) reveals increased evolutionary divergence from year to year ( Fig 1 ) . The estimated transition/transversion bias is 10 . 44 and the majority of the nt changes are transitions with relative rate 25 . 3 for U↔C and 7 . 2 for A↔G . Phylogenetic analysis was performed using the Maximum-likelihood method . In addition to the 19 WNV ORFs sequenced in this study , the North American WNV ORF sequences available from the GenBank database , as of September 2015 , were included in the dataset ( n = 870 , S1 Table ) . We have analyzed the phylogeny of these sequences and identified , as expected , the presence of the common clades representing the North American WNV genotypes NY99 , WN02 and SW/WN03 , previously described in the course of study of WNV evolution in North America [11 , 16–23 , 28–39] ( Fig 2 and S1 Fig ) . The 2012 WNV human isolates from this study are located within six nodes termed here “Node 1” to “Node 6” . Node-specific aa substitutions and geographical origin of isolates are shown in Fig 2 and S4 Table . We have observed that all WNV isolates reported here except BSL53 ( KM012172 ) , which is clustered in Node 6 within the SW/WN03 genotype , belong to the WN02 genotype . All studied isolates carried the common North American WNV aa substitution E-V159A which is fixed in the WN02 and SW/WN03 genotypes and present in all WNV strains collected in the U . S . since 2003 . Analysis of the entire ORF of WNV isolates circulating in the U . S . has shown that two isolates from AZ , BSL05 ( KM012170 ) and BSL80 ( KM012174 ) , clustered together with previously published isolates from AZ in Node 2 , and an isolate from CA , BSL85 ( KM012175 ) , clustered with other isolates from CA in Node 5 . We found that the WNV isolate BSL178 from LA ( KM012181 ) was associated with Node 3 , which mainly consisted of previously published WNV strains collected from TX in 2012 [22 , 23] and two 2012 isolates from CO . WNV strains presented in Node 3 shared up to nine aa substitutions . Surprisingly , many of the published 2012 isolates ( n = 27 ) were clustered within Node 4 together with 13 genetically related WNV isolates from this study collected from 10 states: ND , SD , WY , TX , MS , GA , NM , OH , NE and IL . All WNV strains from Node 4 , except KM012188 and KJ501532 , shared the NS2A-R188K aa substitution in addition to the common E-V159A . Other node-specific aa substitutions are shown in S4 Table . Using different codon-based and branch-site approaches , we detected a number of codons subjected to positive pressure in WNV strains collected in 2012 ( n = 74 for REL , n = 77 for all other methods ) . Analysis was done using the DataMonkey web-server ( www . datamonkey . org ) . We found that eight codons: 379; 1083; 1195; 1238; 1494; 2288; 2389; and 2842; were detected as positively selected by at least two methods . Site 2842 , corresponding to the NS5-K314R aa substitution , was the only site identified as positively selected by all methods ( Table 4 ) . Eleven node-specific aa substitutions identified in the phylogenetic analysis and detected as positively selected by at least one method are shown in Table 5 . We performed evolutionary fingerprint analysis , which models site to site variation in selection pressure across the ORF , for WNV isolates from 2012 ( n = 77 ) ( Fig 3 ) . The colored pixels on this plot show the density of the posterior sample of the distribution for a given rate and ellipses reflect a Gaussian-approximated variance in each individual rate estimate . Points above the diagonal line corresponded to positive selection ( ω>1 ) , and points below the diagonal line corresponded to negative selection ( ω<1 ) . Most of the points are concentrated below the diagonal line which represents the idealized neutral evolution scenario ( ω = 1 ) . The results suggest that WNV strains collected in 2012 were subjected to strong negative purifying selection . In addition , we conducted a supplementary selection pressure analysis using PRIME to detect whether selection for amino acids with differing chemical properties is occurring within the 2012 dataset ( n = 77 ) . Conant-Stadler PRIME analysis allows the non-synonymous substitution rate β to depend not only on the site in question ( like FEL and MEME ) , but also on which residues are being exchanged . Substitution rate analysis identified a single rate class , which suggests that across the 2012 WNV isolates , the rate of substitution between each residue was similar and no particular substitution was favored . PRIME analysis detected an overall substitution rate of 0 . 05 substitutions/codon site . One site , codon 2842 , was negatively selected for volume and positively selected for changes in chemical composition . This codon , corresponding to the NS5-K314R aa substitution , was identified as positively selected by all methods used for study of selection pressure . The time-scale analysis was performed using the North American WNV dataset ( n = 870 , S1 Table ) and the non-parametric Bayesian Skyline plot ( BSP ) model available in BEAST 1 . 8 . 1 . Previously we found that the BSP with the relaxed molecular clock ( UCLN ) was the best-fitted model [21] . The maximum clade credibility tree ( MCC ) was selected and the age for each node containing studied WNV isolates is shown on Fig 4A and S2 Fig . The time to most recent common ancestor ( tMRCA ) for the entire dataset was 14 . 78 years ago . Compared to the maximum-likelihood and Bayesian consensus phylogenetic trees , the MCC tree demonstrated a similar topology . Bayesian coalescent inference of genetic diversity and population dynamics was visualized using the Bayesian Skyline plot available in BEAST ( Fig 4B ) . The plot shows that a period of high genetic variability was observed until 2003 corresponding with the appearance of the new North American genotypes . From 2003–2009 , genetic diversity of the U . S . WNV population decreased slightly , with a maximum decrease occurring around 2008–2009 . A small increase in diversity occurred after 2009 , and the overall diversity of the WNV population then continued to increase through 2012 .
WNV now is the most widespread and common cause of viral encephalitis in the U . S . and worldwide [11 , 12] . After six years of relatively low WNV incidence in the U . S . , a large outbreak was observed in 2012 causing 5 , 674 total disease cases and 286 deaths , the largest number of deaths ever reported [9] . In this study we investigated the genetic variability of 19 WNV strains isolated from human samples collected in 2012 from 13 states of the U . S . ( Table 1 ) . Although humans are considered dead-end hosts for WNV , and therefore , not important for the WNV lifecycle , human isolates represent circulating viruses . Studying human WNV isolates is also important for public health and for the safety of the blood supply . Multiple factors were potentially involved in the magnitude of the 2012 outbreak . In addition to ecological and environmental factors that have been shown to increase viral transmission [40 , 41] , viral genetics and selection of new best-fitted variants may play a significant role in WNV outbreaks . Viral adaptation to domestic mosquitoes and birds has played a major role in the spread of WNV in the U . S . since its introduction in 1999 . WNV has continued to evolve , as illustrated through the displacement of the ancestor genotype WN99 by the new genotype WN02 in 2002 , followed by the appearance and co-circulation of genotype SW/WN03 in 2003 and subtype MW/WN06 in 2006 [11 , 16–23 , 28–39] . Analysis of nucleotide divergence of newly sequenced isolates from this study together with published North American WNV strains ( n = 870 ) demonstrates increasing evolutionary divergence from year to year ( Fig 1 ) . Previous phylogenetic analysis of WNV isolates shows that with limited exceptions , WNV isolates from circulating genotypes in the U . S . were poorly differentiated spatially and temporally [21] . It has been postulated that WNV genetic variations in the U . S . have occurred in some geographic areas which function as distinct niches of evolution . In these areas , the genetic variant accumulates genetic changes while adapting to the local ecological conditions , hosts and vectors , and may either stay in that area or be disseminated to other regions by migrating birds [42] . We observed that isolate BSL178 from LA was grouped in Node 3 together with WNV strains collected from TX in 2012 [22 , 23] and two 2012 isolates from CO . Thirteen other genetically similar human isolates from samples collected in 10 U . S . states for this study clustered with 2012 mosquito and bird isolates from TX [22 , 23] and CO in Node 4 ( Fig 2 ) . Nodes 3 and 4 are good examples of strong temporal phylogenetic structures constituted by well temporally differentiated isolates , and they were composed predominantly of isolates collected in the 2012 epidemic season . In contrast , isolates from AZ and CA clustered within local Nodes 2 and 5 . These nodes are good examples of strong spatial phylogenetic structures , which are supported by high bootstrapping values . The finding of similar isolates across a broad geographic area in 2012 suggests that closely related genetic variants of WNV represented in Node 4 spread over the Atlantic , Mississippi and Central bird flyways , but not the Pacific , and were identified coincident with the largest U . S . WNV outbreak since 2003 . In CA and AZ , both of which are located on the Pacific bird flyway , specimens clustered with local circulating clades suggesting predominantly local scale evolution in this area [21 , 39 , 43] . Previous studies of 2012 U . S . isolates have suggested that viral genetic composition was not a determinant of outbreak intensity at the local level . Duggal et al . noted that the genetic composition of viruses circulating in Texas in 2012 was similar between isolates from a county that experienced a large outbreak ( Dallas County ) and a county that didn’t ( Montgomery County ) [22] . Our data supports this conclusion on a broader geographic basis , because WNV isolates from the Nodes 3 and 4 circulated alongside isolates that were similar to those that circulated in 2008–2011 , and high numbers of disease cases occurred in areas where isolates from these Nodes were not detected at all , such as CA . Rather , increased replication in a favorable environment may have provided opportunity for genetically related co-existing strains to circulate and spread over migratory bird flyways , as has been reported on a local scale in TX and AZ [22 , 23 , 39] . The degree of genetic diversity and fitness of viral population is a balance between positive or negative selection and genetic drift as accumulation of random neutral mutations [44] . Previous studies have shown a low level of positive selection in WNV isolates from the U . S . [21 , 34 , 36] suggesting that most aa changes were the result of genetic drift . In our study of WNV isolates from 2012 , selection pressure analysis revealed only one site that was positively selected by all employed methods , codon 2842 ( NS5314 ) . This site has been previously identified as subject to positive selection in other studies of North American WNV sequences [21 , 22 , 37] . We found that this site is associated with nodes 2 , 4 and 6 ( Tables 4 , 5 and S4 ) and aa substitution NS5-K314R is involved in the emergence of the SW/WN03 genotype [20 , 21] . Site 1195 in NS2A was detected as positively selected by four methods . This site is associated with Node 3 aa substitution NS2A-T52I . Overall for the 2012 isolates , three aa substitutions in Node 3 and five in Node 4 were identified as positively selected by at least one method . Potentially aa substitutions could impact viral fitness and virulence , and the biological significance of those changes in viral proteins warrants further investigation . In general , our results are consistent with previous studies which have demonstrated that only limited positive selection is acting on the population of WNV circulating in the U . S . , and purifying selection is predominant [21 , 22] . In previous studies , results of time-scale analysis were only reported for select genes of WNV or reduced datasets [17 , 21 , 34 , 35 , 45] . In this study we performed comprehensive time-scale analysis using 870 full-length ORFs of WNV strains isolated in the U . S . in 1999–2012 ( Fig 4A ) . We found that the time to most recent common ancestor ( tMRCA ) for the whole dataset ( n = 870 ) was 14 . 78 years ( 95% HPD = 13 . 87–15 . 49 years ) , which is consistent with the value of 15 . 57 years ( 95% HPD = 14 . 23–16 . 98 years ) previously reported in the study of human isolates ( n = 62 ) when strain IS-98 ( AF481864 ) from 1998 was used to root the tree [21] . We calculated the mean nucleotide substitution rate ( MNSR ) , using the BSP model with the relaxed molecular clock , to be 6 . 81 x 10-4 substitutions/site/year ( s/s/y ) , which also correlates with published data [21 , 36] . Analysis of the BSP ( Fig 4B ) shows that genetic divergence had continued to slowly increase through 2012 following a brief period of contraction in 2008–2009 , which also agrees with data published by us and others [21 , 36 , 45] . Overall , our findings in this study suggest that the patterns of WNV genetic evolution in the U . S . following the 2012 outbreak remained consistent with previous trends . Additionally , our observation of the broad geographic distribution of genetically similar isolates suggests that these WNV variants may have spread via migratory birds , and were detected coincident with the largest WNV outbreak since 2003 . The emergence of this genetic variant may potentially mark the beginning of a new genetic shift and spread of a new WNV genotype after 10 years of steady drift . | West Nile virus ( WNV; family Flaviviridae , genus Flavivirus ) is a mosquito-borne virus maintained in a bird-mosquito enzootic cycle . WNV can occasionally infect other animals and humans , which are considered dead-end hosts because they produce too little virus in blood to re-infect mosquitoes . Most human infections ( ~80% ) do not cause symptoms , and when symptoms do occur , they may vary from mild flu-like illness to fatal neuroinvasive disease ( ~1% ) . WNV can be transmitted by transfusion of blood and blood components and by organ transplantation , posing a risk to the blood supply and public health . There is no specific therapy or vaccine for WNV in humans . WNV now is one of the most widely distributed flaviviruses in the world . Comparative studies of WNV genetic sequences have described two major groupings of WNV , lineages I and II , and up to five newer lineages , which correlate well with the geographical point of isolation . Since 1999 , WNV has spread from New York City throughout the U . S . and the Americas including Mexico , Canada , the Caribbean and South America . The emergence of WNV in the U . S . with annual outbreaks represents a unique opportunity to understand how a mosquito-borne virus adapts and evolves in a new environment . Viral adaptation to domestic mosquitoes and birds is considered to have played an important role in the spread of WNV in the U . S . Continuous surveillance of WNV genetic variation is needed to protect public health because the tests used to diagnose infection and screen blood , as well as vaccines and drug therapies currently in development , may not perform as well against newer genetic variants of WNV . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"biogeography",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"animal",
"genetics",
"ecology",
"and",
"environmental",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"population",
"genetics",
"micro... | 2016 | Genetic Variability of West Nile Virus in U.S. Blood Donors from the 2012 Epidemic Season |
The most polymorphic gene family in P . falciparum is the ∼60 var genes distributed across parasite chromosomes , both in the subtelomeres and in internal regions . They encode hypervariable surface proteins known as P . falciparum erythrocyte membrane protein 1 ( PfEMP1 ) that are critical for pathogenesis and immune evasion in Plasmodium falciparum . How var gene sequence diversity is generated is not currently completely understood . To address this , we constructed large clone trees and performed whole genome sequence analysis to study the generation of novel var gene sequences in asexually replicating parasites . While single nucleotide polymorphisms ( SNPs ) were scattered across the genome , structural variants ( deletions , duplications , translocations ) were focused in and around var genes , with considerable variation in frequency between strains . Analysis of more than 100 recombination events involving var exon 1 revealed that the average nucleotide sequence identity of two recombining exons was only 63% ( range: 52 . 7–72 . 4% ) yet the crossovers were error-free and occurred in such a way that the resulting sequence was in frame and domain architecture was preserved . Var exon 1 , which encodes the immunologically exposed part of the protein , recombined in up to 0 . 2% of infected erythrocytes in vitro per life cycle . The high rate of var exon 1 recombination indicates that millions of new antigenic structures could potentially be generated each day in a single infected individual . We propose a model whereby var gene sequence polymorphism is mainly generated during the asexual part of the life cycle .
Plasmodium falciparum is a unicellular parasite that causes malaria in humans . It infects over 300 million people per year and is estimated to have killed 600 , 000–1 . 2 million people in 2010 [1] . One of the most remarkable biological features of P . falciparum is an exceptionally polymorphic parasite antigen expressed on the surface of infected erythrocytes , known as P . falciparum erythrocyte membrane protein 1 ( PfEMP1 ) [2] . PfEMP1 is encoded by a family of hypervariable genes known as var , each representing a different antigenic form , and the parasite is able to vary its antigenic profile by switching expression between different var genes [3] . This allows the parasite to evade the human immune system and has major clinical consequences , as PfEMP1 mediates the cellular interactions and pathological properties of infected erythrocyte [4]–[6] . Each parasite genome contains approximately 60 var genes distributed in clusters across most of the 14 chromosomes . Based on conserved sequences upstream of the coding region , var genes are divided into three main groups . Group A var genes , which are confined to subtelomeric regions , have been shown by in vivo gene expression studies to be involved in the pathogenesis of severe malaria [2] . Group C var genes are found only in internal chromosomal regions while group B var genes occur both within chromosomes and at the subtelomere . Despite being the most polymorphic gene family in P . falciparum , var genes share broad structural similarities and some conserved motifs ( Fig . 1A ) . The first exon ( 4–10 kb ) begins with an N-terminal segment ( NTS ) and is followed by a succession of Duffy Binding Like ( DBL ) and cysteine rich interdomain regions ( CIDR ) domains . The second exon is semi-conserved and encodes the intracellular component of PfEMP1 . Based on distance tree analysis , DBL domains are subdivided into six major classes ( DBLα , β , γ , δ , ε , ζ ) and CIDR domains into four ( CIDRα , β , γ , δ ) [7] . Each class can then be further subdivided into subclasses ( DBLα0 . 1 , DBLα0 . 2 , etc ) [8] . Previous studies have observed ectopic ( non-allelic ) recombination between subtelomeric var genes [9]–[13] and it has been proposed that this might account for a large part of the var gene diversity observed within and between species [14] . However it is not known whether var gene recombination is sufficiently frequent to be the primary driver of var gene diversity , nor whether this recombination occurs mainly in meiosis , which takes place in the mosquito , or during mitosis , which includes the entirety of the intraerythrocytic stages within the human host . It is also unclear whether recombination obeys particular patterns , and how this might relate to var gene structure . Understanding how the parasite manages to generate such an extreme level of sequence diversity while preserving the overall architecture and biological functionality of the var gene repertoire is fundamental to understanding P . falciparum pathogenesis . We used an experimental evolution approach to systematically investigate the mechanisms that drive var gene diversity , performing whole genome sequence analysis on>200 clonal parasites . Parasites cultured in vitro in human erythrocytes were regularly sub-cloned to isolate single infected red blood cells , so that mutations arising in asexually dividing cells could be detected using whole genome sequencing of the expanded progeny derived from these cells ( Fig . 1C ) . We generated these ‘clone trees’ for geographically diverse P . falciparum strains , comparing parasite genomes separated by 20–30 cycles of replication ( S1 Fig . ) . The scale of our dataset allowed us to perform a comprehensive analysis of mitotic var gene recombination using observed recombination events for the first time . This revealed how var gene sequence polymorphism – and , by inference , parasite antigenic variation – can be continuously generated during asexual division over the course of a single human infection .
We began by analysing the 3D7 strain of P . falciparum as this has the most complete genome assembly of any lab isolate . 19 individual parasites were cloned from 3D7 by limiting dilution and cultured until sufficient parasite quantity was reached for whole genome sequencing . From 2 of these 19 subclones , another 3 rounds of clonal dilutions were performed over a combined culturing period of 203 days ( S1 Fig . ) . Whole-genome sequencing of a total of 37 subclones identified 20 de novo single nucleotide substitutions ( hereafter referred to as Single Nucleotide Polymorphisms or SNPs ) distributed throughout the genome , and 40 de novo structural variations comprising 10 duplications , 8 deletions and 22 translocations ( Fig . 2 , S2–S4 Table ) . Our analysis focused solely on de novo mutations , i . e . found in one or more subclone ( s ) but not in the parental clone . As expected , the first generation of subclones contained more mutations than subsequent generations , because they were derived from a standard 3D7 culture that had been growing for several months prior to beginning our experiment . Of the 19 structural variations that affected coding regions , all were in var genes . Of the remaining 21 structural variations , all were in telomeric regions or within internal regions of chromosomes that contained clusters of var genes . From these data we concluded that de novo structural variations of the P . falciparum genome occur relatively frequently during the mitotic intraerythrocytic life cycle and are highly concentrated in and around var genes . Detailed inspection of these structural variations revealed that they took many forms ( Fig . 3B–F ) but all involved ectopic recombination between different var genes located either in the immediate vicinity on the same chromosome , or on different chromosomes . For example , Fig . 3A&B and S2 Fig . show a gene conversion event involving two subtelomeric var genes , in which the end subtelomere of chromosome 4 was duplicated and replaced the subtelomere of chromosome 9 . The translocation breakpoints occurred within var genes on these two chromosomes , referred to as var4 and var9 , creating a chimeric var4/var9 sequence . This gene conversion event was confirmed by capillary sequencing of PCR products , and the var4/var9 chimera was subsequently inherited by downstream parasite clones . There were also instances of two var genes recombining with multiple crossover points ( Fig . 3D , S3 & S4 Fig . ) and recombination between more than two var genes ( Fig . 3E and Fig . S2 ) . Recombination in subtelomeric var genes involved both reciprocal exchange and non-reciprocal gene conversion in which the entire end of one chromosome was duplicated and replaced the corresponding section of the other chromosome . In addition to recombination between subtelomeric var genes on different chromosomes , we found recombination between internal var genes on the same chromosome ( Fig . 3F , S5 & S6 Fig . ) . To establish whether mitotic var gene recombination is a common feature among other P . falciparum strains , we generated similar clone trees for the Dd2 and W2 strains of drug-resistant P . falciparum from Southeast Asia , and for the HB3 strain of drug-sensitive P . falciparum from Honduras ( S1 Fig . ) . The Dd2 , W2 and HB3 clone trees were generated with a total of 7 , 3 , and 8 cloning rounds respectively , over a total of 298 , 119 and 250 days . This generated 56 , 20 and 81 subclones respectively , all of which were whole genome sequenced . Altogether , we observed 20 , 11 and 19 SNPs in the Dd2 , W2 and HB3 clone trees respectively , scattered across the genome ( Fig . 2 ) . As any mutation identified in a clone would have actually occurred in the previous generation , we can only use clones from the second generation onwards for calculating mutation rates ( Fig . 1C ) . From these data we estimate mutation rates of 4 . 07×10−10 , 3 . 63×10−10 and 3 . 78×10−10 SNPs per erythrocytic life cycle per nucleotide for 3D7 , Dd2 and HB3 respectively ( Table 1 ) . The W2 clone tree did not contain enough clones for a reliable mutation rate estimate . Despite geographic and drug-sensibility differences between 3D7 , Dd2 and HB3 , there is no statistically significant difference in SNP mutation rate between these strains ( Kruskal-Wallis , p>0 . 7 ) ; See Methods & Table 1 for rate calculations . The rates calculated here are similar to estimates for 3D7 and Dd2 by Bopp et al 2013 , suggesting that SNP mutation rates are relatively constant across P . falciparum strains . A complete high-quality genome assembly is currently only available for 3D7 , which makes mapping difficult in other strains , especially in var genes . For these strains we therefore focused only on a detailed analysis of recombination events in var exon 1 , which encodes the extracellular domains responsible for antigenic diversity . In Dd2 and W2 , we observed 11 and 13 instances of var exon 1 recombinations in 7 and 6 pairs of recombining var genes respectively ( Table 2 ) . In contrast , there were no recombination events in HB3 clones , despite similar breadth and length of the clone trees and despite the var gene assemblies of HB3 and Dd2 being of a comparable standard . We estimate that the rate at which pairs of var genes undergo exon 1 recombination in the Dd2 clone tree is 2 . 32×10−3 per erythrocytic life cycle , compared with 9 . 48×10−3 , 8 . 47×10−3 and 8 . 81×10−3 SNPs per life cycle in 3D7 , Dd2 and HB3 respectively ( see Methods for rate calculations ) . This means that in the Dd2 strain , for every 48-hour intraerythrocytic generation approximately 0 . 2% of parasites will have undergone a var exon 1 recombination event producing a new chimeric var gene . We further analysed the var gene recombination rate by comparing the ratio of total de novo var recombinations to SNPs observed in the clone trees . Our mutation rate calculations given above exclude mutations found in the first generation of each clone tree because these mutations occurred prior to our beginning the experiment after an unknown length of time in culture ( Fig . 1C ) . However , mutations identified in this first clone tree round must have arisen during the intraerythrocytic stage , at some point between the original cloning of 3D7/HB3/Dd2 in the 1980s and the start of our clone tree experiments . As demonstrated in this study and in Bopp et al [9] , the SNP mutation rate appears to be relatively constant between strains . Therefore , the number of SNPs found in first generation subclones becomes a proxy for the time of culture of the parent , i . e . the more SNPs identified , the longer the parent had been in culture since the original cloning step . Because of this constant SNP mutation rate , we can use all subclones , including the first generations subclones , to measure the ‘var recombination/SNP ratio’ , which indicates the number of var recombinations over a period of time . The overall var exon 1 recombination to SNP mutation ratio is 0 . 25 , 0 . 35 , 0 . 54 and 0 in the 3D7 , Dd2 , W2 and HB3 clone trees respectively ( Table 2 ) . Thus , while the SNP mutation rate is constant between different P . falciparum strains , the structural variant rate in HB3 var genes is significantly lower than in Dd2 ( p = 0 . 03 , Fisher's exact test ) , at least with the specific HB3 isolate used in our laboratory . The dynamic nature of var recombination in Dd2 is illustrated by a pair of internal var genes referred to here as var34 and var45 ( Fig . 4 ) . Recombination between var34 and var45 was first observed in a first generation subclone of the Dd2 clone tree ( Dd2_ ( A ) 1a ) . This was a double crossover recombination which created a new chimeric var gene ( ‘var34/var45’ ) while retaining the original versions of var34 and var45 ( Fig . 3C , Fig . 4 , S7 Fig . ) . Over the next three generations of the clone tree , most of the 31 descendants of the recombined clone inherited var34 , var45 and the var34/var45 chimera , with a few exceptions . Three descendants lost the var34 sequence , two lost the var34/var45 chimera , and two underwent further recombination of var34 and var45 producing new chimeras . In contrast , we observed no changes involving var34 or var45 in the branch of the clone tree that was not descended from the clone in which recombination between var34 and var45 had first occurred . The clone trees demonstrate that var gene recombination during intraerythrocytic growth is extensive and complex , and in at least three strains occurs nearly as frequently as SNPs . To investigate the patterns and mechanisms of var gene recombination , we collated all available data on chromosomal crossovers in var exon 1 ( Table 1 ) based on our clone tree experiments , together with data from previous smaller studies [9] , [15] as well as data obtained by sequencing the progeny of three genetic crosses ( Dd2xHB3 , 3D7xHB3 and 7G8xGB4 ) [16]–[18] ( Pearson & Miles , in prep ) ( discussed in more detail below ) . From these combined datasets we identified 110 crossovers in 46 pairs of var gene exon 1s , generating a mean of 2 . 4 ( range 1–12 ) crossover points per recombining pair of exons ( Table 2 , S4 and S8 Fig . ) . These data revealed a highly structured pattern of ectopic recombination . We found that all of the crossovers were in frame and in the vast majority of cases ( 109/110 ) the recombining sequences were of the same domain class , e . g . CIDRα almost invariably recombined with CIDRα ( Fig . 5 ) . Every var gene begins with a DBLα and CIDRα domain , followed by a variable number and order of other domains . Given that recombination only occurs between the same domain classes , this mechanism would preserve the overall architecture of chimeric var genes . Consistent with this , all observed chimeric var genes possessed a typical number of var domains . We observed only one exception to this , between 7G8var49 and 7G8var56 , in which recombination took place in an area of shared homology between a DBLα and a DBLδ domain [8] , producing a truncated gene with only two domains ( S8 Fig . ) . The new gene sequences generated by this structured process were extremely diverse , as var genes did not necessarily recombine with the most highly homologous var genes to themselves ( S9 Fig . ) . The average sequence identity of the two recombining domains was 68% , and that of the two recombining exons was only 63% . In the terminology used by Rask et al , 65% of recombinations were between domains of different sub-classes ( Fig . 6 ) . But despite this overall diversity , we observed that crossovers invariably took place within a short section of perfect sequence identity ranging in length from 4 to 48 bp ( median 15 bp ) . This section of sequence identity , which we refer to here as an identity block ( S10 Fig . ) , was typically located within a broader section of elevated sequence homology ( median 80 bp ) relative to the average for the two genes concerned ( Fig . 6 ) . When exon 1 var sequences were aligned by BLAST analysis , 94% of crossovers were found to occur in a region of high similarity between the two recombining genes ( S11 Fig . ) , whereas only 59% would be expected if the point of recombination was random ( P<10−5 by chi-squared test ) . We compared the recombination rates found in group A , B and C var genes from our pooled var exon 1 recombination dataset . In total , we observed 19 instances of B×B recombination , 12 instances of C×C recombination and 2 instances of B×C recombination . In contrast we observed no instances of recombination involving group A var genes . The 3D7 reference genome contains 10 group A , 37 group B and 13 group C var genes , so these observations depart significantly from what would be expected if the pairing of ectopic sequences was random across all var genes , ( p = <0 . 0028 , Fisher's exact test ) . These data suggest that group A var genes have a lower recombination rate than group B or C . The prevailing view of var recombination is that it occurs primarily during meiosis , based on the observation of recombination in the progeny of experimental crosses [10] , [19] . An alternative model is that var recombination in these progeny actually occurred during mitosis while the parasites were cultured in vitro prior to and after the cross itself [20] . To test these alternatives we identified and characterised var recombinations in genome sequence data from progeny of all three experimental genetic crosses , and compared them with analogous data from progeny of our clone trees . This comparison strongly supported the latter model – that the vast majority of var recombination occurs during mitosis – for three main reasons ( Fig . 7 ) . First , the recombination in both the ( purely mitotic ) clone trees and the crosses progeny share key features and appear to have been produced through the same process . Recombined var genes were always in frame without SNPs or InDels; crossovers took place in small regions of sequence identity; overall var exon 1 domain architecture was preserved because domains only recombined with other domains of the same class ( with one exception ) , and recombination often involved multiple crossovers between the same two var genes , with an average of 2 . 6 and 2 . 4 crossovers per recombining var pair in the clone trees and crosses progeny respectively ( S7 Table ) . Second , within the crosses progeny we could find only two cases where the var genes in a recombining var pair were derived from different parents . This is not what would be expected if recombination occurred during meiosis when the parasite was diploid . During meiosis , each progeny inherits on average half its chromosomes and associated var genes from each parent . If heterologous chromosomes only recombined at that stage , we should observe chimeric var sequences derived from var genes of different parents or from the same parent in an approximately 50∶50 ratio ( Fig . 7 ) . In contrast , we found that 26 out of 28 recombinations were ‘intra-strain’ , i . e . involving pairs of var genes from the same parent . The simplest explanation is that these recombination events occurred during mitosis while the parental strains were in culture prior to the cross . Finally , knowing the average mitotic var recombination rate calculated using our clone trees allows us to test whether the number of var recombination events observed in the crosses progeny is consistent with the culturing time before and after the cross . In our clone trees and other experiments [9] , the SNP mutation rate is similar across strains ( see above ) . Thus , the number of SNP mutations in a subclone is directly proportional to the number of intraerythrocytic life cycles that occurred before the last cloning step . The number of days of culture can then be inferred from the number of SNP mutations observed . Based on the number of de novo SNPs identified in the crosses progeny and our mitotic SNP mutation rate , we calculated that the P . falciparum strains used for the crosses had undergone 124 and 133 days of in vitro culturing for 3D7xHB3 and HB3xDd2 respectively . i . e . there had been 124 and 133 days in culture from when the original isolates were first sub-cloned ( from NF54 for 3D7 , H1 for HB3 , etc ) , to the day the crosses' progeny were cloned out . For 3D7xHB3 and HB3xDd2 , these estimates could be validated by direct comparison with the actual known culturing data of the parental clones , provided by Tom Wellems ( NIH ) and Lisa Ranford-Cartwright ( University of Glasgow ) . The total recorded culture time , from isolation to progeny sub-cloning , was 89 . 5 and 99 . 5 days for the 3D7 and HB3 parasites used in the 3D7xHB3 cross – an average of 94 . 5 days . The equivalent time for the HB3 and Dd2 parasites used in the HB3xDd2 cross was 89 and 138 days respectively – an average of 113 . 5 days . These actual dates are very similar to the dates prdicted by our mutation analysis , which suggested that the 3D7xHB3 and HB3xDd2 isolates used for the crosses had been cultured for 124 and 133 days , where the real culture times were reassuringly close at approximately 94 . 5 and 113 . 5 days respectively ( S7 Table ) . If we assume that the average mitotic var recombination rate is 6 . 44×10−4 recombinations per life cycle ( the mean from the recombination rates calculated in the 3D7 , Dd2 , W2 and HB3 clone trees ) , then we would expect to find about 0 . 34 var recombinations per progeny sample . Again , these predictions , which are based on purely mitotic recombination , are very similar to what we observe: 0 . 38 , 0 . 20 and 0 . 59 in 3D7xHB3 , HB3xDd2 and 7G8xGB4 , respectively . Therefore , the number of var recombinations found in the crosses progeny is consistent with what is expected from the number of mitotic events that the parasites have gone through both before and after the cross . No additional input of recombination from meiosis is required to explain the data .
By analysing whole genome sequences from hundreds of subclones derived from asexually replicating parasites , we have generated a large dataset of mitotic mutations in the P . falciparum genome . We found that var genes of groups B and C undergo ectopic recombination at a rate which greatly exceeds that of structural variation in other parts of the P . falciparum genome , and is not dissimilar to the SNP mutation rate . Recombining var genes often have low levels of sequence homology , and a pair of var genes can recombine at different positions on different occasions , so the new var sequences generated by this process are extremely diverse . However , the recombinant sequence is invariably in frame and free of errors , and var gene domain architecture is preserved . This process has presumably evolved to generate antigenic diversity in P . falciparum during the course of a single infection to evade the human immune response . Our SNP and structural variant detection method by whole-genome sequencing was validated by PCR , capillary sequencing or both ( S2–S4 Table ) . Sander et al recently published an analysis of var gene recombination in progeny of 3D7xHB3 and HB3xDd2 [19] . They reported that recombination breakpoints are concentrated near low folding free energy DNA 50-mers , with a minimal required sequence identity of around 20 bp with 10% mismatch . Three out of the four progeny samples in which they identified var recombinations were available to us . We independently found the exact same var recombination breakpoint coordinates , and came to the same conclusion that crossover events take place within short regions of high homology . Generating diversity by structured recombination of regional gene segments is somewhat analogous to the way in which the vertebrate immune system generates diversity by V ( D ) J recombination of immunoglobulin and T cell receptor genes . However , our data show that the molecular mechanism of var gene recombination is markedly different from that of V ( D ) J recombination in two respects . First , the site of recombination was not marked by any specific sequence motif equivalent to the 'Recombination Signal Sequence' that targets recombinases in V ( D ) J recombination ( S15 Fig . ) , and we observed instances of the same pair of var genes recombining at different positions on separate occasions ( S8 Fig . ) . Second , V ( D ) J recombination depends on repair of DNA double-strand breaks ( DSBs ) by non-homologous end joining ( NHEJ ) , a molecular pathway that appears to be missing or atypical in P . falciparum [21] , [22] . All the instances of ectopic recombination we observed were achieved without errors or indels , indicating that the mechanism of DSB repair is more similar to homologous recombination ( HR ) than NHEJ , as this is generally high fidelity whereas NHEJ is error-prone . It is interesting to note that , lacking a typical NHEJ system , P . falciparum is presumably reliant on the machinery of HR for DSB repair . This is somewhat paradoxical given that the parasite is haploid for the majority of its life cycle , including the entirety of the mitotic intraerythrocytic stages . It therefore lacks homologous chromosome copies to use as templates for HR-mediated DSB repair . These conditions may be conducive for non-allelic homologous recombination during mitosis , with non-allelic templates being employed in the absence of homologous chromosomes . While the pattern of var gene recombination therefore differs from well-known methods for generating diversity , it is also strikingly different from conventional HR in which crossovers take place in a region of extremely high sequence homology ( >92% ) , typically extending over hundreds of base pairs [23] . Without extensive homology as a cue for pairing sequences from different regions of the genome , some alternative mechanism must exist for co-localisation of var genes before initiation of the recombination-inducing lesion [24] . Var genes are located in more than 30 regions of the genome , i . e . in most subtelomeric regions and in internal regions of four chromosomes . P . falciparum telomeres are tethered to the nuclear periphery where they form several distinct clusters [10] and it has been demonstrated by FISH that both subtelomeric and internal var genes aggregate within these clusters [25] . This nuclear clustering is believed to play an important role in the transcriptional activation and silencing of var genes [25] and it might also be a mechanism of co-localising sequences from different regions of the genome prior to ectopic recombination [10] . Our genome-wide approach in 3D7 subclones identified structural variants in and around var genes , indicating that these regions are more recombinogenic . In these samples , we did not observe any recombination in other variant surface antigen families such as rifin or stevor , despite the fact that there are nearly three times as many rifin genes as var genes in the genome . The rate of var gene recombination differed depending on var gene Ups type and parasite strain . We found a lower rate of recombination in HB3 than in 3D7 , Dd2 or W2 parasites . The HB3 genome does not appear to be unable to recombine var genes per sec , as we did identify a recombined exon 1 var gene in the 'parent' strain used for one of our two HB3 clone trees and all clonal descendants from that parent inherited the recombined var gene . We cannot use this example to measure the recombination rate because we do not know the time in culture prior to performing our first clonal dilution , but the observation indicates that mitotic var exon 1 recombination in HB3 does occur , although perhaps at a rate that is too slow to be detected over the course of our clone trees . We have also observed instances of var recombinations in genetically modified HB3 lines ( Hamilton et al , manuscript in preparation ) . A longer time in culture between clonal dilutions , or more clones per dilution , or both , would be necessary to capture enough var recombination events in this strain to determine its rate accurately . What factors influence var gene recombinogenicity , and whether there is a connection with drug resistance , remains to be determined . It is possible that HB3 is inherently less recombinogenic than other strains , or that the particular HB3 isolate we used in our clone experiments had , for unknown reasons , lost or reduced its capacity to undergo var recombination . Moreover , although var gene recombination in recently adapted P . falciparum field isolates does occur ( Hamilton et al , manuscript in preparation ) , the frequency of such events in the wild and in vivo is currently unknown . The fact that recombination was more common in group B and C var genes is consistent with the observation that group A vars from geographically diverse strains are more conserved than those of groups B and C [26] . Var recombination group-specificity also has implications for the underlying mechanism , raising the possibility that the upstream sequence or some other group-specific sequence feature plays a role in co-localising var genes prior to recombination . Var genes can also be classified into six groups based on their DBLα sequences [27] . However there was no association between var recombination and that specific subgrouping ( Chi-square , p>0 . 40 ) . Whether some specific var genes are inherently more recombinogenic than others , as the Dd2var34/var45 example could suggest , requires an even larger dataset , perhaps involving several years of clone tree evolution , to properly assess . Further studies such as these will also be required to test the hypothesis that specific var genes being actively expressed in a given strain are more likely to recombine . If the latter hypothesis is true , it would partly explain the lack of recombination in group A genes that we observed , as cultured lab strains typically express very few , if any , of that subgroup [28] . It would also fit with the hypothesis that longer and more conserved var genes show a lower activation rate and lower sequence diversity [29] . Further investigation into DNA repair and recombination in P . falciparum will be essential to identify the precise molecular mediators of var recombination and elucidate the mechanism in full . To date , var gene diversity has typically been studied using short DBLα tag sequences from field isolates [30] , [31] , or from 7 sequenced lab strains [8] , [26] , [32] . In either case , these var sequences represent snapshots of diversity at a particular time and place . They reflect the polymorphism found in the wild , under the intense pressure of the human immune system . Our study uses a novel approach to analyse var gene sequences as they are being generated , i . e . without any immunological selection . However , it has to be noted that lethal recombinations , for example between internal and subtelomeric var genes , would never be observed due to the resulting loss of genetic material . But only selection-free experiments such as this one can accurately measure the var recombination rate and the diversity of chimeric sequences being generated . As var gene assemblies from worldwide sampling will become available ( Otto et al , manuscript in preparation ) , it will be of major interest to compare the sequence diversity under selection versus our dataset , for example to differentiate true recombination hotspots from selection hotspots . This will also help modelling studies to define and test recombinational constraints [32] . Comparison with our clone tree data suggests that much of the var gene recombination observed in the progeny of experimental genetic crosses likely arose during mitosis prior to the cross itself ( Fig . 7 ) . The dominance of mitotic recombination is inherent to the fact that the life cycle of P . falciparum consists of a single meiotic event for dozens or hundreds of mitoses . We propose a model whereby var gene sequence polymorphism is mainly generated during the asexual part of the life cycle , while meiosis has the crucial function of creating a new repertoire of ∼60 var genes from two different parents . There is also a clear biological advantage in having extensive recombination in antigen-coding genes take place during mitosis: it allows the continuous generation of antigenic diversity inside a human host over the course of a single infection . An adult with malaria fever typically carries on the order of 107 to 1010 infected erythrocytes ( parasitaemia 5 to 5 , 000 parasites/µl ) , so this implies that on the order of 104 to 108 new mosaic var gene sequences could be generated through recombination every two days in a single infected individual . Var gene expression is a mechanism of immune evasion , enabling parasitized erythrocytes to sequester in small blood vessels and thus avoid circulation through the spleen . But var genes are themselves the targets of host immunity as they elicit a strong antibody response , which the parasite evades by switching expression between different var genes . Ectopic recombination of var genes may be of limited value in the initial phase of infection , as each new var sequence is present in only a tiny fraction of parasites . However it could be of great importance in chronic asymptomatic infection , which is increasingly recognized a crucial problem for malaria elimination [33] . As infection progresses , the parasite population carried by a single individual has the potential to accumulate an almost limitless repertoire of antigenic variants , thus allowing the evolutionary selection of variants which are best equipped to evade the immune response in that particular individual over a protracted period of time .
Our 3D7 ‘parent’ clone was provided by Bob Pinches . It was one of several 3D7 reference stabilates produced in 1989 following the isolation of 3D7 from NF-54 . Thus , its genome should be very close to the canonical 3D7 reference genome . Dd2 , W2 and HB3 parasites were obtained originally from the Malaria Research and Reference Reagent Resource Center ( MR4 ) . All P . falciparum strains were cultured in human O+ erythrocytes with heat-inactivated 10% pooled human serum as in [34] . For the clonal dilution step , parasites were ultra-diluted into a 96-well plate , to reach a theoretical concentration of 0 . 2 to 0 . 5 parasites per well , supplemented with 100 µl of complete medium at 2% haematocrit . The chamber was gassed every 2–4 days with medium changes on day 4 , 12 and 20 and 50% dilution with fresh 2%hct medium on day 8 , 16 and 24 . Positive wells were identified by medium colour change confirmed microscopically after day 16–24 post clonal dilution . The clones were further cultured in either 6-well plates or 10 ml flasks at 4–5%hct . Cultures were pelleted , mixed with RPMI to a total volume of 2 ml and frozen at −80°C for DNA extraction ( see below ) , with most clones also being frozen in glycerolyte for reculture . After 1–2 months growth , one clone was arbitrarily selected for the next round of clonal dilution . DNA extraction was performed with Qiagen whole blood midi kits [12145] following the manufacturer's instructions . DNA was quantified using Qubit 2 . 0 Fluorometer ( Invitrogen , Carlsbad , CA , USA ) according to manufacturer's instructions , with 1 µl from the extraction step . An average of 2 µg DNA ( range 1–5 µg ) per sample was submitted in 100 µl volume for PCR-free library preparation for next generation sequencing on the Illumina HiSeq , as described in [35] , except without the initial PCR amplification step . All whole-genome sequencing output samples ( 100 bp paired-end reads ) were processed through the MalariaGEN pipeline , as described in [35] . Briefly , FASTQ files were mapped to the P . falciparum 3D7 reference genome ( PlasmoDB version 5 . 5 ) using BWA with default parameters ( See S1 Table for coverage of each sample ) . For HB3 , Dd2 and W2 , the resulting BAM files were also mapped to the HB3 and Dd2 genomes from BROAD ( assembled in 1187 and 2837 supercontigs , respectively ) and to a Dd2 assembly made following the PAGIT protocol [36] ( 14 chromosomes +636 contigs , by T . Otto , unpublished data ) . When an in silico control was needed , these reference genomes were modified according to the predicted chromosomal rearrangements . For the discovery of translocations within var genes , samples were also mapped to each var sequence from their respective strain ( sequences from [8] ) . This allows the identification of potential translocations between var genes that are located on the same chromosome . SAMtools mpileup followed by BCFtools were used for SNP detection [37] . Parent samples were analysed with the default parameters except for the following changes: Extended BAQ computation , unlimited read depth , use Bayesian inference , “−P” option <0 . 9 ( probability that the site is a variant ) . Progeny samples were analysed with the following parameters: Extended BAQ computation , unlimited read depth , use Bayesian inference . The strict filter for progeny samples directly removes false positives . As the downstream analysis only keeps de novo SNPs called in progeny but not in parent , the looser filter for parent samples set here is also directed to remove false positives . Each progeny VCF file was then parsed using R with the following criteria: variant Quality >120 , mapping quality >25 , reject bi- or tri-allelic calls , at least 5 non-ref reads , number ref reads less than 20% of all reads . From this output list , each potential SNP was then visualised on LookSeq [38] , an online tool in which sequence coverage is represented by a pile-up of blue reads matching the 3D7 reference genome , with mismatches in red . Finally , 95% of all detected SNPs had maximal quality ( 222 ) and a mapping quality >54 . Note that a “progeny” sample was also analysed as a “parent” if that sample was subcloned . In order to detect translocations involving var genes located on the same chromosome , reads from each sample were mapped with BWA to separate var gene sequences , i . e . each var sequence is considered its own “chromosome” . DELLY identifies structural variation ( >150 bp ) by integrated paired-end and split-read analysis [39] . Paired ends mapping to loci with an increased or reduced spacer distance relative to what is expected based on insert size suggests either a deletion or an insertion , respectively . Face-away reads with increased spacer distance and higher coverage is suggestive of a duplication , and reads mapping to different chromosomes ( ‘trans-locus reads’ ) indicate a translocation ( or recombination ) . DELLY was run on each 3D7-mapped BAM file , as well as all var gene mapped BAM files , with default parameters . For deletions and duplications , outputs were parsed to a minimum of 20 reads and a maximum distance of 50 , 000 . For translocations , outputs were parsed to a minimum of 10 reads . Similar to SNP detection , non-specific hits were discarded by detecting structural variants in progeny that were also identified in parents . All resulting hits were inspected on LookSeq . Progeny clones from the P . falciparum crosses 3D7xHB3 [18] ( 16 samples +2 parents ) , 7G8xGB4 [16] ( 27 samples +2 parents ) and HB3xDd2 [17] ( 30 samples +2 parents ) were whole genome sequenced as part of another genetic study ( Miles , Pearson et al , manuscript in preparation ) . These clones were also searched for var gene translocations . SNPs were validated by PCR amplifying a 150–800 bp fragment containing the putative SNP and capillary sequencing the PCR product ( S6 Table ) . Chimeric var genes identified by DELLY were validated through PCR , by designing primers that bridge the putative translocation site and only amplify product in samples possessing that translocation ( S6 Table ) . The chimeric PCR product was then sub-cloned into a pCR Blunt-end Invitrogen vector backbone and 1–3 positive clones were capillary sequenced . Quantitative real time PCR ( qPCR ) was used to validate copy number variations ( CNV ) in var genes . Primers were designed flanking the putative translocation site in the PF3D7_0421100-0421300 chimera on chromosome 4 . Assay was conducted in triplicate on a 96 well plate using a Roche Light Cycler 480 real-time PCR system . We estimated gene copy number relative to wild type 3D7 , and normalized samples against the AMA1 gene ( which has one copy in 3D7 ) . Conditions as described in [40] , [41] . We estimated the SNP mutation rate per life cycle for each generation in the 3D7 , Dd2 and HB3 clone trees , based on the number of SNPs identified per clone within that generation , and the number of life cycles since the previous clonal dilution: Where is the SNP mutation rate per life cycle for that clonal dilution generation , ∑S is the total number of SNPs identified for all the clones in that clonal dilution generation , ∑C is the total number of clones that were analysed in that clonal dilution generation , L is the number of life cycles that took place between the previous clonal dilution and the current one ( see below ) , and G is the Genome size , taken as 23 . 3×106 bp [42] . The number of life cycles was calculated as follows: Where d is the days in culture between clonal dilutions , 24 to convert to hours , and t is the estimated life cycle time in hours . t has been measured in HB3 and Dd2 at 49 . 7 and 44 . 1 hours , respectively [43] . 48 hours was used for 3D7 . We took into account the number of clones in each clonal dilution generation to produce a weighted average mutation rate ( S5 Table ) of 4 . 07×10−10 , 3 . 63×10−10 and 3 . 78×10−10 SNPs per erythrocytic life cycle per nucleotide for 3D7 , Dd2 and HB3 respectively . We did not calculate a SNP mutation rate in W2 as the clone tree was not as extensive and we did not feel there was enough data for an accurate calculation . We observed one instance of two SNPs occurring ‘back to back’ , i . e . two consecutive nucleotides were mutated from the reference , and one occasion where a single nucleotide separated two SNPs . These were all counted as separate SNP mutations for the rate calculation , though it may be that they represent single ‘mutation events’ at a molecular level . The Bopp et al [9] published mutation rates for 3D7 and Dd2 in the absence of drug selection was 1 . 7×10−9 and 3 . 2×10−9 , respectively . However , they adjusted their raw data to account for deleterious non-synonymous SNPs that may have been eliminated by purifying selection before being identified . Using their methods , our raw mutation rates become 4 . 28×10−9 , 3 . 82×10−9 and 3 . 98×10−9 SNPs per erythrocytic life cycle per nucleotide for 3D7 , Dd2 and HB3 respectively . Thus , there is high consistency both between ours and Bopp et al's SNP mutation rate estimates and between different P . falciparum lab strains . Our calculation of exon 1 var gene recombination in the Dd2 clone tree was very similar to our SNP mutation method described above ( S5 Table ) . Because we often observed multiple crossovers per recombining var gene pair , we calculated two recombination rates: First , we calculated the rate at which pairs of var genes recombine in exon 1 in Dd2 as 2 . 32×10−3 per life cycle , i . e . a pair of var genes will recombine in ∼0 . 2% of infected erythrocytes per life cycle . We did not calculate the rate per life cycle per nucleotide because our analysis was focused on var exon 1 , so dividing by the total nuclear genome was not appropriate . We therefore also calculated ‘per life cycle’ rates for SNP mutations to enable comparison between var recombination and SNPs . Second , we calculated the rate of var exon 1 crossover events at 3 . 45×10−3 per life cycle . This is higher than the rate of recombining var pairs per life cycle because some pairs have >1 crossover . | Malaria kills >600 , 000 people each year , with most deaths caused by Plasmodium falciparum . A family of proteins known as P . falciparum erythrocyte membrane protein 1 , PfEMP1 , is expressed on the surface of infected erythrocytes and plays an important role in pathogenesis . Each P . falciparum genome contains approximately 60 highly polymorphic var genes encoding the PfEMP1 proteins , and monoallelic expression with periodic switching results in immune evasion . Var gene polymorphism is thus critical to this survival strategy . We investigated how var gene diversity is generated by performing an in vitro evolution experiment , tracking var gene mutation in ‘real-time’ with whole genome sequencing . We found that genome structural variation is focused in and around var genes . These genetic rearrangements created new ‘chimeric’ var gene sequences during the mitotic part of the life cycle , and were consistent with processes of mitotic non-allelic homologous recombination . The recombinant var genes were always in frame and with conserved overall var gene architecture , and the recombination rate implies that many millions of rearranged var gene sequences are produced every 48-hour life cycle within infected individuals . In conclusion , we provide a detailed description of how new var gene sequences are continuously generated in the parasite genome , helping to explain long-term parasite survival within infected human hosts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"immune",
"system",
"proteins",
"apicomplexa",
"parasite",
"groups",
"genomics",
"proteins",
"plasmodium",
"antigens",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"antigenic",
"variation",
"dna",
"recombination",
"immunology",
"parasitology",
... | 2014 | Generation of Antigenic Diversity in Plasmodium falciparum by Structured Rearrangement of Var Genes During Mitosis |
Autochthonous cutaneous and visceral leishmaniasis ( VL ) caused by Leishmania martiniquensis and Leishmania siamensis have been considered emerging infectious diseases in Thailand . The disease burden is significantly underestimated , especially the prevalence of Leishmania infection among HIV-positive patients . A cross-sectional study was conducted to determine the prevalence and risk factors associated with Leishmania infection among patients with HIV/AIDS living in Trang province , southern Thailand , between 2015 and 2016 . Antibodies against Leishmania infection were assayed using the direct agglutination test ( DAT ) . DNA of Leishmania was detected by ITS1-PCR using the buffy coat . Species of Leishmania were also identified . Of 724 participants , the prevalence of Leishmania infection was 25 . 1% ( 182/724 ) using either DAT or PCR assays . Seroprevalence of Leishmania infection was 18 . 5% ( 134/724 ) , while Leishmania DNA detected by the PCR method was 8 . 4% ( 61/724 ) . Of these , 24 . 9% ( 180/724 ) were asymptomatic , whereas 0 . 3% ( 2/724 ) were symptomatic VL and VL/CL ( cutaneous leishmaniasis ) . At least five species were identified: L . siamensis , L . martiniquensis , L . donovani complex , L . lainsoni , and L . major . Multivariate analysis showed that CD4+ levels <500 cells/μL and living in stilt houses were independently associated with Leishmania infection . Those who were PCR positive for Leishmania DNA were significantly associated with a detectable viral load , whereas non-injection drug use ( NIDU ) and CD4+ levels <500 cells/μL were potential risk factors of Leishmania seropositivity . A magnitude of the prevalence of underreporting Leishmania infection among Thai patients with HIV was revealed in this study . Effective public health policy to prevent and control disease transmission is urgently needed .
Co-infection of leishmaniasis and human immunodeficiency virus ( HIV ) is a major public health problem globally . Leishmania and HIV each promote the activation of the other , causing host immune impairment . The co-infection results in treatment failure , high relapse , and high mortality rate [1] . Meta-analysis has revealed that the direct agglutination test ( DAT ) gave high sensitivity and specificity for serodiagnosis of VL when compared to other serological tests [2] . However , low sensitivity of serological tests for VL diagnosis has been shown among these patients due to defective host immunity [3] . To increase the sensitivity of Leishmania DNA detection , the polymerase chain reaction ( PCR ) using blood samples has been suggested [4] . The internal transcribed spacer 1 ( ITS1 ) -PCR method has been recommended to detect Leishmania DNA [5] . In Thailand , the first autochthonous VL case was documented in a 3-year-old girl living in a southern province in 1999 [6] . Until 2012 , L . siamensis was firstly reported in a patient with HIV in Trang province . Since then , CL and/or VL have been sporadically reported in immunocompetent and immunocompromised patients predominantly in the south and north of Thailand and about 40% were patients with HIV/AIDS . L . martiniquensis was the predominant causative agent while L . siamensis was indigenously reported in only one Thai patient [7] . Information of the true prevalence of Leishmania infection among Thai patients with HIV , a high risk group , is still lacking . Thus , the objectives of this study were to determine the prevalence and the risk factors associated with Leishmania infection among patients with HIV/AIDS in Trang province , southern Thailand .
A cross-sectional study of Leishmania infection was conducted between February 2015 and February 2016 . Eligible participants were >18 years old and attending an HIV clinic , Trang Hospital , Trang province . They visited the clinic every 6 months for follow-up testing and to receive antiretroviral therapy ( ART ) . They lived in ten districts of Trang province , other nine provinces located in the south , and three provinces in other regions of Thailand . Clinical information of participants was collected from patients’ medical records . Written informed consent was obtained from all participants . All participants were >18 years old . All analyzed data were anonymized . The research protocol was approved by the Ethics Committee of the Royal Thai Army Medical Department and the Ethics Committee of Mahidol University , Thailand . Eight milliliters of EDTA anti-coagulated blood samples were collected . The whole blood was centrifuged at 900 × g for 10 minutes to separate the plasma and buffy coat and was then kept at −20°C until further use . Seropositivity of Leishmania infection was defined as detection of antibodies in individuals who were exposed to Leishmania infection and being either symptomatic or asymptomatic . Asymptomatic Leishmania infection was defined as individuals who experienced no symptoms of VL but presented a positive test by DAT or PCR assays . Symptomatic VL was defined as individuals having a history of fever lasting at least 2 weeks with splenomegaly . One or more of the following clinical characteristics may be observed: hepatomegaly , weight loss , anemia , leucopenia , thrombocytopenia , and hypergammaglobulinemia . Detection of the parasites must be confirmed under microscopic examination or by PCR assay using any clinical samples ( e . g . , bone marrow aspirates , lymph node , blood , and/or other biopsy samples ) . Leishmania antibodies were assayed using the commercial DAT kit ( Biomedical Research ) according to the manufacturer’s instruction . The positive plasma control was obtained from confirmed VL cases using the PCR method . For the negative control , plasma from healthy individuals was used . The cutoff value of positive DAT titers was ≥1:100 following manufacturer recommendation . DNA was extracted from 200 μL of buffy coat sample using Gen UP gDNA Kit ( Biotech ) . Nested PCR was used to amplify the ITS1 region of the ribosomal DNA ( rDNA ) gene of Leishmania . In the primary PCR , primers LITSR and L5 . 8S were used to amplify the 319–348 amplicons [8] . The newly designed secondary primers LITSR2 ( CTG-GAT-CAT-TTT-CCG-ATG-ATT ) and L5 . 8S inner ( GTT-ATG-TGA-GCC-GTT-ATC-C ) generated 230–280 amplicons depending on Leishmania species . PCR reactions were performed using the MJ Mini thermal cycler ( BioRad ) in volumes of 25 μL , containing 12 . 5 pmol of each primer , 0 . 2 mM dNTP , 1 . 5 mM MgCl2 , 1× PCR buffer , 1 U of Taq DNA polymerase , and 4 μL of DNA template . DNA of L . martiniquensis promastigotes ( MHOM/MQ/92/MAR1 ) was used as the positive control . The condition was started by pre-denaturation at 94°C for 3 minutes followed by 35 cycles: denaturation at 94°C for 1 minute , annealing temperature at 54°C for 30 seconds , and extension at 72°C for 30 seconds . Final extension was at 72°C for 5 minutes . PCR products were separated by electrophoresis in 1 . 5% agarose gel stained with SYBR Safe ( Invitrogen ) . The results were visualized and documented by Molecular Imager Gel Doc XR+ System with Imager Lab 3 . 0 ( BioRad ) . Positive PCR products were sent to U2Bio Co . Ltd . , South Korea for sequencing . Chromatograms were validated using BioEdit version 7 . 0 . 1 . The sequences were multiple-aligned with reference Leishmania strains retrieved from GenBank . The phylogenetic tree was constructed by using the neighbor-joining ( NJ ) method using the MEGA program , version 7 . 0 . The reliability was tested by 1 , 000 bootstrap replications and Tajima-Nei was selected for the DNA substitution model of phylogenetic analysis . To determine the risk factors and outcomes of Leishmania infection , standardized questionnaires were used . Enrolled subjects with HIV were interviewed face-to-face covering demographic data , socioeconomic status , clinical symptoms , and associated risk behaviors . The association between potential risk factors and Leishmania infection was assessed by univariate and multivariate logistic regression analysis . Odds ratios and 95% confidence intervals ( CI ) were calculated and p values <0 . 05 were considered statistically significant . All analyses were performed using STATA , version SE14 ( Stata Corporation , College Station , TX , USA ) . ( http://dx . doi . org/10 . 17504/protocols . io . j2dcqa6 )
A total of 724 participants with HIV were enrolled in this study . Of these , 643 ( 88 . 8% ) filled out questionnaires . Living areas of participants were as follows: 570 ( 88 . 6% ) lived in Trang province and 67 ( 10 . 4% ) lived in nine other provinces located in the south . Only five ( 0 . 8% ) were from other regions of Thailand . The mean age was 43 . 6 ± 8 . 5 years . The characteristics of the enrolled subjects are shown in Table 1 . Their clinical characteristics and risk behaviors during the past one year ( S1 and S2 Tables ) included history of injection drug users ( IDUs ) at 16 . 6% while 13 . 7% were non-injection drug users ( NIDUs ) . A total of 512 ( 79 . 6% ) subjects lived in non-stilt houses while 131 ( 20 . 4% ) lived in stilt houses . Most of the individuals ( 68 . 6% ) had CD4+ levels more than 500 cells/μL and only 9 . 6% were less than 200 cells/μL . Three categories of data analysis of the prevalence of Leishmania infection were performed ( Table 1 ) . The first group comprised patients who were either seropositive by DAT analysis with titers of ≥100 or positive by PCR assay . The second category involved patients who were seropositive by DAT analysis with titers of ≥100 , and the last group comprised patients who were positive only by PCR assay . The prevalence of Leishmania infection using positive results either by DAT or PCR assays was 25 . 1% ( 182/724 ) . Seropositive cases comprised 18 . 5% ( 134/724 ) while Leishmania DNA detection by PCR was 8 . 4% ( 61/724 ) . Only 1 . 8% ( 13/724 ) were positive using both methods ( Table 2 ) . Table 3 shows numbers of positive Leishmania infection using DAT , the titer of DAT , and PCR . Thus , the overall prevalence of asymptomatic Leishmania infection was 24 . 9 ( 180/724 ) . Tables 4 and 5 show affected areas of Leishmania infections . Regarding the analysis of those who were either seropositive by DAT or positive by PCR assay , the prevalence of Leishmania infection significantly differed among participants who lived in stilt houses compared to those living in non-stilt houses ( p = 0 . 03 ) , those who developed jaundice ( p = 0 . 02 ) , having opportunistic infection ( p = 0 . 002 ) especially tuberculosis ( p = 0 . 001 ) , and those having low CD4+ levels <500 cells/μL ( p = 0 . 003 ) ( Fig 1 ) . No significant difference was found among age group , gender , educational level , occupation , working outdoors at night , average income , years of HIV diagnosis , viral load , history of going abroad , drug use ( IDUs/NIDUs ) , pet/animal owner , animal shed nearby the house , plantation nearby the house , and bed net use ( S1 and S2 Tables ) . Two patients showed symptomatic VL . The first case was a 39-year-old herdsman living in Phuket province . He had a history of NIDU . He presented with a fever for more than two weeks . Many nodules were observed on the trunk . Laboratory findings revealed CD4+ levels at 173 cells/μL with an undetectable viral load . The DAT titer was 1:6400 . The causative agent was L . martiniquensis , which was identified by the nested ITS1-PCR using the buffy coat and skin biopsy . The patient was treated with amphotericin B . However , he died of disease progression one year after initial VL diagnosis . The second case involved a 41-year-old male who originally lived in Trang province . He worked on a rubber plantation . He was both an IDU and NIDU . He developed epistaxis and bleeding gums . Laboratory findings revealed pancytopenia and CD4+ levels of 622 cells/μL with an undetectable viral load . Intracellular amastigotes were observed from the lymph node biopsy with Giemsa stain . The DAT titer was 1:6400 and the causative agent was L . martiniquensis , which was identified using the nested ITS1-PCR of the buffy coat . The patient was treated with amphotericin B but he died within one week after treatment . Univariate and multivariate analysis of risk factors for acquiring Leishmania infections by DAT titers of ≥100 or positive by PCR assays are shown in Table 6 . Univariate analysis showed that participants who lived in stilt houses had higher risk ( OR = 1 . 58 , 95% CI = 1 . 04–2 . 39 ) of contracting Leishmania when compared with those living in non-stilt houses . CD4+ levels between 200 and 500 cells/μL were at higher risk than those who had CD4+ levels >500 cells/μL ( OR = 1 . 70 , 95% CI = 1 . 29–2 . 97 ) . After adjusting for age , gender , NIDUs , history of traveling abroad , pet owners and raising animals in housing areas , bed net use , animal shed and plantation nearby the house , underlying diseases , viral load , and duration of HIV diagnosis , multivariate logistic regression analysis revealed that those living in stilt houses had greater risk ( OR = 1 . 60 , 95% CI = 1 . 04–2 . 47 ) of acquiring the infection when compared with those living in non-stilt houses . In addition , those who had CD4+ levels 200–500 cells/μL ( OR = 2 . 13 , 95% CI = 1 . 36–3 . 32 ) and <200 cells/μL ( OR = 1 . 98 , 95% CI = 1 . 06–3 . 73 ) had higher risk of contracting Leishmania than those who had CD4+ levels >500 cells/μL . In addition , multivariate analysis of the associated risk factors of Leishmania infection using only seropositivity showed that participants who were NIDU had higher risk of presenting Leishmania seropositivity than those who were not ( OR = 2 . 23 , 95% CI = 1 . 27–3 . 92 ) . In addition , those who had CD4+ levels 200–500 cells/μL were also at higher risk of being seropositive than those who had CD4+ levels >500 cells/μL ( OR = 2 . 09 , 95% CI = 1 . 27–3 . 44 ) ( S3 Table ) . Multivariate analysis of the associated risk factors of Leishmania infection using only positive PCR results showed that those who had a detectable viral load >50 copies/mL were at higher risk of acquiring detectable Leishmania DNA in the blood ( OR = 2 . 31 , 95% CI = 1 . 01–5 . 29 ) than those who had an undetectable viral load after adjusting for those variables as mentioned above ( S4 Table ) . Of 61 samples , nucleotide sequencing was successful for 49 . These sequences , together with 16 reference sequences of different Leishmania species , were included to construct the phylogenetic tree using the NJ method ( Fig 2 ) . The phylogenetic analyses grouped the sequences into five separated clades . The majority of the samples ( 20 ( 40 . 8% ) and 13 ( 26 . 5% ) ) were closely related to L . siamensis and L . martiniquensis , respectively , whereas ten ( 20 . 4% ) were closely related to L . donovani complex , five ( 10 . 2% ) were related to L . lainsoni , and one ( 2 . 1% ) was related to L . major .
This was the first study providing important information of the prevalence of co-infection of Leishmania among Thai patients with HIV who had been regularly attending the HIV clinic in Trang province . The prevalence of Leishmania infection was approximately one-fourth of the 724 participants determined by either DAT or PCR assays . In the past , Leishmania infections were previously reported in five provinces in the south ( Surat Thani [6] , Phang-Nga [9] , Trang [10] , Songkhla [11] , and Satun [12] ) where most people mainly earn their living in agricultural sectors . The climate and humidity in the south are suitable for the sand fly’s habitat where potential sand fly vectors have been reported [13 , 14] . Our study revealed four new affected areas of Leishmania infections in the south ( Phuket , Krabi , Nakhon Si Thammarat , and Phatthalung provinces ) . Thus , at present , overall affected areas cover nine southern provinces . For VL/HIV co-infection , the sensitivities of DAT to detect antibodies against Leishmania infection were 50–84% [15] . Related studies of VL in immunocompetent subjects in endemic areas in India and Iran used DAT titers at different cutoff values ranging from 1:800 to 1:3200 [16–18] . A cutoff value at a titer of 1:100 was previously used to screen VL among HIV-positive patients who developed clinical symptoms in northeast Iran [19] . A low DAT titer of 1:200 was also detected in an immunocompetent VL Thai patient caused by L . martiniquensis [9] . In this study , one patient having a DAT titer of 1:100 also produced a positive PCR result . Therefore , a positive serological test at low titers could have diagnostic value to detect the infection . Thus , the cutoff values of DAT varied when conducted in different study populations as well as areas of study where cross-reactivity of DAT against other blood parasite infections could have occurred . A systematic review revealed that DAT titers detected in symptomatic patients were higher than those of asymptomatic patients [20] . Patients presenting very high DAT titer would have significantly greater disease progression than those presenting low titers [21] . In this study , clinical characteristics of VL were observed in two symptomatic cases that had DAT titers of 1:6400 together with positive PCR results . Thus , a close follow-up for those asymptomatic infections showing positive results of DAT or PCR is needed . In this study , serological and molecular diagnosis among individuals with HIV was not in concordance with other studies [22] . Our results showed that PCR positivity was low when compared with numbers of DAT positivity . The potential reasons among DAT positive individuals who might become PCR negative could include degradation and clearance of Leishmania DNA after infection , which corresponds to development of protective immunity due to the use of antiviral drugs for HIV . A positive PCR test among DAT negative individuals could occur when the individual was bitten by a Leishmania infected sand fly , but either immunity has not yet developed or antibody levels are too low to be detected by the methods employed , especially in HIV-positive populations [23] . Before using HAART , asymptomatic infection in HIV-Leishmania co-infection in Europe was 4–33% [24] . However , the incidence has been reduced to 20% after ART drugs had been given to all individuals with HIV [3] . In this study , all patients with HIV regularly received HAART treatment that could restore TH1 cytokine and antibody production [25] . HAART-treated HIV patients demonstrated a better ability to control Leishmania infection [26] . Many patients with subclinical VL did not develop clinical symptoms after taking HAART medications while some developed the disease [3] . Other risk factors ( e . g . , stage of HIV infection , parasite virulence , drug resistance , nutritional status , age , and gender ) may be involved in disease progression [1 , 27] . Prospective studies are needed to determine disease progression in this population . In this study , most enrolled participants ( 88 . 6% ) were not randomly selected and originally lived in Trang province . Thus , our results do not represent the prevalence of co-infection of Leishmania and patients with HIV of the country , and they do not represent the prevalence in each district of Trang province . However , this study showed the magnitude of the seroprevalence and significant numbers of subclinical results of Leishmania DNA detection circulating in the blood in Thai individuals with HIV . In southern Europe , IDUs were the most important risk factor accounting for more than 90% of all cases [28] . Our results showed no significant difference in prevalence among IDUs , while seroprevalence was associated with NIDUs . Leishmaniasis was also associated with socioeconomic status . In India , housing materials such as mud , plants , and earthen floors were risk factors for VL [20 , 29] . Our study showed that living in stilt houses was an independent associated risk factor for Leishmania infection . The presence of stilts provided an open area under the house that might have increased chances of sand fly bites to humans as well as providing resting sites for sand flies . Sand flies frequently bite at dusk [30] when most people spend their time at the open area of the house . In addition , CD4+ levels play an important role to protect the host from opportunistic infections . Related studies showed that the first episode of symptomatic VL diagnosis involved more than 80% of patients with HIV who had low CD4+ levels [26 , 31] . Our results also confirmed that low CD4+ levels ( <200 cells/μL ) as well as 200 to 500 cells/μL significantly increased the risk of Leishmania infection . In this study , Leishmania DNA detection by PCR assay was associated with detectable viral load . No correlation has been reported between PCR positivity and CD4+ levels , whereas the correlation between HIV viral load and parasitemia was observed among asymptomatic patients [22] . Clinical progression of HIV/AIDS was simultaneously promoted by VL . HIV infection enhances parasite growth by modulating significant cytokine response to Leishmania while the parasite upregulates viral expression [3] . Public health awareness of Leishmania infection in Thailand started when one autochthonous VL was reported in 1996 . However , at that time , the disease was uncommon as well as unfamiliar to most physicians , which could have led to a lot of underreporting of leishmaniasis cases in the past 20 years . Using the molecular method , this cross-sectional study was the first to systematically estimate the prevalence of Leishmania infection among patients with HIV , a high risk group , revealing not only L . siamensis and L . martiniquensis infection but also infections of other species ( e . g . , L . donovani complex , L . lainsoni , and L . major ) that already existed in the affected area . From the phylogenetic analysis , the ITS1 region is one of the more powerful targets used to discriminate the Leishmania species [32] . Our previous study showed that the ITS1 region had the highest sensitivity to detect L . martiniquensis and L . siamensis compared with the other genes: hsp70 , cyt b , and SSU-rRNA [5] . Additionally , we used hsp70-PCR and kDNA-PCR to amplify DNA samples of L . major , L . donovani , L . lainsoni , L . siamensis , and L . martiniquenensis . Unfortunately , negative PCR results were obtained due to lower sensitivities of PCR amplification using these genes . L . martiniquensis and L . siamensis were the predominant species detected in this study . Additionally , L . infantum ( which causes VL ) was previously reported in an HIV negative individual living in Bangkok [33] . The L . donovani complex species are the major causative agents of VL worldwide . The distribution of the complex species might have been introduced by travelers or workers from VL-endemic areas to Thailand [34] . L . major , a causative agent for CL in the Old World , causes zoonotic transmission , especially Afghanistan and India [31] . VL , caused by L . major , has occasionally been reported among patients with HIV [3] . L . donovani complex species and L . major have been reported in China , Bangladesh , India , and Nepal [15] . Thus , distribution of these two species in Thailand could be possible . L . lainsoni infection causes localized CL and was reported in South America in Bolivia , Peru , Suriname , French Guiana , and Brazil [15 , 35] . This is the first report of L . major and L . lainsoni infection among individuals with HIV in Thailand . In addition , no report of L . lainsoni has been documented in the Old World , especially among patients with HIV . To prevent and control VL , understanding disease epidemiology is extremely important . A cohort study conducted in this population as well as studies of potential vectors and animal reservoirs are needed . Moreover , large scale molecular epidemiological studies in other high morbidity areas are required for this emerging disease . | Visceral leishmaniasis ( VL ) in Thailand is caused by two causative agents , Leishmania martiniquensis and Leishmania siamensis . A public health concern brought us to investigate the magnitude of Leishmania infection among individuals with HIV living in an affected area , Trang province , southern Thailand . The results showed a high seroprevalence of Leishmania infection . Using PCR-based technique , DNA detection in the buffy coat was 8 . 4% , and 1 . 8% were symptomatic VL . Asymptomatic Leishmania infection could play an important role in disease transmission . Risk factors associated with Leishmania infection among Thai patients with HIV were firstly described . Those who were NIDU and lived in stilt houses were associated with Leishmania infection . Individuals who had lower immunity with detectable viral load were more likely to contract the infection . Interestingly , not only L . martiniquensis and L . siamensis were identified but also L . donovani complex , L . major , and L . lainsoni , were firstly reported among indigenous Thai people . These findings could lead to effective intervention and prevention methods to control leishmaniasis in Thailand . Further studies are needed to investigate the disease development in asymptomatic infections as well as evaluate the prevalence in other regions of Thailand . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"opportunistic",
"infections",
"parasitic",
"diseases",
"p... | 2017 | Prevalence and risk factors associated with Leishmania infection in Trang Province, southern Thailand |
In Brazil , schistosomiasis is a parasitic disease of public health relevance , mainly in poor areas where Schistosoma mansoni is the only human species encountered and Biomphalaria straminea is one of the intermediate host snails . A nested-PCR based on a specific mitochondrial S . mansoni minisatellite DNA region has been successfully developed and applied as a reference method in Brazil for S . mansoni detection , mainly in host snails for epidemiological studies . The amplification efficiency of LAMP is known to be higher than PCR . The present work aimed to assess the utility of our previously described SmMIT-LAMP assay for S . mansoni detection in human stool and snail samples in a low-transmission area of schistosomiasis in the municipality of Umbuzeiro , Paraíba State , Northeast Region of Brazil . A total of 427 human stool samples were collected during June-July 2016 in the municipality of Umbuzeiro and an overall prevalence of 3 . 04% ( 13/427 ) resulted positive by duplicate Kato-Katz thick smear . A total of 1 , 175 snails identified as Biomphalaria straminea were collected from 14 breeding sites along the Paraíba riverbank and distributed in 46 pools . DNA from human stool samples and pooled snails was extracted using the phenol/chloroform method . When performing the SmMIT-LAMP assay a total of 49/162 ( 30 . 24% ) stool samples resulted positive , including 12/13 ( 92 . 31% ) that were Kato-Katz positive and 37/149 ( 24 . 83% ) previously Kato-Katz negative . By nested-PCR , only 1/46 pooled DNA snail samples was positive . By SmMIT-LAMP assay , the same sample also resulted positive and an additional one was positive from a different breeding site . Data of human and snail surveys were used to build risk maps of schistosomiasis incidence using kernel density analysis . This is the first study in which a LAMP assay was evaluated in both human stool and snail samples from a low-transmission schistosomiasis-endemic area . Our SmMIT-LAMP proved to be much more efficient in detection of S . mansoni in comparison to the 'gold standard' Kato-Katz method in human stool samples and the reference molecular nested-PCR in snails . The SmMIT-LAMP has demonstrated to be a useful molecular tool to identify potential foci of transmission in order to build risk maps of schistosomiasis .
Schistosomiasis has been a public health problem in Brazil for decades . Around 1 . 8 million people , mostly in the coastal states of Brazil , are thought to be infected with Schistosoma mansoni and 25 million living at risk of contracting the disease in America [1] . Nineteen of the twenty-six federal states of Brazil are affected by the disease , especially in the northeastern region of the country . A schistosomiasis control program was implemented more than 40 years ago , decreasing prevalence , morbidity , and mortality over the past years [2] . Nevertheless , parasitological or immunological tests are not effective for detecting S . mansoni infection in low prevalence areas although polymerase chain reaction ( PCR ) -based diagnostic methods have been successfully developed and applied in endemic areas of schistosomiasis in Brazil [3–7] . They are not widely used in low-income countries due to the highly technical requirements and need for skilled personnel , making them unviable for routine application in field conditions . Snails of the genus Biomphalaria are well known for their role as intermediate hosts of S . mansoni which are able to produce hundreds or thousands of cercariae for months . Detection of cercarial shedding by infected snails after exposure of the specimens to light has traditionally been the method used to detect active sites for snail-to-human transmission [8] . This technique has several disadvantages: non-shedding of snail during the prepatent period , lack of experienced personnel for the identification of the acute infection , and difficulty in differentiating the morphology of the cercariae between trematodes species . To avoid these limitations , the detection of S . mansoni DNA in snail has been a good option offering greater sensitivity than classical methods with the advantage of detecting parasite of pooled snail samples . Therefore , several PCR-based assays have been developed to detect snails infected with S . mansoni [9 , 10] . In Brazil , a nested-PCR for monitoring S . mansoni-infected Biomphalaria spp . has been used as the most common technique to identify active foci of schistosomiasis transmission [11 , 12] . However , as noted , PCR-based techniques are difficult to apply in many endemic areas of schistosomiasis . Loop-mediated isothermal amplification ( LAMP ) technology [13] is a powerful tool to apply for point-of-care testing in resource-poor settings because it is a rapid single-step assay which does not require a thermal cycler and results can be visualized by naked eye; additionally , LAMP technology can be used quantitatively using real-time assays , thus potentially providing information about level of infection [14] . LAMP assays have been developed for molecular detection and diagnostics of several Neglected Tropical Diseases ( NTDs ) and applied mainly in those produced by protozoa as human African trypanosomiasis and leishmaniasis [15 , 16] . Additionally , LAMP assays have been successfully described for detecting NTDs produced by helminth parasites , including filariasis , soil-transmitted helminthiases and foodborne trematodiases [17–22] . Recently , several monitoring LAMP-based assays have also been developed for the detection of schistosomes [23–27] . In a previous work , a 620 bp sequence corresponding to a mitochondrial S . mansoni minisatellite DNA region was selected as a target for designing a LAMP-based method to detect S . mansoni DNA . This technique , called SmMIT-LAMP , was developed by our research group to detect S . mansoni DNA in stool samples from infected mice [28] . The specificity of this assay was determined against 16 DNA of different parasites -including several helminths and protozoa- and the limit of detection was established in 1 fg using S . mansoni DNA . With the aim to apply SmMIT-LAMP as a cost-effective molecular tool for the detection of S . mansoni in field applicable conditions , in this study we assess SmMIT-LAMP in human and snail samples collected in an endemic area of Brazil . Moreover , the results obtained by Kato-Katz analysis of human stool samples and nested-PCR performed in snails have been compared with the SmMIT-LAMP assay . It is the first time that a LAMP-based method has been used to identify transmission foci and to evaluate the epidemiological risk of acquiring schistosomiasis .
The study was approved by the Aggeu Magalhães Institute Ethics Committee ( protocol approval no . CAAE 56338916 . 6 . 0000 . 5190 ) . Volunteers were given detailed explanations about the aims , procedures and possible benefits of the study . Written informed consent was obtained from all subjects prior to the collection of biological samples for parasitological and molecular evaluation . Parents or guardians of children who participated in the study provided written informed consent on the child's behalf . All participants with confirmed infection received appropriate treatment . All samples were coded and treated anonymously . Procedures were performed in accordance with the ethical standards laid down in the Declaration of Helsinki as revised in 2013 . The study was conducted during June and July 2016 in the municipality of Umbuzeiro , located in the Agreste region of Paraíba State in the Northeast Region of Brazil ( Fig 1 ) . The municipality of Umbuzeiro covers an area of 181 , 327 km2 and has a population of 9 , 300 inhabitants ( 51 . 28 inhabitants/km2 ) with 3 , 986 and 5 , 314 people living in urban and rural areas , respectively , at 2010 census [29] . This location was chosen for the study because it is a known rural area with a low-endemicity for schistosomiasis and no records of mass treatment of the population within the last ten years . Moreover , this municipality is crossed by the Paraíba River , the only hydrographic basin in the region and the population has work and leisure activities centered on the river . The locality was georeferenced by means of the global positioning system ( GPS ) technology , using a GPS receiver ( Garmin , model eTrex ) with minimum accuracy of 10 meters , configured in the Universal Transverse Mercator ( UTM ) projection Datum SIRGAS 2000 . Using the TrackMaker Pro software , the GPS receiver data was transferred to a computer , making it possible to save files ( map , case distribution , breeding sites and foci ) in format used in the spatial data analysis , which was done by means of the ArcGis software 10 . 1 . The shapefiles of Brazil , Paraíba and Umbuzeiro , which is part of the base layers of the maps presented , as well as all the satellite images of Umbuzeiro in this work were generated using the sensor Sentinel 2 of the European Space Agency ( ESA ) ( https://sentinel . esa . int/web/sentinel/user-guides/sentinel-2-msi ) with Open Access CC-BY License ( http://open . esa . int/ ) . The satellites images were taken in June/2016 . Of the total of 427 human stool samples included in the study , only 162 samples ( collected in a zone at the northeast of the study area where Kato-Katz positive samples were detected ) were tested using the reaction mixture and specific primer set for LAMP assay -SmMIT-LAMP- previously established by Fernández-Soto et al . [28] . All pooled snail samples were also analyzed using the same LAMP assay . The SmMIT-LAMP method amplifies a specific sequence corresponding to a mitochondrial S . mansoni minisatellite DNA region ( GenBank Acc . No . L27240 ) . Briefly , the reaction was carried out in 25 μL reaction mixture containing 40 pmol each of FIP and BIP primers , 5 pmol of each F3 and B3 primers , 1 . 4 mM of each dNTP ( Intron ) , 1x Isothermal Amplification Buffer -20 mM Tris-HCl ( pH 8 . 8 ) , 50 mM KCl , 10 mM ( NH4 ) 2SO4 , 2 mM MgSO4 , 0 . 1% Tween20- ( New England Biolabs , UK ) , 1 M betaine ( Sigma , USA ) , supplementary 6 mM of MgSO4 ( New England Biolabs , UK ) and 8 U of Bst 2 . 0 WarmStart DNA polymerase ( New England Biolabs , UK ) with 2 μL of template DNA . Reaction tubes were placed in a heating block at a constant temperature of 63°C for 60 min and then heated at 80°C for 5 min to stop the reaction . In all SmMIT-LAMP trials positive ( S . mansoni DNA ) and negative ( water ) controls were always included . The LAMP-positive results could be visually inspected by naked eye by color change after adding 2 μL of 1:10 diluted 10 , 000x concentration fluorescent dye SYBR Green I to the reactions tubes . Green fluorescence was clearly observed in successful LAMP reaction , whereas it remained the original orange in the negative reaction . To avoid as much as possible the potential risk of cross-contamination with amplified products , all tubes were briefly centrifuged and carefully opened before adding the fluorescent dye . Data of human and snail surveys and results of parasitological and molecular analysis were used to build risk maps . Based on the number of snails collected in each station a thematic map demonstrating the abundance of snails in the breeding sites and foci of transmission was built . A kernel density analysis was also performed to draw a risk map of schistosomiasis incidence according to the diagnostic methods used for detection . Kernel Density Estimation ( KDE ) is a statistical technique of interpolation , nonparametric method , which produces a continuous surface ( cluster ) of the density calculated at all locations for visual identification of hotspots without changing their local characteristics [38 , 39] . The area unit was defined in m2 and the kernel spatial resolution in 10 meters . Statistical analyses were performed using GraphPad Prism software package ( version 6 , GraphPad Software , Inc . , San Diego , CA , USA; https://www . graphpad . com ) . Comparison of LAMP results with those obtained by microscopy were analyzed by McNemar's test for matched pairs . Comparisons were considered significant at a p-value < 0 . 05 . The diagnostic sensitivity , specificity , positive predictive value ( PPV ) and negative predictive value ( NPV ) were calculated for the SmMIT-LAMP and the Kato-Katz method using the MedCalc statistical program version 15 . 2 . 2 ( MedCalc Software , Ostende , Belgium ) according to the software instruction manual ( www . medcalc . org ) .
A total of 13/427 ( 3 . 04% ) human stool samples were positive by duplicate Kato-Katz thick smears , including samples obtained from 5 males and 8 females ( median age 45; range 14–90; SD 22 . 76 ) . In all Kato-Katz positive slides , the S . mansoni egg counts was very low , as well as the number of egg per gram of feces ( EPG ) with an average from 12 to 180 ( Table 1 ) . Up to 4/13 slides resulted negative in one of the two analyses . The spatial distribution of parasitological positive cases is represented in Fig 1A . Of the total of 427 participants in the study delivering their stool samples , all Kato-Katz positive cases were detected in a zone located at the northeast of the study area ( Fig 1B ) . In that zone , a total of 162 samples having tested previously collected in the parasitological survey , counting the 13 Kato-Katz positive samples and 149 Kato-Katz negative samples . These 162 samples were further subjected to molecular analysis by LAMP assay as described below . The SmMIT-LAMP overall results obtained after testing the stool samples in comparison to Kato-Katz results are showed in Fig 2 . When performing the SmMIT-LAMP assay a total of 49/162 ( 30 . 24% ) stool samples resulted positive , including 12 of the 13 ( 92 . 31% ) previously resulting Kato-Katz positive and , additionally , 37/149 ( 24 . 83% ) previously Kato-Katz negative . The SmMIT-LAMP results for human stool samples resulting Kato-Katz positive are shown in Fig 3 . Despite no child had been positive for S . mansoni by the Kato-Katz method , up to 12 children under 12 years old were positive for S . mansoni using the LAMP assay , which represent 24 . 5% of the total 49 positive cases detected by the molecular technique . MacNemar's test showed a statistically significant relationship between LAMP results and microscopy-detected S . mansoni infections ( p-value<0 . 0001 ) . Considering the microscopy findings by Kato-Katz as the reference standard , the following diagnostic parameters were calculated for the SmMIT-LAMP in this study: 92 . 86% sensitivity ( 95% CI: 66 . 13% -99 . 82% ) ; 80 . 11% specificity ( 95% CI: 73 . 64% -85 . 59% ) ; 26 . 00% positive predicted value ( 95% CI: 20 . 28% -32 . 67% ) and 99 . 33% negative predicted value ( 95% CI: 95 . 75% -99 . 90% ) . A total of 1 , 175 snails were collected with an average number of specimens per breeding site of 83 . 92 ( range 4–370; SD: 109 . 44 ) . All snails were identified as Biomphalaria straminea . None of the snails examined by exposure to artificial light for cercariae to emerge was infected . Testing the 46 pooled snail DNA samples by nested-PCR , only one resulted positive ( no . 45 ) . By SmMIT-LAMP assay , the same pooled snail sample also resulted positive and another pool ( no . 15 ) was positive from a different breeding site . The SmMIT-LAMP and nested-PCR results are shown in Fig 4 and the geographical distribution of breeding sites where pooled snails samples resulted positive by molecular assays is shown in Fig 5 . In these two points , the abundance of snails was the largest of the survey . Additionally , the two-pooled snail positive samples were located in the same area where the highest number of patients with microscopy-positive results were detected . Thus , two potential foci of schistosomiasis transmission were identified in the study area . The distribution of positive cases by both microscopy and SmMIT-LAMP in households included in the study , as well as risk maps of schistosomiasis infection generated with the Kernel density method are shown in Fig 6 . Only one positive result per household was detected when using the Kato-Katz technique ( Fig 5A ) , whereas up to five positive results per household could be obtained when using the SmMIT-LAMP assay ( Fig 5B ) . The maps of potential risk areas of schistosomiasis transmission based on this study , either by Kato-Katz tests ( Fig 5C ) , and SmMIT-LAMP ( Fig 5D ) are shown . Two foci of schistosomiasis transmission were located at the breeding sites where pooled snail samples resulted positive by molecular assays .
Our study was conducted in a known low prevalence area of schistosomiasis in Brazil . Kato-Katz results obtained in the population survey corroborated previous results published from the study area [40] . Only thirteen stool samples ( 3 . 04% ) resulted positive by microscopic analysis , including up to ten samples with light infections ( 1–99 EPG ) , where four of those thirteen did not show the presence of S . mansoni eggs in one of the two slides examined . These data are in line with the known low sensitivity of the Kato-Katz technique for diagnosing schistosomiasis in areas of low prevalence and parasite load [4 , 41] . Molecular assays arise as a potential alternative to traditional parasitological methods in situations were highly sensitive diagnostic test are needed [42] . In this context , we tried to evaluate our SmMIT-LAMP assay in an area where the Kato-Katz previously showed low sensitivity . When performing the SmMIT-LAMP , the number of positive samples detected increased up to 49 , with an overall prevalence of 30 . 24% . Moreover , from the number of samples positive by Kato-Katz , up to 92 . 31% were LAMP-positive , thus indicating a higher performance of the technique . Thus , our SmMIT-LAMP seems to be much more sensitive than microscopic detection of eggs , commonly used as the classical reference test for intestinal schistosomiasis [28] . We assume that LAMP positive results are S . mansoni-infected patients considering the known low sensitivity of the microscopy and the high sensitivity and specificity of the molecular assay . In a previous work of our group , the SmMIT-LAMP was developed and the specificity of the assay was tested against more than 20 DNA of different parasites , including helminths and several protozoa . The specificity of the mitochondrial DNA sequence used for primer design was also confirmed using different bionformatic tools [28] . Additionally , negative controls were always included in all LAMP assays in order to detect potential contaminations with S . mansoni DNA . If any negative control used in the trials showed DNA amplification , the result was invalidated and the reaction was discarded and repeated . Thus , we are confident that the LAMP-positive results could only be due to the presence of S . mansoni DNA in the stool samples . Other studies evaluating molecular assays versus microscopy have reported a similar increment of positive cases [42] , thus representing a better performance of the molecular assays in comparison with parasitological methods . One KK-positive sample was LAMP-negative , resulting in a sensitivity of 92 . 86% . This sample presented absence of S . mansoni eggs in one of the two slides microscopically examined . This sensitivity is in accordance with other molecular assays for the diagnosis of S . mansoni [42] , so the percentage of false negative is entirely aceptable . Furthermore , S . mansoni DNA was extracted using a cost-effective phenol/cloroform method without compromising the sensitivity of the LAMP analysis . In this context , a number of LAMP assays have been previously reported for parasites detection with a minimal or no DNA extraction requirement , including schistosomes [25 , 43 , 44] . Moreover , the SmMIT-LAMP results are easily visualized by color change by naked eye . This is a great advantage for epidemiological surveys in low-income areas compared to other DNA-based molecular methods . Biomphalaria straminea was the sole species identified as intermediate host in the study area . This finding is in accordance with a previous malacological survey in this region [45] . Among the three species of host snails for S . mansoni in Brazil , B . glabrata is considered to be the most epidemiologically important species since its geographical distribution overlaps with the distribution of schistosomiasis in Brasil . However , B . straminea is better adapted to all climatic variations and ecological conditions in Brazil [46] . Besides , B . straminea is considered more resistant to S . mansoni infection than other snail species [47] . Different epidemiological studies have demonstrated the utility of the nested-PCR method for S . mansoni detection in pooled B . straminea samples when the parasitological assays are not effective [11 , 45] . In our work , one pooled sample of B . straminea was detected using this nested-PCR assay as reference , although no one was positive by classical cercarial shedding tests . In order to test the SmMIT-LAMP in detecting S . mansoni in B . straminea , we analyzed the pooled snail samples and compared results with nested-PCR assay . The SmMIT-LAMP assay was originally designing on a sequence corresponding to a specific mitochondrial S . mansoni minisatellite DNA region [48] . This sequence was also previously used for designed a specific PCR-based method for detection of S . mansoni with no cross-reaction with other Brazilian trematodes which have snails of genus Biomphalaria as intermediate hosts [10] . When applying the two molecular methods for snail samples screening , two breeding sites wereidentified as potential foci of schistosomiasis transmission , one detected by both nested-PCR and SmMIT-LAMP and an additional one by SmMIT-LAMP . These two foci of transmission are located in the same study area where the Kato-Katz positive human cases were detected . In recent years , it has been reported the use of LAMP in large-scale screening of pooled field-collected snails for analyzing the transmission of schistosomiasis , as a simple and efficient tool for snails’ surveillance , including S . mansoni in Brazil [49 , 27] . In our study , the SmMIT-LAMP assay was applied for the first time to evaluate the S . mansoni infection not only in pooled field-collected snails but also in human stool samples . Data obtained in both SmMIT-LAMP and Kato-Katz tests were used to create Kernel density-based maps of risk of schistosomiasis . The Kernel density has previously been used to build maps of risk for several helminthiases , including schistosomiasis [50–52] . The risk areas obtained mapped close to the snail breeding sites identified as foci of schistosomiasis transmission by the SmMIT-LAMP and nested-PCR . In those breeding sites the highest contact between the population and the river was observed for work activities ( extraction of sand from the river ) , domestic activities ( washing clothes and dishes ) , and leisure activities ( fishing and children's recreation ) . All these activities are known to be associated with transmission of schistosomiasis [53 , 54] . In addition , this area is commonly used by the inhabitants as a route for crossing the river further increasing the risk of infection . In summary , this is the first study in which a LAMP assay was evaluated in both human stool and snail samples from a low-transmission schistosomiasis-endemic area . Our SmMIT-LAMP proved to be much more efficient in detecting S . mansoni in comparison to the 'gold standard' method ( Kato-Katz ) in human stool samples and the reference molecular nested-PCR in snails . Moreover , SmMIT-LAMP has demonstrated to be a useful molecular tool to identify foci of transmission in order to build risk maps of schistosomiasis . | In Brazil , around 1 . 8 million people , mostly in the northeastern region of the country , are thought to be infected with Schistosoma mansoni . Snails of the genus Biomphalaria serve as intermediate hosts of the S . mansoni . A special program for schistosomiasis control was implemented more than 40 years ago in Brazil , decreasing prevalence , morbidity , and mortality over the past years . PCR-based diagnostic methods have been successfully applied in a few endemic areas of schistosomiasis in Brazil , although they are not still widely used due to the highly technical requirements making them unviable for routine application in field conditions . Loop-mediated isothermal amplification ( LAMP ) technology could be a powerful tool to apply for point-of-care testing in resource-poor settings . In previous work , a LAMP-based method to detect S . mansoni DNA , called SmMIT-LAMP , was developed by our research group to detect S . mansoni DNA testing stool samples from experimentally infected mice . Here , with the aim to apply SmMIT-LAMP as a cost-effective molecular tool for the detection of S . mansoni in field applicable conditions , we assess SmMIT-LAMP in human and snail samples collected in an endemic area of Brazil . The results obtained by Kato-Katz analysis of human stool samples and nested-PCR performed in snails were compared with the SmMIT-LAMP assay . It is the first time that a LAMP-based method has been used to identify potential transmission foci and to evaluate the epidemiological risk of acquiring schistosomiasis . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"biomphalaria",
"helminths",
"malacology",
"tropical",
"diseases",
"geographical",
"locations",
"parasitic",
"diseases",
"animals",
"parasitology",
"gastropods",
"neglected",
... | 2018 | A field survey using LAMP assay for detection of Schistosoma mansoni in a low-transmission area of schistosomiasis in Umbuzeiro, Brazil: Assessment in human and snail samples |
Polymerase basic protein 1 ( PB1 ) is the catalytic core of the influenza A virus ( IAV ) RNA polymerase complex essential for viral transcription and replication . Understanding the intrinsic mechanisms which block PB1 function could stimulate development of new anti-influenza therapeutics . Affinity purification coupled with mass spectrometry ( AP-MS ) was used to identify host factors interacting with PB1 . Among PB1 interactors , the E3 ubiquitin ligase TRIM32 interacts with PB1 proteins derived from multiple IAV strains . TRIM32 senses IAV infection by interacting with PB1 and translocates with PB1 to the nucleus following influenza infection . Ectopic TRIM32 expression attenuates IAV infection . Conversely , RNAi depletion and knockout of TRIM32 increase susceptibility of tracheal and lung epithelial cells to IAV infection . Reconstitution of trim32-/- mouse embryonic fibroblasts with TRIM32 , but not a catalytically inactive mutant , restores viral restriction . Furthermore , TRIM32 directly ubiquitinates PB1 , leading to PB1 protein degradation and subsequent reduction of polymerase activity . Thus , TRIM32 is an intrinsic IAV restriction factor which senses and targets the PB1 polymerase for ubiquitination and protein degradation . TRIM32 represents a model of intrinsic immunity , in which a host protein directly senses and counters viral infection in a species specific fashion by directly limiting viral replication .
Influenza A virus A ( IAV ) is a human respiratory pathogen that causes seasonal epidemics and occasional global pandemics with devastating levels of morbidity and mortality . IAV is a member of the Orthomyxoviridae family and possesses eight segments of negative-sense single-stranded RNA genome . Replication and transcription of these IAV segments is catalyzed by a heterotrimeric RNA-dependent RNA polymerase complex , which consists of an acidic subunit ( PA ) and two basic subunits , PB1 and PB2 [1 , 2] . PB1 is the structural backbone for formation of the IAV polymerase complex [1] . PB1 contains a 14 residue binding site for PA at the N-terminus and a C-terminal domain for PB2 association [3–6] . Since the activity of RNA-dependent polymerases is distinct from enzymes found in host cells , these viral proteins are promising drug targets for interfering with viral replication [7 , 8] . Little is understood about the natural defenses employed by host cells to defend against the IAV polymerase . In this report , we analyze PB1 protein complexes and find a host interactor , tripartite motif-containing protein 32 ( TRIM32 ) , which directly targets PB1 proteins to restrict influenza virus replication . TRIM32 was initially identified as a protein that binds HIV-1 tat ( a key transactivator of viral transcription ) [9 , 10] . TRIM32 contains an N-terminal signature tripartite motif ( TRIM ) consisting of RING , B-box and coiled-coil domains followed by a spacer segment and a series of NHL repeats . The presence of the RING domain is a sign that TRIM family proteins may function as ubiquitin E3 ligases , catalyzing transfer of ubiquitin from an E2 enzyme to form a covalent bond with a substrate lysine . Genetic mutation in the TRIM32 NHL domains causes recessive hereditary muscle disorders , often with a neurogenic component , including limb girdle muscular dystrophy 2H and sarcotubular myopathy [11–14] . These conditions are phenocopied in knockout mice that lack trim32 [15 , 16] and knockin animals that carry a disease associated TRIM32 mutation [17] . In addition , mutations in the TRIM32 B-box domain are responsible for Bardet Biedl syndrome , which has a pleiotropic phenotype often accompanied with retinal degeneration [18 , 19] . TRIM32 is a ubiquitously expressed E3 ligase , which targets several proteins for ubiquitination , including actin [20] , PIASγ [21] , Abl-interactor 2 [22] , c-Myc [23] , PKCζ [24] , dysbindin [25] , X-linked inhibitor of apoptosis ( XIAP ) [26] , desmin filaments [27] , p73 transcription factor [28] , STING [29] and Gli-related Krüppel-like zinc finger protein ( Glis2 ) [30] . Based on this broad substrate specificity , it is not surprising that TRIM32 has versatile activities and is linked to diverse biological processes , including innate immunity [31 , 32] , development and differentiation [15 , 16 , 23 , 33 , 34] , regulation of microRNA [23 , 35] , and tumorigenesis [36 , 37] . However , the role of TRIM32 in intrinsic immunity and viral restriction remains enigmatic . This report characterizes a role for TRIM32 in intrinsic cellular defense against influenza viruses by targeting the influenza polymerase for ubiquitination and degradation .
Mass spectrometry was used to examine the physical interactions between IAV PB1 and endogenous cellular proteins . PB1 protein complexes were immunoaffinity purified from HEK293 cells stably expressing FLAG tagged PB1 derived from the influenza A/Puerto Rico/8/1934 ( PR8 ) . Two independent purifications were analyzed by LC/MS-MS analysis . Controls include our laboratory database of 200 FLAG-tagged non-viral proteins isolated by identical procedures from stably transfected HEK293 cell lines [32 , 38 , 39] . A well-established computational algorithm , known as SAINT , was applied to the dataset [40 , 41] . Twenty-six proteins had SAINT scores above 0 . 89 and were designated high confidence interacting proteins ( HCIP ) , including 18 proteins ( S1 Table ) reported to associate with PB1 or IAV polymerase complexes [42–44] . In a preliminary screen , 6 available GFP- or FLAG-tagged HCIP ( TRIM32 , STUB1 , IQSEC , GALK , PDCD6 and HAUS6/DGT6 ) were expressed in HEK293 cells and examined for their effects on viral replication using the PR8-Gaussia luciferase reporter virus [45] . The most potent inhibition of IAV replication was noted following ectopic expression of TRIM32 ( S1A Fig ) . Interactions between TRIM32 and IAV proteins had not been previously reported . Thus , we first validated the protein interaction . Following PR8 IAV infection of either primary human tracheal epithelial cells or HEK293 cells , endogenous TRIM32 binds to viral PB1 ( Figs 1A and S1B ) . To demonstrate this interaction is direct , TRIM32-HIS and PB1-GST were purified from bacteria . In vitro GST pull down assays confirmed the direct association of PB1 with TRIM32 ( Fig 1B ) . We next examined if TRIM32 can associate with PB1 proteins derived from different IAV strains . PB1 proteins from 6 IAV strains [PR8 , A/Puerto Rico/8/1934 ( H1N1 ) ; WSN , A/WSN/1933 ( H1N1 ) ; NY , A/New York/1682/2009 ( H1N1 ) ; Aichi , A/Aichi/2/1968 ( H3N2 ) ; A/Vietnam/1194/2004 ( H5N1 ) and A/Anhui/1/2013 ( H7N9 ) ] were co-transfected with TRIM32-GFP into HEK293 cells . Co-precipitation was noted between TRIM32 and PB1 from all 6 IAV strains ( Fig 1C ) . Furthermore , the IAV ribonucleoprotein components PA , PB2 and NP failed to interact with TRIM32 , demonstrating the specificity of PB1-TRIM32 interaction ( Figs 1D and S1C ) . Taken together , TRIM32 physically interacts with PB1 polymerases derived from multiple IAV strains . As influenza viruses are intrinsically sensitive to the antiviral action of interferons ( IFN ) , we speculated that IFN might regulate TRIM32 expression . To address this issue , we infected A549 lung epithelial cells with PR8 IAV or treated the cells with IFN . TRIM32 mRNA and protein expression were then examined . However , there is no evidence that IAV infection or IFN stimulation modulate TRIM32 RNA or protein levels ( S1D and S1E Fig ) . We conclude TRIM32 is constitutively expressed in A549 human lung epithelial cells , a model host cell line for IAV infection . To examine TRIM32-PB1 co-localization , we infected primary human tracheal and A549 lung epithelial cells with PR8 IAV . In the absence of viral infection , endogenous TRIM32 is diffusely expressed in cytosolic foci and lesser amounts are detected in the nucleus . Following IAV infection TRIM32 accumulates in the nucleus ( Figs 1E and S2A ) . Similarly , nuclear accumulation of endogenous TRIM32 is noted in A549 cells stably transfected with PB1 ( S2B Fig ) . Influenza A driven TRIM32 nuclear translocation was biochemically confirmed by isolation of nuclear and cytosolic fractions ( Fig 1F ) , indicating that IAV infection triggers translocation of TRIM32 to the nucleus within 4 hr . The evolutionarily conserved CRM1 ( exportin 1 ) receptor is responsible for nuclear export of most proteins [46] . To examine the role of CRM1 in TRIM32 nuclear distribution , A549 cells were treated with CRM1 inhibitor leptomycin B . Addition of leptomycin B causes accumulation of nuclear TRIM32 in the absence of viral infection ( S2C Fig ) , thereby implying that TRIM32 physiologically shuttles between the cytosol and nucleus . The combined data suggest that during IAV replication PB1 retains TRIM32 in the nuclear compartment . To determine which segment of TRIM32 mediates its interaction with PB1 , we generated a series of TRIM32 deletion mutants ( Fig 2A ) . Initial experiments suggested TRIM32 constructs carrying residues 140 to 265 were involved in PB1 binding ( Fig 2B ) . The 140 to 265 segment contains the TRIM32 CC domain plus part of the linker region . Deletion of this segment results in the loss of PB1 binding ( Fig 2C ) , while this segment alone is sufficient for PB1 association ( Fig 2D ) . Thus , TRIM32 residues 140 to 265 are necessary and sufficient for PB1 binding . Assembly of the heterotrimeric IAV polymerase requires interactions between the PB1 N-terminal domain with the PA chain while the PB1 C-terminus associates with the polymerase PB2 protein [6 , 47 , 48] . PB1 truncation mutants were prepared in order to define the PB1 segment which interacts with TRIM32 ( Fig 2E ) . The PB1 C-terminal region consisting of residues 493 to 757 was sufficient for interaction with TRIM32 ( Fig 2F ) and the TRIM32 CC-linker fragment ( Fig 2G ) . These findings led us to question whether TRIM32 could compete with PB2 for PB1 binding . The failure of TRIM32 to block PB1-PB2 association suggests TRIM32 is not a competitive inhibitor of IAV polymerase assembly ( S2D Fig ) . Fine mapping of IAV polymerase interaction sites suggest a stretch of 15 amino acids in the PB1 C-terminus are most important for PB2 binding , leaving an expansive “thumb” surface available for interactions with other proteins [1 , 2] . The combined data identify the essential peptide fragments for TRIM32-PB1 interaction . Four assay systems ( reporter assay , Western blot , immunofluorescence and plaque assay ) were used to evaluate the biological effect of TRIM32 overexpression on IAV replication . First , FLAG-tagged TRIM32 and TRIM65 ( a control TRIM family member with established E3 ligase activity [39 , 49] ) were transfected into A549 cells followed by infection with PR8-Gaussia luciferase reporter virus . TRIM32 selectively restricted IAV replication ( Fig 3A ) without significantly impacting cell viability ( S3A Fig ) . We then determined the effect of TRIM32 on viral infection by examining IAV NP protein expression using Western blot and immunofluorescence . A549 lung epithelial cells stably transfected with TRIM32 were infected with WSN and PR8 IAV . TRIM32 overexpression resulted in viral restriction as noted by decreased levels of NP protein detected by Western blot ( Figs 3B and S3B ) . Similarly , HEK293 cells transfected with FLAG-TRIM32 or GFP-TRIM32 displayed decreased levels of viral protein production ( S3C and S3D Fig ) . Microscopic inspection confirmed that TRIM32 overexpression inhibited PR8 and WSN viral NP protein expression ( Figs 3C and S3E ) . Plaque assays were used to determine the effects of TRIM32 on production of infectious IAV particles . Stable or transient overexpression of TRIM32 consistently reduced viral titers ( Figs 3D and S3F ) . To compare the relative susceptibility of various IAV strains to TRIM32-mediated restriction , A549 stable cell lines transfected with control vector or TRIM32-FLAG were infected with 0 . 001 MOI of 4 different IAV strains . After 18 hr , supernatants were transferred to fresh A549 target cells and the relative fraction of infected cells was determined . TRIM32 transfection results in 70–80% reduction of infectious IAV particles ( S3G Fig ) . Thus , the combined findings indicate TRIM32 is an IAV restriction factor . To complement the above overexpression data , we depleted TRIM32 with siRNA . Decreased TRIM32 expression correlated with increased IAV reporter activity in primary respiratory epithelial cells ( Fig 4A ) and also in HEK293 cells ( S4A Fig ) . In addition , silencing TRIM32 enhanced IAV propagation in A549 cells , as detected by plaque assay and increased NP levels ( Figs 4B , S4B and S4C ) . Knockdown of TRIM32 also enhanced IAV propagation in primary tracheal epithelial cells , as detected by plaque assay ( Fig 4C ) . To exclude off-target effects of siRNA , cells were transfected with wild type TRIM32 or a siRNA resistant TRIM32 rescue construct before infection with PR8 reporter virus . Cells transfected with wild type or rescue TRIM32 constructs displayed comparable levels of TRIM32 expression and both decreased viral replication as demonstrated by reduced reporter activity ( Fig 4D ) . Importantly , when combined with TRIM32 siRNA the rescue construct restored antiviral activity , validating siRNA specificity ( Fig 4D ) . We next examined IAV infection in TRIM32 deficient mouse embryonic fibroblasts ( MEF ) . A 7 fold increase in PR8 luciferase reporter activity was noted in trim32-/- MEF compared to wild type controls and antiviral activity was partially rescued by transfection with human TRIM32 ( Fig 5A ) . In line with these results approximately 10-fold enhancement of viral NP protein expression was observed after PR8 infection of trim32-/- MEF ( Fig 5B ) and significantly more trim32-/- cells were stained with anti-NP antibody ( Fig 5C ) . We also evaluated the impact of TRIM32 on viral propagation using plaque assays . As predicted , trim32 deficient cells produce more infectious viral particles than control MEF ( Fig 5D ) . Finally , TRIM32 deficient and control MEF were infected with 0 . 1 MOI of 4 different IAV strains . Supernatants were transferred to A549 target cells and the relative fraction of infected cells was determined . Trim32-/- cells consistently produced 3 to 6 fold more infectious virus than control cells ( S4D Fig ) . To establish virus restriction specificity , MEF were infected with Sendai virus . TRIM32 deficiency did not alter the level of infection with wild type Sendai or a Sendai-luciferase reporter virus ( Fig 5E ) . The combined data using trim32 deficient cells indicate TRIM32 is an antiviral cellular factor that acts to curb infections with influenza viruses . TRIM32 was reported to regulate NF-κB and IFN driven cytokine production , which may mediate the antiviral activity of TRIM32 [29 , 32 , 50–52] . Thus , we investigated the capacity of trim32+/+ and trim32-/- MEF to produce IFNβ after infection with PR8 IAV ( ΔNS1 ) virus [53] or treatment with the synthetic analog of dsRNA , poly ( I:C ) . Trim32+/+ and trim32-/- MEF show comparable levels of IFNβ mRNA after IAV infection or poly ( I:C ) stimulation ( S5A and S5B Fig ) . Furthermore , similar levels of IAV reporter activity were noted in MEF derived from wild-type and NF-κB deficient p65-/- mice ( S5C Fig ) . The results suggest that TRIM32 does not alter influenza-activated IFN production and indicate other mechanisms are needed to account for TRIM32-mediated IAV restriction . RING domains are critical for E3 ligase function in TRIM family molecules . Thus , we evaluated the role of the TRIM32 RING domain in antiviral activity . The TRIM32 ( C39S ) mutation disrupts the rigid cross-braced architecture of the RING domain and destroys E3 ligase activity [29] . Antiviral activity was measured after transfection of HEK293 cells with TRIM32 or the TRIM32 ( C39S ) mutant . Wild type TRIM32 attenuates infection with the IAV reporter virus . In contrast , ectopic expression of the TRIM32 ( C39S ) mutant protein failed to defend against IAV infection ( Fig 6A ) . Reconstitution of trim32-/- MEF with human TRIM32 restores host restriction of PR8 reporter virus , while reconstitution with the TRIM32 ( C39S ) mutant retains high viral propagation ( Fig 6B ) . Similarly , the TRIM32 ( C39S ) mutant failed to restore antiviral activity in trim32-/- MEF as assayed by Western blotting and immunofluorescence for viral NP proteins ( Fig 6C and 6D ) . Thus , we conclude that E3 ligase activity is critical for the antiviral activity of TRIM32 . To determine whether TRIM32 can directly couple ubiquitin onto PB1 , bacterially derived TRIM32 and PB1 were combined in an in vitro ubiquitination assay . TRIM32 is able to heavily conjugate ubiquitin onto PB1 in vitro ( Fig 7A ) . PB1 ubiquitination is dependent on the presence of ubiquitin plus an E1 , E2 , and an ATP regenerating system . The TRIM32 ( C39S ) RING mutant ( RM ) fails to conjugate ubiquitin onto PB1 ( Fig 7A ) . To examine the role of TRIM32 on PB1 ubiquitination within cells , trim32-/- MEF stably transfected with either TRIM32 or TRIM32 ( C39S ) were infected with PR8 IAV . After treatment with the proteasomal inhibitor , MG132 , PB1 was immunoprecipitated . Ubiquitinated PB1 showing a characteristic high molecular weight ladder was noted from trim32+/+ MEF or trim32-/- MEF reconstituted with TRIM32 ( Fig 7B ) . Ubiquitinated PB1 was not detected in trim32-/- MEF or trim32-/- cells reconstituted with the C39S mutant . The requirement for MG132 treatment to visualize PB1 ubiquitination suggests that PB1 undergoes degradation in the presence of TRIM32 . Proteins modified with K48-linked polyubiquitin are classic targets for proteasomal degradation . To test this possibility , HEK293 cells were cotransfected with PB1 plus wild type ubiquitin or a ubiquitin construct in which all lysine residues except K48 were mutated to arginine ( K48 ) and a ubiquitin construct in which only the K48 residue was mutated to arginine ( K48R ) . After MG132 treatment wild type and K48-only ubiquitin were conjugated onto PB1 , while ubiquitin molecules lacking the K48 residue ( K48R ) were not coupled onto PB1 ( Fig 7C ) . Co-transfection with TRIM32 increased the levels of PB1 ubiquitination ( Fig 7C ) . To examine the role of TRIM32 in PB1 protein turnover , GFP-TRIM32 and GFP-TRIM32 ( C39S ) mutant constructs were transfected into A549 cells stably expressing FLAG-PB1 . Ectopic expression of TRIM32 decreased PB1 expression , while TRIM32 ( C39S ) showed little or no effect on PB1 expression ( Fig 7D ) . To further examine the role of TRIM32 in PB1 degradation , wild type MEF , trim32-/- MEF and reconstituted trim32-/- MEF were treated with a protein synthesis inhibitor ( cycloheximide ) . After 2 or 6 hr , cell lysates were examined by quantitative Western blotting . The data indicate the presence of TRIM32 has a destabilizing effect on PB1 expression ( Fig 7E ) . Finally , we examined whether TRIM32-dependent PB1 ubiquitination had an effect on PB1 polymerase activity . TRIM32 transfection inhibits IAV polymerase activity ( Fig 7F ) . As predicted , silencing TRIM32 with siRNA enhances polymerase activity , while addition of a RNAi resistant TRIM32 rescue construct reduced polymerase activity ( Fig 7F ) . The combined data indicate that TRIM32 mediates K48-linked ubiquitination of PB1 resulting in augmented PB1 degradation and reduced viral polymerase activity .
Humans evolved a broad spectrum of defense strategies to limit IAV infection . Among them , the best characterized antiviral defense systems are the broadly acting TLR and RLR pattern-recognition receptors which detect microbial nucleic acids and other cross-species biomarkers . These innate antiviral defenses indirectly inhibit infection by triggering signaling cascades that lead to the production of interferons and other antiviral effector molecules . In contrast , as demonstrated in this study , TRIM32 operates in a more restricted fashion to directly sense an IAV “danger” signal and to restrict infection in a species specific manner . Thus , TRIM32 provides a form of defense often categorized as intrinsic immunity . Examination of IAV polymerase interacting proteins by AP-MS identified TRIM32 . Associations of TRIM32 with the PB1 subunit of the polymerase complex were noted using 6 distinct PB1 proteins derived from IAV strains of H1N1 , H3N2 , H5N1 and H7N9 origin , suggesting that PB1 has not yet adapted to avoid TRIM32 targeting . Preliminary data suggest TRIM32 conjugates polyubiquitin at multiple PB1 sites , thereby limiting the opportunity for PB1 mutants to avoid targeting by TRIM32 . The functional activity of TRIM32 was demonstrated by a combination of overexpression , mutagenesis and loss-of-function ( RNAi and genetic deletion ) analyses . Comparison between wild type TRIM32 and the C39S ligase defective mutant demonstrated the requirement for TRIM32 E3 ligase activity in antiviral activity . The efficacy of TRIM32 in restricting IAV infection was established with four IAV strains and multiple cell types , including a human lung epithelial cell line and primary human tracheal epithelial cells . The latter represent a model for host cells which are naturally targeted during IAV infection . TRIM32 is expressed in all tissues and displays broad substrate specificity . TRIM32 is known to bind at least 10 different cellular proteins of highly diverse function and localization , most of these interaction partners are also known substrates of TRIM32-mediated ubiquitination [22 , 23 , 25–29] . TRIM32 is a remarkably versatile E3 ligase; it participates in monoubiquitination [20 , 25] and formation of K48-linked or K-63-linked polyubiquitin chains which are covalently conjugated to target proteins [29 , 30] . In addition , TRIM32 can form unanchored polyubiquitin chains [54] . Here we report that in response to influenza infection , TRIM32 targets PB1 for K48-linked ubiquitination . The K48 polyubiquitin tagged PB1 molecules are shuttled to the proteasome system which orchestrates turnover of target proteins . Depending on context , some cell types predominantly express TRIM32 proteins in the nuclear compartment , while in other cell types it primarily resides in a cytoplasmic niche [23 , 24 , 27 , 34 , 55 , 56] . However , TRIM32 nuclear import and export sequences remain undefined . Leptomycin B treatment causes TRIM32 nuclear accumulation suggesting that TRIM32 shuttles between the cytosolic and nuclear compartments . TRIM32 trafficking was previously studied in neural stem cells , where TRIM32 autoubiquitination helps maintain cytoplasmic localization [24] . In another study , ubiquitin conjugating E2 enzymes controlled TRIM32 localization in nuclear or cytoplasmic compartments [57] . The current study suggests an additional mechanism capable of rapidly regulating TRIM32 localization . We propose that the increasing nuclear concentration of PB1 during IAV replication combined with the high affinity of TRIM32 for this substrate affords a mechanism to trap and accumulate TRIM32 in the nucleus . It would be interesting to learn if this trapping hypothesis also applies to other settings where TRIM32 accumulates in the nucleus . Many distinct types of proteins can regulate antiviral responses and they can act at different stages of viral replication [31 , 52] . Among the most studied innate or intrinsic antiviral factors are members of the TRIM family . In fact TRIM32 was among the first TRIMs shown to specifically interact with a viral protein , tat [9] . Although TRIM32 binds to the activation domain of lentiviral tat proteins , the role of TRIM32 in viral restriction was not studied . In a separate report , TRIM32 was shown to target STING ( stimulator of interferon genes ) for ubiquitination [29] . TRIM32 also stimulates NF-κB activity [21] . Both of these pathways can lead to potentiation of type I IFN and other antiviral genes associated with innate immunity . However , genetic deficiency of TRIM32 does not significantly alter the levels of IFN production activated by flu infection . Many TRIM members require interferon to achieve expression and function [50] . In contrast , TRIM32 is constitutively expressed in respiratory epithelial cells where it is positioned to provide early and direct defense against IAV infection . In summary , we demonstrate TRIM32 inhibits IAV polymerase activity , targets PB1 for proteasomal degradation and provides intrinsic cellular restriction against IAV infection . Exploitation of this natural defense pathway may offer potential strategies for controlling IAV infections . Currently approved treatments against influenza are losing effectiveness as new viral strains are often refractory to conventional treatments . Understanding the molecular mechanisms controlling intrinsic antiviral defenses may illuminate novel strategies for preventing and treating viral diseases .
Primary human bronchial tracheal epithelial cells and supporting medium were purchased from Lifeline Cell Technology ( Frederick , MD ) . HEK293 and A549 cells ( American Type Culture Collection , Manassas , VA ) were cultured in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) in 10% CO2 at 37°C . P65+/+ and p65-/- MEF were generously provided by Denis Guttridge ( Ohio State University , Columbus , OH ) . Primary MEF derived from trim32+/+ and trim32-/- mice [16] were transfected with SV40 large T antigen for immortalization . MEF were cultured in DMEM supplemented with 10% FBS . For generation of stable cell lines , human TRIM32 or mutant expression constructs were transfected into HEK293 cells , A549 cells and trim32-/- MEFs cells using Lipofectamine 2000 ( Invitrogen , Carlsbad , CA ) as detailed elsewhere [32] . Two days after transfection , cells were treated with 200 μg/ml hygromycin for 14 days . Single colonies were picked and expanded in 6-well plates . Protein expression levels in each colony were determined by immunoblotting . Influenza A viruses used in this study: A/Puerto Rico/8/34 ( H1N1 ) ( Charles River Labs , Wilmington , MA ) , A/PR8ΔNS1 ( generously provided by Adolfo Garcia-Sastre , Mt . Sinai School of Medicine , NY ) [53] , A/WSN/33 ( H1N1 ) ( kindly provided by Peter Palese , Mt . Sinai School of Medicine , NY ) , A/New York/18/2009 ( H1N1 ) ( BEI Resources , Manassas , VA ) , and A/Aichi/68 ( H3N2 ) ( Charles River Labs ) . Influenza A PR8-GLuc virus containing a Gaussia luciferase ( GLuc ) gene inserted downstream of PB2 was a generous gift from Peter Palese [45] . Sendai virus was purchased from Charles River Labs . The Sendai-luciferase reporter virus was kindly provided by Charles Russell ( St . Jude Hospital , Memphis , TN ) [58] . Mass spectrometry experiments were performed as previously described [38 , 39 , 59] . For protein purification HEK293 stable cell lines expressing FLAG-PB1 were collected from five 15 cm dishes in 10 ml TAP buffer ( 50 mM Tris-HCl [pH 7 . 5] , 10 mM MgCl2 , 100 mM NaCl , 0 . 5% Nonidet P40 , 10% glycerol , phosphatase inhibitors and protease inhibitors ) [32] . Cell lysates were precleared with 50 μl protein A/G resin before addition of 20 μl anti-FLAG resin ( Sigma ) and 16 hr incubation at 4°C on a rotator . The resin was 3X washed and transferred to a spin column ( Sigma ) with 40 μl 3X FLAG peptide ( Sigma ) for 1 hr at 4°C on a rotator . Purified complexes were loaded onto a 4–15% NuPAGE gel ( Invitrogen ) . Gels were stained using the SilverQuest staining kit ( Invitrogen ) and lanes were excised for mass spectrometry analysis by the Taplin Biological Mass Spectrometry Facility ( Harvard Medical School , Boston , MA ) . Two independent FLAG-PB1 purifications were analyzed by AP-MS . The resulting data are presented in S1 Table and were compared with our database of 200 controls from stable 293 cell lines expressing the FLAG tag fused to non-viral proteins handled in identical fashion . The SAINT algorithm ( http://sourceforge . net/projects/saint-apms ) was used to evaluate the MS data [41] . The default SAINT options were low Mode = 1 , min Fold = 0 , norm = 0 . SAINT scores computed for each biological replicate were averaged ( AvgP ) and reported as the final SAINT score . Fold change was calculated for each prey protein as the ratio of spectral counts from replicate bait purifications over the spectral counts across all negative controls . A background factor of 0 . 1 was added to the average spectral counts of negative controls to prevent division by zero . Proteins included in the final interactome list had an AvgP ≥0 . 89 . Selection of the threshold for SAINT scores was based on receiver operating curve analysis performed using publicly available protein interaction data and the FLAG AP-MS data set as a list of true positive interactions . A SAINT score of AvgP ≥0 . 89 was considered a true positive BioID protein with an estimated FDR of ≤2% . All influenza viruses were propagated in MDCK cells and specific pathogen-free ( SPF ) embryonated chicken eggs . Monolayers of MDCK cells were washed with phosphate-buffered saline ( PBS ) and incubated with the respective virus at a multiplicity of infection ( MOI ) of 0 . 001 at 37°C . After 1 hour , the inoculum was aspirated , cells were washed twice and incubated at 37°C with DMEM without serum supplemented with tosylsulfonyl phenylalanyl chloromethyl ketone ( TPCK ) -treated trypsin ( 1 μg/ml; Worthington Biomedical Corporation , Lakewood , NJ ) . 48 hr postinfection ( p . i . ) virus was recovered from supernatants . For SPF eggs , 0 . 2 ml stock influenza virus at 1x103 TCID50 was injected into 11-day-old SPF fertile chicken eggs . The eggs were incubated in a stationary incubator at 35°C . After 72 hr incubation , eggs were cooled at -20°C for 30 min , then clear allantoic fluid was collected . For viral titration , plaque assays were performed as described [60] . Briefly , 1 . 2 × 106 MDCK cells/ml were plated in 6-well plates . MDCK cells were washed twice with DMEM without serum , then serial dilutions of virus were adsorbed onto cells for 1 hr . Cells were then covered with DMEM 2×Avicel RC591 NF mix ( FMC Biopolymer , Philadelphia , PA ) supplemented with TPCK-treated trypsin ( 1 μg/ml ) . Crystal violet staining was performed 72 hr p . i . , and visible plaques were counted . Monoclonal anti-FLAG ( M2 ) and anti-HA antibodies were obtained from Sigma ( St . Louis , MO ) . The polyclonal rabbit anti-TRIM32 was prepared as described elsewhere [23] . Mouse anti-PB1 and NP antibodies were obtained from BEI Resources . Anti-GFP antibody was purchased from Santa Cruz Biotechnology . The anti-lamin A , anti-α-tubulin and anti-ubiquitin antibodies were purchased from Cell Signaling Technology ( Danvers , MA ) . Anti-β-actin was purchased from Abcam ( Cambridge , MA ) . The anti-GST and anti-HIS antibodies were obtained from Bethyl Laboratories , Inc ( Montgomery , TX ) . The 3X FLAG peptide , HA peptide , MG132 and cychlohexmide were purchased from Sigma . Poly ( I:C ) was purchased from Invivogen ( San Diego , CA ) . Anti-Sendai virus was obtained from Charles River Labs . Anti-V5 antibody was bought from Thermo Scientific . cDNA encoding full-length human TRIM32 or TRIM32 mutants were subcloned in frame into mammalian expression vector pCMV-3TAG-8 with a C-terminal 3XFLAG , V5 or HA and the pEGFP-N1 vector containing a C-terminal GFP . cDNA encoding full-length PB1 from PR8 , WSN , Aichi , NY , H5N1 and H7N9 IAV ( S2 Table ) were subcloned into mammalian expression vector pCMV-3TAG-8 with a C-terminal 3XFLAG or HA tag . The cDNA encoding PR8 derived PA , PB2 or NP ( S2 Table ) were subcloned into mammalian expression vector pCMV-3TAG-8 with a C-terminal 3XFLAG or HA . The sources of all constructs are provided in S2 Table . 2x106 A549 cells were prepared for nuclear and cytoplasmic protein extraction according to the manufacturer’s protocol ( Pierce , Rockford , IL ) . Immunoprecipitation and immunoblotting were performed as previously described [32] . For immunoprecipitation with anti-FLAG or anti-HA antibodies , the cell lysates were incubated with EZview red anti-FLAG M2 or anti-HA affinity resin ( Sigma ) for 4 or 16 hr at 4°C . After washing with lysis buffer , proteins were eluted by incubation with 1 mg/ml 3X FLAG or HA peptide for 1 hr at 4°C . For immunoprecipitation with anti-TRIM32 or anti-PB1 , cell lysates were incubated with antibody and Protein A/G plus agarose ( Thermo Fisher Scientific , Cambridge , MA ) at 4°C for 16 hr . After washing with the lysis buffer , SDS-PAGE loading buffer was added and heated ( 95°C for 5 min ) . For immunoblotting , protein samples or 2% whole cell lysate were run on SDS-PAGE and transferred to PVDF membranes ( Bio-Rad , Hercules , CA ) . The membranes were blocked in 5% non-fat milk in 1×Tris-buffered saline and then incubated with diluted primary antibodies at 4°C for 16 hr . Anti-rabbit or anti-mouse IgG antibodies conjugated to horseradish peroxidase ( Pierce ) were used as secondary antibodies . An enhanced chemiluminescence system ( Pierce ) was used for detection . Quantitation of immunoblots was performed using GelQuantNet software . Cells were fixed with 4% formaldehyde in PBS for 10 min , permeabilized with methanol or 0 . 5% Triton X-100 for 15 min , blocked with 2% bovine serum albumin in PBS for 30 min , and then incubated with primary antibodies at 4°C for 16 hr . After three PBS washes , the cells were incubated with Alexa 488-labeled and/or Alexa 595-labeled secondary antibodies ( Invitrogen ) for 1 hr at room temperature . Cells were counterstained with DAPI ( 4' , 6-diamidino-2-phenylindole; Sigma ) . Automated imaging software ( Fiji ImageJ ) with colocalization 2 application package was used to quantitate colocalization . FlexiTube siRNA oligos against TRIM32 were purchased from Qiagen ( Valencia , CA ) . TRIM32 RNAi target sequences were as follows: #1 GACCGTGGTAACTATCGTATA , #2 CACACGATGGTGTTAGCTGAA and #3 CAGCACTCCAGGAATGTTCAA . For siRNA gene knockdown experiments , cells were cultured in 24-well plates and transfected with 30 pmol siRNA and 3 μl lipofectamine 2000 according to the manufacturer’s instructions ( Life Technologies , Grand Island , NY ) . After 48 hr , siRNA transfected cells were analyzed . A siRNA resistant TRIM32 was constructed in pCMV-3tag8-HA vector in which the No . 1 siRNA target sequence was mutated . Mutagenesis primers for the siRNA resistant rescue construct were as follows: 5’-tgaagtactagtcgctgatagaggaaagtacaggatccaagtctttacccgc-3’ and 5’-gcgggtaaagacttggatcctgtactttcctctatcagcgactagtacttca-3’ TRIM32-HIS was expressed from pET28b vector ( Clontech Laboratories , Mountain View , CA ) and PR8-PB1-GST was expressed from pGEX-5X-3 vector ( GE Healthcare Life Sciences , Pittsburgh , PA ) in Escherichia coli BL21 ( DE3 ) pLysS ( Life Technologies ) induced with 0 . 2 mM isopropyl-1-thio-β-D-galactopyranoside and 200 μM ZnSO4 for 16 hr at 18°C as detailed elsewhere [38] . GST-tagged proteins were purified with glutathione Sepharose 4B beads according to the manufacturer’s protocol ( GE Healthcare Life Sciences ) . HIS-tag proteins were purified with Ni-NTA agarose resins according to the manufacturer’s protocol ( Qiagen ) . Total RNA was extracted with RNeasy mini kit ( Invitrogen ) and reverse-transcribed ( 2 μg ) with QuantiTect Reverse Transcription Kit ( Qiagen ) . Human TRIM32 , human IFNβ , human GAPDH , mouse IFNβ and mouse β-glucuronidase mRNA levels were quantitated by RT-PCR with SYBR dyes on a LightCycler 480 ( Roche Life Sciences ) as described elsewhere [38 , 61] . Primers for hTRIM32 were: forward CCGGGAAGTGCTAGAATGCC and reverse CAGCGGACACCATTGATGCT . In vitro ubiquitination assays were performed according to the manufacturer’s manual ( Boston Biochem , Cambridge , MA ) . Ubiquitin ( 5 μg ) , E1 ( 200 ng ) , UBCH5a ( 300 ng ) ( Boston Biochem ) , TRIM32-HIS ( 0 . 8 μg ) and PR8-PB1-GST ( 2 μg ) were incubated with 2 mM ATP ( Sigma ) at 37°C 2 hr in ubiquitin assay buffer ( 20 mM Tris-HCl pH7 . 5 , 5 mM MgCl2 , 2 mM DTT ) . 1X stop solution ( Boston Biochem ) was added to end the reaction , after GST pull down the sample was washed with 1 M Urea for 60 min to exclude potential binding of unanchored polyubiquitin , then the sample was placed in SDS-loading buffer and boiled at 95°C for 5 min . Samples were subsequently analyzed by SDS-PAGE followed by Western blotting . Cells were aliquoted in 24 well plates for 24 hr prior to IAV infection . Human primary and cell lines were infected with 0 . 01 MOI PR8-GLuc and MEF were infected with 0 . 1MOI . After 1 hr the viral inoculum was aspirated , cells were washed twice and incubated at 37°C with 0 . 2% BSA-DMEM without serum . Cells were lysed 12–16 hr p . i . and the luciferase assay was performed using BioLux Gaussia Luciferase Assay Kit ( NEB , Ipswich , MA ) . HEK293 cells were transfected in triplicate with vectors expressing PR8 PB1 , PB2 , NP , PA and the indicated TRIM32 or siTRIM32 oligo in addition to the polymerase I ( PolI ) -driven plasmid transcribing an influenza A virus-like RNA coding for the reporter protein firefly luciferase to monitor viral polymerase activity [62] . Cells were lysed 48 hr after transfection . Luciferase activity was measured with a luciferase assay system ( Promega , Madison , WI ) . A plasmid constitutively expressing Renilla luciferase was transfected as a control . Unless indicated otherwise all experiments were repeated on at least three separate occasions . Data from representative experiments are illustrated . Methods for AP-MS data analysis are detailed elsewhere [41] . Other statistical analyses were done with the two-tailed Student’s t test . Data are presented as mean ± standard deviation . A P value of <0 . 05 was considered significant . Identification of the genes and proteins used throughout this study include: TRIM32 ( BC003154 ) , TRIM65 ( BC021259 . 2 ) , STUB1 ( BC0075456 ) , IQSEC ( BC010267 ) , GALK1 ( BC 001166 ) , PDCD6 ( BC050597 ) and DGT6/HAUS6 ( NM 001270890 . 1 ) . The GenBank references for viral genes include: PR8 PB1 ( EF467819 ) , WSN PB1 ( J02178 . 1 ) , NY PB1 ( CY039907 . 1 ) , Aichi PB1 ( CY121123 ) , H5N1 PB1 ( AY651664 . 1 ) , H7N9 PB1 ( CY147058 . 1 ) , H7N9 PB1 ( GISAID # EPI439508 ) PR8 PB2 ( CY148250 . 1 ) , PR8 PA ( CY148248 . 1 ) and PR8 NP ( CY148246 . 1 ) . | Influenza A virus presents a continued threat to global health with considerable economic and social impact . Vaccinations against influenza are not always effective , and many influenza strains have developed resistance to current antiviral drugs . Thus , it is imperative to find new strategies for the prevention and treatment of influenza . Influenza RNA-dependent RNA polymerase is a multifunctional protein essential for both transcription and replication of the viral genome . However , we have little understanding of the mechanisms regulating viral RNA polymerase activity or the innate cellular defenses against this critical viral enzyme . We describe how the E3 ubiquitin ligase , TRIM32 , inhibits the activity of the influenza RNA polymerase and defends respiratory epithelial cells against infection with influenza A viruses . TRIM32 directly senses the PB1 subunit of the influenza virus RNA polymerase complex and targets it for ubiquitination and proteasomal degradation , thereby reducing viral polymerase activity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | TRIM32 Senses and Restricts Influenza A Virus by Ubiquitination of PB1 Polymerase |
Dynamin Guanosine Triphosphate hydrolases ( GTPases ) are best studied for their role in the terminal membrane fission process of clathrin-mediated endocytosis ( CME ) , but they have also been proposed to regulate earlier stages of CME . Although highly enriched in neurons , dynamin-1 ( Dyn1 ) is , in fact , widely expressed along with Dyn2 but inactivated in non-neuronal cells via phosphorylation by glycogen synthase kinase-3 beta ( GSK3β ) kinase . Here , we study the differential , isoform-specific functions of Dyn1 and Dyn2 as regulators of CME . Endogenously expressed Dyn1 and Dyn2 were fluorescently tagged either separately or together in two cell lines with contrasting Dyn1 expression levels . By quantitative live cell dual- and triple-channel total internal reflection fluorescence microscopy , we find that Dyn2 is more efficiently recruited to clathrin-coated pits ( CCPs ) than Dyn1 , and that Dyn2 but not Dyn1 exhibits a pronounced burst of assembly , presumably into supramolecular collar-like structures that drive membrane scission and clathrin-coated vesicle ( CCV ) formation . Activation of Dyn1 by acute inhibition of GSK3β results in more rapid endocytosis of transferrin receptors , increased rates of CCP initiation , and decreased CCP lifetimes but did not significantly affect the extent of Dyn1 recruitment to CCPs . Thus , activated Dyn1 can regulate early stages of CME that occur well upstream of fission , even when present at low , substoichiometric levels relative to Dyn2 . Under physiological conditions , Dyn1 is activated downstream of epidermal growth factor receptor ( EGFR ) signaling to alter CCP dynamics . We identify sorting nexin 9 ( SNX9 ) as a preferred binding partner to activated Dyn1 that is partially required for Dyn1-dependent effects on early stages of CCP maturation . Together , we decouple regulatory and scission functions of dynamins and report a scission-independent , isoform-specific regulatory role for Dyn1 in CME .
Endocytosis has continued to evolve from a simple mode of ingestion and compartmentalization into a complex , multicomponent process that developed a bidirectional relationship with surface signaling [1 , 2] . In particular , evolutionary steps towards this complexity , which are associated with multicellularity , include the expansion to multiple isoforms of endocytic accessory proteins [3 , 4] and the introduction of dynamin [4 , 5] . Dynamin is the prototypical member of a family of large Guanosine Triphosphate hydrolases ( GTPases ) that catalyze membrane fission and fusion [6–8] . While encoded by single genes in Drosophila and Caenorhabditis elegans , further expansion of endocytic dynamins to three differentially expressed isoforms occurred in vertebrates [9] . Dynamin-1 ( Dyn1 ) , the first identified vertebrate isoform , has been extensively studied , and its mechanism of action as a fission GTPase is well understood [6 , 8 , 10] . The three dynamin isoforms are >70% identical in sequence , with most differences occurring in the C-terminal proline/arginine rich domain ( PRD ) that mediates interactions with numerous SRC Homology 3 ( SH3 ) domain-containing binding partners . Dyn1 and Dyn3 appear to be functionally redundant [11] . However , Dyn2 is unable to substitute fully for Dyn1 or Dyn3 in supporting rapid synaptic vesicle recycling in neurons [12] , and correspondingly , Dyn1 could not fully substitute for Dyn2 to support clathrin-mediated endocytosis ( CME ) in fibroblastic cells , even when overexpressed [13] . A direct comparison of the biochemical properties of Dyn1 and Dyn2 revealed differences in their in vitro curvature generating abilities: Dyn1 can potently induce membrane curvature and independently catalyze vesicle release from planar membrane surfaces , whereas Dyn2 requires the synergistic activity of curvature-generating Bin/Amphiphysin/Rvs ( BAR ) domain-containing proteins [14 , 15] . Less understood but still controversial [7 , 16–18] is dynamin’s suggested role in regulating early stages of CME [19–22] . Based on their differential biochemical properties , it was suggested that Dyn1 might be a more effective fission GTPase , while Dyn2 might be positioned to regulate early stages of CME [14] . However , whether dynamin isoforms play distinct roles in regulating CME has not been studied . Previously assumed to be neuron specific , recent studies have provided strong evidence that Dyn1 is indeed widely expressed but maintained in an inactive state in non-neuronal cells through phosphorylation at Serine 774 ( S774 ) by the constitutively active kinase , glycogen synthase kinase-3 beta ( GSK3β ) [23] . Acute inhibition of GSK3β in retinal pigment epithelial ( ARPE ) cells accelerates CME due to increased rates of clathrin-coated pit ( CCP ) initiation and maturation [23] . The effects of GSK3β inhibition on CME depend on Dyn1 but not Dyn2 , suggesting , unexpectedly , that Dyn1 might selectively function to regulate early stages of CME in non-neuronal cells . As the GSK3β phosphorylation site , S774 , is located within the PRD , its phosphorylation is presumed to alter interactions with dynamin’s SH3 domain-containing binding partners , as has been shown for binding partners enriched in the synapse [24 , 25] . Which interactions are affected in non-neuronal cells and whether these might be dynamin isoform specific is not known . Immunoelectron microscopic studies using an antibody that recognizes both Dyn1 and Dyn2 have localized endogenous dynamin to both flat and deeply invaginated CCPs in A431 adenocarcinoma cells [26 , 27] . Live-cell imaging has shown that , when overexpressed , both Dyn1-eGFP and Dyn2-eGFP are recruited at low levels to nascent CCPs , that their association with CCPs fluctuates , and that they undergo a burst of recruitment prior to membrane scission and vesicle release [17 , 22 , 28–31] . Indeed , when compared directly , transiently overexpressed Dyn1-eGFP and Dyn2-eGFP had indistinguishable profiles for their recruitment to CCPs [30 , 31] . Analysis of the recruitment of genome-edited Dyn2-eGFP to CCPs has similarly revealed a burst of recruitment at late stages of CME , as well as more transient interactions of lower numbers of Dyn2 molecules during earlier stages of CCP maturation [17 , 32] . To date , direct and quantitative comparisons of the nature of Dyn1 and Dyn2 association with CCPs when they are expressed at endogenous levels do not exist , nor is it known how activation of Dyn1 affects its association with CCPs . Here , we explore the isoform-specific behaviors of genome-edited Dyn1 and Dyn2 , both at steady state and in cells where Dyn1 is activated . We provide evidence for an early function of low levels of activated Dyn1 in regulating CCP initiation and maturation rates and that sorting nexin 9 ( SNX9 ) serves as an isoform-selective and activity-dependent binding partner of Dyn1 to regulate CCP maturation . Finally , we show that Dyn1 can be activated , under physiological conditions , downstream of epidermal growth factor receptors ( EGFRs ) to alter CCP dynamics .
Recent studies have shown that Dyn1 is widely expressed in non-neuronal cells [2]; but , like at the neuronal synapse [33] , it is mostly inactive at steady state due to phosphorylation by the constitutively active kinase GSK3β . Dyn1 function and its recruitment to CCPs have been studied in non-neuronal cells , albeit under conditions of overexpression and/or without an awareness of its phosphoregulation [14 , 21] . Therefore , to explore potential isoform-specific functions of Dyn1 and Dyn2 , as well as the role of GSK3β in regulating Dyn1 activity , we generated genome-edited H1299 non-small cell lung cancer cells , which we previously showed partially utilize Dyn1 for CME [23] . Cells expressing endogenously tagged Dyn2-mRuby2 were generated using previously validated Zinc Finger Nucleases ( ZFNs ) [32 , 34] to introduce double-stranded breaks and insert the mRuby tag with complementary flanking regions by homology-driven repair ( HDR ) ( Fig 1A ) . The resulting cells were single-cell sorted for mRuby2 fluorescence to obtain a heterozygous clone ( clone 235 , designated Dyn2-mRuby2end ) expressing a single mRuby2-tagged allele of Dyn2 ( Fig 1B ) . Endogenously tagging Dyn1 was complicated by the fact that the DNM1 gene encodes C-terminal splice variants derived from differential splicing of exons 21 and 22 ( S1A Fig ) , whose differential utilization could lead to partial loss of the fusion tag . Previous studies involving CRISPR/Cas9-mediated knockout and reconstitution with the Dyn1a C-terminal splice variant had confirmed that it fully reconstituted the GSK3β phosphoregulated activity of endogenous Dyn1 in H1299 cells [23] , including its ability to be activated by calmodulin [35] . Therefore , using a Clustered Regularly Interspaced Short Palindromic Repeats-associated nucleases 9 nickase ( CRISPR/Cas9n ) strategy , we targeted the Dyn1 gene at the end of exon 21 and introduced sequences encoding the remaining 19 amino acids of the Dyn1a isoform , followed by a seven amino acid linker [32] , monomeric eGFP fusion tag with stop codon , and finally , the SV40 polyadenylation signal to ensure unique expression of the “a” splice variant ( Fig 1A and S1B Fig ) . Single-cell sorting by fluorescence-activated cell sorting ( FACS ) for eGFP fluorescence , followed by clonal amplification generated a heterozygous clone ( clone 1B6 , designated Dyn1a-eGFPend ) expressing one eGFP-tagged allele of Dyn1a ( Fig 1B ) . Note that although Dyn1 is expressed at very low levels in H1299 cells , it can be readily detected following enrichment by amphiphysin-II SH3 domain pulldown . As a robust fiduciary marker for CCPs , clathrin light chain a ( CLCa ) carrying an N-terminal SNAP-fusion tag was stably introduced in parallel into both cell lines via a lentiviral vector with puromycin selection of SNAP-CLCa expressing cells . As previously reported by several groups , mild overexpression of Fluorescent Probe ( FP ) -CLCa has no effect on CME as measured by transferrin endocytosis [22 , 31 , 32 , 36] and no effect on CCP dynamics compared to AP2 or other markers [20 , 29 , 31] . We then performed live-cell dual-channel total internal reflection fluorescence microscopy ( TIRFM ) and analyzed CCP dynamics and Dyn recruitment using the master–slave ( CLCa–Dyn ) approach introduced with the cmeAnalysis software [22 , 37 , 38] . As expected based on previous studies using either overexpressed [28 , 29 , 31] or endogenously tagged Dyn2 [17 , 22 , 32] , Dyn2-mRuby2end was observed , on average , to gradually accumulate and then exhibit a burst of recruitment coincident with clathrin-coated vesicle ( CCV ) release . This can be seen in class-averaged tracks of bona fide CCPs with lifetimes ranging from 40–60 s ( Fig 1C and 1D ) and in all other CCP lifetime cohorts ( S2A and S2B Fig , S1 movie ) . In contrast , Dyn1a-eGFPendo recruitment was barely detectable above background , and no burst was evident ( Fig 1E and 1F , S2C and S2D Fig , S2 movie ) . This could reflect isoform-specific differences , very low levels of Dyn1 expression relative to Dyn2 , and/or the inactivation of Dyn1 by GSK3β phosphorylation . Thus , we further explored these possibilities . We first tested whether activation of Dyn1 alters CCP dynamics and/or the recruitment of Dyn1a-eGFPend in H1299 cells . As expected based on earlier studies in ARPE cells [23] , we confirmed that acute inhibition of GSK3β by incubation with the specific inhibitor , CHIR99021 , leads to decreased phosphorylation of Dyn1 at S774 within 30 min ( Fig 2A and 2B ) and increased rates of CME , as measured by transferrin receptor ( TfnR ) internalization ( Fig 2C ) . Importantly , the effects of GSK3β inhibition were dependent on Dyn1 expression , as treatment of Dyn1 knockout ( Dyn1KO ) H1299 cells [23] with CHIR99021 had no effect on CME ( Fig 2C ) . To further probe the mechanism by which activated Dyn1 accelerates CME , we introduced mRuby2-labeled CLCa into H1299 parent Dyn1KO cells and measured CCP dynamics by TIRFM . Analysis of the rates of assembly and departure of CCPs revealed that GSK3β inhibition resulted in a significant increase in the rate of coated pit initiation per unit cell area ( Fig 2D ) , as well as an increase in maturation rates ( i . e . , decrease in lifetimes ) of CCPs ( Fig 2E ) . The latter was evident in the change in lifetime distribution of all bona fide CCPs ( Fig 2F ) , which displayed a more quasi-exponential profile than untreated cells , indicative of a less-regulated process during early stages of CCP maturation [22] . Importantly , similar effects were observed for H1299 Dyn1a-eGFPend ( S3A–S3C Fig ) , confirming that the C-terminally eGFP-tagged splice variant , Dyn1a , was functional and activated by dephosphorylation . Again , GSK3β inhibition had no effect on CCP initiation rates or lifetimes in H1299 Dyn1KO cells ( Fig 2G–2I ) , confirming that these changes in CCP dynamics are a result of activation of Dyn1 . We then asked whether GSK3β inhibition and activation of Dyn1 altered its recruitment to CCPs . Surprisingly , there was no significant difference in the average recruitment intensity ( Fig 2J ) of Dyn1 at CCPs . Previous studies had shown that the appearance of dynamin fluctuates at CCPs [21 , 32]; thus , it was possible that GSK3β inhibition induces asynchronous and transient appearances of Dyn1 at CCPs that could be obscured by measuring average recruitment . Therefore , we also quantified the maximum intensity of Dyn1 recruited at any time along a CCP track . Using this orthogonal measurement , we again saw no effect of GSK3β inhibition on Dyn1 recruitment to CCPs ( Fig 2K ) . Together , these data suggest that dephosphorylation and activation of Dyn1 can alter CCP dynamics and CME even when Dyn1 is present at low amounts and that the effects of activation of Dyn1 on CCP dynamics are not likely explained simply by its increased recruitment to CCPs . It remained possible that the extremely low expression levels of Dyn1 in H1299 might limit our ability to detect GSK3β-dependent changes in its recruitment . To test this , we stably overexpressed Dyn1aWT-eGFP in H1299 Dyn1KO cells at approximately 20-fold levels higher than endogenous to generate Dyn1aWT-eGFPo/x cells ( Fig 3A ) . Importantly , overexpression of Dyn1aWT-eGFP itself did not result in any additional increase in TfnR uptake compared to the normal low endogenous levels ( Fig 3B , see also Fig 4G ) . However , as in parental and genome-edited H1299 cells , acute GSK3β inhibition in the Dyn1aWT-eGFPO/X cells resulted in increased rates of TfnR uptake ( Fig 3B ) and alterations in CCP dynamics , including increased rates of CCP initiation and maturation ( Fig 3C–3E ) . Yet similar to the Dyn1a-eGFPend-cells , GSK3β inhibition did not result in significantly enhanced recruitment of Dyn1aWT-eGFP to the membrane , either on average ( Fig 3F and 3G ) or when measured as maximum peak intensity ( Fig 3H ) . Moreover , there was no evidence of a burst of Dyn1 recruitment prior to CCV formation ( Fig 3G ) . Together , these results suggest that the observed changes in CCP dynamics are the result of a scission-independent early role for low levels of Dyn1 in regulating CME . Based on our finding that Dyn1 expression is required for the inhibitory effects of GSK3β on CME , we hypothesized that dephosphorylation of residues in Dyn1’s PRD should be sufficient to enhance CME efficiency . To test this , we introduced point mutations in Dyn1 at the serine residue phosphorylated by GSK3β ( S774 ) and at the priming serine site that is responsible for recruiting GSK3β ( S778 ) . We expressed this mutant as an eGFP fusion in H1299 cells , Dyn1aS774/8A-eGFP , at comparable levels to Dyn1aWT-eGFP ( Fig 3A ) . As predicted , Dyn1S774/8A-eGFP cells exhibited increased rates of CCP initiation ( Fig 4A ) , decreased CCP lifetimes ( i . e . , increased rates of CCP maturation , Fig 4B ) , and changed the lifetime distribution to a quasi-exponential profile ( Fig 4C ) . From these data , we conclude that dephosphorylated Dyn1 is sufficient to account for the effects of GSK3β inhibition on CCP dynamics . Surprisingly , even the nonphosphorylatable Dyn1aS774/8A-eGFP mutant was not efficiently recruited to CCPs and failed to display a pronounced late burst of recruitment accompanying membrane scission ( Fig 4D–4F ) . Interestingly , the changes in CCP dynamics in Dyn1aS774/8A-eGFP-expressing cells were not reflected in significantly increased rates of TfnR uptake , presumably due to compensatory changes that occur upon prolonged expression of activated Dyn1 versus acute activation ( Fig 4G ) . However , unlike parental H1299 cells or Dyn1aWT-eGFP cells , Dyn1KO cells reconstituted with Dyn1aS774/8A-eGFP exhibited significant residual levels of TfnR uptake upon siRNA knockdown of Dyn2 ( Fig 4G ) , consistent with functional activation of Dyn1 . Moreover , upon siRNA knockdown of Dyn2 , we detected an increase in Dyn1aWT-eGFP recruitment to CCPs ( Fig 4H and 4I ) , suggesting its activation as part of a compensatory mechanism to restore CME [23] . From these data , we conclude that Dyn1 is negatively regulated in non-neuronal cells through GSK3β-dependent phosphorylation of S774 and that dephosphorylated , active Dyn1 regulates early stages of CME even when present at low ( nearly undetectable , in the case of parental H1299 cells ) levels on CCPs . Importantly , overexpressed Dyn1 , even when activated by mutation or GSK3β inhibition ( Fig 3B ) , does not fully compensate for loss of Dyn2 function in CME , hence the two isoforms have partially divergent functions . We previously reported that several lung cancer cell lines express high levels of Dyn1 [35 , 39] . For example , A549 non-small cell lung cancer cells express approximately 5-fold higher levels of Dyn1 than Dyn2 [39] , corresponding to approximately 20-fold higher levels of Dyn1 than in H1299 cells ( S4A Fig ) . Reflective of these high levels of Dyn1 expression , siRNA knockdown of both Dyn1 and Dyn2 is necessary for potent inhibition of TfnR uptake in A549 cells ( S4B Fig ) . Therefore , we reasoned that it might be possible to individually knockout Dyn1 and the otherwise essential Dyn2 in A549 cell lines for reconstitution experiments . Thus , we used CRISPR/Cas9n to generate a complete knockout of Dyn1 ( Dyn1KO ) or Dyn2 ( Dyn2KO ) in A549 cells ( Fig 5A , S4C Fig ) and then introduced mRuby2-CLCa to track CCP dynamics . Acute inhibition of GSK3β had no effect on the rates of CCP initiation or maturation in Dyn1KO A549 cells but significantly stimulated the rate of CCP initiation and decreased the lifetimes of CCPs in Dyn2KO A549 cells ( Fig 5B and 5C ) . These data show that the two isoforms differentially regulate early stages of CME and confirm that the effects of GSK3β inhibition on CME depend on Dyn1 but not Dyn2 . To directly and quantitatively compare the relative recruitment efficiencies of the two isoforms to CCPs , we reconstituted these knockout cells with their respective eGFP-tagged isoforms and sorted for expression comparable to their endogenous levels ( i . e . , in these A549 cells we chose cells in which Dyn1a-eGFP levels were approximately 5-fold higher than Dyn2-eGFP ) ( Fig 5A ) . Additionally , we introduced SNAP-CLCa and mRuby2-CLCa in Dyn1a-eGFP and Dyn2-eGFP cells , respectively , so that we could distinguish the two A549 cell lines ( i . e . , Dyn1KO:Dyn1a-eGFP:SNAP-CLCa from Dyn2KO:Dyn2-eGFP:mRuby2-CLCa ) while imaging them in the same TIRFM field of view under the same conditions ( Fig 5D ) . These data directly show the differential recruitment efficiencies of Dyn1 and Dyn2 to CCPs . Live-cell imaging revealed the typical gradual accumulation and burst of Dyn2-eGFP recruitment to CCPs when averaged over the cohort of 40–60-s lifetime CCPs ( Fig 5E ) . However , under identical imaging conditions of the same fluorophore , Dyn1a-eGFP was recruited , on average , at least 10-fold less efficiently , even though it is expressed at higher abundance . The maximum intensity of tagged Dyn2 versus Dyn1 recruitment was also higher , albeit showing only an approximately 3-fold differential ( Fig 5F ) . A likely explanation for the differences in average and peak measurements is that in A459 cells , Dyn1a-eGFP does display a slight burst of recruitment at late stages of CCV formation that is visible when the Dyn1 signal is rescaled ( S4D Fig ) . Finally , to verify our results using an independent method , we performed Western blotting after subcellular fractionation and isolation of membrane versus cytosolic fractions , as confirmed using membrane-associated TfnR and cytosolic MEK1/2 as markers ( Fig 5G ) . Under these fractionation conditions , approximately 90% of Dyn2 is membrane associated , whereas only 50% of Dyn1 sediments with the membrane fraction ( Fig 5G ) . We observed a consistent , approximately 20% increase of membrane-associated Dyn1 upon GSK3β inhibition that was not detected by TIRFM . These biochemical data indicate a greater extent of membrane association of both active and inactive Dyn1 than detected at CCPs by TIRFM . The differences could reflect recruitment of Dyn1 to sites on the plasma membrane other than CCPs , as has been previously reported [40] . The approximately 20% increase in recruitment of activated Dyn1 likely reflects the increase in number of CCPs that occurs upon GSK3β inhibition , rather than an increase in Dyn1 per CCP . Consistent with TIRFM data , the distribution of phosphorylated Dyn1 ( detected with an S774 phosphospecific antibody ) was indistinguishable from total Dyn1 ( i . e . , there was no de-enrichment of phosphorylated Dyn1 in the membrane-bound fractions ) . These data confirmed that dephosphorylation of Dyn1 on S774 by GSK3β inhibition does not enhance its recruitment to CCPs . Thus , the effects of activated Dyn1 on CCP initiation and maturation occur either independently of its direct association with CCPs or , more likely , are manifested by very low levels of CCP-associated dephosphorylated Dyn1 . Dynamin exists as a tetramer in solution [41 , 42] and assembles into higher-order helical oligomers on the membrane . Exploiting Dyn1KO and Dyn2KO A549 cells reconstituted with Dyn1a- or Dyn2-eGFP , respectively , we next assessed the degree to which Dyn1 and Dyn2 form hetero-tetramers in solution . Dyn1- or Dyn2-eGFP were efficiently immunoprecipitated with anti-eGFP nanobodies and the immunobeads were washed with 300 mM salt to disrupt any potential higher-order dynamin assemblies before measuring the fraction of Dyn2 or Dyn1 that coprecipitated . Under these conditions , we pulled down nearly 100% of the eGFP-tagged dynamins but only approximately 30% of Dyn2 with Dyn1-eGFP and <5% of Dyn1 with Dyn2-eGFP ( S5A Fig ) . The difference in the extent of hetero-tetramerization is consistent with the approximately 5-fold higher levels of expression of Dyn1 versus Dyn2 in these cells . Thus , the two isoforms predominantly exist as homo-tetramers in solution . We also examined the relative abilities of Dyn1 and Dyn2 to co-assemble into higher-order structures in vitro . For this , we used a dominant-negative Dyn1 mutant ( Dyn1S45N ) defective in GTPase activity , which , when co-assembled with wild-type dynamin into higher-order oligomers on lipid nanotubes , will inhibit total assembly-stimulated GTPase activity through the intercalation of GTPase-defective subunits adjacent to wild-type subunits [43 , 44] . As expected , Dyn1S45N efficiently co-assembles with Dyn1WT such that , when present at equimolar levels , the total assembly-stimulated GTPase activity is inhibited by 50% . In contrast , at the same concentrations of Dyn1S45N , Dyn2 GTPase activity was significantly less affected ( S5B Fig ) , indicating that Dyn2 less efficiently co-assembles into higher-order oligomers with the mutant Dyn1 . Thus , consistent with their differential recruitment to CCPs , even when present at comparable levels of expression in the same cell type , the two isoforms only weakly interact . Our results establish that Dyn1 and Dyn2 are differentially recruited to CCPs in non-neuronal cells and that , on average , Dyn1 is recruited at much lower levels than Dyn2 . Despite this , acute activation of Dyn1 globally alters CCP dynamics . Thus , we next directly compared the recruitment of Dyn1 and Dyn2 to CCPs to determine whether Dyn1 is recruited at low levels to all CCPs or instead might be recruited at higher levels to a subpopulation of CCPs . Such heterogeneity would be lost by averaging . For this , we took advantage of the higher levels of Dyn1 expression in A549 cells and generated double genome-edited cells expressing Dyn1a-eGFP and Dyn2-mRuby2 . We first used ZFNs to generate Dyn2 mRuby2-edited A549 cells and subsequently introduced a C-terminal eGFP to the Dyn1a splice variant using CRISPR/Cas9 , as described earlier ( Fig 1A , see Materials and methods ) . This yielded an A549 cell line homozygous for endogenously tagged Dyn2-mRuby2 and heterozygous for endogenously tagged Dyn1a-eGFP ( 2 of 3 Dyn1 alleles tagged in these triploid A549 cells ) ( Fig 6A ) . We confirmed that the double genome-edited cells exhibited comparable rates of TfnR uptake , as well as the degree of dependence on Dyn2 for CME , relative to the parent cells ( Fig 6B ) . SNAP-CLCa was introduced into these cells by lentiviral transfection ( Fig 6C ) , and we confirmed that GSK3β inhibition resulted in increased rates of CCP initiation , reduced CCP lifetimes , and altered the lifetime distributions of CCPs ( Fig 6D–6F ) , as in the parental cells . Thus , the genome-edited Dyn isoforms were functionally active . We next assessed the interplay between Dyn1a-eGFP and Dyn2-mRuby2 using three-color live-cell TIRFM imaging at 0 . 5 Hz ( 2 s per frame ) ( Fig 7A , S3 Movie ) . As reported previously , we detected fluctuations of both Dyn1 and Dyn2 at CCPs over their lifetimes ( examples shown in Fig 7B ) and frequently detected a burst of Dyn2 just prior to CCV formation . In many cases , we also detected a burst of Dyn1 recruitment , albeit to a lesser degree . For more quantitative analysis of these data , we applied the 3-channel functionality of our cmeAnalysis package to perform three-color master/slave analyses [22] . Using clathrin as the “master” channel and Dyn1 and 2 as “slave” channels , we determined whether the clathrin tracks contained either Dyn1 , Dyn2 , both , or neither . Individual CCP tracks were considered positive for Dyn1 and/or Dyn2 if the intensities of Dyn1/2 signals detected at the position of the clathrin tag were significantly higher than the local Dyn1/2 background signal around the clathrin tag position for a period of time exceeding random associations , as previously described [22] . This analysis revealed that in double genome-edited Dyn1a-eGFPend/Dyn2-mRuby2end A549 cells , both Dyn2 and Dyn1 could be robustly detected in approximately 75% of all bona fide CCPs ( Fig 8A ) . Moreover , in this population of CCPs , a clear burst of recruitment of both Dyn1a-eGFP and Dyn2-mRuby2 could be detected prior to CCV formation . Importantly , the apparently higher levels of recruitment of Dyn1-eGFP versus Dyn2-mRuby2 in these genome-edited cells is not a reflection of protein levels but rather of imaging conditions and brightness for two different fluorophores ( compare with Fig 5E ) . The remaining CCPs were roughly equally distributed as Dyn1 only , Dyn2 only , and both Dyn1- and Dyn2-negative subpopulations ( Fig 8A ) . Note that the Dyn2 levels in the “Dyn1 only” CCPs were still on average higher than background ( Dyn1/Dyn2 negative ) , reflecting the stringency of our master/slave detection and suggesting that Dyn2 is recruited to >90% of all CCPs , albeit to variable extents . We next compared per cell median lifetimes of CCPs relative to their dynamin isoform composition and found that CCPs bearing higher levels of Dyn2 and Dyn1 exhibited longer lifetimes ( median approximately 80 s ) than single-positive CCPs ( median approximately 38 s ) ( Fig 8B ) . CCPs that failed to detectably recruit either isoform were the shortest lived ( median approximately 20 s ) and likely represent abortive CCPs . These findings are consistent with previous data suggesting that a threshold level of Dyn2 recruitment is required for efficient CCP maturation [22 , 34] . All of these CCP subpopulations showed a significant decrease in CCP lifetimes upon inhibition of GSK3β , consistent with other data that only low levels of Dyn1 are required to alter CCP maturation . Our findings thus far point to isoform-specific functions of Dyn1 and Dyn2 and hence suggest the existence of isoform-specific binding partners . Dyn1 and Dyn2 are >80% identical except for their C-terminal PRDs , which are only 50% identical and likely determine isoform-specific interactions with SH3 domain-containing proteins . The Dyn1KO and Dyn2KO A549 cells provide an opportunity to measure Dyn2 and Dyn1-dependent CME , respectively , without the possibility of compensation . Thus , we measured , by TfnR uptake , the effects of siRNA knockdown of several known SH3 domain-containing binding partners on Dyn2-dependent CME in the Dyn1KO cells and on Dyn1-dependent CME in the Dyn2KO cells . Knockdown of these dynamin partners has only mild effects on TfnR uptake in parental A549 cells and in Dyn1KO cells , whose endocytosis is exclusively Dyn2 dependent ( Fig 9A ) . Whether these mild effects reflect partial redundancy with other dynamin partners , activation of compensatory mechanisms [23] , or that these factors , which were identified primarily as dynamin partners in brain lysates , play only minor roles in TfnR uptake in non-neuronal cells , cannot be discerned from these studies . Interestingly , siRNA knockdown of Grb2 appeared to inhibit TfnR uptake in Dyn1KO cells by approximately 20% , while not affecting TfnR uptake in either parental or Dyn2KO cells . This suggests that Grb2 might preferentially function together with Dyn2 in CME and that its depletion in parental cells can be compensated for by Grb2-independent Dyn1 activity . In contrast , siRNA knockdown of SNX9 only mildly inhibited TfnR uptake in parental A549 and had no significant effect on Dyn2-dependent TfnR uptake in Dyn1KO cells , but decreased TfnR uptake in Dyn2KO cells by >50% ( Fig 9A ) . Thus Dyn1-dependent endocytosis appears to be particularly sensitive to SNX9 knockdown . We next tested whether SNX9 preferentially interacts with Dyn1 versus Dyn2 by GFP pulldown assays using Dyn1KO cells reconstituted with either Dyn1aWT- or Dyn1S774/8A-eGFP and Dyn2KO A549 reconstituted with Dyn2-eGFP . Consistent with previous results [45 , 46] , we confirmed that SNX9 binds both Dyn1 and Dyn2 ( Fig 9B ) . However , the ratio of SNX9 binding to Dyn1 versus Dyn2 was 1 . 7 ± 0 . 6 ( mean ± SEM , n = 3 ) , indicative of a slight preference for Dyn1 . Importantly , SNX9 showed a marked preference for binding to the nonphosphorylated and active Dyn1S774/8A-eGFP . The ratio of SNX9 binding to Dyn1S774/8A versus Dyn1WT was 3 . 6 ± 0 . 9 ( mean ± SEM , n = 3 ) . These data suggested that SNX9 might be a preferential functional partner of activated Dyn1 . To test whether SNX9–Dyn1 interactions were required for the effects of activated Dyn1 on CCP initiation rates , CCP maturation , or both , we asked returned to the Dyn1KO H1299 cells reconstituted with Dyn1WT versus Dyn1S774/8A and tested whether the selective effects of Dyn1S774/8A on CCP dynamics ( Fig 4A–4C ) were dependent on SNX9 . Knockdown of SNX9 decreased the rate of CCP initiation in Dyn1WT but was not required for the enhanced rate of CCP initiation triggered by Dyn1S774/8A expression ( Fig 9C ) . Thus , other , yet-unidentified binding partners are responsible for the Dyn1-dependent effect on CCP initiation . SNX9 knockdown also led to an increase in the median CCP lifetimes in both Dyn1WT- and Dyn1S774/8A-expressing cells ( Fig 9D ) . These data suggest that SNX9 functions in both Dyn1-dependent and independent stages of CCP maturation . Consistent with this , SNX9 knockdown also abrogated the effects of Dyn1S774/8A expression on the lifetime distribution of bona fide CCPs ( Fig 9E ) , reverting the quasi-exponential distribution seen in Dyn1S774/8A to a distribution nearer to control . The strong effect of SNX9 knockdown is also seen in the rightward shift of the lifetime distribution of Dyn1WT cells treated with SNX9 siRNA . Together , these data suggest multiple roles of SNX9 at multiple stages of CME , including the support of Dyn1’s early functions in accelerating CCP maturation . We have shown that strong pharmacological inhibition of GSK3β activates Dyn1 in non-neuronal cells and results in increased rates of CCP initiation and maturation , leading to increased rates of TfnR uptake via CME . However , it is not clear whether this regulatory effect on Dyn1 function modulates CME under more physiologically relevant conditions . To test this , we treated serum-starved A549 cells with epidermal growth factor ( EGF ) , which is known to activate Akt and in turn to phosphorylate and inactivate GSK3β [47] . We confirmed that GSK3β is phosphorylated in EGF-treated cells and that this resulted in reduced levels of phosphorylation of Dyn1 at S774 ( Fig 10A , quantified in Fig 10B and 10C ) . As predicted by the results of inhibitor experiments , EGF treatment of serum-starved cells also increased the rate of CCP initiation ( Fig 10D ) , decreased CCP lifetimes ( Fig 10E ) , and , compared to control cells , resulted in a shift of the lifetime distributions of bona fide CCPs to a more quasi-exponential distribution ( Fig 10F ) . Importantly , the effects of EGF treatment on CCP initiation rate and lifetimes were not seen in A549 Dyn1KO cells ( Fig 10G and 10H ) . These data suggest that Dyn1 can be activated to alter CCP dynamics under physiological conditions through signaling downstream of EGFR .
Our experiments provide further evidence that Dyn1 , in addition to its well-studied roles in membrane fission during synaptic vesicle recycling , also has noncanonical functions as a regulator of the earliest stages of CME in non-neuronal cells . As in neurons , Dyn1 activity is negatively regulated by constitutive phosphorylation and activated by dephosphorylation . When studied at endogenous levels of expression , we show that Dyn1 and Dyn2 have distinct functions in CME , reflected in their quantitatively and qualitatively different recruitment to CCPs . While Dyn1 is expressed at very high levels in the brain , it is , like Dyn2 , also widely expressed , albeit at lower levels , in all tissues and cells [2] . Importantly , we show that acute activation of even low , nearly undetectable levels of Dyn1 can increase the rates of CCP initiation and maturation to accelerate CME . These effects of Dyn1 activation occur well upstream of membrane fission and are not accompanied by a pronounced burst of recruitment prior to CCV formation . Hence , they reflect noncanonical activities of Dyn1 , likely mediated by unassembled tetramers , that are distinct from its well-studied role in fission . As in neurons [33] , Dyn1 is constitutively inactivated in non-neuronal cells by phosphorylation at S774 in the PRD by GSK3β . Acute chemical inhibition of GSK3β activates Dyn1 to alter CCP dynamics and increase the rate of CME . While GSK3β has numerous substrates , we show that Dyn1 is both necessary and sufficient to account for the effects of GSK3β inhibition on CCP dynamics and CME . Specifically , the effects of GSK3β inhibition on CME are dependent on Dyn1 but not Dyn2 expression , and Dyn1KO cells reconstituted with a nonphosphorylatable mutant of Dyn1 show increased rates of CCP initiation and maturation that phenocopy the effects of GSK3β inhibition . Even when Dyn1 is activated , either by GSK3β inhibition or by mutation of S774 and S778 to alanine , CME remains dependent on Dyn2 ( data herein and [23] ) . Thus , the two isoforms play functionally distinct roles in CME . It is unlikely that Dyn1 activation merely releases an inhibitory effect of inactive Dyn1 on CME , for example by competing with Dyn2 , because even high levels of overexpression of wild-type Dyn1 does not inhibit CME ( see for example , [26] ) . Further studies are needed to elucidate the mechanisms underlying these distinct roles . A direct comparison of the in vitro properties of Dyn1 and Dyn2 established that they differ in their curvature generating/sensing properties [14] . While Dyn1 is an efficient curvature generator that is able to tubulate and catalyze fission from planar lipid templates , Dyn2 is a curvature sensor that is able to catalyze membrane fission of highly curved lipid templates but requires the synergistic activity of curvature-generating N-terminal Bin/Amphiphysin/Rvs ( N-BAR ) domain-containing accessory factors to drive curvature generation and fission from planar templates [14 , 15] . Strikingly , these biochemical differences could be ascribed to a single residue ( Y600 in Dyn1 , L600 in Dyn2 ) encoded within hydrophobic loops of the curvature-generating Pleckstrin homology ( PH ) domain of dynamin [14] . Based on these biochemical differences , it was suggested that the unique properties of Dyn2 might enable this isoform to monitor CCP maturation and to catalyze fission only after the development of a narrow membrane neck connecting deeply invaginated CCPs to the plasma membrane . Unexpectedly , the findings presented here and elsewhere [23 , 35 , 39] establish that Dyn1 uniquely functions to regulate the earliest stages of CME , including the rate of CCP initiation and maturation . The mechanisms underlying these Dyn1-specific activities remain to be elucidated . Activation of Dyn1 also altered the shape of the lifetime distribution curve for CCPs from a broad Rayleigh-like distribution with a distinct peak at approximately 30 s to a more exponential distribution . We have previously suggested that the Rayleigh-like shape reflects rate-limiting regulatory processes operating during the first 30 s of CCP progression [20 , 22 , 38] . It is possible that , due to its curvature-generating ability and/or through interactions with other partner proteins , Dyn1 activation accelerates these complex early processes of CCP maturation . Although Dyn1 and Dyn2 exhibit >80% sequence identity within their GTPase , middle , PH domains , and GTPase effector domain ( GED ) , previous studies of the cellular activities Dyn1/Dyn2 chimeras have nonetheless revealed striking isoform-specific functional differences conferred by both the PH and GTPase domains [14 , 48] . Most divergent among mammalian dynamin isoforms is the PRD , which functions to mediate interactions with numerous SH3 domain-containing binding partners , and has been shown to target dynamin to CCPs [40] . Earlier comparative studies of Dyn1 and Dyn2 [13] , as well as Dyn1/Dyn2 PRD chimeras expressed at near endogenous levels [14] , have shown that Dyn2 is more efficiently recruited to CCPs in a PRD-dependent manner . However , these studies did not take into account the negative regulation of Dyn1 by GSK3β phosphorylation . Here , we reproduce and extend these findings by showing that the differential recruitment of Dyn1 is not due to phosphorylation of its PRD , at least on S774 or 778 . Indeed , the recruitment of Dyn1 to CCPs was not significantly enhanced by GSK3β phosphorylation or when S774/S778 were mutated to nonphosphorylatable alanines . Thus , surprisingly , the effects of activated Dyn1 on CCP dynamics appear to occur independent of detectably enhanced recruitment to CCPs . It will be important to identify isoform-specific binding partners for Dyn1 and Dyn2 in non-neuronal cells . To date , most dynamin binding partners , including endophilin , amphiphysin , and intersectin have been identified in brain lysates in which Dyn1 is highly expressed and may play a specialized function in rapid synaptic vesicle recycling . Thus , it is perhaps not surprising that siRNA knockdown of these dynamin binding partners in non-neuronal cells has only mild effects on primarily Dyn2-dependent TfnR endocytosis . Further studies are needed to identify essential non-neuronal effectors of both Dyn2 and Dyn1 function in CME . Unexpectedly , our data suggests that SNX9 , which was first identified as a major binding partner of Dyn2 in HeLa cells [45] , interacts most strongly with dephosphorylated Dyn1 and that the effects of Dyn1 activation on early CCP maturation are dependent on SNX9 . Published findings on SNX9 function in CME are enigmatic . Consistent with our findings in A549 NSCLC cells , siRNA-mediated knockdown of SNX9 has only mild effects on CME in several cell lines studied [46 , 49] . While it has been suggested that these mild effects are due to redundant functions of the distantly related ( 40% sequence identity ) SNX18 [50] , this is not the case in all cell types [49 , 50] . TIRFM studies on the recruitment of overexpressed SNX9-GFP to CCPs have also yielded differing results: It has been reported to be recruited coincident with [46] , after [31] , and before dynamin [50] . Interestingly , one study reported that SNX9 might linger at endocytic “hot-spots , ” where it could function as an organizer of CCP nucleation [51] . Our results further suggest a more complex role for SNX9 at multiple stages of CME . SNX9 is not required for activated Dyn1-dependent increases in the rates of CCP initiation . However , it is required for the effects of Dyn1 activation on accelerating CCP maturation , as indicated by the marked switch from Rayleigh-like to quasi-exponential CCP lifetime distributions , which is reversed by SNX9 knockdown . That SNX9 knockdown alone decreases the rate of CCP initiation and slows CCP maturation in cells expressing Dyn1WT suggests other , potentially Dyn2-dependent and/or dynamin-independent functions in CME . More work is required to define both the multiple functions of SNX9 in CME and to identify Dyn1-specific binding partners required for CCP initiation . Recent studies have shown that Dyn1 is upregulated and/or activated in several cancer cell lines [35 , 39] , leading to the suggestion that Dyn1 might function as a nexus between signaling and CME [2] . Here , we show that Dyn1 can be activated downstream of EGFR to alter CCP dynamics . Previous studies showed that tumor necrosis factor-related apoptosis-inducing ligand ( TRAIL ) -activated death receptors can activate Dyn1 to drive their selective uptake via CME [35] . Similarly , elegant studies on clathrin-mediated endocytosis of the G-protein coupled β-adrenergic receptors have shown that they alter the maturation kinetics of the CCPs in which they reside through delayed recruitment of Dyn2 [52] . These authors did not examine Dyn1 recruitment or function . Further studies will be needed to determine whether other signaling receptors can selectively alter the composition and/or maturation kinetics of CCPs in which they reside and , if so , whether these changes , as suggested by our present data , are at least in part Dyn1 dependent . Based on our results and other recent findings [23 , 35 , 39] regarding Dyn1 function downstream of signaling in non-neuronal cells , it is perhaps surprising that Dyn1 knockout mice develop normally , can live for up to 2 weeks after birth , and exhibit primarily neuronal defects [12] . This could , in part , be due to the redundant function of Dyn-3 , as Dyn1/Dyn3-null mice exhibit a more severe phenotype and die within hours of birth [11] . However , we speculate that kinase-based activation of Dyn1 might function at the level of individual CCPs to foster their initiation and accelerate maturation , perhaps at a threshold of signaling not reached during normal development . Indeed , recent evidence has pointed to cargo-selective roles of Dyn1 in regulating CME and signaling in cancer cells [2 , 35 , 39] . Therefore , it might be interesting to probe Dyn1 or Dyn1/Dyn3 knockout mice for other , potentially more subtle , non-neuronal phenotypes related to signaling in health and disease .
Non-small cell lung cancer cell lines A549 and H1299 were kindly provided by Dr . John Minna ( The Hamon Center for Therapeutic Oncology , Depts . of Internal Medicine and Pharmacology , UTSW ) and were grown in RPMI 1640 ( Life Technologies ) with 5% FBS at 37°C and 5% CO2 and imaged in a temperature-controlled chamber mimicking similar culture conditions . The retroviral expression vector pMIEG was a modified pMIB ( CMV-IRES-BFP ) vector encoding Dyn1 cDNA with N-terminal HA-tag and C-terminal eGFP fusion tag . Point mutations to introduce S774/8A in Dyn1 cDNA were performed by site-directed mutagenesis . The lentiviral expression vector pLVX-puro ( Clontech ) encoded CLCa N-terminally tagged with mRuby 2 [53] or SNAP-tag [54] spaced with a 6 amino acid GGSGGS linker . The constructs were assembled from PCR fragments of mRuby2 , SNAP , and CLCa ( for primers , see S1 Table ) in yeast as described below and subsequently cloned into pLVX-puro . Lentiviruses were generated in 293T packaging cells following standard transfection protocols [55] and were used for subsequent infections . They were prepared to transduce fluorescently tagged ( mRuby2 or SNAP tag ) CLCa fused to its N-terminus . Infected cells were selected using 10 μg/ml puromycin for 4 d , conditions under which uninfected cells perished . The cells were passaged for 2 w before imaging for CME analysis . Retroviruses were also generated in 293T cells and used to stably transduce eGFP tagged Dyn1WT , Dyn1S774/8A , and Dyn2WT proteins . Gene transduction was performed by exposing A549 or H1299 cells to retrovirus-containing cell culture supernatants through two rounds of viral transduction spread across 5 d . The recipient cells were further expanded to confluency in a 10-cm culture dish and FACS sorted for eGFP levels comparable to endogenous Dyn1-eGFP in A549 cells . Transfections for siRNA knockdown experiments were carried out using Lipofectamine 2000 or Lipofectamine RNAi-Max ( Life Technologies ) , following manufacturer’s protocol . For siRNA mediated knockdown , approximately 2 × 105 cells ( H1299 ) or 3 × 105 cells ( A549 ) were plated in each well of a six-well plate . Twenty nmol siRNA was used per well , and two rounds of transfection across 5 days was sufficient to achieve over 90% knockdown . Perturbation of culture conditions by GSK3β inhibitor involved the addition of 10 μM CHIR99021 ( Sigma ) to prewarmed culture media and incubation of cells for 30 min before additional analysis . Growth factor stimulation was performed by adding 20 ng/ml EGF ( Invitrogen ) to prewarmed , serum-free culture media . Cells were analyzed after 10 min of incubation with EGF . Genome-edited A549 and H1299 cells were generated by editing Dyn1 and Dyn2 to carry fusion tags . For the endogenous labeling of Dyn1 with fluorescent reporter proteins , we chose an approach based on site directed introduction of CRISPR/CAS9n-targeted DNA breaks and template assisted homology driven repair . eGFP fused to Dyn1 at its C-terminus was generated by CRISPR/Cas9n nickase strategy targeting the end of exon 21 of the DNM1 gene , inserting the last 19 amino acids of splice isoform “a , ” a seven amino acid linker [32] , monomeric eGFP with a stop codon , and the SV40 polyadenylation signal . In the donor plasmid , this inserted sequence was flanked by approximately 950 base pair homology arms for HDR . The +gRNA pair was designed using publicly available software ( http://crispr . mit . edu/ ) and prepared as described [56] with oligos DNM1-Nuclease-A-f/ DNM1-Nuclease-A-r and DNM1-Nuclease-B-f/ DNM1-Nuclease-B-r , respectively ( S1 Table ) . For assembly of the donor vector , the segments were amplified with oligonucleotides coding approximately 30 nucleotide overhangs . The bacterial artificial chromosome clone RP11-348G11 ( BACPAC Resources Center , Children’s Hospital Oakland Research Institute , Oakland , California ) covering the end of human DNM1 gene was used as template for the left and right homology arms ( Fig 1A ) . The left and right homology arms were amplified using primer pairs DNM1-LH-f/DNM1-LH-r and DNM1-RH-f/DNM1-RH-r , respectively ( see S1 Table ) . The 19 C-terminal amino acids of splice isoform “a” and the linker sequence DPPVATL [32] were covered with oligonucleotides DNM1-C-assembly-f and DNM1-C-assembly-r and amplified with short primers DNM1-Cterm-f and DNM1-Cterm-r . The sequence coding for monomeric eGFP and the SV40 polyadenylation signal were amplified from plasmid peGFP-N1 ( Clontech ) , which carried the A206K mutation [57] with primers DNM1-eGFP-f/DNM1-eGFP-r and DNM1-pA-f/DNM1-pA-r , respectively . The first and last primers ( DNM1-LH-f , DNM1-RH-r ) also included overhangs for the E coli/yeast shuttle vector pRS424 [58] . Dyn2-mRuby genome-edited cells were generated using previously validated ZFNs [32 , 34] . For the DNM2-mRuby2 donor vector , the homology arms were amplified from the published [34] DNM2-eGFP construct ( gift from D . Drubin , University of California , Berkeley ) with primers DNM2-LH-f/DNM2-LH-r and DNM2-RH-f/DNM2-RH-r , respectively . The mRuby2 segment together with the linker sequence , DPPVATL [32] , was amplified from pmRuby2-C1 , a gift from Michael Lin ( Addgene #40260 ) [53] . The PCR products were purified on 1% agarose gels and extracted using standard protocols before transformation into YPH500 yeast cells [59] . Yeast transformation , plasmid extraction , and plasmid validation were performed as described earlier [60] . The guide-RNA plasmids for DNM1-eGFP and donor vectors for both DNM1-eGFP and DNM2-mRuby2 are available from Addgene ( IDs 107795 , 107796 , 107794 , and 107793 , respectively ) . For both the edits , the nCas9 nickase + gRNA pairs or the ZFN nuclease pairs were added at 1 μg DNA concentration , and 2 μg of the donor plasmid was added to this mixture in 150 ul OptiMEM ( Life Technologies ) . This mixture was then added to 5 . 5 μl of Lipofectamine 2000 ( Life Technologies ) in 150 μl OptiMEM ( Life Technologies ) , briefly vortexed , and incubated at room temperature for 15 min . The mixture was then added to cells plated 12 h earlier at 70% confluency ( approximately 3 × 106 cells per well in six-well dish ) with freshly replaced media . Transfect-containing media was replaced by prewarmed fresh media and the cells were allowed to grow for the next 48 h and then passaged for expansion in a 10-cm dish . The expanded cells were sorted as eGFP ( or mRuby2 ) gene-edited single cells into 96 well plates 4 days after transfection using a FACSAria 2-SORP ( BD Biosciences , San Jose , CA ) instrument equipped with a 300-mW , 488-nm laser and a 100-μm nozzle . Clonal expansion ensued by incrementing the culture dish area and maintaining a minimum 50% cell confluency . Single clones were then assayed for edits by western blotting , and cells positive for genome edits were expanded . In order to generate double genome edited A549 cells , the A549 clone , 2C8 , with homozygous Dyn2-mRuby2 knock-in was chosen and subsequently edited for Dyn1 , and cell selection was performed as before using FACS preliminary screen followed by western blotting for validation . Dyn1 KO H1299 cells were generated as previously described [23] and the same strategy was employed to generate A549 Dyn1 KO cells . Briefly , cells plated in six-well plates were transfected with 1 μg each of single-guide RNAs ( sgRNAs ) and Cas9 nickase encoding plasmids and cotransfected with a 20th of peGFP plasmid . eGFP-positive cells were assumed to have harbored both the sgRNA guides and single-cell-sorted by FACS . In addition , Dyn2 KO cells were generated using a similar double nickase strategy with sgRNAs CGATCTGCGGCAGGTCCAGGTGG and CGCCGGCAAGAGCTCGGTGCTGG in the pX335 vector . Complete knockout of Dyn1 and 2 was validated by western blotting . Cells expressing appropriate fluorophores were cultured overnight on an acid-etched and gelatin-coated coverslip , placed in a well in a six-well plate . At the time of imaging , cells were checked for adherence and spreading . When imaging SNAP-tagged proteins , labeling was performed by incubating cells in 1 ml of fresh , prewarmed media containing 1 μl of predissolved SNAP-CELL 647-SiR dye ( NEB ) . After 30 min incubation under standard incubator conditions , the media was aspirated , washed twice with sterile PBS , and reincubated in fresh culture media . The coverslips were mounted on glass slides with spacers and sealed with the same media . For experiments involving the addition of growth factor or inhibitor , cells were preincubated for the appropriate times and the coverslips were mounted as before with the treated media . The coverslips were then imaged using a 60x 1 . 49 NA Apo TIRF objective ( Nikon ) mounted on a Ti-Eclipse inverted microscope with Perfect Focus System ( Nikon ) equipped with an additional 1 . 8x tube lens , yielding at a total magnification of 108x . TIRF illumination was achieved using a Diskovery Platform ( Andor Technology ) . During imaging , cells were maintained at 37°C in RPMI supplemented with 5% fetal calf serum . Time-lapse image sequences were acquired at a penetration depth of 80 nm and a frame rate of 2 Hz ( three or two channels ) or 1Hz ( single channel ) using a sCMOS camera with 6 . 5mm pixel size ( pco . edge ) . The detection , tracking and analysis of all clathrin-labeled structures and thresholding to identify bona fide CCPs was done as previously described using the cmeAnalysis software package [22] . Briefly , diffraction-limited clathrin structures were detected using a Gaussian-based model method to approximate the point-spread function [22] , and trajectories were determined from clathrin structure detections using the u-track software [37] . Subthreshold clathrin-labeled structures ( sCLSs ) were distinguished from bona fide CCPs , based on the quantitative and unbiased analysis of clathrin intensity progression in the early stages of structure formation [22 , 62] . Both sCLSs and CCPs represent nucleation events , but only bona fide CCPs represent structures that undergo stabilization , maturation , and , in some cases , scission to produce intracellular vesicles [22 , 62] ) . We report the rate of bona fide CCP formation , distribution of their lifetimes , and intensity cohorts , as described previously [22] . As these values will depend on day-to-day variations in the threshold , we image experimental and control conditions on the same day and apply the same threshold to both data sets to ensure that effects we detect are due to the specific experimental variable being assessed . We also report mean and maximum signal intensities in two or three channels for each individual CCP . These are average and maximum signal intensities for individual CCPs as they are extracted by the previously described analysis software [22] . The extraction of CCPs is achieved by a new function added to the cmeAnalysis software published in [22] that allows us to link the classification of events , CCPs , or sCLSs to more sophisticated analysis intensity time courses and lifetime . In this study , we focused merely on per-CCP mean and maximum intensity values , which were averaged per movie ( 1–5 cells ) and finally presented as per-movie distributions covering 10–30 cells per experimental condition . Differences between conditions were assessed by comparison of the normal-distributed per-movie distributions using Student t test and a threshold of p < 0 . 01 to mark statistical significance . Control and treatment datasets were statistically analyzed with two-tailed , unpaired Student t tests using Graphpad Prism 5 . 0 ( Graphpad Software , La Jolla , CA ) , from which p values were derived ( * p < 0 . 05 , ** p < 0 . 01 , *** p < 0 . 001 , **** p < 0 . 0001 ) . Error bars representing standard error of the mean ( SEM ) for at least three independent experiments were calculated using Microsoft Excel . An in-cell ELISA approach was used to quantitate internalization of TfnR and EGFR , as previously described [23] , using either anti-TfnR mAb ( HTR-D65 ) [63] or biotinylated-EGF as ligands . Cells were grown overnight in 96-well plates at a density of 2 x 105 cells/well and incubated with 4 mg/ml of D65 or 20 ng/ml of biotinylated-EGF ( Invitrogen ) in assay buffer ( PBS4+: PBS supplemented with 1 mM MgCl2 , 1 mM CaCl2 , 5 mM glucose , and 0 . 2% bovine serum albumin ) at 37 °C for the indicated time points . Cells were then immediately cooled down ( to 4 °C ) to arrest internalization . The remaining surface-bound D65 or biotinylated-EGF was removed from the cells by an acid wash step ( 0 . 2 M acetic acid , 0 . 2 M NaCl , pH 2 . 5 ) . Cells were then washed with cold PBS and then fixed in 4% paraformaldehyde ( PFA ) ( Electron Microscopy Sciences ) in PBS for 30 min and subsequently permeabilized with 0 . 1% Triton X-100/PBS for 10 min . Internalized D65 was assessed using a goat anti-mouse HRP-conjugated antibody ( Life Technologies ) , and internalized biotinylated-EGF was assessed by streptavidin-POD ( Roche ) . The reaction was developed by a colorimetric approach with OPD ( Sigma-Aldrich ) , and color development was stopped by addition of 50 μl of 5M of H2SO4 . The absorbance was read at 490 nm ( Biotek Synergy H1 Hybrid Reader ) . Internalized ligand was expressed as the percentage of the total surface-bound ligand at 4 °C ( i . e . , without acid wash step ) , measured in parallel [23] . Well-to-well variability in cell number was accounted for by normalizing the reading at 490 nm with BCA readout at 560 nm . | Clathrin-mediated endocytosis ( CME ) , a major route for nutrient uptake , also controls signaling downstream of cell surface receptors . Recent studies have shown that signaling , in turn , can reciprocally regulate CME . CME is initiated by the assembly of clathrin-coated pits ( CCPs ) that mature to form deeply invaginated buds before the large Guanosine Triphosphate hydrolase ( GTPase ) , dynamin , catalyzes membrane scission and clathrin-coated vesicle release . Here , we characterize an isoform-specific and noncanonical function for dynamin-1 ( Dyn1 ) in regulating early stages of CME and show that Dyn1 and Dyn2 have nonredundant functions in CME . By genetically introducing fluorescent tags and using live-cell fluorescence imaging , we detected , tracked , and analyzed thousands of CCPs comprising up to three endocytic proteins in real time . We find that Dyn1 , previously assumed to function only at neurological synapses , is expressed but maintained in an inactive state in non-neuronal cells through phosphorylation by glycogen synthase kinase-3 beta ( GSK3β ) . We show that inhibition of GSK3β by a chemical inhibitor or downstream of epidermal growth factor receptor ( EGFR ) signaling activates Dyn1 and accelerates CCP assembly and maturation . These early effects are seen even when Dyn1 is barely detectable on CCPs . We conclude that Dyn1 is an important component of cross-communication between endocytosis and signaling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"phosphorylation",
"gene",
"regulation",
"enzymes",
"cell",
"processes",
"enzymology",
"light",
"microscopy",
"microscopy",
"cellular",
"structures",
"and",
"organelles",
"optics",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"proteins",
"fl... | 2018 | A noncanonical role for dynamin-1 in regulating early stages of clathrin-mediated endocytosis in non-neuronal cells |
Aging is the major risk factor for neurodegenerative diseases such as Alzheimer's disease , but little is known about the processes that lead to age-related decline of brain structures and function . Here we use RNA-seq in combination with high resolution histological analyses to show that aging leads to a significant deterioration of neurovascular structures including basement membrane reduction , pericyte loss , and astrocyte dysfunction . Neurovascular decline was sufficient to cause vascular leakage and correlated strongly with an increase in neuroinflammation including up-regulation of complement component C1QA in microglia/monocytes . Importantly , long-term aerobic exercise from midlife to old age prevented this age-related neurovascular decline , reduced C1QA+ microglia/monocytes , and increased synaptic plasticity and overall behavioral capabilities of aged mice . Concomitant with age-related neurovascular decline and complement activation , astrocytic Apoe dramatically decreased in aged mice , a decrease that was prevented by exercise . Given the role of APOE in maintaining the neurovascular unit and as an anti-inflammatory molecule , this suggests a possible link between astrocytic Apoe , age-related neurovascular dysfunction and microglia/monocyte activation . To test this , Apoe-deficient mice were exercised from midlife to old age and in contrast to wild-type ( Apoe-sufficient ) mice , exercise had little to no effect on age-related neurovascular decline or microglia/monocyte activation in the absence of APOE . Collectively , our data shows that neurovascular structures decline with age , a process that we propose to be intimately linked to complement activation in microglia/monocytes . Exercise prevents these changes , but not in the absence of APOE , opening up new avenues for understanding the complex interactions between neurovascular and neuroinflammatory responses in aging and neurodegenerative diseases such as Alzheimer’s disease .
As the general population ages , age-related diseases are on the increase . Data from the United States ( US ) Department of Health and Human Services show that more than 12 . 9% of the US population is over the age of 65 and this is expected to double by 2030 [1] . Aging is the major risk factor for many diseases including cancer , diabetes , heart disease , and Alzheimer’s disease ( AD ) [2] . Therefore , with an aging population , prevalence of age-related diseases is expected to increase . For instance , more than 5 million people in the US suffer from AD , and this is expected to exceed 10 million in the next 20 years [3] . In order to prevent and treat age-related neurodegenerative diseases , particularly AD , it is essential to better understand the factors that contribute to aging-induced susceptibility . In the healthy aged brain , functional and morphological changes lead to cognitive and sensorimotor control decline that affect the performance of activities of daily living ( ADL ) in older adults and increase vulnerability to the development of neurodegenerative conditions [4 , 5] . Age-related cognitive deficits are partly explained by changes in neural plasticity and synaptic activity . However , overall decline in brain health has also been correlated to other non-neuronal processes such as cerebrovascular dysfunction [6 , 7] and activation of innate immune responses [8] . Understanding how aging affects these processes will likely shed light on developing therapeutic strategies that prevent these age-related changes and decrease the vulnerability to neurodegenerative conditions . Interestingly , physical activity during aging has ameliorative effects on cognitive decline and sensorimotor deficits . Physically active older adults show improved performance in cognitive and sensorimotor tests and have greater brain volume in regions noticeably affected by age in sedentary subjects [4 , 9] . For instance , in humans , exercise enhances cerebral blood flow ( CBF ) , neurogenesis , and angiogenesis in the hippocampal dentate gyrus , increases hippocampal volume , and improves memory [10 , 11 , 12] . In addition , aerobic physical activity greatly improves cardiovascular function and decreases systemic inflammation in older subjects [8 , 9 , 13] . However , a detailed analysis of the processes involved has not been performed . In this study , we demonstrate significant changes in neurovascular integrity and function in the superior region of the cortex and hippocampus . We show a significant loss of pericytes and marked neurovascular decline , correlated with dysfunction of the blood brain barrier and activation of innate immune responses including the complement cascade . Strikingly , exercise almost completely prevented these age-related neurovascular changes and lessened complement induction in myeloid cells , but had little to no effect in the absence of APOE . We propose that exercise is an effective means of mitigating age-related neurovascular decline by directly or indirectly modulating Apoe-expressing astrocytes and/or C1QA-expressing myeloid cells .
To determine gene expression changes in response to normal aging , transcriptional profiling was performed on brain tissue from young ( 4 mo ) and aged ( 21 mo ) C57BL/6J ( B6 ) mice . Because portions of the limbic and higher-order integrative cortical areas ( including frontal association cortex , cingulate cortex , retrosplenial cortex , and the parietal associative cortices ) and the hippocampus are commonly impacted in AD [14] , brains were dissected into three separate regions for RNA sequencing and analysis: ( i ) region 1—frontoparietal cortex and corpus callosum ( FPC/CC ) , ( ii ) region 2—hippocampus ( HP ) , and ( iii ) region 3—rest of the cortex and brain stem ( RB ) . In total , 24 samples were profiled separately—three regions from four different mice from two age groups . To avoid batch effects , all RNA was prepared , and RNA-seq libraries generated at the same time . Samples were barcoded , pooled and sequenced across four lanes of an Illumina Hi-seq ( see Methods ) . For each of the three regions profiled , pairwise analyses comparing young samples to aged samples were performed to determine differentially expressed ( DE ) genes ( see Methods ) . A total of 1 , 045 genes ( 551 up-regulated and 494 down-regulated ) were DE in the FPC/CC ( region 1 , S1 Table ) , 644 genes ( 492 up-regulated and 152 down-regulated ) in the HP ( region 2 , S2 Table ) , and 1 , 137 ( 526 up-regulated and 611 downregulated ) in the RB ( region 3 , S3 Table ) ( Fig 1A ) . Gene set enrichment analysis ( using DAVID , see Methods ) showed significant differences between the regions , i . e . , DE genes from each of the three regions were present in overlapping but not identical Kyoto Encyclopedia of genes and genomes ( KEGG ) pathways ( S4 Table ) . This suggests that normal aging impacts the regions of the brain in different ways and may provide clues as to why specific brain regions are more susceptible to dysfunction in certain neurodegenerative diseases . Of particular interest were the KEGG pathways overrepresented in region 1 , tissue enriched for the FPC and the CC . These pathways included focal adhesion , vascular smooth muscle contraction , gap junction and extracellular matrix ( ECM ) -receptor interaction pathways ( Fig 1B ) . Genes in these pathways were generally down-regulated ( Fig 1C and 1D ) and suggest possible perturbation to the neurovascular unit . To assess possible neurovascular dysfunction in region 1 , intravascular and perivascular deposition of the plasma protein fibrinogen , or fibrin , a marker of vascular compromise , was analyzed by immunostaining . Previous studies have shown intra- and extravascular accumulation of fibrin ( ogen ) in the postmortem brains of AD patients as well as AD mouse models where severe neurovascular dysfunction occur [15 , 16 , 17 , 18] . In aged mice , small and sporadic deposits of extravascular fibrin were found only in the cortex , particularly the neocortex ( region 1 ) ( Fig 1E ) . We also noticed that contrary to young mice , intravascular accumulation of fibrin ( ogen ) was persistent in the brains of aging mice even after intracardial perfusion suggesting possible deposition of fibrin ( ogen ) in the luminal vessel wall as previously observed in human AD brains [15] . Quantification of fibrin ( ogen ) -immunostained area in the cortex , which included intra- and extravascular deposition of this protein , indicated a 3-fold increase in aged compared with young mice ( Fig 1F ) , confirming the transcriptional profiling data that predicted neurovascular dysfunction in the FPC in aged mice . The transcriptional profiling of the aged FPC/CC demonstrated significant down-regulation of extracellular matrix-associated genes as well as important genes for pericyte function suggesting possible disturbances in the interactions and function of the components of the neurovascular unit . To more fully characterize the extent of neurovascular dysfunction in the cortex of aged brains , the major components of the neurovascular unit , including basement membrane ( BM ) , endothelial cells , pericytes , and astrocytes were assessed . Collagen IV ( COL4 ) and laminin ( LAM ) , two major components of the BM ( generated by astrocytes , pericytes , and endothelial cells ) , were significantly reduced in the cortex of aged mice compared to young mice ( S1 and S2 Figs ) despite no decline in CD31 , a marker of endothelial cells ( S1 Fig ) . This suggests BM reduction was not caused by vascular reduction due to endothelial cell loss . Gene profiling showed three pericyte-related genes , Pdgfrβ , Act2 , and Cav2 , which were significantly down-regulated in aged compared to young mice ( Fig 1C ) , suggesting pericytes were negatively impacted by normal aging . Analysis of pericytes , using immunostaining for PDGFRβ showed a greater than 20% reduction in pericyte numbers and almost 50% reduction in pericyte coverage of microvessels in the cortex of aged compared to young brains ( Fig 2A–2C ) . Although transcriptional profiling data in the HP did not show pathways relevant to neurovascular unit dysfunction as being differentially affected by age ( S4 Table ) , genes relevant to the neurovascular unit , including Pdgfrβ ( −1 . 36 ) and Cldn5 ( −1 . 63 ) were down-regulated ( S2 Table ) , suggesting possible dysfunction to the neurovascular unit in the HP also . Therefore , we assessed neurovascular health in the CA1 region of the HP . Similar to our findings in the FPC , COL4+ microvessels and coverage of PDGFRβ+ were also significantly reduced in the hippocampal CA1 region of aged mice when compared with young ( S3 Fig ) . Furthermore , ultrastructural analysis identified several degenerating pericytes in the cortex of aged mice that were not observed in young mice ( Fig 2D ) . Pericytes are critical components of the neurovascular unit and play a key role in the regulation of BM , vessel contractility , and inhibition of vesicular transcytosis through endothelial cells [19] . Increased endothelial vesicular transcytosis was frequently observed in the aged cortex ( Fig 2E ) , further supporting age-related pericyte dysfunction or loss . Fibrin deposition and pericyte loss were associated with an increase in the number of activated microglia/monocytes ( Fig 3A and 3B ) . Ultrastructural analyses suggested that these microglia/monocytes were phagocytic when in close proximity to areas of pericyte degeneration and increased endothelial transcytosis activity ( Fig 3C ) . Pericytes can mediate the attachment of astrocyte endfeet to the vascular surface and regulate the polarization of specific proteins , such as the water channel aquaporin-4 ( AQP4 ) , to the perivascular endfoot region [20 , 21] . AQP4 is expressed by astrocytes that play a key role in the regulation of brain water transport at the neurovascular interface [22] . Immunostaining and immunoblotting showed a significant decrease of AQP4 protein at the neurovascular junctions in the cortex of aged compared to young mice ( Fig 4A and 4B ) . AQP4 decrease was accompanied by an increase in glial fibrillary acidic protein ( GFAP ) immunoreactivity , an indication of astrocyte reactivity ( Fig 4C ) . There was also a noticeable swelling of astrocyte endfeet with enlarged vacuoles in some cortical microvessels of aged mice but not from young mice ( Fig 4D ) . Collectively , our data suggest that normal aging causes significant dysfunction to the cortical neurovascular unit , including BM reduction and pericyte loss . These changes correlate strongly with an increase in microglia/monocytes in the aged cortex . Aging is generally accompanied by cognitive decline and sensorimotor deficits that affect the performance of ADL in the aged population [4 , 5] . Lifestyle choices such as exercise have been shown to have beneficial effects on the aging brain [9 , 13] , including increased brain volume [4 , 11] , improved performance in several cognitive and motor tasks [4] , and neuronal function [9] . Also , a recent study claimed a third of AD cases could be attributed in part to physical inactivity [23] . However , the impact of long-term physical exercise on the health of the neurovascular unit has not been determined . To assess this , mice were provided access to a running wheel from 12 months old ( equivalent to middle aged in humans ) and assessed at 18 months of age ( equivalent to early old age [~60 y old] in humans where risk of AD is greatly increased ) ( Fig 5A ) . Voluntary running was preferred to exercise by forced treadmill to remove any potential confounding effects of stress [24] . In addition , voluntary running in mice can induce adaptive physiological changes in cardiac and skeletal muscle showing it is a good method to assess biological changes as a result of exercise [25] . No differences in average running distance ( ~2 miles/night/mouse ) between the young ( 7 mo ) and aged ( 18 mo ) groups of mice were found after 6 months with the running wheel ( Fig 5B ) , indicating that aged mice were able to maintain their running capacity during the 6 months period . Exercised aged mice were first assessed for overall improvement in behavior and neuronal activity . Physical activity improves ADL in humans , and so common daily behaviors in mice—grip strength , nesting , and burrowing—were assessed in exercised and nonexercised ( sedentary ) mice ( Fig 5C–5E ) . Significant improvements in both grip strength ( Fig 5C ) and nesting behavior ( Fig 5D ) were observed in running aged mice ( 18 mo ) compared to sedentary aged mice ( 18 mo ) and were similar to levels seen in aging ( 12 mo ) mice . Burrowing also appeared to be improved but was not statistically significant ( Fig 5E ) . These results indicate that physical activity improved the capabilities and motivation of old mice to engage and perform typical spontaneous behaviors that seem to be affected by aging . These behavioral improvements induced by exercise were accompanied by stabilization of functional synapses and improved neural plasticity . While sedentary aged mice showed significant decreased levels of the presynaptic proteins such as synaptophysin and the postsynaptic protein PSD-95 ( S4 Fig ) , indicating a possible loss or weakening of functional synapses , running mice demonstrated significant preservation of synaptophysin when compared with younger mice ( S4 Fig ) . Neural plasticity was also evaluated in these mice by examining changes in the expression of the immediate early gene Arc ( activity-regulated cytoskeletal gene ) . Arc transcription is induced by neuronal activity [26] and is immediately up-regulated in the parietal cortex in response to spatial exploration [27] . Arc cortical expression was significantly elevated in aged running mice compared with aged sedentary mice after burrowing behavior ( Fig 5F and 5G ) , indicating that more neurons in the parietal cortex were recruited and activated during spatial exploration in the aged runner mice compared to aged sedentary mice . Next , the effects of exercise on the neurovascular unit were assessed . Exercise significantly reduced vascular leakage . Fibrin levels in aged running mice were similar to that seen in young mice and significantly less than aged sedentary mice ( Fig 6A and 6B ) . COL4 immunostaining in the cortex ( S5 Fig ) and CA1 ( S5 Fig ) demonstrated a significant preservation of the BM in aged runner mice when compared with aged sedentary mice ( S5 Fig ) . No changes were observed in the density of CD31+ microvessels between sedentary and runner mice ( S4 Fig ) , indicating that exercise prevented a deterioration of COL4 coverage in cortical and hippocampal CA1 microvessels during aging without an overall increase in vascular density . Aged runner mice also exhibited significantly higher numbers and vascular coverage of cortical PDGFRβ+ pericytes compared to aged sedentary mice , levels similar to those observed in younger sedentary mice ( Fig 6C and 6D ) . Finally , exercise preserved astrocytic AQP4 protein to similar levels observed in young mice ( Fig 6E and 6F ) and correlated with a decrease in astrocyte reactivity ( S6 Fig ) . In the hippocampal CA1 region , the reduction in vascular coverage of PDGFRβ+ pericytes and AQP4 levels observed in the aged sedentary mice were also prevented by exercise ( S7 Fig ) . Therefore , exercise prevented age-related pericyte loss , neurovascular unit decline , and vascular leakage . Induction of the complement cascade in the brain , particularly in myeloid-derived cells such as microglia/monocytes , has been shown to be a potentially damaging event in aging and disease [28 , 29 , 30 , 31] . Further , the classical pathway of the complement cascade has been strongly implicated in synaptic and neuronal dysfunction [32] , but only limited data are available on the role of complement in neurovascular dysfunction [33] . Analysis of our transcriptional profiling showed increased expression of multiple components of the classical pathway including C1qa ( +1 . 94 ) , C1qb ( +1 . 91 ) , C1qc ( +2 . 02 ) , C3 ( +2 . 71 ) , C4a ( +2 . 42 ) , C4b ( +4 . 63 ) , and C5ar2 ( +2 . 81 ) in the HP of aged compared to young mice ( see S2 Table ) . Since exercise prevented neurovascular dysfunction , synaptic decline , and behavioral deficits , we first assessed the impact of exercise on the number of microglia/monocytes in aged brains . Exercise significantly reduced the numbers of IBA1+ microglia/monocytes in the cortex of aged running mice compared to sedentary mice ( Fig 7A and 7B ) . In fact , the numbers of microglia/monocytes were inversely correlated with pericyte numbers ( Fig 7C ) , suggesting an important link between pericyte loss ( and neurovascular health ) and activation of microglia/monocytes . Next , to examine the effects of exercise on induction of the classical pathway of the complement cascade , we assessed C1qa , a component of the C1 complex , the initiating complex of the classical pathway . The number of C1qa+ microglia/monocytes was elevated 3-fold in the aged mice compared to young mice ( Fig 7D and 7E ) . Importantly , exercise caused a significant reduction in the number of cortical C1qa+ microglia/monocytes ( 35% less , Fig 7E ) . Similarly , in the hippocampal CA1 region , microglia/monocytes were increased in aged sedentary mice when compared with young mice , and this increase was prevented by exercise in aged running mice ( Fig 7F and 7G ) . These results suggest that the positive effects of exercise could be due in part to a reduction in complement activation during aging . To begin to determine possible mechanisms of the age-dependent vascular compromise , pericyte loss , and an increase in innate immune responses , transcript and protein expression levels of APOE were assessed in aging brains . APOE is a strong candidate to mediate age-related neurovascular decline for the following reasons . First , variations in human APOE are the major genetic risk factors for decreased longevity and AD [34 , 35 , 36] . Second , APOE has regulatory effects on cerebrovascular integrity and function through pericyte signaling [37 , 38] . Finally , APOE is an important mediator of innate immune responses [39 , 40 , 41] . Supporting a possible role for APOE in mediating age-related neurovascular dysfunction , Apoe transcript expression was noticeably decreased in the FPC ( Fig 8A and 8B ) and hippocampal CA1 ( Fig 8C and 8D ) in aged mice compared to young mice . Other studies have not reported an age-related decrease in APOE levels in the cortex and hippocampus [42] . Our data suggests this is because the decrease in Apoe transcript expression in the cortex was contrasted with a significant increase in Apoe transcript levels in the white matter regions , such as the CC in aged compared to young mice ( Fig 8E–8G ) . This led to no measurable changes by RNA-seq ( Fig 1 ) , qPCR ( 4 mo versus 24 mo , 2-ΔΔCt = 0 . 91 ) or western blotting ( Fig 8H ) . Interestingly , Apoe was only expressed by astrocytes and not neurons or microglial cells ( S8 Fig ) and may reflect a difference between Apoe expression in normal aging compared to injury or disease where others have reported expression in additional cell types such as microglia [36] . The decline in astrocytic Apoe in the cortex was in contrast to a second astrocytic apolipoprotein , ApoJ ( or Clusterin , Clu ) , where no significant changes were observed in aged compared to young brains ( S9 Fig ) . Therefore , astrocytic Apoe expression is dramatically altered in localized regions of the brain during normal aging . This regional difference in Apoe expression could be due to differences between the type of astrocytes that populate the region ( e . g . , white matter astrocytes are mostly fibrous , while astrocytes in the cortex are mostly protoplasmic ) or to different signals coming from the different environments [43 , 44] . Importantly , exercise preserved Apoe expression in the FPC ( Fig 9A and 9B ) and hippocampal CA1 ( Fig 9C and 9D ) , adding further support for a role of APOE in age-related pericyte loss and neurovascular compromise . To further investigate whether APOE contributes to neurovascular deterioration during normal aging and exercise-dependent preservation of the neurovascular unit , Apoe-deficient mice were exercised and compared to aged sedentary Apoe-deficient mice . Previous studies [37] , confirmed by our study ( S10 Fig ) , show neurovascular compromise including pericyte dysfunction or loss , vascular leakage , and BM reduction in young ( <12 months old ) Apoe-deficient mice that are strikingly similar to aged wild-type ( Apoe-sufficient ) mice . To determine the effect of exercise in the absence of APOE , Apoe-deficient mice were exercised from 12 months of age for 6 months , and components of the neurovascular unit assessed at 18 months of age . The average running distance performed by Apoe-deficient mice was not significantly different from aged Apoe-sufficient running mice ( S10 Fig ) . However , age-related deficits in grip strength and nest construction were not prevented by exercised Apoe-deficient mice ( S10 Fig ) . In the brain , exercise did not restore vascular leakage in Apoe-deficient mice ( Fig 10C and 10E ) . Exercise only partially restored PDGFRβ+ pericyte number in aged Apoe-deficient mice ( Fig 10D and 10F ) , but did not impact the density of COL4+ microvessels ( S10 Fig ) . Furthermore , exercise did not prevent the increase in microglia/monocytes in Apoe-deficient mice that was seen in wild-type mice ( Fig 10D–10G ) . Therefore , overall , exercise had little to no effect on age-related neurovascular dysfunction and microglia/monocyte numbers in Apoe-deficient mice .
Dysfunction of the neurovascular unit in the aging and aged brain is of great interest , since numerous studies have independently correlated the development of AD with vascular dysfunction during aging [2 , 6 , 45 , 46 , 47] , but the mechanisms involved are not known . Studies have shown deterioration of the cerebrovascular ultrastructure along with decreasing CBF and lower metabolic rates of glucose and oxygen in normal human aging [45] . Similarly , exercise has been shown to be beneficial for the brain [4 , 9 , 11 , 12 , 13 , 48] , but the processes positively impacted by exercise are not completely understood . Here , we elucidated the damaging effects of normal aging on the neurovascular unit in the cortex and HP of mice and show that exercise can prevent these detrimental changes ( Fig 11 ) . Transcriptional profiling predicted dysfunction of the neurovascular unit particularly in the FPC , with some relevant genes also DE in the HP . Other studies have also profiled aging brains ( e . g . , [49 , 50 , 51] ) but ours , to our knowledge , is the first to propose dysfunction of the neurovascular unit in aging mice . However , enrichment analyses of the DE genes comparing young and old male and female brains from the human study performed by Berchtold and colleagues ( supplemental tables 3 and 4 from [51] ) reveals overrepresentation of genes in multiple KEGG pathways relevant to the neurovascular unit ( including focal adhesion , vascular smooth muscle , and ECM-receptor interactions ) and neuroinflammation ( including the complement cascade ) . These pathways are strikingly similar to the pathways we have identified in our study ( Fig 1 ) providing compelling evidence that our findings in aging mice are directly relevant to human aging . Genes relevant to the neurovascular unit were more represented in the transcriptional profiles from the FPC compared to the HP , adding further support to the possibility that specific brain regions may be more or less susceptible to aging than others [51] . However , some genes relevant to the neurovascular unit were DE in the HP , and neurovascular decline in CA1 region of the HP was observed . An explanation for region-specific differences in the transcriptional profiles may be due to the compositions of the tissues being profiled . For instance , there are significant differences in vascular density between the cortex and the HP [45 , 52] . In the rat brain , the highest vascular density is found in the neocortex ( frontal and parietal regions profiled here ) where variations in vascular density were shown to impact metabolic and synaptic activity [52] . This could result in increased sensitivity to detect neurovascular dysfunction by transcriptional profiling in the FPC compared to the HP . Also , the density and reactivity of astrocytes ( compared to neurons ) is different in the cortex compared to the HP with hippocampal astrocytes constitutively expressing GFAP , even in healthy brains [44 , 53] . This could also explain in part why , in our study , molecular pathways associated with neuroinflammation were more represented in the HP . Although others have also proposed region-specific changes in response to aging [51] , further work is required to fully determine the extent of region-specific susceptibility in the aging brain . Multiple changes to the neurovascular unit in the FPC and HP were observed in aged mice , one of the most striking being a significant loss of pericytes and their vascular coverage . While it is not known whether aging impacts pericytes directly or indirectly , pericyte function is critical for vessel stability , blood flow , and blood–brain barrier ( BBB ) integrity and function [19 , 54] . Loss of pericytes can cause BBB breakdown ( by increasing endothelial transcytosis ) , reductions in BM , and alterations on AQP4 localization , correlating with an increase in innate immune responses [20 , 55 , 56] , features observed in aged mice . Sporadic perivascular and more frequent intravascular deposits of fibrin , and increased vesicular transcytosis in endothelial cells in the aged cortex suggests vascular compromise is possibly occurring as a consequence of pericyte loss and/or dysfunction . Normal aging in humans is also associated with increased vascular compromise [57 , 58] . In fact , a recent study , using a new high resolution MRI method to map the blood-to-brain transfer constant of gadolinium ( Ktrans ) regionally and quantitatively , found that vascular leakage is an early event in the aged human brain that begins in the HP and is correlated with cognitive decline [58] , supporting our findings in the aged mouse brain . Moreover , discontinuous and dysregulated coverage of BM on cortical microvessels and reduction of AQP4 protein levels at the neurovascular interphase were also observed in aging mice , suggesting that astrocyte–endothelial cell interactions are altered in the aged brain . It has been proposed that pericytes regulate the interactions of astrocytic endfeet with the vasculature [20] , and age-related pericyte dysfunction and degeneration could lead to the disruptions of astrocyte-vascular interactions . Growing evidence supports age-related vascular dysfunction as a major contributing factor in the onset and progression of AD [6 , 7 , 45] . Extensive reductions in the number of pericytes have been correlated with BBB breakdown in postmortem AD brains [59] , and intravascular and perivascular deposition of fibrin has been also found in AD mice [15 , 17] and AD patients [15 , 18 , 60] . Our data suggest that signaling changes between astrocytes , pericytes , and endothelial cells are impacted by aging , underpin neurovascular dysfunction , and may contribute to neurodegenerative diseases such as AD . Our data also showed a direct correlation between pericyte loss , neurovascular decline , and an increase in C1qa ( an initiating molecule of the classical pathway of the complement cascade ) in response to aging . One key role of the neurovascular unit is to control the infiltration of peripheral immune cells ( inflammation ) and to protect the brain against blood-derived protein extravasation [45] . However , it is not clear whether the increase in C1qa+ microglia/monocytes in aging mice is as a result of the proliferation of resident microglia or the infiltration of peripheral monocytes . During aging , a shift from an anti-inflammatory to a proinflammatory state can occur in the brain leading to cognitive decline , suggesting innate immune responses are an important characteristic of aging [8] . Conversely , an age-related increase of proinflammatory markers in the blood is correlated with poor cognitive performance in older adults [8] , suggesting that age-related cognitive decline could be triggered in part by age-related systemic inflammation . Increased levels of C1qa have been identified in the brains of 12 mo mice [30] , indicating that proinflammatory processes are detected early in aging brains . Moreover , age-related cognitive decline in mice is prevented by genetic deletion of C1qa supporting the deleterious effect of this protein and proinflammatory processes in brain function during aging [30] . Further work is required to determine whether peripheral complement-expressing monocytes infiltrate into the aging brain and whether the roles of complement-expressing resident microglia and/or peripheral monocytes are beneficial , damaging , or both . It is also not clear whether complement induction precedes neurovascular decline or occurs as a consequence of it . If the entry of peripheral immune cells into the brain is a key driver of age-related neurovascular decline , targeting peripheral immune cells before they enter the brain may be a feasible route to combat age-related cognitive decline and neurodegenerative diseases such as AD . A major finding of this study is that exercise prevents neurovascular unit decline , including the reduction of age-related fibrin deposition and preservation of pericytes . The positive impact of exercise on neurogenesis , angiogenesis , neuronal activity , and cognition have been well demonstrated [4 , 9 , 11 , 13 , 61] , but little is known about the effects of long-term exercise on cellular and molecular interactions of the neurovascular unit and other glial cells during aging . In our study , long-term exercise during aging prevented deficits in spontaneous and common behaviors in mice that are similar to ADL in humans , which are also affected by aging and improved by physical activity [62 , 63 , 64 , 65] . For instance , changes in grip strength and physical frailty in humans have been strongly associated with decline in cognitive performance [66 , 67 , 68] . It has been proposed that common biological processes mediate age-related decline in physical and cognitive function [67] , suggesting that exercise in aging could be targeting and preventing declines in these common processes . Our data support this hypothesis by demonstrating that long-term exercise in aging mice prevents age-related physical frailty and behavioral deficits along with enhancements in brain structure and function . While it is not yet known whether exercise-induced pericyte protection is direct or indirect , the importance of pericytes in the neurovascular unit underscores the potential of exercise to ameliorate age-associated degenerative changes and preserve long-term neurovascular health . Interestingly , previous reports found that levels of peripheral circulating platelet-derived growth factor subunit B ( PDGFB ) , an essential growth factor for pericytes derived from endothelial cells [69] , are significantly reduced with age [70] , but increased by exercise [71] . The effects of peripheral growth factors on the cerebrovascular system have been demonstrated before . For instance , circulatory IGF1 is necessary for exercise-induced angiogenesis [72] in the young brain , while blood-derived GDF11 induces angiogenesis in the aged subventricular zone during heterochronic parabiosis [73] , suggesting that circulatory factors could be mediating exercise-induced changes in the cerebrovascular system . Further studies are necessary to determine the possible contribution of circulatory factors to exercise-mediated improvements of cerebrovascular health . Another important finding in this study was the significant reduction in the number of aged-induced proinflammatory C1qa+ microglia/monocytes in exercised mice . Induction of C1qa in the CNS promotes synapse elimination and synaptic dysfunction during development , neurodegenerative diseases such as glaucoma and AD , and aging [29 , 30 , 31 , 32 , 74] . Importantly , the positive changes found in exercised mice were accompanied by stabilization of synaptic proteins and improvements in neuronal plasticity along with behavioral enhancements when compared with age-matched sedentary mice , and could be a direct result of a lessening of complement activation . Increased levels of C1QA are found in aged human brains suggesting a possible contribution of the complement cascade to age-related cognitive decline [30] , but the impact of exercise on complement activation in the human brain has not been studied . However , it is well established that exercise in humans increases hippocampal volume and blood flow , induces greater levels of electrical and synaptic activity , and improves memory and cognitive performance in older adults that exercised [9 , 11 , 12] . This further highlights the importance of understanding the mechanisms by which aging and lifestyle directly or indirectly affect the function of astrocytes , pericytes , and microglia/monocyte . Determining the mechanisms by which aging leads to neurovascular decline and how exercise prevents this decline is important , as it could lead to the identification of therapeutic strategies that target similar processes . Here we suggest APOE as a strong candidate for mediating age-related neurovascular unit decline; Apoe expression decreases in the cortex and HP of aged mice , Apoe expression is preserved by exercise , and exercise has little to no effect on behavioral deficits , neurovascular dysfunction , and innate immune responses in aged Apoe-deficient mice . Given that mice deficient in Apoe show vascular permeability , decreased CBF , synapse loss , and cognitive impairments [37 , 75 , 76] , a decrease in Apoe expression in the aging brain would be predicted to impact the health of the neurovascular unit . Supporting this , a previous study shows an activation of a proinflammatory pathway in pericytes in Apoe-deficient mice leading to BM dysregulation and neurovascular breakdown [37] . Interestingly , although exercise had no effect on most components of the neurovascular unit in Apoe-deficient mice , exercise did partially increase pericyte number , but not coverage , indicating exercise has at least some positive effects on pericyte survival independent of APOE but they could be still dysfunctional . Another possibility is that intensity and duration of the exercise in this study was not sufficient for Apoe-deficient mice to reach the beneficial effects found in the wild-type mice . It is important to note that Apoe-deficient mice already had a dysfunctional neurovascular unit prior to running in contrast to the experiments using wild-type mice . Therefore , it is possible that exercise prevents neurovascular dysfunction but it has little to no effect on mice with an already dysfunctional neurovascular unit . This outcome would also be important , as neurovascular dysfunction from events such as brain injury or strokes could predispose to neurodegenerative diseases such as AD , and exercise may not be as beneficial in these circumstances . Further work is needed to fully explore the role of Apoe in age-related neurovascular decline . In humans , APOE has three alleles that encode three different isoforms of the protein APOE2 , APOE3 , and APOE4 that differ by a single amino acid substitution [77] . Human APOE alleles impact neurovascular function to different extents in mice with mice carrying human APOE4 showing altered neurovascular function from a young age compared to mice carrying human APOE2 or human APOE3 [37] . Similar to Apoe deficiency , APOE4 transgene in mice induces the activation of the proinflammatory cyclophilin A ( CypA ) –matrix metalloproteinase 9 ( MMP-9 ) pathway in brain pericytes , leading to the breakdown of the BBB and neurodegeneration [37 , 38] . This proinflammatory pathway is also activated in the cerebrospinal fluid ( CSF ) of cognitively normal APOE4 carriers [78] and in the microvessels of postmortem brains from AD patients [79] . Furthermore , increased pericyte degeneration and BBB breakdown are found in brains of APOE4 carriers with AD supporting the role of APOE4 in neurovascular dysfunction [79] . Interestingly , in humans and transgenic mice , APOE4 carriers have lower levels of APOE protein in the brain compared to noncarriers [80 , 81] , and human patients carrying APOE4 develop AD 8–15 y earlier than carriers of APOE2 or APOE3 [38 , 62] . However , the APOE4 isoform also contributes to AD progression by binding with higher affinity to amyloid-β peptide and disrupting its clearance from the brain [82] , and by contributing to amyloid-β aggregation [83] . Our findings suggest that deficiency or reduced efficiency of APOE in normal aging could cause dysfunction of neurovascular health and increased neuroinflammation , contributing to AD susceptibility , onset , and progression . Experiments to fully explore the role of APOE in age-related neurovascular decline and neuroinflammation , including conditional ablation of Apoe in aging mice and the impact of multiple human isoforms of APOE , are underway . In summary , here we show that exercise in aging mice preserves the integrity and function of the neurovascular unit leading to a healthier and “younger-like” brain . In recent years , there has been a distinct lack of success in developing therapies that target specific components of AD such as plaque deposition and excessive Tau phosphorylation . Our data , supported by data from human studies [4 , 9 , 23 , 84] , point towards focusing efforts on understanding the impact of aging and lifestyle on neurovascular unit decline and neuroinflammation , particularly pericyte dysfunction and loss , and activation of innate immune responses . Understanding these processes will both help encourage a healthy lifestyle that where possible includes exercise and could lead to the development of improved treatments for AD and other neurodegenerative disorders .
All experiments involving mice were conducted in accordance to policies and procedures described in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and were approved by the Animal Care and Use Committee at The Jackson laboratory . All mice were bred and housed in a 12/12-hour light/dark cycle . As part of the Nathan Shock Center of Excellence in the Basic Biology of Aging at The Jackson Laboratory , cohorts of C57BL/6J mice were aged to between 12 and 24 months of age . All mice used in this study were females C57BL/6J ( B6 ) ( Stock number 00664 ) . Female B6 mice homozygous for the Apoetm1Unc mutation ( Apoe-/- ) were obtained from The Jackson Laboratory ( Stock Number 002052 ) . To produce experimental animals , Apoe-/- mice were intercrossed and aged . For this study , mice under 9 months old were classified as “young” mice , while 12-month-old mice were classified as “aging” mice , and 18–24 month old mice were identified as “aged” mice . For the parameters analyzed in this paper , no significant differences were found between 18- and 24-month-old mice . Mice were anesthetized with a lethal dose of ketamine/xylazine and transcardially perfused with 1X PBS to remove any trace of blood from the brain . After perfusion , mice were decapitated with brains carefully dissected out and hemisected in the midsagittal plane . For RNA-seq , the superior region of the cortex ( region 1 ) , the HP ( region 2 ) , and the RB ( region 3 ) were removed and immediately snap-frozen . For other histological procedures , one-half of the hemisected brain was snap-frozen and the other half was fixed by immersion in 4% paraformaldehyde for two nights . After fixation , brains were rinsed in 1X PBS , immersed in 30% sucrose/PBS solution overnight at 4°C , frozen in OCT , and cryosectioned at 20 μm . Tissues were homogenized with TRIzol reagent ( Life Technologies ) and centrifuged to remove debris . Chloroform ( 0 . 2 ml per 1 ml of TRIzol ) was added to the cleared homogenate for phase separation . Total RNA was purified from the aqueous layer using the QIAGEN miRNeasy mini extraction kit ( QIAGEN ) according to the manufacturer’s instructions . RNA quality was assessed with the Bioanalyzer 2100 ( Agilent Technologies ) . Poly ( A ) selected RNA-seq sequencing libraries were generated using the TruSeq RNA Sample preparation kit v2 ( Illumina ) and quantified using qPCR ( Kapa Biosystems ) . Using Truseq V4 SBS chemistry , all libraries were processed for 125 bp paired-end sequencing on the Illumina HiSeq 2 , 500 platform according to manufacturer’s instructions . Samples were subjected to quality control analysis by NGSQCtoolkit v 2 . 3 [85] . Reads with 70% of their bases having a base quality score ≥ 30 were retained for further analysis . Read alignment and expression estimation was performed using RSEM v 1 . 2 . 12 [86] with supplied annotations at default parameters against the C57BL/6J mouse genome ( build-mm10 ) . Bamtools v 1 . 0 . 2 [87] were used to calculate the mapping statistics . Differential gene expression analysis between groups was performed using EdgeR v 3 . 8 . 5 [87 , 88 , 89] following the removal of outlier samples and lowly expressed genes ( cpm [or counts per million] < 1 in less than two samples ) . Normalization was performed using the trimmed mean of M values ( TMM ) . Adjustment for multiple testing was performed using FDR . Genes were considered to be significantly DE at a FDR < 0 . 05 . All raw and processed data is being made available through GEO archives and S1–S3 Tables . Official ENSEMBL gene IDs for all genes included in this study are provided in column 1 of S1 , S2 and S3 Tables . The Database for Annotation , Visualization and Integrated Discovery ( DAVID , [89] ) was used to add functional annotation to DE gene lists and provide statistical assessment of the annotations . We focused on the KEGG pathways [90] . For each list of DE genes , DAVID determines the number of genes in each KEGG pathway and uses a Fisher exact test to determine the probability that the number of genes in each pathway would have occurred by chance [89] . A p-value < 0 . 05 was used to identify significant pathways . Pathways with gene expression changes represented by colors ( red–up-regulated , green–down-regulated ) were generated at the KEGG website [90] . For immunostaining involving antibodies for vascular associated proteins , sections were pretreated with Pepsin as previously described [91] with minor modifications . Sections were hydrated with H2O for 3 min at 37°C followed by treatment of the tissue with 0 . 5 mg/ml of Pepsin ( Sigma-Aldrich ) for 18 min at 37°C . Sections were then rinsed twice with 1X PBS at room temperature ( RT ) for 10 min . After Pepsin pretreatment , sections were rinsed once in 1X PBT ( PBS + 1% Triton 100X ) and incubated in primary antibodies diluted with 1X PBT + 10% normal goat or normal donkey serum over two nights at 4°C . After incubation with primary antibodies , sections were rinsed three times with 1X PBT for 10 min and incubated for two hours in the corresponding secondary antibodies ( 1:800 , Invitrogen ) . Tissue was then washed three times with 1X PBT for 10–15 min , incubated with DAPI and mounted in Poly aquamount ( Polysciences ) . The following primary antibodies were used: goat anti-COL4 ( 1:40 , R&D ) , goat anti-PDGFRβ ( 1:40 , R&D ) , goat anti-CD31 ( 1:40 , R&D ) , rabbit anti-LAM ( 1:200 , Sigma-Aldrich ) , goat anti-mouse APOE ( 1:50 , Santa Cruz Biotech ) , Biotinylated Lycopersicon Esculentum ( Tomato ) Lectin ( 1:200 , Vector ) , rabbit anti-IBA1 ( 1:200 , Wako ) , rabbit anti-GFAP ( 1:200 , Dako ) , rabbit anti-AQP4 ( 1:200 , Sigma-Aldrich ) , mouse anti-synaptophysin ( SYN , 1:200 , Millipore ) , mouse anti-NeuN ( 1:300 , Millipore ) , rabbit anti-fibrinogen ( FIBRIN , 1:200 , DAKO ) . Blocking serum was not included in primary antibody solutions that contained the fibrinogen antibody . For quantitative analysis of FIBRIN and SYN in the cortex , four images were randomly taken in the parietal cortex for each brain for each mouse and opened in ImageJ ( 1 . 47 d ) software as a black and white image as reported previously [92] . Stained intensity and % of area statistics were obtained by generating surface segmentation using the same threshold criteria for all the pictures . For quantification of PDGFRβ+ or IBA1+ cells in the cortex and CA1 , four images were taken for each brain from each mouse with a Zeiss Axio Imager fluorescent microscope and manually counted using the cell counter plugin from the ImageJ ( 1 . 47 d ) software . For quantification of pericyte coverage of microvessels , the length PDGFRβ+ pericytes were calculated using the NeuronJ plugin in ImageJ ( 1 . 47d ) software . All image analyses were performed blind to the experimental conditions . For quantification of COL4+ and CD31+ capillary area in the cortex and CA1 , six images were taken for each brain from each mouse with a Zeiss Axio Imager fluorescent microscope . Vessel area was quantified using an automated method developed in-house , available upon request . Briefly , a segmentation algorithm , modified from the ImageJ plugin VNT ( Vascular Network Toolkit , http://ntwrkanlystlkit . sourceforge . net/ ) , was used for measure blood vessels area on 20x magnification images that were stained with the vascular markers COL4 , LAM , CD31 , and Lectin . This segmentation algorithm was written as an ImageJ macro for automated processing of images that includes the following steps: Gaussian blur ( 2 px ) , “Find edges” , Variance ( 5 px ) , Median ( 3 px ) , Subtract ( σ px ) , Multiply ( 255 ) , Invert and Analyze Particles ( S11 Fig ) . Image analysis was automated and blind to the experimenters . This automated processing had a positive correlation with quantification of vessel area by manual tracing ( in ImageJ ) performed by two independent investigators ( R2 = 0 . 612 , p = 1 . 24e-07 ) ( S11 Fig ) . For in situ hybridization experiments , RNA probes for mouse Apoe ( GE Dharmacon Clone ID 5136415 ) , Arc ( GE Dharmacon Clone ID 3498057 ) , Clu ( GE Dharmacon Clone ID 30550773 ) , and C1qa ( GE Dharmacon Clone ID 3592169 ) were synthesized , labeled with digoxigenin ( Dig ) , and hydrolized by standard procedures . Frozen sections were postfixed ( 4% PFA for 5 min ) , rinsed twice with 1X PBS , and acetylated with 0 . 25% acetic anhydride for 10 min in 0 . 1 M triethanolamine ( TEA ) . Sections were then washed in PBS and incubated overnight at 65°C with hybridization solution ( 50% formamide , 1X Hybe solution [Sigma-Aldrich] , 1 mg/ml yeast RNA ) containing 1 μg/ml Dig-labeled riboprobe . After hybridization , sections were washed by immersion in 0 . 2XSSC ( Sodium Chloride- Sodium Citrate ) at 72°C for 1 hr . Dig-labeled probes were detected with an AP-conjugated anti-Dig antibody ( Roche ) followed by NBT/BCIP ( nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate ) reaction ( Roche ) . After in situ hybridization , sections were incubated in primary antibodies ( GFAP , NEUN and IBA1 ) as previously described [29] and briefly described above . Sections were then incubated with DAPI for nuclei staining and mounted in Aqua PolyMount ( Polysciences ) . For quantitative analysis of Apoe and Clu RNA expression , the % area stained by the riboprobes in the cortex and CA1 was measured with ImageJ ( 1 . 47 d ) software as reported previously [92] . Briefly , six images were taken in the parietal cortex for each brain from each mouse and opened in Image J as a black and white image . All images were converted to one stack and cropped to obtain just the center of the picture . Stained area statistics were obtained by generating surface segmentation using identical threshold criteria for all the pictures . All image analyses were performed blind to the experimental conditions . For quantification of Arc+ or C1qa+ cells in the cortex , four images were taken from each brain for each mouse in the transmission light channel and the fluorescence channel for fluorescently-labeled NEUN or IBA1 and manually counted using the cell counter plugin from the ImageJ ( 1 . 47 d ) software . Mice were perfused with a mix solution of 2% PFA and 2% glutaraldehyde in 0 . 1 M Cacodylate buffer ( Caco ) . After perfusion , brains were fixed overnight in the same solution at 4°C . The samples were then rinsed three times in 0 . 1 M Caco buffer . 100 μm thick coronal sections were cut on a vibrating microtome and post fixed with 2% osmium tetroxide in 0 . 1 M Caco buffer for 2 h at room temperature . Sections were rinsed three times with Caco buffer and then dehydrated through of series of alcohol gradations . The sections were then put into a 1:1 solution of propylene oxide/Epon Araldite ( Electron Microscopy Sciences , Hatfield , PA ) overnight on an orbital rotator . Sections were then flat embedded with 100% Epon Araldite between 2 sheets of Aclar film and polymerized at 65°C for 48 h . Specific areas of the parietal cortex in these sections were then selected , cut out with a razor blade and glued onto dummy blocks of Epon Araldite . 90 nm ultrathin sections were cut on a Diatome diamond knife , collected on 300 mesh copper grids , and stained with Uranyl Acetate and lead citrate . Grids were viewed on a JEOL JEM1230 transmission electron microscope and images collected with an AMT high-resolution digital camera . Twenty cross-sectional blood vessels were imaged per brain/mouse , n = 3 per group young ( 4 mo ) and aged ( 18 mo ) group . Hemisected brains were dissected as described above and the superior region of the cortex containing the parietal cortex was sliced , snap frozen at the time of collection and stored at −80°C . RNA extraction was performed according to the TRIzol ( Invitrogen ) manufacturer’s instructions . Briefly , tissue was homogenized in 1ml of TRIzol reagent per 50–100 mg of tissue sample , followed by separation of phases using chloroform and removal of the aqueous phase for RNA precipitation . The interphase and organic phenol-chloroform phase were saved for protein extraction . For protein isolation , 100% ethanol was added to the interphase/phenol-chloroform phase , centrifuged and the phenol-ethanol supernatant was taken . Protein was precipitated from the saved supernatant with isopropanol , washed with 0 . 3 M guanidine hydrochloride in 95% ethanol and resuspended in a 1:1 solution of 8 M urea ( in Tris-HCl 1 M , pH 8 . 0 ) and 1% SDS using sonication as described previously [93] . Isolated RNA and protein were stored at −80°C until use . Trizol-extracted RNA was used to assess the expression levels of mouse Apoe in young ( n = 4 ) and aged ( n = 4 ) cortex , with all samples normalized to β-Actin . RNA was treated with DNAse and reverse transcribed using the GeneAmp RNA PCR kit ( Applied Biosystems ) . For quantitative PCR , the Quanti-Fast SYBR Green kit ( Qiagen ) was used and reactions were carried out with the following primers: Apoe ( Forward: 5’-GGGCAAACCTGATGGAGAAG-3’ and Reverse: 5’-CCTGGCTGGATATGGATGTTG-3’ ) ; and β-Actin ( Forward: 5’-TGGAATCCTGTGGCATCCATGAAAC-3’ and Reverse: 5’-TAAAACGCAGCTCAGTAACAGTCCG-3’ ) , with each gene being interrogated in triplicate . Ct ( threshold cycle ) was calculated as the mean of the successful replicates for each gene . For normalization , ΔCt values were calculated as Ct ( gene of interest ) minus geometric mean of Ct for the normalizers . The average and standard deviation of the young cortex were then calculated . The ΔΔCt was defined as the ΔCt ( gene of interest ) minus ΔCt ( control mean ) . The fold change was calculated as 2 minus ΔΔCt ( up-regulated ) or –2ΔΔCt ( down-regulated ) . A gene was considered DE if the fold change was greater than two standard deviations away from the mean fold change from the young cohort . Protein samples were separated by SDS-PAGE gel electrophoresis and transferred to nitrocellulose membrane . Before incubation with primary antibodies , membranes were blocked in 5% milk , and after primary antibody incubation the appropriate peroxidase-conjugated antibody ( Millipore ) was used as a secondary antibody . For detection , membranes were treated with the Amersham ECL western blotting analysis system ( GE Healthcare ) and exposed to the High performance chemiluminescence film ( GE Healthcare ) . The primary antibodies used for immunoblotting are: goat anti-mouse APOE ( 1:1 , 000 , Millipore ) , rabbit anti-pan laminin ( LAM , 1:1 , 000 , Sigma-Aldrich ) , mouse anti-PSD95 ( 1:1 , 000 , Millipore ) , and mouse anti-βActin ( 1:2 , 000 , Abcam ) . Fore limb strength was assessed by the suspended grid-grasping test . Mice were timed for how long they can support their body weight by holding onto a metal mesh suspended in midair . One minute was established as the maximum time for the test . Nest construction was assessed as reported previously [94 , 95] . Briefly , singly housed mice were provided with a preweighted nestlet one hour before the dark cycle . The next morning , the nest construction was assessed following the 1 to 5 scoring method established by the Deacon lab ( 2012 ) . To evaluate burrowing , individually caged mice were provided with a one open ended 200 mm long and 70 mm diameter polyvinyl chloride ( PVC ) plastic tube filled with 200 g of mouse food pellets as described previously [94] . The open end was elevated 3 cm off the bottom of the cage with machine screws ( 5 cm long ) . The mice were allowed to burrow for 2 hr , and the amount of food pellets that remained in the tube was calculated . All behavioral tasks were performed at least three times and the average calculated . For voluntary wheel running , mice were given free access to running saucer wheels ( Innovive Inc ) ( day and night , two mice per cage ) . Sedentary mice had no access to wheels . Both groups were housed under these conditions for 6 mo . Animals were tested for running capacity by placing individual mice in a cage with a wireless saucer wheel ( ENV-044 Med Associates Inc . ) for 10 d . Data were collected nightly ( 16 hr ) , analyzed and average distance ran per night for each mouse calculated . Data were analyzed using GraphPad Prism software . Significance was calculated using unpaired t tests for comparisons between two groups and one-way multifactorial analysis variance ( ANOVA ) followed by Tukey posthoc tests for multiple comparisons . p-values are provided as stated by GraphPad Prism software and significance was determined with p-values less than 0 . 05 . | Aging is frequently accompanied with frailty and cognitive decline . In recent years , increasing evidence has linked physical inactivity with the development of dementias such as Alzheimer’s disease . In fact , it is recognized that exercise combats frailty and cognitive decline in older adults , but the biological mechanisms involved are not completely known . Understanding the biological changes that trigger cognitive deterioration during aging and the mechanisms by which exercise improves health and brain function is key to ensuring the quality of life of the elderly population and to reducing risk of dementias such as Alzheimer’s disease . Here , we show that the cerebrovascular system in mice significantly deteriorates with age , and the structure and function of the blood brain barrier is progressively compromised . These age-related neurovascular changes are accompanied by neuroinflammation and deficits in common and spontaneous behaviors in mice . We found , however , that exercise from middle to older age preserves the cerebrovascular health , prevents behavioral deficits and reduces the age-related neuroinflammation in the cortex and hippocampus in aged mice . Mice deficient in Apoe , a gene associated with longevity and Alzheimer’s disease , are resistant to the beneficial effects of exercise , suggesting a possible mediating role for APOE in the maintenance and function of the neurovascular system during aging . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | APOE Stabilization by Exercise Prevents Aging Neurovascular Dysfunction and Complement Induction |
Visceral leishmaniasis is a major neglected tropical disease , with an estimated 500 , 000 new cases and more than 50 , 000 deaths attributable to this disease every year . Drug therapy is available but costly and resistance against several drug classes has evolved . Despite all efforts , no commercial , let alone affordable , vaccine is available to date . Thus , the development of cost effective , needle-independent vaccines is a high priority . Here , we have continued efforts to develop live vaccine carriers based on recombinant Salmonella . We used an in silico approach to select novel Leishmania parasite antigens from proteomic data sets , with selection criteria based on protein abundance , conservation across Leishmania species and low homology to host species . Five chosen antigens were differentially expressed on the surface or in the cytosol of Salmonella typhimurium SL3261 . A two-step procedure was developed to select optimal Salmonella vaccine strains for each antigen , based on bacterial fitness and antigen expression levels . We show that vaccine strains of Salmonella expressing the novel Leishmania antigens LinJ08 . 1190 and LinJ23 . 0410 significantly reduced visceralisation of L . major and enhanced systemic resistance against L . donovani in susceptible BALB/c mice . The results show that Salmonella are valid vaccine carriers for inducing resistance against visceral leishmaniasis but that their use may not be suitable for all antigens .
The leishmaniases are regarded as neglected tropical diseases . The causative protozoan parasites are transmitted through the bite of sandfly vectors . Currently an estimated 12 million people are infected , while 350 million people in 88 countries worldwide are at risk to develop one of the diseases associated with Leishmania parasites ( http://www . who . int/leishmaniasis/burden/en/; [1] ) . The most severe form is visceral leishmaniasis ( VL; also known as kala azar in India ) a disease that is fatal if untreated . An estimated 500 000 new cases and 50 000 deaths are reported every year , with 90% occurring in Bangladesh , Nepal , India , Sudan , Ethiopia and Brazil ( [2] ) . VL caused by L . infantum/chagasi is zoonotic with dogs being the main reservoir; however , in areas endemic for L . donovani ( e . g . India and Sudan ) the disease is anthroponotic . In many cases infection remains asymptomatic , most likely indicating immune control . However , patients with symptomatic VL experience fever , fatigue , weight loss and weakness often accompanied by hepato-splenomegaly and anaemia and , if untreated , may die from bacterial co-infections , internal bleeding and anaemia ( reviewed by ( 2] ) . Chemotherapy is available , but due to high toxicity , adverse side effects and emerging parasite resistance , treatment options are limited [3]–[6] . Long treatment regimens and associated costs are additional critical factors preventing patient access and compliance . For example paromomycin , a newly registered drug , is given by intra-muscular injections over a period of 21 days . Though the cheapest drug available , treatment still costs between 5 and 10 US$ per course , making this drug too expensive in relation to household income [5] . This economic burden of treatment is likely to remain for the foreseeable future . Thus , developing a vaccine for VL ( and indeed for other forms of leishmaniasis ) is high on the agenda of the World Health Assembly ( resolution EB118 . R3 , Geneva 05/07 ) . Vaccination is considered possible because of the efficacy of the century-old practice of leishmanization against old world cutaneous leishmaniasis ( CL ) , a treatment that affords life long protection as proven during its large scale use to protect military personnel in Israel , Iran and the former Soviet Union [7]–[9] . However , in some individuals , development of non-healing lesions , exacerbation of chronic disease and immunosuppression as a result of this procedure has been observed [10] . The unsatisfactory safety profile , its questionable efficacy against infection with heterologous species and logistic hurdles render leishmanization problematic . Vaccines that relied on autoclaved or merthiolate-killed whole promastigotes formulated with or without Bacillus Calmette-Guerin as adjuvants were developed to remedy some of the shortcomings of leishmanization but a recent meta-analysis of clinical studies evaluating these vaccines did not support their efficacy [11] . Clinical testing of vaccines based on recombinant Leishmania antigens or fractionated parasite material is much less advanced , although numerous antigenic proteins have been shown to have vaccine potential in pre-clinical models ( see reviews by [12]–[14] ) . These antigens were usually discovered by classical approaches , i . e . by screening with immune or hyperimmune sera from patients or infected animals . Antibody reactivity may not be an ideal criterion since protection is cell mediated and is thought to depend on both CD4+ and CD8+ T lymphocytes [15]–[18] . More recently , however , parasite genome information has become available and vaccine-antigen discovery exploiting this information has been promoted [19] . Recombinant DNA technology enables the formulation of subunit vaccines consisting of one or few specified antigens as DNA- and vectored vaccines , the latter exploiting viruses or bacteria as vaccine vehicles ( summarised by [12] , [20] , [21] ) . Indeed , expressed sequence-tag based vaccine antigen discovery has been explored [22] . However , of 100 ORFs tested only 14 showed detectable protective effects when tested in a high dose infection model of murine CL . This was probably not surprising given that gene expression is regulated mainly post transcriptionally in Leishmania and suggests a need to improve sequence selection criteria . Here , we adapted a reverse vaccinology [23] approach to define novel candidate vaccines , starting from proteomic data sets that were generated recently [24] and ignoring whether or not proteins would be recognized by sera from infected hosts . Moreover , we optimized recombinant attenuated Salmonella as a vaccine carrier platform since they had been explored before as vectors for anti-Leishmania vaccines [25]–[27] and have already been developed for vaccination purposes in humans [28]–[30] .
Female BALB/c mice were purchased from Harlan UK , Charles River UK or bred and maintained under specific pathogen-free conditions in individually ventilated cages in the animal facilities of the School of Biological Sciences at the University of Edinburgh and the University of York . Animals were used at 6–9 weeks of age and were age matched within each experiment . All animal experiments adhered to the UK Animals ( Scientific Procedures ) Act 1986 and were conducted under Project Licenses granted by the UK Home Office and with local ethical approval ( License # PPL 60/03581 to TA and PPL 60/03708 to PK ) . To inducibly express antigens on the surface of Salmonella , the E . coli adhesin involved in diffuse adherence ( AIDA ) autotransporter system was adapted and a variant of plasmid pKRI143 [31] was constructed , pAIDA0 , as previously described [32] . Briefly , the sequence encoding cholera toxin B subunit signal peptide was followed by SpeI/BglII sites for in frame directional cloning of ORF of interest fused with downstream sequences coding for a hemagglutinin epitope ( HA ) -tag and the transporter domain of AIDA , all under the control of the in vivo inducible Mg2+ responsive PpagC promoter [33] . Vaccine antigen ORFs encoding L . donovani KMP-11 ( LinJ35_V3 . 2260 ) , ORF LinJ08 . 1190 ( LinJ08_V3 . 1190 ) , ORF LinJ09 . 1180 ( LinJ09_V3 . 1180 ) , ORF LinJ23 . 0410 ( LinJ23_V3 . 0420 ) , ORF LinJ25 . 1680 ( LinJ25_V3 . 1670 ) and ORF LinJ35 . 0240 ( LinJ35_V3 . 0140 ) were amplified from L . donovani ( MHOM/INI/03BHU-55 ) genomic DNA using primers shown in table 1 . ORF nomenclature and accession numbers are indicated in Table 2 . Amplifications were carried out with the Platinum® Pfx DNA Polymerase kit ( Invitrogen ) . PCR products were digested with SpeI and BglII and cloned into the equally digested pAIDA0 for transformation into SL3261 and E . coli JK321 ( UT5600 zih::Tn10 dsbA::kan ) [34] , respectively . To differentially regulate protein expression levels , point mutations were introduced into the Shine-Dalgarno ribosomal binding sequence ( RBS; underlined ) using site directed mutagenesis . Forward primer for RBS3 ( 5′-GATCAATCTAGATTTAAGAAGCAGATATACATATGATTAAATTAAAATTTGGTG-3′ ) , RBS4 ( 5′-GATCAATCTAGATTTAAGAAGGGAATATACATATGATTAAATTAAAATTTGGTG-3′ ) and RBS5 ( 5′-GATCAATCTAGATTTAAGAAAGAAATATACATATGATTAAATTAAAATTTGGTG-3′ ) were designed to amplify the cholera toxin signal peptide , HA-tag and antigen while simultaneously introducing the mutated Shine-Dalgarno sequence upstream of the signal peptide . The resulting PCR product was SpeI/BglII digested and re-ligated into pAIDA-Antigen . All resulting surface expression plasmids were subsequently named psVAC[# of RBS mutation]-antigen . For expression of antigens in the salmonella cytosol L . donovani ORFs were amplified using primers described in Table 1 . Resulting PCR products flanked by 5′ NdeI and 3′ BamHI sites were digested and first cloned downstream of a PpagC promoter into a pBR322-derived plasmid series already containing mutated Shine-Dalgarno sequences ( RBS1 – AGGAA , RBS2 – GGGAA and RBS3 – AGCAG ) described in [35] for transformation into SL3261 . The resulting plasmids were subsequently named pcVAC[# of RBS mutation]-antigen . Preparation of live vaccine stocks , immunizations and determination of bacterial fitness by in vivo colonisation have been performed exactly as described before [32] . For generating recombinant proteins , Leishmania antigen ORFs were cloned into pET28a ( + ) ( Novagen ) . All antigens were amplified using the NdeI and BamHI site containing primers described above . Recombinant proteins were purified as described previously [32] . KMP-11 , the only soluble protein was directly purified on a Nickel column ( 1 ml , HisTrap FF , GE Healthcare ) . All other antigens formed inclusion bodies which needed to be isolated and dissolved prior purification under denaturing conditions with an on-column refolding step [32] . Recombinant protein containing fractions eluted from columns ( see Fig . S1 ) were pooled and protein concentrations determined using amidoblack [36] ) . Proteins LinJ08 . 1190 , LinJ09 . 1180 , LinJ23 . 0410 and LinJ25 . 1680 became insoluble when imidazole was removed; hence 50 µl/well of a 50 µg/ml protein eluate was used to coat 96-well plates ( MaxiSorb , Nunc ) for ELISA . Plates were sealed and stored at 4°C until needed . For T cell re-stimulation assays imidazole was removed by dialysis against TBS/150 mM NaCl and subsequently concentrated by ultrafiltration using Centricons® ( Millipore ) of appropriate pore size . ELISA for antigen-specific antibodies of different isotypes ( IgG1 and IgG2a ) from mouse serum has been performed as previously described [32] . In brief , serial dilutions of individual sera were analysed . To estimate relative antibody concentrations , titers were determined corresponding to the value of the serum dilution giving a half maximal ELISA signal . L . major promastigotes were grown in semi-defined medium until late stationary phase was reached . Two million parasites were injected into the left hind footpad and lesion size was measured as the difference in thickness between infected and uninfected footpad using a calliper . For determination of parasite numbers in organs mice were sacrificed by cerebral dislocation and organs ( spleen , draining lymph node , footpad ) were removed and homogenized . The single cell suspensions were adjusted to equal volumes and subjected to serial dilutions in 96-well tissue culture plates filled with SDM medium [24] supplemented with 20 µg/ml hygromycin and 50 µg/ml kanamycin , which was carried out in quadruplets . After 14 days at 27°C , parasite growth was scored microscopically and parasite load in the infected organs was calculated using the dilution where at least 2 of 4 wells ( >37 . 5% ) were positive [37] . This dilution was multiplied by the total volume ( in multiples of 0 . 1 ml ) to derive the total number of parasites per organ . Mice were killed by cervical dislocation and livers and spleens were removed and weighed . The body-mass index ( BMI ) was calculated as the organ weight in percentage of body weight . To determine parasitic burden in spleen and liver , impression smears were prepared on microscopic glass slides , fixed in methanol and stained with Giemsa . The number of parasites per 1000 host cell nuclei was counted using a light-field microscope and an immersion oil lens . Leishman-Donovan units ( LDU ) were calculated by multiplication of the number of parasites/1000 nuclei with the organ weight [38] . Liver sections were processed for immunohistochemistry as described in detail elsewhere [39] . Briefly , confocal microscopy was performed on acetone fixed 8 µm frozen sections stained with Alexa 488-conjugated F4/80 ( eBioscience , United Kingdom and purified rabbit anti-mouse inducible nitric oxide synthase ( iNOS ) ( Abcam , United Kingdom ) detected with donkey anti-rabbit Alexa 647 . Sections were counterstained with 4′ , 6′-diamidino-2-phenylindole ( DAPI ) , and mounted in Pro-Long Gold antifade ( Invitrogen ) for examination on a LSM META 510 confocal microscope ( Zeiss ) . Quantification of NOS2 staining was performed on randomly selected fields for each mouse , using Adobe Photoshop CS3 to determine the area of iNOS reactivity ( as number of positively stained pixels ) relative to total granuloma area ( as pixels stained with F4/80 ) . Granuloma maturation was assessed from hematoxylin-eosin ( H&E ) -stained tissue sections as described elsewhere [39] . Statistical analysis was performed using GraphPad Prism Program ( Version 4 . 0 , GraphPad Software , San Diego , California ) . Depending on data passing normality tests , ANOVA was performed with appropriate post-tests for pairwise comparisons or Mann-Whitney tests were computed . P values less than 0 . 05 were considered significant .
For the selection of novel antigen candidates , we conducted a bioinformatic analysis of a proteomic dataset that compared the proteomes of pro- and amastigote stages of L . mexicana [24] . This data set was chosen because to date this is the only dataset containing information on truly intracellular parasites and because a comparison with data from a proteomic analysis of L . donovani axenic amastigotes suggested a very high degree of overlap with respect to abundant proteins [24] . From a total of 509 proteins that reflect the set of highly abundant proteins , we selected five novel antigen candidates based on abundance , conservation throughout the genus and lack of homologies to host proteins ( Figure 1 ) . These criteria and , in addition , predicted subcellular localization were found before to be valuable to identify antigens for induction of protective T cell responses from complex organisms , operationally defined here as expressing ≫103 different protein antigens , e . g . to select antigens for vaccines against Helicobacter pylori [40] . A further , Leishmania-relevant criterion was the expression of the potential antigen in appropriate life cycle stages . Preference was given to proteins expressed in the disease-causing intracellular amastigote stage , but , since early stages of infection after transmission of promastigotes were also considered relevant , antigen expression in both life cycle stages was not an exclusion criterion . Four selected antigens were present in the proteomic datasets of both stages while the LinJ23 . 0410 corresponding protein was present only in the amastigote dataset . Homologues of the encoding genes were found in all cases in L . major , L . infantum , L . donovani and L . braziliensis genomes with a very high degree of conservation ( ranging from 78 . 9% to 95 . 8% identity of amino acid sequence , increasing to 87 . 6% to 99% when including conserved substitutions ) . Sequence homologies to proteins of mouse and human ( human as final target and mouse as a model host ) were excluded by BLAST searches . This approach was biased and preferentially excluded similar sequence-dependent epitopes . It was used here because it was assumed to enhance the likelihood of antigens to be recognized by T cells as “foreign” and to reduce the risk of potential autoimmune sequelae . Novelty and expressability in our salmonella expression systems were additional final selection criterion but we also included the well characterized antigen KMP-11 as a reference vaccine antigen . This antigen has been shown to be protective against L . donovani , when administered as a DNA vaccine [41] , [42] . Subcellular localization and protein amount are not only useful criteria to select T cell vaccine antigens , they are also crucial parameters to consider in the construction of recombinant live vaccine carriers - such as bacteria - to induce antigen-specific cell mediated immunity [43] , [44] . Thus , two expression systems were adapted that directed antigens either to the cytosol or the surface of Salmonella and allow induced expression via the in vivo inducible promoter PpagC . We choose to control antigen production at the translational level and introduced a set of point mutations into a canonical ribosomal binding site ( RBS ) creating a set of four plasmid cassettes each for cytosolic and surface antigen expression . These mutations resulted in staggered protein expression levels when Salmonella strains carrying the respective plasmids were grown under conditions that activate the PpagC promoter ( Figure S2 ) . Heterologous protein expression can greatly reduce fitness of the carrier bacteria in vivo , thereby critically affecting the amount of total antigen delivered to the immune system and thus vaccine immunogenicity . This relationship is schematically shown in Figure 2A ( left panel ) and , as an example , is shown for vaccine strains engineered for cytosolic expression of LinJ23 . 0410 ( Figure 2A , right panel ) . Colonisation of the Peyer's patches seven days after oral administration of 109 CFU was determined as a measure of bacterial fitness . Expression of LinJ23 . 0410 was clearly negatively correlated with the number of CFU found in Peyer's patches , i . e . vaccine strain fitness . Use of a canonical , non-mutated RBS ( RBS0 ) resulted in high amounts of protein but greatly reduced bacterial fitness . Introduction of point mutations ( RBS1 , 2 , 3 ) lowered expression levels from intermediate ( RBS1 ) to very low ( RBS2 and 3 ) which brought fitness back to the level of the empty carrier strain ( Figure 2A right panel ) . A reduction of bacterial fitness far below 104 CFU in this assay , based on past experience ( JS and TA unpublished ) , rendered vaccine strains non-immunogenic with respect to the recombinantly expressed antigen . Thus , out of 48 bacterial strains constructed and evaluated as shown for the example above , 10 strains were selected for further testing . Their respective fitness and antigen expression characteristics were as shown in Figure 2B . Interestingly , antigens LinJ08 . 1190 , LinJ09 . 1180 , LinJ25 . 1680 and LinJ35 . 0240 could not be expressed in the cytosol ( data not shown ) but vaccine strains could be obtained , with the exception of LinJ35 . 0240 , when the antigens were targeted to the bacterial surface . In consequence , only two vaccine strains expressing the antigens KMP-11 and LinJ23 . 0410 cytosolically could be included in the panel ( Figure 2B , right panel ) . In addition , eight surface expression strains were selected ( Figure . 2B , left panel ) . Surface expression of antigen LinJ35 . 0240 could not be detected via western blot despite a clear influence on bacterial fitness ( Figure 2B left panel ) . Based on the latter , it was therefore decided to include psVAC5-35 . 0240 as an example for the respective antigen . All selected strains were next tested in vivo for their ability to protect BALB/c mice against visceralising L . major infection . These mice are highly susceptible to L . major infection , and have been suggested to provide a good mouse model for VL . Mice were vaccinated with a single dose of Salmonella vaccine strains , the carrier control SL3261 or treated with PBS . Mice were subsequently challenged with 2×106 late-stationary phase L . major promastigotes into the left hind footpad . Lesion size was monitored over a course of several weeks after which mice were randomized and selected for analysis of parasitic burden in footpad , lymph node and spleen . A pilot study involving all 10 selected vaccine strains showed that vaccination with Salmonella carrying antigens LinJ08 . 1190 and LinJ23 . 0410 reduced lesion size and parasitic burden compared to the controls ( see Figure S3 ) . Interestingly , vaccination with antigen LinJ25 . 1680 expressing Salmonella exacerbated disease while the other vaccines including the KMP-11-expressing strains had no effect on disease progression compared to controls ( Figure S3 ) . Thus , the presumably protective vaccine strains psVAC5-08 . 1190 , pcVAC1-23 . 0410 and psVAC0-23 . 0410 as well as a mixture of these ( from hereon named ‘vaccine allstars’ ) , were further evaluated ( Figure 3 ) . Vaccination , especially with psVAC5-08 . 1190 and vaccine allstars , significantly delayed the onset and progression of footpad swelling in mice challenged nine weeks later ( Figure 3A ) . Five weeks after infection , five animals per groups were selected randomly and parasitic burden in spleen , popliteal lymph node and footpad was determined . Parasite numbers in footpads and lymph nodes were not significantly different in the vaccine groups ( Figure 3B , C ) although a trend towards lower burdens was notable in mice vaccinated with psVAC0-23 . 0410 , psVAC5-08 . 1190 and vaccine allstars ( Figure 3B , C ) . The discrepancy between lesion size and parasite burden was surprising but is not without precedence . The inverse situation has been described in murine L . major infection when analysing TNR-p55 receptor deficient mice [45] or when mapping susceptibility loci [46] , [47] . However , mechanisms are currently not fully understood . The parasitic burden in the spleen was assessed as a surrogate marker of protection against visceral leishmaniasis . Immunisation with the psVAC5-08 . 1190 and allstars vaccines significantly reduced parasite numbers in the spleen compared to challenged only mice and a similar trend was noted for the surface expressing psVAC0-23 . 0140 vaccine ( Figure 3D ) . Of note , five animals amongst those vaccinated with psVAC5-08 . 1190 and vaccine allstars had no detectable parasites in the spleen ( Figure 3D ) . Hence , a single oral dose of Salmonella vectored vaccines that delivered both LinJ08 . 1190 with LinJ23 . 0410 significantly reduced visceral L . major parasite burdens in these highly susceptible BALB/c mice . Since conservation of the antigens among Leishmania species was a key selection criterion , we hypothesised that antigens which were protective against L . major would also protect against the causative agent of human VL , L . donovani . To test this hypothesis , we immunised BALB/c mice with strains psVAC5-08 . 1190 and vaccine allstars . Leishmania surface antigen KMP-11 had been shown to be protective against L . donovani in mice [42] . Therefore and despite its poor performance in previous experiments , Salmonella strain pcVAC1-KMP , expressing KMP-11 in the cytosol , was included together with the carrier strain SL3261 and sham-immunisation in this study . Mice vaccinated with a single oral dose were challenged intravenously with 3×107 L . donovani amastigotes six weeks later . A characteristic for L . donovani infection in BALB/c mice is hepato-splenomegaly and the organ-specific control of the infection . Half of the mice were sacrificed on day 28 p . i . , when liver parasite burden has usually reached its peak before the onset of self cure and when splenic parasite burden has begun to increase . The remaining animals were analysed at day 68 p . i . to assess long term control , particularly in the spleen . An increased ratio of liver/spleen weight to body weight is an indirect measure of L . donovani infection induced inflammation and disease severity . Thus , body and organ weights were determined at necropsy ( Table S1 ) . The ratio for both liver ( Figure 4A ) and spleen ( Figure 4B ) increased between day 28 and day 68 in non-vaccinated animals and mice treated with either the carrier salmonella alone or the pcVAC1-KMP vaccine . In contrast , in animals vaccinated with psVAC5-08 . 1190 or the allstars vaccine , this ratio either increased less dramatically or not at all ( Figure 4A , B ) . Mice immunized with psVAC5-08 . 1190 or the allstars vaccine had a mean liver parasite burden of 84 . 20±39 . 30 and 69 . 75±20 . 74 LDU , respectively at day 68 p . i . significantly reduced in comparison to the non-immunized group ( 361 . 0±66 . 79 LDU ) , the SL3261 carrier ( 189 . 0±63 . 79 LDU ) or the pcVAC1-KMP treated animals ( 232 . 2±30 . 02 LDU; Figure 4C ) . Of note , the decrease noted after SL3261 treatment in comparison with the naïve controls was also significant ( Figure 4C ) . The effects of the vaccines on splenic parasite burdens followed the same pattern ( Figure 4D ) . Mice immunized with psVAC5-08 . 1190 or the allstars vaccine controlled parasite replication while numbers increased significantly between day 28 and 68 in all other study groups ( Figure 4D ) . Immunisation with pcVAC1-KMP also did not protect mice from L . donovani infection and parasite burden increased over time ( 69 . 80±18 . 67 to 182 . 8±61 . 53 ) , which was similar for SL3261 treated mice ( Figure 4D ) . In summary , a single oral dose of salmonella vectored vaccines delivering LinJ08 . 1190 and/or LinJ23 . 0410 significantly reduced hepato-splenomegaly and visceral infection in mice infected with L . donovani , the causative agent of human VL . To assess immune responses during vaccination and infection , we measured antigen-specific antibody isotype titres as a surrogate of the underlying CD4+ T cell response , given the known correlation between IL-4 and IgG1 responses and between IFNγ and IgG2a [48] . Serum was assessed in vaccinated mice four weeks after immunisation and on day 28 and 68 post infection with L . donovani to test for antigen-specific antibodies . Four weeks after vaccination but before infection , vaccine antigen-specific antibody titers were below the limit of detection ( Figure 5A–F ) . In agreement with the fact that KMP-11-specific antibodies are produced during human VL [49] , infected non-vaccinated mice or SL3261 carrier immunized mice generated anti-KMP-11 antibodies ( Figure 5A , B ) . This anti-KMP-11 response was very similar in the pcVAC1-KMP vaccinated group ( Figure 5C ) . In contrast , vaccines expressing LinJ08 . 1190 and/or LinJ23 . 0410 primed animals for the production of specific antibodies that became detectable after the boosting infection on day 28 and 68 post infection ( Figure 5D–F ) but no antibodies against the respective recombinant proteins were detectable by ELISA ( detection limit of assay was at titers ≤20 ) during infection in naïve , SL3261 or pcVAC1-KMP treated animals ( not shown ) . This indicated that LinJ08 . 1190 and/or LinJ23 . 0410 were not naturally immunogenic during infection of BALB/c mice . Next , the ratios of vaccine antigen-specific IgG1 and IgG2a were calculated for each mouse and time point ( Figure 6 ) to seek evidence for a bias in type 1 vs . type 2 immune response . Over the course of infection significant and different skewing was noted between the treatment groups . Anti-KMP-11 IgG1 to IgG2a ratios were above 1 in pcVAC1-KMP vaccinated mice which was therefore not different from the response to KMP in infected only or SL3261 vaccinated mice . In comparison , anti-vaccine antigen specific IgG1 to IgG2a ratios , however , were significantly different in sera from psVAC5-08 . 1190 or allstars vaccinated mice with values around 1 or below ( Figure 6; p<0 . 05 ) . Finally , to assess the underlying cellular response in a more direct manner , we examined the level of granulomatous inflammation in infected mice that were either unvaccinated or had been vaccinated with control SL3261 Salmonella or with allstars ( Figure 7 ) . At day 28 p . i , there was a small but significant increase in the number of granulomas observed in the liver of allstars vaccinated mice ( Figure 7A ) . We next measured the maturation stage of each granuloma , using established scoring criteria [39] . Granuloma maturation was similar between all groups of mice at day 28 p . i . ( with a small but not significant trend towards enhanced maturation in allstars vaccinated mice ) . By day 68 p . i . , however , mice vaccinated with either SL3261 or allstars showed enhanced granuloma maturation compared to non vaccinated mice . Although the results of this analysis are in keeping with the enhanced ability of these vaccinated mice to reduce parasite burden , it was not a sufficiently sensitive technique to discriminate between the resistance induced by SL3261 and allstars ( c . f . Figure 4D ) . Finally , we measured the area within each granuloma that stained positive for iNOS , as one measure of functional capacity at these inflammatory foci . There were no significant differences in the iNOS response between vaccinated and non-vaccinated mice at either time point by this criterion ( Figure 7D ) . Hence , the main tissue correlate of protection induced by allstars vaccination was an increase in the rapidity of granuloma formation , suggesting that vaccination may have heightened the frequency of CD4+ and/or CD8+ T cells able to facilitate this focal inflammatory response .
We had previously reported on the proteome of the intracellular amastigote stage of L . mexicana [24] which showed extensive overlap with proteins identified in L . donovani axenic amastigotes [50] . Because of this overlap , the former proteomic dataset was exploited here to adapt a reverse vaccinology approach to develop a vaccine against VL . We applied the criteria of protein abundance , within parasite genus conservation , and absence of homologous proteins in host organisms to select novel candidate vaccine antigens aimed to induce cellular immunity . These criteria may not be optimal though to select targets for inducing antibody-dependent immunity . Four of five selected candidates could be expressed in recombinant form and when delivered by recombinant Salmonella two reduced and one exacerbated disease progression in a murine L . major infection model . These results suggest that the frequency of identifying immunologically relevant proteins by this method is high and may well be superior to previous strategies that relied on mRNA expression and genome data for antigen selection [51] , [52] with a hit frequency of ∼15% . Leishmania like other kinetoplastids regulate gene expression mostly post-transcriptionally and mRNA abundance data alone may not be informative to predict protein abundance . However , as in shown in other systems [40] and Leishmania [53] actual protein abundance in amastigotes is highly relevant if the protein is to become a target of the immune response [54] . Analysis of the proteome data sets suggested that bias in codon usage indicates translational bias and therefore is highly correlated with protein abundance [24] . Hence , codon usage may be used to rank ORFs and serve as a substitute parameter for protein abundance in the absence of real protein expression data to refine pure in silico selection of candidate antigens . This becomes particularly relevant for selecting membrane proteins that are severely underrepresented in current proteomic data sets . The two protective antigens , LinJ08 . 1190 and LinJ23 . 0410 , expressed by Salmonella carriers were immunogenic in these vaccines yet , based on antibody responses , were not a target of the immune response to L . donovani infection , at least not in mice . This is noteworthy since many Leishmania vaccine antigens including KMP-11 currently favoured by other groups have been identified using sera from patients [55]–[58] . Our findings with the salmonella vectored KMP-11 vaccine suggest that these immunoselection approaches may introduce an extra hurdle for vaccine development since the natural antigen-specific response may be skewed and , possibly , even be disease exacerbating [59] , [60] . The requirement for an additional type 1 immune response inducing adjuvants , IL-12 , to achieve protective effects with a KMP-11 DNA vaccine in the murine L . major model [61] is in good agreement with this idea . Furthermore , a fusion protein called LEISH-F1 - also known as Leish-110F , Leish-111f or MML – was derived from the sequence of three immunoselected parasite antigens . LEISH-F1 is the most advanced protein-based subunit Leishmania vaccine in trial to date and has shown promising effects when tested in a therapeutic setting against human American CL [62] . However , this is not the case when used to prevent visceral canine disease after high dose experimental infection [63] or to treat naturally acquired VL in dogs [64] . In contrast , Leishmune® , a vaccine based on a glycoproteic fraction of L . donovani that was not immunoselected , is licensed for the prevention of canine VL in Brazil and has shown efficacy in the field [65] . Interestingly , the Leishmune® vaccine antigens are poorly recognized by sera from dogs suffering from VL and vaccination therefore is not interfering with sero-surveillance programs [65] . Thus , reverse vaccinology based approaches as presented here are likely to significantly broaden the choice of protective antigens . A number of subunit vaccine delivery platforms , including purified proteins or mixtures of glycans and glycoproteins , recombinant DNA , viral and bacterial vectors have been evaluated experimentally in murine models of leishmaniases ( review by [12] . However , very few have entered or passed clinical testing and amongst them no vectored vaccine . We have chosen Salmonella as a carrier since these bacteria had already been positively evaluated by several groups in experimental models of leishmaniases [25]–[27] . Moreover , they are being developed as recombinant carriers against a number of pathogens including Helicobacter pylori , Hepatitis B virus and Plasmodium falciparum [66] , [67] . In the context of a major neglected disease such as VL , their main advantages are their excellent safety profile , simple and low-cost production at industrial scale , possibility to store as lyophilized product at room temperature , and oral application route , thus reducing the requirements for extensive infrastructure . In addition , Salmonella are potent inducers of long-lived cell-mediated immunity including CD8+ T cells [28] , [68] . Induction of CD8+ T cells is particularly efficient by vaccines delivered by viral or bacterial carriers and may be a crucial characteristic of anti-Leishmania vaccines , since both CD4+ and CD8+ T cells are required for optimal anti-leishmanial immunity and granuloma formation [15]–[18] . While we do not yet have formal proof that our vaccines induced antigen-specific CD8+ T cells , bioinformatics analysis using CD8 T cell epitope/HLA-binding peptide prediction algorithms suggested epitopes presentable by major HLA alleles e . g . of human populations in VL endemic areas in India [69] . In the context of VL , Salmonella have the additional property to generate viscerotropic immune responses which may explain that the main protective effect was observed at the level of visceralizing infection in the L . major model . Moreover , depending on serovar , S . enterica exhibits broad or narrow host ranges . Serovar Typhimurium that was used here has the potential to deliver vaccine antigens in humans [70] as well as in dogs [71]–[73] while attenuated S . enterica Typhi can be engineered to deliver human vaccines [30] , [66] . In summary , we report the identification of two novel candidate vaccine antigens against VL by reverse vaccinology and the optimized construction of live Salmonella carriers . These VL vaccines could potentially be used to combat VL in the zoonotic , as well as the anthroponotic cycle of the disease . | The leishmaniases are tropical diseases that affect the poorest of the poor . They are caused by Leishmania species , protozoan parasites transmitted by blood sucking insects and the visceral form of the disease is fatal . Vaccines that would tremendously boost disease control strategies need to be designed cost-efficiently and for the existing infrastructure . Salmonella-based live vaccines could fulfil these requirements as they can be cheaply produced on an industrial scale and the lyophilized product can be stored at room temperature and upon rehydration is ready for oral , needle-free application . Salmonella , like Leishmania , are intracellular pathogens that primarily target host macrophages . The bacteria induce a viscerotropic immune response . Herein lies a potentially significant advantage of using attenuated Salmonella as delivery vehicles for parasite antigens for vaccination against visceral leishmaniasis . We used in vivo inducible promoters and optimized expression systems to construct attenuated Salmonella carriers that deliver novel vaccine antigens and show a host protective effect in small rodent models of visceral leishmaniasis . These proof-of-concept studies should serve to further promote exploration of live Salmonella as a cost effective and widely applicable carrier for vaccination against leishmaniases . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"immunizations",
"medicine",
"infectious",
"diseases",
"immunity",
"leishmaniasis",
"neglected",
"tropical",
"diseases",
"biology",
"microbiology",
"parasitic",
"diseases"
] | 2011 | Single Dose Novel Salmonella Vaccine Enhances Resistance against Visceralizing L. major and L. donovani Infection in Susceptible BALB/c Mice |
Deoxyribonuclease II ( DNase II ) is a well-known acidic endonuclease that catalyses the degradation of DNA into oligonucleotides . Only one or a few genes encoding DNase II have been observed in the genomes of many species . 125 DNase II-like protein family genes were predicted in the Trichinella spiralis ( T . spiralis ) genome; however , none have been confirmed . DNase II is a monomeric nuclease that contains two copies of a variant HKD motif in the N- and C-termini . Of these 125 genes , only plancitoxin-1 ( 1095 bp , GenBank accession no . XM_003370715 . 1 ) contains the HKD motif in its C-terminus domain . In this study , we cloned and characterised the plancitoxin-1 gene . However , the sequences of plancitoxin-1 cloned from T . spiralis were shorter than the predicted sequences in GenBank . Intriguingly , there were two HKD motifs in the N- and C-termini in the cloned sequences . Therefore , the gene with shorter sequences was named after plancitoxin-1-like ( Ts-Pt , 885 bp ) and has been deposited in GenBank under accession number KF984291 . The recombinant protein ( rTs-Pt ) was expressed in a prokaryotic expression system and purified by nickel affinity chromatography . Western blot analysis showed that rTs-Pt was recognised by serum from T . spiralis-infected mice; the anti-rTs-Pt serum recognised crude antigens but not ES antigens . The Ts-Pt gene was examined at all T . spiralis developmental stages by real-time quantitative PCR . Immunolocalisation analysis showed that Ts-Pt was distributed throughout newborn larvae ( NBL ) , the tegument of adults ( Ad ) and muscle larvae ( ML ) . As demonstrated by DNase zymography , the expressed proteins displayed cation-independent DNase activity . rTs-Pt had a narrow optimum pH range in slightly acidic conditions ( pH 4 and pH 5 ) , and its optimum temperature was 25°C , 30°C , and 37°C . This study indicated that Ts-Pt was classified as a somatic protein in different T . spiralis developmental stages , and demonstrated for the first time that an expressed DNase II protein from T . spiralis had nuclease activity .
Deoxyribonucleases ( DNases ) are typically divided into two distinct categories , namely , Deoxyribonuclease I ( DNase I ) and Deoxyribonuclease II ( DNase II ) , based their biochemical properties during DNA degradation [1] . There are also many different subclasses . DNase II ( EC 3 . 1 . 22 . 1 ) is a well-known acid endonuclease that catalyses the dissection of DNA molecules into oligonucleotides by single-strand nicking and a double-strand cleavage mechanism [2] . DNase II generates 5′-hydroxyl groups and 3′-phosphate groups without divalent metal ions , but DNase I requires divalent metal ions for its catalytic activity and produces 5′-phosphate groups and 3′-hydroxyl groups [2] . DNase II activity was first observed in 1947 [3] , and many studies have biochemically characterised these enzymes in mammalian systems [4] , [5] . However , the nucleotide and amino acid sequences of these genes were unclear until the human DNase II gene was cloned in 1998 [6] . Soon after this initial report , DNase II or DNase II homologues were identified in vertebrates , invertebrates , and non-metazoans [7] . Trichinella spiralis ( T . spiralis ) is an intracellular pathogen of skeletal muscle and one of the most widespread zoonotic parasitic nematodes in the world [8] . It is especially prevalent in China , Argentina , and some eastern European countries [9] . To date , eight species and four genotypes have been classified in the genus Trichinella . The complete basic life cycle in a single host includes adult worms ( Ad ) , newborn larvae ( NBL ) , and muscle larvae ( ML ) . Approximately 11 million people in 55 countries carry the infection , which is transmitted by eating of poorly cooked or raw infected meat [8]; the infection has a 0 . 2% mortality rate [9] . Trichinellosis is not only a serious public health threat but also an important economic factor in animal production and food safety [9] . Compared with enzymes from other species including C . elegans , the DNase II-like protein family in T . spiralis has expanded remarkably , with an estimated 125 genes in the genome [10] . Based on comparative protein sequence analyses , around half of these genes encode excretory-secretory ( ES ) products that are implicated in host-parasite interactions , and these proteins have been suggested as vaccine candidates for the control and prevention of trichinellosis [11] . A histidine residue that is surrounded by a highly conserved 5-mer , DHSKW [12] , has been proposed as the core catalytic centre of most DNase II family members [13] . Of the 125 genes predicted as the DNase II-like protein family in T . spiralis , only plancitoxin-1 ( 1095 bp , GenBank accession no . XM_003370715 . 1 ) possesses one predicted active site involved in DNA cleavage and located in C-terminus . To date , neither its expression nor its activity has been explored . In the present study , we cloned and characterised the plancitoxin-1 gene from T . spiralis . The cloned sequences were shorter than the predicted sequences in GenBank and named after plancitoxin-1-like ( Ts-Pt , 885 bp , GenBank accession no . KF984291 ) . Meanwhile , the DNase activity of the recombinant Ts-Pt protein ( rTs-Pt ) was examined .
Parasites were prepared from different stages of the T . spiralis ( ISS534 ) life cycle , as previously described [14] . Briefly , mice were experimentally infected per os with 400 L1 infective larvae , and T . spiralis ML were recovered at 35 and 60 days post-infestation ( dpi ) using a standard pepsin-hydrochloric acid digestion method . Rats were experimentally infected per os with 6000 L1 infective larvae , and Ad were isolated from the small intestines at 1 ( Ad1 ) , 3 ( Ad3 ) , and 6 dpi ( Ad6 ) . NBL were obtained from female Ad at 6 dpi and were incubated overnight in RPMI-1640 medium ( Gibco , USA ) containing 200 U/mL penicillin ( Sigma , USA ) and 200 µg/mL streptomycin ( Sigma , USA ) at 37°C and 5% CO2 . BALB/c mice ( 18±2 g ) and Wistar rats ( 180±20 g ) aged 6–8 weeks and New Zealand white rabbits ( 2±0 . 2 kg ) aged 12 weeks were obtained from the Experimental Animal Centre of College of Basic Medical Sciences , Jilin University ( Changchun , China ) . Animals were free of specific pathogens and were housed and fed in compliance with the National Institutes of Health guidelines ( publication no . 85-23 , revised 1996 ) . Animals were reviewed and approved by the Ethical Committee of Jilin University affiliated to the Provincial Animal Health Committee , Jilin Province , China ( Ethical Clearance number IZ-2009-008 ) . Total RNA from ML , Ad and NBL was extracted using a Trizol RNA extraction kit ( Invitrogen , USA ) and transcribed into first-strand cDNA with a SuperScript II RT cDNA synthesis kit ( Invitrogen , USA ) , according to the manufacturer's instructions . The transcription levels for Ts-Pt were evaluated in different T . spiralis developmental stages with a forward primer ( 5′-GAATAATACTGTCAACTGGAAT-3′ ) , reverse primer ( 5′-TTTAGGAATGCTGTGAATTAG-3′ ) and SYBR Premix Ex Taq II ( Tli RNaseH Plus ) ( TAKARA , China ) . Real-time quantitative PCR was performed on an ABI Prism 7500 sequence detection instrument ( Applied Biosystems , Inc . ) , as previously described [15] . The housekeeping gene GAPDH ( glyceraldehyde-3-phosphate dehydrogenase , GenBank accession no . AF452239 ) was amplified with the forward primer 5′-GCTCCTATGTTGGTTATGGG-3′ and the reverse primer 5′-TTTGGGTTGCCGTTGTAG-3′ . The relative expression of Ts-Pt in different developmental stages was determined using the 2−ΔCt method . Three independent experiments were performed . All of nucleotide and amino acid sequences in this study were from the National Center for Biotechnology Information ( NCBI ) ( http://www . ncbi . nlm . nih . gov/ ) . DNAMAN ( version 6 . 0 . 3 . 48 ) was used to analyse the homology between Ts-Pt and other 124 predicted DNase II-like protein family genes . Conserved domains were predicted at http://www . ncbi . nlm . nih . gov/Structure/cdd/wrpsb . cgi . The molecular weight and theoretical pI were analyzed by the software ProtParam from ExPASy ( http://web . expasy . org/protparam/ ) . Online softwares were used to analyze the rare codon ( http://people . mbi . ucla . edu/sumchan/caltor . html ) and recombinant protein solubility ( http://biotech . ou . edu/ ) . Signal peptide , transmembrane domain , and N-linked glycosylation sites were predicted by the SignalP program , TMHMM program , and NetNGlyc program , respectively ( http://www . cbs . dtu . dk/services/ ) . Multiple sequence alignments for DNase II protein families from various organisms was performed with CLASTALX ( version 2 . 1 ) . Phylogenetic analysis of amino acid sequences was carried out using PHYLIP ( version 3 . 695 ) . And a neighbor joining tree was generated by bootstrap analysis with 1000 replicates using PHYLIP-NEIGHBOR . Then , the phylogenetic tree was visualized and edited using FigTree ( version 1 . 3 . 1 ) . Forward ( 5′-TTTTGGATCCATGGACGCACGTCGGCCGGTAT-3′ , BamHI site underlined ) and reverse primers ( 5′-CCCAAGCTTTCAATATGGTGGAATAGGACAAAGT-3′ , HindIII site underlined ) were used to amplify the Ts-Pt gene from cDNA and genomic DNA from larvae . The recombinant plasmid was constructed from a linearised pET-28a ( + ) expression vector ( Novagen , Germany ) and the target gene . DNA sequencing was performed on an automated DNA sequencer . rTs-Pt was expressed from an E . coli Rosetta ( DE3 ) strain after induction with 1 mM isopropyl-β-D-thiogalactopyranoside ( IPTG ) , and the protein was purified by affinity chromatography using a His-Trap purification kit ( GE , USA ) , per the manufacturer's instructions . Infected sera were collected at 35 dpi from BALB/c mice infected experimentally per os with 400 L1 infective T . spiralis larvae . The antisera against rTs-Pt was produced in a rabbit injected subcutaneously with approximately 500 µg of purified rTs-Pt mixed with complete Freund's adjuvant ( FCA , Sigma , USA ) . Three additional booster injections containing 250 µg of rTs-Pt mixed with incomplete Freund's adjuvant ( IFA ) was injected intradermally at 2-week intervals . Antibodies from blood serum were affinity purified using Protein A Sefinose ( Sangon , China ) , according to the manufacturer's instructions . Affinity-purified antibodies were used for following western blotting and immunolocalisation experiments . rTs-Pt , crude somatic extracts and excretory/secretory ( ES ) products of T . spiralis were subjected to SDS-PAGE on a 12% polyacrylamide gel and subsequently transferred to a nitrocellulose membrane ( Millipore , USA ) . After blocking in TBST-B [25 mM Tris , pH 8 . 0 , 125 mM NaCl , 0 . 05% Tween 20 ( V/V ) , 3 . 7% BSA] for 2 h at 37°C or overnight at 4°C , the membrane was incubated with the primary antibodies ( T . spiralis-infected mice serum and rabbit anti-rTs-Pt serum ) at a dilution of 1∶200 in TBST-B for 2 h at room temperature . Secondary antibodies conjugated to horseradish peroxidase ( goat anti-rabbit IgG and goat anti-mouse IgG ) ( Dingguo , China ) were diluted 1∶5000 in TBST-B and incubated with the membrane for 1 h at room temperature . The membrane was reacted with ECL ( enhanced chemiluminescence ) reagent ( Pierce , USA ) and exposed to BioMax film . A modified method described by Detwiler and Macintyre was used for SDS-PAGE zymography activity gels [16] . Briefly , an SDS-PAGE gel containing 50 µg/mL salmon sperm DNA ( Sigma , USA ) in the separation gel ( 12% ) but not in the concentration gel ( 4% ) was prepared . The samples were incubated in loading buffer without β-mercaptoethanol at 37°C for 15 min and electrophoresed at 4°C . After electrophoresis , the gels were shaken gently in 2 . 5% Triton X-100 ( Sigma , USA ) at 4°C for 30 min with 4 changes in buffer and subsequently rinsed in 50 mM sodium acetate ( pH 5 . 4 ) reaction buffer . For the DNase reaction , the gels were incubated at 37°C for 36 h in reaction buffer . For the enzymatic property study , the gels were incubated in reaction buffer at different temperatures in the presence of metal ions and EDTA , with or without inhibitor , or in different pH buffer solutions . The gels were stained with ethidium bromide , visualised with UV light , and subsequently stained with Coomassie brilliant blue . DNase bands were excised from SDS-PAGE zymography gels and stored in ultrapure water . LC-MS/MS was performed by ProtTech , Inc . ( Phoenixville ) . Briefly , the sample was cleaned by washing with water and digested in-gel with trypsin in digestion buffer ( 100 mM ammonium bicarbonate , pH 8 . 5 ) . The peptides were extracted with acetonitrile , completely dried , re-dissolved , and analysed by a NanoLC-ESI-MS/MS . The MS data were used to search against the non-redundant protein database ( NR database , NCBI ) with the ProTech ProtQuest software suite . Immunostaining of worms was performed as described previously [17] . Briefly , whole T . spiralis ML , Ad and NBL were immersed in fixative solution ( 3 . 7% formaldehyde 10 min , cold 100% MeOH 5 min ) , permeabilised by incubation in PBS containing 1% Triton X-100 for 5 min and blocked with 3% BSA in PBST ( PBS containing 0 . 1% Triton X-100 ) . The worms were then incubated with rabbit anti-rTs-Pt polyclonal antibody at 4°C overnight . Following 3 washes in PBST , the worms were incubated with Alexa Flour 594-labeled goat anti-rabbit IgG fluorescent antibody ( Invitrogen , USA ) at room temperature for 1 h . The worms were washed 3 times in PBST , stained with Hoechst 33342 ( Invitrogen , USA ) for 10 min , washed 3 additional times in PBST , mounted with 70% glycerol on slides , and observed under a fluorescence microscope . The transcription data were expressed as the means ± standard deviation ( SD ) , and the differences among groups were analysed with a one-way ANOVA and Student's t-test . P values were denoted as follows: *p<0 . 05 and **p<0 . 01 ( p<0 . 05 or less was considered statistically significant ) .
The Ts-Pt gene , comprising an-885 bp complete cds sequence , was obtained by PCR ( data not shown ) , and has been deposited in GenBank ( accession no . KF984291 ) . The Ts-Pt sequence was 210 bp shorter than previously predicted plancitoxin-1 ( Tsp_09974 , GenBank accession no . XM_003370715 . 1 ) . Sequence analysis revealed that the Ts-Pt gene encoded a protein of 294 amino acids with a predicted molecular mass ( Mr ) of 33 . 2 kDa and theoretical pI of 9 . 17 ( http://web . expasy . org/protparam/ ) . By sequence alignment , the Ts-Pt protein revealed less than 20% sequence identity to other 124 predicted DNase II-like protein family genes except AY790263 ( 22 . 13% ) , AY790264 ( 21 . 84% ) , AY790265 ( 21 . 18% ) , AY790266 ( 23 . 76% ) , Tsp_02430 ( 20 . 37% ) , Tsp_07454 ( 20 . 55% ) , Tsp_11476 20 . 57% ) , Tsp_11488 ( 20 . 06% ) , Tsp_11491 ( 21 . 90% ) , Tsp_11501 ( 20 . 56% ) , Tsp_12136 ( 22 . 95% ) , Tsp_12346 ( 20 . 29% ) , Tsp_12347 ( 21 . 84% ) . Although the sequence homology of the Ts-Pt gene was quite low compared with genes from other species , the NCBI non-redundant protein sequence database indicated that the protein shared a deeper homology with DNase II or DNase II-like proteins from a wide variety of eukaryotic and prokaryotic species ( data not shown ) . The phylogenetic relationships between Ts-Pt and previously reported DNase II family members were determined based on sequence similarities . A total of 33 DNase II sequences from various organisms ( Text S1 ) including human , bovine , horse , porcine , mouse , rat , chicken , fugo , Drosophila , Xenopus laevis , Xenopus tropicalis , Acanthaster , zebrafish , Anopheles gambiae , C . elegans , C . briggsae , Burkholderia pseudomalle , Dictyostelium fasciculatum , Dictyostelium discoideum , canarypox virus , fowlpox virus , and T . spiralis were used to reconstruct the phylogenetic relationships . Ts-Pt and Acanthaster plancitoxin-1 belonged to the same clade ( Fig . 1 ) and had the similarity with 41% . In addition , the protein contained four potential N-linked glycosylation sites ( Asp-X-Thr/Ser ) at positions 19 , 49 , 209 and 262 . Neither a signal peptide nor a transmembrane domain was found in the derived amino acid sequence ( http://www . cbs . dtu . dk/services/ ) . Sequence analysis showed that the Ts-Pt nucleotide sequence contained 26 rare codons ( http://people . mbi . ucla . edu/sumchan/caltor . html ) . An E . coli Rosetta ( DE3 ) strain was used to express the target protein , which had an estimated zero percent chance of solubility when overexpressed in E . coli ( http://biotech . ou . edu/ ) . The recombinant protein of approximately 36 kDa was observed in inclusion bodies after induction with IPTG , and it appeared as a single band on SDS-PAGE after purification ( Fig . 2A ) . Western blot analysis revealed that rTs-Pt can be recognized by serum from rTs-Pt immunized but not pre-immunized rabbit . Serum from T . spiralis-infected mice also showed reactivity to rTs-Pt , which implied the potential antigenicity of native Ts-Pt ( Fig . 2B ) . To determine whether Ts-Pt was expressed as a somatic or secretory protein , the crude somatic extracts and ES products of T . spiralis ( Ad , NBL , and ML ) were reacted with the antibody against rTs-Pt . As shown in Fig . 2 , the antibody against rTs-Pt recognised T . spiralis crude somatic extracts but not ES products ( Fig . 2C ) . To quantify transcription of the Ts-Pt gene in different T . spiralis developmental stages , real-time quantitative PCR was performed . Although there were statistically significant differences in mRNA when any two groups other than Ad1 and NBL were compared , the Ts-Pt gene was expressed at all developmental stages ( Fig . 3 ) . In general , its expression in Ad1 , Ad3 , NBL , and ML was lower than in Ad6 . For Ad , the expression of Ts-Pt mRNA rapidly increased to its maximum level at 6 dpi . Ts-Pt was predicted to have DNase II activity based on sequence homology . We determined the nuclease activity of Ts-Pt using rTs-Pt protein purified by His affinity . rTs-Pt protein catalysed the degradation of salmon sperm DNA in 50 mM sodium acetate and appeared as a single black band of ∼36 kDa under UV light ( Fig . 4 ) . To further examine the catalytic properties of rTs-Pt , we analysed rTs-Pt activity in conditions with various temperatures and pH and in the presence or absence of metal ions , EDTA , and nuclease inhibitors . In a narrow range of slightly acidic conditions , rTs-Pt had it optimal nuclease activity , which disappeared in an alkaline environment at 37°C ( Fig . 5D ) . The single black band was observed clearly at 25°C , 30°C , and 37°C in 50 mM sodium acetate , weakly observed at 20°C and 42°C , and not observed at 16°C or 50°C ( Fig . 5E ) . High concentrations of metal ions inhibited rTs-Pt activity at 37°C in 50 mM sodium acetate ( Fig . 5A ) , and nuclease inhibitors and aurintricarboxylic acid ( ATA , Sigma ) had the same effect ( Fig . 5C ) . Nuclease activity was not be affected by EDTA ( Fig . 5B ) . The protein band was successfully identified by NanoLC-MS/MS . Four peptides were matched and characterised as Ts-Pt ( Table 1 ) . The MS data were only analysed by BLAST with the NCBI reference ( accession no . XM_003370715 . 1 ) because the sequence submitted by us was not released . To evaluate the localisation of Ts-Pt in worms , immunofluorescence staining was performed . As shown in Fig . 6 , strong red fluorescence signals were observed in the entire bodies of NBL and the teguments of Ad and ML after reaction with rTs-Pt antibody . No fluorescence was observed in worms stained with the negative control rabbit serum .
From the perspective of protein function , DNases are divided into three groups: the Mg2+-endonucleases , the Ca2+/Mg2+ endonucleases , and the cation-independent endonucleases . DNase II was discovered more than 50 years ago and belongs to the third group . Recently , DNase II has been purified , and its physical , molecular and enzymatic properties have been thoroughly examined . Three acidic DNase II enzymes—DNase IIα , DNase IIβ , and L-DNase II—have been identified since the human DNase II gene was first cloned in 1998 [18] . DNase II is a monomeric nuclease that contains two copies of a variant HKD motif ( H-x-K-x ( 4 ) -D , where x represents any amino acid residue ) in the N- and C-termini , suggesting a putative catalytic mechanism ( Fig . 7 ) . The two variant HKD motifs compose the catalytic centre of DNase II in a pseudodimeric way . In T . spiralis , the endonuclease activity of at least three proteins was demonstrated in the excretory/secretory ( ES ) products [19] . p43 , SS1 , and AAK85403 have significant similarity to human DNase II [20] . Although the MYC-Tsp43 plasmid facilitated this expression as a recombinant protein in C2C12 myoblasts in the presence of the DNase inhibitor ATA [21] , the T . spiralis protein responsible for nuclease activity is unknown . Of the 125 DNase II-like protein family genes in the T . spiralis genome , only plancitoxin-1 contains the HKD motif . In our work , the Ts-Pt gene encoded a protein of 294 amino acids , which is 70 amino acids shorter than the NCBI reference sequence of plancitoxin-1 . Compared with plancitoxin-1 , which has a variant HKD motif in the C-terminus domain , Ts-Pt had two motifs in the N- and C-terminus domains ( data not shown ) . Generally , the putative active site sequences for DNase IIα and β are DHSK and DHAK , respectively , in the C-terminus domains [12] . However , mouse , rat , fugo , and zebrafish DNase IIβ enzymes contain a DHSK motif . Thus , it was difficult to classify Ts-Pt as a DNase IIα or DNase IIβ . One method to identify potential nuclease activity is zymography . In this assay , DNA is incorporated into an SDS-polyacrylamide gel as a special substrate , and the loss of substrate from the gel matrix reflects nuclease activity [22] . Nuclease activity at a given molecular weight is recognised as a dark band of enzyme degradation in a white background in the gel under UV light , which is not stained by ethidium bromide . In previous studies , all proteins were obtained from animal , plant , bacteria , fungi , and parasite tissues and ES products . Soluble nucleases expressed via prokaryotic and eukaryotic expression systems can be used for nuclease zymography [23] . In this study , we have changed several expression parameters including temperature , IPTG concentration , induction time , bacterial host , expression vector , and medium but still failed to obtain soluble rTs-Pt ( data not shown ) . However , it is surprising that inclusion bodies could be used for nuclease zymography , showing limited degradation of DNA . The theoretical principle of this phenomenon cannot be elaborated completely in this paper; however , it is possible that the inclusion bodies were denatured by SDS and subsequently electrophoresed in SDS gels containing DNA . After electrophoresis , most misfolded proteins were allowed to renature in the gel by washing with 2 . 5% Triton X-100 to remove SDS . DNA was digested by the proteins that properly refolded in the reaction buffer . In vitro , DNase II cuts DNA in 50 mM sodium acetate buffer ( pH 4 . 6–5 ) with the addition of EDTA , and it exhibits weakly Ca2+/Mg2+-dependent endonuclease activity but strongly cation-independent endonuclease activity below pH 7 [24] . In this report , we showed that the nuclease rTs-Pt was active in a narrow pH range ( pH 4–5 ) and an optimum temperature range ( 25°C , 30°C , and 37°C ) . rTs-Pt activity was suppressed by metal ions in concentrations greater than 20 mM , including K+ , Na+ , Ca2+ , Co2+ , Cu2+ , Mg2+ , Mn2+ , Ni2+ , and Zn2+ . The chelator EDTA had no effect on rTs-Pt activity , but a general nuclease inhibitor , ATA , affected rTs-Pt activity . Excluding Xenopus tropicalis DNase II , Burkholderia pseudomalle DNase II , Dictyostelium fasciculatum DNase II , canarypox virus DNase II-like protein , and fowlpox virus DNase II , all the proteins mentioned in Fig . 1 contain a signal peptide at the amino terminus , as predicted by the SignaIP 4 . 1 program analysis ( http://www . cbs . dtu . dk/services/ ) . For T . spiralis , no signal peptide or transmembrane domain was predicted in Ts-Pt . Western blotting showed that rTs-Pt was expressed as a somatic protein in different developmental stages of T . spiralis and most highly expressed in Ad at 3 dpi . In vertebrates , DNase II , which is usually involved in various development processes and DNA degradation during cell death , is important for organismal homeostasis [7] . In C . elegans , there are three DNase II homologues , NUC-1 , CRN-6 ( K04H4 . 6 ) , and CRN-7 ( F09G8 . 2 ) . These proteins play differential roles in apoptotic DNA degradation and development in C . elegans , and they have important regulative action in DNA degradation [25] . Ts-Pt was located in the entire bodies of NBL and the teguments of Ad and ML . It might remove damaged DNA during the growth and development of the parasite . Moreover , the tegument of the parasite is constantly in contact with the host cell . Thus , it may be associated with host-parasite interactions during infection . In conclusion , we characterised a DNase II protein , Ts-Pt , which was expressed in the different developmental stages of T . spiralis . The recombinant protein was purified from a prokaryotic expression system , and it was determined to have nuclease activity in vitro via DNase zymography . To better understand the exact function and mechanisms of Ts-Pt in vivo , more detailed functional and mechanistic studies are needed . | Deoxyribonuclease II ( DNase II ) is classified into a unique family of nucleases and mediates the degradation of DNA associated with apoptosis . Although DNase II activity was first observed in 1947 , and has been studied biochemically and enzymatically since the 1960s , only recently has genetic information on the enzyme been reported . Compared with enzymes from other species , including C . elegans , the DNase II-like protein family of the parasitic nematode T . spiralis has expanded remarkably , with an estimated 125 genes found in the draft genome of T . spiralis . However , none of these proteins have been confirmed by biochemical studies . This study describes Ts-Pt , a DNase II protein that is expressed in different T . spiralis developmental stages . The recombinant protein purified via a prokaryotic expression system displayed in vitro nuclease activity , as determined by DNase zymography . The exact function and mechanisms of Ts-Pt should be further explored in vivo . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"trichinellosis",
"parasitic",
"diseases",
"medicine",
"and",
"health",
"sciences"
] | 2014 | Characterisation of a Plancitoxin-1-Like DNase II Gene in Trichinella spiralis |
The disease severity of Entamoeba histolytica infection ranges from asymptomatic to life-threatening . Recent human and animal data implicate the gut microbiome as a modifier of E . histolytica virulence . Here we have explored the association of the microbiome with susceptibility to amebiasis in infants and in the mouse model of amebic colitis . Dysbiosis occurred symptomatic E . histolytica infection in children , as evidenced by a lower Shannon diversity index of the gut microbiota . To test if dysbiosis was a cause of susceptibility , wild type C57BL/6 mice ( which are innately resistant to E . histiolytica infection ) were treated with antibiotics prior to cecal challenge with E . histolytica . Compared with untreated mice , antibiotic pre-treated mice had more severe colitis and delayed clearance of E . histolytica . Gut IL-25 and mucus protein Muc2 , both shown to provide innate immunity in the mouse model of amebic colitis , were lower in antibiotic pre-treated mice . Moreover , dysbiotic mice had fewer cecal neutrophils and myeloperoxidase activity . Paradoxically , the neutrophil chemoattractant chemokines CXCL1 and CXCL2 , as well as IL-1β , were higher in the colon of mice with antibiotic-induced dysbiosis . Neutrophils from antibiotic pre-treated mice had diminished surface expression of the chemokine receptor CXCR2 , potentially explaining their inability to migrate to the site of infection . Blockade of CXCR2 increased susceptibility of control non-antibiotic treated mice to amebiasis . In conclusion , dysbiosis increased the severity of amebic colitis due to decreased neutrophil recruitment to the gut , which was due in part to decreased surface expression on neutrophils of CXCR2 .
Amebiasis , caused by intestinal infection of Entamoeba histolytica , is one of the leading causes of parasite infection-related mortality and morbidity around the world [1] . Although disease severity ranges from self-limited mild abdominal symptoms to life-threatening systemic disease , determinant factors of infection outcome are still undefined [2] . Even in the same patient , invasive symptomatic disease can develop after long-term asymptomatic colonization [3–5] , and conversely patients with amebic liver abscess after medical treatment can be asymptomatic cyst passers [6] . These results suggest that not only the genetics of host and pathogen are important for determining clinical symptoms of infected individuals , but also the gut environment surrounding E . histolytica . In fact , recent human and animal data indicate that the gut microbiome plays an important role in the pathogenesis of E . histolytica infection . Our group reported that the presence of Prevotella copri in gut flora is associated with susceptibility of children to E . histolytica induced diarrheal disease [7] . Also , in an animal model , we demonstrated gut colonization with segmented filamentous bacterium exerts a protective effect via enhancing the induction of IL-23 in bone marrow-derived dendritic cells [8 , 9] . It is of interest to us to better understand the impact of the gut microbiome on the severity of amebic colitis , potentially by its modulation of intestinal mucosal immunity . Neutrophils are important in protecting the host from E . histolytica tissue invasion into intestine and liver [10–16] . Neutrophils kill E . histolytica in vitro in the presence of TNF-α and IFN-γ mainly via oxygen free radicals [17] . Antibody-depletion of neutrophils in vivo promoted tissue invasion by E . histolytica [14 , 16] , and neutrophil chemotaxis toward leptin plays an important role in protecting host from intestinal tissue invasion [16] . Dysbiosis is known to affect neutrophil function . For example , in an animal model it has been shown that the severity of sickle cell disease is relieved under antibiotic induced dysbiosis , due to a decrease in the number of activated aged neutrophils [18] . However the effect of dysbiosis on neutrophil mediated protection against infectious diseases has not been investigated . Here we demonstrate the impact of dysbiosis on disease severity of E . histolytica infection in humans and in a mouse model of amebic colitis . We go on to demonstrate that one mechanism by which dysbiosis increases susceptibility is by blockading neutrophil recruitment to the gut via down-regulation of CXCR2 .
In order to assess the impact of dysbiosis on outcome of intestinal infection of E . histolytica , we collected stool samples from children cohorts followed from birth in an urban slum , Mirpur in Dhaka Bangladesh [19 , 20] . First , we compared gut microbiome diversity in stools collected from children with symptomatic amebic colitis with those from children who showed asymptomatic E . histolytica colonization . The Shannon diversity index during E . histolytica infection was lower in symptomatic cases than with colonization ( Fig 1a ) . We confirmed that ages of children when stool samples were collected were not different between 2 groups ( Fig 1b ) . Next , we examined microbiome diversity of stools prior to E . histolytica infection ( no E . histolytica confirmed by PCR ) in symptomatic amebic colitis cases , and compared them with uninfected control children ( although we could not use stools prior to E . histolytica infection in asymptomatic colonization cases due to the lack of available samples ) . The Shannon diversity index prior to amebic colitis was significantly lower than that in children who did not develop E . histolytica infection ( S1 Fig ) . Although other factors , such as host/pathogen genetic factors are likely also to be important determinants , our results suggest that decreased microbiota diversity in the gut is one of the determinants for disease severity of E . histolytica infection . We used the murine model of amebic colitis to assess whether prior dysbiosis with decreased microbiome diversity has an impact on disease severity of amebic colitis . Wild type C57BL/6 mice were treated with an antibiotic cocktail consisting of ampicillin , neomycin , metronidazole and vancomycin for 2 weeks prior to E . histolytica challenge ( antibiotic pre-treated mice ) . We confirmed the microbiome diversity was decreased by antibiotic pre-treatment before E . histolytica challenge ( Fig 2a ) . Antibiotic pre-treated mice showed more severe weight loss and higher clinical scores than untreated control mice at early time points ( until day 3 ) after E . histolytica challenge ( Fig 2b & 2c ) . Although obviously bloody stools were not documented in any mice , fecal occult blood ( FOB ) was significantly higher in antibiotic pre-treated mice ( Fig 2d ) . Also , we found that E . histolytica DNA was still detected at later time points ( after 4 days on challenge in stool [Fig 2e] and at day 9 in cecal contents [S3a Fig] ) in antibiotic pre-treated mice , whereas E . histolytica was rapidly cleared in untreated control mice ( Fig 2e ) . Interestingly , despite delayed clearance of E . histolytica , stool lipocalin-2 , which is a neutrophil derived protein reflecting neutrophil associated gut inflammation [21] , was lower in antibiotic pre-treated mice than those in untreated control mice at early time points of infection ( Fig 2f ) . These results indicated that antibiotic-induced dysbiosis increased susceptibility to amebic colitis in the mouse model , consistent with what had been observed in infants . In order to assess the tissue invasion of E . histolytica in cecum at an earlier time point , mice were sacrificed 24 hours after E . histolytica challenge . While E . histolytica culture of cecal contents was positive in all antibiotic pre-treated mice and most untreated control mice ( Fig 3a ) , E . histolytica burden in the cecal lumen was significantly higher in antibiotic pre-treated mice ( Fig 3b ) . Histopathological examination with hematoxylin and eosin stain demonstrated that E . histolytica trophozoites invaded into mucosa with epithelial cell disruption in antibiotic pre-treated mice , whereas trophozoites were localized within gut lumen in control mice ( S2 Fig ) . Although E . histolytica could not be identified at crypts or submucosa in the tissue , immunohistochemistry ( IHC ) using anti-E . histolytica migration inhibitory factor ( anti-EhMIF ) , which is a secreted protein [22] , revealed a higher density of EhMIF at crypts of the cecum in antibiotic pre-treated mice ( Fig 3c & 3d ) , consistent with more severe epithelial damage and invasion by E . histolytica in antibiotic pre-treated mice . These results strongly suggested that more aggressive E . histolytica invasion had already occurred within 24 hours after challenge at the site of infection in antibiotic pre-treated mice , although E . histolytica burden in stool was not different until 4 days after E . histolytica . Next , we checked the mucosal barrier upon E . histolytica challenge . The gut mucosal barrier is the first host defense against intestinal infection by E . histolytica , as previously shown in MUC-2-deficient mice [23] . Recently , we reported that IL-25 mediated mucosal barrier function plays an important role in the protection from intestinal E . histolytica infection [24] . Also , it is known that secretion of IL-25 could be reduced by antibiotic induced dysbiosis [25 , 26] . At 24 hours after E . histolytica challenge , we found less mucus-containing MUC2 positive goblet cells by histopathology and less expression of the Muc2 gene ( Fig 3e–3h ) . Also , we confirmed that IL-25 upon E . histolytica challenge was lower in these mice ( Fig 3i ) . Eosinophils were not different at this time point of infection between antibiotic pre-treated and untreated control mice ( Fig 3j & 3k ) . These results indicate that disturbance of IL-25 and the mucosal barrier were associated with tissue invasion of E . histolytica in antibiotic pre-treated mice . As shown above , despite the more severe tissue damage with delayed clearance of E . histolytica , stool lipocalin-2 [21] , was lower in antibiotic pre-treated mice than those in untreated control mice at early time points of infection ( Fig 2f ) . At later time points , lipocalin-2 was higher in antibiotic-treated and E . histolytica infected mice , in accordance with pathogen burden ( S3a & S3b Fig ) . These results indicate that neutrophil activation upon E . histolytica challenge was suppressed early after infection in the antibiotic pre-treated mice , giving a potential explanation for the action of antibiotics in increasing amebiasis severity . Neutrophil myeloperoxidase activity was also measured , and was lower in cecal tissue at 24 hours after E . histolytica challenge ( Fig 4a ) . Next , we measured IL-1β level by ELISA , in order to assess the recognition of E . histolytica by the intestinal epithelial cells and innate immune system . IL-1β is a pro-inflammatory cytokine , and it has been shown that NLRP3-inflammasome caspase-1 mediated IL-1β is secreted from a macrophage cell line upon contact with E . histolytica [27 , 28] . It was also reported that NLRP3-inflammasome activation could be disturbed by antibiotic induced dysbiosis , resulting in more severe airway infection by influenza virus [29] . Opposed to our hypothesis , however , IL-1β as well as the neutrophil chemoattractant chemokines CXCL1 and CXCL2 at 24 hours after E . histolytica challenge were significantly higher in antibiotic pre-treated mice compared to those in untreated control mice ( Fig 4b–4d ) although they were elevated in response to E . histolytica challenge compared to baseline levels in both antibiotic pre-treated and untreated control mice ( S4a–S4c Fig ) . Considered together , disturbed neutrophil activation upon E . histolytica challenge in antibiotic pre-treated mice was not caused by impaired recognition of E . histolytica as manifest by IL-1β or chemokines , leading us to ask if it might be caused by decreased neutrophil responses to these chemokines . As presented above , neutrophil activation was suppressed despite tissue invasion by E . histolytica in antibiotic treated mice . These mice also paradoxically expressed higher neutrophil chemoattractant chemokines . We hypothesized that inhibition of CXCR2 expression on neutrophils , which is the main receptor for the chemokines CXCL1 and CXCL2 , could explain the relative lack of gut neutrophil abundance . In order to assess this hypothesis , we checked the number and expression of surface markers of the Ly6G high neutrophil population ( Fig 3j ) in the blood and cecal lamina propria of mice at 24 hours after E . histolytica challenge by flow cytometry . As expected , gut but not systemic neutrophil numbers were depressed in antibiotic treated mice ( Fig 4e & 4f ) . In addition , CXCR2 expression on neutrophils was lower both in blood and cecum ( Fig 4g ) , whereas expression of the other surface molecules were either not different or higher in antibiotic pre-treated mice ( S5c & S5d Fig ) . Next , cecal tissue at 24 hours after challenge was stained by anti-Ly6G antibody in order to assess neutrophil localization in cecal tissue ( Fig 4h ) . We confirmed mucosal thickness was not different between 2 groups ( Fig 4i ) , then compared the frequency of Ly6G positive neutrophils in the mucosa ( cells per 10 crypts ) and submucosal tissue ( cells per unit area ) . Neutrophils in the mucosa were lower in antibiotic pre-treated mice ( Fig 4j ) whereas neutrophil number in submucosal tissue was not different between the 2 groups ( Fig 4k ) , suggesting that efficient neutrophil migration from submucosal tissue to the site of infection , which plays an important role in protection from tissue invasion of microorganisms [30 , 31] , was disturbed in antibiotic pre-treated mice . These results indicate that lower expression of CXCR2 on neutrophils was a potential cause of impaired neutrophil recruitment to the infection site , which resulted in a susceptible phenotype to E . histolytica challenge in the antibiotic pre-treated mice . In order to assess the impact of chemokine mediated neutrophil recruitment via CXCR2 on E . histolytica infection , we tested if blocking of CXCR2 would render control ( non antibiotic treated ) mice susceptible to E . histolytica . CXCR2 was neutralized using a monoclonal antibody ( rat anti-mouse CXCR2 IgG2A ) injected intraperitoneally 2 hours before E . histolytica challenge [32] . Neutrophil number after E . histolytica challenge in cecum , but not in peripheral blood , was suppressed in anti-CXCR2 pre-treated mice compared to isotype control treated mice , although both are not statistically significant ( Fig 5a & 5b ) . E . histolytica burden at 24 hours after challenge was significantly higher in anti-CXCR2 pre-treated mice compared to isotype control treated mice , although lower than in the antibiotic pre-treated mice ( Fig 5c ) . Also , tissue invasion of E . histolytica was more severe ( Fig 5d ) , and E . histolytica culture from cecal contents was positive in all anti-CXCR2 pre-treated mice as seen in antibiotic pre-treated mice ( Fig 5e ) . Interestingly , the impact of CXCR2 blocking on tissue MPO activity ( Fig 5f ) as well as E . histolytica burden and histopathological score was less profound than that seen by antibiotic pre-treatment , suggesting that the low expression of CXCR2 was not the sole explanation for the higher susceptibility of antibiotic pre-treated mice to E . histolytica challenge .
The most important finding of this study is that gut microbiome dysbiosis increases susceptibility to amebic colitis in humans and in the mouse model , and that one mechanism of this increased susceptibility is downregulated neutrophil recruitment to the gut . In children who developed symptomatic colitis by E . histolytica , the gut microbiota had lower diversity than those in children who showed colonization . Moreover , we found that children developing amebic colitis had lower microbiome diversity in the month preceding amebic colitis than those without E . histolytica infection , although these data came from 2 different cohort studies at the same location . We extended these studies in the mouse model of amebic colitis , demonstrating that mechanisms of dysbiosis-mediated increased susceptibility were downregulation of neutrophil CXCR2 and attendant failure of neutrophil recruitment to the site of infection . Although changes in gut microbiota can induce or conversely reduce gut inflammation , dysbiosis ( represented by lower microbiota diversity ) in this case reduced the neutrophil activation in the gut , allowing more severe intestinal damage by E . histolytica ( Fig 6 ) . Antibiotic use is pervasive in children in low and middle income countries and a likely cause of dysbiosis [33] . In fact , we found that antibiotics were frequently used in participants of PROVIDE study in Bangladesh with an average of 25 . 5 new prescriptions over first 2 years of life , although such data was not available in NIH birth cohort . It is also known in mouse studies that the gut microbiome is essential not only for the development of the immune system but also for maintenance of homeostasis [34 , 35] . In the present study , we sought to mimic our cohort observation in mice by administration of an antibiotic cocktail ( ampicillin , neomycin , vancomycin and metronidazole ) [29 , 36–40] . Strikingly , C57BL/6 mice , which are naturally resistant to E . histolytica infection [41] , had increased susceptibility after antibiotic induced dysbiosis . We confirmed that mucus secretion was disturbed with decreased IL-25 level at the site of infection in antibiotic pre-treated mice , which might allow more severe tissue invasion of E . histolytica [42] . Interestingly , neutrophil number and activation at the site of infection were suppressed as assessed by fecal lipocalin-2 , myeloperoxidase activity and neutrophil number despite more tissue damage by E . histolytica in antibiotic pre-treated mice . We concluded that insufficient neutrophil recruitment as well as suppressed mucus secretion at the site of infection allowed E . histolytica tissue invasion at 24 hours after infection ( Fig 3 ) , which were followed by more aggressive colitis with more severe weight loss and delayed clearance of E . histolytica from stool ( Fig 2 ) . Upon intestinal damage by infection , DAMPs and PAMPs from damaged tissue and pathogen respectively are recognized by intestinal epithelial cells and antigen presenting cells ( APCs ) . This is followed by the production of neutrophil chemoattractant chemokines including CXCL1 and CXCL2 mainly produced by activated tissue residential macrophages [43] , which in turn induce neutrophil recruitment to the site of infection [44] . However , CXCL1 and CXCL2 as well as IL-1β upon E . histolytica infection were significantly higher in antibiotic pre-treated mice compared to untreated control mice , indicating that increased susceptibility to infection was not due to poor recognition of PAMPs nor DAMPs [29] but instead due to failure of neutrophil response to chemokines . We checked the expression of CXCR2 , which is the main chemokine receptor on neutrophils for CXCL1 and CXCL2 [45] , because expression of CXCR2 is critically important for neutrophil associated protection against various infectious diseases [32 , 46–48] . We found that CXCR2 expression on neutrophils both in blood and cecum after E . histolytica challenge was lower in antibiotic pre-treated mice than that in untreated control mice . Finally , we confirmed that anti-CXCR2 pre-treated mice showed significantly higher E . histolytica burden than untreated control mice at 24 hours after E . histolytica challenge . From these results , we concluded that down-regulation of CXCR2 is an important but not sole mechanism by which dysbiosis diminished neutrophil mediated protection . Furthermore , our results indicate that dysbiosis might affect the susceptibility to infection by other organisms or inflammatory bowel disease [32 , 46–48] by altering neutrophil recruitment or function , by suppressing CXCR2 expression [49] . On the other hand , our results was contradict with the previous paper reported by Houpt et al . which demonstrated that C57BL/6 mice remained resistant to E . histolytica infection after depletion of neutrophils [14] . Although this might be explained by the fact that Houpt et al . did not disrupt the microbiome with antibiotics prior to infection of C57BL/6 mice with antibiotics , further investigations for elucidating the impact of microbiota on neutrophil function , including not only recruitment of neutrophils but also amebicidal activity , will be needed in future study . There are some limitations to be considered . First , our results in Bangladeshi children have not been generalized to children in other countries , although they do represent two different cohorts ( PROVIDE study and NIH birth cohort ) performed at the same district ( Mirpur , Dhaka , Bangladesh ) . Second , we created dysbiosis in the murine model by administering a broad spectrum antibiotic cocktail . However , the composition of the gut microbiome is different between human and mice [50] . The bacterial species which influence disease severity of amebiasis are therefore yet to be fully identified although one paper has shown the relationship between the presence of Prevotella copri and disease severity of amebic colitis [7] . A third limitation is that we do not understand the mechanism of microbiota-mediated changes in CXCR2 expression on neutrophils upon E . histolytica infection . Fourth we have not assessed amebicidal activities of neutrophils although our results suggested that they are suppressed during antibiotic induced dysbiosis ( Fig 5 ) . In the future the colonization of mice by specific bacteria [8 , 51] and assessing amebicidal activity of their neutrophils [17] may be used to identify the bacteria responsible for resistance and to elucidate mechanisms of protection . Finally , we could not assess neutrophil CXCR2 in the bloodstream of the children with E . histolytica infection to test the extent that it was downregulated as in the mouse model . Future work should assess dynamically the impact of antibiotic use on neutrophil function during amebic infection in a human cohort . In conclusion , we demonstrated that antibiotic-induced dysbiosis mediated susceptibility to amebic colitis through decreased neutrophil recruitment to the gut , at least partially due to decreased surface expression of neutrophil CXCR2 . Demonstration that antibiotics can impair neutrophil chemotaxis to a site of infection may be of broader importance than just amebiasis for the insight that it provides into gut microbiome regulation of the immune system .
Stool samples and the clinical information from children who developed amebic colitis came from Performance of Rotavirus and Polio Vaccines in Developing Countries ( PROVIDE ) study , whereas those from children who showed E . histolytica colonization were collected in NIH birth cohort . Details of the PROVIDE study and NIH birth cohort were described previously [19 , 20] . In brief , the PROVIDE study was performed at a peri-urban slum of Mirpur in Dhaka and included 700 infants enrolled at the first week of their age , and followed up at least until 53 weeks ( 2 year ) of age . NIH birth cohort is also prospective cohort study of enteric infections in infants from the same study site as PROVIDE . In both studies , written informed consents were provided by parents or guardian on behalf of all infant participants at enrollment . Field workers visited the child’s household twice weekly to capture any diarrhea event and collected stool samples from these as well as a monthly surveillance stool sample . The diagnosis of amebic colitis was confirmed by qPCR assay [7] . The protocol and informed consent ( English and Bangla ) and all amendments were reviewed and approved by the Research Review Committee ( RRC ) and Ethics Review Committee at the International Centre for Diarrhoeal Disease Research , Bangladesh ( icddr , b ) and the Institutional Review Board of the University of Virginia prior to implementation . Trial registration: ClinicalTrials . gov NCT01375647 for PROVIDE study and NCT02764918 for NIH birth cohort . All animal experiments conducted in this study were carried out in accordance with the Animal Welfare Act and the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All procedures were approved by the Institutional Animal Care and Use Committee of the University of Virginia ( Protocol Number: #4126 ) . Female 6-week-old C57BL/6 mice ( Jackson Laboratories , Charles River ) were housed in a specific-pathogen-free facility in micro-isolator cages and provided autoclaved food ( Lab Diet 5010 ) and water . Mice were treated with an antibiotic cocktail , consisting of 1 . 0 gram per liter of ampicillin ( Sigma ) , neomycin ( Sigma ) and metronidazole ( Sigma ) and 0 . 5 gram per liter of vancomycin ( Hopsra ) in drinking water , as previously used to induce dysbiosis [29 , 36–40] . All antibiotics except metronidazole were continued until harvest , with metronidazole discontinued 72 hours prior to E . histolytica challenge due to potential amebicidal activity . Twenty microgram of rat monoclonal anti-mouse CXCR2 antibody ( clone # 242216 , R&D ) or 20 μg of rat IgG2A isotype ( R&D ) control was injected into the peritoneal cavity 2 h before infection [32 , 52] . Trophozoites for intracecal injections were originally derived from laboratory strain HM1:IMSS ( American Type Culture Collection ) that have been sequentially passaged in vivo through mouse cecum . Cecal contents of infected mice were cultured in complete trypsin-yeast-iron ( TYI-33 ) medium supplemented with Diamond vitamin mix ( JRH Biosciences ) , 100 U/ml of penicillin and 100 μg/ml of streptomycin , and bovine serum ( Sigma-Aldrich ) [41 , 53] . Prior to injection , trophozoites were grown to log phase , and 2 x 106 parasites were suspended in 150 μl culture medium and injected intracecally [41] . Body weight and clinical score were monitored weekly before E . histolytica challenge and every 24 hours after challenge . Clinical score was calculated as the total score of 6 different variables [54] ( S1 Table ) . DNA was extracted from 50 mg of stool as described later . For ELISA , 50 mg of stool was reconstituted in 1 mL of PBS containing 0 . 1% Tween 20 ( 100 mg/ml ) and vortexed for 20 min to get a homogenous fecal suspension . These samples were then centrifuged for 10 min at 12 , 000 rpm at 4°C . Clear supernatants were collected and stored at -20°C until analysis . Lipocalin-2 levels were estimated in the supernatants using Duoset murine Lcn-2 ELISA kit ( R&D Systems ) [21 , 55] . Additionally , inflammation-associated rectal bleeding was assessed by examination of blood in the stool by Hemooccult II SENSA ( Beckman Coulter ) [56] . Two hundred microliter of cecal contents were collected at sacrifice of the infected mice , and cultured in complete TYI-S-33 medium with supplemental antibiotics for 3 days at 37°C . E . histolytica infection was confirmed by the direct microscopic examination of cultured tube at 72 hours . Infection rate was presented by the ratio of infected mice among total number of mice challenged by E . histolytica . After fixation for 24 hours in Bouin’s solution , mouse cecal tissue was washed and stored in 70% ethanol . Paraffin embedding , H&E and PAS staining were processed by the University of Virginia Research Histology core . Immunochemistry ( IHC ) staining was performed by the University of Virginia Biorepository and Tissue Research Facility . IHC staining was performed using the Dako Autostainer Universal System with a primary antibody directed against E . histolytica macrophage migration inhibitory factor [22] , rabbit anti-MUC2 polyclonal antibody ( Cat . No . PA5-21329 ) or rat anti-mouse Ly6G ( clone No . 1A8 , Biolegend ) . Scoring was based on intensity and abundance of EhMIF staining in mucosa by crypt invasion of E . histolytica , and based on abundance of mucin containing cells for goblet cell score ( staining scale was between 0 and 5 ) . Histopathological scoring was done by three independent blinded scorers . Myeloperoxidase ( MPO ) activity was determined as previously described [23 , 57] . In brief , cecal tissues were homogenized in hexadecyltrimethylammonium bromide buffer . Samples were centrifuged at 10 , 000 g ( 10 minutes , 4 degree ) , and the supernatant was collected , and total protein concentration was assessed by a BCA assay according to the manufacturer’s instructions ( Pierce ) . Two hundred μl of 0 . 53 μmol/L o-dianisidine dihydrochloride with 1% hydrogen peroxide were added to 7 uL of supernatant and the absorbance was determined at 450 nm . Results were normalized to total protein concentration . Resected cecal tissue was bead beaten for 1 min in 250 μl of lysis buffer I ( 1× HALT protease inhibitor ( Pierce ) , 5 mM HEPES ) . Lysis buffer II ( 250 μl ) was added ( 1× HALT protease inhibitor , 5 mM HEPES , 2% Triton X-100 ( Sigma-Aldrich ) ) and the tubes were inverted gently . Tissue samples were incubated on ice for 30 min , followed by a 5 min spin at 13 , 000 x g at 4°C . The supernatant was removed to a fresh tube , and total protein concentration was assessed by a BCA assay according to the manufacturer’s instructions ( Pierce ) . IL-1β was measured using Ready-Set-Go ! ELISA kit ( eBioscience ) . CXCL1 , CXCL2 and IL-25 were measured using R&D Systems Duoset ELISA kits . All procedures are performed according to the manufacturers’ instructions . All data were expressed relative to total protein concentration . For DNA extraction from stool ( 50 mg ) or cecal contents ( 200 μL ) , QIAamp Fast DNA Stool Mini Kit ( Qiagen ) was used . All samples were bead beaten for 2 min prior to DNA extraction . All other procedures were performed according to the manufacturer’s instruction . For RNA purification , RNeasy isolation kit ( Qiagen ) was applied for 50 mg of cecal tissue sample . RNA was reverse transcribed with the Tetro cDNA synthesis kit ( Bioline ) . All procedures were performed according to the manufacturer’s instructions . For the quantification of E . histolytica , a standard curve was prepared from trophozoites , and quantitative PCR ( qPCR ) targeting small subunit ribosomal RNA gene [58] was utilized . Probe , primers and annealing temperature ( AT ) were as follows: Eh-probe: Fam/TCATTGAATGAATTGGCCATTT/BHQ; Eh-forward: ATTGTCGTGGCATCCTAACTCA; Eh-reverse: GCGGACGGCTCATTATAACA , AT: 62 . 4°C . MUC2 gene expression was quantified by qPCR using Sensifast SYBR & Fluorescein Mix ( Bioline ) . Gene expression was normalized to the GAPDH gene expression . Primer sequences and annealing temperature ( AT ) used in this study were as follows; MUC2 gene ( forward: 5’- GCTGACGAGTGGTTGGTGAATG - 3’; reverse: 5’ - GATGAGGTGGCAGACAGGAGAC - 3’; AT: 60 . 0°C ) and GAPDH ( forward: 5’ -AAC TTT GGC ATT GTG GAA GG - 3’; reverse: 5’ -ACA CAT TGG GGG TAG GAA CA – 3’; AT: 62 . 4°C ) . DNA was extracted from fecal material using a modified QiaAmp stool DNA extraction protocol which incorporates a 3 min “bead-beating” step as per standard study protocols [59] . Human DNA was removed from E . histolytica positive samples using a Microbiome DNA Enrichment Kit used by the manufacturer’s direction ( NEB ) . The 255bp V4 region was completely sequenced in both forward and reverse orientation using the Miseq V3 kit ( also used by the manufacturer’s direction ) . The sequencing library was prepared using phased Illumina-eubacteria primers to both amplify the V4 16S region rDNA ( 515–806 ) , add the adaptors necessary for illumina sequencing and the GOLAY index necessary for de-multiplexing after parallel sequencing [60 , 61] . As a positive control , DNA extracted from the HM-782D Mock Bacteria Community ( ATCC through BEI Resources ) was added , and as a control for reagent and laboratory contamination a no-template control reaction was added . Sequencing , quality control and OTU assignation using the QIIME pipeline was performed by the Institute for Genome Sciences Core facility ( Baltimore ) . The data was then visualized and Shannon Diversity Scores determined using the Seed Program [62] . Stool samples were collected from PROVIDE study children . Single cell suspensions from cecal lamina propria were prepared as previously described [54 , 63] . Briefly the tissue was thoroughly rinsed in Hank’s balanced salt solution ( HBSS ) supplemented with 5% FBS . Epithelial cells were removed by gentle shaking for 40 min at 37°C in HBSS with 15 mM HEPES , 5 mM EDTA , 10% FBS and 1 mM dithiothreitol . Halfway through the incubation , cecal tissue samples were transferred to fresh buffer . Next , cecal tissue samples were thoroughly chopped using scissors and digested in RPMI 1640 containing 0 . 17 mg ml–1 Liberase TL ( Roche ) and 30 μg ml–1 DNase ( Sigma ) . Samples were digested for 30 min at 37°C with shaking . Samples were then spun down at 300 x g and resuspended in HBSS with 5% FBS and 25 mM HEPES before passage through a 100 μm cell strainer followed by a 40 μm cell strainer ( both Fisher Scientific ) . Cells were counted and the density adjusted to 5 × 106 cells per ml . Cell suspensions ( 200 μl ) were aliquoted into each well of a 96-well plate for antibody staining . For staining , cells were initially blocked with TruStain fcX ( anti-mouse CD16/32 antibody , BioLegend ) for 15 min at room temperature . Cells were spun down and resuspended in LIVE/DEAD Fixable Aqua ( Life Technologies ) for 30 min in the dark place at room temperature . Cells were washed twice and stained with fluorochrome conjugated antibodies . Flow cytometry was performed on an LSR Fortessa cytometer ( BD Biosciences ) and all data analysis was performed via FlowJo ( Tree Star ) . Data was analyzed as the ratio to CD45 positive cells or cell counts calculated by counting beads ( Molecular Probes ) . Fluorescent conjugated antibodies used for flow cytometry are shown in S2 Table . ANOVA was used for differences among multiple groups . Welch’s unequal variance t-test , Mann-Whitney U-test or chi-squared test were used as appropriate for comparing valuables in 2 groups . A p value below 0 . 05 was considered significant . All statistical tests were done using GraphPad Prism software . | Amebiasis , caused by intestinal infection of Entamoeba histolytica , is one of the leading causes of parasite infection-related mortality and morbidity around the world . However , pathogenesis , such as determinant factors of infection outcome , is still unclear although recent data indicate that the gut microbiome plays an important role . In the present study , we firstly found that dysbiosis , which was represented by a lower Shannon diversity index of the gut microbiota , occurred symptomatic E . histolytica infection in children living in endemic area . In mouse model , we demonstrated that dysbiosis induced by antibiotic pre-treatment increased the severity of amebic colitis due to decreased neutrophil activity as well as decreased IL-25 associated mucosal defense in the gut . Moreover , we demonstrated surface expression on neutrophils of CXCR2 was diminished in mice with dysbiosis , which resulted in decreased neutrophil recruitment to the gut . This study is of fundamental importance in amebiasis research for the discovery of a mechanism of microbiome-mediated resistance to amebiasis via neutrophil trafficking to the gut . The work is importantly of broad interest in infectious diseases and immunology for the discovery that neutrophil mediated protection can be disturbed by dysbiosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"blood",
"cells",
"antimicrobials",
"children",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"microbiome",
"drugs",
"immunology",
"microbiology",
"parasitic",
"protozoans",
"animal",
"models",
"colitis",
"age",
"groups",
"model",
"organisms",
"protozoans"... | 2017 | Microbiome-mediated neutrophil recruitment via CXCR2 and protection from amebic colitis |
Viral production from infected cells can occur continuously or in a burst that generally kills the cell . For HIV infection , both modes of production have been suggested . Standard viral dynamic models formulated as sets of ordinary differential equations can not distinguish between these two modes of viral production , as the predicted dynamics is identical as long as infected cells produce the same total number of virions over their lifespan . Here we show that in stochastic models of viral infection the two modes of viral production yield different early term dynamics . Further , we analytically determine the probability that infections initiated with any number of virions and infected cells reach extinction , the state when both the population of virions and infected cells vanish , and show this too has different solutions for continuous and burst production . We also compute the distributions of times to establish infection as well as the distribution of times to extinction starting from both a single virion as well as from a single infected cell for both modes of virion production .
The earliest events in infection are stochastic . Whether exposure to virus leads to systemic infection or complete elimination of the virus can be a matter of luck , particularly when exposure is to low levels of virus . For example , the transmission probability for HIV infection is – per coital act [1]–[3] . In 80% of HIV infections of heterosexuals , a single viral strain is transmitted or founds the infection [4] . In most cases after sexual exposure to HIV , infection fails to take off . When it does take off it likely does so from a single infectious virion or a single infected cell . Whether exposure to virus , be it HIV or the common cold , results in persistent infection or elimination hinges on numerous poorly understood factors including antibody and innate immune responses , virus specific cytotoxic T lymphocyte responses , the spatial distributions of these components [5] , and pure chance . Here we study some simple viral infection models in a stochastic setting using HIV as a model system . The models that we consider are relevant for the earliest stages of infection before target cells are depleted to any extent and before immune responses are stimulated . Thus , we consider models with no immune response in which the number of target cells , , is held fixed and where we consider only variations in the number of virions , , and the number of infected cells , , with and being non-negative integers . We derive exact analytic expressions for the extinction probability , i . e . , the probability that the virus and all infected cells are completely eliminated from the host , for two related models that differ in the manner in which virus is produced . We also present simulation results for the conditional mean time to observable infection . The extinction problem is related to the classic “gambler's ruin” problem [6] , which Pascal [7] first solved and then posed to Fermat , hoping in vain to stump him , and to Huygens [8] who thought there might be some applicability to disease and wrote “For what can there be in common between the Value of a Chance in a Game , and the Knowledge and Cure of a Distemper ? And how can the nicest Determination of the former , any way influence or illustrate the latter ? ” More recently Tan and Wu [9] developed a -dimensional stochastic infection model for HIV that incorporated target cells and latently infected cells and studied it via Monte Carlo simulations . They noted that there was positive probability that the virus could be eliminated by the process [9] . Monte Carlo approaches were also used by Kamina et al . [10] and Heffernan and Wahl [11] to study the probability that an infection would not become established after exposure to a viral inoculum of a given size . Tuckwell and Corfec [12] , [13] developed similar multi-dimensional models to study early infection but modeled them as diffusion processes via simulation of stochastic differential equations . Merrill [14] modeled early infection as a branching process that kept track of the number of infected cells but not of virions . Lee et al . [15] also modeled only infected cell dynamics during acute infection but focused on the stochastic changes in HIV genetic sequences starting from an infection initiated by a single HIV genome . Tuckwell et al . [16] studied the probability of viruses entering a host infecting one or more target cells before being cleared , but did not carry out a detailed analysis including infected cells . Haeno and Iwasa [17] developed a stochastic model of early infection in order to study the generation of drug resistant virus in an exponentially expanding viral population . Ribeiro and Bonhoeffer [18] also develop a stochastic simulation of early infection in which only infected cells are follwed to study the best time to start antiretroviral therapy in a model with stochastic generation of drug resistant mutants . In this manuscript we model early infection as a discrete random process in which both the number of virions and the number of infected cells are followed . The form of the models that we develop are similar to those used in epidemiology to study the spread of infection from person to person [19] . As such we will find that the basic reproductive number , , first introduced in epidemiology to denote the average number of people infected by one infected person put into a population of susceptibles , plays a role in our analysis . Here will denote the average number of new cells infected by one cell during its lifetime when placed in a population of fully susceptible cells . As in epidemiology , we will find that when infections will surely die out and when there is a positive probability that the infection will die out . Our goal here is not simply to reiterate these well known results but rather to uncover basic features of early HIV infection and to study the differences between continuous and burst viral release .
One of the simplest infection models consists of virions , ( V ) , target cells ( T ) , and productively infected cells ( I ) with transitions [20]: ( 1 ) where denotes the empty set and indicates that infected cells or virus is being cleared . The symbols above the arrows denote the rates of the various processes , where is the rate constant characterizing infection , is the death rate of infected cells , is the viral burst size , i . e . , the total number of virions produced by an infected cell over its lifetime , is the rate at which infected cells produce virus , and is the virion clearance rate [20] . In some models , particularly those in which a cytolytic lymphocyte response may affect lymphocyte lifespan , the symbol is used to denote the virion production rate rather than [20] . Here , where we focus on the earliest events in infection , before there is an immune response , using for the virion production rate allows us to simplify some expressions . Also , because we are focusing on early infection we neglect variations in the number of target cells . This is justified because , as we show below , only a tiny fraction of target cells need be infected to insure that the infection will persist . Thus the model above can be written: ( 2 ) We call the model specified by Eq . ( 2 ) the “continuous production” model because once a cell is infected it produces virus continuously throughout its life . A slightly different but related model is given by the set of transitions ( 3 ) We call the model specified by Eq . ( 3 ) the “burst” model because once a cell is infected it releases all its virus in a single burst simultaneous with its death . Although an infected cell may not burst as in a lytic phage infection of bacteria , HIV may be rapidly produced towards the end of an infected cell's lifespan as in other retroviral infections [21] . Also , because we are studying very early infection , before immune responses begin , we assume death of a cell is due solely to viral cytopathic effects and hence ignore the possibility that death occurs before virions are released . Both models have identical mean-field kinetics given by: ( 4 ) where and are the concentrations of virus and infected cells . At the deterministic level the burst and continuous production models make the same predictions . Note that this model differs from the “standard” model of viral infection in that viral clearance occurs at rate rather than at rate , i . e . , the model keeps track of the fact that one virus is lost every time a cell is infected . However , since is a constant the model is equivalent to the standard model in which in the standard model incorporates virion loss due to infection [22] . Note that the origin ( ) is a steady state of the deterministic system . The origin is a stable steady state provided the basic reproductive ratio , where is the number of new cells infected by an infected cell during its lifetime with ( 5 ) Although this is easily seen by calculating the determinant of the linear system specified in Eq . ( 4 ) it is worth noting that for HIV , ( ) is large compared to and virions become “slaved” to infectious cells [23] , so that , which results in . We show that if virus and all infected cells will be eliminated with certainty . Unlike deterministic models , for there is still a finite probability that the virus and all infected cells will be eliminated stochastically . We shall also see that once the virus “takes off” , it roughly satisfies the slaving approximation , while before it takes off the dynamics are fundamentally stochastic .
We consider systems which can be fully specified by a state vector . For both the burst and continuous production models , where and are the number of virions and infected cells , respectively . Upon a transition the state is incremented by one of the transition vectors where is the maximum number of transitions the system can make out of any state . For the continuous production model we have and , , , and . The rate of the reaction is given by . Thus , for the continuous production model there are four types of reactions: ( 1 ) infection with rate , ( 2 ) viral production with rate , ( 3 ) death of a infected cell with rate , and ( 4 ) virion clearance with rate . The probability that the reaction is the next reaction is given by Gillespie's algorithm [24]: ( 6 ) where ( 7 ) For the continuous production model , , , , , and . The time of the next reaction is a random variable with distribution . For the burst model , , , , and the corresponding reaction rates are , , and . Our goal is to determine the probability that an exposure to virus eventually evolves to “extinction” , i . e . , . Throughout this article , we refer to the loss of all virus and infected cells from the host as “extinction” and to the decay of virus as “clearance” . Stochastic extinction is a multi-dimensional analogue to the classic gambler's ruin problem first solved by Pascal [7] . The extinction probability , from state , satisfies [6] , [25]–[27]: ( 8 ) ( 9 ) Equation ( 8 ) can be understood from Figure 1 . If the system starts out in state on the first transition the state will jump to one of the states , , with probability . Clearly , then the extinction probability from state is the weighted sum of the extinction probabilities at the neighboring sites where the weights are just the probabilities of making the individual transitions . Note that is always a solution since . Although the general solution is intractable we will show that if processes of virion and infected cell extinction are independent , the functional equation for can be reduced to an algebraic one . Since each virus and infected cell acts independently in our model , we assume: ( 10 ) where and are the probabilities that a process initiated with a single virion or single infected cell , respectively , results in extinction . Using Eq . ( 10 ) , Eqs . ( 8–9 ) can be reduced to algebraic equations for and . In the following we carry out this program for both the continuous and burst models . For the continuous production model , substituting Eq . ( 10 ) into Eqs . ( 8–9 ) yields ( 11 ) where . We convert this system of equations to a pair of algebraic equations by first setting and obtaining a first equation and then setting to obtain another . Note that and . Thus we obtain the pair of equations ( 12 ) ( 13 ) where is the probability that a virion infects a cell . Note from the definition of , . Substituting Eq . ( 12 ) into Eq . ( 13 ) , we obtain a quadratic equation with solutions , and . Substituting into Eq . ( 12 ) , we find and . Since probabilities need to be less than or equal to 1 , ( 14 ) ( 15 ) for the continuous production model single cell and single virion extinction probabilities . Thus , if , , whereas if , and . For the burst model , as we show in the next section , a similar analysis yields ( 16 ) ( 17 ) where is a positive real root of ( 18 ) or equivalently of ( 19 ) Noting that for and using Descartes' rule of signs shows that there are either 2 or 0 real positive roots . Since is one root , there is exactly one other positive root of Eq . 18 , which we denote . Note that if then and there is only one root , . Figure 2 shows the single virion extinction probability , , as a function of for , 5 , and 25 for both the burst and continuous models . For large the extinction probabilities for both models converge to the diagonal line ( ) connecting the upper left to the bottom right corners . For both models the single virion extinction probability , , is a function of and and that for , i . e , for . Also in both models if then and in both cases extinction is certain if . This is not a new result and could be derived from a branching process approach where the process would be subcritical if and then extinction would be guaranteed . Results along this line in the context of epidemiological models are summarized in Britton and Lindenstrand [28] and Britton [29] . Britton [29] also points out that Reed and Frost in a series of unpublished lectures from 1928 study an epidemiological model where all infections are assumed to occur exactly at the end of the infectious period , which is analogous to the burst model where infection can only be transmitted from one cell to another at the end of the infected cell's life . The main difference between the two models is that . The difference between the two models is most easily understood in the limit where the probability of a virion infecting a cell rather than being cleared approaches 1 . Note that in the burst model since the number of cells infected by a single infected cell must be less than or equal to the number of virions produced , and in the continuous production model this is true for the mean . As we find that and . In the limit virus is not cleared in either model but disappears only by infecting another cell . In the burst model all infected cells result in the creation of new virions . Thus , for the burst model , the extinction probability approaches zero as . By contrast , for the continuous model there is a chance that an infected cell will die before it produces any virus . In the infinite limit the single virion extinction probabilities become equal for the two models , i . e . , . We have been focusing on the single virion extinction probability . Note that and that for , , and . In the large limit if then for both models . Thus in the large limit the probability of stochastically clearing the infection is effectively zero if any cells at all are infected , since each infected cell is assumed to produce an arbitrarily large amount of virus in this limit . The earliest stages of HIV and SIV infection have a characteristic “eclipse” phase during which the virus remains below the limit of detectability of current assays . Here we explore the role stochastic effects play in determining the length of the eclipse phase . Using Gillespie's stochastic simulation method [24] we compute the mean time to detectability following a one virion challenge . In Figures 3–10 we use the following parameters for illustrative purposes: , /day , /day [30] , and /day [31] . For these parameter values , , which is lower than the median value of found by Ribeiro et al . [32] , Stafford et al . [33] and Little et al . [34] during primary HIV infection . However , these estimates relied on data obtained after the virus was observable and in the case of Stafford et al and Little et al . mainly after the viral load peak . At earlier stages of infection , could be different . A value of of about has recently been estimated for SIV infection in rhesus macaques [35] . However , not all virions are infectious . In the formulation given above we have assumed all virions are equivalent and hence equally infectious . Although one could generalize the model to include both infectious and noninfectious virions , following only infectious virions has the advantage of allowing one to track smaller numbers of virions in simulations . For virus isolated during chronic infection , approximately one in to virions appear to be infectious [36]–[39] , suggesting that if we model only infectious virus values of between 5 and 50 might be reasonable . As our default , we have chosen a value of consistent with these estimates . Recent work has suggested that virus isolated early in infection has a higher ratio of infectious to noninfectious virus [40] , and thus depending on the source of infecting virus larger values of might be appropriate . Our choices of default values of and are based on estimates derived from data obtained during chronic infection [30] , [31] , and thus they too might not be appropriate for the earliest stages of infection . Lastly , the value of was chosen so that with the other parameter choices a sensible value for was obtained . Thus , while the parameter choices studied here are reasonable guesses based on what we know about HIV infection dynamics , there is some uncertainty about them . Figures 3 and 4 show and for the continuous production and burst models , respectively . As expected , in both cases infection can persist ( top panels in Figures 3 and 4 ) or go extinct ( bottom panels in Figures 3 and 4 ) . Here we have arbitrarily defined persistent infection as . Although it is mathematically possible for the virus to be cleared by chance with , at this point the probability of stochastic extinction is on the order of . This is because for , . Also , virus becomes detectable in plasma with conventional assays when its concentration is 50 copies/ml . Assuming that deterministic equations are appropriate at this point , one finds that if virus and infected cells are at quasi-steady state then . Thus , if each infected cell produces 50 , 000 virions [35] , lives about a day while productively infected [30] and has a clearance rate ( ) of about 23/day [31] , then when , will be approximately 50 copies/ml assuming virus distributes through approximately 1 . 5 liters of extracellular body water in a 7 kg macaque . Thus , by the time the eclipse phase of SIV infection should be over . For HIV infection the volume of distribution is about 10-fold larger ( a 70 kg human has about 15 liters of extracellular body water ) and thus virus detectability would be delayed until is about 10-fold larger . Nonetheless , the probability of extinction would still remain . In the realization that leads to persistent infection in the continuous production model , the initial virus quickly infects a single cell and that cell starts producing new virions . Thus , begins fluctuating from time zero as virions are produced and cleared stochastically ( Figure 3 ) . Further , these released virions infect new cells and rises substantially over the first 2 days of infection . By contrast , in the burst model , in the illustrated realization that leads to persistent infection ( Figure 4 ) , after the first virus infects a cell that cell lives about 1 . 25 days . No additional virus is produced until this cell dies and thus stays at zero until day 1 . 25 at which time a burst of virus is produced . While some of this newly produced virus infects new cells , the rest gets cleared and returns to zero until another cell dies at approximately 1 . 8 days . Additional cells are infected at this point and rises due to this and subsequent bursts of virus . Realizations that lead to extinction are shown in the lower panels of Figures 3 and 4 . Note the y-axes are scaled differently than in the cases that lead to persistent infection . In the continuous production case , by chance most of the produced virus is cleared and thus never gets above 3 . Also , the number of infected cells remains small , reaching , before these cells sequentially die and extinguish the infection . In the burst model , even though in the realization shown 10 virions are produced in each of three bursts , the first two bursts only lead to the infection of 1 cell each and virions in the last burst are all cleared without infecting any cells leading to the extinction of the infection . Because a particular realization may not be representative of a stochastic process , we show in the left column of Figure 5 100 realizations of the continuous production model that lead to infection starting from a single virion , and in the right column 100 realizations that lead to extinction . Figure 6 is the same as Figure 5 except the initial condition is , i . e . , the infection is started by the introduction of a single infected cell . During sexual transmission of HIV it is not known whether infected cells or virus particles penetrate epithelial layers and initiate infection . For the burst model , Figures 7 and 8 show 100 realizations each of infection and clearance for the initial conditions ( ) and ( ) respectively . It can be seen that in none of the burst model realizations that lead to extinction were there ever more than a single infected cell . By contrast , the continuous model had several realizations in which 2 or 3 cells were infected but still went to extinction . Infected cells in the burst model always produce infectious virions ( here ) . Infected cells in the continuous model realizations that led to extinction never produced more than 4 infectious virions total even though there were as many as 3 infected cells . The differences in the two models are fairly evident in the sets of realizations that lead to extinction . The differences in the realizations that lead to infection are not evident to the naked eye because the particle numbers start to get large and the models converge towards mean-field dynamics . In a stochastic model each infection can have a different course and the scenarios described above even with 100 realizations need not be representative . We thus ran simulations until 100 , 000 realizations resulted in infection . For the continuous production model this occurred after a total of 429 , 639 simulations had been performed . Of these 429 , 639 simulations 329 , 639 resulted in extinction and 100 , 000 in infection . The resulting fraction of simulation that went extinct , 0 . 767 , is in accord with our calculation of . Note that for the continuous model . With and we find the extinction probability is . For the burst model the single virion extinction probability , Eq . ( 18 ) , gives . To check this , we performed simulations using the burst model until realizations resulted in infection . To achieve this a total of 306 , 592 simulations were performed . Of these 306 , 592 simulations , 206 , 592 resulted in extinction and 100 , 000 resulted in infection , yielding a 67 . 38% chance of extinction , in accord with the predicted value , . ( The expected value for the number of extinctions in 306 , 592 Bernoulli trials is 206 , 336 and the standard deviation is 259 . 8 . ) The analytical results that we have derived for the extinction probabilities do not provide any information about dynamics . Thus , the stochastic process could take hours , days or months before extinction is reached . To gain insight into these dynamics , we have plotted in Figures 9 and 10 , for infections starting with a single virion , the fraction of simulations that go extinct at various times after infection , with Figure 9 for the continuous production model and Figure 10 for the burst model . These histograms represent the distributions of time to extinction conditioned on the process ultimately going extinct . Both the continuous production and burst models have a sharp initial decay in their conditional distributions of times to extinction . One might expect the extinction rate to be proportional to , since that is the rate at which virions are cleared . However , from the graphs one can deduce that the initial decay occurs on a time scale given by . Since new cells are infected at rate it is not completely self-evident that the initial decay should be given by . The fact that the initial decay is given by rather than just can be understood in terms of a simple 3-state Markov chain , where the initial state is , a single virion , i . e . , , represents the extinction of the infecting virion , i . e . ( 0 , 0 ) , and representing the virion infecting a new cell , i . e . ( 0 , 1 ) . Given that extinction occurs , consider the conditional distribution of times for the system to make the transition from to . The probability that the system remains in state given that it was in state at time , is just where and are the transition rates from to and to , respectively . The probability flux from to is just . Let be the conditional probability that the system makes the transition into for the first time at time , given that it was in state at time . Then . The conditional distribution of first passage times from to is then , , where is the probability that the system transitioned into from . This is exactly analogous to the sharp initial decay with rate from the single virion initial condition . After the initial transient , the distributions of times to extinction display long tails that decay roughly with rate . In both models the long tails are caused by the infection of cells . Once a cell is infected it takes much longer to reach extinction , on average , than before any cells are infected . The difference between the two models is largely due to the difference between the single infected cell extinction probability , in the continuous and burst models , i . e . , and for our default parameter values . Extinction from an infected cell is much less likely for the burst model than for the continuous model . Thus there is substantially more probability in the tails ( of the distribution of times to extinction starting from a single virion ) for the continuous model than for the burst model . For the continuous model we have derived approximate analytic solutions to the full problem that we shall present elsewhere . To further highlight the difference between the models , we examined the time needed to obtain a 95% probability of extinction given that the process goes extinct . For the default parameter values , the burst model reaches 95% ( conditional ) probability that the infection is extinct after about 2 . 5 hours , whereas the continuous model reaches this probability of extinction after about a half day . Thus , there is a significant difference in the behavior of systems governed by the continuous production and burst models . Note also that the conditional distribution of times for an arbitrary number of virions to go extinct can be inferred from the conditional single virion distribution of extinction times . The time to extinction is difficult to determine experimentally , while the time to observable infection is not . Thus , we have studied the time it takes for infection to reach , which as we have argued above is essentially the time for SIV to be detectable in a rhesus macaque , and which is also a measure of the time to reach a state comparable to established infection . For both the continuous and burst models we generated a 100 , 000 realizations in which is reached . For these simulations , the distribution of times until 32 cells are infected is shown in Figure 9 and Figure 10 for the continuous production and burst models , respectively , and with infections initiated either with a single virion or with a single infected cell . The mean time to reach 32 infected cells in the burst model is 2 . 46 days and in the continuous production model 1 . 75 days for either initial condition . Here the two initial conditions give essentially the same result . In an infection started with a single virion , if the virion is cleared the process goes extinct . Since we have conditioned on this not occurring , the initiating virion must infect a cell , and hence it quickly generates the same state as initiating infection with a single infected cell . One also expects the burst and continuous models to converge to statistically indistinguishable behavior once the particle numbers are sufficiently high , well before there are 32 infected cells . The differences in the mean time to reach 32 infected cells starting from a single infected cell is substantial . This is because the early dynamics are dominated by stochastic effects . In Figure 11 we have plotted the mean time to infection from ( and ) for the two models . For the differences are substantial but decrease with increasing .
The dynamics of acute HIV and SIV infection have been modeled deterministically by a number of authors [32] , [33] , [41]–[43] , and in some cases these models have been used to fit data and extract best-fit parameter values . However , despite the success of these models they do not properly capture the very earliest dynamics of infection where stochastic effects may play a large role . Recent data has convincingly established that a large fraction of infections are established by one or a few infectious virions or infected cells [4] , [44]–[48] . If during sexual transmission only a few virions or infected cells are actually transmitted from one infected person to another then one would expect that a large fraction of sexual encounters between an infected and uninfected person might not lead to successful viral transmission . Epidemiological studies support this and have concluded that HIV transmission occurs at frequencies of between 1 in 100 and 1 in 1 , 000 coital acts [3] . Similarly , experimental studies of SIV infection by intrarectal inoculation of virus has shown that at low doses not every encounter with virus leads to detectable infection and that there is substantial variability in the number of inoculations needed to establish detectable infection [47] . Further , as with HIV when infection was detectable , in most cases it appeared that only one or a few viral genomes established the infection . Lastly , one study aimed at detecting HIV-1 at the earliest possible moments in infection using a qualitative assay that could detect the presence of 4 HIV-1 RNA copies/ml with 95% accuracy showed that in some individuals a period of intermittent low-level viremia preceded the period of steadily rising viremia previously studied with deterministic models [49] . Intermittent low level viremia and frequent extinction of infection is precisely what would be expected by a stochastic model as shown by our stochastic stimulations . A number of previous authors have also performed stochastic simulations of HIV infection [10]–[14] , [18] , [50] . What is novel here is that we have shown that the stochastic extinction probability , , for early infection models is amenable to exact solution under the assumption that clearance of each infecting virion and infecting cell occurs independently . We validated the predictions of this analysis via stochastic simulations based on the standard model of viral infection . That our model and simulations agree is not surprising as in the basic target-cell limited model each virion and infected cell acts independently . One can think of situations where this does not hold; for example , if a threshold number of infected cells is required to generate an immune response that then rapidly clears the infection . Thus , while mathematically it is fairly clear when the independence assumption holds , and most current models of early HIV dynamics that ignore immune responses are consistent with this assumption , whether real viral extinction processes are in fact independent is an experimental question . There is at least one report of experiments on rhesus macaques in which it appears that repeated low dose challenges are cleared independently , suggesting that immune responses are not generated during exposures that lead to viral extinction as assumed by our models [51] . Although we have not done so here , one can use our analytical results on extinction probabilities to explore the parameter ranges that give rise to different probabilities of extinction . For example , if one assumes that extinction occurs 99% of the time so as to yield a 1% chance of infection in a coital act , in which say 1 infectious virion is transmitted to an uninfected individual , then one requires that . Then for the continuous production model , Eq . ( 15 ) , predicts that with one requires , and with one requires . While values of in the literature are higher than this [32]–[34] they were obtained from viral load measurements obtained after the viral level has reached 50 HIV RNA copies/ml or higher . Thus , very early in infection may be much smaller than determined later in infection or may be larger than assumed here . Experimental validation of these possibilities is needed . To further explore potential parameter ranges , Chen et al . [35] estimate that in SIV infection 50 , 000 virions can be released from an infected cell . Further , Ma et al . [40] showed that when 10 SIV particles taken from a recently infected macaque were injected intravenously into two other macaques , both became infected , indicating that the ratio of infectious particles to virions was between 0 . 1 and 1 in this experiment . To see if these numbers make sense in the context of our extinction calculation , assume that of the 50 , 000 virions released were infectious . Also , assume there is only a 0 . 1% chance of infection per coital act as frequently cited for stable couples with low prevalence of high-risk cofactors [3] . Then by Eq . ( 15 ) with , we find , which is in the range estimated by Stafford et al . [33] and Ribeiro et al . [32] for acute HIV infection . This example shows that various parameter estimates in the recent literature are consistent with the findings of our model . However , the fact that the two monkeys injected intravenously with 10 SIV particles became infected is not consistent with the 0 . 1% infection rate per coital act assumed above . Clearly , sexual transmission and direct injection of virus into the blood stream are very different events . Further , if and infectious particles , then from the definition of , Eq , ( 5 ) , ) = , and an estimate of can be made if a value of is assumed . In our simulations we used which yields ( for infectious virions ) , but higher values of are possible depending upon whether one is estimating clearance from blood or lymphoid tissue as recently discussed by De Boer et al . [52] . Clearly , direct measurements of these parameters during acute infection still needs to be done , but these example provide some guidelines as to what we might expect . Our calculations focused on determining and , the probabilities of an infection starting from one virion or from one infected cell going extinct , respectively . Once these probabilities are determined it is straightforward to analyze circumstances where more than one virion or one infected cell initiates infection . For example , assume that infectious virions are transmitted to a recipient and initiates infection . Frequently only one viral genome is identified by sequencing [4] . One explanation for this observation is that of the virions lead to extinction and only one virion founds a successful infection . If we assume that successful infection only occurs in 1 per 1 , 000 coital acts [3] , then or . Further , the probability of only one viral genome founding the infection , given that infection occurs , is given by the conditional binomial distribution , i . e . , , which with , occurs with probability 0 . 9995 . Thus , even if 10 infectious virions are transmitted , if successful infection is rare , as in this example , one is almost assured that only one virus will grow and found the infection . Another unique aspect of our work is that we show in a stochastic setting continuous viral production can be distinguished from viral production that occurs in a burst . In at least one lentivirus , visna virus , the greatest fraction of virus production occurs towards the end of the viral life cycle [21] , more consistent with a burst model than a model with constant continuous production . For HIV it has not yet been established whether a burst or continuous production model is most appropriate . One might envision viral production from a highly activated CD4+ T cell to occur in a process approximating a burst , whereas production from an infected resting CD4+ T cell or from an infected macrophage , where infected cell life spans might be weeks rather than days [53] , might be continuous . In simple deterministic models , such as the standard model of viral infection , burst versus continuous production can not be distinguished , and give rise to identical dynamics . Here we show that the probability of extinction is different for continuous production and burst production and that the time to establish infection differs between these two modes of production . Our core result is that with the burst model one obtains lower extinction probabilities ( see Figure 2 ) and longer times to the establishment of infection than with the continuous production model ( see Figures 9 , 10 and 11 ) , even when the mean number of virions produced is the same . In the continuous production model virus production starts as soon as a cell is infected and these released virions can infect other cells leading to a more rapid establishment of infection than with the burst model . Further , with continuous virion production there is more heterogeneity in the number of virions an infected cell produces owing to the variability in infected cell lifespans . In fact , there is a chance an infected cell will die before producing any virions . This in turn leads to a greater chance of the process going extinct . In epidemic models a similar effect has been noted , where for , increased variability in individual infectiousness increases the probability of stochastic extinction [54] . In the continuous production model we have assumed that the rate of virion production is constant . In prior work using deterministic models to describe HIV dynamics , more realistic models of viral production have been studied in which the rate of viral production varies continuously over the cell's lifespan [55]–[57] . In such models the rate of viral production is described by a function , where denotes the age or length of time a cell has been infected . Our continuous production and burst model are two choices of possible functions , i . e . = constant and being a Dirac delta function . Clearly many other choices are possible . Such age-structured HIV production models have not yet been analyzed in a stochastic context . In the burst model we first assumed that each cell produces exactly virions . As this is unlikely to be true , we then generalized this by allowing to be a random variable . Viral production at the individual cell level still remains to be measured and thus nothing is known about in vivo burst size distributions . Further , in both cases the burst size was not coupled to the cell's lifespan . Another possible extension of our model is to allow the lifespan of a cell to be influenced by the rate of viral production or the viral burst size . Cells that use resources to produce virus rapidly might die sooner . Alternatively , one could envision that the amount of virus produced by a cell is influenced by the cell's lifespan . For example , if a cell produces virus at a constant rate and then releases it in a burst , then a cell that lives longer would have the opportunity to make more virus . Couplings between cell lifespan and viral production have been studied previously in deterministic models by a number of authors [55]–[58] . Because our model is derived from the standard model of viral infection it carries over features and limitations of that model . In particular , both the standard model and our continuous production model assume that once a cell is infected it begins producing virus immediately . Also , in the burst model even if a cell lives an infinitesimal amount of time after being infected it releases a full burst of virus . In reality , many steps of the viral life cycle need to be completed before viral production can occur . It is straight forward to refine our models so that infected cells wait a period of time before they can begin to produce virus . This has been done previously in the context of differential equation models [59] , [60] . Obviously , in the stochastic model the waiting time distributions until extinction or infection are strongly affected by such a modification . On the other hand , the extinction probabilities remain unaltered if one assumes infected cells before they begin to produce virus have negligible death rates . Including death of such cells will require a modification of the extinction probability calculations . The analysis we have presented assumes that target cell levels remain constant . This assumption is valid early in infection if we assume the system is well-mixed as there are approximately CD4+ T cells in a human [61] , and perhaps 10-fold less in a macaque , and our model only follows the infection process until 32 cells are infected . At longer times , once the number of infected cells get large enough to have an impact on target cell numbers , stochastic fluctuations would be of no significance in the context of a well-mixed system and deterministic models should be appropriate . If HIV is introduced into the blood , say through transfusion or by intravenous drug use , then the well-mixed assumption with no target cell limitation would seem appropriate . However , in sexual transmission , one could envision that spatial effects near the site of transmission are important and in the region that the entering virions or infected cells find themselves in target cells maybe rare and hence limiting . Thus , it might be of interest to study the nonlinear problem in which target cell numbers vary . One could also envision situations in which immune responses are included in the model , such as in studies of vaccine-induced protection , and in which stochastic effects are important in describing the early immune response . Such extensions of our model remain to be developed . Lastly , our model has not yet addressed the question of how the initial infecting agents , infectious virions or infected cells , get access to target cells . In experiments involving intrarectal or intravaginal challenge of rhesus macaques large numbers of virions have been introduced , e . g . to in the experiments by Keele et al . [47] . Nonetheless , only one or a few viral genomes were seen to expand in infected animals . Whether larger numbers cross epithelial barriers and are then rapidly eliminated or whether the barrier itself prevents all but a few viral genomes to gain access to target cells and expand is not known , Thus , models and further experiments examining these early steps are still required . In conclusion , we have developed stochastic models of early viral infection in which continuous viral production and burst viral production are distinguished . The models capture the stochastic aspects of some of the earliest events in infection and provide quantitative insights into the possibility that early infection will go extinct rather than become established . We provide analytical solutions for the extinction probability and , via simulation , insights into the distribution of times until infection goes extinct or becomes established . | The dynamics of HIV infection and treatment has been extensively studied using ordinary differential equation models . Recent work on HIV transmission has suggested that most sexually transmitted infections are started by a single virus or infected cell . This observation coupled with the fact that successful HIV transmission only occurs in 1 per 100 to 1 per 1000 coital acts suggests that early events in infection are stochastic . Here we develop a stochastic model of HIV infection and use it to characterize the dynamics of early infection when virus is released from cells either continuously or in a burst . We show that these mechanisms of viral production produce different early dynamics , with different probabilities of extinction and different distributions of time to establish infection . In deterministic models , these modes of viral production are indistinguishable . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
] | [
"mathematics/statistics",
"biophysics",
"immunology",
"virology"
] | 2011 | Stochastic Theory of Early Viral Infection: Continuous versus Burst Production of Virions |
Primary microcephaly is a congenital neurodevelopmental disorder of reduced head circumference and brain volume , with fewer neurons in the cortex of the developing brain due to premature transition between symmetrical and asymmetrical cellular division of the neuronal stem cell layer during neurogenesis . We now show through linkage analysis and whole exome sequencing , that a dominant mutation in ALFY , encoding an autophagy scaffold protein , causes human primary microcephaly . We demonstrate the dominant effect of the mutation in drosophila: transgenic flies harboring the human mutant allele display small brain volume , recapitulating the disease phenotype . Moreover , eye-specific expression of human mutant ALFY causes rough eye phenotype . In molecular terms , we demonstrate that normally ALFY attenuates the canonical Wnt signaling pathway via autophagy-dependent removal specifically of aggregates of DVL3 and not of Dvl1 or Dvl2 . Thus , autophagic attenuation of Wnt signaling through removal of Dvl3 aggregates by ALFY acts in determining human brain size .
Primary microcephaly has mostly been reported as an autosomal recessive trait coupled with mild to severe intellectual deficit [1 , 2] . The developing brain of higher mammals begins with a pseudostratified layer of apical neuroepithelial ( NE ) progenitor ( AP ) cells , which are attached to the apical and pial surfaces maintaining their polarity . At the onset of neurogenesis , NE cells turn into radial glial cells ( RGCs ) that will generate , directly or indirectly , all neurons . The RGCs undergo self-renewing cell divisions , later switching from symmetric to asymmetric divisions , giving rise to RGC daughter cells and differentiating basal progenitor ( BP ) cells which maintain their proliferative state and will later differentiate into neuronal cells [3 , 4] . The number of proliferative division rounds of both APs and BPs prior to their differentiative division is critical for establishing proper brain size and development [3 , 5 , 6] . Therefore , it is not surprising that most genes known to date to be associated with MCPH are involved in the processes of mitosis , cell cycle regulation , DNA replication and primary cilia formation and stabilization . It is believed that premature transition between symmetrical to asymmetrical divisions during brain development is the main cause for primary microcephaly [5–7] . This premature transition results in an insufficient number of precursor cells within the neuronal stem cell ( NSC ) population , and eventually leads to reduced number of neurons in the cortex [5] . To date , 16 loci and genes have been associated with autosomal recessive primary microcephaly ( MCPH ) , [5 , 8–13] and two genes , KIF11 [14] and DYRK1A [15] , have been linked to autosomal dominant primary microcephaly . Most of the known MCPH genes are expressed predominantly in neuronal tissues during embryonic development and have been implicated in neuronal differentiation [5 , 9 , 14 , 15] . We now demonstrate that autosomal dominant primary microcephaly can be caused by a dominant mutation in ALFY , encoding a master scaffold protein which facilitates removal of aggregated intracellular proteins . Through Drosophila and in-vitro experiments , we show that ALFY controls Wnt signaling by regulating DVL3 aggregation , likely in an autophagy-dependent manner , unveiling novel molecular pathways of normal brain development and primary microcephaly .
A large kindred presented with apparently autosomal dominant isolated primary microcephaly with mild to moderate intellectual disability ( Fig 1A ) . Head circumference of all affected individuals was <3 standard deviations below the mean per age . None of the patients had apparent dysmorphic features or ocular malformations , and thorough physical examination revealed no further abnormalities or failure to thrive . Magnetic resonance imaging ( MRI ) demonstrated microcephaly with no structural defects . Genome-wide linkage analysis followed by fine mapping identified a ~9 Mb haplotype on chromosome 4 , which was shared by and unique to the affected individuals in the kindred ( Fig 1A ) . Maximal LOD score was 3 . 44 at D4S1534 ( θ = 0 ) . Within the 9 Mb locus , whole exome sequencing for individual II:4 ( Fig 1A ) identified no homozygous mutations and only 2 heterozygous variations: in ALFY ( termed also WDFY3 ) and in CXCL11 . While the CXCL11 variation was found in 4 of 200 Israeli Arab healthy controls , none of the controls had the ALFY variation . The ALFY variation segregated within the kindred as expected . Thus , the only variation common and unique to the affected individuals of the kindred was a missense mutation in ALFY: c . 7909C>T g . Chr4:85636503G>A , NM_014991 . 4 , p . R2637W ( Fig 1B ) . The missense mutation , within an extremely conserved residue ( Fig 1C ) of the PH-BEACH domain , results in substitution of a hydrophilic positively charged arginine to aromatic hydrophobic tryptophan . Reconstruction modeling based on the resolved 3D structure of the PH-BEACH domain of neurobeachin , [16] which has high homology with the PH-BEACH domain of ALFY ( 49% identity , 65% similarity over 265 amino acids of the PH-BEACH domain ) , predicted that the arginine at position 2637 normally protrudes into a putative binding pocket . Although phospholipid binding by the PH-BEACH domain of ALFY has never been demonstrated , it is thought that such binding might occur at this site , based on known function of PH domains [17] . Replacement of arginine with tryptophan at this position is predicted to dramatically alter the putative binding domain function ( Fig 1D ) . ALFY is a nuclear protein , which upon accumulation of protein aggregates in the cell shuttles to the cytoplasm , where it serves as a scaffold protein facilitating autophagy-mediated removal of such cytosolic protein aggregates [18–21] . The 1200 amino acid C-terminal of the 3526 amino acid long ALFY is sufficient for these functions , in line with the putative functional domains ( PH-BEACH , WD40 and FYVE ) within this fragment [18–21] . ALFY is extremely conserved throughout evolution , and its conservation is highest at the C-terminal ( Fig 1C ) . Its homology with its Drosophila ortholog Bluecheese ( Bchs ) is relatively high ( 48% identity and 64% similarity in full protein sequence ) , and extremely high surrounding the mutation ( Fig 1C ) . To explore the functional effect of the human ALFY ( hALFY ) mutation , we utilized the Drosophila model system , since ALFY shares high degree of similarity with its fly ortholog bchs [22 , 23] . In line with the dominant heredity of the human phenotype , we generated transgenic flies over-expressing EGFP-tagged wild type and mutant forms of the 1226 amino acids of the C-terminal of hALFY under the control of the upstream activating sequence ( UAS ) Gal4 system , allowing tissue-specific expression . While ubiquitous expression of the wild type hALFY C-terminal construct under the control of actin promoter resulted in no abnormal phenotype , in flies ubiquitously expressing the human mutant allele , the larvae appeared to develop normally and were able to reach pupae stage but did not eclose . All pharates of the mutant hALFY allele were similar to the wild type ones in terms of size and morphology . To study possible effects of the hALFY mutation on Drosophila brain development , we micro-dissected brains of age matched wild type hALFY and mutant hALFY transgenic pharates just prior to eclosing , and visualized their morphology using confocal microscopy . Unlike the normal morphology of brains of pharates expressing wild type hALFY , brains expressing mutant hALFY were 40–60% smaller in volume , denser , very fragile and malformed , and in some cases disintegrated during dissection ( Fig 2C ) . Interestingly , while hALFY wild type brains displayed normal elongated neurons , in flies expressing mutant hALFY , clusters of disorganized cells containing aggregates of EGFP-labeled ALFY were evident ( Fig 2C , white arrows ) . Although the expression of the mutant allele was ubiquitous , no clear effect was observed other than in brain and neuronal tissues . Since the Drosophila eye is of neuronal origin and was previously used to explore phenotypes in mutants of the ALFY ortholog Bluecheese ( Bchs ) , [20 , 22] we used GMR-Gal4 promoter to drive gene expression in the eyes . While expression of the wild type hALFY C-terminal allele in the Drosophila eye had no effect on eye development , exogenous expression of the human mutant allele resulted in a severe rough eye phenotype ( Fig 2B ) : the omatids were disorganized and of variable size and shape with many omatids fused together; the bristles were disorganized with irregular numbers stemming from in between omatids . The phenotype observed was evident immediately after eclosing . To unravel the molecular pathway through which the ALFY mutation causes the disease phenotype , we tested mutant ALFY’s ability to exert its known function in aggregate removal . To that end , we transiently transfected HEK293T cells with the first exon of the huntingtin gene coupled with 103 PolyQ expansion repeats , tagged with an enhanced green fluorescent protein ( EGFP tagged Htt-poly103Q ) [19 , 24] . Cells were co-transfected with plasmids expressing tdTomato-tagged constructs of either wild type or mutant hALFY C-terminus , which has previously been shown to harbor all the essential functional domains of the protein [19 , 20 , 25] . As seen in Fig 3A , the mutation had no effect on PolyQ aggregate encapsulation and removal by hALFY . We therefore set out to elucidate other , possibly novel autophagy-related functions of ALFY , which might unravel the molecular pathway through which the ALFY mutation causes the microcephaly phenotype . Many of the known MCPH genes encode proteins that are associated with formation , stabilization and function of primary cilia [1 , 5 , 9] . Among their known functions , primary cilia mitigate Wnt signaling [26] . The Wnt signaling pathway , in turn , has been shown to be negatively regulated by autophagy-mediated degradation of disheveled [27] . The three human disheveled genes ( DVL1 , DVL2 and DVL3 ) encode hub proteins known to control Wnt signaling [28 , 29]: upon Wnt signaling stimulation , DVL proteins , in their polymer form , recruit the β-catenin destruction complex to the membrane receptors LRP6 and Frizzled , releasing β-catenin , leading to its accumulation in the cytoplasm . β-catenin then shuttles to the cell nucleus , where it serves as a transcription factor activating transcription of Wnt signaling genes [30] . In contrast , ubiquitin-mediated autophagy of DVL proteins abrogates this process , attenuating Wnt signaling [27 , 31] . The specific molecular pathway mediating DVL autophagy is yet unknown . We thus hypothesized that ALFY , a known master scaffold autophagy-related protein , might be the missing link , regulating DVL autophagy , thus controlling Wnt signaling . To test whether hALFY ( and its mutation ) might affect Wnt signaling , we utilized the TOP Flash reporter assay: SH-SY5Y neuroblastoma cells were co-transfected with Renilla luciferase ( normalizer construct ) controlled by HSV TK promoter , and with firefly luciferase under the control of the TCEF/LEF promoter , a known reporter of Wnt signaling activation . Upon addition of Wnt3a conditioned medium to the cell culture , the Wnt pathway is activated , and its activity can be measured by the quantification of firefly luciferase in proportion to the control Renilla luciferase . In addition to the TOP Flash reporter assay constructs , the neuroblastoma cells were co-transfected with different hALFY constructs: wild type hALFY , mutant hALFY , and the negative controls hALFY Δ-FYVE and hALFY-mock , all tagged with either Flag or tdTomato . The constructs were generated using the hALFY wild type construct as template , as follows: hALFY mutant—by inserting a single point mutation recreating the exact p . R2637W substitution observed in our patients; hALFY Δ-FYVE , introducing a stop codon eliminating translation of the last of the five WD40 domains and the FYVE domain; hALFY-mock–inserting a premature stop codon eliminating all functional domains of the ALFY protein ( Fig 3B , detailed in Methods ) . As seen in Fig 3C , expression levels of firefly luciferase were ~40% lower in cells transfected with wild type ALFY in comparison with cells overexpressing hALFY mutant , hALFY Δ-FYVE or hALFY-mock . This ALFY-mediated attenuation of Wnt signaling was abrogated by the addition of wortmannin , a PI-3-kinase inhibitor , known to inhibit autophagy processes [19] . Thus , ALFY attenuates Wnt signaling , probably in an autophagy-dependent manner . Note that cells transfected with mutant hALFY exhibited slightly higher ( albeit not statistically significant ) luciferase levels compared to cells transfected with Δ-FYVE and mock controls , suggesting a possible dominant negative effect of the mutation . We next set out test whether this ALFY-controlled Wnt signaling occurs via the canonical pathway . Activation of the canonical Wnt signaling pathway is known to result in higher levels of cytoplasmic β-catenin [30] . In line with the TOP Flash reporter assay data , β-catenin endogenous protein levels were lower in cells transfected with wild type hALFY as compared to cells expressing mutant hALFY ( Fig 3D ) . Here too , introduction of the autophagy inhibitor wortmannin abrogated the differential effect seen ( Fig 3C ) . These data confirm that ALFY regulates Wnt signaling most likely through an autophagy-mediated process , demonstrating that this occurs specifically through the canonical Wnt signaling pathway . The high levels of β-catenin in cells transfected with mutant hALFY in comparison to wild type hALFY further suggest a dominant negative effect of the mutation ( Fig 3D ) . It has been previously shown that autophagy negatively regulates canonical Wnt signaling by promoting disheveled degradation [27] . Thus , we speculated that the ALFY autophagy-mediated attenuation of the Wnt signaling pathway that we identified might be mediated through disheveled degradation . To test this hypothesis , we transiently co-transfected neuroblastoma ( SH-SY5Y ) cells with constructs harboring human DVL1 , DVL2 or DVL3 with a FLAG epitope tag in the N-terminus and EGFP in the C-terminus ( Fig 4A ) . As expected , over-expression of the different DVL isoforms resulted in dynamic aggregate formation [32] of the respective DVL proteins , as could be seen in green puncta in confocal analysis ( Fig 4B ) . We then repeated the above experiment , co-transfecting the cells also with either wild type or mutant human ALFY . As seen in Fig 4B , both wild type and mutant ALFY co-localized specifically with DVL3 rather than with DVL1 or DVL2 ( minimal co-localization with DVL1 was seen , but was significantly lesser than with DVL3 and might be secondary to the massive overexpression levels ) . As DVL proteins are continuously generated , to test their removal we repeated the above experiment with the addition of cycloheximide 24 hours following transfection , preventing further de-novo translation of DVL proteins . Amounts of DVL proteins were assessed by western blotting using antibodies to the FLAG tag epitope of the DVLs . As seen in Fig 4C , in the presence of wild type hALFY , DVL3 protein levels were lower ( though not significantly ) than in the presence of the Δ-FYVE or mock controls lacking hALFY’s crucial functional domains . In line with likely dominant negative effect of the mutation , this effect was even more apparent and significant when comparing the effect of wild type hALFY to that of mutant hALFY , with significantly higher DVL3 levels in the presence of mutant hALFY . Thus , hALFY acts in DVL3 aggregate removal . In contrast , there was no visible effect of hALFY on aggregate removal of DVL1 and DVL2 ( Fig 4C ) . We repeated the experiment exploring the effect of hALFY on endogenous DVL protein levels . As shown in Fig 4D , the levels of endogenous DVL3 were significantly lower in cells over-expressing wild type hALFY in contrast with cells over-expressing either mutant , Δ-FYVE or mock negative control hALFY constructs . Moreover , this effect of wild type hALFY was abolished in the presence of Wortmannin . In line with the previous results , hALFY over-expression did not significantly alter levels of DVL1 or DVL2 . Thus , our data taken together with the known role of ALFY in autophagy , demonstrate specific targeting and autophagy-mediated removal of DVL3 by hALFY .
Through linkage analysis and whole exome sequencing , we demonstrated that dominantly inherited human congenital microcephaly with mild to moderate intellectual disability is caused by a mutation in hALFY ( termed also WDFY3 ) . hALFY , encoding an autophagy master scaffold protein , is ubiquitously expressed and its expression levels in mouse tissues ranges from low ( heart , kidney , lungs , skeletal muscle and pancreas ) to extremely high levels ( brain , spleen ) [19] . In mice , expression of the ALFY homolog wdfy3 is maintained in the neocortex of the developing embryo [33] , specifically in dividing neuronal progenitor cells ( RGCs ) , and is required for a specific subset of progenitor divisions [34]; its Drosophila homolog , Bchs , is expressed in the outer cortical regions of the nervous system and in the ommatidial cluster of the developing eye after neuronal differentiation [22] . It is noteworthy that this expression profile of ALFY is similar to that of other microcephaly genes [5 , 8 , 9 , 14 , 15 , 35] . Null mutants of ALFY homologs have been generated and studied both in flies and in mice: Bchs knockout files develop normally and display normal motor , feeding , and grooming behaviors [22] . However , the life span of those mutants is 40–50% shorter than that of wild type flies . Furthermore , the mutant flies display a progressive neurodegenerative phenotype and reduction of approximately 40% in brain volume [22] . Mice with null mutations in wdfy3 have also been recently generated [34] . While heterozygous mutant mice show no abnormal phenotype , homozygous mutants are embryonic lethal , with a clear brain phenotype: the cerebral cortex is visibly thinner and tangentially longer compared to wild type controls . The proliferative regions of the cortical ventricular and subventricular zones as well as the intermediate zone are thinner in mutants , while the cortical plate and marginal zone are not affected in thickness [34] . The tangential expansion but lateral thinning of the neocortical neuroepithelial seen in those mice , suggest an imbalance in the mode of cortical progenitor cell divisions , favoring proliferative at the expense of differentiative divisions [34] . Although the embryonic phenotype of the mouse mutants has been well studied , the molecular mechanisms underlying normal and abnormal ALFY-related neurodevelopment are yet unknown . To address this issue , we extended our studies , exploring the effects of human wild type and mutant ALFY in the drosophila model system and in vitro experiments . First , we verified the dominant effect of the human ALFY mutation through Drosophila experiments: while transgenic flies expressing the human wild type ALFY had no abnormal phenotype , transgenic flies expressing human mutant ALFY driven by a ubiquitous actin promoter were lethal and had an evident neuro-specific phenotype with smaller brain size , similar to that seen in bchs null mutant flies [22] . As no abnormalities were evident in the transgenic flies aside from the brain phenotype , it is likely that the late embryonic lethality was secondary to the neurological defect . It is noteworthy that the brain-specific microcephaly phenotype seen in flies expressing the human mutant ALFY strikingly recapitulates the dominant phenotype seen in our patients , in line with the extreme conservation of ALFY throughout evolution . The dominant effect of the hALFY mutation was further demonstrated through studies of the Drosophila eye , a well-studied model system of neurodevelopment [36 , 37] Eye-specific GMR-promoter driven expression of mutant hALFY generated a rough eye phenotype , as compared with no abnormal phenotype in flies expressing wild type hALFY . The Drosophila rough eye phenotype , disrupting the regular ommatidia arrangement , has been demonstrated in mutations in various genes controlling neurodevelopment , including , for instance , lgl1 [38] , whose mouse null mutants demonstrate failure of neural progenitor cells to exit the cell cycle and differentiate , and instead , continue to proliferate and die by apoptosis [39] . Thus , our data suggest a possible role of ALFY in controlling proper transition of neuronal progenitor cells from proliferation to differentiation . This point will be dealt with later in the discussion . Since ALFY is an autophagy-mediating scaffold protein , capable of removing Htt-Poly103Q aggregates [18 , 19 , 21] , we tested whether this function is affected by the mutation . As seen in Fig 3A , mutant hALFY was as effective as wild type hALFY in Poly103Q aggregate removal . This is in line with the disease phenotype being one of primary congenital microcephaly rather than a progressive neurodegenerative disease . However , the question of molecular mechanisms through which the ALFY mutation exerts its effect remained unclear . ALFY encodes a large nuclear protein , which shuttles to the cytoplasm upon formation of protein aggregates: upon generation of aggregates in the cytoplasm , they are ubiquinated and bound by P62 . P62 , a nuclear protein , is thought to bind ALFY through its PH-BEACH domain and shuttles the protein complex to the ubiquinated aggregates . ALFY then binds through its WD40 domains the ATG complex of autophagy proteins , and through its FYVE domain LC3-bound membranes , thus putting together autophagosomes . This is followed by fusion of autophagosomes with lysosomes to generate autolysosomes , where aggregated proteins are disintegrated [18 , 40] . Thus , ALFY serves as a scaffold protein recruiting the autophagy machinery to generate the autophagosome , enabling degradation of cytosolic protein aggregates . All genes associated to date with human MCPH are associated with processes of cell cycle control , such as DNA replication , primary cilia formation and stabilization , and centriole duplication [5 , 8 , 9 , 15 , 35 , 41–54] . Such processes govern properly timed transition from symmetric to asymmetric cell division in the developing brain . Taking into consideration ALFY’s role in autophagy and the known association of MCPH with defects in cell cycle regulation , we set out to elucidate a novel , yet unraveled , function of ALFY that might integrate those two processes . The Wnt signaling pathway is one of the major molecular pathways controlling cell division and the transition between symmetrical and non-symmetrical cell division [26 , 55–59] . In fact , proliferating cells are known to demonstrate high Wnt signaling activity [56 , 58 , 60] . It has previously been shown that autophagy negatively regulates the Wnt signaling pathway through degradation of disheveled ( DVL ) proteins , hub proteins controlling Wnt signaling [27] . We thus hypothesized that ALFY might be the missing link connecting the Wnt signaling pathway and DVL autophagy . First , using the TOP Flash assay we showed that ALFY normally attenuates Wnt signaling , an effect abrogated by either the ALFY mutation seen in the patients , or null mutations introduced such that ablate its crucial domains . Furthermore , through studies of β-catenin levels , we showed that the effect of ALFY is specifically in controlling the canonical Wnt signaling pathway . Repeating the experiments with and without a known autophagy inhibitor , we showed that this ALFY-mediated regulation of the canonical Wnt signaling is most likely achieved through autophagy . Finally , we showed that ALFY co-localizes with DVL3 and facilitates the removal of DVL3 aggregates from the cytoplasm . Both co-localization and aggregate removal experiments demonstrated that the effect was very specific: ALFY did not co-localize with or act in removal of aggregates of DVL1 and DVL2 . Thus , we show a novel molecular mechanism of regulation of the canonical Wnt signaling pathway through ALFY-mediated regulation of DVL3 aggregate removal . Our data heavily suggest that the hALFY mutation acts through a dominant negative effect rather than haploinsufficiency: over-expression of the mutant hALFY ( and not wild type hALFY ) was sufficient to generate a Drosophila phenotype regardless of the endogenous Bchs protein , practically ruling out haploinsufficiency as a mechanism through which the mutation acts . In the cell culture experiments , Wnt signaling activity , as exhibited by both TOP Flash luciferase activity and beta-catenin levels , was consistently slightly higher in cells expressing mutant hALFY than in those expressing the Δ-FYVE or hALFY-mock control lacking hALFY’s crucial functional domains . This , again , suggests a likely dominant negative rather than dosage effect . The same is true also for the effects on DVL3 aggregate removal further supporting our claim . Regarding the mechanism through which the ALFY mutation causes primary microcephaly , taking all our data together with that of the previously described Wdfy3 null mutants , we suggest the following model ( Fig 5 ) : during normal brain development the apical progenitor ( APs ) cell layer undergoes a series of symmetrical proliferative divisions , triggered by Wnt signaling . This cell proliferation constitutes the initial clonal expansion of cortical brain cells . At this point , ALFY which is expressed in this layer of proliferating RGCs [30] , attenuates Wnt signaling through DVL3 aggregate removal , leading to inactivation of the Wnt signaling resulting in transition to asymmetrical differentiative cell division , generating the basal progenitor ( BPs ) cell layer . Expansion of this second cell layer is essential in primate brain development and cortex expansion [3] . This BPs layer will eventually differentiate to form functional neurons . In the mutant ALFY this attenuation is abrogated . Therefore , Wnt signaling proceeds unharmed , "locking" the AP cells in further consecutive rounds of symmetrical cell divisions , at the expense of formation of BPs , eventually resulting in fewer neurons and thus microcephaly . This is in line with the 50% reduction in BP cell number and increase in proliferative symmetrical division at the expense of asymmetrical divisions , resulting in tangential expansion and radial thinning of the cortex seen in Wdfy3 null mutant mice[34] . Since the AP cell layer is confined in space , this expansion of APs is limited and will not compensate for the lack of crucial BP cells . It is the BP cell layer that demonstrates the most significant difference between primates and rodents and is believed to be responsible for determining the final size of the primate brain [3 , 61] . Our data together with the known functional roles of ALFY , suggest a novel , previously undescribed role for autophagy in neurodevelopmental processes and proper brain development , and demonstrate that defective autophagy-related processes can cause MCPH .
The study was approved by the Soroka Medical Center institutional review board ( approval number 5071G ) and the Israel Ministry of Health National Helsinki committee ( approval number 920100319 ) . Written informed consent , per the study approved protocol , was obtained from all individuals studied or their legal guardians . The study protocol was approved by Soroka Medical Center institutional review board . Informed consent was provided by all participants or their legal guardians . Genome-wide linkage analysis of six family members ( five affected , one unaffected ) was done ( ABI PRISM Linkage Mapping Set MD10 , Applied Biosystems ) and analyzed using GeneScan software , as previously described . [62] Fine mapping analyzing nine affected and seven unaffected individuals was carried out as previously described [62] using polymorphic markers indicated in the family pedigree ( Fig 1A ) . PCR products were separated on 6% polyacrylamide gel and visualized by silver staining , and haplotypes were manually constructed and analyzed . Maximal LOD score was calculated using SUPERLINK . [62] Whole exome sequencing was performed as previously described [63] for individual II:4 ( Fig 1A ) and data was analyzed using QIAGEN’s Ingenuity Variant Analysis software ( http://www . qiagen . com/ingenuity ) from QIAGEN Redwood City . Using their filtering cascade , we excluded variants observed with an allele frequency ≥1 . 0% of the genomes in the 1000 genomes project , National Heart , Lung , and Blood Institute ( NHLBI ) , Exome Sequencing Project ( ESP ) or the Allele Frequency Community . In addition , we excluded variants that appeared in our in-house whole exome sequencing database of 100 Bedouin control samples . Furthermore , we kept variants that are predicted to have a deleterious effect upon protein coding sequences ( eg , frameshift , in-frame indel , stop codon change , missense or predicted to disrupt splicing by MaxEntScan ) and variants that are experimentally observed to be associated with a phenotype: pathogenic , possibly pathogenic or disease-associated according to the Human Gene Mutation Database ( HGMD ) . Screening for the CXCL11 3-nucleotide deletion mutation was done using PCR amplification followed by separation on a 6% polyacrylamide gel and visualization by silver staining ( details available upon request ) . For the ALFY mutation , screening was done using BtsCI restriction analysis of PCR amplicons ( Forward primer 5’-CCCTCCCATATCTTCCCATAA-3’; Reverse 5'-GCCGCAGATTAACTTCTTGA-3’ ) , demonstrating a 240 bp wild type allele versus 100 bp and 140 bp fragments for the mutant allele . All clones were generated with a modified pUAST vector which contains an EGFP fragment and an attb1 site which enables site-directed insertion of the constructs . Plasmids containing the C-Terminal part of hALFY , which contains all functional protein domains ( PH-BEACH , WD40 and FYVE domain ) , were a kind gift from Prof . Anne Simonsen and were used as template for PCR amplification . Primers were designed to match ALFY gene sequence , as well as to contain a homologous sequence for the pUAST plasmid at the 5' end and an EGFP homology sequence at the 3' end ( Forward primer 5'-CTGCCAAGAAGTAATTATTGAATACAAGAAGAGAACTCTGAATAGGGAATTGGGAATTCCAAAATGTTAACAGGATCAAGAAGGAATC-3' and reverse primer 5'-GCTCGACCAGGATGGGCACCACCCCGGTGAACAGCTCCTCGCCCTTGCTCACCATGAATTCTCTACAACAATTTCGAGGCCCATCTTC-3' ) . The 3756 bp PCR amplicon was inserted into the pUAST vector using homologous recombination in yeast . The PCR product was cloned in frame with EGFP sequence at the 3' end of hALFY to generate ALFY-EGFP fusion protein . Transgenic flies were generated by Genetic Services Inc . ( Biotech Center , Boston , Massachusetts , USA ) in a site specific manner . To reduce variability of expression due to position effects , the final wild type and mutant constructs were inserted at the same attp2 site on the 3rd chromosome , which is known for its strong and stable expression levels . [64] One line for the wild type and two lines for the mutant construct were generated . The different lines were balanced to form a balanced lethal stock . Constructs containing the C-terminal part of human ALFY gene as well as Htt-Poly103Q were a kind gift from Prof . Simonsen . Tagged versions of hALFY gene ( N-terminal FLAG tag , N-terminal tdTomato and a C-terminal EGFP tag ) , containing all functional domains , spanning 2285–3526 amino acids of ALFY were obtained and used as a template for generating the different constructs used in our experiments . The same technique was employed with different primers to generate the different constructs . We performed PCR amplification using modified back to back primers containing the desired mutations , followed with DpnI treatment to remove template plasmid DNA . Finally , the mutated constructs were validated with Sanger sequencing and restriction analysis . We performed this mutagenesis on all available constructs ( FLAG , tdTomato and EGFP ) . To generate the hALFY mutant ( R2637W ) constructs we inserted a single point mutation , replicating the mutated Arg to Trp substitution seen in our patients . To that end we used modified back to back primers containing the mutation ( WDFY3-pMut-F-5'-CTCTGGAGATGGAtGGAATTACCTC-3' , WDFY3-pMut-R-5'-GAGGTAATTCCaTCCATCTCCAGAG-3' ) . The ALFY Δ-FYVE constructs were generated by inserting a premature stop codon prior to the last WD40 domain with mutated primers ( ALFY Cter-Δ-FYVE-F-5'-GGAGCCAGCAGATCATCTGaTGCTGCATGTCGGAGATG-3' , ALFY Cter-Δ-FYVE-R-5'-CATCTCCGACATGCAGCAtCAGATGATCTGCTGGCTCC-3' ) , resulting in truncated protein lacking the crucial WD40 and FYVE domains . Finally , the ALFY-mock constructs were generated by inserting a premature stop mutation prior to the PH-BEACH domain using mutated primers ( ALFY Cter-mock-F-5'- GAAGAGCCGTAAGTTAacACAGTAAAGAG-3' , ALFY Cter-mock-R-5'- CTCTTTACTGTgtTAACTTACGGCTCTTC-3' ) . The inserted mutation results in a 133 amino acid truncated protein ( of the 1226 amino acids full length ALFY ) lacking all known functional domains . The different DVL genes were PCR amplified from brain cDNA library to obtain the full length constructs ( DVL1-F–5'-GCGGAGACCAAGATTATCTACC-3' , Acc65I-DVL1-R–5'-GGTACCtCATGATGTCCACGAAGAACTCGCAG-3' , DVL2-F-5'-GCGGGTAGCAGCACTGGGG-3' , Acc65I-DVL2-R- GGTACCtCATAACATCCACAAAGAACTC-3' , DVL3-F-5'-GGCGAGACCAAGATCATCTACCAC-3' , Acc65I-DVL3-R-5'-GGTACCtCATCACATCCACAAAGAACTC-3' ) followed by a second round of PCR using modified primers to insert a FLAG epitope at the N terminal end of the protein as well as Acc65I restriction sites ( Acc65I-Flag-DVL1-F–5'-GTACCATGgattacaaggatgacgatgacaagGCGGAGACCAAGATTATCTACC-3' , Acc65I-Flag-DVL2-F-5'-GGTACCATGgattacaaggatgacgatgacaagGCGGGTAGCAGCACTGGGG-3' , Acc65I-DVL3-Flag-F-5'-GGTACCATGgattacaaggatgacgatgacaagGGCGAGACCAAGATCATCTACCAC-3' ) . The modified PCR fragments were cloned into a pEGFP-N2 vector in frame with a C terminal EGFP protein using Acc65I restriction enzyme . Further information regarding the different cloning will be delivered upon request . Wnt3a conditioned medium was obtained by collecting growth medium from mouse fibroblast secreting the active form of Wnt3a glycoprotein ( ATCC CRL-2647 ) according to the manufacturer protocol . In brief: cells were split and grown in T-75 flasks in normal DMEM medium ( without G418 ) for 4 days and growth medium was collected and filter sterilized to achieve the first batch of medium . The medium was replaced and cells were grown for additional 2–3 days and again medium was collected and filter sterilized to achieve the second batch . The two batches were than mixed to generate the Wnt3a conditioned medium . To assess the activation of the Wnt signaling we utilized the standard , largely accepted TOP Flash reporter assay . Twelve hours prior transfection SH-SY5Y neuroblastoma cells were plated in 24-well tissue plates to approximately 60–70% confluence 12 hours prior to transfection . Cells were then transfected with lipofectamine2000 ( Invitrogen ) according to manufacturer protocol with 400 ng of TOP flash plasmids ( Millipore ) , 400 ng the different FLAG-ALFY constructs and 50 ng of internal normalizer pGL4 . 74 [hRluc/TK] vector ( Promega ) . The experiments were repeated with the negative control FOP Flash ( Millipore ) to verify the activity of the TOP Flash reporter system and indeed FOP Flash luciferase expression levels were zero . Autophagy inhibition was achieved by adding wortmannin ( 0 . 2μM final concentration , ab120148 –abcam ) for the entire duration of the experiment . Following transfection , cells were incubated with the plasmid mixture over night and the following day the growth medium was replaced with Wnt3a conditioned medium ( with or without wortmannin ) for a period of 24 hours . Finally , after 24 hours of induction , the cells were harvested , lysed and the luciferase reporter gene assay was conducted ( Dual-Luciferase Reporter Assay System , Promega , E1910 ) . Each transfection was repeated 3 times ( 3xALFY wild type , 3xALFY Mutant etc . ) and each transfected well was measured 3 times for luciferase activity . Average values per transfected well were calculated to obtain a precise measurement of Wnt signaling activity per single well which was later used to calculate the average value for each experiment . Luciferase activity was measured using TECAN infinite M200 instrument . Protein lysates were heated for 10 min in 95°c and loaded onto 8% polyacrylamide gel . Following electrophoresis at 150V for 1 . 5 hours , proteins were transferred to nitrocellulose membranes for 1 hour at 300 mA . The nitrocellulose membranes were blocked by incubation in TTBS ( 0 . 02M Tris , pH7 . 5 , 0 . 15M NaCl , 0 . 9mM Tween 20 –Bio Lab ) containing 3% BSA for 1 hour at room temperature . The blocked membranes were incubated for 1 hour at room temperature with primary antibodies , followed by 3 TTBS washes 10 min each . The membranes were then incubated with the appropriate secondary antibodies for 1 hour at room temperature , washed 3 times in TTBS and visualized using ChemiDoc MP imaging system ( Bio Rad ) . Antibody binding was visualized using an enhanced chemiluminescence detection kit ( SuperSignal West Pico chemiluminescent substrate–Thermo scientific ) . All experiments were repeated several times and were all normalized to wild type hALFY . Transgenic flies for both wild type and mutant hALFY were generated ( Genetic Services Inc . ) and F1 progeny were analyzed for eye phenotype in terms of pigmentation , size , shape and surface texture as well as for neuronal brain phenotype . For each of the experiments , one line for the wild type allele and two different lines for the mutant alleles were generated , with similar results . All fly lines were cultured and maintained on standard cornmeal/agar medium at 25°c . Actin- and GMR-Gal4 driver strains were obtained from Bloomington stock center and used to drive ubiquitous and eye expression of the ALFY constructs , respectively [65] . Representative digital eye images were taken with a Leica 205C dissection microscope and Leica DSC290 HD camera system . In addition , samples were examined with a JEOL JSM 7400F scanning electron microscope ( SEM ) . Brains were analyzed using confocal microscope and Z stack images of the brains were taken using bright field and 488nm filter . Flies were anesthetized with carbon dioxide gas , rinsed once in 70% ethanol , twice in S2 media ( Schneider’s Insect Medium , Sigma ) , submerged in cold S2 media on a dish , and dissected with Dumont #5 stainless steel forceps . Once the brains were exposed , membranes , trachea , and other tissues were removed , taking care not to touch the brain as was demonstrated in a video of the dissection procedure http://www . janelia . org/team-project/fly-light#5064 . [66] The brains were transferred to 4% paraformaldehyde in PBS medium for 20 min followed by 2 washes in PBS and were mounted onto slides containing Vectashield mounting medium without DAPI . Drosophila brain volume was evaluated using the EGFP Z stack images obtained before . Using Icy software , [67] we selected a representative image from each Z stack acquired , which contained the largest part of the brain in the stack . The brain area of that stack was then calculated using the EGFP pixels to determine brain borders . Once brain size was evaluated , the following formula was used to evaluate brain volume: Volume ( mm3 ) = ( AreaofBrain ( mm2 ) ) * ( NumberofZstacks ) * ( Zstackinterval ( mm ) ) All stacks were measured at 5μm ( 0 . 005mm ) intervals , analyzing brains of 18 wild type hALFY and 31 mutant hALFY transgenic flies . The values obtained are a close approximation of the actual brain volumes; the process was performed in similar manner for both wild type and mutants . | One of the major events in human evolution is the significant increase in brain volume in the transition from primates to humans . The molecular pathways determining the larger size of the human brain are not fully understood . Hereditary primary microcephaly , a neurodevelopmental disorder in which infants are born with small head circumference and reduced brain volume with intellectual disability , offers insights to the embryonic molecular pathways determining human brain size . Previous studies have shown that human microcephaly can be caused by mutations in genes affecting cell division processes , such as cell cycle regulation , DNA replication , primary cilia formation and centriole and centrosome duplication . We now show a novel molecular pathway determining human brain size: human primary microcephaly can be caused by a mutation in ALFY , a gene that encodes an autophagy scaffold protein . In fact , transgenic flies over expressing the mutant form of human ALFY recapitulate the human disease phenotype of microcephaly . We show the molecular pathway through which ALFY regulates cell division and differentiation: we demonstrate that ALFY normally controls removal of aggregate of DVL3 , and through this regulates Wnt signaling , a major molecular pathway in embryogenesis . Thus , Wnt signaling , controlled by ALFY-mediated aggregate removal of DVL3 , determines human brain size and human microcephaly . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"death",
"invertebrates",
"autophagic",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"enzymes",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"neuronal",
"differentiation",
"enzymology",
"animals",
"cell",
"differe... | 2016 | ALFY-Controlled DVL3 Autophagy Regulates Wnt Signaling, Determining Human Brain Size |
The IκB kinase ( IKK ) complex is a key regulator of signal transduction pathways leading to the induction of NF-κB-dependent gene expression and production of pro-inflammatory cytokines . It therefore represents a major target for the development of anti-inflammatory therapeutic drugs and may be targeted by pathogens seeking to diminish the host response to infection . Previously , the vaccinia virus ( VACV ) strain Western Reserve B14 protein was characterised as an intracellular virulence factor that alters the inflammatory response to infection by an unknown mechanism . Here we demonstrate that ectopic expression of B14 inhibited NF-κB activation in response to TNFα , IL-1β , poly ( I:C ) , and PMA . In cells infected with VACV lacking gene B14R ( vΔB14 ) there was a higher level of phosphorylated IκBα but a similar level of IκBα compared to cells infected with control viruses expressing B14 , suggesting B14 affects IKK activity . Direct evidence for this was obtained by showing that B14 co-purified and co-precipitated with the endogenous IKK complex from human and mouse cells and inhibited IKK complex enzymatic activity . Notably , the interaction between B14 and the IKK complex required IKKβ but not IKKα , suggesting the interaction occurs via IKKβ . B14 inhibited NF-κB activation induced by overexpression of IKKα , IKKβ , and a constitutively active mutant of IKKα , S176/180E , but did not inhibit a comparable mutant of IKKβ , S177/181E . This suggested that phosphorylation of these serine residues in the activation loop of IKKβ is targeted by B14 , and this was confirmed using Ab specific for phospho-IKKβ .
Nuclear factor-κB ( NF-κB ) is critical for the innate and adaptive immune responses to infection . Various stimuli , such as the pro-inflammatory cytokines interleukin ( IL ) -1 and tumour necrosis factor ( TNF ) , activate signaling pathways leading to NF-κB-dependent gene expression [1 , 2] . Several of these signaling pathways converge on the IKK complex [3–5] , and this complex is therefore a prime target for anti-inflammatory drugs . It is also a logical target for pathogens aiming to minimize the host response to infection . The IKK complex , or signalosome , comprises a heterodimer of IKKα and IKKβ in association with NF-κB essential modifier ( NEMO also called IKKγ ) [6 , 7] and is critical for NF-κB activation induced by pro-inflammatory cytokines [8–10] . The IKK complex is activated by upstream kinases , such as transforming growth factor-β ( TGFβ ) -activated kinase-1 ( TAK1 ) , which phosphorylates IKKβ at Ser177 and Ser181 located in the activation loop [2 , 4 , 5] . Once activated , IKKβ phosphorylates the inhibitor of NF-κB ( IκBα ) [11] to initiate IκBα degradation . Phosphorylated IκBα ( phospho-IκBα ) is recognized by an F-box/WD protein , β-transducin repeats-containing proteins ( β-TrCP ) , which functions as a receptor subunit of the SCF family ubiquitin ligase complex , and binds to the phosphorylated E3 recognition sequence on IκBα [12–15] . This poly-ubiquitinated IκBα remains associated with NF-κB but is degraded selectively via the 26S proteasome [16] . After IκBα degradation , NF-κB is translocated into the nucleus to induce transcription of responsive genes [17] . Poxviruses have developed strategies to modulate important cellular signaling pathways to evade host responses [18–20] . These viruses target many of the primary mediators of immune system including IL-1 , IL-18 , interferons ( IFNs ) , TNF , complement , and chemokines [20–23] . Many of the genes encoding vaccinia virus ( VACV ) immunomodulators show amino acid similarity to host proteins that function in the immune system . However , others lack such similarity; for instance , the intracellular virulence factor N1 [24] , anti-apoptotic protein F1 [25 , 26] , and secreted chemokine binding protein [27] . VACV and other poxviruses interrupt the activity of NF-κB in several ways [21 , 28] . One strategy is to secrete proteins from the infected cell to bind cytokines , chemokines , or IFNs and prevent these reaching their receptors on cells . Another strategy is to express intracellular factors to regulate signaling pathways leading to NF-κB activation . Among these intracellular inhibitors , VACV proteins A52 and A46 antagonize IL-1R and toll like receptor ( TLR ) signaling [29–31] and N1 is a virulence factor [24] that is reported to interfere with NF-κB and IRF3 activity [32] . In addition , the crystal structure of N1 reveals it is a Bcl-2-like protein and N1 was shown to protect cells from apoptosis [33] . VACV protein K1 also inhibits NF-κB activation during infection [34] . Lastly , protein M2 downregulates ERK-mediated NF-κB induction in virus-infected cells [35] . B14R is one of the few VACV genes that are located in the terminal region of the virus genome and yet is conserved in many orthopoxviruses [36] , suggesting an important function . However , B14 lacks sequence identity with proteins from outside poxviruses . An initial characterization of B14 showed it is an intracellular virulence factor that is expressed early during infection and affects the inflammatory response to infection in a murine model by an unknown mechanism [37] . In this study , the mechanism of action of B14 has been investigated . Data presented show that B14 associates with and inhibits the activity of the IKK complex and thereby inhibits NF-κB activation from multiple signaling pathways . This mechanism of action is consistent with the in vivo phenotype of a virus lacking the B14R gene [37] .
Bioinformatic analyses indicated that B14 is a member of a family of poxvirus proteins that include B14 , K7 , C6 , and A52 [38] . Subsequently , the A46 protein was shown to be related to A52 and was added to this family [29] . Given that proteins A46 and A52 are intracellular inhibitors of TLR signaling pathways [29–31] , the presence of B14 in the same family suggested that B14 might also act to regulate signaling pathways leading to NF-κB activation . To investigate the effect of B14 on NF-κB activation , a plasmid containing a luciferase reporter gene linked to a NF-κB-dependent promoter was transfected into HeLa cells and these cells were stimulated with IL-1β ( Figure 1A ) , TNFα ( Figure 1A ) , or PMA ( Figure 1B ) . Luciferase activity was increased greatly by addition of each stimulant but the level reached was reduced in a dose-dependent manner in the presence of B14 . Similar findings were observed using HEK 293 cells ( unpublished data ) . Moreover , B14 decreased poly ( I:C ) -induced NF-κB dramatically ( p-value = 0 . 0006; 95% decrease ) ( Figure 1E ) . In contrast , B14 did not reduce luciferase activity from ISRE ( Figure 1C ) and AP-1 ( Figure 1D ) reporter genes induced by IFNα and PMA , respectively . Notably , B14 increased PMA-induced AP-1 activity slightly ( p = 0 . 02; 1 . 5-fold increase; Figure 1D ) . We also observed a small but significant reduction in poly ( I:C ) -induced ISRE activity in the presence of B14 ( p-value = 0 . 01; 29% decrease; Figure 1E ) . However , it is uncertain if these relatively small changes seen with these reporter assays are relevant biologically . As a control we also expressed A20 , a de-ubiquitinating enzyme that downregulates NF-κB and IRF3 [39–41] and observed strong inhibition of both pathways ( Figure 1E ) . Therefore , B14 is a specific downregulator of NF-κB but did not inhibit AP-1 or IRF responsive gene expression . The fact that B14 inhibits multiple pathways leading to NF-κB activation suggests that B14 might act at a position at or downstream of the site at which these pathways converge , namely the IKK complex . To examine how B14 downregulates NF-κB activation , we searched for interactions between B14 and potential ligands using a luminescence-based mammalian interactome mapping ( LUMIER ) assay [42] . Components of the IKK complex were included in the assay because several pathways leading to NF-κB activation converge on this complex . HA-tagged B14 and A20 were transfected into cells together with different proteins fused with luciferase , and cell extracts were immunoprecipitated with anti-HA mAb . The immunoprecipitates were then tested for luciferase activity ( Figure 2A ) . B14 interacted with IKKα , IKKβ , and NEMO but not with TBK1 , IKKɛ , p65 , and A20 . As expected , A20 showed no interaction with any proteins screened in the assay except itself [43] . These observations indicated that B14 interacts with the IKK complex . Previously , VACV protein N1 was reported to interact with and inhibit the IKK complex [32] and therefore FLAG-tagged N1 was also included in this assay . Surprisingly , no interaction between N1 and IKKα , IKKβ , or NEMO was observed ( Figure 2B ) , although FLAG-B14 and IKKα each co-precipitated with the IKKs . FLAG-GFP was included as a negative control and did not bind to the IKKs . Collectively , these data show that B14 , but not N1 , associates with the IKK signalosome . To investigate these protein interactions further , we fractionated VACV-infected cell extracts by size exclusion chromatography ( SEC ) and blotted the fractions with antibody to B14 . B14 eluted in two peaks corresponding to approximately 160 kDa and 700–900 kDa , despite having a monomeric size of 17 kDa . Given that the IKK complex also has a mass of between 700–900 kDa [1] , we immunoblotted the column fractions with antibodies to IKK components and found that the IKK complex co-purified with the first B14 peak ( Figure 2C ) . The column fractions were also blotted with Ab to N1 and this showed that N1 eluted with a mass of approximately 60–70 kDa ( Figure 2C ) , quite distinct from the IKK complex and also distinct from the expected position of the 28-kDa N1 homo-dimer [24] . Therefore , B14 , but not N1 , co-purified with the IKK complex in the VACV-infected cell lysates . The possible interaction between B14 and the IKK complex was investigated further by immunoprecipitation . HeLa cells were infected with a VACV strain expressing an HA-tagged version of B14 ( vB14-HA ) or VACV lacking gene B14R ( vΔB14 ) [37] , and cytoplasmic extracts were prepared . B14-HA was immunoprecipitated with anti-HA mAb and immunoprecipitates were analysed by immunoblotting with Abs against IKKα and IKKβ , NEMO , or HA ( Figure 2D ) . The anti-HA mAb precipitated B14-HA together with IKKα , IKKβ , and NEMO from the vB14-HA infected cell lysates ( Figure 2D , lane 4 ) . The interaction between B14 and the IKK complex was also seen in the reciprocal immunoprecipitation using antibody to NEMO ( Figure 2D , lanes 5 and 6 ) and anti-IKKα/β ( unpublished data ) . In contrast , B14 and the IKK complex were not co-immunoprecipitated with a control mAb against glycogen synthase kinase ( GSK ) -3β ( Figure 2D , lanes 7 and 8 ) . In summary , B14 and the IKK complex co-purified and co-precipitated when each component was expressed at natural levels . To identify which of the IKK components interacts with B14 , mouse embryo fibroblasts ( MEFs ) lacking IKKα or IKKβ were analysed as above for HeLa cells ( Figure 3 ) . In vB14-HA-infected wild type MEFs B14 co-precipitated with the IKK complex ( Figure 3A , lane 4 ) . In the absence of IKKα or IKKβ , the anti-NEMO mAb still precipitated a complex of IKKβ-NEMO and IKKα-NEMO , respectively ( lanes 5 and 6 ) . However , B14 was co-precipitated from IKKα but not IKKβ null MEFs , indicating that B14 was incorporated in the IKKβ-NEMO complex ( lane 5 ) and that IKKβ was needed for B14 to be part of the IKK complex . As a control , an anti-FLAG mAb did not immunoprecipitate any proteins ( Figure 3A , lanes 7–9 ) . The interaction between B14 and IKKβ was also investigated by SEC of extracts from wild type , IKKα , or IKKβ null MEFs ( Figure 3B ) . B14 only co-purified with the IKK complex of 700–900 kDa when IKKβ was present , but was present in the second peak of 160 kDa in all samples . So , IKKβ is necessary for B14 to co-purify or co-precipitate with the IKK complex . Upon stimulation , the IKK complex phosphorylates IκBα and this is then removed quickly via the proteasome system . Therefore , we examined the level of IκBα in cells stimulated with TNFα in the presence and absence of B14 ( Figure 4 ) . The amount of IκBα was reduced dramatically at 20 min after TNF treatment but had recovered to the original level by 50 min . However , in the presence of B14 the level of IκBα was greater at 20 min post-stimulation with TNF . Thereafter , the level of IκBα recovered to that before stimulation . Equal loading of samples was demonstrated by blotting for α-tubulin . Therefore , B14 increased IκBα stability after TNF stimulation , implying a negative effect on IKK activity . The above experiment was performed in cells expressing B14 after transfection . To investigate whether the endogenous levels of B14 could affect IKK activity during virus infection , the phosphorylation status of IκBα was investigated in cells infected with VACV strains that do or do not express B14 . Cells were infected with wild type ( vB14 ) , deletion mutant ( vΔB14 ) , or revertant ( vB14-rev ) viruses [37] at 2 p . f . u . /cell and at 2 and 4 h p . i . , cytoplasmic fractions were prepared and analysed by immunoblotting ( Figure 5 ) . The level of IκBα was indistinguishable in infected or uninfected cells , and similarly there was no difference following infection with viruses that did or did not express B14 . However , following infection by all viruses , the level of phospho-IκBα was increased , but the increase was noticeably higher in cells infected with vΔB14 , compared to vB14 and vB14-rev . To show that each virus caused equivalent infection , cell extracts were immunoblotted with antibody to the VACV intracellular protein N1 [24] , and N1 was detected at similar levels in each sample at 2 h p . i . and at slightly higher levels in each sample later during infection ( 4 h ) ( Figure 5 , bottom panel , lanes 2–4 and 6–8 ) . In contrast , B14 was present in vB14- and vB14-rev-infected cells only . As expected , each VACV protein was absent in mock-infected cells . This suggested that B14 reduces IKK activity during VACV infection . The effect of B14 on IKK activity in the absence of other VACV-encoded NF-κB inhibitors was investigated next using an in vitro kinase assay . Plasmids expressing TRAF2 or HA-tagged IKKβ were co-transfected with or without pCI-B14 . TRAF2 acts as an intracellular stimulator of IKK activity . Extracts from transfected cells were immunoprecipitated with anti-HA mAb . The activity of the immunoprecipitated IKK complex was studied using a synthetic IκBα peptide substrate and 32P-γ-ATP followed by SDS-PAGE and autoradiography . Notably , the level of the phospho-IκBα peptide was reduced in the presence of B14 , indicating B14 inhibited IKK activity ( Figure 6 ) . Coomassie blue staining of the SDS-polyacrylamide gel indicated that similar amount of the immunoprecipitated HA-IKKβ and substrate peptides were applied in the assay ( Figure 6 , lower panels ) . To study the effect of B14 on IKK activity further , the IKK complex was activated by overexpression of either IKKα or IKKβ ( Figure 7A ) , and B14 was found to inhibit this activation significantly and in a dose-dependent manner ( Figure 7A ) . This indicated that B14 acts at , or downstream of , the IKK signalosome . The site of action was investigated further using IKK constitutively active mutants , IKKα SS/EE and IKKβ SS/EE that contain mutations in the activation loop [2] ( Figure 7B ) . B14 inhibited IKKα SS/EE significantly and in a dose-dependent manner . In contrast , there was only a small ( 15% ) reduction of IKKβ SS/EE-induced NF-κB activation in the presence of the highest amount of B14 . These findings imply that once IKKβ is activated , B14 can no longer prevent NF-κB activation and also suggest a model in which B14 inhibits activation of the IKK complex by preventing phosphorylation of IKKβ in the activation loop . This hypothesis was tested directly by using Ab to detect IKKβ that has been phosphorylated in the activation loop at serine 177 and 181 ( Figure 8 ) . HA-tagged IKKβ was transfected into 293 T cells either alone or together with increasing concentrations of B14 . In the absence of transfected HA-IKKβ no phospho-IKKβ was detected , but after addition of HA-IKKβ , phospho-IKKβ was observed easily and was reduced in a dose-dependent manner as the concentration of B14 increased . Notably , while the amount of phospho-IKKβ decreased in the presence of B14 , the amount of total HA-IKKβ remained fairly constant and blotting for tubulin confirmed equal loading of samples . Therefore , B14 inhibits NF-κB activation by preventing phosphorylation of IKKβ in the activation loop .
In this study , VACV protein B14 is shown to inhibit the IKK complex and to downregulate NF-κB-dependent gene expression , which is crucial for the innate and adaptive immune response to infection [3 , 4] . Our previous in vivo study , using recombinant VACVs that do or do not express B14 , demonstrated B14 is an intracellular virulence factor that modulates the inflammatory response in vivo [37] . The activity of B14 described here is consistent with this phenotype: downregulation of NF-κB-dependent expression of pro-inflammatory cytokines will alter recruitment of inflammatory cells to sites of infection and so diminish the ability of the host to fight infection . Notably , a virus lacking the B14R gene was attenuated compared to parental virus [37] . The IKK complex is critical for activation of NF-κB [3 , 44–46] and therefore is a logical target for modulation by pathogens [4 , 47] . B14 is one of several VACV proteins that inhibit signaling pathways leading to NF-κB activation , but these proteins all have non-redundant functions because when the gene encoding each inhibitor is deleted individually , the deletion mutant displays an in vivo phenotype [24 , 29–31 , 37] . Therefore , these proteins must each have distinct functions . In this regard , B14 differs from A46 and A52 in that it targets a broader array of immune signaling pathways; for instance , A46 and A52 inhibit IL-1 but not TNF-induced signaling , whereas B14 inhibits both ( Figure 1A ) . Also A46 and A52 target the signaling pathways upstream of the IKK complex [29–31] , whereas B14 targets the activity of the IKK complex . B14 also differs from N1 in that N1 was reported to inhibit signaling pathways leading to NF-κB activation [32] and to IFN responses via TBK1 [32] , whereas B14 did not inhibit IFN responses induced by either IFNα or poly ( I:C ) ( Figure 1E ) . N1 was reported to target to the IKK signalosome by binding to the kinase complex when both components were overexpressed [32] . However , three independent experiments shown here contradict this: first , N1 did not bind to IKK components in the LUMIER assay ( Figure 2B ) ; second , N1 did not co-purify with IKK during biochemical fractionation of infected cells ( Figure 2C ) ; and third , N1 did not co-precipitate with IKK components using the anti-NEMO mAb ( unpublished data ) . In addition , we showed previously that under the conditions tested N1 did not affect NF-κB activation in VACV-infected cells [33] . Therefore , B14 , but not N1 , associates with the IKK complex and thereby inhibits NF-κB responsive gene expression . Concerning the site of action of B14 , it is clear that B14 shuts down expression of reporter genes with NF-κB-responsive promoters in response to multiple stimuli ( Figure 1 ) and that within infected cells the overall level of IκBα is not altered by virus infection ( Figure 5 ) or by the expression of B14 in resting cells ( Figure 4 ) . However , in the presence of B14 there is a reduced level of phospho-IκBα in the infected cell lysates ( Figure 5 ) and a reduced degradation of IκBα in TNFα-stimulated cells . These findings suggest a possible effect of B14 on the IKK activity . Direct evidence for the reduced phosphorylation of IκBα by IKK in the presence of B14 was provided by an in vitro kinase assay using a synthetic IκBα peptide substrate and IKK that had been immunoprecipitated from cells ( Figure 6 ) . Therefore , the mechanism of action of B14 lies upstream of IκBα phosphorylation . Consistent with this , B14 was found to co-purify with the IKK complex from infected cells and to co-precipitate with the IKK complex using specific antibodies either against tagged B14 , NEMO ( Figures 2 and 3 ) , or against IKKα/β ( unpublished data ) . Notably , the assembly of the IKK complex was not interrupted by B14 . Furthermore , use of IKK null MEFs revealed that IKKβ is the target of B14 in the complex and B14 did not bind to or disrupt the IKKα-NEMO complex ( Figure 3B ) . These findings indicate that the inhibitory effect of B14 on the activity of the IKK complex is not due to disassembly of the IKK complex . B14 inhibited NF-κB activation driven by overexpression of either IKKα or IKKβ ( Figure 7A ) or by expression of the constitutively active IKKα SS/EE mutant in which the ser176 and ser180 in the activation loop were mutated to glutamic acid ( Figure 7B ) . In contrast , B14 was unable to inhibit NF-κB gene expression by a similar constitutively active IKKβ SS/EE mutant ( Figure 7B ) , indicating IKKβ but not IKKα is the target for B14 . Furthermore , B14 associated with the IKK complex via IKKβ ( Figure 3 ) and inhibited phosphorylation of IKKβ in the activation loop ( Figure 8 ) , thereby downregulating the activity of the IKK complex . However , once the IKKβ subunit is activated , B14 may not be inhibitory . B14 co-purified with the IKK complex but was also present in a 160-kDa complex , much larger than the mass of monomeric B14 ( 17 . 3 kDa ) . Consistent with these findings , recombinant B14 made in Escherichia coli was oligomeric ( unpublished data ) . Whether B14 is the only protein in the 160-kDa complex or whether it is complexed with other unidentified cellular or viral protein ( s ) is unknown . However , its presence in this complex suggests B14 might have function ( s ) additional to that described here . For instance , the slight increase of PMA-induced AP-1 activity in the presence of B14 ( Figure 1D ) may result from interaction of B14 with an unidentified protein ( s ) . Alternatively , this may be a consequence of the downregulation of NF-κB responsive genes that negatively regulate AP-1 activity . There is ample precedent for small VACV proteins having more than one immunomodulatory activity . For instance , protein A52 is both a TLR inhibitor and an activator of p38 kinase to modulate IL-10 [48] . In summary , VACV virulence factor B14 inhibits the IKK signalosome by preventing phosphorylation of IKKβ in the activation loop , resulting in inhibition of NF-κB-dependent gene expression . This mechanism of action fits with the observed increased inflammatory response in vivo to infection with a virus lacking gene B14R [37] . Overall our findings reveal a novel strategy used by VACV to modulate cellular signaling pathways to aid viral immune evasion . The B14 may be an interesting target to develop anti-inflammatory therapeutics directed against the IKK complex .
Human embryonic kidney ( HEK ) 293 cells ( a gift from Dr . Paul Farrell , Imperial College London ) , wild type , IKKα null , and IKKβ null MEF cells ( provided by Dr . Michael Karin , UCSD ) were cultured in Dulbecco's modified Eagle's medium ( DMEM , Gibco BRL ) supplemented with 10% heat-treated foetal bovine serum ( FBS , heat-treated at 56 °C for 30 min , Harlan Sera-Lab ) , 50 IU/ml penicillin and 50 μg/ml streptomycin ( Gibco BRL ) and 2 mM L-glutamine ( Gibco BRL ) . HeLa cells were maintained in Minimum Essential Medium ( MEM Gibco BRL ) supplemented with 1 x non-essential amino acid solution ( Sigma ) and identical chemicals as DMEM . The cells were incubated in a humidified incubator ( Heraeus ) with 5% CO2 . Expression vectors , VACV strains that do or do not express B14 , and rabbit anti-serum against B14 have been described previously [37] . Plasmids expressing IKKs and IKK constitutively active mutants were kindly provided by Dr . Alain Chariot ( University of Liège ) and Dr . Richard Gaynor ( Lilly Corporate Center ) , respectively . Reporter and TRAF2 plasmids were gifts from Dr . Andrew Bowie ( Trinity College Dublin ) . Anti-IKKγ ( NEMO ) ( BD Biosciences ) , anti-HA ( Cambridge Biosciences ) , anti-GSK3β ( BD Biosciences ) mAbs were used for immunoprecipitation or immunoblotting . For immunoblotting , rabbit polyclonal anti-IKKα ( Cell Signaling ) , anti-IKKα/β ( Santa Cruz ) , NEMO ( Cell Signaling ) , and IκBα ( Santa Cruz ) were used . In addition , murine mAb anti-P-IκBα ( Cell Signaling ) , α-tubulin , IKKα and IKKβ ( Upstate ) were used . The anti-N1 polyclonal Ab was described previously [24] . Lastly , rabbit mAb against phospho-IKKα/β ( 16A6 , Cell Signaling ) was used to detect IKKβ that is phosphorylated at Ser177/181 . HeLa cells ( 8 × 104 per well ) were seeded and then transfected with 100 ng of reporter plasmids , 50 ng of pSV-β-galactosidase ( Promega ) , and the indicated amount of expression vectors with FuGENE 6 ( Roche ) . The total amount of DNA ( 400 ng ) was kept constant by supplementation with pCI ( Promega ) . After overnight incubation , the transfected cells were simulated with 100 ng/ml of IL-1β , TNFα ( Peprotech ) , or 50 ng/ml of PMA ( Sigma ) for 8 h . Cells were harvested in passive lysis buffer ( Promega ) , and the relative stimulation of NF-κB activity was calculated by normalizing luciferase activity with β-galactosidase activity . HEK 293 cells ( 6 × 104 per well ) were seeded into 24-well tissue culture plates overnight before transfection . Reporter plasmids ( 90 ng ) , 10 ng of pTK-Renilla luciferase ( pRL-TK , a gift from Dr . Andrew Bowie ) , and the indicated amount of expression vectors were delivered into cells with FuGENE 6 . The total amount of DNA ( 500 ng ) was kept constant by supplementation with pCI ( Promega ) . After 24 h , cells were harvested in passive lysis buffer ( Promega ) , and the relative stimulation of NF-κB-dependent gene expression was calculated by normalizing luciferase activity with Renilla luciferase intensity . In case of stimulations , the cells were incubated with stimuli described previously and 5 μg/ml of poly ( I:C ) ( Invitrogen ) for 12 h before lysis . In all cases , data shown are from one of three to five independent experiments with similar qualitative results . Data from experiments performed in triplicate are expressed as means ± SD . HeLa cells in 10-cm dishes were infected by vB14-HA or vΔB14 ( 10 p . f . u . per cell ) . At 4 h p . i . , the infected cells were washed once with ice-cold PBS and lysed with IP buffer ( 50 mM HEPES [pH 7 . 5] , 100 mM NaCl , 1 mM EDTA , 10% [v/v] glycerol , 0 . 5%[v/v] Nonidet P-40 containing 1 mM phenylmethylsulphonyl fluoride , 0 . 01% [v/v] aprotinin , and 1 mM sodium orthovanadate ) . For immunoprecipitation , the indicated antibodies were pre-incubated with protein G-Sepharose ( Amersham ) at 4 °C for 1 h . Then equal amounts of the beads were added and incubated with the cell lysates overnight at 4 °C . The immune complexes were washed , boiled with 30 μl of 5 x sample buffer , and analysed by immunoblotting . Proteins were resolved and transferred to nitrocellulose membranes ( Hybond ECL , Amersham ) . After transfer , the membranes were rinsed once in PBS and then incubated with blocking buffer ( PBS containing 5% Marvel milk powder ) for 30 min at RT . The primary Ab was added to the blocking buffer and incubated for 1 h on a rocking platform . The membranes were washed five times , for 6 min , with PBS , and then HRP-conjugated secondary Ab ( Sigma ) was added in blocking buffer . After 45-min incubation , the membranes were washed as above and then incubated with chemiluminescence reagent ( ECL , Amersham ) for signal detection . The membranes were wrapped in Saran wrap and exposed to X-ray film ( Kodak ) . At 4 h p . i . , the infected cells were washed once with ice-cold PBS and lysed with CE buffer ( 10 mM 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ( HEPES ) , 10 mM KCl , 0 . 1 mM EDTA ( pH 8 . 0 ) , 0 . 1 mM EGTA , 1 mM DTT , 0 . 05 % NP-40 , 20 mM β-glycerophosphate , 1 mM Na3VO4 , 1 x protease inhibitor cocktail II ( CalBio ) for 30 min at 4 °C . The extract was centrifuged at 10 , 000 x g for 1 h at 4 °C . Five hundred μl of the supernatant was loaded onto a Superose 6 gel filtration column ( Amersham ) that had been equilibrated in gel filtration ( GF ) buffer ( 25 mM Tris-HCl [pH 7 . 6] , 150 mM NaCl , 0 . 2% NP-40 , 1 mM DTT ) . Approximately 10%–15% of each fraction was analysed by SDS-PAGE followed by immunoblotting with the indicated Ab . The column was calibrated in the GF buffer using protein standards kits from Amersham . HEK 293 cells were transfected using the indicated vectors overnight and lysed in lysis buffer ( 20 mM Tris [pH 7 . 4] , 150 mM NaCl , 1 mM EDTA , 1 mM EGTA , 1% Triton X-100 , 2 . 5 mM sodium pyrophosphate , 1 mM β-glycerophosphate , 1 mM Na3VO4 , 1 μg/ml leupeptin , and 1 mM phenylmethylsulfonyl fluoride ) . The HA-tagged IKK proteins in the cell lysate were immunoprecipitated using anti-HA mAb . The precipitate was washed three times in lysis buffer and twice in kinase buffer ( 20 mM Hepes/KOH [pH 7 , 4] , 25 mM β-glycerophosphate , 2 mM dithiothreitol , 20 mM MgCl2 ) . The kinase assay was performed in a final volume of 20 μl of kinase buffer containing 10 μM ATP , 5 μCi of [γ-32P] ATP and 1 μg of IKK substrate peptide ( Upstate ) derived from IκBα sequence ( KKKKERLLDDRHDSGLDSMKDEE ) . After incubation for 10 min at 30 °C , the reaction was stopped by the addition of 5× SDS sample buffer . Proteins were separated by SDS-PAGE and stained by Coomasie blue . 32P-labelled proteins were visualized by autoradiography . For LUMIER assays [42] , 293 ET cells were transfected with a pair of putative interactors fused to Renilla luciferase or HA/FLAG antibody tags . Post-nuclear supernatants from cells lysed in IP buffer ( 10% glycerol , 150 mM NaCl , 20 mM Tris-HCl [pH 7 . 4] , 0 . 1 % Triton-X100 , and inhibitors ) were incubated with HA or FLAG agarose ( Sigma ) . After washing , proteins were eluted for 30 min with 150 μg/ml FLAG peptide or 100 μg/ml HA peptide in Renilla lysis buffer ( Promega ) . The ratio between luciferase activity in eluates and lysates is presented as fold binding over a control reaction . | Vaccinia virus ( VACV ) is the live vaccine used to eradicate smallpox and is also the most intensively studied poxvirus . Like many poxviruses , VACV produces a wide variety of proteins that inhibit parts of the host response to infection . Consequently , the virus can escape destruction by the immune system and be passed on to additional hosts . Here we report a new VACV immune evasion mechanism mediated by protein B14 , a protein that contributes to virus virulence . B14 functions by interacting with a cellular protein called IKKβ , which is critical for mounting an innate immune response to infection , and also plays important roles in cancer and cell death . B14 prevents IKKβ being activated and consequently the cellular signaling pathway leading to activation of nuclear factor kappa B ( NF-κB ) is not induced . Without activation of NF-κB the host cell cannot produce other molecules that amplify the innate immune response to infection . This mechanism of action of B14 fits nicely with the observed increase in the host response to infection by a VACV strain lacking the B14R gene . Lastly , an increased understanding of how B14 inhibits IKKβ function may lead to development of novel drugs against this important cellular enzyme . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"viruses",
"immunology",
"virology"
] | 2008 | Inhibition of IκB Kinase by Vaccinia Virus Virulence Factor B14 |
Most of the somatic cells of adult metazoans , including mammals , do not undergo continuous cycles of replication . Instead , they are quiescent and devote most of their metabolic activity to gene expression . The mutagenic consequences of exposure to DNA–damaging agents are well documented , but less is known about the impact of DNA lesions on transcription . To investigate this impact , we developed a luciferase-based expression system . This system consists of two types of construct composed of a DNA template containing an 8-oxoguanine , paired either with a thymine or a cytosine , placed at defined positions along the transcribed strand of the reporter gene . Analyses of luciferase gene expression from the two types of construct showed that efficient but error-prone transcriptional bypass of 8-oxoguanine occurred in vivo , and that this lesion was not repaired by the transcription-coupled repair machinery in mammalian cells . The analysis of luciferase activity expressed from 8OG:T-containing constructs indicated that the magnitude of erroneous transcription events involving 8-oxoguanine depended on the sequence contexts surrounding the lesion . Additionally , sequencing of the transcript population expressed from these constructs showed that RNA polymerase II mostly inserted an adenine opposite to 8-oxoguanine . Analysis of luciferase expression from 8OG:C-containing constructs showed that the generated aberrant mRNAs led to the production of mutant proteins with the potential to induce a long-term phenotypical change . These findings reveal that erroneous transcription over DNA lesions may induce phenotypical changes with the potential to alter the fate of non-replicating cells .
During replication , DNA lesions exert deleterious effects by either blocking the DNA polymerase or allowing for mutagenic bypass of the lesion , which may be of major importance for evolution , hereditary diseases and cancer [1] . However , outside the unnatural environment of the laboratory , few cells undergo continuous cycles of division , and most cells exist instead in a non replicating state [2] . For example , several of the organs of multicellular organisms consist principally of non-dividing cells , the lifespan of which is limited by the functional differentiation associated with their normal physiology . These cells do not replicate their genome , but must nonetheless express a large number of genes for their physiological maintenance , which depends on the fidelity of both DNA transcription and mRNA translation . DNA lesions may be caused by a plethora of physical and chemical agents present in the natural environment . RNA polymerases would therefore be expected to encounter such lesions frequently , but much less is known about the interaction of the transcription machinery with such lesions than about the effects of these lesions on replication . Most studies concerning the relationships between RNA polymerases and DNA lesions focus on bulky or distortive DNA damages . Such damage generally arrests elongation and is eliminated by transcription-coupled repair ( TCR ) . This subpathway of the nucleotide excision repair pathway removes RNA polymerase II ( RNApolII ) -arresting lesions from the transcribed strand ( TS ) of genes by recruiting the DNA excision machinery [3] . However , some DNA lesions are bypassed by an elongating RNApolII in vitro , which can miscode at the lesion site and produce mutant transcripts with high efficiency via a process known as transcriptional mutagenesis ( TM ) [4] , [5] . Interestingly , it was recently reported that even distortive DNA lesions , such as 8 , 5′-cyclo-2′-deoxyadenosine and cyclo-pyrimidine dimer , are bypassed at low frequency by human RNApolII in vivo , leading to the production of mutant transcripts [6] . If these events occur in cells , then each round of transcription of the sequence including the lesion would result in the production of an mRNA with a targeted change that will be translated multiple times to produce a relatively large population of mutant proteins . TM may therefore induce major phenotypical changes and important biological outcomes , particularly in cells that are not dividing [4] , [7] . A frequently occurring DNA lesion results from the direct oxidation of guanine to generate 7 , 8-dihydro-8-oxoguanine ( 8OG ) [8] . In Escherichia coli , 8OG is bypassed by the RNA polymerase , leading to TM events due to the insertion of adenine or no nucleotide opposite to this lesion [9] . Several in vitro studies have indicated that 8OG could also be the source of TM in human cells , as it does not represent a strong block for an elongating RNA polII and , in various experimental conditions , the bypass of this lesion has been shown to result in the erroneous incorporation of adenine opposite to the lesion [10]–[12] . The tendency of 8OG to induce TM in murine cells was also reported in a recent study [13] . In this study , we focused on the outcome of 8OG-mediated TM in mammalian cells , including human cells in particular . A Photinus pyralis luciferase ( Ppluc ) reporter system has been used to examine the occurrence of 8OG-mediated TM in diverse mammalian cells and to investigate the effects of DNA repair capacity on these TM events in human and mouse cells . Two types of construct were used in this study , in which 8OG was introduced into the transcribed strand of the Ppluc gene , opposite either a thymine ( 8OG:T ) or a cytosine ( 8OG:C ) . In cells , the 8OG:T mispair constitutes a poor substrate for DNA repair mechanisms . Consequently , many rounds of transcription over 8OG occur before the complete removal of this lesion from the transcribed strand of this type of construct [14] , [15] . This would result in an amplification of 8OG-induced TM events , thereby facilitating studies of such events . However , most of guanine oxidation process leads to 8OG:C pairs in DNA , which are rapidly processed through hOGG1-mediated base excision repair ( BER ) [14] . We therefore also investigated the ability of 8OG-mediated TM events to induce a transient phenotypical change with 8OG:C-containing constructs . An analysis of the relative Ppluc activity expressed from 8OG:T mispair-containing constructs showed that the extent of 8OG-mediated TM is , similarly to what has been found for DNA polymerase , largely dependent on the context sequence and , probably on the relative distance of the lesion from the promoter . Ppluc mRNAs expressed from 8OG:T mispair-containing constructs were extracted from cells with high levels of TM and sequenced , to identify the spectrum of RNApolII misinsertion events induced by transcription over the 8OG lesion . Quantification of the Ppluc activities expressed from 8OG:C-containing constructs confirmed the hypothesis that the extent of TM depends strongly on the DNA repair capacity of the cell . Such quantification also showed that TM was a potential source of a long-term phenotypical change , even in cells with a normal DNA repair background . With both types of construct , we assessed the effect of the level of reporter gene expression on TM by modifying the amount of reporter mRNA produced , using a dose-dependent doxycycline-responsive promoter . Furthermore , the Ppluc activities expressed from both types of construct in various TCR-deficient cells provided insight into the role of this mechanism in the repair of an 8OG lesion in the transcribed strand of a gene . These observations may have potentially important implications for the etiology of diseases , including those affecting non-dividing cells in particular .
We investigated the effect of 8OG on transcription and phenotypical change in mammalian cells by using a reporter assay to measure the levels of active Ppluc generated from expression constructs derived from the pBDA6 plasmid ( Figure S1 ) and containing DNA lesions at defined positions on the TS of the gene ( Figure 1 ) . Five sets of three constructs were generated with the following nomenclature ( lesion-free strand ( LFS ) or 8OG-containing strand ( 8OG ) / amino acid specified on the NTS ) codon number and composed as follows: ( i ) a wild-type construct with the wild-type sequence of the Ppluc gene [ ( LFS/Lys ) 5 , ( LFS/Lys ) 297 , ( LFS/Glu ) 344 , ( LFS/Asp ) 422 and ( LFS/Lys ) 445]; ( ii ) an 8OG-containing construct [ ( 8OG/Stop ) 5 , ( 8OG/Stop ) 297 , ( 8OG/Ala ) 344 , ( 8OG/Ala ) 422 and ( 8OG/Stop ) 445] in which the 8OG was introduced into the TS of the specified codon and ( iii ) a mutant construct [ ( LFS/Stop ) 5 , ( LFS/Stop ) 297 , ( LFS/Ala ) 344 , ( LFS/Ala ) 422 and ( LFS/Stop ) 445] ( Figure 1 ) . In three of the mutant constructs , a lysine codon within the Ppluc gene ( codon number 5 , 297 or 445 ) was replaced by a premature stop codon , resulting in the production of an inactive C-terminally truncated protein [16] . The other two mutant constructs , specifying alanine at codon 344 or 422 , resulted in the production of an inactive form , E334A or D422A , of the Ppluc protein ( Branchini , B . R . personal communication ) . In transfected cells , expression of the Ppluc gene from these different constructs was driven by the dose-dependent doxycycline-responsive Ptight promoter and protein activity was normalized with respect to the Renilla reniformis luciferase ( Rrluc ) . Both luciferases are independently translated from the same polycistronic mRNA , with Rrluc translation initiated at an internal ribosome entry site ( IRES ) located between the two open reading frames ( Figure S1 ) . For each assay and for each cell line ( Table 1 ) , normalized Ppluc activities measured after transfection with wild-type constructs was set as the 100% reference point for quantifying relative Ppluc activities expressed from the same cell line transfected with mutant or 8OG-containing constructs . The relative Ppluc activity of the cell lines transfected with mutant constructs was very low and varied from 0 . 001% to 0 . 022% ( Table 2 and Table 3 ) . These results confirm that the method used to generate the constructs was appropriate for this study and that expression of the Ppluc gene from mutant constructs resulted in the production of inactive proteins . The 104– to 105–fold difference in Ppluc activity between wild-type and mutant constructs is large enough for measurement of the intermediate levels of activity potentially generated by the TM events induced by 8OG . The extent of 8OG-induced TM was determined with ( 8OG/Stop ) constructs , which contain an 8OG:T mispair in codon 5 , 297 or 445 ( Figure 1 ) . Transcription through the lesion and the insertion of adenine or cytosine opposite to the 8OG would result in a Ppluc mRNA encoding lysine or glutamine at the corresponding codon . The insertion of a lysine residue at this position results in fully active wild-type Ppluc , whereas the insertion of a glutamine residue at position 5 , 297 or 445 leads to the production of a Ppluc protein with activity levels 5% to 315% that of the wild-type Ppluc ( Table 2 ) . Alternatively , base excision repair ( BER ) of this 8OG would result in the production of a Ppluc mRNA containing a premature stop codon , which would therefore not give rise to an active Ppluc ( Figure 2 ) . Ppluc activities were 100 to 1 , 000 times higher in normal cells transfected with ( 8OG/Stop ) constructs than in normal cells transfected with ( LFS/Stop ) constructs . The relative Ppluc activities expressed from ( LFS/Stop ) constructs were very low and similar in all cell lines tested , whereas the relative activity of Ppluc measured in normal human ( MRC5V1 and VA13 ) and murine ( MEF ) cells transfected with ( 8OG/Stop ) constructs depended strongly on the position of the lesion in the TS of the Ppluc gene ( Table 2 ) . These relative activities are indeed ranging from less than 1% , if the 8OG was located at codon 5 , to more than 50% if the 8OG was located at codon 445 ( Table 2 ) . In cells transfected with ( LFS/Stop ) constructs , activities of the coexpressed Rrluc were systematically high and similar to those measured in cells transfected with wild-type constructs , thus ruling out the involvement of nonsense-mediated decay in the modulation of relative Ppluc activity , because the same transcript encodes both luciferases . The repair of 8OG:T mispairs in normal cells therefore seems to depend largely on the sequence context and , possibly , on the distance between the promoter of the transcribed gene and the mispair . Differential 8OG:T mispair repair in a transcribed gene could potentially be affected by the level of expression of the gene . We tested this hypothesis by lowering the level of Ppluc/Rrluc mRNA production by decreasing the amount of doxycycline in the recovery medium for transfected cells ( Table 2 ) . A comparison of Rrluc activities shows that production of the reporter mRNA under control of the pTight promoter decreased by a factor of about 10 when the concentration of doxycycline wasdecreased from 2 µg/ml to 1 ng/ml ( data not shown ) . The relative Ppluc activities expressed in MRC5V1 cells transfected with ( 8OG/Stop ) constructs were similar for both doxycycline concentrations . Similar results were also obtained with 8OG:C-containing constructs ( Table 3 ) . Taken together , these results indicate that , over the range tested , the expression level of the mRNA does not affect the 8OG repair process and that 8OG-induced TM events occur at similar frequency whether the gene is strongly or weakly expressed . The Ppluc relative activity expressed in cells transfected with ( 8OG/Stop ) constructs is directly correlated with the efficiency of 8OG:T mispair repair . It has been shown in vitro that 8OG can be removed from an 8OG:T mispair-containing DNA molecule by either hOGG1-driven BER or by the mismatch repair system ( MMR ) , in an hMSH2/hMSH6-dependent manner [14] , [15] . The role of OGG1-driven BER in the differential 8OG:T repair efficiency was deciphered by quantifying the relative Ppluc activity expressed from normal ( MEF ) or Ogg1-deficient ( MEF ogg1 −/− ) murine cell lines transfected with ( 8OG/Stop ) constructs [17] . No significant difference in relative Ppluc activity was observed between MEF and MEF ogg1 −/− cells transfected with the ( 8OG/Stop ) 445 construct , whereas the relative Ppluc activities of MEF ogg1 −/− cells transfected with ( 8OG/Stop ) 5 or ( 8OG/Stop ) 297 were significantly higher by factors of 5 and 2 . 7 , respectively , than those of MEF cells transfected with the same constructs ( Table 2 ) . The impact of MMR on the differential repair efficiency of an 8OG:T mispair was assessed by using our constructs to transfect hMLH1- ( HCT116 ) , hMSH6- ( DLD-1 ) or hMSH2-deficient ( LoVo ) cells . Relative Ppluc activities expressed in hMSH2-deficient cells transfected with ( 8OG/Stop ) constructs were higher than those in normal cells , whereas the relative activities of hMSH6- and hMLH1-deficient cells were lower than those in normal cells ( Table 2 ) . The only significant differences with respect to normal cells ( MRC5 and VA13 ) were obtained for hMSH2- and hMSH6-deficient cells transfected with the ( 8OG/Stop ) 445 construct , indicating a possible key role of these proteins in the 8OG MMR-dependent repair pathway . These results indicate that both MMR and BER are involved in repairing 8OG:T mispairs in vivo . The high relative Ppluc activities expressed in cells transfected with the ( 8OG/Stop ) 445 construct and the consistently high levels of Rrluc activity in cells transfected with ( 8OG/Stop ) , in which Rrluc activity levels were similar to those in cells transfected with wild-type constructs , suggest that the presence of an 8OG on the TS of a gene does not block transcription and that the human and murine RNApolII enzymes incorporate adenine or cytosine opposite to 8OG . However , detectable enzyme activity cannot be viewed as direct evidence for TM , as both nucleotide insertions result in the production of Ppluc enzymes with various degrees of activity . As mentioned above , the 8OG:T mispair in the ( 8OG/Stop ) 445 construct constitutes a very poor substrate for DNA repair and many rounds of transcription occur before the removal of the lesion from the DNA template . Thus , analyses of Ppluc mRNA sequences produced in cells transfected with the ( 8OG/Stop ) 445 construct should provide an accurate description of the spectrum of misinsertion events occurring during transcription over an 8OG lesion . We identified the nucleotides inserted opposite 8OG by the human RNApolII by sequencing partial cDNA subclones obtained from RNA extracted from MRC5V1 cells transfected with the ( 8OG/Stop ) 445 construct . The major cDNA type ( 85% ) harbors an AAA lysine codon at position 445 , the expected sequence when adenine is incorporated opposite to 8OG through TM ( Figure 1 and Figure 2 ) . The other two minor cDNA types contain a TAA stop codon ( 3% ) , reflecting the transcription of repaired ( 8OG/Stop ) 445 molecules , or a CAA glutamine codon ( 12% ) . Thus , in human cells , RNApolII can generate mutated transcripts containing an adenine residue at the position corresponding to the lesion during transcription over 8OG . The use of ( 8OG/Stop ) constructs provided important insight into the repair of an 8OG:T mispair in cells and the spectrum of nucleotide insertions occurring opposite to the 8OG lesion during in vivo transcription by human RNApolII over this lesion . However , 8OG:T mispairs occur only rarely in vivo , because guanine oxidation mostly generates 8OG:C pairs . The ability of 8OG to induce a phenotypical change through TM was investigated with ( 8OG/Ala ) constructs containing an 8OG:C pair at codon 344 or 422 ( Figure 1 ) . Active Ppluc proteins can be produced from these constructs only through the insertion of an adenine residue opposite to the 8OG , resulting in the production of an mRNA with the wild-type Ppluc gene sequence . Although 8OG:C pair is a good substrate for OGG1-mediated repair , levels of relative Ppluc activity in human ( MRC5V1 and VA13 ) and murine ( MEF ) cell lines transfected with ( 8OG/Ala ) constructs were from 57- to 1300-fold higher than those obtained following transfection of these same cell types with ( LFS/Ala ) constructs ( Table 3 ) . Thus , in vivo , the murine and human RNApolII enzymes can transcribe through an 8OG lesion , inducing the misincorporation of adenine opposite to this lesion , resulting in a significant phenotypical change . The magnitude of this phenotypical change may depend on the DNA repair capacity of the cells , as repair of the 8OG would convert codon 344 or 422 to an alanine codon , leading to the production of inactive Ppluc . We assessed the extent to which the phenotypical change depended on the DNA repair capacity of the cells by transfecting mouse cells lacking OGG1-mediated BER with ( 8OG/Ala ) constructs [17] . The relative Ppluc activities of MEF ogg1 −/− cells transfected with ( 8OG/Ala ) 344 or ( 8OG/Ala ) 422 were 36 . 4- and 74 . 3-fold higher , respectively , than those for the normal parental cell line ( MEF ) transfected with the same constructs ( Table 3 ) . These findings thus demonstrate that the impact of TM on the phenotype depends on the DNA repair capacity of the cells ( Table 3 ) . An 8OG lesion in a TS might also be repaired by pathways other than OGG1-mediated BER , possibly including TCR , as cells from patients with Cockayne syndrome have been shown to be defective for both TCR and the repair of oxidative lesions [18] . Nonetheless , the role of TCR in the repair of oxidative lesions , such as 8OG , remains debatable , as several papers addressing this question have recently been retracted [19]–[21] . In our system , the TCR-mediated repair of 8OG should be revealed by a higher level of phenotypical change in TCR-deficient cells transfected with ( 8OG/Ala ) constructs and higher relative Ppluc activities in cells transfected with ( 8OG/Stop ) constructs . However , the relative Ppluc activities expressed from CS- and XP/CS-derived cells transfected with ( 8OG/Ala ) or ( 8OG/Stop ) constructs fell within the same range as those for normal cells transfected with these constructs ( Table 2 and Table 3 ) , ruling out the possibility of TCR-mediated repair of 8OG . The change in phenotype observed for normal cells transfected with ( 8OG/Ala ) should not be permanent , as the 8OG lesion responsible for inducing this change should be repaired over time . We evaluated the magnitude of the phenotypical change induced by 8OG over time by assessing the production of active Ppluc at various times after the transfection of MRC5V1 cells . Higher levels of active Ppluc were consistently expressed from ( 8OG/Ala ) constructs than from ( LFS/Ala ) constructs , over a period of at least seven days after transfection ( Figure 3 ) . The observed differences were significant for up to four days after transfection , but the difference observed on day 7 was not significant as , for each construct , only one of the six replicates displayed levels of Ppluc activity above the background , a phenomenon similar to the so-called “mutagenesis jackpot” [22] . Similar decreases in Ppluc activities were observed with wild-type and ( 8OG/Ala ) constructs , but these results clearly indicate that the TM process induced by 8OG can lead to a long-term phenotypical change in the affected cells .
In vitro studies have shown that 8OG does not block the progression of the mammalian RNApolII and that non mutagenic cytosine insertions opposite to this lesion are favored , although the insertion of a certain number of adenine residues is also detected [10]–[12] , [23] . Analysis of the cDNA population generated from the Ppluc mRNA produced in MRC5V1 cells transfected with the ( 8OG/Stop ) 445 construct revealed that in vivo transcription of 8OG generates two distinct populations of transcripts . The largest of these two populations consisted of mutated mRNA molecules containing an adenine residue incorporated opposite to the 8OG during transcription . The other population consisted of transcripts in which a cytosine residue was incorporated at the position corresponding to the lesion , probably due to non mutagenic transcription over 8OG . This type of cDNA could potentially result from faithful transcription over across the 8OG lesion , but may also result from the transcription of ( 8OG/Stop ) 445 molecules repaired by MMR . Indeed , it has been shown that the binding of hMSH2/hMSH6 to an 8OG:T mispair can promote excision of the 8OG-free strand and that adenine and cytosine are inserted with similar efficiencies opposite to the 8OG during repair synthesis , resulting in 8OG:A- or 8OG:C-containing molecules [15] . For 8OG:A , a two-step pathway has been proposed in which the incorporated adenine is excised by hMYH and a cytosine is inserted during repair synthesis [1] , [24] . The resulting 8OG:C-containing DNA is then used as a substrate for hOGG1 [25] , which can replace the 8OG by a guanine residue , creating a glutamine codon ( 3′-GTT-5′ ) in the TS of the Ppluc gene ( Figure 2 ) . Saxowsky et al . recently reported cytosine incorporation to be the major event observed during 8OG bypass by murine RNApolII , with adenine incorporation observed in about 10% of transcripts [13] . This apparent discrepancy may be due to differences in sequence context . As reported above , our results clearly indicate that sequence context may have a major influence on the outcome of 8OG-induced TM in mammalian cells . The nature of the nucleotide paired with the 8OG in the DNA template may also account for this difference . Indeed , if 8OG is placed opposite a cytosine residue , about 28% of the transcripts contain an adenine at the position corresponding to the lesion after the expression of their reporter gene in MEF ogg1 −/− cells [13] . Our findings are consistent with those of Saxowsky et al . , because we found that 15 to 20% adenine-containing transcripts were produced when 8OG:C pair-containing constructs were expressed in MEF ogg1 −/− cells ( Table 3 ) . Therefore , these results indicate that , in vivo , adenine insertion by human RNApolII during transcription across an 8OG lesion is a major event . Our findings clearly demonstrate that TM can be induced during transcription over an 8OG lesion in vivo . Consequently , 8OG must be removed from the DNA before RNApolII encounters this lesion , to avoid the production of mutant transcripts and mutant proteins . Significant DNA repair pathway activity is therefore required in conditions of non-growth in the absence of DNA replication . The use of an 8OG:T mispair-containing construct was crucial for analysis of the specificity of base incorporation opposite to this lesion during transcription by RNApolII . This mispair probably occurs rarely in cells , because guanine oxidation in DNA mostly results in the production of 8OG:C pairs , the best substrate for OGG1-mediated repair [26] . However , even in cells not deficient for any of the known DNA repair pathways , significant amounts of active Ppluc protein were expressed in cells transfected with 8OG:C pair-containing constructs . Thus , in mammalian cells , the oxidation of a guanine residue in the TS of a gene may lead to major phenotypical changes , as the only difference between the ( LFS/Ala ) and ( 8OG/Ala ) constructs is the replacement of a normal guanine residue by 8OG ( Figure 1 ) . This simple and only difference allows for cells transfected with ( 8OG/Ala ) constructs to express non negligible amounts of active Ppluc protein through TM , rendering them phenotypically different from the same cells transfected with ( LFS/Ala ) constructs , which produce no active Ppluc protein . Thus , when there is an 8OG residue present in the TS of an expressed gene , RNApolII continually produces transcripts containing a G to A base substitution at the same position , potentially leading to a phenotypical change . The long-term consequences of this phenotypical change for the cell depend on the time required to repair the lesion inducing them and , particularly , the half-life of the mutant protein produced ( Figure 4 ) . Unexpectedly , we continued to detect active Ppluc protein ( the “mutant” form in this case ) for up to seven days after transfection with ( 8OG/Ala ) constructs . Knowing that Ppluc is not a very stable protein , as its half-life was estimated to be of no more than four hours in mammalian cells [27] , TM must therefore continue for a prolonged period of time in human cells . This finding has important implications concerning the role of TM in the etiology of diseases , particularly those affecting non-dividing cell populations , caused by the generation of mutant proteins by TM . In mammalian cells , the main pathway for the removal of 8OG from DNA involves the OGG1 glycosylase protein . It has been shown that ogg1−/− mice accumulate 8OG lesions in their DNA with aging , leading to a moderate tissue-specific increase in spontaneous mutation rate; these findings demonstrate the antimutator role of the OGG1 BER pathway [17] , [28] . The relative activity of Ppluc expressed in MEF ogg1 −/− cells transfected with ( 8OG/Ala ) constructs reflects the higher level of mutant transcript production in these cells , leading to a more pronounced phenotypical change than observed in the normal parental cells . This implies that a deficiency or decrease in the activity of this enzyme , as observed in several diseases [29] , [30] , may induce a phenotypical change in some cells of the body , contributing to the etiology of the disease . We also investigated the role of TCR in the repair of 8OG lesions in the TS in vivo . The role of this process in removing non bulky oxidative lesions , such as 8OG , from the TS is , as aforementioned , quite controversial [19]–[21] . It is thought that repair events of this type involve the blockage of RNApolII elongation by a lesion on the TS , providing a signal for the recruitment of the TCR machinery . In this regard , the role of TCR in the removal of 8OG from a TS has been investigated in several studies focusing on RNApolII interactions at sites containing this lesion [10]–[12] , [23] . These studies concluded that , both in vitro and in vivo , 8OG only weakly blocks elongation by the mammalian RNApolII . In this study , the two luciferases were generated by independent translation from the same polycistronic mRNA , with the translation of Rrluc initiated at an IRES located between the two open reading frames . A blockage of RNA polII elongation during the transcription of this mRNA would thus result in very weak Rrluc luminescence . However , the observation that cells transfected with 8OG-containing or wild-type constructs had similar levels of Rrluc protein activity strongly suggests that , in vivo , 8OG does not represent a strong block to an elongating mammalian RNApolII . Furthermore , in the five TCR-deficient human cell lines obtained from patients with Cockayne syndrome or XP/CS , the relative Ppluc activities resulting from the expression of ( 8OG/Ala ) constructs were not significantly different from those in cells with normal DNA repair capacities . This suggests that the 8OG lesion in these constructs was repaired equally efficiently in TCR-deficient and normal cells . These results represent direct evidence that TCR does not play an important role in the repair of 8OG lesions in human cells , consistent with the most recent results obtained in vitro [11] , [12] , [31] . Furthermore , Saxowsky et al . also reported that 8OG was repaired equally efficiently in murine TCR-deficient and normal cells [13] . These independent in vivo observations are clarifying a controversial area of the DNA repair field . The relative activity of Ppluc expressed in MEF ogg1 −/− cells transfected with ( 8OG/Ala ) constructs suggests that other DNA repair activities may also be involved in the repair of 8OG:C pairs in cells . The activities involved may include that of hNTH1 , as this BER N-glycosylase has been shown to cleave duplex oligonucleotides containing 8OG [32] . Alternatively , 8OG may be removed from the DNA by glycosylases of the NEIL family [33] , [34] . Accessibility to 8OG may influence the efficiency of lesion repair and , consequently , the magnitude of 8OG-mediated TM events . In our system , the reporter gene is under the control of a dose-dependent doxycycline-responsive promoter , facilitating the modulation of expression levels . The relative activity of Ppluc expressed in MRC5V1 cells cultured with low doses of doxycycline were similar to that obtained in the presence of high doses of the transcription inducer . This observation reveals that expression level and thus accessibility to the lesion does not play a major role in the modulation of TM-mediated events occurring in cells . It has frequently been reported that 8OG:T mispairs may be processed by both OGG1-mediated repair and MSH2/MSH6-dependent MMR pathways [15] , [26] . Nonetheless , the difference in affinity of the OGG1 protein and of the MSH2/MSH6 complex for an 8OG:T mispair suggests that mispairs of this type are most likely to be processed in an MSH2/MSH6-dependent manner [15] . The relative activity of Ppluc expressed in OGG1-deficient or MMR-deficient cells transfected with 8OG:T-containing constructs suggests that both pathways play a role in the repair of this type of mispair in mammalian cells . However , the efficiency of 8OG:T mispair repair depends on its location within the transcribed gene . It therefore seems likely that the recognition of this mispair , by MSH2/MSH6 or OGG1 , and the efficiency of its removal by MMR or BER may depend strongly on sequence context . We cannot rule out the possibility that the efficiency of these mechanisms also depends on transcription factors , as the efficiency of 8OG:T repair seemed to be correlated with the distance of this mispair from the transcriptional initiation sequence of the gene , creating a polarity gradient . A similar polarity gradient phenomenon has been reported during meiotic gene conversion in fungi . Indeed , the non reciprocal transfer of information from one chromatid to another during yeast meiosis often varies linearly from one end of the studied gene to the other ( for review see [35] ) . This phenomenon was shown to be initiated from promoter-containing regions of the chromosome and to be dependent upon MMR . It remains unclear whether the observed polarity gradient along a transcribed reporter gene is a general feature of DNA repair mechanisms or due exclusively to the specific sequence context at codons 5 , 297 and 445 of the Ppluc gene . It would also be interesting to use this reporter system to investigate whether the great variability of 8OG-induced TM is correlated with similar levels of variability in DNA polymerase errors during replication . The potential outcomes of TM include a number of deleterious events initiated by mutant proteins , such as cell death and changes in cellular physiology [5] , [7] . An “error catastrophe” scenario [36] , in which age-related cell death may result from the corruption of genes required for normal cellular function and viability , may result from the accumulation of TM-generated mutant proteins . Indeed , age-dependent deficiencies in the import of OGG1 into the nuclear and mitochondrial compartments results in the accumulation of oxidative lesions , such as 8OG , which may lead to an age-related increase in the production of mutant or misfolded proteins [37] . Furthermore , some neurological disorders are characterized by aggregates of misfolded and aberrant proteins associated with an increase in DNA oxidation [38] , mainly due to a decrease in hOGG1 activity in neuronal cells , resulting in the accumulation of large amounts of 8OG in the genomes [39] . These aggregates are very resistant to cellular degradation [40] and have a dominant-negative effect on cell survival . Indeed , the addition of aggregated proteins to the culture medium of human neuroblastoma cells is sufficient to induce apoptotic cell death [41] , because these aggregates act as nucleation points for the normal protein [42] . Additionally , it has also recently been suggested that hypomorphic alleles of hOGG1 are associated with Alzheimer's disease cases and that defects in OGG1 may play an important role in the disease in a significant number of AD patients [43] . Thus , as depicted in Figure 4 , aberrant proteins with a dominant-negative effect produced through TM-related events may play an important role in the pathogenesis of neurodegenerative diseases , such as Alzheimer's disease and Parkinson's disease .
The cell lines used in this study are described in Table 1 . They were cultured at 37°C , under an atmosphere containing 5% CO2 , in minimal essential medium ( MEM ) supplemented with 10% fetal calf serum , 2 mM L-glutamine , 0 . 3% amphotericin B ( Fungizone ) , 100 IU penicillin and 100 µg/ml streptomycin . XPCS2BA and XPCS1LV diploid fibroblasts were transformed , in our laboratory , with the pLAS-wt-plasmid carrying the TAg of SV40 [44] . The pBDA6 luciferase reporter vector ( Figure S1 ) contains the Photinus pyralis ( firefly ) and the Renilla reniformis ( sea pansy ) luciferase genes ( Ppluc and Rrluc , respectively ) organized in a bicistronic operon . Rrluc gene translation is initiated from the IRES located between the Ppluc and Rrluc open reading frames , and both luciferase proteins are thus translated from the same mRNA . Transcription to generate this polycistronic mRNA is initiated from the dose-dependent doxycycline-responsive Ptight promoter . When cells are transfected with this plasmid , the presence of doxycycline in the culture medium allows the transcriptional activator ( rtTA ) to bind the Ptight promoter , leading to the production of both luciferases . This plasmid also contains the ampicillin resistance gene ( AmpR ) and an origin of double-strand DNA replication ( ColE1 ) , allowing its propagation in bacterial cells . The production of circular single-stranded DNA corresponding to the coding strand of the Ppluc gene is initiated from the f1 origin of replication ( f1 ori ) . The other elements of this plasmid are two SV40-polyadenylation sites ( SV40pA ) and an intervening sequence ( IVS ) directing the correct processing and stabilization of the mRNA in mammalian cells , and a PCMV promoter for expression of the rtTA gene . This vector is deprived of mammalian origin of replication , to prevent artifacts generated by mutagenic replication of the 8OG-containing constructs . The pBDA6 vector was constructed in several steps ( Figure S2 ) . First , nucleotides 6060 to 2614 from pIRES ( Clontech Laboratories ) and nucleotides 79 to 4360 from pTet-On ( Clontech Laboratories ) were amplified by PCR , using the Pfu Turbo DNA polymerase ( Stratagene ) . The oligonucleotide primers used for these reactions were designed to create AgeI and PacI sites at either end of the amplified fragments . The two PCR products were then digested with AgeI and PacI ( New England Biolabs ) and ligated together , using T4 DNA ligase ( Roche ) , to generate the pBDA5/1 plasmid . Nucleotides 4336 to 5084 of the pBDA5/1 plasmid were replaced by nucleotides 2590 to 343 from pTRE-Tight , resulting in the replacement of the PCMV promoter by the PTight dose-dependent doxycycline-responsive promoter ( Clontech Laboratories ) and generation of the pBDA5/2 plasmid . The pBDA6 final construct was obtained by amplifying the Ppluc and Rrluc genes from pBI-Luc ( Clontech Laboratories ) and pRL-CMV ( Promega ) , respectively . The Ppluc fragment was inserted upstream from the IRES , between the NsiI and NheI sites , whereas the Rrluc fragment was inserted into the NotI site downstream from the IRES . All the variants ( pBDA6-luc K5X , K5Q , K297X , K297Q , E344A , D422 , K445X and K445Q ) , differing from the original by a single point mutation in the Ppluc gene , were obtained by directed mutagenesis , through the PCR of overlapping extensions technique [45] . The fragments generated were then digested with NheI and NsiI and inserted into pBDA6 digested with the same enzymes . All PCR amplifications were performed using the Pfu Turbo DNA polymerase ( Stratagene ) . The name of the pBDA6 indicates the change of amino-acid sequence of the Ppluc protein at the specified codon . The Rrluc and Ppluc genes of all the plasmids were sequenced by Genome Express ( Meylan , France ) . All plasmid constructs were introduced into the DH12S strain of Escherichia coli . Bacteria were grown in LB supplemented with ampicillin ( 100 µg/ml ) ( Sigma ) . We produced eighteen constructs: ( LFS/Lys ) 5 , ( 8OG/Stop ) 5 , ( LFS/Stop ) 5 , ( LFS/Gln ) 5 , ( LFS/Lys ) 297 , ( 8OG/Stop ) 297 , ( LFS/Stop ) 297 , ( LFS/Gln ) 297 , ( LFS/Glu ) 344 , ( 8OG/Ala ) 344 , ( LFS/Ala ) 344 , ( LFS/Asp ) 422 , ( 8OG/Ala ) 422 , ( LFS/Ala ) 422 , ( LFS/Lys ) 445 , ( 8OG/Stop ) 445 , ( LFS/Stop ) 445 and ( LFS/Gln ) 445 . The first part of the name of each construct indicates the strand transcribed: lesion-free strand ( LFS ) or 8OG-containing strand ( 8OG ) . The second part of the name indicates the amino acid specified by the non-transcribed strand and the single strand DNA of the pBDA6 variant used as a template for DNA synthesis . The index number corresponds to the codon number in the Ppluc gene . Single-stranded DNA was prepared , and templates constructed , as previously described [46] , [47] . The primers used to initiate DNA polymerization reactions for the template construction are listed in Table S1 . Cells were transfected with constructs by nucleofection methods , using the NHDF nucleofector kit ( Amaxa ) . Cells were first treated with trypsin and washed twice in 1×PBS ( Gibco ) . For each transfection , 300 , 000 cells ( or 1 , 000 , 0000 cells for normal MEF and LoVo cells ) were resuspended in 100 µl of NHDF solution and mixed with 300 ng ( or 1 µg for normal MEF cells ) of the construct concerned . The mixture was then subjected to electroporation program U23 of the Amaxa nucleofector device . Immediately after the electric shock , cells were resuspended in 3 ml of MEM ( Gibco ) supplemented with 10% fetal calf serum ( Gibco ) , 2 mM L-glutamine ( Gibco ) and 2 µg/ml ( or 1 ng/ml when specified ) doxycycline ( Sigma ) and placed in 6-well plates in an incubator maintained at 37°C , under an atmosphere containing 5% CO2 . The medium was removed from each well 24 hours after transfection , and cells were washed twice with cold 1×PBS . Cells were lysed by incubation for 45 minutes in 500 µl of Passive Lysis Buffer ( Promega ) , placed at −20°C for 30 minutes and then thawed to room temperature . Luciferase activity was measured with the Dual-Luciferase Reporter Assay System ( Promega ) , using 80 µl of “Luciferase Assay Reagent” , 80 µl of lysed cells and 80 µl of “Stop and Glo” reagent . Luminescence , in relative light units ( RLU ) , was determined over a 10-second period , in a Femtomaster FB12 luminometer ( Zylux Corp . ) . Ppluc activity was normalized with respect to Rrluc activity for each transfection , using the following formula: ( RLUPp/RLURr ) . For each set of transfections with the same cell line , the relative Ppluc activity of cells transfected with 8OG-containing ( or LFS/Stop or LFS/Ala ) constructs was calculated as follows: [ ( RLUPp/RLURr ) construct/ ( RLUPp/RLURr ) 100%]×100 with the 100% being the normalized Ppluc activity in cells transfected with the corresponding wild-type construct ( LFS/Lys for codon 4 , 297 and 445 , LFS/Glu for codon 344 and LFS/Asp for codon 422 ) . RNA was extracted from MRC5V1 cells 24 hours after transfection with the ( 8OG/Stop ) 445 construct . RNA was extracted with Tri-Reagent solution ( Sigma ) , according to the manufacturer's instructions . Contaminating DNA was eliminated from the RNA solution by two treatments with the DNA-free kit ( Ambion ) . We then used about 50 ng of RNA for RT-PCR with Superscript II ( Invitrogen ) as a reverse transcriptase and Taq DNA polymerase for amplification ( New England Biolabs ) , using LBRT1 and LBRT2 as primers [9] . For each RNA preparation , the absence of DNA contamination was checked by amplification reactions in the same conditions but with the omission of the reverse transcriptase . The cDNA was subcloned by ligating a Sau3A/HincII fragment of the RT-PCR product between the BamHI and HincII sites of pUC18 ( all restriction enzymes were from New England Biolabs ) . Subclones were then amplified with Clo18L and Clo18U [9] . The DNA amplified from the subclones was sequenced by Genome Express ( Meylan , France ) . | The DNA molecule is used as a template for duplication , to transmit genetic information to the progeny of a given cell , but also as a template for the transcription machinery . This machinery converts genetic information from the DNA form to the RNA form used for protein synthesis . Chemical alterations of the DNA molecule caused by endogenous or environmental stresses are responsible for the generation of mutations . Indeed , these lesions can induce replication errors when DNA is duplicated during cell division . These mutations have been shown to be responsible for many genetic diseases and other sporadic diseases , such as cancer . However , less is known about their effects on transcription . We report here that a specific DNA lesion may lead to erroneous transcription events , ultimately leading to the production of aberrant proteins . The magnitude of these errors seems to depend largely on the DNA sequences surrounding the lesion and the capacity of the cell to repair this lesion . We also show that the production of aberrant protein from the erroneous transcription products may affect the phenotype of the cells concerned . Lesion-induced transcription errors may also play a role in the development of neurodegenerative diseases , such as Alzheimer's and Parkinson's diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/translation",
"mechanisms",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"molecular",
"biology/dna",
"repair"
] | 2009 | Transcriptional Mutagenesis Induced by 8-Oxoguanine in Mammalian Cells |
Understanding tumor invasion and metastasis is of crucial importance for both fundamental cancer research and clinical practice . In vitro experiments have established that the invasive growth of malignant tumors is characterized by the dendritic invasive branches composed of chains of tumor cells emanating from the primary tumor mass . The preponderance of previous tumor simulations focused on non-invasive ( or proliferative ) growth . The formation of the invasive cell chains and their interactions with the primary tumor mass and host microenvironment are not well understood . Here , we present a novel cellular automaton ( CA ) model that enables one to efficiently simulate invasive tumor growth in a heterogeneous host microenvironment . By taking into account a variety of microscopic-scale tumor-host interactions , including the short-range mechanical interactions between tumor cells and tumor stroma , degradation of the extracellular matrix by the invasive cells and oxygen/nutrient gradient driven cell motions , our CA model predicts a rich spectrum of growth dynamics and emergent behaviors of invasive tumors . Besides robustly reproducing the salient features of dendritic invasive growth , such as least-resistance paths of cells and intrabranch homotype attraction , we also predict nontrivial coupling between the growth dynamics of the primary tumor mass and the invasive cells . In addition , we show that the properties of the host microenvironment can significantly affect tumor morphology and growth dynamics , emphasizing the importance of understanding the tumor-host interaction . The capability of our CA model suggests that sophisticated in silico tools could eventually be utilized in clinical situations to predict neoplastic progression and propose individualized optimal treatment strategies .
Cancer is not a single disease , but rather a highly complex and heterogeneous set of diseases that can adapt in an opportunistic manner , even under a variety of stresses . It is now well accepted that genome level changes in cells , resulting in the gain of function of oncoproteins or the loss of function of tumor suppressor proteins , initiate the transformation of normal cells into malignant ones and neoplastic progression [1] , [2] . In the most aggressive form , malignant cells can leave the primary tumor , invade into surrounding tissues , find their way into the circulatory system ( through vascular network ) and be deposited at certain organs in the body , leading to the development of secondary tumors ( i . e . , metastases ) [3] . The emergence of invasive behavior in cancer is fatal . For example , the malignant cells that invade into the surrounding host tissues can quickly adapt to various environmental stresses and develop resistance to therapies . The invasive cells that are left behind after resection are responsible for tumor recurrence and thus an ultimately fatal outcome . Therefore , significant effort has been expended to understand the mechanisms evolved in the invasive growth of malignant tumors [2] , [4]–[7] and their treatment [8] , [9] . It is generally accepted that the invasive behavior of cancer is the outcome of many complex interactions occurring between the tumor cells , and between a tumor and the host microenvironment [3] . Tumor invasion itself is a complex multistep process involving homotype detachment , enzymatic matrix degradation , integrin-mediated heterotype adhesion , as well as active , directed and random motility [4] . In recent in vitro experiments involving glioblastoma multiforme ( GBM ) , the most malignant brain cancer , it has been observed that dendritic invading branches composed of chains of tumor cells are emanating from the primary tumor mass; see Figure 1 . Such invasive behaviors are characterized by intrabranch homotype attraction and least-resistance paths of cells [4] . Although recent progress has been made in understanding certain aspects of the complex tumor-host interactions that may be responsible for invasive cancer behaviors [4] , [10]–[12] , many mechanisms are either not fully understood or are unknown at the moment . Even if all of the mechanisms for cancer invasion could be identified , it is still not clear that progress in understanding neoplastic progression and proposing individualized optimal treatment strategies could be made without the knowledge of how these different mechanisms couple to one another and to the heterogeneous host microenvironment in which tumor grows [13] . Theoretical/computational cancer modeling that integrates distinctly different mechanisms for tumorigenesis , when appropriately linked with experimental and clinical data , offers a promising avenue for a better understanding of tumor growth , invasion and metastasis . A successful model would enable one to broaden the conclusions drawn from existing medical data , suggest new experiments , test hypotheses , predict behavior in experimentally unobservable situations , and be employed for early detection and prognosis [13] . Indeed , cancer modeling has been a very active area of research for the last two decades ( see Refs . [13] and [14] for recent reviews ) . A variety of interactions between the tumor and its host microenvironment have been investigated [15]–[32] via continuum [25]–[28] , [32] , discrete [16] , [20] , [33] or hybrid [19] , [21]–[23] mathematical models . Very recently , multiscale mathematical models [22] , [23] , [28] have been employed to study the effects of the host microenvironment on the morphology and phenotypic evolution of invasive tumors and it has been shown that microenvironmental heterogeneity can dramatically affect the growth dynamics of invasive tumors . Although these simulated tumors predicted certain invasive characteristics ( e . g . , development of protruding surfaces ) , no dendritic invasive branches emerged from these numerical studies . In response to the challenge to develop an “Ising” model for cancer growth [13] , we generalize here a cell-based discrete cellular automaton ( CA ) model that we have developed [16]–[18] , [20] , [21] to investigate the invasive growth of malignant tumors in heterogeneous host microenvironments . To the best of our knowledge , this generalized CA model is the first to investigate the formation of invasive cell chains and their interactions with the primary tumor mass and host microenvironment . Our generalized cellular automaton model takes into account a variety of microscopic-scale tumor-host interactions , including the short-range mechanical interactions between tumor cells and tumor stroma , the degradation of extracellular matrix by the invasive cells and oxygen/nutrient gradient driven cell motions and thus , it can predict a wide range of growth dynamics and emergent behaviors of invasive tumors . In particular , our CA model robustly reproduces the salient features of dendritic invasive growth observed in experiments , which is characterized by least-resistance paths of cells and intrabranch homotype attraction . The model also predicts nontrivial coupling between the growth dynamics of the primary tumor mass and the invasive cells , e . g . , the invasive cells can facilitate the growth of primary tumor in harsh microenvironment . Moreover , we show that the properties of the host microenvironment can significantly affect tumor growth dynamics and lead to a variety of tumor morphologies . These emergent behaviors naturally arise due to various microscopic-scale tumor-host interactions , which emphasizes the importance of taking into account microenvironmental heterogeneity in understanding cancer . Further refinement of our model could eventually lead to the development of a powerful in silico tool that could be utilized in the clinic . As a demonstration of the capability and versatility of our CA model , we mainly consider invasive tumor growth in two dimensions , although the model is easily extended to three dimensions . Indeed , the algorithmic details of the model are given for any spatial dimension .
We now provide specific details for the CA model to study invasive tumor growth in confined heterogeneous microenvironment . In what follows , we will simply refer to the primary tumor as “the tumor” and explicitly use “invasive” when considering invasive cells . After generating the automaton cells by Voronoi tessellation of RSA sphere centers , an ECM macromolecule density is assigned to each automaton cell within the growth-permitting region , which represents the heterogeneous host microenvironment . Then a tumor is introduced by designating any one or more of the automaton cells as proliferative cancer cells . Time is then discretized into units that represent one real day . At each time step: The aforementioned automaton rules are briefly illustrated in Figure 3 . We note that non-invasive tumor growth can be studied by imposing a mutation rate . This enables us to compare the growth dynamics of invasive and non-invasive tumors and in turn to investigate the effects of the coupling between the growth dynamics of the primary tumor mass and the invasive cells . Although we only consider spherical growth-permitting regions here , the CA rules given above allow growth-permitting regions with arbitrary shapes . The important parameters mentioned in the bullet points above are summarized in Table 1 . In the following , we will employ our CA model to investigate the growth dynamics of malignant tumors with different degrees of invasiveness in a variety of different heterogeneous microenvironments . To characterize quantitatively the morphology of simulated tumors , we present several scalar metrics that capture the salient geometric features of the primary tumor , dendritic invasive branches or the entire invasive pattern . These metrics include the ratio of the invasive area over the tumor area ( defined below ) , the specific surface of the invasive pattern , the asphericity of the primary tumor and the angular anisotropy metric for the invasive branches . The metrics are computed for all simulated tumors and compared to available experimental data . We note that the invasive pattern associated with a neoplasm includes both the primary tumor and the invasive branches . Following Ref . [4] , the tumor area is defined as the area of the circumcircle of the primary tumor ( see Figure 4 ( a ) ) and the invasive area is the area of the region between the effective circumcircle of the invasive pattern and the circumcircle of the primary tumor ( see Figure 4 ( a ) ) . The radius of the effective circumcircle of the invasive pattern is defined to be the average distance from the invasive branch tip to the tumor center . The ratio as a function of time reflects the degree of coupling between the primary tumor and the invasive cells . If is linear in , there is no coupling; otherwise the two are coupled . The specific surface [34] for the invasive pattern is defined as the ratio of the total length of the perimeter of the invasive pattern over its total area . In general , is inversely proportional to the size of the tumor and thus , large tumors have small values . Moreover , given the tumor size , tumors with a large number of long dendritic invasive branches possess a large value of . And is minimized for perfectly circular tumors with , where is the radius . Since depends on the size of the tumor , which makes it difficult to compare tumors with different sizes , in the calculations that follow we employ a normalized with respect to for an arbitrary-shaped tumor with effective radius ( i . e . , the average distance from tumor edge to tumor center ) . For simplicity , we will still refer to the normalized specific surface as “specific surface” and designate it with symbol . The asphericity of the primary tumor is defined as the ratio of the radius of circumcircle of the primary tumor over its incircle radius [45] , i . e . , ( see Figure 4 ( b ) ) . A large value indicates a large deviation of the shape of primary tumor from that of a perfect circle , i . e . , the tumor is more anisotropic . To quantify the degree of anisotropy of the invasive branches , we introduce the angular anisotropy metric . In particular , the entire invasive pattern is evenly divided into sectors with lines emanating from the tumor center ( see Figure 4 ( c ) ) . The angular anisotropy metric is defined as ( 1 ) where is the average length of the invasive branches within the th sector and ( 2 ) is the average length of all invasive branches . For tumors with invasive branches of similar lengths that are uniformly angularly distributed , the metric is small . Large fluctuations of both invasive branch length and angular distribution can lead to large values . In the following , we use to compute for the simulated invasive tumors .
To verify the robustness and predictive capacity of our CA model , we first employ it to reproduce quantitatively the observed invasive growth of a GBM multicellular tumor spheroid ( MTS ) in vitro [4] . In particular , the boundary of the growth-permitting region is considered to be vascularized , i . e . , a growing tumor can receive oxygen and nutrients from the growth-permitting region . A constant radially symmetric nutrient/oxygen gradient in the growth-permitting region with the highest nutrient/oxygen concentration at the vascular boundary is used . Initially , approximately 250 proliferative tumor cells are introduced at the center of the growth-permitting region with homogeneous ECM and tumor growth is started . This corresponds to an initial MTS with diameter which is consistent with the in vitro experiment set-up [4] . The following values of the growth and invasiveness parameters are used: , , , , , . Note that the value of corresponds to a cell doubling time of 40 hours , which is consistent with the reported experimental data [4] . A small value of the ECM density is used , which corresponds to the soft DMEM medium used in the experiment [4] . In the visualizations of the tumor that follow , we use the following convention: the ECM in the growth-permitting region is white , and gray outside this region . The ECM degraded by the tumor cells is blue . In the primary tumor , necrotic cells are black , quiescent cells are yellow and proliferative cells are red . The invasive tumor cells are green . Figure 5 ( a ) and ( b ) respectively show the morphology of simulated MTS and a magnification of its invasive branches with increasing branch width towards the proliferative core . Specifically , one can clearly see that within the branches , chains of cells are formed as observed in experiments [4] ( see Figure 1 ) . The invasive cells tend to follow one another ( which is termed “homotype attraction” ) since paths of degraded ECM are formed by pioneering invasive cells and it is easier for other cells to follow and enhance such paths than degrading ECM to create new paths by themselves . In other words , invasive cells tend to take paths with “least resistance” . We note that no CA rules are imposed to force such cellular behaviors . Instead , they are emergent properties that arise in our simulations . The ratio of the invasion area over the primary tumor area as a function of time for the simulated tumor is computed and compared to the reported experimental data [4] ( see Figure 5 ( c ) ) . One can clearly see that our simulation results agree with experimental data very well . Moreover , the deviation of from a linear function of indicates that the growth of primary tumor and the invasive branches are strongly coupled [4] . Other metrics for tumor morphology such as the specific surface of the invasive pattern , the sphericity of the primary tumor and the angular anisotropy metric for the invasive branches are computed from our simulation results and from the image of invasive MTS in Figure 1 ( a ) at 24 hours after initialization . The values are given in Table 2 , from which one can see again a good agreement . Thus , we have shown that our CA model is both robust and quantitatively accurate with properly selected parameters . Having verified the robustness and predictive capacity of our CA model , we now consider three types of distributions of the ECM density , i . e . , homogeneous , random and sinusoidal-like , to systematically study the effects of microenvironment heterogeneity on invasive tumor growth ( see Figure 6 ) . These ECM density distributions represent real host microenvironments in which a tumor grows . ( Details about these ECM distributions are given in the following sections . ) Again , the boundary of the growth-permitting region is considered to be vascularized with a constant radially symmetric nutrient/oxygen gradient in the growth-permitting region pointing to the tumor center . We note that although generally the nutrient/oxygen concentration field in vivo is more complicated than exhibited here , previous numerical studies that considered the exact evolution of nutrient/oxygen concentrations have shown a decay of the concentrations toward the tumor center [22] , [23] . Since the directions of cell motions are determined by the nutrient/oxygen gradient , our constant-gradient approximation is a very reasonable one . In the beginning of the simulation , a proliferative tumor cell is introduced at the center of the growth-permitting region and tumor growth is initiated . The growth parameters for the primary tumors in all cases studied here are the same and are given in Table 1 . The invasiveness parameters and ECM densities are variables and specified in each case separately . The values of the growth parameters for the CA model were chosen to be consistent with GBM data from the medical literature [16] , [20] . Specifically , the value of the base probability of division is , which corresponds to a cell doubling time of 4 days [46] , [47] . This value is used for all of the cases of invasive growth that follow . Since our CA model takes into account general microscopic tumor-host interactions , we expect that the general growth dynamics and emergent behaviors predicted by the model will qualitatively apply to other solid tumors . We note that all of the reported growth dynamics and emergent properties of the simulated tumors for any specific set of growth and invasiveness parameters are repeatedly observed in 25 independent simulations .
We have developed a novel cellular automaton ( CA ) model which , with just a few parameters , can produce a rich spectrum of growth dynamics for invasive tumors in heterogeneous host microenvironment . Besides robustly reproducing the salient features of branched invasive growth , such as least-resistance paths of cells and intrabranch homotype attraction observed in in vitro experiments , our model also enables us to systematically investigate the effects of microenvironment heterogeneity on tumor growth as well as the coupling between the growth dynamics of the primary tumor and the invasive cells . In particular , we have shown that in homogeneous ECM with low densities ( i . e . , soft microenvironment ) , both the shape of the primary tumor and invasive pattern are isotropic . For high cellular motility cases , the invasive cells form extended dendritic invasive branches; while for low cellular motility cases , the invasive cells clump near the primary tumor surface and form a bumpy concentric-like shell that facilitates the growth of the primary tumor . Tumors growing in a highly rigid homogeneous ECM can develop anisotropic shapes , facilitated by the invasive cells that degrade the ECM; both the tumor size and the extent of invasive branches are much smaller . In heterogeneous ECM , both the primary tumor and invasive pattern are significantly affected during the early growth stages , i . e . , anisotropic shapes and patterns are developed to avoid high density/rigid regions of the ECM . If the characteristic length scale of the heterogeneities is comparable to the macroscopic tumor size , such effects can persist in later growth stages . In addition , invasive cells with large motility can significantly diminish the anisotropy effects by their ECM degradation activities . We emphasize that we did not manipulate the behavior of cells by imposing artificial CA rules to give rise to these complex and rich growth dynamics . Instead , these are emergent behaviors that naturally arise due to various microscopic-scale tumor-host interactions that are incorporated into our CA model , including the short-range mechanical interaction between the tumor cells and tumor stroma , and the degradation of extracellular matrix by the invasive cells . It is noteworthy that the growth dynamics of tumors in a heterogeneous microenvironment is distinctly different than those in a homogeneous microenvironment . This emphasizes the importance of understanding the effects of physical heterogeneity of the host microenvironment in modeling tumor growth . Here we just make a first attempt to take into account a simple level of host heterogeneity , i . e . , by considering the ECM with variable density/rigidity . Currently , the invasion of the malignant cells into the host microenivronment is considered to be a consequence of invasive cell phenotype gained by mutation , and is not triggered by environmental stresses . However , the effects of environmental stresses can be taken into account . For example , a CA rule can be imposed that if the division probability of a malignant cell is significantly reduced by ECM rigidity , i . e . , it is extremely difficult to push away/degrade ECM to make room for daughter cells , the malignant cell leaves the primary tumor and invades into soft regions of surrounding ECM . This would lead to reduced tumor invasion ( i . e . , development of the dendritic invasive branches ) in soft microenvironments but enhanced invasion in rigid microenvironments [48] . Indeed , we have very recently generalized the CA model reported here to explicitly take into account the pressure exerted on the tumor due to the deformation of its surrounding ECM as well as the local geometry of the tumor-host interface to study mechanical-stress induced tumor morphology instability [49] . Moreover , the spatial-temporal evolution of more complicated and realistic nutrient/oxygen fields can be incorporated into our CA model . This can be achieved by solving the coupled nonlinear partial differential equations governing the evolution of the nutrient/oxygen concentrations as was done in Refs . [19] and [21] . Since the CA rules are given for any spatial dimension , our model is readily generalized to three dimensions . In addition , the model can be modified to incorporate other host heterogeneities , such as stromal cells , blood vessels and the shape anisotropy of the host organ [20] , [21] . As currently implemented , a single 2D simulation takes less than 0 . 5 hours on a 32-bit 1 . 56 Gb Memory 1 . 44 GHz dual core Dell Workstation . We expect that a 3D simulation will take no longer than 24 hours on a supercomputer when a proper parallel implementation is used . Such an in silico tool not only enables one to investigate tumor growth in complex heterogeneous microenvironment that closely represents the real host microenvironments but also allows one to infer and even reconstruct individual host microenvironment given limited growth data of tumors ( such as shape and size at various times ) . Such microstructural information of the individual host would be extremely valuable for developing individualized treatment strategies . For example , based on the host microstructure one can design special encapsulation and transport agents that maximize drug delivery efficiency [13] . In our current CA model , the microscopic parameters governing tumor invasion are variable and can be arbitrarily chosen within a feasible range as given in Table 1 . Given sufficient and reliable experimental data of invasive tumor growth , the parameters in our CA model could be uniquely determined and thus , the model could produce robust predictions about neoplastic progression . Although the current CA model is specifically implemented to reproduce and predict the growth dynamics of invasive solid tumors in vitro , further refinement of the model could eventually lead to the development of a powerful simulation tool that could some day be utilized clinically . For example , more complicated and realistic host heterogeneities such as the vascular structure , various stromal cells , the corresponding spatial-temporal evolution of the nutrient/oxygen concentrations as well as environmental stress-induced mutations should be incorporated as we described earlier . If the robustness of the refined model could be validated clinically , we would expect it to produce quantitative predictions for in vivo tumor growth , which could potentially be valuable for tumor prognosis and proposing individualized treatment strategies . | The goal of the present work is to develop an efficient single-cell based cellular automaton ( CA ) model that enables one to investigate the growth dynamics and morphology of invasive solid tumors . Recent experiments have shown that highly malignant tumors develop dendritic branches composed of tumor cells that follow each other , which massively invade into the host microenvironment and ultimately lead to cancer metastasis . Previous theoretical/computational cancer modeling neither addressed the question of how such chain-like invasive branches form nor how they interact with the host microenvironment and the primary tumor . Our CA model , which incorporates a variety of microscopic-scale tumor-host interactions ( e . g . , the mechanical interactions between tumor cells and tumor stroma , degradation of the extracellular matrix by the tumor cells and oxygen/nutrient gradient driven cell motions ) , can robustly reproduce experimentally observed invasive tumor evolution and predict a wide spectrum of invasive tumor growth dynamics and emergent behaviors in various different heterogeneous environments . Further refinement of our CA model could eventually lead to the development of a powerful simulation tool for clinical purposes capable of predicting neoplastic progression and suggesting individualized optimal treatment strategies . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"physics",
"mathematics",
"materials",
"science",
"biology"
] | 2011 | Emergent Behaviors from a Cellular Automaton Model for Invasive Tumor Growth in Heterogeneous Microenvironments |
Interactions among genes and the environment are a common source of phenotypic variation . To characterize the interplay between genetics and the environment at single nucleotide resolution , we quantified the genetic and environmental interactions of four quantitative trait nucleotides ( QTN ) that govern yeast sporulation efficiency . We first constructed a panel of strains that together carry all 32 possible combinations of the 4 QTN genotypes in 2 distinct genetic backgrounds . We then measured the sporulation efficiencies of these 32 strains across 8 controlled environments . This dataset shows that variation in sporulation efficiency is shaped largely by genetic and environmental interactions . We find clear examples of QTN:environment , QTN: background , and environment:background interactions . However , we find no QTN:QTN interactions that occur consistently across the entire dataset . Instead , interactions between QTN only occur under specific combinations of environment and genetic background . Thus , what might appear to be a QTN:QTN interaction in one background and environment becomes a more complex QTN:QTN:environment:background interaction when we consider the entire dataset as a whole . As a result , the phenotypic impact of a set of QTN alleles cannot be predicted from genotype alone . Our results instead demonstrate that the effects of QTN and their interactions are inextricably linked both to genetic background and to environmental variation .
As we identify more genetic loci that underlie complex traits , the challenge remains to understand and predict the effects of the causal genetic variants upon individuals' phenotypes . The relationship between genotype and phenotype is rarely simple . The effect of an allele often depends upon the environment , resulting in gene-environment interactions ( GxE ) . GxE is a well-documented occurrence in many species , including humans [1]–[5] . Gene-gene interactions also take place that render the effect of one locus dependent upon the genotype at another locus . Genetic interactions can occur between characterized loci ( epistasis ) [6] , [7] , or between one known locus and other unknown loci ( genetic background effects ) [8] . If individuals vary in their environmental exposure and genetic makeup , as they almost always do in nature , then GxE and genetic interactions will create differences in the effects of alleles among individuals . Therefore , to understand allelic effects , we must also understand the scope and prevalence of genetic and environmental interactions . However , standard approaches for the identification of causative loci , such as association analysis and linkage mapping [9] , measure the average effects of alleles in populations . Without very large sample sizes , population averages cannot account for potential individual-to-individual variation created by complex interactions [10] . Some study of interactions on an individual-to-individual basis has occurred through the use of near isogenic lines [11] , but there are still few examples that illustrate the impact of interactions from one individual to the next at the resolution of single-nucleotides [7] , [12] . To better understand the effects of GxE and genetic interactions at the resolution of single nucleotides , we took advantage of four naturally occurring quantitative trait nucleotides ( QTN ) known to cause variation in yeast sporulation efficiency [13] , [14] . We engineered allele replacement strains that carry all possible combinations of these QTN in two genetic backgrounds , and we then systematically measured the phenotypes of these strains in eight environments . Our results provide a detailed picture of how segregating QTN , environmental variation , and genetic background all combine to shape variation in a quantitative trait through complex relationships .
Our phenotype of interest , yeast sporulation , is a cell fate decision executed by diploid yeast cells in response to a shift from fermentative to respiratory conditions [15] . Yeast cells switch to primarily aerobic respiration when faced with only a non-fermentable carbon source . When this environmental change is accompanied by a reduction in a critical nutrient such as nitrogen , a fraction of yeast cells in a culture will initiate meiosis and enclose the meiotic products in a protective spore wall . Our QTN all affect the proportion of cells in a culture that initiate meiosis ( the sporulation efficiency ) after a shift from glucose ( fermentable ) to acetate ( non-fermentable ) media . The QTN include a coding polymorphism in RSF1 ( a positive regulator of respiration ) [16] , both coding and non-coding polymorphisms in IME1 ( the master regulator of sporulation ) [17] , and a non-coding polymorphism in RME1 ( a direct repressor of IME1 ) [18] , [19] . Each of these genes encodes a transcription factor . Each QTN has two alleles: a reference allele found in the wild oak tree isolate YPS606 , and an allele that reduces sporulation efficiency from the vineyard isolate UCD2120 [14] . ( In the rest of this article , we denote the QTN with the labels: rsf1 , rme1 , ime_coding , and ime_nc . ) Both the patterns of phenotypic variation in sporulation efficiency and the sequence variation of the causal genes indicate that sporulation efficiency is subject to purifying selection in oak strains and disruptive selection in vineyard strains [14] , [20] . The change in phenotype caused by the QTN therefore represents genotype-phenotype variation that has occurred due to a shift in selective pressures between two habitats . To broadly test for GxE effects , we first measured the sporulation efficiency of each parent strain genetic background ( designated oak and vineyard ) carrying two QTN genotype combinations: either all the QTN alleles of the oak parent , or all the QTN alleles of the vineyard parent . We generated environmental variation by growing the strains in eight different fermentative media conditions ( Table 1 ) prior to the induction of sporulation in acetate ( see Methods ) . In all eight environments and across both genetic backgrounds , the oak QTN alleles collectively increase sporulation efficiency , and the vineyard QTN alleles collectively decrease sporulation efficiency ( Table 2 ) . However , the environments vary with respect to the proportion of the phenotype that the QTN explain ( Table 3 ) . For example , in grape juice , we can explain 99% of the phenotypic difference between the parents by placing the vineyard QTN alleles into the oak background . However , in raffinose , the same allele replacement explains only 55% of the parental difference . The phenotypic difference explained by the QTN also depends on the genetic background . For example , placing the oak QTN alleles into the vineyard background explains 90% of the difference between the parent strains in raffinose , but we explain only 55% of the difference between the parents if we conduct the reciprocal experiment that places the vineyard QTN alleles into the oak background . Because the phenotypic difference created by the QTN varies across both the environments and genetic backgrounds , our results imply genetic interactions among the QTN , the environmental treatments , and uncharacterized loci in the two parent genetic backgrounds . To further investigate the extent of these interactions , we measured the phenotypes of strains with all 16 possible QTN genotype combinations in both genetic backgrounds ( 32 total strains ) . We calculated a correlation matrix of the eight environments from their effects on the phenotype rank-order of the 32 strains so that we can broadly compare the QTN effects across environmental treatments and genetic backgrounds ( Table 4 ) . Because the QTN alleles always act in the same direction regardless of condition , all the environments were positively correlated ( Spearman's ρ = 0 . 69 to 0 . 99 ) . The differences in correlations therefore reflect changes in the rank order ( and therefore relative magnitude ) of QTN effects . We used hierarchical clustering to construct a dendrogram that reflects the correlations between environments ( Figure 1 ) . Sucrose , fructose , and glucose are the most highly correlated environments ( Spearman's ρ>0 . 99 for all pair wise comparisons ) and cluster closely . We did not detect significant differences between these three environments in either genetic background . Their values were therefore pooled and averaged as “glucose-like” ( YGlu ) for all subsequent analyses . Maltose and raffinose cluster separately from YGlu and are slightly less correlated with glucose ( ρ = 0 . 93 , 0 . 96 , respectively ) . Both the oak and vineyard genetic backgrounds sporulate more efficiently in raffinose than in YGlu . ( Figure 2A ) The effect of maltose , however , depends upon the genetic background ( Figure 2B ) . Sporulation efficiency of the vineyard background is similar in maltose and YGlu , but the oak background sporulates more efficiently in maltose . Therefore , there is an interaction between the genetic background and maltose . Galactose also shows a background:environment interaction . The oak background sporulates similarly in galactose and YGlu , but the vineyard background sporulates more efficiently in galactose ( Figure 2C ) . In Figure 1 , galactose clusters distinctly from all other environments . When we run the clustering algorithm separately for each genetic background , this separation disappears ( Figures S1 , S2 ) . Therefore , the disparity of galactose relative to the other environments appears to result from the background:environment interaction . Synthetic oak exudate and grape juice also cluster distinctly from the other environments , but these two conditions are highly correlated with each other ( ρ = 0 . 97 ) . The sporulation efficiency of both genetic backgrounds tends to be lower in exudate and grape juice than in the other environments ( Figure 1 , Tables S1 , S2 ) . There are also QTN:environment interactions that occur in both exudate and grape juice relative to YGlu . For example , in the oak background , the rsf1 vineyard allele has a much larger effect in exudate and grape juice than it does in YGlu ( Figure 3 ) . To quantify this difference in QTN effect , we constructed a linear model of sporulation efficiency in the oak background that incorporates the main effects of single QTN , the effects of the environments , and the QTN:environment interactions ( see Methods ) . In this model , the differential effect of rsf1 is manifested as a QTN:environment interaction in exudate and grape juice relative to YGlu ( exudate: effect = −34±2% , t-test , P<2e-16; grape juice: effect = −25±2% , t-test P<2e-16; all errors reported in the text are the standard errors of coefficient estimates ) . The effect of rsf1 is the largest of any single QTN in both grape juice and exudate ( Tukey's HSD , maximum adjusted P = 0 . 002 ) . However , in YGlu the rsf1 QTN does not even have a significant main effect in the oak background ( effect = −1 . 8±1% , t-test P = 0 . 08 ) . The effect of rsf1 in the vineyard background reveals a different story . In the vineyard background , the rsf1:environment interaction is not significant in exudate or grape juice ( exudate: effect = 4 . 8±2 . 8% , t-test P = 0 . 1; grape juice: effect = −0 . 3±0 . 03% , t-test , P = 0 . 92 ) . Instead , rsf1 has a large main effect in YGlu as well as exudate and grape juice ( Figure 3 ) . Therefore , the effect of rsf1 can be best explained as an environment:rsf1:background interaction that reduces the effect of rsf1 in YGlu relative to exudate and grape juice , but only in the oak strain background . How do the effects of these environment and background interactions compare with the role played by QTN:QTN interactions ? We previously demonstrated significant QTN:QTN epistasis in the oak background and the glucose environment [14] . In that context , epistasis appears to play a large role in shaping phenotypic variation . However , the differences in rsf1's effect across backgrounds and environments imply that the QTN:QTN interactions might occur only in certain environments or backgrounds . We therefore tested for all possible QTN:QTN interactions across all eight environments and both genetic backgrounds . To do so , we modeled variation in sporulation efficiency in a standard linear framework using all phenotypic measurements across QTN genotypes , genetic backgrounds , and environments ( see Methods ) . A completely saturated model that incorporates all possible effects and interactions between environment , background , and QTN has an adjusted R2 of 0 . 99 . All of the parameters in the model are controlled variables , so this R2 indicates that 1% of the variation in our experiment is due to experimental error . We then constructed a reduced model that explains most of the variation , but with fewer parameters and only two and three-way interactions ( Figure 4 , adjusted R2 = 0 . 963 , see Methods ) . This model ( the global model ) captures the predominant interactions in the data ( Table S3 ) . For example , it contains a significant positive interaction term between galactose and the vineyard background ( effect = 25 . 6±2 . 4% , t-test P<2e-16 ) . This term is expected given the higher sporulation efficiencies we observe in the vineyard background in galactose ( Figure 2C ) . There are also significant interactions between rsf1 and both exudate ( effect = −28±2 . 8% , t-test P<2e-16 ) and grape juice ( effect = −13 . 8±2 . 8% , t-test P = 9e-7 ) , which are expected due to the larger effect of rsf1 in these two conditions ( Figure 3 ) . The most striking result from the global model is the lack of two-way QTN:QTN interactions . Three QTN:QTN interaction terms were left in the model after stepwise regression ( Table S3 ) . Only one of these , a negative interaction between rsf1 and ime1_coding ( effect = −7 . 6±1 . 8% , P = 3 . 9e-5 ) , passed either Bonferonni correction or permutation testing . This result stands in contrast to what we observe within a single condition . In line with our previous data in glucose [14] , we find abundant QTN:QTN interactions when YGlu is modeled alone ( Table S4 ) . For example , the rme1:ime1_coding interaction is large ( effect = −29 . 4±2 . 3% , t-test P<2e-16 ) . However , when all the environments and both backgrounds are analyzed together in the global model , the same rme1:ime1_coding interaction is small and only marginally significant ( effect = −3 . 9±1 . 6% , t-test P = 0 . 02 ) . In the place of QTN:QTN interactions , the global model contains several significant three-way QTN:QTN:environment and QTN:QTN:background interactions . This suggests that significant QTN:QTN interactions cause variation in sporulation efficiency , but the interactions only occur in particular environments and genetic backgrounds . To examine this possibility further , we modeled each environment-background combination separately and observed that QTN:QTN interactions varied widely . For example , the rme1:ime1_coding interaction that is strong in YGlu in the oak strain is marginal in exudate ( Figure 5 , YGlu effect = −29 . 4±2 . 3% , t-test P<2e-16 ; exudate effect = −4 . 2±2 . 1% , t-test P = 0 . 052 ) . This interaction is present in maltose ( effect = −29±3 . 9% , t-test P = 1 . 6e-8 ) , but not in the vineyard background ( Figure 6 , effect = 0 . 00±2 . 1% , t-test P = 0 . 23 ) . Taken together , these results show that the vineyard alleles of rme1 and ime1_coding act synergistically in the oak strain and specifically in YGlu and maltose , as the combination of two vineyard QTN produces a larger change in phenotype than could be expected from their individual effects . However , in exudate , or in the vineyard background , the effects of these same QTN alleles remain independent . Therefore , the synergistic interaction between the vineyard alleles is not intrinsic to the alleles themselves , but instead depends upon the specific context of the environment and genetic background . Our measurements of sporulation efficiency therefore indicate that QTN:QTN interactions are not widespread , but QTN:QTN:environment and QTN:QTN:background interactions are common . In a linear model that ignores genetic background and environment , no interactions between the QTN are significant ( adjusted R2 = 0 . 4 ) . This QTN-only model correctly identifies that individuals with all vineyard alleles tend to sporulate poorly , but it does not provide the ability to accurately predict the phenotypes of individuals with intermediate genotypes ( Figure 7 ) . Ultimately , the effects of the QTN and their interactions are shaped by the environmental and genomic context in which they occur . Knowledge of the environment and genetic background is therefore crucial to accurately predict the effects of QTN across individuals ( compare Figure 4 to Figure 7 ) .
In this set of experiments , we measured sporulation efficiency in a variety of isogenic strains that differed with respect to QTN genotypes , genetic background , and growth environment . Overall , our results show that a complex set of genotype:environment:background interactions shape variation in sporulation efficiency . Our results also shed light on the general effects of environment on sporulation efficiency in the context of natural variation . We found that carbon sources with similar effects on yeast catabolite repression tended to have similar effects on sporulation efficiency . For example , glucose and fructose both cause strong catabolite repression in yeast [21] , and their effects on sporulation efficiency are highly correlated ( Table 4 ) . Sucrose , a disaccharide composed of glucose and fructose , is likewise highly correlated with glucose . Raffinose and galactose , which cause weaker catabolite repression [22] , cluster less closely with glucose . One surprising result is the GxE we observed in maltose relative to glucose ( Figure 2B ) . Since maltose is composed of two glucose molecules , one might expect the effect of maltose to be as similar to glucose as that of sucrose or fructose . One possible explanation for the GxE in maltose arises from the fact that maltose catabolism genes commonly display copy number variation among yeast isolates [23]–[25] . We observed a slow growth phenotype of the oak strain in maltose and mapped this phenotype to the MAL1 multigene locus ( K . Lorenz and B . Cohen , unpublished results ) . We suspect that this locus is responsible for the maltose:background interaction we observe for sporulation efficiency , and it may also modulate the QTN effects and QTN:QTN interactions in maltose , but confirmation of this hypothesis awaits the cloning of the causative polymorphism . Exudate and grape juice produce lower sporulation efficiencies than the other environments . This result occurs in spite of the fact that exudate is composed of exactly the same ingredients as YGlu , but with reduced concentrations of peptone and yeast extract . This reduction of nutrient concentrations not only reduces sporulation efficiency in both genetic backgrounds , but it also alters the effect of rsf1 in the oak background relative to the other QTN ( Figure 3 ) . The fact that exudate consists of the same ingredients as YGlu but produces different effects on sporulation efficiency suggests that QTN effects are shaped not only by nutrient type , but also by nutrient concentrations . Drops in nitrogen concentration are well-known to strengthen the signal to sporulate , so the difference in peptone concentration between exudate and rich media may explain some the differences in sporulation efficiency through nitrogen sensing . Across multiple environments , the unknown polymorphisms in the genetic background not only interact with the environment but also alter the effects of the known QTN . The known QTN used in this study were mapped in glucose and explain ∼90% of the segregating variation in that condition [14] . The interactions we observe here suggest that the remaining unmapped loci may have stronger effects ( and be easier to map ) in non-YGlu environments . For example , the known QTN only explain half of the phenotypic difference in the vineyard background in grape juice ( Table 3 ) . Presumably , the remaining unknown polymorphisms that regulate sporulation efficiency have larger effects in this environment-background combination than they do in YGlu . An attractive experiment to identify new QTN governing sporulation efficiency would therefore be to map the phenotype in grape juice using a cross of the original oak parent with a new version of the vineyard parent strain that is fixed for all four known oak QTN . It is possible , however , that the new polymorphisms uncovered by this experiment would not reside in the sporulation pathway per se , but would instead be metabolic factors specific to grape juice catabolism . Despite the fluctuations in QTN effects across environments and backgrounds , the direction of QTN effects remain consistent . Vineyard alleles always decrease sporulation efficiency relative to oak alleles . Without accounting for changes in the environment or differences in genetic background , we can therefore safely predict that a strain with all four vineyard alleles will sporulate poorly relative to a strain carrying all oak alleles . However , because the effect magnitudes of the QTN change across environments and backgrounds , we cannot predict the sporulation efficiency of intermediate allelic combinations ( Figure 7 ) . This case reminds us of the situation unfolding in human association studies , where it appears that high-risk individuals can be identified as carriers of collections of disease associated polymorphisms , even though it is more difficult to predict the actual phenotypic outcome of a particular individual with intermediate sets of alleles [26] . In this case of yeast sporulation efficiency , complexity occurs because the relative importance of particular alleles and their interactions are not constant across individuals , but instead vary with the individuals' genetic background and environment . If context dependencies on allelic effects are common , how can we achieve better predictive power when environment and background are unknown ? Environment and genetic background presumably influence the phenotype just as all genetic changes must: through effects on cell physiology . It might be possible to account for the physiological effects of environment and background using a biomarker or physiological indicator that is correlated with , but upstream of , the phenotype of interest . Biochemical markers are used in medicine to inform calculations of disease risk and diagnosis [27] . Inclusion of a physiological marker into the genetic model may condition the model to unknown parameters and therefore increase the accuracy of genotype-phenotype predictions . Although such a model could improve predictive power , it still does not increase our understanding of how various physiological forces in the cell combine to quantitatively alter phenotype . Perhaps improved understanding could arise from interpreting QTN effects through a framework rooted in cell biology and biochemistry , rather than through an abstract linear model . Biochemical and gene regulatory pathways have long been theorized to naturally generate non-linear effects through the basic thermodynamic properties of proteins and DNA [28] , [29] . We have modeled sporulation efficiency in glucose through a thermodynamic framework , and this method shows promise in revealing the molecular basis of genetic interactions [30] . However , thermodynamic modeling requires detailed knowledge of molecular mechanism of the proteins involved , and this information is not available for most traits . Also , the challenge of applying this approach to multiple environments is nontrivial [31] . A more traditional method to deal with statistical interactions is to eliminate them through data transformations . We experimented with a number of scale transformations for our dataset , but found that the best transformation for reducing the complexity of the interaction terms varied from one environment:background combination to the next . Furthermore , data transformations that reduced the number of interaction terms sometimes had undesirable effects , such as increasing the dependence of the variance upon the mean . More importantly , scale transformations that worked well on some subsets of the data still required numerous interaction terms to provide a global model . None of the data transformations we tried improved the three-way interaction fit obtained on the natural scale ( Figure S3 ) . Although data transformations may be appropriate to obtain simpler predictive models in single background:environment combinations , they do not account for the non-linear dynamics that create complexity across conditions and backgrounds . Regardless of the approach taken in the future , our results clearly show that the genetic architecture of sporulation efficiency is environment-dependent . QTN effects cannot be understood without taking into account contextual factors such as the environment's influence on cell physiology . We expect that quantitative biochemical measurements will be required to illuminate what is happening inside the cell and bridge the missing link between genotype and phenotype .
Each of the 32 strains were grown for 15 hours in growth media ( except for grape juice , in which we instead grew the yeast for 54 hours ) . After the growth period , we diluted each culture 1∶50 into 1% potassium acetate to induce sporulation . We tested three replicates of each QTN genotype - environment - genetic background combination . One exception is the strain carrying only the ime_nc vineyard QTN allele in the vineyard background grown in sucrose , for which there were only two measurements due to a sample failure . The experimental design is balanced such that the genotype frequencies of the four QTN do not vary across environments or backgrounds , so any significant interactions between QTN reflect physiological effects rather than differences in allele frequency [32] . Sporulation efficiency was calculated by flow cytometry on samples of 15 , 000 cells per replicate using methods we have described elsewhere [20] . The raw data of sporulation efficiencies for each replicate is available as a supplementary data file ( Dataset S1 ) . Each of the eight environmental treatments was composed of a different growth medium prior to the induction of sporulation in acetate ( Table 1 ) . Six of the environments consisted of rich yeast media ( 1% yeast extract , 2% peptone ) supplemented with 2% of a sugar or polysaccharide: glucose , fructose , sucrose , maltose , galactose , or raffinose . The other two environments were synthetic oak exudate and chardonnay grape juice . Synthetic oak exudate is composed of the same nutrients as rich media , but contains yeast extract and peptone at ten-fold reduced concentrations ( Table 1 ) . Exudate also contains a mixture of fructose , sucrose , and glucose at a total concentration of 2% [33] . After each environmental treatment , sporulation was induced for 30 hours in 1% potassium acetate , which provides a non-fermentable carbon source but no source of nitrogen . First , we created allele replacement strains in each parental background that carry single QTN alleles from the opposite parent [34] . These strains were created by backcrosses of initial haploid ura3− allele replacement transformants with their prototrophic diploid parents . Ura3+ progeny from the backcross of each allele replacement were then intercrossed to generate strains carrying multiple QTN alleles from the opposite parent . Each cross was performed in triplicate . We confirmed after each cross that the QTN co-segregated with variation in sporulation efficiency in glucose , and we also ensured that the phenotypes resulting from replicate crosses were identical . This assured us that no new mutations governing sporulation efficiency had arisen elsewhere in the genome during the crossing scheme . Once a strain with the desired QTN alleles from the opposite parent was created , this strain was backcrossed once more to its original wild type parent strain . Individual homothallic diploid progeny from this final cross were isolated and genotyped until we obtained three replicates of every possible QTN allele combination . Genotyping was based on the restriction digest of PCR amplicons [14] . The selected strains were arrayed in a 96-well plate such that all the strains from both genetic backgrounds can be assayed in a single block . We generated a matrix of the Spearman rank correlations of the means of each of the 32 strains across each environment . A distance matrix was then defined as 1−ρ , where ρ is the matrix of pair wise Spearman rank correlations . We carried out hierarchal cluster analysis with the complete linkage clustering method as implemented in the hclust function in the statistical package R . We also split the data by genetic background , then calculated rank correlations and clustered separately for the oak and vineyard genetic backgrounds . All statistical analyses were performed in R . In all linear models , the strain with all oak QTN alleles was treated as the intercept , so the additive effects represent the effect of a single vineyard QTN placed into a strain with oak QTN alleles at all other loci . We chose this reference point because the oak strain probably best resembles the genotype of the common ancestor of the two parent strains [14] , [35] . To compare the effects of QTN:QTN interactions in single environment-background combinations , we created linear models of QTN effects including all possible interaction terms within each condition , and significant coefficients were calculated by t-tests of the coefficient's estimated effect versus its standard error . All interaction terms reported in the text are significant by Bonferonni correction ( P = 0 . 05/N , where N is the number of coefficients in the model ) . To analyze QTN effects across all environments and both backgrounds , we constructed a linear model in which the oak genetic background , oak QTN alleles , and the glucose environment are treated as intercepts . Therefore , coefficients in the model represent the effects of the vineyard genetic background , vineyard QTN alleles , and non-glucose environments . The simplest additive model therefore takes the following form:Where EFF is sporulation efficiency , Oak is the oak strain phenotype in glucose ( the y-intercept ) , BG is the effect of genetic background , RME , RSF , IMEC , and IMENC are the effects of the vineyard QTN alleles , ENV is the effect of non-glucose environments , and e represents the error across the multiple replicates of each combination of strain and environmental treatment . Sucrose and fructose were not significantly different from glucose , so these three conditions were pooled into a single treatment . We found that models with increasing levels of interaction terms were often significant , but very little improvement to the fit or explanatory power of the model was gained by adding four-way interactions ( Table S5 ) . We therefore limited our analysis to models with three-way interaction terms to reduce saturation without much sacrifice of explanatory power . To select a specific model with a subset of the three-way terms , we used stepwise regression as implemented in the stepAIC function in R . We then took the output from stepwise regression and manually removed terms from the model if their treatment contrast P-values did not pass a model-wide Bonferonni correction . Table S3 displays the coefficients in our final model and the P-values of each coefficient . The significance of the QTN:QTN interactions in this model were also tested by creating 10 , 000 null linear models from random permutations of the entire dataset . The critical P-value from these permutations was P = 0 . 016 . Probit and Logit transformations , which are common used for frequency data , provide good fits with fewer interaction terms in some individual conditions . However , we chose to model the data on the raw scale . The Probit and Logit transformations obtain a fit by weighting the explanatory power at extremely high and low values of sporulation efficiency at the expense of intermediate values ( Figure S3 ) . For example , under the Probit transformation , a difference in sporulation efficiency from 1% to 2% is as great in magnitude as a raw difference of 40% to 50% . No transformation eliminated interactions altogether , and transformations did not improve the overall fit of the model across multiple environments . The raw scale allows more intuitive interpretation of the model coefficients , and our reduced model performs well on values of sporulation efficiency between ∼5 and 95% . Some extreme values are fit below zero or above 100% . However , with one exception ( a data point at 92% ) , all data points fitted to higher than 100% have actual values greater than 96% . All data points predicted to be below zero have actual values less than 5% . We therefore simply bounded all predicted values between 0% and 100% . To model the specific QTN:environment interactions in exudate and grape juice , we conducted an analysis of variance on only the additive effects ( no QTN:QTN interactions ) of the four vineyard QTN separately in each genetic background . This model focused on the additive effects because the phenotypes of vineyard background strains carrying multiple vineyard QTL approach zero in a non-linear fashion ( Tables S1 , S2 ) . The model took the form:Where EFF is sporulation efficiency , GEN is the genotype across the four QTN alleles , ENV is the environment , and e is the error . To confirm significant differences in the rank order of QTN effects in different environments , we took the estimated QTN effects from an analysis of variance in each environment separately and computed Tukey's Honest Significant Difference to determine the rank-order of the QTN within each environment . The reported P value is the largest adjusted P value among all the possible comparisons between the effect of rsf1 and the effects of other QTN . | Phenotypic variation among individuals is caused by naturally occurring genetic differences , or alleles . The relationship between an allele and the phenotype is extremely complex; for example , the effect of an allele often depends upon both the environment and the individual's genetic background . To better understand these complex relationships , we examined the effects of four quantitative trait nucleotides ( QTN ) in three genes that cause variation in sporulation efficiency between vineyard and oak tree strains of yeast . We measured the effects of the QTN while varying both the genetic makeup of the strains and their growth environments . We found that the effects of each of the four QTN alleles depended upon the genotypes at the other QTN , the growth environment , and whether the strain carried the oak or vineyard parent genome . There were no simple rules that describe the effects of the alleles across all environments; instead , detailed models were needed to account for environmental and genetic variation in order to predict the effects of alleles in specific individuals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/complex",
"traits"
] | 2010 | Gene–Environment Interactions at Nucleotide Resolution |
Malaria control relies heavily on pyrethroid insecticides , to which susceptibility is declining in Anopheles mosquitoes . To combat pyrethroid resistance , application of alternative insecticides is advocated for indoor residual spraying ( IRS ) , and carbamates are increasingly important . Emergence of a very strong carbamate resistance phenotype in Anopheles gambiae from Tiassalé , Côte d'Ivoire , West Africa , is therefore a potentially major operational challenge , particularly because these malaria vectors now exhibit resistance to multiple insecticide classes . We investigated the genetic basis of resistance to the most commonly-applied carbamate , bendiocarb , in An . gambiae from Tiassalé . Geographically-replicated whole genome microarray experiments identified elevated P450 enzyme expression as associated with bendiocarb resistance , most notably genes from the CYP6 subfamily . P450s were further implicated in resistance phenotypes by induction of significantly elevated mortality to bendiocarb by the synergist piperonyl butoxide ( PBO ) , which also enhanced the action of pyrethroids and an organophosphate . CYP6P3 and especially CYP6M2 produced bendiocarb resistance via transgenic expression in Drosophila in addition to pyrethroid resistance for both genes , and DDT resistance for CYP6M2 expression . CYP6M2 can thus cause resistance to three distinct classes of insecticide although the biochemical mechanism for carbamates is unclear because , in contrast to CYP6P3 , recombinant CYP6M2 did not metabolise bendiocarb in vitro . Strongly bendiocarb resistant mosquitoes also displayed elevated expression of the acetylcholinesterase ACE-1 gene , arising at least in part from gene duplication , which confers a survival advantage to carriers of additional copies of resistant ACE-1 G119S alleles . Our results are alarming for vector-based malaria control . Extreme carbamate resistance in Tiassalé An . gambiae results from coupling of over-expressed target site allelic variants with heightened CYP6 P450 expression , which also provides resistance across contrasting insecticides . Mosquito populations displaying such a diverse basis of extreme and cross-resistance are likely to be unresponsive to standard insecticide resistance management practices .
Malaria mortality has decreased substantially in sub-Saharan Africa over the last decade , attributed in part to a massive scale-up in insecticide-based vector control interventions [1] . As the only insecticide class approved for treatment of bednets ( ITNs ) and the most widely used for indoor residual spraying ( IRS ) , pyrethroids are by far the most important class of insecticides for control of malaria vectors [2] . Unfortunately pyrethroid resistance is now widespread and increasing in the most important malaria-transmitting Anopheles species [3]–[5] and catastrophic consequences are predicted for disease control if major pyrethroid failure occurs [6] . With no entirely new insecticide classes for public health anticipated for several years [5] , [6] preservation of pyrethroid efficacy is critically dependent upon strategies such as rotation or combination of pyrethroids with just three other insecticide classes , organochlorines , carbamates and organophosphates [6] , [7] . In addition to logistical and financial issues , insecticide resistance management suffers from knowledge-gaps concerning mechanisms causing cross-resistance between available alternative insecticides , and more , generally how high-level resistance arises [8] . With strongly- and multiply-resistant phenotypes documented increasingly in populations of the major malaria vector Anopheles gambiae in West Africa [9]–[13] such information is urgently required . Of the four classes of conventional insecticide licensed by the World Health Organisation ( WHO ) , pyrethroids and DDT ( the only organochlorine ) both target the same para-type voltage-gated sodium channel ( VGSC ) . This creates an inherent vulnerability to cross-resistance via mutations in the VGSC target site gene [14]–[16] , which are now widespread in An . gambiae [5] . In contrast , carbamates and organophosphates cause insect death by blocking synaptic neurotransmission via inhibition of acetylcholinesterase ( AChE ) , encoded by the ACE-1 gene in An . gambiae . Consequently , target site mutations in the VGSC gene producing resistance to pyrethroids and DDT will not cause cross-resistance to carbamates and organophosphates . The carbamate bendiocarb is being used increasingly for IRS [17] , [18] , and has proved effective in malaria control programs across Africa targeting pyrethroid- or DDT-resistant An . gambiae [18]–[20] . A single nucleotide substitution of glycine to serine at codon position 119 ( Torpedo nomenclature; G119S ) in the ACE-1 gene , which causes a major conformational change in AChE , has arisen multiple times in culicid mosquitoes [21] , [22] , and is found in An . gambiae throughout West Africa [23]–[25] . The G119S mutation can produce carbamate or organophosphate resistance [26] but typically entails considerable fitness costs [27]–[30] . This is beneficial for resistance management because in the absence of carbamates or organophosphates , serine frequencies should fall rapidly [29] , [31] . In Culex pipiens , duplications of ACE-1 create linked serine and glycine alleles , which , when combined with an unduplicated serine allele , creates highly insecticide resistant genotypes with near-full wild-type functionality , thus providing a mechanism that can compensate for fitness costs [28] , [31] . Worryingly , duplication has also been found in An . gambiae [23] though the consequences of copy number variation for fitness in the presence or absence of insecticide are not yet known in Anopheles . Though far from complete , information is available for metabolic resistance mechanisms to pyrethroids and DDT in wild populations of An . gambiae [5] , [6] , [32]–[34] . Indeed , a specific P450 enzyme , CYP6M2 , has been demonstrated to metabolize both of these insecticide classes , suggesting the potential to cause cross-resistance in An . gambiae [32] , [35] . By contrast little is known about metabolic mechanisms of carbamate resistance in mosquitoes and , as a consequence , potential for mechanisms of cross-resistance are unknown . A particularly striking and potentially problematic example of insecticide resistance has been found in one of the two morphologically identical , but ecologically and genetically divergent molecular forms comprising the An . gambiae s . s . species pair ( M molecular form , recently renamed as An . coluzzii [36] ) in Tiassalé , southern Côte d'Ivoire . The Tiassalé population is resistant to all available insecticide classes , and displays extreme levels of resistance to pyrethroids and carbamates [11] . The VGSC 1014F ( ‘kdr’ ) and ACE-1 G119S mutations are both found in Tiassalé [11] , [25] . Yet kdr shows little association with pyrethroid resistance in adult females in this population [11] . ACE-1 G119S is associated with both carbamate and organophosphate survivorship [11] , but this mutation alone cannot fully explain the range of resistant phenotypes , suggesting that additional mechanisms must be involved . Here we apply whole genome microarrays , transgenic functional validation of candidates , insecticide synergist bioassays , target-site genotyping and copy number variant analysis to investigate the genetic basis of ( 1 ) extreme bendiocarb resistance and ( 2 ) cross-insecticide resistance in An . gambiae from Tiassalé . Our results indicate that bendiocarb resistance in Tiassalé is caused by a combination of target site gene mutation and duplication , and by specific P450 enzymes which produce resistance across other insecticide classes .
Our study involved two microarray experiments ( hereafter referred to as Exp1 and Exp2 ) , involving solely M molecular form An . gambiae ( Table S1 ) , to identify candidate genes involved in bendiocarb resistance ( full microarray results for Exp1 and Exp2 are given in Table S2A ) . In Exp1 gene expression profiles of female mosquitoes from bendiocarb-susceptible laboratory strains ( NGousso and Mali-NIH ) and a bendiocarb-susceptible field population ( Okyereko , Ghana ) , none of which were exposed to insecticide , were compared to those of Tiassalé females . Two Tiassalé groups were used: either without insecticide exposure ( Figure 1A ) , or the survivors of bendiocarb exposure selecting for the 20% most resistant females in the population [11] ( Figure 1B ) . We used a stringent filtering process to determine significant differential expression ( detailed in the legend to Figure 1 ) , which included criteria on both the probability and consistency of direction of differential expression , and also required a more extreme level of differential expression in the Tiassalé-selected than Tiassalé ( unexposed ) vs . susceptible comparisons . Inclusion of this third criterion enhanced the likelihood that genes exhibiting differential expression are associated with bendiocarb resistance , rather than implicated via indirect association with another insecticide . Moreover , the requirement for significance in comparisons involving both bendiocarb-exposed and unexposed Tiassalé samples ( Figure 1A , B ) negates the possibility that any differential expression identified was a result solely of induction of gene expression by insecticide exposure . In Exp1 145 probes were significant , out of a total of 14 914 non-control probes , with almost all ( 143/145 ) expressed at a higher level in the resistant samples ( Table S2B ) . Functional annotation clustering analysis detected two significant clusters within the significantly over-expressed genes ( Table S2C ) . The larger cluster was enriched for several P450s and the functionally-related genes cytochrome b5 and cytochrome P450 reductase . Of these , CYP6P3 , CYP6P4 , CYP6M2 and cytochrome b5 are evident amongst the most significant and/or over-expressed probes in Figure 2A . Of the five physically-adjacent CYP6P subfamily genes in An . gambiae , CYP6P1 and CYP6P2 were also significant ( Table S2B ) , and CYP6P5 only marginally non-significant according to our strict criteria ( five out of the six comparisons q<0 . 05 ) . The four probes for the ACE-1 target site gene exhibited the strongest statistical support ( lowest q-values ) for resistance-associated overexpression in the Exp1 dataset ( Figure 2A ) . Experiment 2 employed a simpler design in which bendiocarb resistant samples from Kovié ( Togo ) were compared to the same Okyereko field samples used in Exp1 and to a second field population from Malanville ( Benin ) . Significant differential expression was determined according to the first two criteria employed for analysis of Exp1 ( Figure 1 ) . The likelihood of specificity of results to the bendiocarb resistance phenotype was enhanced because all three populations used in Exp2 exhibit resistance to pyrethroids and DDT , all are susceptible to organophosphates , but only the Kovié population is resistant to bendiocarb . In Exp2 2453 probes were significantly differentially expressed ( Table S2D ) ; likely reflecting the lower number of pairwise comparisons available for stringent filtering than in Exp1 . Consequently we do not consider results from Exp2 alone in detail . Nevertheless it is interesting to note that the lowest q-values and highest fold-changes were both for alcohol dehydrogenase genes ( Figure S1 ) , and the latter is the physical neighbour and closest paralogue of the highly overexpressed alcohol dehydrogenase in Exp1 ( Figure 2A ) . Sixteen probes , representing only seven genes , were significant in both Exp1 and Exp2 ( Figure 2B ) , including all replicate probes for three of the CYP6 P450 genes highlighted previously . Of these , CYP6M2 was most highly over-expressed , second only to Ribonuclease t2 . However , results for Ribonuclease t2 were much more variable , with differential expression dramatically high compared to lab strains , but moderate or low compared to wild populations ( Table S2E ) . Evidence for specific involvement in bendiocarb resistance is suggested by significance of two of the CYP6M2 probes in the ( relatively low-powered ) direct comparison of bendiocarb selected vs . unselected samples within Exp1; the other two CYP6M2 probes and two of those for ACE-1 were marginally non-significant ( 0 . 05<q<0 . 10; Figure S2 ) . Five genes were chosen for further analysis: ACE-1 and CYP6P3 from Exp1; CYP6M2 and CYP6P4 from Exp1+Exp2; and CYP6P5 , which we included because of a suspected type II error in the microarray analysis ( see above ) . qRT-PCR estimates of expression , relative to the susceptible Okyereko population , showed reasonable agreement with microarray estimates albeit with some lower estimates ( Figure S3 ) . CYP6M2 and CYP6P4 exhibited up to eight and nine-fold overexpression , and ACE-1 six-fold compared to Okyereko , though high variability among biological replicates for the P450 genes resulted in relatively few significant pairwise comparisons ( Figure 3 ) . Nevertheless the hypothesis that fold-changes should follow the rank order predicted by the level of bendiocarb resistance in each comparison ( i . e . Tiassalé selected>Tiassalé unexposed>Kovié ) was met qualitatively for all genes ( Figure 3 ) . For functional validation via transgenic expression in D . melanogaster , we chose CYP6P3 and CYP6M2; both of which have been shown to metabolize pyrethroids [34] , [35] , and CYP6M2 also DDT [32] . The capacity of each gene to confer resistance to bendiocarb , to the class I and II pyrethroids permethrin and deltamethrin , respectively , and to DDT and was assessed by comparing survival of transgenic D . melanogaster , exhibiting ubiquitous expression of CYP6M2 or CYP6P3 ( e . g . UAS-CYP6M2/ACT5C-GAL4 experimental class flies ) , to that of flies carrying the UAS-CYP6M2 or CYP6P3 responder , but lacking the ACT5C-GAL4 driver ( e . g . UAS-CYP6M2/CyO control class flies ) . For CYP6M2 the relative expression level of the experimental flies was 4 . 0 and for CYP6P3 4 . 3 ( Table S3 ) . As indicated by elevated LC50 values ( Figure S4 ) , expression of either CYP6M2 or CYP6P3 produced pyrethroid resistant phenotypes , and CYP6M2 expression also induced significant DDT resistance ( Table 1 ) . Assays for CYP6P3 with DDT did not produce reproducible results ( data not shown ) . Flies expressing the candidate genes exhibited greater survival across a narrow range of bendiocarb concentrations ( Figure S4 ) . However , at a discriminating dosage of 0 . 1 µg/vial [37] a resistance ratio of approximately seven was exhibited for CYP6M2/ACT5C: CYP6M2/CyO flies ( Mann-Whitney , P = 0 . 0002; Figure 4 ) with a much smaller , but still significant , ratio of approximately 1 . 4 ( Mann-Whitney , P = 0 . 019 ) for CYP6P3/ACT5C: CYP6P3/CyO flies . Caution is required in quantitative interpretation of the resistance levels generated , both because of the non-native genetic background and also ubiquitous expression of genes that may be expressed in a tissue-specific manner [38] . Nevertheless , the bioassays on transgenic Drosophila show that each P450s can confer resistance to more than one insecticide class . Recombinant CYP6M2 and CYP6P3 were expressed in E . coli with An . gambiae NADPH P450 reductase and cytochrome b5 . An initial experiment , using 0 . 1 µM P450 and 2 hour incubation with bendiocarb , demonstrated metabolism of bendiocarb by CYP6P3 ( 64 . 2% mean depletion ±4 . 0% st . dev ) but no metabolic activity of CYP6M2 ( 0±11 . 0% ) . Further investigation of CYP6P3 activity across a range of incubation times ( Figure 5a ) and enzyme concentrations ( Figure 5b ) supported the initial observation , with metabolism plateauing at a maximum of 50% . An . gambiae from Tiassalé are classified as resistant to all classes of WHO-approved insecticides ( <90% bioassay mortality 24 hours after a 60 min exposure ) , with resistance phenotypes stable across wet and dry seasons ( Figure 6 , Table S4 ) . Nevertheless , resistance varies markedly among insecticides ( Table S4 ) , with notably higher prevalence for bendiocarb and DDT than the organophosphate fenitrothion . The synergist PBO , which is primarily considered an inhibitor of P450 enzymes , exerted a significant influence on bioassay mortality ( Table S4 ) for four of the five insecticides tested , with only DDT not significantly impacted ( Figure 6 ) . The synergising effect of PBO was strongest for bendiocarb , with a near five-fold increase in mortality , equivalent to an odds ratio for PBO-induced insecticidal mortality exceeding ten ( Figure 6 ) . However , for all of the insecticides , apart from fenitrothion , over 20% of the population survived even with PBO pre-exposure . The ACE-1 G119S substitution is the only non-synonymous target site mutation known in An . gambiae [23] , and the resistant ( serine ) allele is common in Tiassalé with an estimated frequency of 0 . 46 ( N = 306 ) . All occurrences of serine are in heterozygotes ( 95% confidence limits for heterozygote frequency: 0 . 87–0 . 94 ) , which underlies a dramatic deviation of genotype frequencies from Hardy-Weinberg equilibrium ( ÷2 = 135 . 5 , P≈0 ) . To examine the independence of putatively P450-mediated resistance and AChE target site insensitivity , we typed the G119S locus in females from the diagnostic ( 60 min ) bendiocarb assays with and without pre-exposure to PBO . In either case absence of the 119 serine allele appears to almost guarantee mortality to bendiocarb ( Table S5 ) , as previously observed for fenitrothion bioassays in Tiassalé [11] . However , the strong bendiocarb resistance association of G119S was reduced significantly by PBO pre-exposure ( homogeneity ÷2 = 8 . 3 , P = 0 . 004 ) with the probability of survival for heterozygotes reduced to approximately 50% ( Table S5 ) . To investigate whether heterozygote survivorship might be linked to copy number variation , via a difference in numbers of serine and glycine alleles , we examined the qPCR dye balance ratio for live and dead individuals within the heterozygote genotype call cluster ( Figure 7A ) . In many individuals called as heterozygotes , a markedly higher ratio of 119S: 119G dye label than the 1∶1 expected for a true heterozygote is evident ( Figure 7A ) , and surviving heterozygotes exhibited a significantly higher serine: glycine dye signal ratio than those killed ( t-test , P = 1 . 5×10−5 ) . We designed an additional qRT-PCR diagnostic to investigate copy number more directly in a portion of the surviving and dead individuals typed as G119S heterozygotes . The difference in copy number was highly significant between survivors and dead ( Figure 7B ) , with 15/16 survivors but only 5/16 dead females exhibiting a copy number ratio in excess of 1 . 5 ( Table S5 ) , consistent with possession of an additional allele . These results show that independent of the enzymes inhibited by PBO survival , females heterozygous for the G119S mutation ( i . e . most individuals in Tiassalé ) depends upon Ace-1 copy number variation and possession of additional resistant serine alleles .
The major biochemical mechanisms of carbamate resistance in mosquitoes have previously been identified as modified AChE ( via point substitutions , most notably G119S ) and less frequently esterase-mediated metabolism [7] . PBO-induced increases in carbamate mortality have been reported in wild mosquito populations exhibiting low to moderate resistance levels , including M form An . gambiae from West Africa [12] , [39] , [40] . The significant synergizing effect of PBO in the present work and these previous studies is consistent with a role of P450s in carbamate resistance , but should not be taken alone as direct proof [41] because PBO exposure can also inhibit some esterases [42] , [43] . However , our microarray data clearly identified over-expression of multiple CYP6 P450 genes , whereas only a single carboxylesterase gene ( COEAE6G ) was significant , and expressed at a lower level ( Table S2B ) . Taken together , the synergist data and transcriptional profiles indicate that a substantial proportion of the Tiassalé population is dependent upon the action of P450s for resistance to bendiocarb . Near-equivalent synergism of permethrin and deltamethrin , coupled with identification and functional validation of shared candidate genes , suggests the same conclusion for pyrethroids . For fenitrothion , the effect of PBO is also consistent with P450 involvement , but in the absence of specific candidate genes , additional supporting evidence will be required to confirm this hypothesis . Genes from the CYP6P cluster emerged as strong candidates for involvement in P450-mediated detoxification . CYP6P3 overexpression has been linked repeatedly with pyrethroid resistance in An . gambiae [33] , [34] , as has its orthologue in An . funestus CYP6P9 [44] , [45] and both enzymes can metabolise class I and II pyrethroids [34] , [35] , [45] . We demonstrate that CYP6P3 can produce significant resistance to both classes of pyrethroid and , to a lesser extent bendiocarb , in D . melanogaster . We also show that recombinant CYP6P3 can metabolise bendiocarb in vitro; the third mosquito P450 to metabolise a carbamates , after An . gambiae CYP6Z1 and CYP6Z2 which have been demonstrated to metabolise the insecticide carbaryl [46] . Interestingly CYP6P4 , which , in contrast to CYP6P3 , was also significantly overexpressed in the Togolese Kovié population , is the orthologue of the resistance-associated CYP6P4 gene in An . funestus [44] , and along with CYP6P3 was recently found to be overexpressed in DDT-resistant samples of both M and S molecular forms of An . gambiae from Cameroon [47] . Although we were unable to obtain data for the impact of CYP6P3 expression on survival with DDT exposure in D . melanogaster , the potential of CYP6P genes to act on DDT merits further investigation . It is also interesting to note that both cytochrome b5 and cytochrome P450 reductase , both important for P450-mediated insecticidal detoxification [48] are overexpressed in Tiassale , suggesting a possible role in resistance for co-expression of these genes with the CYP6 P450s . CYP6M2 was overexpressed in Tiassalé , Kovié , and also in the Tiassalé bendiocarb-selected vs . control comparison . CYP6M2 expression generated Drosophila phenotypes significantly resistant to bendiocarb , DDT , and class I and II pyrethroids . Overexpression of CYP6M2 has been linked repeatedly to pyrethroid [33] , [34] and DDT resistance [32] , [47] in An . gambiae , and is known to metabolise both these classes of insecticide [32] , [35] . Our data now suggest a role in bendiocarb resistance , and overall provide strong evidence for involvement in resistance to three classes of insecticide . The biochemical mechanism of involvement remains unclear however because CYP6M2 did not metabolise bendiocarb in vitro , though we cannot rule out the possibility that some unknown , and thus currently , absent co-factor might be required . Sequestration also seems unlikely since CYP6M2 does not appear to bind bendiocarb . A role in breakdown of secondary bendiocarb metabolites certainly remains plausible , though at present knowledge of such mechanisms for any insecticide in mosquitoes is very limited [49] , [50] . High variability in CYP6M2 expression among biological replicates , especially evident in qRT-PCR , suggests that the regulatory mechanism ( s ) generating overexpression is far from fixation in Tiassalé . Further work is required to determine whether the cause of overexpression might be gene amplification , as seen for insecticide-linked CYP6P genes in An . funestus [44] and CYP6Y3 in the aphid Myzus persicae [51] or a cis regulatory variant , or both , as documented for CYP6G1 in D . melanogaster [52] . In either case , the actual level of expression in individuals possessing causal regulatory variant ( s ) may be much higher than we detected from pooled biological replicates . As a consequence , it is possible that CYP6M2 ( and other key P450s ) might be expressed at too high a level for PBO to fully inhibit at the dosage applied , resulting in only partial synergy . Indeed it is interesting that CYP6M2 generated significant DDT resistance in transformed Drosophila in our study and has been shown metabolise DDT [32] yet PBO provided only very slight and non-significant synergy for DDT-exposed Tiassalé females . An inadequate concentration of PBO might be important , but it is worth noting that levels of DDT resistance in West African An . gambiae can be extreme and are likely to be underpinned by additional mechanisms [32] such as the significantly resistance-associated kdr L1014F target site mutation in Tiassalé [11] . Whilst incomplete synergy of highly expressed P450 enzymes might be a partial explanation , our results point to target site mechanisms as a key factor underpinning survival following PBO and bendiocarb exposure . Possession of the ACE-1 119 serine variant appears to be a near-prerequisite for bendiocarb-survival in Tiassalé , as documented previously for fenitrothion [11] . This is apparently not the case in all An . gambiae populations , with some individuals lacking the serine mutation surviving a standard 60 min exposure [12] , [39] . Over 90% of Tiassalé mosquitoes are heterozygous for G119S , which could be consistent with fitness costs for individuals lacking a fully-functional wild-type allele since the serine allele exhibits lowered activity [28] . It is apparent though that possession of the ACE-1 G119S mutation represents only a portion of the target site mediated resistance mechanism . Tiassalé females generally showed much higher expression of ACE-1 than all other populations in our experiments , reaching approximately six-fold in the highly resistant bendiocarb-selected group compared to the Okyereko susceptible group . Following PBO-mediated P450 inhibition , survival of G119S heterozygotes was reduced to approximately 50% and our results show that individuals exhibiting a higher ACE-1 copy number and more copies of the serine allele had a significant survival advantage . Together these results indicate that the primary explanation for the ubiquitous heterozygosity found in Tiassalé is an elevated copy number of expressed ACE-1 alleles . At least in individuals possessing additional serine alleles , this enhances carbamate resistance , and can apparently generate resistance independently of P450 activity . Extra copies of ACE-1 alleles have been found in West African An . gambiae , and lack of sequence variation suggests that duplication is a very recent event [23] . Consequences of ACE-1 duplication have not been documented previously in Anopheles but Cx . pipiens possessing two G119S resistant alleles and a wild type susceptible allele can exhibit near maximal fitness in the presence and absence of organophosphate treatment [30] . If this fitness scenario is similar in An . gambiae ACE-1 duplicates could spread rapidly , or may have already done so but have been largely undetected by available diagnostics . The estimated copy numbers we detected in some individuals suggests that more ACE-1 copies may be present in An . gambiae than are known in Cx . pipiens , perhaps more akin to the high level of amplification found in spider mites Tetranychus evansi [53] . This raises the possibility of a potentially multifarious set of resistant phenotypes dependent upon the number and G119S genotype of the copies possessed by an individual , understanding of which will benefit from further application of the DNA-based qPCR diagnostic we have developed . Extreme levels of resistance to single insecticides , and multiple resistance across different insecticidal classes represent major problems for control of disease vectors , and pest insects generally . Tiassalé An . gambiae show exceptionally high-level carbamate resistance and the broadest insecticide resistance profile documented to date . Our results indicate that overexpression of specific CYP6 enzymes and duplicated resistant ACE-1 alleles are major factors contributing to this resistance profile . Results from the less resistant Kovié population show that at least some of the mechanisms are not restricted to Tiassalé and could be quite widespread in West Africa . The involvement of CYP6P3 and CYP6M2 in resistance to multiple insecticide classes parallels the cross resistance engendered by CYP6 genes in other insect taxa [54] , [55] and is extremely concerning because resilience to standard resistance management strategies is likely to be increased greatly . Further work is now required to understand the biochemical role of CYP6M2 in detoxification of bendiocarb and also to better understand any associated fitness costs of elevated CYP6P gene expression . In addition , whilst we have demonstrated involvement of elevated expression of the CYP6 P450s in insecticide resistance , the impact of structural variants within these genes remains to be investigated and is very poorly understood for P450-mediated insecticide resistance in mosquitoes . In spite of a major impact of PBO on three distinct insecticide classes , too many females remained alive to suggest that PBO provides a resistance-breaking solution . Nevertheless , we suggest that this preliminary conclusion may be worth further testing: ( i ) using higher PBO concentrations; ( ii ) in females old enough to transmit malaria , which are usually less insecticide resistant [56]–[58]; or ( iii ) in less resistant populations . Monitoring the spread of ACE-1 duplications should be an immediate priority , whereas modification of AChE-targeting insecticides to reduce sensitivity to the G119S substitution [59] , [60] represents an important longer-term goal .
Our study involved Anopheles gambiae samples for bioassays coupled with target site genotyping and copy number analysis , and two microarray experiments . The first ( Exp1; see Figure 1A , B ) compared samples from laboratory strains or field populations entirely susceptible to carbamates , with bendiocarb-resistant females from Tiassalé , which were also the subject of bioassays . Exp2 ( see Figure 1C ) involved a comparison of a population moderately resistant to bendiocarb ( Kovié ) with two fully carbamate susceptible field populations . Sample site details and resistance profiles for each population or strain used in the microarrays are given in Table S1 . For field populations , larvae were collected and provided with ground TetraMin fish food . Emerged adults were provided 10% sugar solution . All 3–5 day old females for subsequent gene expression analysis were preserved in RNALater ( Sigma ) . With the exception of a selected group from the Tiassalé population ( below ) , all samples were preserved without exposure to insecticide . The Tiassalé selected group were survivors of exposure to 0 . 1% bendiocarb ( using WHO tubes and papers ) for 360 min which induces approximately 80% mortality after 24 h ( 11 ) ; unexposed controls were held for 360 min with control paper , which did not induce mortality . All mosquitoes used in the study were identified as An . gambiae s . s . M molecular form using the SINE-PCR method [61] . The effect of the insecticide synergist piperonyl butoxide ( PBO ) , a primary action of which is to inhibit P450 monooxygenase enzymes [41] , was evaluated using WHO bioassays . Eight replicates of 25 adult female An . gambiae emerging from larvae obtained from an irrigated rice field in Tiassalé were exposed to five insecticides ( permethrin , deltamethrin , DDT , bendiocarb and fenitrothion ) . Immediately prior to each 60 min insecticide exposure , mosquitoes were exposed to 4% PBO paper for 60 min . 100 females were exposed to PBO alone as control . Chi-squared tests were used to compare the mortality with and without PBO . A TaqMan qPCR assay [62] run on an Agilent Stratagene real-time thermal cycler was used to genotype PBO-exposed samples for the ACE-1 G119S polymorphism , with qualitative calling of genotypes based on clustering in endpoint scatterplots . G119S genotype call data for samples not exposed to PBO was taken from a prior publication [11] . Following qualitative genotype calling , endpoint dR values for each dye were exported , and the data from individuals called as heterozygotes was analyzed quantitatively to investigate the possibility of sub-grouping within this genotype cluster . Specifically we tested whether surviving and dead mosquitoes , heterozygous for G119S , might possess different numbers of serine and alleles by comparing FAM ( serine label ) /VIC ( glycine label ) dye ratios using an unequal variance t-test . To further quantify the copy number variation suggested by the TaqMan genotyping results we designed a qRT-PCR to amplify fragments from three different exons of the ACE-1 gene , with normalisation ( for varying gDNA concentration among samples ) provided via comparison with amplification of a fragment from each of two single-copy genes CYP4G16 and Elongation Factor . Primer details are given in Table S6 and qRT-PCR conditions are the same as listed below for gene expression analysis . Relative copy number levels for Ace-1 were estimated relative to two pools of samples ( N = 4 each ) from the Kisumu laboratory strain by the ΔΔCT method [63] . ΔΔCT values for each test sample are the mean for the three ACE-1 amplicons following normalisation to both single copy genes and subtraction of the average normalised Kisumu values . Test samples were 16 ACE-1 G119S heterozygote survivors and 16 dead , chosen at random from those genotyped by the TaqMan assay . ΔΔCT values were compared between survivors and dead using an unequal variance t-test . Total RNA was extracted from batches of 10 mosquitoes using the Ambion RNAqueous-4PCR Kit . RNA quantity and quality was assessed using a NanoDrop spectrophotometer ( Thermo Fisher Scientific ) and a 2100 Bioanalyzer ( Agilent Technologies ) before further use . Three biological replicate extractions of total RNA from batches of 10 mosquitoes for each sample population or colony ( except Ngousso where there were N = 2 replicates ) were labelled and hybridised to Anopheles gambiae 8×15 k whole genome microarrays using previously described protocols [32] . Exp 2 employed a fully-interwoven loop design ( Figure S6 ) , optimal for study power [64] whilst , owing to the large number of comparisons and unbalanced replication , a pairwise full dye-swap design was used for Exp1 with indirect connection through the ( resistant ) Tiassalé groups ( Fig . 1 A , B ) . Exp1 was analysed using GeneSpring GX v9 . 0 software ( Agilent ) , which is readily applied to dye swap experiments , while the R program MAANOVA [65] , with LIMMA [66] for normalisation prior to ANOVA , was used to analyse the interwoven loop in Exp2 , using previously-described custom R-scripts [32] . For both experiments , the basic significance threshold for any single pairwise comparison was a q-value with false discovery rate ( FDR ) set at 0 . 05 ( i . e . an FDR-corrected threshold for multiple testing ) . Full details of the criteria applied to determine overall significance within and across Exp1 and 2 are given in Figure 1 . Within Exp1 , the direct comparison of Tiassalé bendiocarb-selected vs . Tiassalé control comparison was analysed separately and not used to determine overall significance , owing to the lower power expected for a within-population experiment involving the same level of replication as the cross-population comparisons [34] . Significantly over-expressed genes emerging from Exp1 were studied at functional level using the software DAVID Bioinformatics resources 6 . 7 [67] . Microarray data are deposited with ArrayExpress under accession numbers E-MTAB-1903 ( Exp1 ) and E-MTAB-1889 ( Exp2 ) . Quantitative real-time PCR was used to provide technical replication of results from the microarray experiments for a subset of significantly over-expressed genes . Samples were converted to cDNA using oligo ( dT ) 20 ( Invitrogen ) and Superscript III ( Invitrogen ) according to the manufacturer's instructions and purified with the QIAquick PCR Purification Kit . Three pairs of exon-spanning primers were designed for each gene of interest and from each triplicate a pair was chosen that produced a single peak from melt cure analysis , and PCR efficiency closest to 100% , determined using a cDNA dilution series obtained from a single sample . Primers details are listed in Table S7 . All qRT-PCR reactions were run on an Agilent Stratagene real-time thermal cycler and analysed using Agilent's MXPro software ( Mx3005P ) . The PCR conditions used throughout were 10 min for 95°C , 40 cycles of 10 s at 95°C and 60°C respectively , with melting curves run after each end point amplification at 1 min for 95°C , followed by 30 s increments of 1°C from 55°C to 95°C . The same RNA samples used for microarrays from Tiassalé ( selected and unexposed ) , Kovié and Okyereko plus an additional two replicates ( N = 5 for all but the Tiassalé selected group where N = 3 ) were used . Expression levels for each gene of interest were estimated relative to the Okyereko population ( chosen as the reference bendiocarb susceptible group because it was present in both microarray experiments ) by the ΔΔCT method following correction for variable PCR efficiency [63] , and normalisation using two stably-expressed genes ( Rsp7 and Elongation Factor ) ; primers and efficiencies are listed in Table S7 . Statistical significance of over-expression of each group relative to Okyereko was assessed using equal or unequal variance t-tests as appropriate , depending on results of F-tests for homoscedasticity . cDNA clones containing the open reading frames for CYP6M2 and CYP6P3 ( sequences from the An . gambiae Kisumu laboratory strain ) were PCR-amplified using high fidelity AccuPrime Pfx polymerase ( Invitrogen ) . PCR primers contained EcoRI and NotI restriction sites within the forward and reverse primers , respectively . PCR products were gel-purified using the GenElute Gel Extraction Kit ( Sigma ) and subsequently digested with the aforementioned restriction enzymes ( New England Biolabs ) . The pUAST-attB plasmid ( obtained from Dr . Konrad Basler , University of Zurich ) digested with EcoRI and NotI was gel purified , as noted above , and incubated with PCR-amplified , restriction enzyme-digested products of the CYP6M2 or CYP6P3 clone and T4 DNA ligase ( New England Biolabs ) . Ligation mixtures were transformed into competent DH5α cells , and individual colonies were verified using PCR . The EndoFree Plasmid Maxi Kit ( Qiagen ) was utilized to obtain large amounts of plasmids for subsequent steps . pUAST-attB clones containing the CYP6M2 or CYP6P3 insertion were sent to Rainbow Transgenic Flies , Inc . ( Camarillo , CA , USA ) for injection into Bloomington Stock #9750 ( y1 w1118; PBac{y+-attP-3B}VK00033 ) embryos . The PhiC31 integration system in this stock enables site-specific recombination between the integration vector ( pUAST-attB ) and a landing platform in the fly stock ( attP ) [68] . Upon receiving the injected embryos , survivors were kept at 25°C , and Go flies that eclosed were sorted by sex prior to mating . To establish families of homozygous transgenic flies , Go flies were crossed with w1118 flies , and G1 flies were sorted based on w+ eye color ( as a marker for insertion events ) . G1 w+ flies were crossed inter se to obtain homozygous insertion lines . The following D . melanogaster stocks were obtained from the Bloomington Drosophila Stock Center ( Bloomington , IN , USA ) : y1 w1; P{Act5C-GAL4}25FO1/CyO , y+ , w* ( BL4414 ) ; P{GawB}Aph-4c232 ( BL30828 ) , and w1118 ( BL3605 ) . Virgin females from CYP6M2 or CYP6P3 insertion stocks were crossed with Act5C-GAL4/CyO ( ubiquitous Actin5C driver ) flies for expression studies . For each class within a cross ( control and experimental ) , 8–10 two-day-old flies were obtained and flash-frozen in liquid nitrogen , and then stored at −80°C in triplicate . Total RNA was extracted using TRI Reagent ( Sigma ) , and 1 µg of RNA was treated with RNase-Free DNaseI ( Fisher Scientific ) . For each synthesis , a 10 µL reaction was created using 1 µL DNase-treated RNA; three technical replicates were performed for each biological replicate . Primers for amplification of cDNA product , used at a concentration of 0 . 75 µM , were: Cyp6M2_Forward: 5′-ACGAGTTCGAGCTGAAGGAT-3′ , Cyp6M2_Reverse: 5′-GTTACACTCAATGCCGAACG-3′ , Cyp6P3_Forward: 5′-TATTGCAGAGAACGGTGGAG-3′ , Cyp6P3_Reverse: 5′TACTTCCGAAGGGTTTCGTC-3′ . Relative expression was compared using Actin primers [69] at a concentration of 0 . 50 µM . qRT-PCR reactions were performed using USB VeriQuest SYBR Green One-Step qRT-PCR Master Mix ( 2X ) on a 7500 Fast Real-Time PCR System ( Applied Biosystems ) . Cycling conditions used were 50°C for 10 minutes and 95°C for 10 minutes , followed by 40 cycles of 90°C for 15 seconds and 56°C for 30 seconds , with the fluorescence measured at the end of each cycle . Recombinant CYP6M2 and CYP6P3 were commercially co-expressed with An . gambiae NADPH P450 reductase and cytochrome b5 in an E . coli system by Cypex ( Dundee , UK ) . Using previously described methodologies [35] a first experiment showed that CYP6M2 was unable to metabolise bendiocarb ( 10 µM ) after a 2 hour incubation and thus only CYP6P3 was investigated in subsequent experiments . For time course measurements , reactions were performed in 200 µL with 10 µM insecticide , 0 . 1 µM CYP6P3 membrane in 200 mM Tris-HCl pH 7 . 4 and started by adding the NADPH regenerating system ( 1 mM glucose-6-phosphate ( G6P ) , 0 . 25 mM MgCl2 , 0 . 1 mM NADP+ , and 1 U/mL glucose-6-phosphate dehydrogenase ( G6PDH ) ) . Reactions were incubated for a specified time at 30°C with 1200 rpm orbital shaking and stopped by adding 0 . 2 mL of acetonitrile . Shaking was carried for an additional 10 min before centrifuging the reactions at 20000 g for 20 min . 200 µl of supernatant was used for HPLC analysis . Reactions were performed in triplicate and compared against a negative control with no NADPH regenerating system to calculate substrate depletion . An additional experiment with different enzyme concentrations was performed , using the methods above , for 20 mins with P450 concentrations of: 0 . 2 , 0 . 1 , 0 . 075 , 0 . 05 , 0 . 025 and 0 . 0125 µM . The reactions were performed in parallel against a negative control ( −NADPH ) . In each experiment the supernatants were analyzed by reverse-phase HPLC with a 250 mm C18 column ( Acclaim 120 , Dionex ) and a mobile phase consisting of 35% acetonitrile and 65% water . The system was run at a controlled temperature of 42°C with 1 ml/min flow rate . Bendiocarb insecticide was monitored at 205 nm and quantified by measuring peak areas using OpenLab CDS ( Agilent Technologies ) . Retention time was around 14 . 9 minutes . An appropriate amount of insecticide was added to 100 µl of acetone and placed into individual 16×200 mm glass disposable culture tubes ( VWR Scientific ) . Tubes were then placed on their sides and rotated continuously , coating the entire interior of the tube , until all acetone was evaporated . A total of 8–12 control and 8–12 experimental transgenic flies , aged 3–5 days post-eclosion , were added to each tube . Flies from experimental and control classes were mixed in single insecticide-coated vials for assays , to ensure equivalent exposure to insecticide . The tubes were capped with cotton balls saturated with a 10% ( w/v ) glucose/water solution . Tubes were then incubated at 25°C for 24 h , after which mortality was assessed . Linear regression models were used to fit dose-response curves , from which LC50 values ( and confidence intervals ) were estimated using Prism v5 . 0 . However , for bendiocarb this was not possible owing to a very sharp inflection in the dose-response profile . Instead differences between lines were assessed at a diagnostic dose of 0 . 1 µg bendiocarb/vial , applied previously to Apis mellifera [37] , [70] , using Mann-Whitney U tests . | Malaria control depends heavily on only four classes of insecticide to which Anopheles mosquitoes are increasingly resistant . It is important to manage insecticide application carefully to minimise increases in resistance , for example by using different compounds in combination or rotation . Recently , mosquitoes resistant to all available insecticides have been found in Tiassalé , West Africa , which could be problematic for resistance management , particularly if common genetic mechanisms are responsible ( ‘cross-resistance’ ) . Tiassalé mosquitoes also exhibit extreme levels of resistance to the two most important classes , pyrethroids and carbamates . We investigated the genetic basis of extreme carbamate resistance and cross-resistance in Tiassalé , and the applicability of results in an additional population from Togo . We find that specific P450 enzymes are involved in both extreme and cross-resistance , including one , CYP6M2 , which can cause resistance to three insecticide classes . However , amplification of a mutated version of the gene which codes for acetycholinesterase , the target site of both the carbamate and organophosphate insecticides , also plays an important role . Mechanisms involved in both extreme resistance and cross resistance are likely to be very resilient to insecticide management practices , and represent an alarming scenario for mosquito-targeted malaria control . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"genetic",
"mutation",
"gene",
"expression",
"genetics",
"biology",
"evolutionary",
"biology"
] | 2014 | CYP6 P450 Enzymes and ACE-1 Duplication Produce Extreme and Multiple Insecticide Resistance in the Malaria Mosquito Anopheles gambiae |
Diverse classes of proteins function through large-scale conformational changes and various sophisticated computational algorithms have been proposed to enhance sampling of these macromolecular transition paths . Because such paths are curves in a high-dimensional space , it has been difficult to quantitatively compare multiple paths , a necessary prerequisite to , for instance , assess the quality of different algorithms . We introduce a method named Path Similarity Analysis ( PSA ) that enables us to quantify the similarity between two arbitrary paths and extract the atomic-scale determinants responsible for their differences . PSA utilizes the full information available in 3N-dimensional configuration space trajectories by employing the Hausdorff or Fréchet metrics ( adopted from computational geometry ) to quantify the degree of similarity between piecewise-linear curves . It thus completely avoids relying on projections into low dimensional spaces , as used in traditional approaches . To elucidate the principles of PSA , we quantified the effect of path roughness induced by thermal fluctuations using a toy model system . Using , as an example , the closed-to-open transitions of the enzyme adenylate kinase ( AdK ) in its substrate-free form , we compared a range of protein transition path-generating algorithms . Molecular dynamics-based dynamic importance sampling ( DIMS ) MD and targeted MD ( TMD ) and the purely geometric FRODA ( Framework Rigidity Optimized Dynamics Algorithm ) were tested along with seven other methods publicly available on servers , including several based on the popular elastic network model ( ENM ) . PSA with clustering revealed that paths produced by a given method are more similar to each other than to those from another method and , for instance , that the ENM-based methods produced relatively similar paths . PSA was applied to ensembles of DIMS MD and FRODA trajectories of the conformational transition of diphtheria toxin , a particularly challenging example . For the AdK transition , the new concept of a Hausdorff-pair map enabled us to extract the molecular structural determinants responsible for differences in pathways , namely a set of conserved salt bridges whose charge-charge interactions are fully modelled in DIMS MD but not in FRODA . PSA has the potential to enhance our understanding of transition path sampling methods , validate them , and to provide a new approach to analyzing conformational transitions .
Conformational transition paths are represented by sequences of ( snapshots of ) conformers in 3N-dimensional configuration space , making it difficult to examine—both visually and quantitatively—their character without resorting to dimensionality reduction in a collective variable ( CV ) space . Native contacts analysis ( NCA ) , for example , is a general approach frequently used to characterize protein folding pathways [40] and enables dimensionality reduction via a projection onto 2D native contacts ( NC ) space . NCA has the property that structural contacts are defined without reference to another structure , making NC space projections particularly useful when good reaction coordinates are not known a priori . Another common approach is principal component analysis ( PCA ) , a tool that can be used to visualize conformational dynamics in a lower-dimensional subspace spanned by several principle components ( PCs ) [41 , 42] . An important aspect of PCA is that motion along PCs can be viewed in real space , helping make complicated dynamical motions visually tractable . Using NCA , PCA or other CV approaches cannot , however , guarantee that important dynamical motions will be captured in the projections—whether ( and what ) dynamical information is lost depends on the projection itself . It is clear that a quantitative method that can examine a full 3N-dimensional trajectory would help mitigate biases inherent to selecting a coordinate projection . We propose a general computational method named Path Similarity Analysis ( PSA ) to quantitatively compare 3N-dimensional macromolecular transition paths , which is based on the idea of measuring the geometric similarity between pairs of paths using path similarity metrics . Once distances are assigned to all pairs of paths , trajectories are then clustered by similarity . The structural determinants responsible for the difference between any two trajectories are extracted at the atomic level by exploiting properties of the underlying metric . Here we introduce the PSA approach , examine its suitability , performance , and limitations as a computational approach to quantifying path similarity and apply it to a toy system and conformational transitions of two proteins . Path similarity analysis ( PSA ) exploits the properties of a ( path ) metric function , δ , that measures a distance between a pair of piecewise-linear or polygonal curves , i . e . , an ordered set of vertices connected by edges . A metric δ applied to curves A , B , C has the properties δ ( A , B ) ≥ 0 ( 1a ) δ ( A , B ) = 0 ⇔ A = B ( 1b ) δ ( A , B ) = δ ( B , A ) ( 1c ) δ ( A , C ) ≤ δ ( A , B ) + δ ( B , C ) . ( 1d ) In particular , Eq 1b , the identity property , is essential since it implies that , given two curves A and B , if B were to be continuously deformed so as to monotonically decrease the distance δ ( A , B ) , then δ ( A , B ) → 0 as B → A . That is , two curves must become identical as their mutual distance approaches zero so that decreasing values of δ correspond to increasing similarity . The other properties—non-negativity ( Eq 1a ) , commutativity ( Eq 1c ) and triangle inequality ( Eq 1d ) —guarantee that δ behaves in the same way as any other metric usually used in structural comparisons ( such as root mean squared distance ) even though it compares whole paths and not just individual conformations . PSA does not require the use of true metrics and can be used with any path distance function or other dissimilarity measure where only Eqs 1a–1c are satisfied . The triangle inequality ( Eq 1d ) , which is a generalization of the transitive property , says that when two objects , A and B , in some metric space , are each close to a third object , C , in the same space , then A is close to B in the sense that the triangle inequality , d ( A , B ) ≤ d ( A , C ) + d ( B , C ) , provides an upper bound on their distance apart . The triangle inequality is therefore important when comparing more than two objects , which is the common scenario when analyzing many conformational transitions . Although in the following we only consider true metrics , we also explore several distance functions that violate the triangle inequality in S1 Text . In the main part of this study , we consider two candidates for δ—the Hausdorff metric [43–45] and the discrete Fréchet metric [46 , 47]—and illuminate situations where one might be selected in favor of the other . Given two paths as input , both metrics locate two points , one per path , corresponding to some notion of a maximal deviation between the paths . An important property of these metrics is that they are sensitive only to path geometry; they are insensitive to dynamical motions and associated physical time scales along paths . We provide a brief overview of these two path metrics in the context of conformational transitions . To investigate the applicability of the Hausdorff and Fréchet metrics to the problem of quantifying transition paths , we generated trajectories using an abstract toy system and we simulated conformational transitions of two globular proteins , the enzyme adenylate kinase ( AdK ) in its ligand-free form and diphtheria toxin ( DT ) . The toy model was designed to gain an intuition for the path metrics and their applicability to highly fluctuating paths in high dimensions . AdK’s closed/open transition ( Fig 2A ) is a standard test case that captures general , essential features of conformational changes in proteins [12] . Alongside AdK in our analysis of transition ensembles , we also examined closed → open DT transitions ( Fig 2B ) , which serves as a more challenging example due to the difficultly of capturing the putative unfolding and refolding required for conformational change [72] . AdK is divided into three domains: the ATP-binding ( or “LID” ) domain , residues 122–159 in the mesophilic Escherichia coli sequence ( AKeco ) , and the AMP-binding ( or “NMP” or “AMPbd” ) domain , residues 30–59 , move relative to the CORE domain [73–77] around conserved hinges [78] ( Fig 2A ) . The conformational change can occur in the ligand-free ( apo ) state as demonstrated in multiple experimental studies [78–81] and corroborated by computational analyses ( reviewed by Seyler and Beckstein [12] ) . Therefore , the apo AKeco enzyme is a particularly suitable model system for studying general conformational transitions [12] . We produced transition paths between an open conformation of AdK [represented by chain A of PDB id 4AKE [77] from the Protein Data Bank [82] ( PDB ) ] , and a closed conformation ( chain A of 1AKE [83] with ligand removed ) . DT is believed to undergo a transition from an inactive closed conformation to an active open one , which includes a 180° rotation of a mobile domain [84] ( Fig 2B ) . An open conformation was captured in a domain-swapped dimeric structure [85] and compared to the closed monomeric structure [86] . DT is divided into three domains , with the receptor-binding ( R ) domain , residues 380–535 , being responsible for the majority of the opening and unrolling conformational motion about the translocation ( T ) domain , residues 179–379 , and the catalytic ( C ) domain , residues 1–178 . The conformational transition of a DT monomer was simulated previously and considered challenging for simulation methods [39 , 72] . We simulated transition pathways of DT between a closed and open conformation based on chain A from the monomeric structure ( PDB id: 1MDT [86] ) and chain A from the domain-swapped dimeric structure ( PDB id: 1DDT [85] ) , respectively .
We first describe the toy model system used to supply simple transitions for testing purposes . We then summarize the path generation—using a variety of enhanced path-sampling methods—of closed → open transitions of AdK and DT , which serve as more realistic representations of conformational transitions .
We simulated one- and eight-particle cluster transitions in the double-barrel potential energy landscape between a starting state ( defined as a center-of-mass location below z = 0nm ) and a final state ( z ≥ 4nm ) . Eight-particle simulations at zero and 250 K are shown in Fig . 5 . The particles were weakly confined to one of two potential energy barrels separated by a 2 kBT barrier at 250 K ( Fig 5A and 5D ) and evolved under the influence of thermal diffusion and drift due to a linearly decreasing ramp potential in the z direction ( Fig 5B and 5E ) . Simulations were run at temperatures between 0 K and 600 K in 50 K increments , with eight runs at each temperature . Trajectories were initialized such that two distinct groups of paths would be produced at zero temperature: for each temperature , we initialized half of the simulations to one side of the central barrier at ( x0 , y0 ) = ( 0 nm , 0 . 4 nm ) and the other half at ( 0 nm , −0 . 4 nm ) . At zero temperature , trajectories initiated at the same point progressed along identical paths due to the absence of thermal diffusion . Two trajectory groups were formed ( Fig 5A and 5B ) , consistent with what was expected from the initial conditions . A clustered heat map of the Fréchet distances between the T = 0 K trajectories clearly showed two well-defined clusters ( Fig 5C ) , containing four trajectories each , in both the structure of the dendrogram as well as the color division in the heat map . Due to thermal perturbations , higher-temperature trajectories exhibited substantial wandering ( Fig 5D and 5E ) and even produced a transition across the central barrier ( blue trajectory in Fig 5E ) . In contrast with the zero temperature case , both the number of clusters and the clusters themselves were much more vaguely defined . Two clusters with four trajectories per cluster ( red and green/blue trajectories , Fig 5D–5F ) were still formed , although the blue trajectory , which underwent a barrier-crossing transition near z = −0 . 5 nm , is an outlier in the cluster with the three green trajectories . Trajectory categorization for the toy model with PSA did not depend strongly on the dimensionality ( cluster size ) as thermal noise alone appeared to have a much more substantial influence ( S1 Fig ) . In particular , we could not discern meaningful differences in the center of mass motions between one- and eight-particle clusters from the data . Furthermore , in the eight-particle case at 250 K , performing PSA using the full ( 24-dimensional ) configuration space trajectories did not produce a different clustering than PSA applied only to the center of mass trajectories . The same analysis as above was carried out with the Hausdorff distance instead of the Fréchet distance to assess their relative discriminative powers . Both metrics produce similar results at temperatures below 300 K , each identifying two distinct pathways ( S2 Fig ) . Between 350 K and 500 K , however , Hausdorff and Fréchet distance measurements started to become substantially uncorrelated ( S3 Fig ) . This effect is likely due in part to the sensitivity of the Fréchet metric to backtracking ( Fig 1 ) , which may be amplified when the typical energy of thermal perturbations become comparable to the height of a potential barrier ( 2kBT at 300 K ) . High-temperature simulations ( ≥300 K ) began to explore both tubes as if they were a single pathway ( S2 Fig and S4 Fig ) . Taken together , PSA was able to distinguish groups of paths in the presence of stochastic thermal motions as long as the thermal energy was lower than the energy scale of distinguishing features in the underlying energy landscape . The dimensionality of the problem did not appear to be an important factor . Fréchet and Hausdorff distances discriminated paths equally well with some small differences at high temperatures that likely reflect trajectory backtracking . In order to compare a selection of fast transition path sampling methods , three distinct trajectories were generated for the closed → open AdK transition as described in Methods . We applied PSA to transition path ensembles containing hundreds of trajectories to highlight several approaches to handling the statistical nature of dynamical path-sampling methods and illustrate the portability of our analyses to other systems . Ensembles of the AdK and DT closed → open transitions were analyzed . DT was selected in part to make contact with a previous study by Farrell et al . [39] as well as provide a more challenging example to demonstrate the ease with which PSA can filter erroneous trajectories from an ensemble . We focused on two methods , DIMS MD and FRODA , because they differ fundamentally in their energetic considerations yet still share several salient features: Heavy-atom representations were used for both methods for both AdK and DT . Both methods can generate path ensembles by employing a form of stochastic dynamics , and they both drive transitions ( toward a target structure ) with similar rmsd-to-target progress variables ( DIMS uses the heavy-atom rmsd-to-target for the soft-ratcheting coordinate; FRODA attempts to gradually decrease the Cα rmsd to the target ) . Furthermore , our in-house implementations of DIMS MD methods allowed us to efficiently generate large numbers of transitions . Four unique ensembles and 800 total trajectories were generated: 200 pathways per method per protein . Details about trajectory alignment for both AdK and DT are provided in S6 Text of the Supporting Information . For AdK , transition path trajectories generated with DIMS formed one cluster that was distinct from a second cluster containing all FRODA trajectories ( see S8 Fig in the Supporting Information ) . The mean Fréchet distance 〈δF〉 between DIMS and FRODA trajectories was 2 . 9 ± 0 . 1 Å , significantly higher than the mean within the FRODA ( 2 . 2 ± 0 . 1 Å ) and DIMS ensemble ( 1 . 4 ± 0 . 2 Å ) . DIMS generated paths with smaller Fréchet distances among themselves than FRODA , while paths produced by a given method were notably more similar among themselves than when compared with paths from the other method , with no difference between Fréchet and Hausdorff distance ( S10A Fig ) . These observations imply that while FRODA produced paths that sampled a larger region of AdK’s configuration space than DIMS , each method generated a unique pathway that can be viewed as a tube in configuration space whose diameter was smaller than the typical distance between the tubes . While the AdK analysis was relatively straightforward , the DT heat map immediately revealed nine erroneous FRODA trajectories producing Fréchet distances upwards of 5 Å from any other path ( see S9 Fig for the original clustering ) . Erroneous paths were removed by specifying a distance cutoff and re-clustering using the trimmed FRODA ensemble . Visual inspection of the omitted trajectories confirmed that they either stopped short of the target or that they came somewhat near the target but continued to dramatically wander in its vicinity . All the DIMS trajectories and all of the FRODA trajectories formed two large , separate clusters ( Fig 8 ) . It is not immediately obvious how one would tune one particular path-generating algorithm to increase its likelihood to produce a path characteristic of another algorithm , although we already observed that the variation of the pulling speed in the rTMD method led to quantitatively ( Fig 6 ) and qualitatively ( Fig 7 ) different paths . In particular , fast pulling ( rTMD-F ) generated paths similar to linear interpolation ( LinInt ) , whereas slow pulling ( rTMD-S ) paths were more similar to DIMS and MDdMD . Thus , although we do not yet understand the general relationship between the pathways sampled by different algorithms , PSA appears to be a useful tool to tackle this question . The ultimate goal is , of course , to find a method that reliably samples transitions realized in the real system . The analyses presented here should also aid in identifying any overlap between different sampling methods and experimental data ( e . g . from femtosecond structural biology experiments ) when such data become available . PSA is a general approach that can operate on the full 3N-dimensional trajectories without requiring any system-specific knowledge . It provides a very broad means to categorize transitions as distinct from one other . But as described so far , it is difficult to relate the global PSA analysis to physically relevant differences at the molecular level . To address this question we introduce the new concept of “Hausdorff pairs” ( or “Fréchet pairs” ) that allows us to pinpoint conformations that may be more likely to exhibit geometric ( structural ) features relevant to conformational change . By construction , the Hausdorff and the Fréchet distances identify a point-wise distance between two particular conformers , one on each path , as the global distance between the paths . The path metrics therefore induce a map between a conformer on one path to a conformer on another whose separation distance is , in some sense , a maximal deviation between the paths . We term such a pair of conformers a Hausdorff pair ( δH-pair ) or a Fréchet pair ( δF-pair ) . These conformers can be examined at the molecular or atomic level to reveal the specific structural discrepancies that give rise to large deviations in configuration space between pairs of paths . As an explicit example , we identified three Hausdorff pairs for the DIMS and FRODA closed → open AdK transition ensembles and projected them in AA space ( Fig 9A ) . We first segregated the full set of Hausdorff distance measurements into: ( 1 ) mutual distances among DIMS paths , ( 2 ) mutual distances among FRODA paths , and ( 3 ) inter-method distances measured between a DIMS and a FRODA path . A total of N ( N−1 ) /2 = 79800 δH-pairs were identified for the ensemble of N = 400 paths . In order to present representative data for the whole ensemble , we identified the two δH-pairs associated with the median and maximum Hausdorff distances for each comparison ( 1 ) , ( 2 ) , and ( 3 ) as defined above . As a typical example we explicitly examined the median δH-pair identified for the inter-method comparisons and projected the atomic displacements onto each structure to locate regions of large deviation ( see structures in Fig 9A ) . It became apparent that the NMP domain in the DIMS structure was closer to the LID domain because a number of evolutionary conserved salt bridges ( D33–R156 , R36–D158 , D54–K157 ) persisted late into the transition due to the strong electrostatic interaction between the acidic and basic moieties [99] ( Fig 9B ) . FRODA , on the other hand , operating on purely geometric principles and neglecting Coulomb interactions , does not account for the influence of salt bridges on the transition and the associated δH-pair structure exhibited broken salt bridges in the inter-LID/NMP region ( Fig 9C ) . It is therefore not surprising that the FRODA trajectory did not show the “salt-bridge zipper” [99] , which manifested itself as discerning difference between the DIMS and FRODA trajectories . With salt bridges located across the NMP domain but primarily on the side of the LID , the LID is relatively free to move to an open configuration , whereas the NMP domain is prevented from fully opening until the salt bridges are broken . These considerations are consistent with the tendency of DIMS paths to primarily favor a LID-opening pathway ( Fig 9A , blue circles ) , while FRODA paths ( Fig 9A , green circles ) sampled the region around LinInt ( Fig 9A , black dashed line ) corresponding to simultaneous LID/NMP-opening . A Hausdorff pair describes the two frames at which the two trajectories in question differ most . Additionally , the regions where trajectories differ to varying degrees from each other might also be of interest . This kind of information is provided by the set of nearest neighbor distances along a path . Eq 3 defines the nearest neighbor distance of point pk on path P from path Q as δh ( k; P ∣ Q ) ≔ δh ( pk ∣ Q ) ≔ minq ∈ Q d ( pk , q ) and the nearest neighbor distance of point qk on path Q from path P as δh ( k; Q ∣ P ) . In general , these two distances are not symmetric , i . e . δh ( k; P ∣ Q ) ≠ δh ( j; Q ∣ P ) for any conformations j , k . When δh ( k ( ξ ) ; P ∣ Q ) and δh ( j ( ξ ) ; Q ∣ P ) are plotted against a suitable common order parameter ξ , the regions of large and small differences between trajectories can be quantified . For example , in S11 Fig , the nearest neighbor distances of the three pairs of trajectories corresponding to the median Hausdorff pairs in Fig 9A showed that the DIMS and FRODA trajectories primarily differed in the first ∼ 60% of the transition , which corresponds to LID-opening in DIMS and simultaneous LID/NMP-opening in FRODA . The DIMS trajectories differed almost uniformly along the whole path by only ⪅ 1 . 3 Å , suggesting that they follow a similar path perturbed by thermal fluctuations . The FRODA trajectories differed by ∼ 2 Å during the middle half of the transition but practically coincided at beginning and end , showing that FRODA can accurately connect two given endpoint structures even with its stochastic component enabled . The Hausdorff-pair and nearest neighbor distance analysis naturally followed from the formulation of PSA . Even though only Cα atoms were used to distinguish DIMS from FRODA trajectories hence the level of detail of PSA was primarily restricted to conformational differences in the protein backbone , atomic-scale analysis of Hausdorff-pairs was able to reveal the molecular determinants responsible for the structural differences . | Many proteins are nanomachines that perform mechanical or chemical work by changing their three-dimensional shape and cycle between multiple conformational states . Computer simulations of such conformational transitions provide mechanistic insights into protein function but such simulations have been challenging . In particular , it is not clear how to quantitatively compare current simulation methods or to assess their accuracy . To that end , we present a general and flexible computational framework for quantifying transition paths—by measuring mutual geometric similarity—that , compared with existing approaches , requires minimal a-priori assumptions and can take advantage of full atomic detail alongside heuristic information derived from intuition . Using our Path Similarity Analysis ( PSA ) framework in parallel with several existing quantitative approaches , we examine transitions generated for a toy model of a transition and two biological systems , the enzyme adenylate kinase and diphtheria toxin . Our results show that PSA enables the quantitative comparison of different path sampling methods and aids the identification of potentially important atomistic motions by exploiting geometric information in transition paths . The method has the potential to enhance our understanding of transition path sampling methods , validate them , and to provide a new approach to analyzing macromolecular conformational transitions . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] | [
"distance",
"measurement",
"crystal",
"structure",
"molecular",
"dynamics",
"engineering",
"and",
"technology",
"applied",
"mathematics",
"condensed",
"matter",
"physics",
"simulation",
"and",
"modeling",
"algorithms",
"interpolation",
"mathematics",
"network",
"analysis",
... | 2015 | Path Similarity Analysis: A Method for Quantifying Macromolecular Pathways |
Effectors of the bacterial type III secretion system provide invaluable molecular probes to elucidate the molecular mechanisms of plant immunity and pathogen virulence . In this report , we focus on the AvrBs2 effector protein from the bacterial pathogen Xanthomonas euvesicatoria ( Xe ) , the causal agent of bacterial spot disease of tomato and pepper . Employing homology-based structural analysis , we generate a three-dimensional structural model for the AvrBs2 protein and identify catalytic sites in its putative glycerolphosphodiesterase domain ( GDE ) . We demonstrate that the identified catalytic region of AvrBs2 was able to functionally replace the GDE catalytic site of the bacterial glycerophosphodiesterase BhGlpQ cloned from Borrelia hermsii and is required for AvrBs2 virulence . Mutations in the GDE catalytic domain did not disrupt the recognition of AvrBs2 by the cognate plant resistance gene Bs2 . In addition , AvrBs2 activation of Bs2 suppressed subsequent delivery of other Xanthomonas type III effectors into the host plant cells . Investigation of the mechanism underlying this modulation of the type III secretion system may offer new strategies to generate broad-spectrum resistance to bacterial pathogens .
Plants have evolved sophisticated innate immune systems to counter the attack of various microbial pathogens through a combination of diverse molecular mechanisms [1] . Plant innate immunity is controlled by two overlapping signaling pathways . The first pathway , PAMP-Triggered Immunity ( PTI ) , is a basal defense response that is triggered by the recognition of pathogen-associated molecular patterns ( PAMPs ) through a set of specialized plant extracellular receptor kinase proteins [2]–[5] . Plants use PTI to suppress the growth of non-pathogens . However , successful bacterial pathogens can interfere with PTI via effector proteins that are delivered into plant cells through the type three secretion and translocation system ( TTSS ) . Many bacterial TTSS effectors have identified virulence functions that modulate the pathways involved in PTI , making the plants more susceptible to the proliferation of microbial pathogens [1] . Most of these TTSS effector proteins are not homologous , and the majority have no obvious biochemical function , although a few have been shown to have enzymatic activity [6]–[9] . Characterizing the biochemical functions of pathogen effectors and identifying the plant targets of each effector will shed light on bacterial pathogenesis and plant immunity . In response to effector proteins , plants have evolved a second layer of defense signaling pathways controlled by resistance genes ( R genes ) . The plant R proteins directly or indirectly recognize the bacterial TTSS effectors and initiate effector-triggered immunity ( ETI ) [10] . This response is often a localized , programmed cell death-related defense response , also known as the hypersensitive reaction ( HR ) [11] . Despite intensive study of the molecular mechanisms of PTI and ETI , the interplay between these two primary defense mechanisms remains elusive [12] , [13] . The TTSS machinery of phytopathogenic bacteria encoded by the clustered hrp ( hypersensitive reaction and pathogenicity ) genes is essential for the delivery of effectors to the interior of the plant cell [14] . Mutations in the pathogen that block the TTSS will subsequently prevent the translocation of the type III effectors and impair the virulence of the pathogen on host plants [14]–[16] . Therefore , the TTSS plays a critical role in bacterial pathogenesis . The translocation of TTSS effectors can be quantitatively measured by monitoring adenylate cyclase enzyme activity in plant cells by fusing the effector protein with the calmodulin-dependent adenylate cyclase domain ( Cya ) of Bordetella pertussis cyclolysin [17] , [18] . Despite intensive characterization of the TTSS in model bacterial pathogens , including several Pseudomonas and Xanthomonas species , detailed information describing the establishment and regulation of the TTSS is still missing . It is also not clear if plants have evolved defense mechanisms that can recognize the establishment of bacterial TTSS . However , a recent report demonstrated that PTI of the host plant can inhibit the injection of bacterial type III effectors [19] , suggesting that the suppression of TTSS may contribute to the plant immunity . Xanthomonas euvesicatoria ( Xe ) is the causal agent of bacterial leaf spot disease of pepper and tomato , which can deliver more than 28 TTSS effectors into plant cells [20] , [21] . One type III effector AvrBs2 is highly conserved not only in Xe strains but also in many other Xanthomonas pathovars that cause disease in a wide range of crops [22] , [23] . The presence of avrBs2 in many of these pathogens makes a significant contribution toward their virulence [22] . Previous analyses have determined that the avrBs2 gene encodes a protein containing a domain homologous to the E . coli glycerolphosphodiesterase ( GDE ) and the agrocinopine synthase ( ACS ) of Agrobacterium tumefaciens . However , it has not been shown whether AvrBs2 possesses GDE or ACS enzyme activity and whether such activity is relevant to AvrBs2 function [23] , [24] . Pepper plants ( Capsicum annuum ) carrying the bacterial leaf spot disease resistance gene ( Bs2 ) are resistant to strains of Xe that contain AvrBs2 . This host-pathogen interaction results in a resistance response that inhibits the growth of Xe [22]–[25] . The Bs2 gene has been isolated by map-based cloning and encodes a protein that belongs to the largest class of plant disease resistance proteins . The protein contains a central putative nucleotide-binding site ( NBS ) and a carboxyl-terminal leucine-rich repeat ( LRR ) region [25] . Bs2 has been shown to associate with the molecular chaperone SGT1 through its LRR domain to specifically recognize AvrBs2 and trigger the HR in plants [26] . However , it is still not clear whether Bs2 recognizes AvrBs2 directly or indirectly in planta . In addition to the Bs2 gene , two other pepper resistance genes , Bs1 and Bs3 , have been identified that confer resistance to Xe strains carrying the avrBs1 and avrBs3 effector genes , respectively [27] . Near-isogenic lines carrying the Bs1 , Bs2 , and Bs3 genes have been generated by introgression of individual or combinations of Bs genes into the susceptible pepper cultivar Early Cal Wonder ( ECW ) [28] , [29] . The avrBs1 and avrBs3 genes have also been identified and cloned [23] , [30]–[32] . The Bs1 gene has not been cloned [32] , but Bs3 , which encodes a flavin monooxygenase enzyme , has recently been isolated from the pepper genome [33] . In this study , the pepper and Xe pathosystem is used to study the interaction between Bs2 and AvrBs2 . We demonstrate that the catalytic sites of the putative GDE domain of AvrBs2 are under purifying selection , and that the GDE catalytic sites are required for AvrBs2 virulence function but not the activation of Bs2 . Although we were unable to demonstrate the GDE enzymatic activity using purified , full-length AvBs2 , we determine that the AvrBs2 GDE catalytic site could functionally replace the GDE catalytic site of BhGlpQ ( Borrelia hermsii ) [34] . We also identify a minimum domain of AvrBs2 that included the GDE homologous region and a carboxyl Bs2 activation domain . Therefore , we are able to genetically separate the virulence function of AvrBs2 , which is dependent on its GDE catalytic site , from the Bs2 activation , which is independent of the GDE catalytic site . Finally , we describe a novel plant disease resistance phenotype related to the AvrBs2/Bs2 host-pathogen interaction . When AvrBs2 activates the Bs2 R gene function , the TTSS is reduced in the delivery of effectors to the plant host . Investigation of the mechanism of the AvrBs2 virulence function and TTSS suppression during its recognition by Bs2 could offer new strategies to generate broad-spectrum resistance to the Xe bacterial pathogen .
Previous characterization of AvrBs2 ( YP 361783 ) from Xe revealed a domain [amino acids ( aa ) 280 to 340] with homology to a bacterial GDE [23] . To further characterize this Xe AvrBs2 domain , we searched the current GenBank database with the BLASTP program using the full-length AvrBs2 protein as a query . This search allowed us to compile remote homologs from plants , animals , fungi , and bacteria that contain GDE domains homologous to AvrBs2 . In Figure 1A , selected GDE ( or putative GDE ) proteins from plants [AtGDE ( NP_177561 ) ] and OsGDE [ ( AP003274 ) ] , human [HsMIR16 ( NP_057725 ) ] , fungi [ScGDE1 ( NP_015215 ) ] , and bacteria [TmGDPD ( TM1621 ) of Thermotoga maritima , BhGlpQ ( ADD63790 ) from Borrela hermsii , and AgtACS ( AAO15364 ) from Agrobacterium tumefaciens] aligned with the GDE domain of AvrBs2 ( aa 274 to 328 ) are shown . Several AvrBs2 homologs from Xanthomonas pathogens of tomato , euvesicatoria ( Xe ) ( YP_361783 ) ; alfalfa , campestris pv . alfalfae ( Xca ) citrus , axonopodis pv . citri ( Xac ) ( NP_640432 ) ; cabbage , campestris pv . campestris ( Xcc ) ( NP_635447 ) ; and rice , oryzae pv . oryzae ( Xoo ) ( YP_449177 ) or oryzae pv . oryzicola ( Xoc ) ( ZP_02241238 ) were included in the alignment . The overall sequence identity between AvrBs2 and the different GDEs in this region was approximately 33% ( with >37% sequence similarity ) ( Figure 1A ) [35] . The putative GDE domain in AvrBs2 aligned well with the glycerophosphodiester phosphodiesterase ( GdPd ) protein from Thermotoga maritima , for which the three-dimensional crystal structure had been previously determined ( PDB ID: 1O1Z ) [36] . The GDE domains of AvrBs2 and TmGdpd share 60% amino acid sequence similarity and 47% identity . The high amino acid sequence similarity between the GDE domains of AvrBs2 and TmGdpd predicts that these two proteins will have similar three-dimensional structures . A homology-based modeling method was employed to generate a three-dimensional structural model for AvrBs2 ( aa 274 to 328 ) using the solved crystal structure of TmGdpd as a template [36] , [37] . The resulting three-dimensional structural model of AvrBs2 closely matched the solved crystal structure of Tm 1o1z A ( Figure 1B ) . Both structures consist of two antiparallel beta-sheets capped by nine putative alpha-helices . Recently , GDE enzyme activity and the putative catalytic sites of the human GDE ( HsMIR16 ) have been characterized [38] , [39] . Point mutations in the GDE catalytic sites ( E97A , D99A , and H112A ) in HsMIR16 eliminated GDE enzyme activity [38] , [39] . The putative catalytic sites of HsMIR16 are conserved in all of the GDE homologs , including the six AvrBs2 homologs ( Figure 1A ) . In the three-dimensional structural model of AvrBs2 , the catalytic sites are present in regions of high structural homology between the two proteins ( TmGdpd in blue and AvrBs2 in red ) , which suggests that AvrBs2 utilizes the same residues for enzymatic function ( Figure 1B ) . To investigate whether the AvrBs2 protein possesses GDE enzyme activity , both the wild type and the catalytic mutants of avrBs2 were expressed in E . coli as GST-AvrBs2 fusion proteins . The fusion proteins were assayed for GDE enzyme activity using a method that was originally adapted for E . coli and Borrelia GDEs , with glycerophosphocholine as a substrate [40] , [41] . However , we were unable to detect GDE enzyme activity of AvrBs2 with this substrate . Because the GDE catalytic sites of the BhGlpQ enzyme were conserved with predicted catalytic sites in AvrBs2 ( Figure 1A ) , we hypothesized that if we replaced the core GDE catalytic site of the active BhGlpQ enzyme [41] ( 24 amino acids ) with the putative GDE catalytic site of AvrBs2 , we might be able to detect enzyme activity with glycerophosphocholine substrate in vitro . To test this possibility , the GDE catalytic site of BhGlpQ was replaced with either the wild-type AvrBs2 catalytic site or a GDE catalytic site mutant ( E304A/D306A ) ( Figure 1C ) . The GDE enzyme activities of purified GST:BhGlpQ ( positive control ) , GST:BhGlpQ-AvrBs2-WT , and GST:BhGlpQ-AvrBs2-E304A/D306A were analyzed using an indirect coupled enzyme assay [41] . The higher light absorbances at 340 nm for GST:BhGlpQ ( positive control ) and GST:BhGlpQ-AvrBs2-WT compared to the inactive GST:BhGlpQ-AvrBs2-E304A/D306A indicated that AvrBs2 had a functional GDE catalytic site ( Figure 1C and 1D ) . To test whether the GDE catalytic site of AvrBs2 is important for Xe virulence in susceptible bs2 plants or for Bs2 disease resistance activation , we mutated the GDE catalytic sites E304A , D306A and H319A by site-directed mutagenesis of the wild-type avrBs2 gene ( Figure 2A ) . We replaced the chromosomal copy of avrBs2 in strain Xe GM98-38-1 with various avrBs2 mutants by homologous recombination . The effects of these mutations on AvrBs2 virulence function and/or Bs2-activation were evaluated by in planta bacterial growth assays in near-isogenic pepper and tomato lines with and without the R gene Bs2 ( Figure 2B ) . In pepper and tomato lines without Bs2 , the Xe strain with wild-type avrBs2 was more virulent and grew approximately five-fold higher than the null strain Xe without avrBs2 ( Figure 2B ) . The Xe strains with mutations in GDE domain ( E304A/D306A and H319A ) lost AvrBs2 virulence function and were similar to the null strain Xe without avrBs2 ( Figure 2B ) . However , on near-isogenic pepper and transgenic tomato lines with Bs2 [25] , Xe strains carrying the AvrBs2 GDE mutants were still able to activate Bs2-based resistance , similar to the Xe strain carrying wild-type avrBs2 ( Figure 2B ) . These results demonstrate that the putative GDE catalytic sites of avrBs2 are required for its virulence function but not for recognition by Bs2 . Additionally , we tested two control Xe strains that contain point mutations ( R403P and A410E ) [24] that evade Bs2 activation while maintaining most of the virulence functions of AvrBs2 ( Figure 2A ) . Similar to previously reported results in pepper plants without Bs2 [24] , these mutants were intermediate in virulence between Xe carrying wild-type avrBs2 and Xe without avrBs2 . However , the mutants were unable to activate Bs2 resistance in pepper plants containing Bs2 ( Figure 2B ) . Another method for assaying the induction of plant immunity is to challenge a plant with a high-density bacterial dose that triggers a macroscopic hypersensitive cell death reaction , or HR response . High-density inoculations ( 2×108 CFU/ml ) of pepper with Bs2 caused a similar , strong brown necrosis with the Xe strain with wild-type avrBs2 and the Xe strains with avrBs2 GDE mutations ( E304A/D306A and H319A ) ( Supplemental Figure S1 ) . However , high-density inoculations of pepper plants containing Bs2 with the Xe avrBs2 mutant strain ( A410E ) caused a light brown necrosis , suggesting that this mutant maintained a low level of Bs2 activation capability ( Supplemental Figure S1 ) , similar to previously reported [24] . To test whether the GDE mutations had a negative effect on AvrBs2 delivery by Xe TTSS , the TTSS effector delivery reporter Cya [18] was utilized to quantitatively measure the translocation of two different AvrBs2 GDE mutant Xe effectors . The AvrBs2 GDE mutations caused no reduction of detectable effector delivery ( Supplemental Figure S2A ) . Additionally , the Xe ( avrBs2-Cya ) wild type and catalytic site mutant strains were not altered from the non-Cya strains in the activation Bs2 HR ( Supplemental Figure S2B ) . Demonstrating that the GDE domain of AvrBs2 is required for virulence prompted us to evaluate the natural variations in various avrBs2 alleles with respect to the evolutionary selection . In addition to the previously published avrBs2 homologs [ ( Xe in pepper ( YP_361783 ) , Xca in alfalfa and Xcc in cabbage ( NP_635447 ) ] [23] , three additional uncharacterized homologs of avrBs2 ( Xanthomonas axonopodis pv . citri [Xac] ( NP_640432 ) , Xanthomonas oryzae pv . oryzae [Xoo] ( YP_449177 ) , and Xanthomonas oryzae pv . oryzicola [Xoc] ( ZP_02241238 ) from newly released genome sequences were aligned using the CLUSTALW program [35] . The overall sequence identity of the different avrBs2 homologs in Xanthomonas was high ( >70% ) . Phylogenetic analysis by maximum likelihood ( PAML ) software was used to determine which evolutionary model acts on these six homologs of avrBs2 from different Xanthomonas pathovars that have adapted to cause disease in different host plant species [42] . This statistical analysis of nucleotide changes with respect to amino acid changes calculated an average rate of non-synonymous ( KA ) and synonymous ( Ks ) substitutions per site for all six avrBs2 homologs . The ratio ( ω ) = KA/Ks measures the difference between the two rates . For neutral amino acid changes or neutral selection , the ω ratio is 1 . 0 . For advantageous amino acid changes or adaptive selection , the ω ratio is >1 . 0 , and for deleterious amino acid changes or purifying selection , the ω ratio is <1 . 0 [42] , [43] . The average ω ratio over all six homologs was estimated to be 0 . 1534 , indicating a strong purifying selection on the Xanthomonas pathovars to maintain avrBs2 for its contribution to pathogenic virulence in a range of different host plant species . In addition , PAML analysis revealed a significant variation in the ω ratio over the length of the avrBs2 sequence . Sliding window analysis using the SWAKK program [43] was used to determine the distribution of variation in the ω ratio across avrBs2 from Xe and Xcc . The low ω over the GDE-virulence region is consistent with purifying selection to maintain the virulence function of avrBs2 ( Figure 2C ) . Although the ω for the TTSS signal peptide remained below one , there was an increase in ω in this region , possibly associated with differences in TTSS effector delivery for specific Xanthomonas pathovars as they infect different host plants ( Figure 2C ) . Having established that the GDE catalytic sites are required for AvrBs2 virulence function but not Bs2-activation , we generated additional deletions of the N-terminus of AvrBs2 to define a minimal region required for Bs2 activation . The deletions were cloned into a binary vector and screened for HR in stable transgenic Bs2 Nicotiana benthamiana using Agrobacterium-mediated transient expression ( Figure 3A ) . The previously reported [44] avrBs2 deletion construct ( aa 97 to 520 ) was still able to trigger a Bs2 HR; the N-terminal deletion ( aa 271 to 520 ) produced a similar result ( Figure 3A and 3B ) . Further deletions at either the amino or the carboxyl terminus of the minimal domain failed to elicit a Bs2-dependent HR . Thus , the fragment ( aa 271 to 520 ) was the minimal region required for Bs2 activation . Interestingly , the minimal Bs2 recognition region included the GDE domain , although an active catalytic site was not required for Bs2 activation . We confirmed the Agrobacterium-mediated transient expression HR response of these AvrBs2 mutants on Bs2 pepper ( Supplemental Figure S3B ) . Also , we detected similar protein expression for all clones using C-terminal HA epitope tags and immunoblot analysis ( Figure S3A ) . The previously identified AvrBs2 loss-of-Bs2-recognition mutations ( R403P and A410E ) [24] are within the minimal Bs2 activation domain but are C-terminal to the GDE homologous region . To identify other residues in AvrBs2 near the point mutations of R403P and A410E that might play a role in Bs2 activation , a collection of randomly selected single amino acid mutations in the C-terminal region of the minimal Bs2 activation domain was generated . These fragments were cloned into the same binary vector used for the deletion constructs and used in Agrobacterium transient expression experiments . We identified one additional point mutant ( Y419A ) that had lost the ability to trigger HR ( Figure 3A and 3C ) . In the AvrBs2 three-dimensional structural model ( Figure 1B ) , the Y419A mutation and the two other mutations ( R403P and A410E ) that also disrupt AvrBs2 activation of Bs2 are located on the loops that do not closely align with the solved crystal structure template ( 1O1Z ) . In Supplemental Figure S3A and S3B we confirm the Agrobacterium-mediated transient expression HR response of these AvrBs2 mutants on Bs2 pepper and confirm protein expression . To further evaluate the role of Y419A , we replaced the wild type avrBs2 allele of Xe with the Y419A mutant by double homologous recombination . The effects of Y419A on AvrBs2 virulence and/or Bs2-activation were evaluated by in planta bacterial growth assays ( Supplemental Figure S4A ) . On Bs2 pepper the Xe Y419A mutant strain was intermediate between Xe carrying wild-type avrBs2 and Xe without avrBs2 . High-density inoculations of pepper plants containing Bs2 with the Xe avrBs2 mutant Y419A caused a light brown necrosis , suggesting that this mutant maintained a low level of Bs2 activation ( Supplemental Figure S4B ) similar to the Xe mutant A410E ( Supplemental Figure S1 ) . This deletion analysis defined a minimal Bs2 activation domain that included the GDE region , but did not require an active GDE catalytic site . The results of the mutagenesis assays suggest that the critical amino acids for Bs2 recognition are located near the C-terminal end of the minimal Bs2-activation domain . Therefore , the general AvrBs2 structure but not the putative GDE enzymatic activity , was required for Bs2 activation . It has long been known that cognate effector/R protein interactions result in a hypersensitive reaction that is specified by the interacting gene pairs . The intensity and the color of the collapsing host tissue and the timing of cell death are specific to the interacting gene pairs . The activation of HR by AvrBs2/Bs2 interactions is slow; macroscopic cell death symptoms appear at 48 hours post-infection ( hpi ) . The Xanthomonas effector AvrBs1 activates a rapid Bs1-dependent HR visible at 18 hpi [30] . When we inoculated the Xe ( avrBs2 , avrBs1 ) strain delivering both AvrBs1 and AvrBs2 into a pepper line containing both Bs1 and Bs2 R genes , we observed that AvrBs2 activation of a slower Bs2-HR was epistatic to the AvrBs1 activation of a more rapid Bs1-HR ( Figure 4 ) . Control strains Xe ( avrBs1 ) and Xe ( avrBs2 ) along with control pepper ( Bs1 ) and pepper ( Bs2 ) were included for comparison to detect the epistatic , slow Bs2-HR at 48 hpi instead of the expected faster Bs1-HR at 18 hpi ( Figure 4 ) . The epistasis of the Xe activated slower Bs2 HR over the Xe activated faster Bs1 HR was also confirmed by measuring electrolyte leakage ( Supplemental Figure S5A and S5B ) . To test whether the Bs2 activation dependent suppression of the AvrBs1/Bs1 fast HR phenotype could be activated in trans , we co-inoculated a mixed inoculum of two strains of Xe containing either avrBs1 , avrBs2 or no effector onto pepper ( Bs1 , Bs2 ) . Again we observed the Bs2 activation dependent suppression of the AvrBs1/Bs1 fast HR phenotype ( Supplemental Figure S6A ) . Control inoculations with single Xe effectors , either by individual or mixtures , gave the expected responses on pepper plants with and without the corresponding R gene ( Supplemental Figure S6 ) . Additionally , the epistasis of the Xe activated slower Bs2 HR over the Xe activated faster Bs1 HR was again confirmed by measuring electrolyte leakage ( Supplemental Figure S7A ) . We hypothesized that this suppression might be accounted for by one of the following: ( i ) Bs2 activation disrupts Bs1 activation or ( ii ) Bs2 activation disrupts TTSS-mediated translocation of AvrBs1 or ( iii ) Bs2 activation causes a reduction or loss of induction of AvrBs1 . To test the first hypothesis , three Agrobacterium strains containing either 35S-avrBs1 , 35S-avrBs2 alone or a 35S-avrBs1/35S-avrBs2 tandem construct were inoculated on pepper containing both the Bs1 and Bs2 R genes . If Bs2 activation disrupts Bs1 activation , then suppression of AvrBs1/Bs1-dependent HR should occur . However , we did not observe alteration of the fast , Bs1 HR by the slow Bs2 HR activation when both effectors were transiently expressed ( Supplemental Figure S8A ) . The fast Bs1 HR for the co-expressed AvrBs2 and AvrBs1 on pepper ( Bs2 , Bs1 ) was confirmed by measuring electrolyte leakage ( Supplemental Figure S8B ) . In addition , immunoblot analysis detected similar levels of expression for both HA epitope tagged effectors after 24 hours ( Supplemental Figure S8C ) . Therefore , when AvrBs1 and AvrBs2 were simultaneously expressed in plant cells , the Bs2/AvrBs2-dependent HR no longer suppressed the Bs1/AvrBs1-dependent HR . This finding is not consistent with the first hypothesis . To test our second hypothesis , whether Bs2 activation modulates subsequent Xe TTSS effector delivery , the TTSS effector delivery reporter Cya [18] was utilized to quantitatively measure the translocation of two different Xe effector-reporters for avrBs1 and xopX . In this assay , the type three secretion and translocation signal peptides for each effector were translationally fused to the reporter Cya . Using homologous recombination , the reporters were marker-exchanged in tandem with the corresponding chromosomal allele of different Xe strains so that the wild-type copy of the particular effector was also maintained [18] . Pairs of effector-Cya reporter strains with and without avrBs2 included the pair of strains Xe ( avrBs1 ) and Xe ( avrBs1 , avrBs2 ) with either AvrBs11-212-Cya reporter ( Figure 5A ) or XopX1-183-Cya reporter ( Figure 5B ) . Pairs of Xe Cya reporter strains , with and without avrBs2 , were inoculated on pepper ( no R genes ) , pepper ( Bs2 ) and pepper ( Bs1 ) . Plants were sampled eight hours post-inoculation to avoid in planta multiplication of the reporter strains [18] . Eight hours post-inoculation is also before visible R gene-mediated HR . Because each effector-Cya reporter construct has a unique rate of translocation , each reporter construct was evaluated separately . When the translocation of AvrBs1 and XopX Cya reporters was assessed in the presence of Bs2/avrBs2 , the detectable levels of cyclic AMP for both effector-Cya reporters were significantly reduced in comparison to all other combinations where Bs2 was not activated including the Bs1/AvrBs1 interaction ( Figure 5A , 5B ) . Additionally , we tested three other pairs of effector-Cya reporter strains with and without avrBs2 that included the pair of strains Xe ( avrBs3 ) and Xe ( avrBs3 , avrBs2 ) with either AvrBs21-212-Cya reporter , AvrBs31-212-Cya reporter or XopX1-183-Cya reporter ( Supplemental Figure S9 ) . Again only Bs2 activation was associated with reduced levels of effector-Cya reporter delivery to the host . This is consistent with the hypothesis that the Bs2 activation disrupts general TTSS-mediated translocation of effectors . To preclude the possibility that Bs2 activation might block calmodulin dependent Cya elevation of in planta cyclic AMP levels , we tested Agrobacterium transient expression of 35S-AvrBs2:Cya in the presence and absence of Bs2 at 15 hpi in N . benthamiana . Similar elevated levels of cyclic AMP were observed in the presence and absence of Bs2 activation ( Supplemental Figure S10A ) . Additionally , we evaluated the effect of the GDE catalytic site mutations in AvrBs2 on the TTSS disruption by Bs2 activation with the AvrBs3-Cya reporter Xe strain . The set of four effector-Cya reporter Xe strains ( avrBs2 , avrBs2-E304A/D306A , avrBs2-H319A and without avrBs2 ) with the AvrBs31-212-Cya reporter were tested on pepper with or without Bs2 . The loss of the GDE catalytic sites in AvrBs2 did not alter the TTSS repression effect of the Bs2/AvrBs2 interaction ( Supplemental Figure S10B ) . To preclude the possibility that Bs2 activation causes a reduction or loss of induction of TTSS effectors in Xe , AvrBs2-Cya , an effector that is also disrupted in delivery to the host by Bs2 activation ( Supplemental Figure S9A ) , was tested for reduction in protein level . Immunoblot assays of high titer inoculation of pepper ( w/o Bs2 ) and pepper ( Bs2 ) with Xe ( avrBs2 ) , Xe ( avrBs2-Cya ) , Xe ( avrBs2-E304A/D306A:Cya ) and Xe ( avrBs2-H319A:Cya ) detected no reductions of protein levels associated with Bs2 activation ( Supplemental Figure S10C ) . Although these results do not support hypothesis ( iii ) as a broad mechanism targeting all TTSS effectors it does not preclude an AvrBs1 specific targeting for degradation or loss of induction by Bs2 activation . While both 35SAvrBs2:HA and 35S-AvrBs1:HA transiently expressed in pepper were detected in immunoblot analysis we were only able to detect Xe expressed AvrBs2:HA but not AvrBs1:HA ( data not shown ) . Low Xe expression of AvrBs1 may contribute to the overall low levels of TTSS delivered AvrBs1-Cya reporter compared to all other effector-Cya reporters evaluated . There is also a Bs2 activation specific reduction in the detectable Xe delivered AvrBs1-Cya reporter that should correlate with a Bs2 activation specific reduction in the Xe delivered AvrBs1 . This indirect evidence is all consistent with a Bs2 activation dependent reduction in TTSS delivery of an already lowly expressed AvrBs1 resulting in a lack of the minimal amount of AvrBs1 required to activate a confluent Bs1 HR . These results led us to conclude that plant cells undergoing a Bs2/AvrBs2 incompatible reaction were able to modulate subsequent effector delivery by the Xe TTSS .
Several classes of bacterial TTSS effectors have been characterized based on their enzymatic activities targeting host proteins [6]–[9] . In this study , we identified a GDE domain present in AvrBs2 that is highly conserved in homologs from several species of Xanthomonas . In addition to generating a three-dimensional structural model of the GDE domain of AvrBs2 using the crystal structure of a bacterial GDE , we demonstrated that the putative GDE catalytic site of AvrBs2 could functionally replace the catalytic site of the bacterial GDE from Borrelia hermsii ( BhGlpQ ) . We further demonstrated that Xe strains with mutations in the putative GDE catalytic site of AvrBs2 had reduced bacterial growth in susceptible bs2 plants , suggesting that glycerolphosphodiesterase activity has an important virulence function in this pathogen . An evolutionary analysis supports this conclusion and demonstrates that the GDE domain in AvrBs2 is under strong purifying selection . Interestingly , the catalytic mutations in GDE did not interfere with the ability of the plant to recognize AvrBs2 through the cognate R protein Bs2 and trigger disease resistance . This finding suggests that recognition of AvrBs2 is independent of its GDE enzyme activity . Genes with GDE domains have been identified in species across the animal , plant , fungal and bacterial kingdoms [45]–[47] . Although the exact biological functions of most GDE genes are unknown , it has been documented that GDE enzyme activity is directly linked to bacterial pathogenesis in other systems [45]–[47] . For example , in Borrelia species , some but not all spirochetes carry GDE genes . It has been demonstrated that spirochetes carrying GDE genes were able to achieve high cell densities ( >108/ml ) in the blood , whereas spirochetes lacking GDE genes grew too much lower densities ( <105/ml ) [41] , [48] . These results clearly suggest that the GDE gene product could contribute to bacterial virulence , although the exact mechanism is still unclear [40] . Genes similar to GDE have been identified in plants; their products may contribute to plant cell wall biogenesis [49]–[51] . It is possible that bacterial pathogens interfere with the functions of endogenous plant GDEs by either blocking or competing for the same substrates . This hypothesis could be tested in future studies as more information is revealed about plant GDEs and their endogenous substrates . In this study , we purified the GST-AvrBs2 fusion protein from E . coli and subjected it to a common procedure used to test bacterial proteins for GDE enzyme activity [41] . However , GDE enzyme activity was not detectable using the recombinant GST-AvrBs2 . This result could be due to the buffer conditions or the substrates employed , which may not be optimal for AvrBs2 enzyme activity in vitro . Interestingly , the in vitro GDE enzyme activity of the Arabidopsis putative GDE ( AT4G26690 ) was not confirmed by using a similar testing condition as described in this report [51] . It may suggest that certain plant GDEs prefer different substrates compared to E . coli GDE . Our results ( Figure 1C and 1D ) confirmed that AvrBs2 has a functional GDE catalytic site . However , the amino acid sequences flanking the GDE catalytic site may be important for substrate binding . Since the flanking sequences in AvrBs2 are different from BhGlpQ , AvrBs2 could have a different substrate specificity and not use glycerophosphocoline as substrate . It is also possible that AvrBs2 requires other plant co-factors to activate its proper folding or its GDE enzyme activity . It is not unusual for a bacterial TTSS effector protein to require plant co-factors for full enzyme activity [1] , [6] . For example , the bacterial TTSS effector AvrRpt2 requires plant cyclophilin to activate its protease activity [1] , [6] . In this study , however , it was not possible to test whether AvrBs2 required plant cofactors for its GDE enzyme activity by mixing plant total protein extracts because of the high background of endogenous plant GDE activity . By using chimeric proteins , we confirmed that AvrBs2 did possess the functional GDE catalytic site that is essential for GDE enzyme activity . Because the GDE domain is required for the virulence function of AvrBs2 , it is possible that AvrBs2 fulfills its virulence function through the GDE-activated hydrolysis of substrates in plant cells . Further investigation to identify the substrates for AvrBs2 enzyme function may help to elucidate the mechanism of the AvrBs2 virulence function and the modulation of Xe TTSS . We demonstrated that AvrBs2 carries a GDE domain with catalytic sites required for promoting bacterial virulence . However , GDE activity is not required for the activation of Bs2-dependent disease resistance . Through further genetic analyses , two overlapping AvrBs2 domains were identified: one corresponding to the GDE homologous region and one to a minimal Bs2-activating domain that includes the GDE domain and a C-terminal region . We confirmed that the previously identified mutations in this C-terminal region of AvrBs2 no longer activated Bs2-dependent resistance [24] and several novel mutations were identified that compromised Bs2 activation while having little effect on bacterial virulence . These results show that Xanthomonas can overcome Bs2 resistance without losing the virulence function of AvrBs2 . These findings are significant for optimizing the deployment of Bs2 resistance in field studies because it is important to understand how Xe strains can overcome Bs2 activation but retain the AvrBs2 virulence function . For example , anticipatory breeding could be used to identify new Bs2 alleles that recognize the AvrBs2 loss-of-recognition mutants ( R403P , A410E and Y419A ) . This scheme would allow us to use molecular breeding to stay ahead of evolving pathogens . In this study , we used the AvrBs2/Bs2 system to identify a potentially novel mechanism in plant disease resistance . AvrBs2-dependent activation of Bs2 triggers an unknown plant immunity mechanism , resulting in the suppression or modulation of the TTSS of the bacterial pathogen . In host plants containing the two R genes Bs1 and Bs2 , we observed epistasis of the Bs2 activity with a slow , 48-hour HR over the Bs1 activity with a rapid , 18-hour HR when avrBs1 and avrBs2 were present in either a single Xe strain or during co-infection into the appropriate pepper plants . A Cya reporter assay demonstrated that this interference was most likely due to the inhibtion of the bacterial TTSS following the AvrBs2/Bs2 interaction . This general inhibition of the subsequent Xe TTSS effector-reporter delivery could be detected as early as one hour after inoculation of Xe delivering wild-type AvrBs2 to Bs2 pepper plants . Recently , it has been reported that the pre-inoculation of non-pathogenic Pseudomonas fluorescens or flg21 ( a 21-amino-acid peptide from bacterial flagellin ) induces PAMP-triggered immunity ( PTI ) in Nicotiana tabacum ( tobacco ) plants [19] . The PTI subsequently inhibited the HR triggered by the secondary inoculation with Pseudomonas carrying TTSS effector genes [19] . Effector-Cya assays confirmed that HR suppression was caused by the restriction of injection of the TTSS effectors into plant cells . From this result , the authors concluded that PTI could directly or indirectly inhibit the injection of TTSS effectors into plant cells [19] . In this report , we demonstrated that the effector-triggered immunity , which was triggered by the interaction of Bs2 and AvrBs2 , led to the suppression of the delivery of TTSS effectors into plant cells . It would be interesting to test whether the mechanism of the PTI-based suppression of TTSS is similar to that of the AvrBs2/Bs2 interaction . Because almost all Gram-negative pathogens , some symbiotic bacteria and several phytopathogenic bacteria have similar TTSS machineries [52]–[54] , it is possible that the conserved components of the TTSS machinery also serve as PAMPs that are specifically recognized by plant extra- or intracellular receptors , triggering plant immunity [55] . It would be intriguing to test the hypothesis that the interaction of AvrBs2 with Bs2 directly or indirectly modifies the plant cell walls , subsequently blocking the penetration of the TTSS pilus across the plant cell walls . It would also be interesting to explore whether the TTSS suppression triggered by AvrBs2/Bs2 is common in other R protein/effector interactions in other plant species . Answering these questions may reveal whether plants employ TTSS suppression as a general immune response to help inhibit the growth of invasive bacterial pathogens .
Escherichia coli strains DH5α , Top10 , BL21 ( DE3 ) and DB3 . 1 as well as Agrobacterium tumefaciens strain C58C1 were grown on Luria-Bertani agar containing the appropriate antibiotics at 37°C ( for E . coli ) and 28°C ( for A . tumefaciens ) . Xanthomonas strains were grown on nutrient yeast glucose agar [56] containing the appropriate antibiotics at 28°C . The Xanthomonas strains used were GM98-38 Xe ( avrBs3 ) , GM98-38-1 Xe ( avrBs2 , avrBs3 ) [24] , 85–10 Xe ( avrBs2 , avrBs1 ) [31] and 69–1 Xe ( avrBs2 ) [25] . Various constructs in E . coli were transferred to Xanthomonas and A . tumefaciens C58C1 by tri-parental mating with DH5α ( RK600 ) acting as helper strain [57] . Electrolyte leakage of 1 . 5 cm2 pepper leaf disc post inoculation with Xe strains at 2×108 CFU/ml and rocked gently in 4 ml water for 1 hour . Conductance was measured with an Thermo Orion conductance meter ( model 105A+ ) in microSiemens/cm ( uS ) . Nicotiana benthamiana , tomato cv . VF36 , Bs2 transgenic Nicotiana benthamiana and VF36 and pepper lines ECW-0 ( no R gene control ) , ECW-20R ( Bs2 ) , ECW-10R ( Bs1 ) and ECW-123R ( Bs1 , Bs2 and Bs3 ) were grown in the greenhouse before and after inoculation at 24°C under 16 hours light/8 hours dark cycles . The MODELLER software package [37] was used to create a comparative protein structural model for AvrBs2 using the solved crystal structure of 1o1z A as a template . The Chimera package was used to perform structural alignments and generate molecular graphics images [58] . The full-length avrBs2 gene was amplified as a BamHI-SalI fragment by using the following primer set: 5′-caccGGATCCATGCGTATCGGTCCTCTGCAACCTTC-3′ and 5′-GTCGACATCCGTCTCCGTCTGCCTGGCCT-3′ . The resulting PCR fragment was cloned into the same sites of the protein expression vector pGEX4T-1 ( GE Healthcare , NJ ) . The GDE positive control gene Borrelia hermsii BhGlpQ was amplified from a plasmid provided by Dr . Tom Schwan ( University of Montana , Missoula , MT , USA ) by using the following primer set: 5′-caccGGATCCTGTCAGGGCGAAAAAATGAGTCA-3′ and 5′-GTCGAC TGGTTTTATTTTTGTGATGAA-3′ . The PCR product was cloned into the BamHI/SalI sites of pGEX4T-1 ( GE Healthcare , Piscataway , NJ ) . An overlap extension PCR method was applied to generate the chimeric genes BhGlpQ-avrBs2-wt and BhGlpQ-avrBs2-E304A/D306A . The catalytic domain of wild-type avrBs2 was first amplified with the following primer set: 5′-caccGGATCCTGTCAGGGCGAAAAAATGAGTCA-3′ and 5′-GCACGCCATCGGAACTGACTTCGACGTCCAGCTCTAGGTAGTCAGCTCCTAAGGCAT-3′ . The catalytic domain of avrBs2-E304A/D306A was amplified with the following primer set: 5′-caccGGATCCTGTCAGGGCGAAAAAATGAGTCA-3′ and 5′-GCACGCCATCGGAACTGACTTCGACGGCCAGCGCTAGGTAGTCAGCTCCTAAGGCAT-3′ . The derived PCR products were used as templates for another round of amplification with the following primer set: 5′-caccGGATCCTGTCAGGGCGAAAAAATGAGTCA-3′ and5′-GTTTGTTGTTGTATCAAGTTCTGGATCGTGCATCAACACCGGCACGCCATCGGAACTGA-3′ . The resulting product was the N-terminal chimera with BhglpQ genes carrying the GDE catalytic domain from either the wild-type or the mutant avrBs2 gene . The other portion of the DNA sequence of the BhGlpQ gene was amplified with the following primer set: 5′-TCAGTTCCGATGGCGTGCCGGTGTTGATGCACGATCCAGAACTTGATACAACAACAAAC-3′ and 5′-GTCGACTGGTTTTATTTTTGTGATGAA-3′ . The resulting two portions of the chimeric BhglpQ gene were re-amplified with the following primer set: 5′-caccGGATCCTGTCAGGGCGAAAAAATGAGTCA-3′ and 5′-GTCGAC TGGTTTTATTTTTGTGATGAA-3′ . The PCR products were purified by a gel-purification kit ( Bioneer , CA ) and cloned into the BamHI/SalI sites of pGEX4T-1 ( GE Healthcare , NJ ) . The DNA sequences of all clones were confirmed by sequencing . The protein expression constructs were transformed into E . coli strain BL21 ( DE3 ) by electroporation and were grown in liquid LB medium supplemented with 50 µg/ml ampicillin at 28°C/220 rpm to OD600 = 0 . 4; 0 . 5 mM IPTG was added to the culture for 6 hours to induce protein expression . The cells were harvested and disrupted by sonication in cold PBS buffer ( 147 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 2 mM KH2PO4 , pH = 7 . 4 ) supplemented with 1% Triton X-100 . The cell debris was cleared by centrifugation at 12 , 000 g for 20 min . The soluble GST fusion proteins were purified using Glutathione Sepharose following the protocol provided by the manufacturer ( GenScript USA Inc . , NJ , USA ) . The fusion proteins were eluted in 50 mM Tris-Cl , pH = 8 . 0 , supplemented with 10 mM reduced glutathione . All protein samples were stored on ice before the enzyme assays . The enzyme activity of the purified GST-fusion proteins was determined using an enzyme-coupled spectrophotometric assay to measure the amount of G3P that was released by the glycerophosphodiester phosphodiesterase reaction . The reaction mixture contained 0 . 2 M hydrazine-glycine buffer , pH = 9 . 0 , 0 . 5 mM NAD , 10 U/ml G3P dehydrogenase ( Sigma G6880 ) , 10 mM CaCl2 , 0 . 5 mM Sn-glycerol-3-phosphocholine ( G5291 ) , and the GST-fusion proteins at several pre-set concentrations . The reaction mixture was incubated at 30°C in a 96-well plate for 1 h until the oxidation of G3P by G3P dehydrogenase was complete . The G3P concentration was determined from the absorbance change at 340 nm by using the BioTek plate reader ( BioTek Instruments , Inc . , VT , USA ) . Mutants formed by homologous recombination of the genomic copy of avrBs2 in Xe were constructed as previously described [18] , [59] . The avrBs2 open reading frame was first PCR amplified with a SalI site at the 5′-end and a BamHI site at the 3′-end and cloned directionally into pBluescript KS+ . This intermediate construct was mutagenized using the QuikChange Site-Directed Mutagenesis kit ( Stratagene , CA ) to incorporate the two GDE catalytic site mutations ( E304A/D306A , H319A and Y419A ) using overlapping forward and reverse primers for the E304A/D306A sequence ( 5′-CAATCTGGCGCTGGCCGTCGAAG-3′ ) , H319A sequence ( 5′- GTGTTGATGGCCGATTTCAG-3′ ) and for the Y419A sequence ( 5′- GCCAAGTACGCCACGGGCGG-3′ ) . The resultant mutant constructs were digested with Not1 and BamH1 , and T4 DNA polymerase was used to create blunt ends . The blunt-ended fragments were then cloned into the suicide vector pLVC18L , which has a col E1 replicon and contains the highly efficient mob region from pRSF1010 [18] , cut with XbaI and SmaI , and filled using T4 DNA polymerase to make pLVC18avrBs2 ( E304A/D306A , H319A and Y419A ) . The three constructs were then mobilized into Xe ( avrBs2 , avrBs3 ) and rescued by tetracycline selection of a single recombination event into the genomic copy of avrBs2 . Second-site resolution crossover events were identified as tetracycline-sensitive single colonies from cultures grown in the absence of tetracycline . PCR amplification and sequencing were used to confirm a double homologous recombination event for either the E304A/D306A , H319A or Y419A . All bacterial growth assays in pepper and tomato were performed as previously described [25] . Two mutant strains Xe ( E304A/D306A and H319A ) were further modified by homologous recombination to add Cya as a C-terminal translational fusion as previously reported [18] . Double homologous genomic recombination was used to delete the avrBs2 locus in strains 85–10 Xe ( avrBs2 , avrBs1 ) and 69–1 Xe ( avrBs2 ) to make Xe ( avrBs1 ) and Xe ( no effector ) respectively using p815:avrBa2:GM as previously described [23] . All avrBs2 deletions and mutations were first cloned into pENTR/D-TOPO ( Invitrogen ) as previously described [59] . Each construct began with a start codon and ended without a stop codon so that the HA epitope and stop codon of the destination vector would be maintained after transfer . For Agrobacterium-mediated transient expression from the 35S promoter and C-terminal HA epitope tagging , pMD1 was first digested with Xho1 . The HA epitope and the stop codon linker ( 5′- CTCGAGTATCCCTACGACGTACCAGACTACGCATAGCTCGAG-3′ ) were cloned in and then re-opened at the Sma1 site , and the ccdB cassette A ( Invitrogen ) was cloned in to create the destination vector pMD1-Des-HA . All pENTR-avrBs2 constructs were then transferred to pMD1-Des-HA using LR clonase ( Invitrogen ) . For AvrBs1:HA and AvrBs2:HA Agrobacterium-mediated transient expression constructs both full length effectors were cloned into pENTR/D-TOPO with N-terminal XbaI site and a Cterminal HA epitope tag ( 5′- GGATCCTACCCATACGATGTTCCTGACTATGCGGGCTATCCCTATGACGTCCCGGACTATGCAGGATAGGAGCTC-3′ ) followed by a SacI site . These were then subcloned into pMD1 . The pMD1-AvrBs2:HA construct was further modified by re-opening at the single BsaI site and the ccdB cassette B ( Invitrogen ) cloned in to create a destination vector . The HindIII-EcoRI 35S-nosTerminator fragment was cloned into pENTR/D TOPO and then the AvrBs1:HA XbaI-SacI fragment was subcloned in . This pENTR-35S-AvrBs1:HA was transferred into the pMD1-AvrBs2:HA destination vector using LR clonase ( Invitrogen ) to create a double effector binary vector for Agrobacterium transient expression . The binary deletion and mutation constructs were transferred to Agrobacterium ( C58C1 ) for transient expression in Nicotiana benthamiana and pepper , as previously described [25] . Immunoblot analysis protocol was previously described [26] . Two effector-Cya reporters from avrBs1 and avrBs3 were made by directional cloning PCR products into Gateway-compatible pENTR/D-TOPO ( Invitrogen ) and then translationally fused to Cya by LR clonase ( Invitrogen ) into the suicide destination vector pDDesCya [59] . The effector PCR products of 1352 base pair for avrBs1 and 950 bp for avrBs3 included the promoter region and the first 212 codons of AvrBs1 and the first 107 codons of AvrBs3 were used to create AvrBs11-212-Cya and AvrBs31-107-Cya , respectively . The two previously constructed pDDesCya effector-Cya reporters for AvrBs21-98-Cya and XopX1-183-Cya , along with AvrBs31-107-Cya and AvrBs11-212-Cya , were introduced into Xe by genomic single recombination rescues of these constructs . This recombination still maintained the wild-type genomic copy of the particular effector [18] . The pairs of effector-Cya reporter strains with and without avrBs2 included the three-strain pairs of Xe ( avrBs3 ) and Xe ( avrBs3 , avrBs2 ) with either reporter AvrBs21-98-Cya , AvrBs31-107-Cya or XopX1-183-Cya . Also included were the two-strain pairs of Xe ( avrBs1 ) and Xe ( avrBs1 , avrBs2 ) with either XopX1-183-Cya or AvrBs11-212-Cya . Additionally the pDDesCya with AvrBs31-107-Cya was introduced into strains Xe ( avrBs2-E304A/D306A or H319A ) by genomic single recombination rescues of these constructs . The Cya was added to the C-terminus of Xe catalytic mutants of AvrBs2 as previously described [18] . The 35S- avrBs2:Cya construct was made by replacing the BamHI-SacI GFP fragment from pMD1- avrBs2:GFP [18] . with a BamHI-SacI Cya fragment . This construct was introduced into Agrobacterium for transient expression as previously described [26] . Plant cyclic AMP ( cAMP ) levels eight hours post-inoculation were measured as previously described [18] . Sampling at eight hours post-inoculation will avoid in planta multiplication of the reporter strains . Eight hours post-inoculation is also long before the development of any R gene-mediated HR . | The bacterial pathogen Xanthomonas euvesicatoria ( Xe ) is the causal agent of bacterial leaf spot disease of pepper and tomato . This pathogen is capable of delivering more than 28 effector proteins to plant cells via the type three secretion and translocation system ( TTSS ) . The AvrBs2 protein is a TTSS effector of Xe with a significant virulence contribution that depends on a conserved glycerolphosphodiesterase ( GDE ) domain . Additionally , activation of the resistance protein Bs2 by AvrBs2 modulates the TTSS of Xe and suppresses the subsequent delivery of TTSS effectors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"agriculture",
"biology"
] | 2011 | Computational and Biochemical Analysis of the Xanthomonas Effector AvrBs2 and Its Role in the Modulation of Xanthomonas Type Three Effector Delivery |
Developmental programming links growth in early life with health status in adulthood . Although environmental factors such as maternal diet can influence the growth and adult health status of offspring , the genetic influences on this process are poorly understood . Using the mouse as a model , we identify the imprinted gene Grb10 as a mediator of nutrient supply and demand in the postnatal period . The combined actions of Grb10 expressed in the mother , controlling supply , and Grb10 expressed in the offspring , controlling demand , jointly regulate offspring growth . Furthermore , Grb10 determines the proportions of lean and fat tissue during development , thereby influencing energy homeostasis in the adult . Most strikingly , we show that the development of normal lean/fat proportions depends on the combined effects of Grb10 expressed in the mother , which has the greater effect on offspring adiposity , and Grb10 expressed in the offspring , which influences lean mass . These distinct functions of Grb10 in mother and pup act complementarily , which is consistent with a coadaptation model of imprinting evolution , a model predicted but for which there is limited experimental evidence . In addition , our findings identify Grb10 as a key genetic component of developmental programming , and highlight the need for a better understanding of mother-offspring interactions at the genetic level in predicting adult disease risk .
Growth during prenatal and postnatal development influences adult health status . In humans , low birth weight is associated with an increased risk of metabolic diseases , including obesity and diabetes [1] . In addition to overall size , disproportionate growth during development is also a risk factor for common adult diseases , including coronary heart disease and high blood pressure [2] . It is well-established that environmental stresses , such as poor maternal diet [3] , can influence offspring growth , but the genetic control of developmental programming is not fully understood . Additionally , most studies have focused on prenatal development , but development in the postnatal period is also critical in influencing adult health status ( reviewed in [4] ) . The genetic control of growth and nutrient acquisition in utero is mediated in part by imprinted genes , defined by their expression from a single parental allele [5] . In the placenta , paternally expressed genes generally promote growth and maternally expressed genes suppress growth . For many genes , this pattern is thought to reflect conflict between the parental genomes , played out within the offspring , over maternal resource allocation [6] . Females maximise lifetime reproductive success by evenly distributing their resources to offspring ( since a mother is equally related to all of her offspring ) . However , when there is multiple paternity within or among litters , males are only related to some offspring produced by a female , and so it is in his genetic interest for his offspring to extract as much of his mate's resources as possible . The functions of a number of imprinted genes are consistent with the predictions of this “parental conflict hypothesis . ” For example , deletion of the paternally expressed Igf2 P0 transcript in the mouse placenta causes a failure in the supply of nutrients to the developing fetus , resulting in growth restriction [7] . Conversely , deficiency of the maternally expressed gene Grb10 causes placental overgrowth and increases placental efficiency [8] . Conflict between the parental genomes appears to be conserved in humans [9] . Postnatally , the majority of imprinted genes influence aspects of metabolism or behaviour . Grb10 is an intriguing example of an imprinted gene with diverse postnatal functions . We have previously generated and characterised Grb10 knockout mice to show that Grb10 is expressed from the maternally inherited allele in most peripheral tissues but expression in the central nervous system ( CNS ) is from the paternally inherited allele [10] . Consistent with these sites of parent-of-origin specific expression , we have demonstrated that maternally expressed Grb10 controls placental and fetal growth [8] , [11] , as well as adiposity and energy homeostasis in adulthood [12] , [13] , while paternally expressed Grb10 tempers social dominance [10] . This reciprocal imprint , with the parental alleles of Grb10 influencing distinct physiological and behavioural processes , raises questions about the evolution of imprinting at this and other loci . The presence of imprinted gene expression beyond weaning , i . e . , after the cessation of parental provisioning , is not consistent with the original predictions of the parental conflict hypothesis [14] . However , theoretical analyses have demonstrated that there are conditions that can favour imprinted expression later in life , but the expectations for patterns of imprinting in adults are not as clear as those associated with imprinting during the phase of parental care [15] , [16] . It is also possible that selection for imprinting at a locus occurs only during development and the imprint simply persists into adulthood . Postnatal nutrient provisioning by mothers occurs through the mammary gland , but unlike the placenta , the genome of the offspring's father is not represented in this tissue , precluding it as a site of direct parental conflict over allocation of resources to offspring . Despite this , we find that Grb10 is expressed and imprinted in the mammary epithelium during lactation . We show that Grb10 controls postnatal nutrient supply through expression of the maternally inherited allele in the lactating mammary gland , but at the same time , expression of the maternally inherited allele in the offspring controls nutrient demand . Proportionate growth as in wild-type ( WT ) animals requires a combination of Grb10 expression in the mother , which influences offspring adiposity , and Grb10 expression in the offspring , which influences lean mass . Together , these findings have two key implications . Firstly , our data suggest that Grb10 mediates both sides of the mother-offspring interaction . The coordinated pleiotropic effects of the gene suggest a possible role for Grb10 in mother-offspring coadaptation [17] . Coadaptation has been shown theoretically to potentially favour the evolution of imprinting [18] , but there is limited empirical evidence to support this . Secondly , when coupled with our previous observations that Grb10-associated control of lean/fat proportions influences adult energy homeostasis , our data identify Grb10 as a key genetic mediator of developmental programming . Moreover , the need for expression of Grb10 in both mother and offspring to achieve WT lean/fat proportions suggests that a better understanding of mother-offspring interactions at the genetic level will be required for more accurate prediction of adult disease risk .
We have previously described two mouse models of Grb10 ablation , generated by the integration of a LacZ reporter gene-trap cassette [10] , [11] . Here we more precisely map the gene trap loci ( Figures 1A and S1 ) , and also confirm and directly compare the patterns of fetal LacZ expression ( Figure 1B ) . Both the Grb10Δ2-4 and Grb10KO alleles ablate Grb10 protein with essentially identical phenotypic consequences , yet expression from their LacZ reporter genes is not always equivalent . Maternal transmission of each allele produces similar LacZ expression patterns at embryonic day 14 . 5 ( e14 . 5; Figure 1B ) . Contrastingly , LacZ expression in the CNS is apparent after paternal transmission of the Grb10KO allele ( Grb10KO+/p embryos ) but is not detected in the CNS of Grb10Δ2-4+/p embryos . Expression in adult Grb10Δ2-4+/p brain is detectable but weak relative to Grb10KO+/p ( Figure S2 ) . Gene-trap cassette integration in the Grb10Δ2-4 allele is coincident with the deletion of 36 kb of endogenous sequence , while only 12 bp are deleted in Grb10KO ( Figures 1A and S1 ) [11] . We considered that the differences in LacZ expression between Grb10KO+/p and Grb10Δ2-4+/p brains might be attributed to a tissue-specific enhancer perturbed by cassette integration in the Grb10Δ2-4 allele . A screen of the Grb10 genomic sequence for similarity to regulatory elements in Transfac , using PReMod [19] , identified a single element of 70 bp , called cis-regulatory module 1 ( CRM1 ) , that is highly conserved among vertebrates ( Figure 2A ) , and represents a candidate enhancer element . CRM1 proved hypersensitive to DNase I digestion in mouse adult brain , but not liver where Grb10 is not transcribed ( Figure 2B ) , consistent with CRM1 being an enhancer . PReMod identified potential recognition sequences within CRM1 for Signal transducer and activator of transcription ( STAT ) 5a/STAT5b , Tst-1 , TCF11 , and Pax family members ( Figure S3A ) . We reasoned that only factors spatially overlapping with Grb10 expression could potentially regulate Grb10 in vivo , and using public expression data [20] , we ruled out all but STAT5b . Using mRNA in situ hybridisation , we confirmed that Stat5b brain expression overlaps extensively with Grb10 ( Figure 2C ) . In a cell transfection assay , CRM1 demonstrated enhancer capability in the presence of constitutively active STAT5b ( Figure 2D ) . The STAT5 recognition sequences [21] , but not those of Tst-1 , TCF11 , or Pax , are 100% conserved between mouse , human , chimpanzee , cow , and chicken ( Figure S3B ) . Together , these data suggest STAT5 promotes Grb10 expression in mouse brain . Although CRM1 is not within the deleted sequence of Grb10Δ2-4 , its 5′ end is within 366 bp of the deletion . STAT5 binding , or its affect on Grb10 transcription , might therefore be perturbed by the deletion in the Grb10Δ2-4 allele , which could account for the observed expression differences in the CNS . One explanation is that the deletion alters the local chromatin conformation , reducing the interactions between CRM1 and the Grb10 promoter . Consistent with this , the deletion includes at least one binding site in brain for CTCF [22] , a regulator of chromatin architecture . Grb10 was previously identified as a STAT5-responsive gene in mammary epithelial cells [24] . More recently , genome-wide ChIP-seq mapping of STAT5 binding sites in mammary tissue identified three binding sites within the Grb10 locus , including one coincident with CRM1 , confirming that CRM1 can bind STAT5 in vivo [23] . We examined Grb10 expression in the mammary epithelia of our mouse models utilising LacZ reporter activity as a readout , predicting that reporter expression from the Grb10Δ2-4 allele would be weaker than from the Grb10KO allele because of a perturbation of CRM1 activity , similar to the differences observed in the CNS . We first demonstrated pregnancy-dependent expression of Grb10 in mammary epithelium using Grb10KOm/+ females . No reporter activity was detected at day 7 . 5 of gestation ( G7 . 5 ) , a subset of epithelial cells were LacZ-positive at G12 . 5 , and widespread epithelial expression was observed at day 6 of lactation ( Figure 1C ) , an expression profile comparable with other transcriptional targets of STAT5 signalling [24] . Mammary epithelial expression is restricted to the maternally inherited copy ( no LacZ staining was detected in Grb10KO+/p females ) , consistent with Grb10 imprinting in other peripheral tissues; expression of the Grb10 maternal allele is widespread during fetal development [8] , [10] , [11] and in neonatal tissues ( Figure S4 ) , but more restricted in the adult [10] , [12] . Comparable with expression differences in the CNS , no LacZ expression was detected in Grb10Δ2-4m/+ epithelium . The functional significance of imprinting in the mammary gland , which regulates nutrient allocation in the postnatal period , has not been widely considered . We were therefore intrigued by the pregnancy-dependent , imprinted expression of Grb10 observed in the mammary gland and its potential functional importance . Pup growth is the ultimate correlate of gland function [25] , and we thus compared growth of WT ( +/+ ) pups born to WT and Grb10KOm/+ dams . At e17 . 5 , WT embryos of Grb10KOm/+ females are 10% smaller than those of WT females , due to an increased litter size [8] . Despite this embryonic growth disadvantage , WT pups born to Grb10KOm/+ dams gained more weight postnatally than those born to WT dams , after standardising litter size ( Figure 3A ) . We initially interpreted this as an enhanced provisioning capacity of Grb10KOm/+ dams , suggesting that Grb10 functions in mothers to suppress nutrient supply postnatally . However , WT pups born to Grb10KOm/+ dams also had Grb10KOm/+ siblings , which were 16%±5 . 4% larger than their WT littermates at birth ( Figures 3B , 3C , and S5A ) . It was therefore necessary to consider whether these larger siblings might be impacting on WT growth , and to separate as far as possible these effects from the genotype of the dam . We used a cross-fostering strategy to differentiate the postnatal growth effects of Grb10 ablation in pup from those in dam and sibling ( Figure 3C ) . Similar approaches have been used previously to differentiate between parental and offspring effects on traits , including in the context of genomic imprinting ( e . g . , [26]–[29] ) . Pup weights at days 1 , 8 , and 15 were modelled as described in Materials and Methods . Data were analysed using a generalised linear mixed model ( GLMM ) . Cross-fostering had no main effect on any trait ( p>0 . 2 for all traits; Table S1 ) , consistent with other studies , and therefore was not included in the final models ( see also Figure S5B ) . The effect of Grb10 ablation in the dam alone was assessed by comparing the growth of pure WT litters raised by WT or Grb10KOm/+ nurses . Reduced weight gain was observed in pups raised by Grb10KOm/+ nurses ( Figure 3D , purple line ) compared to those raised by WT nurses ( black ) , demonstrating that Grb10KOm/+ nurses exhibit compromised nutrient supply , and therefore the function of Grb10 in the mother is to promote nutrient provisioning postnatally . Compromised supply from Grb10KOm/+ nurses was confirmed by our modelling , which showed a significant effect of nurse genotype at day 8 ( F1 , 21 = 15 . 08; p<0 . 001 ) and day 15 ( F1 , 21 . 6 = 25 . 60; p<0 . 0001 ) , but not day 1 ( F1 , 19 . 9 = 0 . 53; p = 0 . 4744 ) ( Table 1 , which presents hypotheses tested and key findings from the data; Tables S1 , S2 , S3 , which present the results of the models in full ) . The observations in our initial experiment that WT pups born to Grb10KOm/+ dams gained more weight postnatally than those born to WT dams could therefore not be attributed to dam genotype alone , but were likely to be influenced by sibling genotype . In support of this , WT pups with Grb10KOm/+ siblings gained more weight than WT pups with only WT siblings , when raised by WT nurses ( Figure 3D , red and black , respectively ) . This is consistent with the idea that Grb10KOm/+ pups exhibit increased demand for nutrients , to which WT nurses respond with improved provisioning , enabling increased weight gain in WT littermates . However , because all litters born to Grb10KOm/+ dams contained at least one Grb10KOm/+ pup , this result is also consistent with a maternal effect in which exposure to the Grb10KOm/+ uterine environment promotes increased postnatal weight gain in both WT and Grb10KOm/+ pups . Although the effects of siblings and the dam are correlated , we examined whether the postnatal growth of an individual is influenced by the frequency of Grb10KOm/+ pups in their litter . This analysis supported the conclusions that the pattern of postnatal growth reflects a demand effect , with postnatal growth of a pup increasing as a function of the frequency of Grb10KOm/+ siblings , with this relationship being significant for growth from day 1 to day 8 ( β = 0 . 97 , 21 degrees of freedom [df] , p = 0 . 022 ) and from day 1 to day 15 ( β = 1 . 34 , 21 . 5 df , p = 0 . 018 ) . Thus , Grb10 has pleiotropic and complementary roles in dam and pup , enhancing nutrient supply in dams and suppressing demand in pups . We next considered the compound effects of Grb10 ablation in nurse and pup , hypothesising that more demanding Grb10KOm/+ pups might not reach their full size potential when suckling from Grb10KOm/+ nurses with reduced supply . Supporting this , Grb10KOm/+ pups raised by WT nurses gained more weight than those raised by Grb10KOm/+ nurses during the postnatal period , despite being similar in weight at day 1 ( Figure 3E , blue and brown , respectively ) . Modelling the data confirmed significant differences in weight between Grb10KOm/+ pups raised by WT and Grb10KOm/+ nurses at day 8 ( t = 4 . 78 , 43 . 6 df , p<0 . 0001 ) and day 15 ( t = 5 . 47 , 52 . 3 df , p<0 . 0001 ) , but not at day 1 ( t = 1 . 76 , 43 . 7 df , p = 0 . 086 ) ( Tables 1 and S2 ) . Grb10KOm/+ pups suckling from Grb10KOm/+ nurses were larger than their WT littermates at day 1 , but their growth trajectories converged within a few days ( Figure 3E , brown and grey , respectively ) . However , Grb10KOm/+ pups remained larger than WT siblings throughout the experimental period when suckling from WT nurses ( Figure 3E , blue and red , respectively ) . Modelling of the pup/nurse interaction confirmed these observations , with the contrast between Grb10KOm/+ pups suckling from WT nurses compared to other combinations being significant at day 8 ( F1 , 62 . 3 = 39 . 33; p<0 . 0001 ) and day 15 ( F1 , 70 . 4 = 38 . 83; p<0 . 0001 ) , while the difference between Grb10KOm/+ pups with Grb10KOm/+ nurses was significantly different to that of WT pups with Grb10KOm/+ nurses at day 1 ( t = 6 . 28 , 113 df , p<0 . 0001 ) , but not day 8 ( t = 1 . 48 , 115 df , p = 0 . 14 ) or day 15 ( t = 0 . 25 , 119 df , p = 0 . 80 ) ( Table S3 ) . Thus , oversized Grb10KOm/+ neonates rapidly adjust to WT size after birth , but only when the nurse genotype is also Grb10KOm/+ , implying a role for Grb10 in influencing mother-offspring coadaptation . Our cross-fostering experiments show that WT body size is achieved through the complementary actions of Grb10 in mother and offspring . In addition to overall size , disproportionate growth can also be a risk factor for adult disease . In adulthood , Grb10Δ2-4m/+ and Grb10KOm/+ mice have an altered body composition , exhibiting increased lean mass and reduced adiposity relative to WT animals , which results in enhanced glucose metabolism [12] , [13] . The finding in the present study that growth is influenced by Grb10 in both mother and pup prompted us to examine the lean/fat ratios of animals used in the cross-fostering study . More specifically , we set out to ask whether the increased lean/fat ratio of a Grb10KOm/+ mouse is a result of Grb10 depletion within that mouse , or if the genotype of the nurse also contributes to this phenotype . To address this question , we analysed the body composition of cross-fostered pups at the end of the growth study period using dual-emission X-ray absorptiometry ( DXA ) . Ablating Grb10 in either nurse or pup increased the lean/fat ratio compared to WT pups raised by WT nurses ( Figure 4A ) . When Grb10 was ablated in offspring alone , the increased lean/fat ratio was caused by a gain in lean mass , with fat mass unchanged ( Figure 4B and 4C; compare red with blue ) . Conversely , Grb10 ablation in nurse alone caused a reduction in adipose tissue , while lean mass remained unchanged ( Figure 4B and 4C; compare black with purple ) . Therefore , Grb10 expressed in the mother has the major influence on adipose deposition , while offspring Grb10 largely influences lean mass . A WT lean/fat ratio requires functional Grb10 in both mother and pup . Since Grb10KOm/+ neonates adjust to WT size when suckling from a Grb10KOm/+ nurse , we asked whether this normalisation effect was also reflected in the lean/fat ratios . Consistent with the adjustment in body size , the lean/fat ratio of Grb10KOm/+ animals raised by Grb10KOm/+ nurses was comparable to that of their WT littermates , and also to that of WT pups raised by WT nurses ( Figure 4D; compare brown with grey and black ) . This provides further evidence that the pleiotropic functions of Grb10 in mother and pup are complementary . To inform on the mechanism through which Grb10 regulates postnatal nutrient supply , we examined Grb10KOm/+ mammary gland gross morphology at different stages , but did not detect any obvious differences from WT glands ( Figure S6A ) . We also measured various parameters in histological sections of glands harvested at day 5 of lactation . Since we had shown that Grb10KOm/+ pups demonstrate increased nutrient demand , we compared glands from Grb10KOm/+ dams to glands from Grb10KO+/p dams . Unlike WT dams , both Grb10KOm/+ and Grb10KO+/p dams raise comparable mixed genotype litters , but the absence of Grb10 expression from the paternally-expressed allele in mammary glands ( Figure 1C ) means that Grb10KO+/p females are effectively WT for Grb10 in this tissue . No differences were observed in total abdominal gland weight or surface area ( Figure S6B and S6C ) . To gain more detailed insight into gland structure , we quantified the total number , total area , mean area , mean perimeter , mean Feret diameter and mean minimum Feret diameter for lumina and adipocytes , but found no differences ( Figures S6D and S6E ) . These analyses were repeated on a separate cohort of glands isolated 48 hours after a forced wean at day 15 of lactation , and we made similar observations ( unpublished data ) . In support of these morphometric analyses , no significant differences were found between WT and Grb10KOm/+ glands , in immunofluorescence experiments using antibodies to markers of luminal epithelial ( cytokeratin-18 [CK18] ) and myoepithelial ( CK14 ) cells ( Figure S7A and S7B ) . Moreover , fluorescence activated cell sorting ( FACS ) showed no differences in the number or proportions of the same key cell types ( Figure S7C and S7D ) . As a measure of milk letdown from dam to pups , pup weight gain was assessed following a period of separation from the dam . We found no evidence that the reduced nutrient provisioning of Grb10KOm/+ dams observed in our earlier experiments was due to compromised milk letdown since , if anything , they were able to transfer more milk to pups than Grb10KO+/p dams ( Figure S8A ) . Although Grb10 is almost exclusively expressed from the paternally inherited allele in adult brain , and therefore Grb10KOm/+ females are unlikely to demonstrate perturbed maternal behaviour , we confirmed that pup retrieval and nest building behaviours are comparable between Grb10KOm/+ and Grb10KO+/p dams ( Figure S8B–S8E ) . The protein and fat content of milk was also comparable between nurses used in the cross-fostering study ( Figure S9A and S9B ) . Together these data suggest that milk letdown , maternal behaviour or the proportions of fat and protein in milk are not the basis for reduced provisioning in Grb10KOm/+ dams . However , prolactin expression in the pituitary glands of Grb10KOm/+ nurses was significantly elevated relative to WT nurses , when raising WT litters , whereas pituitary growth hormone levels were unchanged ( Figure S9C ) . Reduced provisioning causes pups to suckle more vigorously , promoting maternal pituitary prolactin expression that normally stimulates increased milk production [30] . Our data suggest that mammary glands of Grb10KOm/+ females are resistant to elevated prolactin .
In mice , most imprinted genes are expressed in the placenta and many have been shown experimentally to influence placental development and function . The parental conflict hypothesis , which , at a gross level , predicts that paternally expressed genes promote growth while maternally expressed genes suppress growth , is consistent with the functions of several genes imprinted in the placenta , including Grb10 [8] , [11] . While the conflict hypothesis can potentially explain the occurrence of imprinting at many loci , the functions of a considerable number of imprinted genes cannot be easily reconciled with the predictions of the model , such as those involved in maternal care behaviours . Recent extensions to the model have considered cases in which asymmetries between genes inherited from mothers and fathers can arise from various patterns of interactions with kin , but they do not provide strong predictions about the nature of imprinting in adult tissues [15] , [31] . There are also alternative models that do not consider conflict , such as the maternal-offspring coadaptation model , which describes how the combination of alleles expressed in mothers and their offspring jointly determines offspring fitness [18] . Whether each of these models could potentially account for the complex patterns of Grb10 expression and imprinting is unclear , but we present here some of the strongest empirical evidence that coadaptation could play a role , without necessarily ruling out alternative hypotheses . Earlier studies involving reciprocal crosses , and cross-fostering , between two distinct mouse strains established that parent-of-origin effects such as genomic imprinting could contribute to coadaptation between genotypes ( e . g . , [26] , [29] ) . These studies provide evidence that maternal provisioning is influenced by maternal and offspring genotypes [27] and that provisioning is optimal when mother and offspring are of the same genotype [29] . A study mapping quantitative trait loci on adult mouse body weight and organ weights indicated that a number of imprinted loci had small but detectable effects on these traits [28] . This genome-wide mapping is complementary to our approach , the manipulation of a single imprinted gene , which shows that Grb10 can contribute functionally to body weight and proportions through actions in both mother and offspring . Grb10 is an intriguing model with which to study imprinted gene function and evolution , because its two parental alleles are expressed in different tissues where they influence distinct physiological and behavioural processes [10] . Our earlier work characterising the same knockout mice used in this study established that maternally expressed Grb10 regulates fetal and placental growth [8] , [11] , consistent with the conflict hypothesis , as well as glucose homeostasis in adulthood [12] . In the present study , we demonstrate that Grb10 also controls postnatal growth through imprinted expression in mammary epithelium . The two archetypal mammalian tissues differ fundamentally in that the placenta contains both maternal and paternal genetic contributions , as in the offspring , but the mammary gland shares only maternal genes with offspring . Thus , while they are functionally analogous in supporting offspring growth , the placenta can be a site of direct conflict between the maternal and paternal genomes , whereas the mammary gland is not [14] . Consequently , our findings do not appear to fit with the simple predictions of the conflict hypothesis . The finding that WT body size and proportions require the combined and complementary actions of Grb10 in mother and pup ( Figure 5 ) , provides support for the coadaptation model of imprinting evolution , although further work would be needed to confirm that the effects of Grb10 lead to increased fitness . However , even if coadaptation explains the imprinting of Grb10 in pups , the coadaptation process would not favour imprinted expression in the mammary gland . Either a different hypothesis is needed to explain why the gene is imprinted in this tissue , or the pattern of expression in the mammary gland could simply reflect the selection that led to imprinting earlier in life . It should also be noted that our study compares WT Grb10 with a single knockout allele . Strong support for the coadaptation theory would come from an analysis of different allelic variants . Pups from combinations where mothers and offspring are expressing the same allelic variant would be expected to have higher fitness than those from combinations expressing different alleles . Conflict between the parental genomes , coadaptation of processes regulating postnatal nutrient acquisition , or indeed some other driving force , could have provided the initial selective pressure for the evolution of imprinting . However , a mechanism for imprinted gene regulation , once evolved , could facilitate development of novel gene functions , and thus the two models for the evolution of imprinting at the Grb10 locus , conflict and coadaptation , need not be mutually exclusive [32] . Indeed , expression of both the maternally and paternally inherited alleles of Grb10 in different tissues , influencing different phenotypes , strongly supports this notion , since the same selective forces presumably could not have led to both patterns of expression . Evidence for the later evolutionary acquisition of novel functions or imprinted expression of a gene in adult tissues could be viewed as consistent with this idea . In this context it is interesting that in at least one marsupial species , the tammar wallaby , Grb10 is expressed in a range of fetal and adult tissues , including lactating mammary gland , but its expression appears not to be imprinted [33] . Future work on the expression and functions of Grb10 in non-mammalian species will help to unravel the likely evolutionary course of this pleiotropic gene . The mechanism through which Grb10 regulates postnatal nutrient supply is not clear , but its pregnancy-dependent expression in the mammary epithelium is consistent with Grb10 controlling supply through this tissue . Our data suggest that Grb10 might mediate the response of the epithelium to pituitary prolactin , but further work will be needed to define its mode of action . We could not detect any phenotype associated with Grb10KO in analyses of mammary glands at the cellular level or in analyses of milk fat and protein content . One intriguing possibility is that Grb10 in mammary epithelium could regulate the release of signalling molecules that influence growth or metabolism of suckling offspring . Evidence exists for such lactocrine signalling playing a role in offspring development [34] . Environmental influences on the programming of adult health status during development have been widely investigated . In humans , maternal and child undernutrition are risk factors for high glucose concentration and blood pressure in adulthood [3] . Chemical insults can also adversely affect development with implications for adult health . Bisphenol A ( BPA ) , which mimics oestrogen , has garnered much attention recently because of its widespread use in the manufacture of baby bottles , coatings for food cans and other commonly used items . Effects of exposure during development to environmentally relevant levels of BPA , using rodent models , include increased body weight , advanced puberty , and altered reproductive function , as well as a possible predisposition to mammary and prostate cancers [35] . This is effected , at least in part , through alterations to the fetal epigenome , including reduced levels of DNA methylation , which is associated with changes in gene transcription [36] , [37] . The genetic control of developmental programming of adult health status is relatively poorly understood . We have previously shown that Grb10 determines the proportions of lean and fat tissue during development , and that altering the lean/fat ratio by ablating Grb10 affects glucose homeostasis in adulthood [12] . In the present study , we find that this programming of adult metabolic state during postnatal development is not the result of Grb10 acting in the pup alone , but is achieved through the combined actions of Grb10 in mother and offspring . Interestingly , two recent mouse studies of nutrient restriction during gestation have demonstrated up-regulation of Grb10 expression in offspring , consistent with our findings that Grb10 plays a key role in responding to nutrient supply [38] , [39] . Our work highlights the need for a much better understanding of genetic interactions between mother and offspring in predicting adult health status .
Experiments involving mice were conducted under a UK Home Office licence granted following local ethical review . Embryos were isolated at e14 . 5 , bisected and assayed for β-galactosidase activity as described previously [10] . Abdominal mammary glands ( gland number 4 ) were isolated from gravid and lactating females , all 10 weeks old and virgins at mating . For lactating females , litters were standardised to seven pups at birth . Glands were mounted on APTS-subbed slides , fixed for 2 hours in 2% ( w/v ) paraformaldehyde ( PFA ) , 0 . 25% ( v/v ) glutaraldehyde , 0 . 01% ( v/v ) Igepal CA-630 in 0 . 1× PBS , and further fixed for 2 hours in 2 mM MgCl2 , 0 . 01% ( w/v ) sodium deoxycholate , 0 . 02% ( v/v ) Igepal CA-630 in 0 . 1× PBS . Glands were incubated in X-gal as described for embryos for 18 hours at 28°C . Glands were cleared in acetone for 6 . 5 hours , dehydrated through an ascending ethanol series and stored in xylene . In situ hybridisations and LacZ expression analyses performed on adult brain sections have been described previously [10] . The Stat5b probe was amplified using forward ( 5′-GCAAGCATTGTCATTGTCTCCG-3′ ) and reverse ( 5′-CCATTCCTACCACCTAATCCTCAG-3′ ) primers . The Grb10 probe has been described elsewhere [10] . ExactPlus [40] was used to detect sequence conservation among Grb10 homologs , obtained from the UCSC Genome Browser . The minimum length of exact match to seed was 6 bp; the minimum number of species to seed was equivalent to the number for that alignment; and the minimum number of species to extend a hit was 2 . Murine Grb10 exons were excluded from the alignment . 15 bp of intronic sequence flanking each exon were also excluded to eliminate conserved splice recognition sequences . Custom tracks were submitted to the UCSC Genome Browser . The genomic sequence of murine Grb10 was submitted to PReMod [19] for analysis . CRM1 was assessed for DNase I hypersensitivity in adult brain and liver , using a standard method [41] with 0 , 120 and 200 units of DNase I . A 799 bp genomic DNA probe A was amplified using forward ( 5′-GGGTGTTTGTCCTTGATGCT-3′ ) and reverse ( 5′-CTGACCCCCAGAATGTGTTT-3′ ) primers , and radiolabelled using [α-32P]dCTP and a Roche High Prime labelling kit . Probe A was cloned into the KpnI and SacI restriction enzyme sites of the pGL3-Promoter vector ( pGL3-Pro , Promega ) , to generate pGL3-Pro-CRM1 . The constitutive STAT5b expression construct , pRSV-puroSTAT5b1*6 , has been used elsewhere [42] . Transfections were performed on ∼60% confluent NIH/3T3 cells in six-well plates , using Lipofectamine 2000 ( Invitrogen ) . Cells were co-transfected with 0 . 1 µg pRL-SV40 ( Promega ) , encoding Renilla luciferase . 48 hours after transfection , luciferase activity was quantified using the Dual-Luciferase Reporter Assay System ( Promega ) and a Microlumat Plus luminometer ( EG&G Berthold ) . Activity levels for each well were normalised to that of Renilla luciferase . Grb10Δ2-4 and Grb10KO mice were generated and maintained as described previously on a mixed C57BL6:CBA strain background [10] , [11] . WT mice of the same genetic background were used as controls . 7-week-old virgin dams were mated with WT sires and removed to separate cages following the observation of a cervical plug . At birth , pup paws were tattooed to permit identification . Litters were standardised to five to seven pups at birth by arbitrary pup selection . Where appropriate , pups were cross-fostered at birth to nurses that had given birth on the same day . Pups were weighed daily . Figure 3A: +/+ born to +/+ , n = 20 pups ( three litters ) ; +/+ born to m/+ , n = 18 pups ( six litters ) . Figure 3D/3E: +/+ born to +/+ raised by +/+ , n = 31 pups ( five litters ) ; +/+ born to +/+ raised by m/+ , n = 27 pups ( four litters ) ; +/+ born to m/+ raised by +/+ , n = 9 pups ( four litters ) ; +/+ born to m/+ raised by m/+ , n = 18 pups ( six litters ) ; m/+ born to m/+ raised by +/+ , n = 12 pups ( four litters ) ; m/+ born to m/+ raised by m/+ , n = 14 pups ( six litters ) . Figure S5B: +/+ raised by biological +/+ dam , n = 20 pups ( three litters ) ; +/+ raised by nurse +/+ dam , n = 31 pups ( five litters ) . Pups were culled on postnatal days 15 or 17 , and genotyped following tissue biopsy [11] . A subset of animals from the growth studies were analysed by dual-emission X-ray absorptiometry ( DXA ) ( PIXImus scanner , Lunar ) [12] . For Figure 4A–4C , animals were analysed at day 17: +/+ born to +/+ raised by +/+ , n = 19 pups ( three litters ) ; +/+ born to +/+ raised by m/+ , n = 20 pups ( three litters ) ; +/+ born to m/+ raised by +/+ , n = 5 pups ( two litters ) ; m/+ born to m/+ raised by +/+ , n = 6 pups ( two litters ) . For Figure 4D , animals were analysed at day 15: +/+ born to +/+ raised by +/+ , n = 13 pups from two litters; +/+ born to m/+ raised by m/+ , n = 18 pups from six litters; m/+ born to m/+ raised by m/+ , n = 16 pups from six litters . Pup weight was modelled using restricted maximum likelihood in the Mixed Procedure in SAS ( SAS Institute ) with pup , biological dam , nurse , and the interaction between pup and nurse genotypes fitted as categorical fixed effects and a litter ID as a random effect . Degrees of freedom were determined using the Kenward-Roger Degrees of Freedom Approximation [43] , which , in this model , effectively corrects the denominator degrees of freedom for the random effects to avoid pseudoreplication when using replicates sampled from the same litters . The influence of Grb10KOm/+ siblings on individual growth was determined by adding a regression variable to the model described above that accounts for the frequency of Grb10KOm/+ among the siblings of that individual ( so it is measured on the rest of the litter not including the genotype of the focal individual being considered; i . e . , it is the frequency that an individual pup experiences within their litter ) . Abdominal glands for carmine alum staining were fixed in Carnoy's fixative ( 60% ethanol , 30% chloroform , 10% acetic acid ) for 3 hours , transferred to 70% ethanol for 15 minutes , and hydrated slowly . Glands were stained in carmine alum ( 0 . 2% w/v ) carmine dye , 0 . 5% ( w/v ) aluminium potassium sulphate ) for 18 hours , dehydrated , and stored as above . Abdominal ( number 4 ) glands for morphometric analyses were isolated from Grb10KOm/+ and Grb10KO+/p females at day 5 of lactation , all 8–10 weeks old at conception and previously unmated . Note that these dams are the same as those used in the behavioural and milk letdown experiments . Wet weights were recorded and glands were spread across APTS-subbed slides , photographed on grids and surface areas measured using ImageJ . Glands were fixed overnight in 4% PFA ( w/v ) in PBS , dehydrated and sectioned at a thickness of 8 µm . Sections were stained with haematoxylin and eosin , and photographed at 200× magnification . Measurements were made using ImageJ . The combined measurements from three fields per sample were used . Grb10KOm/+ , n = 4; Grb10KO+/p , n = 3 . Southern blots and PCR were performed using standard protocols . Primer sequences are available upon request . RNA isolation from day 1 neonatal tissues , cDNA synthesis , PCR , and sequencing were performed as described previously [44] . 30 cycles of amplification were used with forward ( 5′-GCTGGACTCTGGTGGAACAC-3′ ) and reverse ( 5′-GGCACACATACAGCTTCTTCC-3′ ) primers . WT and Grb10KOm/+ abdominal glands were isolated from females 48 hours after a forced wean at day 15 of lactation . All females were 7 weeks old and virgins at mating . The litter size of all females was normalised to 7 pups on the day of birth . Glands were sectioned and fixed as described for haematoxylin and eosin staining . Immunofluorescence was performed essentially as described [45] , using the following antibodies: CK14 ( LL002 , Abcam ) , CK18 ( Ks 18 . 04 , Progen Biotechnik ) , Alexa555-conjugated goat anti-mouse IgG1 cross-adsorbed ( Invitrogen ) , and Alexa488-conjugated goat anti-mouse IgG3 cross-adsorbed ( Invitrogen ) . Sections were counter-stained with DAPI . Images were taken using a Zeiss LSM510META confocal laser scanning microscope . Cells were counted by assigning DAPI-stained nuclei as either luminal epithelial- ( CK18+ ) or myoepithelial-associated ( CK14+ ) . Counts from three fields were combined for each sample . WT , n = 3; Grb10KOm/+ , n = 3 . Single cells were prepared from fourth mammary fat pads of mice aged 17–25 weeks , at day 7–10 of gestation , with pregnancy confirmed during harvest of mammary tissue [46] , [47] . Cell suspensions at 106 cells/ml were stained with anti-CD24-FITC ( clone M/69 at 1 . 0 µg/ml; BD Biosciences ) , anti-Sca-1-APC ( clone D7 at 1 . 0 µg/ml; eBioscience ) , anti-CD45-PE-Cy7 ( clone 30-F11 at 1 . 0 µg/ml; BD Biosciences ) , and anti-CD49f-PE-Cy5 ( clone GoH3 at 5 . 0 µl/ml; BD Biosciences ) . Cells were sorted at low pressure ( 20 psi using a 100 µm nozzle ) on a FACSAria ( Becton Dickenson ) equipped with violet ( 404 nm ) , blue ( 488 nm ) , green ( 532 nm ) , yellow ( 561 nm ) , and red ( 635 nm ) lasers and using FACSDiva software . There was no intervening culture period between cell isolation , staining , and flow sorting . Mammary epithelial cell subpopulations were defined as in [47] . At postnatal day 17 , nurses used in the cross-fostering study were anaesthetised with an intraperitoneal injection of 0 . 1% ( w/v ) xylazine , 0 . 5% ( w/v ) ketaset in 0 . 9% ( w/v ) NaCl at 16 µl/g mouse . Milk collection was aided by an intraperitoneal injection of 200 µl 10 IU/ml oxytocin ( Sigma Aldrich ) in 0 . 1% PBS . A vacuum pump was used to harvest 100–200 µl milk . Milk fat and protein content were analysed as described [48] . After fat removal , protein was diluted 1/5 in 50 mM Tris HCl ( pH 8 . 0 ) , 150 mM NaCl , 1% ( v/v ) Igepal CA-630 , boiled for 10 minutes , and mixed with an equal volume of reducing sample buffer . Samples were run on a 15% Criterion Tris-HCl gel ( Bio-Rad Laboratories ) alongside pre-stained molecular weight markers . Fixation and drying were performed using standard methods . After milk collection , nurses were humanely killed by cervical dislocation and pituitary glands harvested . RNA was extracted from pituitaries using TRI Reagent ( Sigma Aldrich ) . Total RNA ( 1 µg ) was DNase-treated with RQ1 RNase-free DNase I ( Promega ) . cDNA was synthesised using random hexamers with Superscript III RNase H− Reverse Transcriptase ( Invitrogen ) . Real-time PCR ( qRT-PCR ) was used to measure expression of prolactin and growth hormone normalised to β-actin . Reactions were performed in duplicate and analysed as described previously [8] . Grb10KO+/p dams were considered to be a better control than WT dams in these experiments because both Grb10KOm/+ and Grb10KO+/p dams have mixed genotype litters , and therefore demand is more equally matched than with WT dams that have only WT litters . Nurturing behaviour of dams was quantified , including time to initiate nest building activity , time to settle on the nest , and time to retrieve pups , following removal of pups from the nest and separation from the dam for 1 hour ( essentially as in [49] , but performed at 3–4 hours into the light cycle on postnatal day 1 ) . As a surrogate measure of milk letdown , pups were weighed immediately following removal from the dam , immediately before their return to the nest , and at intervals for up to four hours thereafter [49] . | Experiences during early life can impact on health status in adulthood; low birth weight , for example , is linked to an increased risk of diabetes and obesity in later life . Such developmental programming can be influenced by environmental factors such as diet , but the importance of genetics in this process is not well understood . Using the mouse as a model , we investigate the gene Grb10 , which is imprinted , meaning that it is expressed from only one of its two copies . We show that Grb10 is a key mediator of developmental programming , controlling supply and demand of nutrients in the postnatal period and influencing growth and body composition . Specifically , we find that Grb10 determines the proportions of lean and fat tissue during development , and that this is dependent on the combined actions of Grb10 in the mother and offspring . Our findings have two main implications . First , they suggest that the functions of Grb10 in mother and offspring are coadapted , providing support for a coadaptation model for the evolution of imprinted genes . Second , they highlight the need for a better grasp of how maternal and offspring genetics interact during development if we are to understand more fully the causes of complex adult disorders such as obesity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"developmental",
"biology",
"histology",
"animal",
"genetics",
"model",
"organisms",
"organism",
"development",
"genomic",
"imprinting",
"genetics",
"epigenetics",
"molecular",
"genetics",
"biology",
"anatomy",
"and",
"physiology",
"molecular",
"cell",... | 2014 | Developmental Programming Mediated by Complementary Roles of Imprinted Grb10 in Mother and Pup |
Snakebites are a major Collective Health problem worldwide . In Brazil , Bothrops jararaca snake venom ( BjV ) evokes hemostatic disturbances , bleeding manifestations , and redox status imbalance . Specific antivenom therapy , although efficacious to revert most snakebite-induced manifestations , is incapable of treating secondary manifestations , such as oxidative/nitrosative stress . Searching for new complementary therapies that could attenuate physiological derangements triggered by envenomation , we elected to test quercetin-3-rutinoside ( rutin ) by its potential as both a potent antioxidant and a hemostasis modulatory compound . The activity of rutin was evaluated both on the biological activities of crude BjV in vitro , and in vivo by the ability of rutin ( 14 . 4 mg/kg b . w . ) to modulate hematological , hemostatic and redox status markers altered by BjV injection ( 1 . 6 mg/kg b . w . , s . c . ) in mice . In vitro , rutin failed to inhibit BjV-induced platelet aggregation and biological activities of major BjV enzymes ( metalloproteinases , phospholipases A2 , serine proteases , and L-amino acid oxidases ) . On the other hand , rutin attenuated local hemorrhage , and the increase in reactive species , prevented the fall in RBC counts and fibrinogen levels , diminished tail bleeding and shortened prothrombin time ( PT ) evoked by envenomation . Furthermore , rutin reduced tissue factor ( TF ) activity and altered the protein expression of TF in liver , lungs , heart and skin . In conclusion , the disturbances in redox status and hemostatic system induced by B . jararaca envenomation were modulated by rutin , suggesting it has a great potential to be used as an ancillary therapeutic agent for snakebites .
Worldwide , snakebites are considered a major public health issue , and , acknowledging that , the World Health Organization ( WHO ) declared snakebite envenomation as a neglected tropical disease . In Brazil , Bothrops snakes ( lance-headed vipers ) are responsible for approximately 20000 snakebites/year , and Bothrops jararaca is considered by WHO as a species of high medical importance since it causes numerous snakebites [1 , 2] . Bothrops jararaca venom ( BjV ) is composed by a complex mixture of proteins , and most of them are grouped into the following families: snake venom metalloproteinases ( SVMP ) , snake venom serine proteinases ( SVSP ) , type-C lectins , phospholipases A2 ( PLA2 ) and L-amino acid oxidases ( LAAO ) [3] . These proteins are responsible for the toxic activity of BjV , and evoke clinical disturbances in snakebite victims , such as intense inflammatory reactions ( edema , local bleeding , and necrosis ) at the site of the bite , and systemic bleeding ( gingival bleeding , ecchymosis , petechiae , hematuria , epistaxis , and hemoptysis ) [4] . Furthermore , envenomation induces thrombocytopenia , platelet dysfunction , consumptive coagulopathy , secondary fibrinolysis , a late mild fall in red blood cells ( RBC ) counts , and neutrophilic leukocytosis [5 , 6] . A recent study of our group [6] showed that rats injected with BjV displayed increased tissue factor ( TF , NCBI Reference Sequence: NP_001984 . 1 ) activity in plasma and tissues , and altered the protein expression of TF at the site of venom injection . TF is a 47-kDa transmembrane glycoprotein that initiates the extrinsic pathway of the coagulation cascade , and is present in platelets , monocytes , macrophages , endothelial cells , and microparticles of these cells . In normal physiological conditions , TF is found on its encrypted ( inactive ) form . However , under pro-inflammatory circumstances and/or injury , TF is decrypted and rapidly initiates blood coagulation in vivo [7] . Several mechanisms have been attributed to modulate TF function , and among them , the controversial regulatory activity of PDI , which is a thiol-isomerase and oxi-reductase chaperone that is required for thrombus formation in vivo and that has been reported to form an allosteric Cys186–Cys209 disulfide bond in TF , which is essential to its coagulating activity [7–9] . Envenomation by different genera of snakes–e . g . , Bothrops [10] , Daboia [11] , Echis [12] , and Crotalus [13]–have also been shown to evoke oxidative/nitrosative stress ( ONS ) , characterized by an imbalance between the pro- and antioxidant systems . ONS may lead to extremely deleterious effects to living organisms [10] and it has been associated with several pathophysiological conditions , including hematological disturbances and inflammatory reactions resulting in leukocyte activation . In fact , increased levels of reactive oxygen and nitrogen species ( ROS/RNS ) may induce apoptotic signaling in RBC [14] and platelets [15] , shortening their life span . During envenomation , the increase in reactive species has been attributed to both the ischemia-reperfusion tissue injury and the inflammatory reaction that occur following venom injection [10 , 16] . Moreover , snake venom PLA2 and LAAO evoke lipid peroxidation and L-amino acid deamination , respectively , generating reactive species . Patients bitten by B . jararaca manifest ONS even one month after antivenom administration [10] . Although antivenom therapy is the recommended treatment for snakebites , it cannot directly block the secondary complications , such as reactive species generation , triggered directly or indirectly by venom toxins . On that account , plant-based compounds are widely used in the treatment of several diseases , including those associated with ONS . Glycosides from quercetin group are remarkably abundant dietary flavonoids and display antioxidant action , such as metal ion chelation and reactive species scavenging , and block reactions and systems that generate reactive species [17] . Besides these actions , rutin ( quercetin-3-rutinoside ) binds to PDI , inhibiting its activity and , interestingly , the administration of rutin or isoquercetin has been shown to prevent thrombus formation in mice and humans [8 , 18 , 19] . However , if rutin administration also affects TF function or protein expression has never been evaluated . Reasoning that the hemostatic disorders evoked by B . jararaca envenomation are multifactorial [6 , 20] , and that the concurrent imbalance in the redox status might be an additional mechanism to aggravate them , we hypothesized that ROS/RNS generation and TF encryption/decryption could be involved in the etiopathogenesis of hemostatic disorders during envenomation and might be mitigated by rutin administration . By using rutin , a potent antioxidant that controls thrombus growth and regulates ROS/RNS generation , we could address the link between generation of reactive species , TF decryption/encryption and hemostatic disturbances during B . jararaca envenomation; in addition , we investigated whether rutin could be used as a putative ancillary treatment to B . jararaca envenomation . We show herein that rutin has various beneficial effects during B . jararaca envenomation .
Lyophilized venom from adult specimens of B . jararaca snakes was obtained from the Laboratory of Herpetology , Instituto Butantan . Bothrops antivenom was kindly donated by Instituto Butantan ( lot: 1305077 ) . Rutin was obtained from Sigma-Aldrich ( USA ) . All other reagents were of analytical grade or better . Antibodies: Anti-β-actin ( A5316 ) , anti-GADPH ( G8795 ) and PDIa1 ( P7372 ) were obtained from Sigma-Aldrich ( USA ) . The anti-TF antibody ( ab151748 ) was obtained from Abcam ( USA ) . The anti-mouse IgG Alexa Fluor 488 ( A11001 ) and anti-rabbit IgG Alexa Fluor 647 ( A21245 ) were purchased from Thermo Fisher Scientific ( USA ) . Male Swiss mice , weighing 30–35 g , were obtained from the Animal Facility of Instituto Butantan , and were maintained with free access to food and water . The experimental procedures involving human donors , rats , and mice were in accordance with National Guidelines , and were approved , respectively , by the National Human Research Ethics Committee ( Plataforma Brasil , CAAE: 51368615 . 5 . 0000 . 0065 ) , and the Institutional Animal Care and Use Committee from Instituto Butantan ( CEUAIB 4388061115 ) and the Faculdade de Medicina , Universidade de São Paulo ( protocol 188/15 ) . All procedures involving animals were in accordance with the National Guidelines of Conselho Nacional de Controle de Experimentação Animal ( CONCEA ) [21] . To evaluate whether rutin directly inhibits SVMP , SVSP , LAAO and PLA2 present in BjV , the following substrates were used , respectively: azocoll , DL-BAPNA , L-leucine , and soybean lecithin [22] . The direct inhibition of the procoagulating activity of BjV by rutin was also tested in human and mouse plasmas , using the values of the minimum coagulant dose ( MCD ) for comparison [6] . Fixed concentrations of rutin ( 0 . 25 mg/mL for SVMP , SVSP and LAAO , and 1 . 25 mg/mL for PLA2 ) or two-fold serially diluted solutions ( from 0 . 5625 to 9 . 0 mg/mL for MCD ) of rutin were employed . BjV was two-fold serially diluted ( from 0 . 0156 to 1 mg/mL for SVMP , SVSP , LAAO and MCD , or from 0 . 078 to 5 . 0 mg/mL for PLA2 ) and incubated with or without rutin for 30 min at 37°C . As positive controls for the inhibition of SVMP or SVSP , 13 mM Na2EDTA or 8 mM AEBSF , respectively , was incubated with BjV for 1 h at 37°C . Results were expressed as the percentage of inhibition of the enzymatic activities by rutin . To test if BjV-induced platelet aggregation was inhibited by rutin , mouse blood ( 6 vol . ) was collected from the caudal vena cava into ACD anticoagulant ( 1 vol . ) ( 85 mM trisodium citrate , 71 . 4 mM citric acid , 111 mM dextrose , pH 6 . 2 ) and kept at 37°C . Thereafter 1 vol . of anticoagulated blood was gently homogenized with 2 vol . of washing solution [6 vol . of Dulbecco’s Modified Eagle’s Medium ( DMEM , pH 7 . 4 ) to 1 vol of ACD] , laid onto 1 mL of Histopaque 1077 ( Sigma-Aldrich , USA ) , and centrifuged at 700 g for 30 min at room temperature . The upper layer was added up to 2 mL of DMEM/ACD containing prostaglandin E1 ( 0 . 5 μL , 200 μg/mL ) and centrifuged at 9500 g for 90 s at room temperature . Supernatants were discarded and the pellet of platelets was resuspended in DMEM/ACD/PGE1 . Centrifugation and resuspension of platelets were carried out 2 more times and finally platelets were resuspended in DMEM , and platelet count was adjusted to 5 x 105 platelets/mL . Washed platelets were stimulated by BjV alone or BjV+rutin ( final concentrations: BjV , 24 . 4 μg/mL and rutin , 220 μg/mL ) . Platelets were also pre-incubated with rutin for 15 min at 37°C and then BjV was added . Platelet aggregation was recorded for 5 min using a Chrono-log aggregometer ( USA ) [22] . BjV was dissolved in sterile saline ( 0 . 8 mg/mL ) ; rutin ( 7 . 2 mg/mL ) was dissolved in a solution containing 1 vol . of propylene glycol and 1 vol . of saline . The solutions were prepared immediately before use . The dose of BjV ( 1 . 6 mg/kg b . w . , s . c . route ) was selected based on previous tests to determine a dose that evoked hemostatic disturbances in mice similar to those observed in rats [6 , 23] and humans [5] . The dose of rutin ( 14 . 4 mg/kg b . w , s . c . route ) was established based on previous experiments , showing that this dose was effective to modulate redox status and hemostatic parameters . Animals were randomly allocated in four experimental groups ( n = 6/group/time interval ) ( Fig 1 ) and received ( 4 . 0 mL/kg b . w . , s . c . ) : saline ( saline control , i . e . , the negative control ) ; rutin ( rutin control , 14 . 4 mg/kg b . w . of rutin ) ; BjV+saline ( positive control , 1 . 6 mg/kg b . w . of BjV ) ; and BjV+rutin ( rutin treatment , 14 . 4 mg/kg b . w . of rutin and 1 . 6 mg/kg b . w . of BjV ) . The solutions were incubated at 37°C for 30 min before administration . At 3 , 6 and 24 h after injection , animals ( 5–6 mice/group ) were anesthetized with isoflurane ( induction and maintenance at 2 . 5% ) , and blood and tissues were collected ( Fig 1 ) . Blood was collected from the caudal vena cava into plastic syringes to obtain complete blood cell counts ( CBC ) , blood smears and plasma . Blood was added to flasks containing the following anticoagulants: CTAD ( 75 mM trisodium citrate , 42 mM citric acid , 139 mM dextrose , 15 mM theophylline , 3 . 7 mM adenosine , 0 . 2 mM dipyridamole , and 2 μM imipramine ) , 3 . 2% trisodium citrate , or 269 mM Na2EDTA . Bothrops antivenom ( 1 vol . to 100 vol . of blood ) was added to all flasks to avoid in vitro clotting of the blood induced by BjV . Fragments of liver , lungs , heart ( 50–100 mg ) and skin ( 0 . 5–0 . 75 cm2 around the site of s . c . injection ) were collected in flasks containing RIPA buffer ( 50 mM Tris-HCl , 150 mM NaCl , 1% Triton X-100 , 1% sodium deoxycholate , 0 . 1% SDS , 2 mM Na2-EDTA , and a protease inhibitor cocktail , pH 7 . 5 ) or HBSA buffer ( 20 mM HEPES-NaOH buffer , pH 7 . 5 , containing 100 mM NaCl , 0 . 02% NaN3 and 1 mg/mL bovine albumin ) , both containing Bothrops antivenom , and frozen at -80°C . Tissues were disrupted using FastPrep-24 ( #6004–500 , MP Biomedicals ) for 60 s at 6 . 5 m/s , followed by 3 cycles of freezing in dry ice and thawing in a water bath at 37°C . Finally , samples were centrifuged at 13000 g for 10 min at 4°C , and supernatants were collected and frozen at -80°C . Protein concentration in the supernatants was determined by bicinchoninic acid assay [24] . As reference samples , pools of normal plasma and each organ ( liver , lungs , heart and skin ) were obtained and assayed on each determination . The measurement of reactive species levels in plasma was assayed based on previous reports [25 , 26] by detecting the fluorescence of DCFH-DA . The results were analyzed using a DCF standard curve ( from 0 . 097 to 6 . 250 nM DCF ) . The total antioxidant capacity ( TAC ) was determined based on the CUPRAC colorimetric assay [27] . As a standard curve , 2-fold serially dilutions of reduced L-glutathione ( from 12 . 5 to 300 μM ) were used . Results obtained in each assay were normalized using a normal mouse plasma pool as reference , and were expressed in arbitrary units . CBC was performed in an automated cell counter BC-2800 Vet ( Mindray , China ) , and differential counts were carried out in blood smears stained panchromatically . Levels of plasma hemoglobin were measured by the peroxidase method [28] . The results were quantified based on a standard curve of hemoglobin ( from 3 . 125 to 200 μg/mL ) and normalized using a reference plasma pool as described above on DCF method . Plasma fibrinogen was analyzed using a colorimetric assay [29] . Prothrombin time ( DiaPlastin , DiaMed , Brazil ) , TF activity ( Actichrome TF Kit , Sekisui , USA ) , and TF antigen ( mouse-TF ELISA kit , Elabscience , USA ) were assayed in plasma according to manufacture instructions . Both tail bleeding and local hemorrhage were assessed in the same mice ( 5–11 mice/group ) . The tail bleeding assay was used to evaluate hemostasis in vivo , and was evaluated by determining the volume of blood ( hemoglobin ) leaked after tail amputation; the hemorrhagic activity of venom was evaluated by measuring the area and color intensity of hemorrhage at the subcutaneous region surrounding the site of injection [30 , 31] . Briefly , animals received the treatments s . c . at the dorsum , and after 3 h they were anesthetized and maintained under isoflurane anesthesia during the whole procedure of tail bleeding . A distal 10-mm segment of the tail was amputated and the tail was immediately immersed in 50 mL of isotonic solution ( 154 mM NaCl , 2 mM CaCl2 ) at 37°C for 15 min . Thereafter , mice were euthanized , and the suspension was centrifuged at 1900 g for 15 min at room temperature . The pelleted blood cells were resuspended in 1 mL of saline , homogenized , and 100 μL of this mixture were added to 5 mL of von Kampen-Zijlstra reagent ( 200 mg K3[Fe ( CN ) 6] , 50 mg KCN , 120 mg KH2PO4 , 50 mg NaCl and 1 mL Triton X-100 per liter ) . The suspension was read at 540 nm and the results were expressed as mg of hemoglobin per sample [32] . The analysis of the area and intensity of hemorrhage was carried out as described elsewhere [33] , using the images of the subcutaneous of mouse skin captured with ImageScanner III ( GE Healthcare , USA ) , and the results were expressed as hemorrhagic units . Tissue supernatants were electrophoresed in 10% SDS-PAGE gels and blotted to evaluate protein expression of TF , PDI , β-actin and GADPH . Nitrocellulose membranes were blocked with 5% nonfat dry milk , and thereafter incubated with primary antibodies anti-TF , anti-β-actin and anti-GAPDH ( 1:5000 ) , and later on with secondary antibodies ( anti-mouse IgG conjugated with Alexa Fluor 488 and anti-rabbit IgG conjugated with Alexa Fluor 647 , 1:5000 ) . Fluorescence emission was captured in a ChemiDoc MP system ( Bio-Rad , USA ) , and images were analyzed using ImageLab software ( version 5 . 2 . 1 , Bio-Rad ) . Relative quantification was performed as described elsewhere [6 , 34] , with normalization based on the values of reference pools for each organ . TF activity in tissue samples was measured by a coagulant assay [35] using organ fragments immersed in HBSA . Tissue supernatants , obtained as mentioned previously , were incubated with a pool of normal plasma ( 10% rat , 90% human ) and after adding 25 mM CaCl2 , the clotting time was measured in a Start4 coagulometer ( Stago , France ) . Once mouse TF is known to be less efficient to activate human factor VII/VIIa , and the volume of plasma required for clotting assays is relatively high , the mixture of plasma pools from normal rats and humans [36–38] was used as a surrogate to assay mouse TF in tissue samples . TF activity was calculated based on a standard curve of commercial thromboplastin ( DiaMed , Brazil ) , and was normalized using the reference sample of tissues . Results were also used to calculate the TF activity/ TF protein expression ratio . The a priori calculation of sample size ( 6 mice per group ) was determined in G*Power 3 software ( http://www . gpower . hhu . de ) , taking a two-way ANOVA design , a β error of 20% ( power 0 . 8 ) , an α error of 0 . 05 , and an effect size of 0 . 2 ( based on previous data about platelet count and fibrinogen levels in BjV-injected animals and controls ) . Normal distribution and homoscedasticity of the results were analyzed using the software STATATM , version 10 , and data were transformed whenever necessary . Depending on the statistical analysis , one-way , two-way ANOVA , or Kruskal-Wallis test was used , followed by post-hoc tests ( Bonferroni , Student-Newman-Keuls or Dunn’s tests ) . The softwares SPSS ( version 22 ) and SigmaPlot ( version 12 . 0 ) were employed for these analyses . Post-hoc power analysis of ANOVA’s were greater than 0 . 87 . SPSS was also used to determine the Pearson correlation coefficients and ROC curves between two variables . Results were considered statistically significant when p< 0 . 05 , and the data were expressed as mean ± standard error of mean ( s . e . m . ) . Statistical tests used for each variable were described in figure legends .
Rutin minimally interfered in the activity of SVMP , SVSP , PLA2 and LAAO . Furthermore , the clotting activity of BjV–represented by MCD values in human and mouse plasmas–was similar in the presence or absence of rutin ( Fig 2A ) . As positive controls , Na2EDTA , a SVMP inhibitor , inhibited 86% of the collagenolytic activity of SVMP , completely inhibited the clotting activity of BjV in mouse plasma and diminished the clotting activity of BjV by 3-fold in human plasma; on the other hand , AEBSF , an inhibitor of SVSP , decreased the catalytic activity of SVSP by 98 . 4% , and also blocked the coagulant capacity of BjV by 2-fold in human and mouse plasma . These data showed that rutin failed to inhibit SVMP , SVSP , PLA2 and LAAO in vitro , as well as the coagulant activities of BjV . The inhibitory activity of rutin on BjV-induced platelet aggregation was also tested ex vivo , showing that when rutin was incubated with washed platelet suspensions for 15 min and later stimulated with BjV , the same extent of platelet aggregation induced by BjV alone ( approximately 60% ) was observed ( Fig 2B ) . On the other hand , if BjV was incubated with rutin for 30 min , and both were added simultaneously to the platelet suspensions , platelet shape change was more prolonged , leading to a mild reduction in the extent of platelet aggregation ( ≈ 50% ) . These results indicated that rutin possibly bound to components in BjV important to induce platelet activation , however this fragile binding was rapidly broken in contact with platelets , so that platelet activation and aggregation occurred normally . As shown in Fig 3A for the DCF assay , in comparison with the saline group , a marked increase in ROS/RNS levels was observed following BjV injection from 3 to 24 h , but it was more evident at 3 h ( p< 0 . 05 ) . Rutin tended to limit the rise in ROS/RNS levels in animals treated with BjV , particularly at 6 and 24 h , but the differences among BjV+saline and BjV+rutin groups were not statistically significant ( p = 0 . 178 at 6 h and p = 0 . 238 at 24 h ) . Concomitantly , but not in direct parallel with the rise in ROS/RNS levels , there was a statistically significant drop in TAC levels ( Fig 3B ) in mice injected with BjV+saline and BjV+rutin at 3 h ( p<0 . 05 ) . Although TAC and DCF plasma levels varied inversely in the experimental groups , only a modest correlation was noticed ( r = -0 . 245 , p = 0 . 0399 , n = 71 ) , evidencing that they evaluated different aspects of ONS . Rutin failed to restore TAC levels in the BjV+rutin group , even though it decreased ROS/RNS levels . Altogether , these data showed that BjV generated reactive species , particularly hydroxyl and peroxynitrite radicals detected by DCF [39] , and decreased TAC levels . The reduction in TAC levels reinforced the idea that low non-enzymatic antioxidant compounds have been consumed during envenomation . Rutin partially controlled the generation of reactive species , but rutin alone did diminish TAC levels in vivo . Erythron values did not vary importantly among groups , except at 24 h when values of RBC ( Fig 4A ) , hematocrit ( Fig 4B ) and hemoglobin ( Fig 4C ) decreased around 25–30% in BjV+saline group compared to all control groups ( p< 0 . 001 ) . Importantly , rutin prevented this fall at 24 h , and the BjV+rutin group showed RBC values similar to those of controls . The morphological analysis of RBC in blood smears showed that half of the animals in the BjV+saline group showed anisocytosis and polychromasia at 24 h , whereas they were not observed in the BjV+rutin group . We also analyzed plasma hemoglobin levels to verify if the decrease in RBC parameters was due to intravascular hemolysis . As shown in Fig 4D no statistically significant difference among groups was noticed at any time period , indicating that BjV evoked RBC disturbances , and that rutin could successfully prevent them . An increase in WBC and neutrophil counts ( p< 0 . 001 ) was noticed in the BjV+saline group at 6 h ( Fig 4E and 4F ) . Mice that received rutin showed higher WBC , neutrophil , and monocyte counts ( Fig 4E , 4F and 4H ) in comparison with the saline group at 3 and 6 h ( p< 0 . 05 ) . Moreover , in the control rutin group at 6 h , lymphocyte counts ( Fig 4G ) were also elevated ( p< 0 . 05 ) . BjV ( alone or with rutin ) markedly decreased platelet counts ( Fig 5A ) in comparison with the saline control ( p< 0 . 001 ) at all time periods . Simultaneous to the fall in platelet counts , an increase was noticed in the mean platelet volume ( MVP , Fig 5B ) over time ( p< 0 . 05 ) , showing that the consumption of circulating platelets was followed by the release of new larger platelets into the blood stream . Pearson correlation coefficients and ROC curves between DCF or TAC , and hematological and biochemistry data evidenced that only platelet counts showed a statistically significant correlation with TAC ( r = 0 . 429 , p = 2 . 48 × 10−5 , n = 72 ) or DCF ( r = -0 . 352 , p = 2 . 59 × 10−3 , n = 71 ) , implying that thrombocytopenia was mildly associated with ONS . In fact , the values of the area under the curve ( AUC ) for ROC curves between DCF ( 0 . 728 ± 0 . 063; cutoff 1 . 3 ) or TAC ( 0 . 618 ± 0 . 072; cutoff 1 . 1 ) and platelet counts were modestly high , and thereby they could not be considered good discriminators for the development of thrombocytopenia . BjV induced a marked drop in fibrinogen levels ( Fig 5C ) , which tended to recover at 24 h ( p< 0 . 05 ) . Interestingly , rutin prevented animals from fibrinogen consumption during envenomation , once no statistically significant differences were noticed between the BjV+rutin and saline groups over time ( p = 0 . 935 at 3 h , p = 0 . 233 at 6 h , and p = 1 . 000 at 24 h ) . Once rutin blocked the drop in fibrinogen levels induced by BjV occurring as soon as 3 h after injection , we explored if rutin could attenuate other major hemostatic disturbances induced by BjV at this time period . No statistically significant correlation was noticed between DCF or TAC levels , and fibrinogen levels . Considering the hemostatic disturbances evoked by BjV at 3 h after the envenomation , we investigated what was occurring to the extrinsic pathway of the coagulation cascade at this time period . As expected , 3 h after envenomation the BjV+saline group showed a prolongation of prothrombin time ( >300 s ) ( p< 0 . 05 ) , and despite the difference noticed between the BjV+rutin and saline control groups ( p< 0 . 05 ) , rutin markedly shortened prothrombin time in comparison with BjV+saline group ( Fig 6A ) . Furthermore , increased TF activity ( p< 0 . 05 ) ( Fig 6B ) and unaltered TF antigen levels ( p = 1 . 000 , Fig 6C ) were noticed in the BjV+saline group , leading to a tendency towards the increase in TF activity/antigen ratio ( Fig 6D ) . On the other hand , the BjV+rutin group displayed an increase in both the activity and antigen levels of TF ( p< 0 . 05 ) , showing thereby a lower TF activity/antigen ratio compared with the BjV+saline group . Thus , these results indicate that envenomation triggered the coagulation cascade inducing not only the direct activation of coagulation factors , but also TF decryption , and that both factors possibly contribute to the consumptive coagulopathy . Rutin in turn failed to decrease TF activity in plasma , but , on the contrary , did augment the levels of TF antigen in envenomed animals . These findings demonstrated that rutin did not interfere in TF decryption , and favored the increase in TF levels in plasma . Nonetheless , and most importantly , rutin markedly reduced prothrombin time , suggesting that its beneficial activity was not associated with the modulatory activity of TF encryption/decryption . The changes induced by BjV in blood coagulation and platelets were also echoed by increased tail bleeding ( Fig 7A ) , a test that evaluates hemostasis in vivo . On the other hand , the BjV+rutin group showed a marked reduction in tail bleeding–which was not statistically different from the saline control group ( p = 1 . 000 ) –showing that rutin was able to circumvent hemostatic alterations induced by BjV . Rutin alone did not interfere in tail bleeding . Furthermore , the well characterized local hemorrhage induced by BjV at the site of venom injection ( Fig 7B and 7C ) was also remarkably attenuated by rutin ( p< 0 . 05 when compared to BjV+saline ) . These findings imply that rutin has in vivo actions that counteract the toxic activities of BjV . Protein expression of endogenous proteins ( GAPDH and β-actin , supporting information , S1 Fig ) and TF ( Fig 8A ) was analyzed in samples of liver , lungs , heart , and skin ( at the site of venom injection ) at 24 h , and the profile of protein expression changed depending on the organ . BjV failed to evoke statistically significant differences in TF , GAPDH and β-actin expression in regard to the saline control in most organs , except for decreasing GAPDH expression in the skin . However , most of protein expression changes were induced by rutin alone , which upregulated or downregulated all studied proteins systemically . The same profile of changes was obtained for PDI , supporting information , S2 Fig ) . In the BjV+rutin group , protein expression was similar to that of the BjV+saline group in the liver , whereas in the lungs and skin it was similar the rutin control . Only in the heart , protein expression behaved differently from the BjV+saline or rutin control groups . These results evidence that BjV exerted little influence on protein expression of TF at 24 h . On the other hand , surprisingly rutin alone induced a drastic change in protein expression , from either endogenous proteins ( GAPDH and β-actin ) or TF . In order to investigate if rutin could alter TF decryption systemically , TF activity was also assayed in organs at 24 h . In addition , TF activity/protein expression ratio was also calculated using both the TF activity and TF protein expression values reported above ( Fig 8B and 8C ) . As noticed for protein expression results , no alterations were observed in TF activity and TF activity/protein expression ratio in all organs in the BjV+saline group . However , animals injected with rutin ( rutin control and BjV+rutin groups ) showed lower TF activity ( p< 0 . 05 ) in the lungs , heart and skin when compared to the saline control or BjV+saline groups ( Fig 8B ) . Interestingly , in the rutin control group the TF activity/protein ratio ( Fig 8C ) decreased significantly only in the lungs ( p< 0 . 05 ) in comparison with the saline control and BjV+saline groups , although the same paradigm was also detected in the liver and skin . The BjV+rutin group had decreased TF activity/antigen ratio in the lungs and heart when compared to the BjV+saline group ( p< 0 . 05 ) .
Bothrops jararaca venom displays pro-inflammatory , hemorrhagic and anti-hemostatic activities [22] that elicit clinical manifestations in patients bitten by B . jararaca snakes . However , the mechanisms of action whereby clinical manifestations develop are not completely understood , as well as why some patients develop complications resulting from severe hemostatic disturbances [40] and from ONS [10] . Herein we scrutinized the complex relationship between ONS and hemostatic disturbances , mainly related to TF , and the potential of a natural antioxidant , rutin , to be used as an ancillary treatment to snakebites . Firstly , we investigated the effect of rutin on BjV , since both were pre-incubated prior to injection in mice . Rutin failed to directly inhibit the main protein families tested in vitro , indicating that it modulates the very pathophysiological events evoked by snake envenomation . BjV increased the levels of ROS/RNS and decreased the levels of antioxidants , which are features of redox status imbalance , in mice at least for 24 h , which are in agreement with previous reports [41 , 42] . Recently , ONS has been considered critical for the pathophysiology of snake envenomation [43–45] , and in one prospective study in human patients bitten by Bothrops snakes [10] the levels of oxidative stress markers were altered for long periods of time . Once the presence of ONS is an incipient observation in B . jararaca snake envenomation , the coexistence of the redox status imbalance and the development of hemostatic disturbances led us to wonder whether the former could be interfering or be associated with the latter , as reported elsewhere for other genus of snakes [11 , 12] . In mice , BjV induced various hemostatic and hematological disturbances . Thrombocytopenia is a characteristic manifestation in B . jararaca envenomation , and it is not due to the action of SVMP or SVSP , nor evoked by platelet consumption at the sites of hemorrhagic spots; moreover it is not directly associated with fibrinogen consumption [6 , 23] . Although rutin inhibits platelet aggregation induced by physiological agonists [17 , 18] , our results evidence that rutin failed to directly inhibit BjV-induced platelet aggregation and to protect animals from platelet consumption . Additionally , even when redox status parameters were unaltered during envenomation , thrombocytopenia could be observed , indicating that thrombocytopenia is not directly caused by ONS during envenomation . On the contrary , BjV-induced coagulopathy–characterized by fibrinogen consumption , secondary fibrinolysis , and moderate falls in levels of factor II and X [6 , 23]–was drastically reduced by the use of rutin . Fibrinogen consumption is considered the primary cause of prolonged coagulation time in B . jararaca envenomation [6] , and herein rutin prevented BjV-induced coagulopathy . In addition , SVMP-induced intravascular thrombin generation—by factor II and X activators—has a pivotal role in causing fibrinogen consumption during B . jararaca envenomation in rats [6 , 23] . Our results confirmed that BjV induced an increase in TF activity in plasma in the acute phase of envenomation , possibly triggering the coagulation cascade and contributing thereby to the hemostatic disturbances . Interestingly , the alteration in TF in plasma was not due to an increase in TF antigen concentration , but to an increase in TF activity , suggesting TF decryption . Coagulant and non-hemorrhagic SVMP [46 , 47] might have a role in increasing TF by stimulating monocytes and endothelial cells , and in fact a coagulant SVMP has been shown to modulate TF activity in mononuclear cells in vivo [48] . Thus , our data demonstrate that the raise in circulating levels of TF activity is a remarkable response to BjV-induced systemic injury , possibly leading to true disseminated intravascular coagulation . However , even though there are snake venom enzymes that cause direct consumption coagulopathy , the contribution of TF activation has yet to be determined . Nonetheless , TF activation may be important to hemostatic complications resulting from snake envenomation , as recurrent coagulopathy and thrombocytopenia following antivenom therapy do occur , and may be due not exclusively to low doses of antivenom used in the treatment . It is however still unclear which mechanism is responsible for the activation of TF during envenomation and how important it is to consumptive coagulopathy . Considering that Bothrops snake venoms and their coagulant SVMP can up-regulate TF activity in vitro and in vivo [6 , 46 , 48] , and that rutin prevented coagulopathy , we investigated whether rutin was controlling TF activity . Our initial rationale by choosing rutin was its properties as an antioxidant , a direct inhibitor of thrombin [49] , and an antithrombotic drug that prevents thrombus formation in vivo [8 , 18 , 19] by inhibiting extracellular and membrane-anchored PDI , an enzyme that modulates TF activity . Rutin administration failed to decrease TF activity in plasma , indicating that the amelioration induced by rutin in the coagulopathy of the envenomation is not due to a direct interference in the TF activity in blood stream . Furthermore , it was essential to elucidate whether BjV altered not only blood parameters , but also TF at the site of venom injection and systemically . Since TF is key player in hemostasis , we investigated their protein expression during the late phase of envenomation and how rutin would affect TF . In fact , BjV has been reported to alter TF protein expression in the skin during the acute phase of envenomation ( up to 6h ) [6 , 50] , but herein we showed that BjV no longer induced TF alterations at 24 h . This seems a conservative approach by the organism to maintain hemostasis , since TF must be rigidly controlled to avoid systemic activation of coagulation . On the other hand , surprisingly , rutin not only controlled TF activity , particularly in lungs and heart , but also induced marked alterations in protein expression of endogenous and interest proteins . By Western blotting , rutin itself modified PDI protein expression in liver , lungs , and heart when administered to naive animals ( without BjV administration ) , and thus , we could not evaluate PDI importance during envenomation . However , rutin has various advantageous actions other than in PDI , such as in vascular tonus , endothelium metabolism , and inflammatory reaction , and altogether , independently of the mechanism of action , the current findings support the view that rutin protects animals from the consumptive coagulopathy . Thus , rutin seems a beneficial therapy to be used after B . jararaca envenomation , although extrapolation of our findings to human patients deserve additional studies and the design of clinical trials . It is well established that B . jararaca envenomation induces an acute inflammatory reaction , leading to neutrophilia . Although mice have a lower neutrophil:lymphocyte ratio in blood circulation compared to humans , a mild rise in absolute neutrophil counts in mice , similar to that of human patients , was noticed at 3 h ( equivalent to the mean admission time in human patients at the hospital ) and at 6 h . [4 , 5 , 51] . However , rutin per se also induced neutrophilia , which may be explained by its ability to reduce rolling , adhesion and transmigration of WBC [52 , 53] , justifying the increase in the circulating pool of neutrophils in either the presence or absence of BjV . Therefore , rutin could be considered an anti-inflammatory drug , as it prevents neutrophils of committing themselves into the inflammatory reaction , and could be favorably used to diminish the systemic and local lesions resulting from envenomation . However , even though rutin showed the capacity to augment the circulating pool of neutrophils , they do not seem to be primarily implied in the inflammatory reaction induced by BjV , inasmuch as neutrophil depletion does not remarkably interfere in the natural course of the local inflammatory reaction and local hemorrhage induced by BjV in mice [54] . Furthermore , in patients bitten by B . jararaca snakes , total and differential leukocyte counts do not differ drastically between mild , moderate and severe cases , whose classification depends on the extent and rate of spreading of local swelling at the site of bite , and the occurrence of bleeding and shock [5] . Local injury and bleeding are not exclusively caused by the proteolytic activity of snake venom enzymes , but they also emerge from inflammatory events [32 , 55] . Blood platelets continuously survey vascular integrity [56–59] , and the coexistence of thrombocytopenia and increased neutrophil migration into tissues during B . jararaca envenomation [51 , 60 , 61] might further contribute to blood leakage induced by hemorrhagic toxins and inflammatory mediators at the site of BjV administration [55] . In fact , platelet depletion increases local bleeding , but not edema formation , induced by BjV [62] . Once local bleeding and petechiae development occurs early ( < 3h ) , platelet dysfunction [20] and neutrophil transmigration are likely to have an important role to bleeding development after B . jararaca envenomation , but further studies are necessary to confirm this hypothesis . Furthermore , we also observed a mild fall in erythron values induced by B . jararaca envenomation at 24 h , which was also abrogated by rutin . In humans [5] , this fall has been associated with local and systemic bleedings , which in turn is correlated with the hemostatic disturbance . In rats [23] , such a fall in RBC was related to aforementioned mechanism and microangiopathic anemia [23] . Nonetheless , our results do not support the view that mice also manifested intravascular hemolysis , however , the results confirm the occurrence of bleeding and hemorrhage in the envenomation , which in turn were mitigated by the use of rutin . Several studies have reported the use of plant-based or natural antioxidants to inhibit snake venoms in vitro or to treat snake envenomation [63 , 64] . In fact , the potential of rutin as an ancillary therapy and an anti-hemorrhagic compound was already described in Bothrops envenomation studies [63 , 65] , particularly in a pioneering study from Seba in 1949 [66] , which showed that rutin ingestion retarded or abrogated the development of local hemorrhage and inflammatory reactions induced by Bothrops atrox venom . Importantly , the administration of rutin to envenomed mice led to a tendency to decrease ROS/RNS levels and increase the antioxidant capacity , which is in accordance with its potent activities of scavenging and neutralizing reactive species , and chelating metal ions [52 , 67 , 68] . Other beneficial effects reported for rutin , which could alleviate the inflammatory manifestations provoked by B . jararaca bites include its analgesic and anti-inflammatory effects [17 , 69 , 70] . A limitation of the current manuscript has been not to investigate the activity of rutin when administered after envenomation or concomitantly with Bothrops antivenom . Future studies will certainly focus on its therapeutic use after the venom has initiated the cascade of pathophysiological events , and investigate by which mechanistic action rutin preponderantly inhibits coagulopathy . Using antivenom associated with rutin will be a proof of concept to demonstrate that rutin is really effective in B . jararaca envenomation . In conclusion , important impairments of hemostasis , blood cells and redox status were observed during B . jararaca envenomation . Rutin successfully prevented the development of consumptive coagulopathy , bleeding and local hemorrhage , but failed to mitigate thrombocytopenia . Furthermore , rutin decreased the levels of reactive species and blocked the fall in RBC counts . In lungs and heart , rutin altered TF protein expression and activity . Our results demonstrated that rutin ameliorated coagulation disorders , as well as secondary complications not treated by the antivenom , thus indicating that rutin has indeed a great potential as an ancillary treatment for snakebites . In fact , clinical trials have already been using quercetin and its derivatives as an ancillary therapy to cancer [71–74] and other illness [75] , and high oral doses of quercetin ( up to 1 g/ day during 12 weeks ) showed no toxicity [76] . However , even if quercetin showed no toxicity , further studies are necessary to understand the mechanisms of action and putative side effects of rutin , particularly in humans , prior to evaluating its therapeutic use in snake envenomation . | Snakebite is a neglected disease and a major health issue in tropical countries . Bites by Bothrops snakes ( jararacas ) are frequent throughout South and Central Americas and usually lead to inflammatory and bleeding manifestations in patients , which may be highly severe in some cases . Antivenom therapy is highly effective and the only officially recommended treatment for snakebites , but it is well known that it is not quite effective in blocking various secondary complications . We reasoned that rutin–an easily available and inexpensive plant-based compound–would have the ability to favorably minimize venom-induced low circulating fibrinogen levels , low blood platelet counts , local hemorrhage , bleedings , and raised production of reactive oxygen and nitrogen species during jararaca envenomation . We observed that rutin had remarkably beneficial effects , much higher than we originally expected , in most of the evaluated tests , and such effect was apparently not due to inhibition of the snake venom . Therefore , our results indicate that rutin has a great potential as an ancillary drug in concert with antivenom therapy to treat snakebites , particularly in countries where antivenom availability is scarce . Although we studied the effects of rutin on an experimental model , and further in vivo studies are necessary , our results are extraordinarily encouraging . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"platelet",
"aggregation",
"tropical",
"diseases",
"vertebrates",
"animals",
"toxicology",
"toxic",
"agents",
"protein",
"expression",
... | 2018 | Rutin (quercetin-3-rutinoside) modulates the hemostatic disturbances and redox imbalance induced by Bothrops jararaca snake venom in mice |
NKG2D plays a major role in controlling immune responses through the regulation of natural killer ( NK ) cells , αβ and γδ T-cell function . This activating receptor recognizes eight distinct ligands ( the MHC Class I polypeptide-related sequences ( MIC ) A andB , and UL16-binding proteins ( ULBP ) 1–6 ) induced by cellular stress to promote recognition cells perturbed by malignant transformation or microbial infection . Studies into human cytomegalovirus ( HCMV ) have aided both the identification and characterization of NKG2D ligands ( NKG2DLs ) . HCMV immediate early ( IE ) gene up regulates NKGDLs , and we now describe the differential activation of ULBP2 and MICA/B by IE1 and IE2 respectively . Despite activation by IE functions , HCMV effectively suppressed cell surface expression of NKGDLs through both the early and late phases of infection . The immune evasion functions UL16 , UL142 , and microRNA ( miR ) -UL112 are known to target NKG2DLs . While infection with a UL16 deletion mutant caused the expected increase in MICB and ULBP2 cell surface expression , deletion of UL142 did not have a similar impact on its target , MICA . We therefore performed a systematic screen of the viral genome to search of addition functions that targeted MICA . US18 and US20 were identified as novel NK cell evasion functions capable of acting independently to promote MICA degradation by lysosomal degradation . The most dramatic effect on MICA expression was achieved when US18 and US20 acted in concert . US18 and US20 are the first members of the US12 gene family to have been assigned a function . The US12 family has 10 members encoded sequentially through US12–US21; a genetic arrangement , which is suggestive of an ‘accordion’ expansion of an ancestral gene in response to a selective pressure . This expansion must have be an ancient event as the whole family is conserved across simian cytomegaloviruses from old world monkeys . The evolutionary benefit bestowed by the combinatorial effect of US18 and US20 on MICA may have contributed to sustaining the US12 gene family .
Human cytomegalovirus ( HCMV ) is a clinically important pathogen , which is particularly associated with high levels of morbidity and mortality in immuno-compromised individuals . Systemic HCMV infection results in a higher incidence of graft rejection in transplant recipients and a wide range of end-organ disease including pneumonia , enteritis , hepatitis and retinitis ( specifically in HIV-AIDS ) . The virus is the major cause of congenital birth defects , with long-term sequelae including mental retardation and sensorineural hearing loss [1] . A correlation has been established between infections and two common and aggressive brain tumors ( medulloblastoma and glioblastoma multiforme ) [2] , [3] , which remains controversial [4] , while HCMV has also been implicated in cardiovascular disease , arthritis and in imprinting characteristic changes on the immune repertoire [2] , [5] , [6] . Nevertheless , the vast majority of HCMV primary infections are subclinical and are followed by life-long asymptomatic persistence . Thus , while a competent immune response is unable to eliminate this herpesvirus , in most individuals it is effective at limiting virus replication and preventing disease . HCMV has the largest genome ( ∼236 kbp ) of any characterized human virus and is a paradigm of viral immune evasion . A substantial proportion of its coding capacity is dedicated to evading or modulating immune defenses , and includes genes that target the antigen presentation and processing pathway ( US2 , US3 , US6 , US11 , and miR US4-1 ) [7] , [8] , [9] , [10] , [11] , [12] , [13] , human leukocyte antigen ( HLA ) -G ( US10 ) [14] , T-cell receptor signaling ( UL11 ) [15] , TNF-related apoptosis-inducing ligand ( TRAIL ) death receptor signaling ( UL141 ) [16] , interferon signaling or its downstream effects ( UL83 ) [17] , [18] , [19] and dendritic cell cytokine secretion ( UL7 ) [20] . The virus also encodes interleukin-10 ( UL111A ) [21] and interleukin-8 ( UL146 ) homologs [22] . Natural killer ( NK ) cells play a critical role in the control of HCMV infections; individuals with genetic defects in their NK cell response exhibit extreme susceptibility to the virus [23] , [24] , [25] , [26] . The prognosis of bone marrow and renal transplant patients infected with HCMV infection has recently been demonstrated to correlate with the genotype of specific NK cell receptors ( KIR ) [27] . The significance of the NK cell response is also reflected in the great lengths to which HCMV goes to evade it . To date , UL16 , UL18 , UL40 , UL83 , UL141 , UL142 , and miR-UL112 have all been identified as NK cell evasion functions [28] , [29] , [30] , [31] , [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] , [40] . NK cells constitute a heterogeneous population that express a mosaic of inhibitory and activating receptors , each capable of recognizing and responding to specific ligands presented by potential target cells [41] . The activating receptor NKG2D is remarkable in being expressed ubiquitously on all NK cells and capable of recognizing at least 8 distinct ligands: the major histocompatibility complex class I ( MHC-I ) chain-related molecules ( MICA and MICB ) and the UL16-binding proteins 1–6 ( ULBP1-ULBP6 ) [41] , [42] , [43] . NKG2D ligands ( NKG2DL ) can be induced on the cell surface during times of cellular stress , including genotoxic damage , growth stimulation or viral infection . Within 24 h post infection ( p . i . ) , HCMV activates all NKG2DL with the exception of ULBP4 [44] . More specifically , the HCMV major immediate early ( IE ) proteins IE1 and IE2 ( encoded by genes UL123 and UL122 , respectively ) have been implicated in activating transcription of the MICA/B promoters [45] . HCMV counters this up regulation through ( i ) the sequestration of MICB , ULBP1 , ULBP2 , and ULBP6 in the ER by the UL16 protein [33] , [34] , [35] , [36] , [43] , [46] , ( ii ) the retention of MICA and ULBP3 by the UL142 protein within the cis-Golgi [32] , [47] , [48] , and ( iii ) the microRNA miRUL112 targeting the MICB transcript [28] . The high degree of sequence polymorphism exhibited by both MICA ( 80 alleles ) and MICB ( 33 alleles ) has the potential to provide an added challenge to the virus . Indeed , it has been proposed that this degree of MICA/B diversity may have been selected as a mechanism by which cells evade HCMV infection [48] . We sought initially to explore whether the rapid kinetics of NKG2DL expression during HCMV infection provides a window of opportunity for NK cell recognition prior to the expression of virus-encoded immune evasion proteins . In characterising the up regulation of NKG2DL , we identified a differential effect on MICA and MICB by the HCMV IE gene products , and yet HCMV immune evasion functions were effective in preventing surface expression of these NKG2DL through the early phase of lytic infection . However , the full complement of HCMV immune evasion functions targeting the NKG2DL had not been defined . We now describe the identification and characterization of two novel NK cell evasion genes , HCMV US18 and US20 act individually and in concert to suppress expression cell surface MICA . The US12 gene family comprise some 10 genes arranged sequentially ( US12-US21 ) and uninterrupted through a 9 kb stretch if the US genomic region [49] . US18 and US20 are the first members of this large gene family to be assigned a function .
The regulation and function of NKG2DL exhibits the properties of an intrinsic host immune defense system designed to sense cellular changes associated with danger or pathogen-associated molecular patterns . The existence of eight dedicated ligands provides great scope for the NKG2D receptor to sense and react to a broad range of stimuli . HCMV IE genes have previously been implicated in the transcriptional activation of NKG2Ls [45] , [50] . Therefore , we first examined the capacity of IE1 and IE2 to regulate the various NKG2DL . When expressed using adenovirus ( Ad ) vectors , IE1 induced relatively modest increases in MICA and MICB , but provided for a major up regulation in ULBP2 both at the level of total protein expression and specifically on the cell surface ( Fig . 1A and B ) . In contrast , IE2 induced strong activation of MICA and MICB , yet only a small increase ULBP2 levels . IE1 and IE2 were thus found to differentially activate individual ligands recognized by the NKG2D activating receptor . Consistent with their capacity to up regulate NKG2DL , IE1 and IE2 were each able individually to sensitize cells to NK cell-mediated cytotoxicity ( Fig . 1C ) . We concluded that the infected cell responds to differential sensing of the expression of IE1 and IE2 by promoting the up regulation of ULBP2 and MICA/B , respectively . We hypothesized that HCMV-infected cells may be vulnerable to NK cell surveillance during the early phase if there were a temporal window between the activation of NKG2Ls by IE1 and IE2 and the expression of effective HCMV counter measure . To investigate this possibility , the expression of MICA , MICB , ULBP1 , ULBP2 and ULBP3 was monitored throughout the course of productive HCMV infection ( Fig . 1D , Fig . S1 , Fig . S2 ) . MHC Class-I ( MHC-I ) expression was included as an infection control . At 6 h p . i . , there was no difference between mock- and virus-infected cells in MHC-I or NKG2DL expression . However , cell surface down-regulation of MHC-I expression was clearly observed through the early and late phases of a productive replication cycle ( 24 to 120 h p . i ) . At no stage during the replication cycle did HCMV infection induce an up regulation MICA , MICB , or ULBP2 expression at the cell surface . Thus , we were unable to detect the hypothesized temporal window during which the infected cell may be vulnerable to NK cytotoxicity through activation of NKG2D . As the NKG2DLs must first respond to IE gene expression , they should be activated with similar kinetics to those of a standard HCMV-encoded early gene . To be effective , the corresponding HCMV counter measure must therefore also be expressed with early phase kinetics . UL16 is clearly essential during this process , as its deletion ( strain Merlin ΔUL16 ) increased cell surface MICB , ULBP1 and ULBP2 levels as early as 24 h p . i . ( Fig . 1D , Fig . S1B , Fig . S2 ) . In contrast , deletion of UL142 ( ΔUL142 ) had no overt effect on the expression of any NKG2DL tested , including MICA and ULBP3 ( Fig . 1D , Fig . S1B , Fig . S2 ) . Moreover , fibroblasts infected with the high-passage HCMV laboratory strain AD169 , which lacks UL142 and several other genes , showed a level of control of cell surface MICA ( Fig . S3 ) comparable to that induced by strain Merlin ( Fig . 1D , Fig . S1B ) . The efficient control of MICA expression in cells infected with strain AD169 and strain Merlin ΔUL142 implied the existence of additional HCMV functions capable of targeting MICA . We concluded that , although HCMV activates MICA cell surface expression , this response is counteracted in the context of productive HCMV infection by the action of a previously uncharacterized function . HCMV contains 170 canonical protein-coding genes , of which only 45 are essential for replication in fibroblast cells in vitro [51] , [52] , [53] . In order to map HCMV functions to particular parts of the genome , we generated 10 mutants in which blocks of non-essential genes totaling 56 genes were deleted ( Fig . S4A ) . To facilitate detection of novel NK evasion functions , these mutants were generated on a strain Merlin background that already lacked UL16 and UL18 ( ΔUL16ΔUL18 ) and contained a UL32-GFP fusion reporter . Screening cells infected with these block deletion mutants identified one mutant lacking US18–22 ( ΔUS18–22 ) , which led to a marked increase in cell surface MICA detected by flow cytometry ( Fig . 2A ) . This mutant was also found to replicate less efficiently than the parent virus ( data not shown ) . Infection with this mutant also resulted in a very marked increase in the total cellular abundance of MICA/B ( Fig . 2B ) . Specific antibodies to other NKG2DLs were then used to see if this effect was specific by flow cytometry . As the block deletion mutants were made on a ΔUL16ΔUL18 background , the parental HCMV control up regulated MICB and ULBP2 relative to mock-infected cells ( Fig . 2C , Fig . S4B ) , whereas the ΔUS18–22 mutant up regulated cell surface levels of MICA ( Fig . 2C , Fig . S4B ) relative to both mock- and parental HCMV-infections . These results suggest the potential for an HCMV NK cell evasion function that acts by promoting degradation of an NKG2DL . To map the MICA-suppressing function more precisely , the 5 genes in the US18–US22 region were expressed individually in fibroblasts using an Ad vector , and expression monitored by detection of the V5 tag by intracellular flow cytometry ( Fig . S5 ) . Expression of US18 or US20 led to reductions in MICA/B levels by immunoblotting ( Fig . 3A ) . Interestingly , expression of the US18–US22 genes individually was each associated with a marginal reduction in MICA cell surface expression; thus genes adjacent to US18 and US20 could also be contributing to the suppression of MICA ( Fig . S6 ) . In the context of an HCMV infection , deletion of US18 and US20 led to an increase MICA/B levels as monitored by immunoblotting ( Fig . 3B ) . Deletion of both US18 and US20 led to a further increase in MICA/B levels ( Fig . 3B ) . No differences in MICA/B glycosylation were observed in cells infected with the different HCMV US18 and US20 deletion mutants when samples were treated with Endoglycosidase H/EndoH ( E ) or Peptide N-Glycosidase F/PNGase F ( P ) ( Fig . 3B ) . Almost all of the MICA/B was EndoH-resistant and therefore consistent with its localization in a post-ER compartment ( e . g . cis-Golgi , lysosomes , or cell surface expression ) . By flow cytometry , deletion of US18 alone had no discernible effect on cell surface expression of MICA , MICB , ULBP2 or MHC-I ( Fig . 3C , Fig . S7 ) . In contrast , deletion of US20 induced a modest but significant increase in surface levels of MICA compared to parental virus , but MICB , ULBP2 and MHC-I expression were unchanged . However , deletion of both US18 and US20 caused a much more dramatic increase in MICA levels , similar to that observed for US18–22 block deletion mutant ( Fig . 2C and 3C ) . These mutants had no appreciable difference in replication efficiency ( data not shown ) . Thus , absence of either US18 or US20 was partially compensated by the other gene , whereas absence of both genes had a more than additive effect on MICA expression . The HF cell lines ( HF-TERT and HF-CAR ) were immortalized lines derived from the same donor HFs , which have a MICA genotype , MICA*016/027 [54] . The same pattern of MICA cell surface expression was obtaining using the primary parent HF cell line infected with the US18 and US20 deletion mutants ( data not shown ) . To determine whether US18 and US20 could target a range of MICA alleles , we performed a similar analysis in a number of donor dermal fibroblast cell lines with different MICA genotypes . These experiments revealed a greater dependence on US20 for regulating MICA , as there was little ( D43 , MICA*007/010 ) or no difference ( D45 , MICA*002/009 and NP , MICA*004/004 ) between the single US20 deletion mutant and US18 and US20 double deletion mutant ( Fig . 3D , Fig . S8 ) . To complement findings obtained with exogenous US18 and US20 gene expression , the equivalent C-terminal V5 antigenic tag was added to the genes in the context of the HCMV genome . In contrast to transcriptomic studies [52] , pUS18-V5 expression was expressed at lower levels than pUS20-V5 at all stages of the HCMV replication cycle , as gauged by the intensity of staining from the V5 epitope tag ( Fig . 4A ) . Therefore , functional differences between the US18 and US20 single HCMV deletion mutants ( Fig . 3C ) may reflect this difference in detectable expression levels , or perhaps pUS18 is subject to a more rapid cellular turnover than pUS20 . However , comparable expression between pUS18 and pUS20 was observed when using Ad vectors ( Fig . 4B ) . Analysis of pUS18 and pUS20 expression by immunofluorescence staining in the context of HCMV revealed that both proteins were located in subcellular punctate structures ( Fig . 4C ) . This pattern of expression was similar to that observed using the US20 RAd ( Fig . 4D ) , but contrasted with the ER-like expression pattern detected for the US18 RAd ( Fig . 4D ) . We concluded that , whether expressed from Ads or in the context of HCMV , both proteins reside within an intracellular compartment from which they are able to regulate MICA expression . Next , we sought to determine the fate of MICA in HCMV-infected cells . Cells were infected with HCMV strain Merlin in combination with chemical inhibitors of the major protein degradation pathways that act via the proteasome ( MG132 ) or lysosome ( folimycin ) ( Fig . 5A and B , Fig . S9A ) . Inhibition of lysosomal , but not proteasomal , degradation greatly increased the cellular levels of MICA/B as assessed by immunoblotting ( Fig . 5A ) . Addition of either MG132 or folimycin had no effect on cell surface MICA or MICB expression in either mock or HCMV-infected cells ( Fig . 5B , Fig . S9A ) . Treatment with a further two lysosomal inhibitors , leupeptin ( a protease inhibitor ) and chloroquine ( an inhibitor of lysosomal acidification ) led to an increase in MICA/B levels by immunoblotting ( Fig . 5C ) , but no effect on cell surface MICA or MICB expression by flow cytometry ( Fig . 5D , Fig . S9B ) . Folimycin and chloroquine affect the acidification of the lysosomes and the Golgi apparatus , while leupeptin is a protease inhibitor targeting lysosomal proteases . Therefore , these inhibitors will prevent degradation of proteins within the lysosome by preventing the action of proteolytic enzymes requiring an acidic pH , but will not redirect them from this compartment to the surface . Therefore HCMV appears to target MICA for lysosomal degradation , and this protein is retained within the cell following lysosomal inhibition . To examine further the effect of these genes on MICA , the genes were delivered to a U373 cell line expressing a MICA-yellow fluorescent protein ( YFP ) fusion protein ( Fig . 6 ) . MICA-YFP expression was suppressed efficiently by US18 or US20 , but could be restored by treatment with the lysosomal inhibitor , folimycin ( Fig . 6 ) . In the presence of folimycin , US18 expression caused MICA-YFP to be redistributed to punctate intracytoplasmic structures , in which MICA-YFP was co-localized with the US18 protein ( pUS18; Fig . 6 ) . Although US20 expression also induced MICA-YFP to traffic to similar structures , the US20 protein ( pUS20 ) did not localize to this compartment ( Fig . 6 ) . Expression of US19 as a control had no effect on MICA distribution ( Fig . 6 ) . To investigate whether these intracellular structures represented lysosomes , we performed similar experiments using a lysosomal staining reagent ( Lysotracker Red DND-99 ) ( Fig . 7A ) . These experiments showed that the MICA-YFP signal in both pUS18- and pUS20-expressing cells treated with folimycin co-localized with the lysosomal staining . The pUS18 staining was also lysosomal , whereas that of pUS20 was not . We also investigated whether pUS18 and pUS20 co-localized with the lysosomes in the context of HCMV infection ( Fig . 7B ) . This analysis did not clearly reveal an association of pUS18 and pUS20 staining with that of lysosomes , although there was some co-localization in a proportion of cells with a strong cytopathic effect . In many cells , pUS20 appeared to be present in an intracellular compartment adjacent to the lysosomal staining . We concluded that both proteins promote proteolysis of MICA in the lysosome . Before designating a virus gene as being an NK cell evasion function , it is important to monitor its biological activity during infection . The effects of US18 and US20 on NK cell recognition were therefore analyzed using the HCMV deletion mutants described above ( Fig . 8A , Fig . S10 ) . Relative to uninfected cells , infection with strain Merlin elicited robust protection against NK cells in all donors tested . A significant increase in NK cell degranulation was associated with loss of US18 or US20 ( Fig . 8A ) or the US18–22 ‘block’ deletion ( Fig . S10B ) , and an additive effect was observed when both genes were absent . We concluded that US18 and US20 are effective in suppressing NK cell activation in the context of a productive HCMV infection . The use of a MICA blocking antibody led to a small decrease in NK degranulation in response to targets infected with the US18 and US20 double deletion virus , whilst the same antibody had no effect on NK activation in response to Mock or HCMV-infected targets ( Fig . 8B ) . These data suggest that either the blockade of MICA was incomplete or that US18 and US20 target other cellular molecules capable of regulating NK cell activation .
HCMV-infected fibroblasts are extraordinarily resistant to NK cells as assessed in vitro by cytolysis or CD107-mobilization assays [29] , [30] . Nonetheless , NK cells play a critical role in combating infections in vivo . We set out to determine whether the rapid up regulation of NKG2DL transcription driven by expression of the major HCMV IE genes would result in transient exposure of the activating ligands MICA , MICB , or ULBP2 early during infection [44] , [45] , thereby creating a window of opportunity for NK cell recognition . These studies revealed an interesting differential effect; HCMV IE1 preferentially unregulated the NKG2DL ULBP2 , whereas IE2 predominantly activated MICA/B . Although IE1 and IE2 are encoded from within the same transcriptional unit by differential splicing , they are functionally distinct . IE1 is not essential for viral replication , but enhances replication by targeting intrinsic barriers to viral infection ( PML-bodies , hDaxx , STAT-2 , and p107 ) , whereas IE2 is essential and is a potent transcriptional transactivator [55] . It is logical that individual NKG2DL should be differentially sensitive to specific triggers of cellular stress . Indeed , the dichotomy presented by the major HCMV IE genes provides a tractable experimental system by which to explore the cellular mechanisms that underpin NKG2DL regulation . Despite their up regulation by IE genes , HCMV was observed to suppress NKG2DL expression efficiently right through the early and late phases of the replication cycle . Activation of NKG2DL expression during infection could only be detected when the viral genome had been engineered to delete relevant NK cell evasion functions; studies using such deletion mutants elegantly reveal the relative contributions made by individual NK cell evasion genes in mounting a comprehensive defense . While the control of MICB , ULBP1 and ULBP2 by UL16 was fully consistent with previous findings [46] , unexpectedly the UL142 deletion mutant had no overt effect on MICA or ULBP3 expression . Strain AD169 controlled MICA expression despite containing a 15 kbp deletion that includes UL142 ( this study; [56] ) . These observations prompted us to search for additional HCMV functions capable of targeting MICA , which were mapped to US18 and US20 using a combination of HCMV ‘block’ deletion mutants on a ΔUL16 , ΔUL18 , UL32-GFP background and adenovirus vectors expressing individual HCMV genes . Human viruses are known to eliminate host proteins selectively by recruiting cellular proteasomal or lysosomal proteolytic pathways . A specific role for the lysosome in the proteolytic degradation of MICA in HCMV-infected cells was identified . The functions of US18 and US20 were validated in the context of HCMV productive infection . Deletion of US18 and US20 individually was associated with either no effect or a modest but significant increase in MICA cell surface expression relative to HCMV-infected cells , and both of these deletions enhanced NK cell activation in a functional assay . pUS18 and pUS20 are thus able individually to compensate for each other's loss in regulating MICA . Deletion of both genes together reduced the efficiency of MICA down regulation in a more than additive fashion , and yet consistently resulted in only a modest increase in NK cell activation above that observed for the single US18 and US20 deletions . One possible explanation of this is that the MICA antibodies used may not recognize all glycosylated isoforms of MICA with equal efficiency . Also , NK cells are controlled by thresholds , and thus cannot be expected to respond with linear kinetics to levels of MICA expression . Neither US18 nor US20 has been assigned a function previously . However , both genes belong to the HCMV US12 family , which consists of ten tandemly arranged genes ( US12–US21 ) encoding distantly related seven-transmembrane domain proteins [46] . A low level of amino acid sequence similarity has been noted between some US12 gene family members and the transmembrane BAX-inhibitor motif containing protein ( TMBIM ) superfamily [57] , [58] , [59] . The 6 TMBIM family proteins ( TMBIM1–6 ) have functions relating to apoptosis and regulation of ER stress [60] , and TMBIM6 ( BAX-inhibitor 1 ) regulates lysosomal degradation of Gb3 and P450 2E1 [61] , [62] . Although a poxvirus-encoded TMBIM4 homolog ( viral GAAP , Golgi anti-apoptotic protein ) has been shown to suppress apoptosis [63] , no such function has yet been assigned to any US12 family member . The significance , if any , of the similarity of US12 family proteins to TMBIM family proteins in evolutionary and functional terms remains unknown . The same evaluation applies to the more marginal similarities to G protein-coupled receptors [53] . The HCMV genome contains 15 families of related genes , each potentially acquired via a process of gene duplication [51] , [64] . The US12 family may have arisen as a genomic “accordion” , which involves the rapid expansion of a single gene under a strong selective pressure , to an array of related genes , which may collapse subsequently if the selective pressure wanes . This feature was recently demonstrated experimentally in a poxvirus ( vaccinia virus ) to facilitate selective adaption to the host's intrinsic antiviral defences [65] . Expansion of the US12 family is likely to have long pre-dated the speciation of humans , as the family is well conserved in cytomegaloviruses of chimpanzee and Old World primates ( Rhesus and Cynomolgus macaques ) , which encode the same genomic arrangement of 10 recognizable US12 family genes . Several US12-related genes are also present in cytomegaloviruses of New World primates ( Green Monkey and Owl Monkey ) , although these vary in number ( 11 and 7 genes respectively ) [49] , [51] , [66] , [67] , [68] , [69] . In support of a genomic accordion expansion , the ten US12 family members ( US12–US21 ) are encoded tandemly in a gene cluster that is transcribed in three transcriptional cassettes: US12–US17 , US18–US20 , and US21 independently [70] , [71] , [72] . The presence of the US12 family in primate cytomegalovirus genomes implies that the selective pressure that initially induced the expansion has been maintained in some form , and the various numbers and arrangements of US12-related genes indicates that the family has continued to adapt and diverge in order to acquire an expanded functional range . Interestingly , the US6 gene family ( US6–US11 ) contains multiple members ( US6 , US10 , and US11 ) that target the MHC-I antigen-processing pathway , with the individual proteins targeting various parts of the pathway in order to provide greater effectiveness and resistance to mechanisms of host resistance [13] , [14] , [73] , [74] . In this context , it is interesting to note that efficient down regulation of MICA requires both US18 and US20 . Further functional studies may provide additional insights into the evolution and functions of the US12 family . US18 and US20 now join with UL16 , UL142 and miR-UL112 in expanding to five the set of HCMV genes that have been demonstrated to counter the action of the single NK cell activating receptor , NKG2D . In addition to its role in regulating the function of NK cells , NKG2D is also expressed ubiquitously on γδ-T cells and a subset of αβ-T cells . The number of HCMV-encoded gene products targeting the NKG2D pathway emphasizes the importance of this activating receptor in the immune response to HCMV .
Healthy adult volunteers provided blood and dermal fibroblasts for this study following written informed consent ( approved by the Cardiff University School of Medicine Ethics Committee Ref . no: 10/20 ) . Human fetal foreskin fibroblasts immortalized by human telomerase ( HF-TERT ) , HF-TERTs transfected with the Coxsackie-adenovirus receptor ( HF-CAR ) , donor dermal fibroblasts ( primary and TERT-immortalized ) , and U373 MICA-YFP-expressing cells ( were maintained at 37°C in 5% CO2 in growth medium ( Dulbecco's minimal essential medium ( DMEM ) supplemented with penicillin/streptomycin and 10% fetal calf serum ( Invitrogen , Paisley , UK ) ) [75] . U373-MICA-YFP cells were constructed by transfecting full-length MICA allele ( Genbank Accession no . AAD52060 ) fused to YFP , ( a gift from Professor Dan Davis , University of Manchester , UK , generated as previously described [76] ) into U373 cells with Effectene ( Qiagen , Manchester , UK ) and adding drug selection using 0 . 75 mg/ml G418 ( Invitrogen ) . DNA was extracted from donor blood samples using the Qiagen Blood and tissue DNA extraction kit . MICA typing was performed by the Anthony Nolan Trust , as previously described [54] . HCMV deletion mutants were generated by recombineering of the bacterial artificial chromosome ( BAC ) of HCMV strain Merlin ( GenBank accession number GU179001 . 1 ) , as described previously [77] . Strain Merlin contains the complete genetic complement of HCMV , and is frame shifted in two genes ( RL13− , UL128− ) . The regions encompassing the sites recombineered in BACs were verified by PCR and sequencing . A list of HCMV constructs , indicating deleted genes , other modifications , and the primers used in their production , is shown in Text S1 . The HCMV strain AD169 ( varUK; Genbank accession number BK000394 ) previously described was also used in some experiments [78] , [79] . Recombinant adenoviruses ( RAds ) were generated as described previously [80] . Briefly , HCMV genes were amplified from the strain Merlin BAC by using primers containing arms of homology to the adenovirus BAC vector ( pAL1141 ) and recombineered into pAL1141 . RAd-IE1 and IE2 contained IE1 and IE2 from HCMV strain AD169 cloned into the adenovirus BAC vectors . A list of RAds used and the primers used to generate them is shown in Text S2 . Cells were seeded in growth medium at appropriate cell densities ( 1×106 cells for a 25 cm2 flask , 5×105 cells per well for a 6-well plate , 5×104 cells per well for a 24-well plate ) . The following day , the cells were infected with virus at the required multiplicity of infection in an appropriate volume of growth medium ( 2 ml for a 25 cm2 flask , 1 ml per well for a 6-well plate , 250 µl per well for a 24-well plate ) for 2 h on a rocker at 37°C in 5% CO2 . The inoculum was then replaced with fresh growth medium ( 7 ml for a 25 cm2 flask , 4 ml per well for a 6-well plate , 1 ml per well for a 24-well plates ) , and the cells were incubated for the required times . Fetal calf serum was omitted from the growth medium for HCMV infections . For inhibitor studies , cells were treated 12 h prior to harvesting with proteasomal ( MG132 10 µM ) or lysosomal inhibitors ( folimycin 1 µM , leupeptin 200 µM or chloroquine 100 µM ) in DMEM . Cells were harvested by washing once with phosphate-buffered saline ( PBS ) and treating with 1× trypsin/EDTA for 1 min at 37°C . After neutralizing the trypsin with growth medium , the cells were washed once in flow cytometry ( FC ) buffer ( 1% bovine serum albumin and 0 . 05% sodium azide in PBS ) . The cells were resuspended in an unconjugated primary antibody ( murine IgG , Sigma Aldrich , Poole , UK , 1∶1000 dilution in FC buffer; anti-MICA clone AM01 , anti-MICB clone BM02 , anti-MICA/B clone BAM01 , anti-ULBP2 clone BUM01 , BAMOMAB GmBH , Graefelfing , Germany , 1∶400 dilutions in FC buffer; anti-ULBP1 Clone 170818 , MAB1380 , anti-ULBP3 Clone 166510 , MAB1517 , R&D Systems , Abingdon , UK , 1∶200 dilutions in FC buffer; anti-MHC-I , clone W632 , AbD Serotec , Kidlington , UK , 1∶2000 dilution in FC buffer ) and incubated for 30 min at 4°C . The cells were then washed twice with FC buffer and incubated in an Alexa Fluor 647 conjugated anti-mouse IgG secondary antibody ( Invitrogen , 1∶500 dilution in FC buffer ) for 30 min at 4°C . After 3 further washes in FC buffer , the cells were fixed in 2% paraformaldehyde ( PFA ) for 10 min and analyzed by using an Accuri Cflow cytometer ( BD Biosciences , Oxford , UK ) . A forward scatter and side scatter dot plot was used to gate both on viable cells and infected cells ( Fig . S1A ) . The median fluorescence intensity ( MFI ) was used in subsequent analysis . Cells were washed once with ice-cold PBS , scraped into 4 ml ice-cold PBS , recovered by centrifugation ( 1 , 600 rpm for 3 min ) , and stored at −20°C . They were then resuspended in 300 µl Triton X-114 extraction buffer ( 2% Triton X-114 , 1∶100 protease inhibitors , and 1 mM DTT in PBS ) and sonicated . The lysate was clarified by microcentrifugation at 13 , 000 rpm for 1 h at 4°C in a 1 . 5 ml non-stick tube . The supernatant was transferred to a fresh 1 . 5 ml non-stick tube , incubated at 37°C for 10 min , and microcentrifuged at 13 , 000 rpm for 5 min at 37°C . The upper phase was removed , leaving a bead of Triton-X114 of approximately 60 µl . PBS was added up to a total volume of 300 µl and the samples sonicated . The extracted protein was precipitated by adding 1 . 2 ml precipitation reagent ( Merck protein precipitation kit 539180 ) and incubating at −20°C for 1 h , and collected by centrifuging at 13 , 000 rpm for 10 min at 4°C and aspirating the supernatant . The precipitated proteins were washed once with 500 µl wash buffer ( Merck protein precipitation kit 539180 ) and collected by centrifuging at 13 , 000 rpm for 10 min at room temperature and aspirating the supernatant . The pellets were allowed to air dry for a few minutes and then resuspended either in 1× Nupage gel sample buffer ( Invitrogen ) plus 10 mM DTT for direct immunoblotting or in 1× denaturing buffer ( NEB ) for EndoH and PNGase F treatment . Samples were denatured at 95°C for 10 mins or for analysis of US18 and US20 expression at 50°C for 10 mins . Denaturation at 95°C was found to lead to high molecular weight smears of US18 and US20 in resulting immunoblots , presumably due to formation of aggregates as noted for other polytopic membrane proteins . Protein samples ( 20 µl ) were separated on ready-made 10% Nu-PAGE polyacrylamide gels ( Invitrogen ) by SDS-PAGE and transferred to Hybond-P nitrocellulose ( GE Life Science ) by semi-dry blotting . Nitrocellulose membranes were prepared by treating with 5 ml Pierce MISER antibody extender ( Fisher Scientific ) for 10 min and washing 7 times with distilled water . They were then blocked in 5% milk in Tris-buffered saline containing 0 . 05% TWEEN-20 and 0 . 05% Triton X-100 ( TBS-T-T ) overnight at 4°C . The membranes were then incubated with primary antibodies diluted in 5% milk in TBS-T-T overnight at 4°C . They were then washed 5 times for 5 min in TBS-T-T , and incubated in anti-mouse IgG-HRP conjugate ( Insight Biotechnology Ltd , Wembley , UK , 1∶5 , 000–1∶10 , 000 ) for 1 h at room temperature . After washing a further 5 times for 5 min in TBS-T-T , the membranes were incubated for 5 min in SuperSignal West Pico Chemiluminescent substrate ( Fisher Scientific ) before being exposed to Hyperfilm-MP film ( GE Life Science , Little Chalfont , UK ) for development . The blots were stripped in Pierce Stripper buffer ( Fisher Scientific ) for 10 min , washed 7 times in TBS-T-T , reblocked , and reprobed . Cells were grown and infected with virus in glass-bottomed 24-well plates . At various times p . i . , the cells were washed with PBS , and fixed with 2% PFA for 10 min . In some instances , cells were incubated with LysoTracker Red DND-99 ( Cat . no . L7528 , 1∶1000 in complete DMEM , Invitrogen ) for 30 min , washed once with complete DMEM , then PBS and and fixed with 2% PFA for 10 min . The fixed cells were washed twice with IC buffer ( 0 . 2% saponin , 1% BSA , 0 . 05% sodium azide in PBS ) and incubated with primary antibodies ( anti V5-tag , AbD Serotec ) diluted in IC buffer for 1 h at 4°C . The cells were then washed 3 times with IC buffer and incubated with Alexa Fluor 594 or 350 ( Lysotracker experiments ) -conjugated anti-mouse IgG antibodies ( Invitrogen ) for 1 h at 4°C . The cells were washed once with IC buffer and where required incubated for 10 min with DAPI ( 0 . 5 µg/ml ) diluted in IC buffer . They were then washed twice further in IC buffer , before the addition of 2% PFA and analysis by fluorescent microscopy . Cells were harvested by washing once with phosphate-buffered saline ( PBS ) and treating with 1× trypsin/EDTA for 1 min at 37°C . After neutralizing the trypsin with growth medium , the cells were washed with PBS and fixed with 2% PFA for 10 mins . The fixed cells were washed twice with IC buffer and incubated with anti-V5 tag antibody ( 1∶2000 ) or control mouse IgG diluted in IC buffer for 1 hr at 4°C . The cells were washed twice with IC buffer and incubated with Alexa Fluor 647-conjugated anti-mouse IgG antibodies ( 1∶500 ) for 30 mins at 4°C . They were then washed three times in IC buffer , before the addition of 2% PFA and analysis by flow cytometry . NK cytotoxicity was assessed by standard 51Cr chromium release assay as described previously [39] , using labelled RAd-infected fibroblast target cells . The effector∶target ( E∶T ) ratio was adjusted by using the number of NK cells present in the CD3+-depleted IFN-α-activated peripheral blood mononuclear cells ( PBMC ) used as effectors . Specific lysis ( % ) was calculated as [ ( experimental mean release - spontaneous mean release ) / ( maximum mean release - spontaneous mean release ) ]×100 . The mean and SEM were determined from the results of triplicate or quadruplicate samples . NK degranulation assays were performed in a similar manner to that described previously [39] , [81] . Briefly , PBMC ( approved by the Cardiff University School of Medicine Ethics Committee Ref . no: 10/20 ) and incubated overnight with IFN-α ( 1000 IU/ml ) and IL-15 ( 15 ng/ml , Milenyi Biotech ) . PBMC ( 0 . 5–1×106 ) were incubated for 6 h with 0 . 5–1×105 fibroblast targets per well in a 96 well plate at an effector∶target ( E∶T ) ratio of 10∶1 , with the addition of 3 µl per well FITC-conjugated anti-CD107 antibody ( cat . no . 555800 , clone H4A3 , BD Biosciences ) or 3 µl per well FITC-conjugated isotype control ( cat . no . 555748 , BD Biosciences ) , adding 1 µl/well BD GolgiStop ( BD Biosciences ) 1 h after beginning the incubation . ( For antibody blocking experiments , targets were pre-incubated with anti-MICA ( Clone 159277 , mouse IgG2B , MAB1300 , R&D Systems ) or isotype control ( MICB non-blocking antibody , Clone 236511 , mouse IgG2B , MAB1599 , R&D Systems ) antibodies at a concentration of 10 µg/ml for 30 min prior to incubation with PBMC ) . PBMC were harvested and stained with conjugated antibodies against CD3 ( anti-CD3 PE-Cy7 , cat . no . 737657 , Beckman Coulter , High Wycombe , UK ) and CD56 ( anti-CD56 PE , cat . no . A07788 , Beckman Coulter ) , and fixed in 2% PFA before analysis by flow cytometry ( BD Biosciences Accuri C Flow ) ( Fig . S10A ) . | Human cytomegalovirus ( HCMV ) is a herpesvirus that infects most people in the world , usually without producing symptoms . However , infection is life-long and must be kept in check by the immune system . When the immune system is weakened , the outcome of HCMV infection can be very serious . Thus , HCMV is the major cause of birth defects resulting from infection of the fetus during pregnancy , and it can cause severe disease in people with a weakened immune system , especially transplant recipients and HIV/AIDS patients . One type of immune cell , the natural killer ( NK ) cell , is crucial in controlling cells in the body that are abnormal . They do this by recognizing cells , which have special stress proteins on their surface , and killing them . When cells are infected with HCMV , they start to make these stress proteins . However , the virus has evolved ways to stop NK cells from killing infected cells by quickly stopping the stress proteins from reaching the surface . We have now identified two HCMV genes that target a major stress protein ( called MICA ) and cause its rapid destruction . Removing these two genes from HCMV renders infected cells very susceptible to killing by NK cells . This discovery might help the development of new ways to fight HCMV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"virology",
"biology",
"and",
"life",
"sciences",
"microbiology",
"viral",
"diseases"
] | 2014 | Two Novel Human Cytomegalovirus NK Cell Evasion Functions Target MICA for Lysosomal Degradation |
In everyday life , we have to decide whether it is worth exerting effort to obtain rewards . Effort can be experienced in different domains , with some tasks requiring significant cognitive demand and others being more physically effortful . The motivation to exert effort for reward is highly subjective and varies considerably across the different domains of behaviour . However , very little is known about the computational or neural basis of how different effort costs are subjectively weighed against rewards . Is there a common , domain-general system of brain areas that evaluates all costs and benefits ? Here , we used computational modelling and functional magnetic resonance imaging ( fMRI ) to examine the mechanisms underlying value processing in both the cognitive and physical domains . Participants were trained on two novel tasks that parametrically varied either cognitive or physical effort . During fMRI , participants indicated their preferences between a fixed low-effort/low-reward option and a variable higher-effort/higher-reward offer for each effort domain . Critically , reward devaluation by both cognitive and physical effort was subserved by a common network of areas , including the dorsomedial and dorsolateral prefrontal cortex , the intraparietal sulcus , and the anterior insula . Activity within these domain-general areas also covaried negatively with reward and positively with effort , suggesting an integration of these parameters within these areas . Additionally , the amygdala appeared to play a unique , domain-specific role in processing the value of rewards associated with cognitive effort . These results are the first to reveal the neurocomputational mechanisms underlying subjective cost–benefit valuation across different domains of effort and provide insight into the multidimensional nature of motivation .
Neuroeconomic theories highlight that a key component of motivation is evaluating whether potential rewards are worth the amount of effort required to obtain them [1 , 2] . Behaviours are executed if they have sufficient “subjective value” ( SV ) , which is based on how much a potential reward is discounted—or devalued—by the effort required to obtain that outcome [3] . A characteristic of these cost–benefit valuations is that they are inherently highly subjective and thus vary across individuals [4–6] . Some people are willing to invest a quantum of effort for a reward that others would not . However , not all types of effort are subjectively evaluated in the same manner . Some individuals may be willing to overcome physically demanding challenges but be averse to mental effort , while others might show the opposite profile . Understanding the mechanisms that underlie cost–benefit valuations across different domains of effort is crucial to understanding the variability in people’s motivation [7 , 8] , but little is known of the neural or computational basis of these mechanisms . Current theories of value-processing suggest that the computation of SV occurs in a common , domain-general network of brain regions [9] . Single-cell and neuroimaging studies have implicated areas within the basal ganglia and parieto-prefrontal cortices in the computation of SV for rewards that are devalued by costs such as risk , delays , or probability [9 , 10] . Separately , research on effort-based decision making has implicated areas including the anterior cingulate cortex ( ACC ) ( area 32’ ) , dorsolateral prefrontal cortex ( dlPFC ) ( areas 8/9 ) , anterior insula ( AI ) , intraparietal cortex ( area 7 ) , and several amygdala nuclei [11–16] . However , critical unanswered questions are whether these effort-sensitive areas compute the subjective value of discounted rewards associated with effort costs and whether these areas are differentially sensitive to the nature of those costs . To date , most research on effort-based decision making has either focused on the cognitive or physical domains in isolation [4 , 17–19] . The only previous study to have examined the neural mechanisms associated with different types of effort required participants to perform a cognitively or physically demanding task [20] . Importantly , however , participants in that study were not engaged in the choice of whether it was worthwhile to invest effort for reward . Thus , although this study was useful in examining how the brain motivates the exertion of different effort costs , the neural substrates that underlie the subjective valuation of reward—and the decision of whether to engage in an effortful action—remain unknown . Increasingly , these decision processes are being recognised as a critical component of motivated behaviour , with evidence that aberrant effort-based decision making may be a key element of motivational disorders such as apathy [18 , 19 , 21] . Here , we used the computation of SV as a key operation to understand cost–benefit decision making across the domains of cognitive and physical effort [9 , 22–24] . In contrast to classical accounts , recent research in animals suggests that the mechanisms that underpin cognitive and physical effort discounting might be separable . For example , animal studies of the amygdala have causally linked it to motivation and the devaluation of reward by effort costs [25 , 26] . Recently , however , a novel rodent decision-making task showed dissociable effects of amygdala and frontal lesions on cognitive effort–based decisions [27] . Specifically , amygdala and ACC inactivations caused changes to behaviour during a cognitive effort task [27] that were different to those in physical effort tasks [2 , 25 , 26] . Furthermore , amygdala inactivation influenced individual animals differently , suggesting that the amygdala may play a distinct role in subjectively valuing rewards associated with cognitive effort . Such findings suggest that the computation of SV in the context of effort may not be within a domain-general network of valuation areas , as is often argued [9] . To establish whether the SV of different effort costs are processed within domain-general or domain-specific brain systems , the current study directly examined whether the neural mechanisms underlying subjective reward valuation are sensitive to different types of effort . We first trained participants on two tasks that were closely matched on many properties that are known to influence the valuation of a reward ( e . g . , probability , duration prior to outcome ) [28] but differed in whether cognitive or physical effort was required to obtain rewards . In each , we parametrically varied effort in one domain while holding the demands of the other constant . Then , while being scanned with functional magnetic resonance imaging ( fMRI ) , participants chose between a fixed low-effort/low-reward “baseline” option and a variable higher-effort/higher-reward “offer . ” Central to our paradigm was the use of computational models to calculate the SV of each effort and reward combination relative to the baseline option for individual subjects , which allowed us to calculate subject-specific discounting parameters for each of the cognitive and physical effort tasks . Using model-based fMRI , we then identified regions in which blood oxygen level–dependent ( BOLD ) activity correlated with these parameters . This revealed that cognitive and physical effort discounting occurred in largely overlapping neural areas , but in addition , the right amygdala contributed uniquely to cognitive effort valuation .
In the cognitive effort task [4] , we employed a rapid serial visual presentation ( RSVP ) paradigm [29] , in which participants fixated centrally while monitoring one of two target streams to the left and right of fixation for a target number “7” ( Fig 1A ) . Each target stream was surrounded by three distractor streams . The target stream to be monitored was indicated at the beginning of the trial by a central arrow and , during the trial , participants had to simultaneously monitor the central stream for a number “3 , ” which would be a cue to switch their attention to the opposite target stream . We parametrically varied the amount of cognitive effort over six levels by increasing the number of times attention had to be switched between streams from one to six . We previously confirmed that this task was able to manipulate perceived cognitive effort while controlling for physical demands and reinforcement rates [4] . In the physical effort task , participants exerted one of six different levels of force on a handheld dynamometer ( Fig 1B ) . The effort levels for each participant were defined as proportions of their individually calibrated maximum voluntary contraction ( MVC ) ( 8% , 13% , 18% , 23% , 28% , and 33% ) , as determined at the beginning of the experiment . The duration of each of the cognitive and physical effort trials was identical ( 14 s ) , ensuring that participants’ choices were not due to temporal discounting [30 , 31] . Participants were first trained on each of the cognitive and physical effort tasks outside the scanner in counterbalanced order . They undertook an extensive training session of 60 trials for each task to familiarise themselves with the effort associated with each level in each domain and so that we could estimate performance measures for each task ( see Materials and Methods ) . Participants were told that their reimbursement at the end of the study would be contingent on performance and that for each trial that they performed well , they would be awarded one credit , which would be later converted into a monetary amount . This training resulted in participants being rewarded on over 80% of trials , and a repeated-measures ANOVA revealed that , although there was a significant effect of effort ( F ( 1 . 7 , 57 . 2 ) = 7 . 48 , p < . 005 ) , neither the main effect of domain nor its interaction with effort were significant ( p > . 05; S1 Fig ) . Importantly , this indicates that the reinforcement rates did not differ between tasks and ensured that subsequent effort-based decisions in the two domains could not be confounded by participants’ belief that they would be differentially successful at obtaining rewards across the two tasks . The critical choice phase occurred after the training phase , while participants were being scanned with fMRI ( Fig 1C ) . During this phase , participants made cost–benefit decisions for the cognitive and physical effort tasks separately . On each trial , they were presented with a fixed low-effort/low-reward “baseline” option and a variable high-effort/high-reward “offer . ” The baseline option was an opportunity to perform the lowest level of effort for one credit , while the offer presented a higher number of credits ( 2 , 4 , 6 , 8 , or 10 credits ) for having to invest a greater amount of effort ( levels 2–5 ) . Importantly , by providing participants with the identical range of reward options for both cognitive and physical effort , we could disentangle how cognitive and physical effort differentially devalued the identical rewards . In addition , in order to eliminate the effect of fatigue on participants’ decisions , they were not required to execute their choices within the scanner . Instead , they were instructed that they would be required to perform a random selection of ten of their choices at the conclusion of the experiment and that their remuneration would be based on these randomly selected trials . Because separate decisions were made for the cognitive and physical tasks , we were able to estimate the extent to which the same amount of reward was devalued within each domain for each participant . An important feature of our design was that we temporally separated the presentation of the offer from that of the response cue . Thus , participants did not know which button corresponded to the baseline or offer until the onset of the response prompt . This ensured that we could examine activity time-locked to a cue from which SV would be processed independently , with activity related to these events not confounded by preparatory motor activity . Using the modelling parameters derived above , we computed the SV for the effort and reward combinations on every trial and used the difference in value between the SV of the chosen offer and the value of the baseline as a parametric regressor modelled to the onset of the offer cue [37] . Many studies have shown that regions we hypothesised would be engaged by cost–benefit valuations are sensitive to the difference in the SV of two options rather than to the SV of an offer per se [31 , 37] . Thus , we fitted the SV difference on each trial within the cognitive domain as a parametric modulator time-locked to the onset of each cognitive offer and performed the corresponding analysis for offers in the physical domain . This allowed us to examine activity covarying with SV for the cognitive and physical domains separately . These parametric modulators were defined based on the discounting parameters estimated for each participant’s choice behaviour . We considered significant those voxels which survived whole brain–level , voxel-wise corrections for multiple comparisons ( p < 0 . 05 , corrected for family-wise error [FWE] ) .
We used model-based fMRI to determine whether shared or separate neurocomputational mechanisms underlie cost–benefit valuation in the cognitive and physical domains . Computational modelling revealed that individuals were differentially sensitive to cognitive and physical effort . Neuroimaging data showed that activity in several areas previously implicated in effort processing covaried with the subjective value of rewards independent of effort domain . This included the dACC , dmPFC , dlPFC , IPS , and anterior insula . Importantly , activity within many of these areas also covaried with absolute reward and effort levels , suggesting an integration of these parameters within these areas . However , in contrast to the view that SV is processed in an entirely domain-general manner , an ROI analysis revealed that the right amygdala appeared to process SV uniquely for rewards associated with cognitive and not physical costs . Importantly , none of these results could be explained by choice difficulty or perceived risk . Together , these data indicate that cost–benefit valuation in the human brain is underpinned mostly by a common , domain-independent mechanism but that the amygdala may play an important role in valuing rewards associated with cognitive effort . These results therefore suggest that the classical view of a domain-general set of brain regions for valuation cannot fully account for the subjective valuation of rewards associated with all effort costs [9] . To our knowledge , no study to date has examined the neural correlates of SV associated with cognitive versus physical effort in a single paradigm . The only study that has addressed the nature of cognitive and physical effort examined the processing of raw magnitudes of effort and reward without considering individuals’ subjective valuations and did not require subjects to make choices about whether the effort was worth exerting to obtain the reward [20] . Such an approach is common in the literature and assumes that rewards have a similar effect across individuals to exert the associated effort [11 , 13 , 43 , 44] . However , preferences vary depending on subject-specific cost–benefit valuations , and SVs potentially afford a more sensitive measure of capturing individual differences in motivation [9 , 22 , 23 , 45] . Furthermore , SV has been proposed as an important entity in understanding apathy in healthy individuals as well as those with clinical disorders of motivation [8 , 21 , 46] . Defining the neural and computational mechanisms that underlie the choice to exert effort for reward is therefore crucial to understanding the variability in motivated behaviour across individuals . In the present study , by parametrically varying effort across six levels in both domains , we were able to computationally model SVs for individual participants and therefore more closely examine the key computations that underpin choice behaviour and motivation . Our paradigm had several other advantages . First , the protocol involved manipulating effort in two separate domain-specific tasks , as opposed to requiring participants to exert a combination of both forms of effort to attain specific rewards in each trial [20] . We were therefore able to examine choice behaviour for identical rewards in each domain independently . Second , although many studies have examined the processing of effort and reward , the majority may have been confounded by motor execution for the choices or preparatory activity related to an upcoming effortful exertion . In the design used here , it was possible to investigate activity specifically related to decisions based on SV by temporally separating the choice process from the preparation or execution of the effortful act . Third , by controlling the temporal parameters of both the cognitive and physical effort tasks , it was possible to eliminate delay discounting as an explanation of choice behaviour [5 , 47 , 48] . Fourth , by using computational modelling approaches , we were able to examine activity that varied with SV . Finally , by ensuring that reinforcement rates were similar for the six levels of effort within and across domains , it was possible to ensure that probability discounting could not have contributed to our findings . Thus , the study reported here isolates the effect of SV on choice and motivation independently of many effects that can confound studies examining effort-based decision making . As such , we can effectively rule out the possibility that several regions that we identified were only related to the energisation of behaviour and not to motivation or the valuation of behaviour [49] . Our model comparisons indicated that individuals valued rewards differently when associated with cognitive and physical effort . This was demonstrated by the winning model , which specified separate discounting functions requiring separate discounting parameters for cognitive and physical effort . This conclusion was also supported by the more general pattern of the computational modelling results , which showed that the models assuming equal reward devaluation across cognitive and physical effort ( i . e . , those assuming a single discounting parameter ) provided poorer fits than those that assumed separate discounting parameters . This finding that different functions best fitted cost–benefit valuations for cognitive and physical effort most likely reflects differential sensitivities to effort in the two domains . Our finding that a parabolic function best accounts for participants’ choice behaviour in the physical effort task is in keeping with previous observations [33] . In contrast , effort discounting in the cognitive domain has been much less studied [6] , and it is likely that the specific shape of a discounting function will depend on the specific cognitive faculty being tested ( e . g . , attention versus working memory ) . However , the key point for the present study is that , in the tasks that we used , identical rewards were valued distinctly across both domains . Strikingly , despite rewards being devalued at different rates and in a mathematically distinct manner across the two domains , a largely overlapping network of regions was involved in processing the SV of rewards devalued by both the cognitive and physical effort cost . It is important to note that this finding does not rely on the generalisability of these specific functions to other cognitive or physical effort–based tasks . However , the fact that effort discounting in our task is best described by separate functions does considerably strengthen this result , as it implies that any differences between cognitive and physical effort cannot simply be a matter of scale ( e . g . , some participants finding one task more effortful than the other ) . Rather , it suggests a possible difference in the underlying mechanism between the two processes . Furthermore , the separate discounting functions render the imaging results more compelling by showing that the SVs computed from entirely different functions nevertheless engage overlapping brain regions . Regardless , a question that remains is whether the same pattern of results would be achieved in a cognitive and physical effort task that were best described by the identical discounting function . Exploratory analyses using a single function to model choice across both domains revealed a pattern of domain-general and domain-specific effects that were essentially similar to those of the primary analyses . However , it remains for future studies to verify the conclusions from our study in the case of cognitive and physical effort tasks that are best described by identical discounting functions . Interestingly , most of the domain-general areas that encoded subjective value also showed a significant negative effect of reward and a significant positive effect of effort . The findings that many domain-general areas that encode SV also encode raw reward and effort levels are not incompatible—indeed , one interpretation is that these regions integrate the reward and effort on offer into a value signal . Although many previous studies have examined the neural basis of processing SV [9] , we believe this is one of the first demonstrations that regions of the brain can process a SV formed from costs that devalue rewards at different rates . Furthermore , although some of these domain-general regions may be involved in processing decision difficulty in certain contexts [50] , this is not always the case [51] , and none of the regions identified in the present study were found to encode choice difficulty across both the cognitive and physical domains . The key to elucidating the neural basis of cost–benefit decision making will be understanding how this domain-general network learns or forms a valuation of rewards associated with different forms of effort [15 , 22] . A central role of the dorsal ACC/dmPFC in value-based decision making and motivation is considered by some to be in signalling the value of a behaviour in comparison to alternatives [52 , 53] . The study reported here extends this notion by showing that this region not only processes the SV of an offer but also integrates effort and reward information independent of the nature of the effort cost [50 , 52 , 54] . In addition , single-unit studies have shown that dACC/dmPFC neurons signal the net value of rewards associated with effort , and the necessity of this region in cost–benefit valuation has been demonstrated by lesion studies that report that inactivation of medial prefrontal cortex impairs an animal’s ability to overcome effort costs [15 , 16 , 25 , 48 , 55] . Recently , several human studies have also shown this region to be important in calculating choice value for effortful rewards . Although the majority of these have been in the physical domain [11 , 13 , 14] , a recent investigation reported a similar pattern for cognitive effort [56] . Neurons sensitive to reward information have been identified in the dlPFC [55 , 57–59] , and the activity of lateral prefrontal areas in humans correlates with predicted SVs that guide decision making [60] . Lateral intraparietal neurons have been found to signal expected value [61] , and parietal activity has been reported in tasks requiring value comparisons [62 , 63] . Lastly , insular activity is negatively correlated with the SV of rewards associated with higher effort [14 , 64] , and dopaminergic responses , which play an important role in motivated decision making , exhibit greater variability in the insula with less willingness to expend effort for reward [65] . Our findings extend this body of data by showing that the process of subjective reward valuation occurs independent of the nature of effort costs , and suggest that it is underpinned by activity in a the dACC/dmPFC , dlPFC , IPS , and anterior insula . Do these regions of domain-independent areas comprise a network for subjective valuation ? Tracer studies in macaque monkeys and neuroimaging studies in humans suggest that these domain-independent regions are monosynaptically connected . The upper bank of the dorsal anterior cingulate sulcus is connected to the anterior portions of the insula , several amygdala nuclei , and BA 9/46 in the lateral prefrontal cortex . Similar projections exist between each of these locations and the other domain-independent regions within this putative network [42 , 66–69] . In addition to the connectional anatomy , it has been noted that these same domain-independent regions are activated during a variety of different cognitive and motor control tasks [70 , 71] . It has been argued that this multiple-demand ( MD ) network is involved in flexibly controlling the cognitive processes required across a large number of tasks [70] . In this context , our results could be taken as support for the notion that this network is activated independent of the nature of the cost or associated behavioural domain . However , our findings also suggest a more nuanced interpretation of the functional properties of the MD network . In our study , activity in this network was influenced by the value of working and not by the demand alone . Moreover , as highlighted above , these areas contain single neurons that respond to reward valuations , and the BOLD signal in these regions has been shown to scale with subjective reward valuations in studies investigating temporal discounting or probabilistic reward-based decisions . Thus , a more refined account might be that the MD network is crucial for motivating behaviours across different domains of behaviour . Such a notion would explain why these regions are activated during many cognitive and motor tasks in which motivation must be sustained for successful performance [72] . Importantly , we found evidence of domain specificity for cognitive effort valuation , specifically in the right amygdala . The amygdala is known to play an important role in reward valuation , and single-unit recordings have demonstrated that neurons here encode the value associated with individual items [26 , 73–75] . Recent evidence points to the amygdala as playing a crucial role in effort-based decision making in rodents , with neurophysiological data showing that the amygdala plays an important role in valuing effort [40 , 41] . Recently , some have proposed that the amygdala is sensitive to different types of effort costs [27] and also highlighted the key role for this region in the flexible control of cognitive processes . However , drawing a definitive conclusion , especially in humans , requires comparisons across species and across tasks . Substantial differences exist between the paradigms used in valuation studies and include differences in reinforcement schedules , training intervals , reward magnitudes , and contrast effects . Furthermore , previous effort-based tasks have not tightly controlled the contributions from each domain to their manipulations of effort , thus making it difficult to compare the relative contributions of the two domains . Indeed , such discrepancies may even underlie varying amygdala involvement in cost–benefit decision-making tasks across cognitive and physical effort . In our study , we designed each of our closely matched tasks to hold all features constant except for the type of effort involved , which was maximised in each domain relative to the other . We were therefore able to provide more direct evidence that the human amygdala may be differentially involved in cognitive over physical effort valuation . Nevertheless , while our result is consistent with the preceding studies noting potentially dissociable roles of the basolateral amygdala for cognitive and physical effort–based decisions , the finding of amygdala domain specificity does deserve replication in future studies and would be even more compelling if it was demonstrable at a whole-brain level . Interestingly , previous studies have shown that the VS and vmPFC are engaged when processing value [20 , 39 , 76] . Here , we found no such activity for either cognitive effort , physical effort , or the conjunction . This was the case even after specifically probing these areas with regions of interest defined on the basis of previous studies . A key difference between this study and all previous studies implicating the VS and vmPFC in value processing is that previous tasks required effort to be exerted while participants were being scanned , and most of the effects may have been related to the execution of the effortful task rather than to the choice of whether the effort was worth exerting . This may suggest that the VS and vmPFC process value primarily when value may guide or motivate the execution of a behaviour that will be followed immediately by a rewarding outcome , rather than in the evaluation of whether resources should be allocated to a task at all . Rewards in real life are rarely obtained without effort . Our model-based fMRI approach revealed that effort discounting in the cognitive and physical domains is underpinned by largely shared neural substrates but that the amygdala uniquely contributes to cognitive effort valuation . Importantly , neither delay nor probability discounting can account for our results . It has been postulated that disorders of diminished motivation—such as apathy and abulia—which are manifest in multiple neurological and psychiatric conditions , may be characterised as a diminished willingness to exert effort for reward [46 , 77] . Our findings may therefore help us understand the neural basis for such disorders of motivation by providing an insight into their multidimensional nature and identifying potential neural foci that might be manipulated to modulate motivation [78] .
This study was approved by the Central University Research Ethics Committee of the University of Oxford ( MSD-IDREC-C1-2014-037 ) . We recruited 38 young , healthy , right-handed participants . All participants had no history of neurological or psychiatric illness and were not taking regular medications . Four participants were excluded: 2 for failing to provide responses on a high proportion of trials while being scanned ( over 9% ) , and a further 2 because of excessive head motion within the scanner ( more than 5 mm of translation ) . The final group of 34 participants ( 23 females ) had a mean age of 24 y ( range 19–39 ) . All participants were behaviourally trained on a cognitively effortful task and a physically effortful task prior to being scanned . These extensive training sessions were aimed at familiarising participants with the effort associated with all levels for both tasks . The training phase for each task began with 18 practice trials ( 3 per effort level ) and was followed by a further 60 trials to reinforce behaviour ( 10 per effort level ) . Behavioural analyses of task performance were conducted on the latter 60 trials . After training , we scanned participants while they made economic decisions based on how much effort they would be willing to trade off for varying levels of reward . The order of training in the physical and cognitive effort tasks was counterbalanced across participants . | Rewards are rarely obtained without the motivation to exert effort . In humans , effort can be perceived in both the cognitive and physical domains , yet little is known about how the brain evaluates whether it is worth exerting different types of effort in return for rewards . In this study , we used functional magnetic resonance imaging ( fMRI ) to determine the neural and computational basis of effort processing . We developed two novel tasks that were either cognitively or physically effortful and had participants indicate their preference for a low-effort/low-reward versus a higher-effort/higher-reward version of each . Our results showed distinct patterns of reward devaluation across the different domains of effort . Furthermore , regardless of the type of effort involved , motivation was subserved by a large network of overlapping brain areas across the parieto-prefrontal cortex and insula . However , we also found that the amygdala plays a unique role in motivating cognitively—but not physically—effortful behaviours . These data impact current neuroeconomic theories of value-based decision making by revealing the neurocomputational signatures that underlie the variability in individuals’ motivation to exert different types of effort in return for reward . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"decision",
"making",
"prefrontal",
"cortex",
"brain",
"social",
"sciences",
"neuroscience",
"magnetic",
"resonance",
"imaging",
"cognitive",
"psychology"... | 2017 | Neurocomputational mechanisms underlying subjective valuation of effort costs |
One of the central goals of evolutionary biology is to explain and predict the molecular basis of adaptive evolution . We studied the evolution of genetic networks in Saccharomyces cerevisiae ( budding yeast ) populations propagated for more than 200 generations in different nitrogen-limiting conditions . We find that rapid adaptive evolution in nitrogen-poor environments is dominated by the de novo generation and selection of copy number variants ( CNVs ) , a large fraction of which contain genes encoding specific nitrogen transporters including PUT4 , DUR3 and DAL4 . The large fitness increases associated with these alleles limits the genetic heterogeneity of adapting populations even in environments with multiple nitrogen sources . Complete identification of acquired point mutations , in individual lineages and entire populations , identified heterogeneity at the level of genetic loci but common themes at the level of functional modules , including genes controlling phosphatidylinositol-3-phosphate metabolism and vacuole biogenesis . Adaptive strategies shared with other nutrient-limited environments point to selection of genetic variation in the TORC1 and Ras/PKA signaling pathways as a general mechanism underlying improved growth in nutrient-limited environments . Within a single population we observed the repeated independent selection of a multi-locus genotype , comprised of the functionally related genes GAT1 , MEP2 and LST4 . By studying the fitness of individual alleles , and their combination , as well as the evolutionary history of the evolving population , we find that the order in which these mutations are acquired is constrained by epistasis . The identification of repeatedly selected variation at functionally related loci that interact epistatically suggests that gene network polymorphisms ( GNPs ) may be a frequent outcome of adaptive evolution . Our results provide insight into the mechanistic basis by which cells adapt to nutrient-limited environments and suggest that knowledge of the selective environment and the regulatory mechanisms important for growth and survival in that environment greatly increase the predictability of adaptive evolution .
Increasingly , the fields of evolutionary and molecular biology are fusing in a research program that has been termed the “functional synthesis” [1] . The power of this approach is exemplified by the molecular reconstruction of ancestral proteins enabling the study of the functional properties [2] and evolutionary histories [3] of individual genes . By contrast , the evolution of pathways and networks comprising multiple genes has thus far been less amenable to functional studies . This is due in part to the difficulty of inferring and engineering ancestral states of genetic networks . An alternative approach to the study of genetic network evolution is the study of long-term natural selection in laboratories . Experimental evolution using microbes has a number of useful features including the ability to monitor evolution in real time and to measure fitness in the relevant environmental condition [4] that makes it ideally suited to the study of gene network evolution . Uniquely among experimental methods of long-term selection , continuous culturing using chemostats [5] , [6] enables establishment of a precise and invariant selective pressure in which cell growth is continuously constrained by the rate of provision of a growth limiting nutrient . In contrast to evolution experiments using serial dilution [4] , [7] , [8] , in which cells undergo repeated cycles of feast and famine , the unchanging nutrient-poor environment of a chemostat reduces fitness to a single component –continuous growth in a nutrient-poor environment– facilitating testing and interpretation of the functional basis of beneficial mutations . Moreover , in chemostats , large population sizes can be maintained ( in excess of a billion cells ) during the long-term selection thereby minimizing the effects of genetic drift and population bottlenecks . Despite recent progress in our understanding of the molecular basis of adaptive evolution in chemostats [9]–[14] many questions remain . Does selection target particular loci and preferentially utilize distinct types of alleles ? What is the functional basis of adaptation and are there mechanistic relationships between beneficial mutations ? Does increased environmental complexity result in increased heterogeneity within a population ? To what extent does epistasis constrain adaptive landscapes ? Here , we describe the results of experimental evolution of the budding yeast , Saccharomyces cerevisiae , in different nitrogen-limited chemostat environments . Variation in nitrogen availability is frequently encountered in natural ecologies and use of this selection enables comparison with previous adaptive evolution studies in other nutrient-limited environments using chemostats [9] , [12] , [14] . Importantly , for the goal of understanding genetic network evolution the molecular mechanisms underlying nitrogen utilization in budding yeast have been extensively studied [15] , which facilitates interpretation of the functional effects of adaptive mutations . In nitrogen-limited chemostats , the steady-state nitrogen concentration in the culture is extremely low and cells grow continuously in a nitrogen-poor environment . Under these conditions , expression of a set of coordinately regulated genes , the nitrogen catabolite repression ( NCR ) regulon , is activated by the GATA transcription factors , GLN3 and GAT1 [16] . NCR genes encode a number of transporter and catabolic enzymes for import and assimilation of diverse nitrogen sources , the expression of which is repressed during growth in a nitrogen-rich environment by the negative regulators GZF3 and DAL80 [16] . Despite the greatly simplified and invariant selective conditions of a chemostat we find evidence for at least three distinct adaptive strategies in nitrogen-limited chemostats that operate with different levels of environmental specificity . Consistent with earlier studies in other nutrient limitations [9] , [12] , [17] , comparative analysis among the different nitrogen-limited conditions revealed selection for copy number variant ( CNV ) alleles that result in increased abundance of transporters specific for the molecular form of nitrogen provided in each environment . We show that these alleles are also selected when multiple nitrogen sources are simultaneously present in the environment and that their inordinate fitness effects likely limit the accumulation of genetic diversity , even in environments with increased environmental complexity . Novel alleles at some loci are recurrently selected in different nitrogen-limited environments , including VAC14 and genes with related functions , pointing to a role for remodeling of phosphatidylinositol-3-phosphate production and vacuole biogenesis in adaptation to nitrogen-limitation . By integrating our results with previous studies we find that variation in a subset of loci is selected in both nitrogen-limited chemostats and glucose-limited chemostats providing evidence for a general adaptive strategy in nutrient poor environments through remodeling of the TORC1 and Ras/PKA pathways . We also report a striking example of clonal interference in which independent lineages , defined by mutations in three functionally related loci , GAT1 , MEP2 and LST4 co-evolve in a single population undergoing adaptive evolution in an ammonium-limited chemostat . By studying the individual and interactive effects of these alleles as well as reconstruction of lineage dynamics , we demonstrate that the order of mutations is constrained by epistatic interactions . We propose that this three-locus genotype comprising functionally related gene products represents a gene network polymorphism ( GNP ) , which may be a more frequent outcome of adaptive evolution than previously appreciated .
Initially , we studied populations evolving in seven different nitrogen-limited environments . To identify phenotypically distinct clones within each adapted population of ∼1010 cells following 250 generations of selection we performed batch culture growth rate assays on an unbiased sample of 94 clones from each population and selected three individuals that exhibited growth characteristics distinct from each other and the ancestral strain for further characterization ( Figure S1 and methods ) . We determined the relative fitness of each clone in the appropriate nitrogen-limited chemostat environment and typically observed large increases in fitness ( >10% ) ( Figure 1A ) . This is consistent with mutation and selection rapidly moving strains towards a fitness optimum . It is clear that the ancestral genotype differs in its distance to the fitness optimum with respect to different nitrogen limited environments: fitness increases in clones selected from ammonium- , arginine- and glutamine-limited chemostats are around 25% whereas fitness increases in clones evolved in urea- and allantoin-limited chemostats exceed 80% . In general , individuals from the same population had similar fitness . A minority of clones did not show increased fitness using this assay for reasons that are not clear , but may be indicative of frequency-dependent selection . The majority of evolved clones were unaltered in their ability to grow in nitrogen-rich conditions or showed decreased fitness ( typically less than 4% ) ( Figure S2 ) . Thus , mutations selected in the nitrogen-poor environments are uniquely beneficial in nitrogen-poor environments and exhibit antagonistic pleiotropy in nitrogen-rich environments . To identify mutations associated with increased fitness we first analyzed the genomes of selected clones , and entire populations , using array comparative genomic hybridization ( aCGH ) . We observed multiple copy number variants ( CNVs ) , including duplicated and deleted genomic regions , typically greater than ∼10 kb , in individual clones and entire populations ( Figure S3 ) . Previously , we reported identification of amplification alleles that include the GAP1 locus in clones adapted to glutamine- or glutamate-limitation [10] . A subset of CNVs present in other nitrogen-limited environments include compelling candidates that are likely to underlie selection of the amplified allele . These include a CNV containing the allantoin permease ( DAL4 ) in allantoin-limited conditions , a CNV including the urea permease ( DUR3 ) in urea-limited conditions and a CNV including the proline permease ( PUT4 ) in proline-limited conditions ( Figure 1B ) . Our ability to detect these CNV alleles in population samples using aCGH ( Figure 1B ) indicates that they are at high frequency following 250 generations of selection . Consistent with previous studies [9] , [18] , CNVs are frequently proximal to retrotransposon sequences ( Figure 1B ) , which may increase their spontaneous rate of generation . Previously , we , and others , have identified the repeated selection of copy number variants ( CNVs ) at the HXT6/7 [9] , [17] and SUL1 [9] locus in yeast strains selected from glucose- and sulfur-limited chemostats respectively . In E . coli evolved in lactulose-limiting conditions the lac operon , which includes the lactose permease ( lacY ) , is frequently amplified [19] . Collectively , these findings make clear that in diverse nutrient-limiting conditions , increased production of specific nutrient transporters is a rapid route to increased fitness . The spontaneous rate at which amplification CNVs are generated appears to depend on context [20]; however , estimates of gene amplification rates suggest that they are on the order of nucleotide substitution rates [21] . Selection for spontaneously generated amplification alleles appears to be an expedient means of increasing production of specific nutrient transporters and these alleles are strongly selected in nutrient-poor conditions . It is notable that we did not detect amplification alleles containing the known high affinity ammonium transporter gene , MEP2 , in the ammonium-limited population or the arginine transporter , CAN1 , in the arginine-limited population ( Figure S3 ) . It remains to be determined if amplification of MEP2 or CAN1 is beneficial in ammonium- or arginine-limited conditions or if these amplification alleles are deleterious for functional or genetic reasons . Moreover , we cannot exclude the possibility that amplification alleles were present at an earlier stage in these populations but were subsequently out-competed . We observed additional copy number variants and entire chromosomal aneuploidies that include genes without obvious connections to growth in nitrogen-limited conditions ( Figure S3 ) . We identified 7 aneuploid clones among the 18 analyzed clones ( ∼40% ) . The recurrent observation of aneuploidy in adaptive evolution studies [9] , [18] and as a mechanism of genetic suppression [22] suggests that they are likely to be adaptive , although the mechanistic basis for the selective advantage of aneuploidies remains to be determined . We quantified the DNA content of all clones , using flow cytometry , and found that in populations adapted to allantoin- and urea-limitation a high frequency of cells had a 2N DNA content ( Figure S4 ) . These individuals are still of a haploid mating type ( MATa ) as demonstrated by successful mating with MATα cells . The resulting triploid cells underwent sporulation , but typically yielded poor spore viability ( <10% ) consistent with massive unbalanced chromosome content in the meiotic products of triploids ( Figure S4 ) . The maintenance of a MATa mating type in diploid cells recovered from chemostat selections indicates that they are the result of failed cytokinesis and not due to spontaneous mating type switching and subsequent mating . We did not detect a fitness advantage in the chemostat that is attributable to the diploid state per se ( Figure 1A ) consistent with previous studies [23] . Although the high frequency of diploid cells is consistent with selection , the lack of a detectable fitness effect in a wild type diploid cell suggests that selection for diploidization may require the prior acquisition of at least one mutation that is advantageous when increased in copy number as a result of a whole genome duplication . To study the functional basis of adaptation we performed genome-wide transcriptional profiling of evolved clones in the same chemostat environment as they had been selected . Divergence in the transcriptome between clones adapted to different nitrogen environments was qualitatively similar to that seen between clones adapted to glucose- and phosphorous-limited environments [9] ( Figure S5 ) . Some of the transcriptional variation in clones adapted to nitrogen-limited environments is a direct result of altered copy number due to CNVs as we detected a small but significant positive correlation between DNA copy number and mRNA abundance ( Figure S6A ) . In general , mRNAs corresponding to transporter genes found within CNVs were increased in abundance , consistent with increased DNA copy number resulting in increased transporter abundance ( Figure S6A ) , providing further evidence that these genes drive selection of the CNV . As previously observed [24] , DNA copy number in disomic or trisomic chromosomes of aneuploid cells is proportional to mRNA abundance level ( Figure S6B ) . In some cases this may explain the selection for a specific aneuploidy . For example , a clone recovered from the glutamine-limitation adaptation contains an additional entire copy of chromosome XI , which contains GAP1 [10] . However , other chromosomal aneuploidies do not have an obvious connection to nutrient transport making it unclear how , or why , the large-scale increase in expression of genes along duplicated chromosomes of adapted clones contributes to fitness . To identify all mutations acquired during the selection experiments we performed whole genome sequencing of 18 clones from the seven populations ( see methods ) . We found an average of 4 SNPs per clone that together represent a broad range of classes ( Figure 2A and Table S1 ) . The average number of SNPs is higher than expected ( ∼1 . 0 ) based on the measured spontaneous nucleotide substitution rate [25] but is consistent with the average number of acquired SNPs ( ∼3 . 3 ) reported for equivalent selections in glucose- or phosphorous-limited environments [13] , [26] , [27] . Whether this reflects an increased mutation rate under conditions of stress , as reported for E . coli [28] , or heterogeneity in the number of mitotic events a particular lineage undergoes in a chemostat , remains to be determined . We detected a marginal but statistically significant bias towards SNPs in coding regions: 60/72 SNPs ( 83% ) were found in coding regions , while 72% of yeast genome is coding ( exact binomial test , p = 0 . 035 ) . Although the majority of base changes in coding regions were non-synonymous ( 52/72; 72% ) this is not significantly different than the expected frequency ( 79% ) of non-synonymous mutations [14] ( exact binomial test , p = 0 . 1912 ) . We also identified 8 indels ( 7 deletions and 1 insertion ) of one or two base pairs ( Table S1 ) . The average number of indels per clone ( ∼0 . 44 ) is higher than that expected on the basis of the known spontaneous rate of indel events ( ∼0 . 06 ) [25] . All CNVs detected using aCGH were also identified on the basis of sequence read depth . Furthermore , we detected additional deleted genomic segments of several hundred base pairs suggesting that whole genome sequencing has superior sensitivity to aCGH for CNV detection [26] ( Table S1 ) . In lineages that had undergone diploidization we detected both homozygous and heterozygous point mutations ( Table S1 ) , which allowed us to distinguish mutations that occurred prior to , and after , diploidization , respectively . In sum , comprehensive genome characterization indicates that in individual clones evolving in nitrogen-limited environments , multiple mutations are acquired in a short period of time that range from single nucleotide substitutions to complete duplication of the genome ( Figure 2B ) . Whereas sequencing of clonal isolates provides information on individual lineages , deep sequencing of entire populations provides a means of assessing the genetic diversity in a population at a particular time point in the evolutionary history of the population [29] . We sought to identify all alleles that had risen to appreciable frequencies following 250 generations of selection using whole genome sequencing of entire populations ( Table S2 ) . We identified fixed and non-fixed alleles and estimated their frequencies on the basis of sequence read counts ( Figure S7 ) . Despite sequence read depths in excess of 300-fold , we detected few additional mutations in populations that were not identified in clones . Populations typically contained less than 10 SNPs at frequencies >5% ( Table 1 ) . A single exception was identified; in the population adapted to allantoin-limitation we found 486 mutations , which is likely the result of mutator phenotype due to loss of function in the mismatch repair gene , MSH2 , which we estimate to have a frequency of ∼6% in the population ( Table S3 ) . We were surprised by the low genetic diversity in populations adapted to individual nitrogen sources ( see Table 1 ) especially since previous analyses of E . coli populations evolving in glucose-limited chemostats have suggested the presence of multiple ecotypes [30] , [31] . We hypothesized that the low genetic diversity within populations may be a related to the presence of a single nitrogen source in the environment . To study the effect of increasing the complexity of environments on genetic variation in adapting populations , we performed additional long-term selection experiments using mixtures of 2–4 different nitrogen sources . Following the same period of selection we did not detect increased genetic complexity , as assessed by population deep sequencing , in these selections compared with populations adapted to a single nitrogen source ( Table 1 ) . We performed aCGH on clones and populations evolved in the presence of mixed nitrogen sources and detected CNVs that include transporter genes specific to individual nitrogen sources present in each environment ( Figure S8 ) . However , we did not detect any lineages containing multiple CNVs that would improve transport of more than one of the available nitrogen sources in an environment , suggesting that lineages underwent specialization in the mixed environments . The highest frequency CNVs in populations adapted to mixed nitrogen sources transport non-preferred nitrogen sources ( proline , allantoin and urea ) ( Figure S8 ) , which also tend to be associated with the greatest individual fitness increases ( Figure 1A ) . Collectively , our observations in single and mixed nitrogen-limited environments are consistent with a highly skewed distribution of fitness effects in which CNV alleles that include transporter genes have large fitness effects and therefore a high probability of sweeping to fixation . The large effect sizes of these CNV alleles limits genetic diversity even in environments of increased complexity . High throughput sequencing of clones and populations revealed that genetic variation at a number of loci was repeatedly selected in different nitrogen-limited selections ( Figure 3A ) . In addition to amplification of permease genes in conditions in which they increase import rates of nitrogen-containing compounds , we find that inactivating alleles are selected in conditions in which their function provides no benefit or may be deleterious . As we previously reported , this is the case for GAP1 , which is amplified in glutamine- and glutamate-limited conditions and deleted when the nitrogen source is not an amino acid such as allantoin and urea [10] ( Figure 3A ) . Similarly , amplification alleles containing PUT4 , which encodes a proline permease , are selected in environments in which proline is a nitrogen source , but an inactivating mutation in PUT4 was found in the arginine-limited environment . We hypothesize that loss of function mutations in these genes are selected as the NCR-derepressing conditions of a nitrogen-limited chemostat result in their high expression , which is futile in the absence of the substrate ( s ) they transport . We identified six loci that acquired point mutations in multiple nitrogen-limitation selections . The most striking of these was VAC14 , which is mutant in 8 of the 11 different selective environments . Sequence variants in VAC14 are predominantly loss of function mutations and in two populations we found multiple independent VAC14 alleles ( Figure 3A ) . VAC14 encodes a scaffold component of the protein complex regulating inter-conversion of phosphatidylinositide-3-phosphate ( PI3P ) to phosphatidylinositide-3 , 5-bisphosphate ( PI ( 3 , 5 ) P2 ) [32] . Interestingly , an additional repeatedly mutated locus , FAB1 , encodes the 1-phosphatidylinositol-3-phosphate 5-kinase that functionally interacts with VAC14 . When all mutations identified in clones and populations are considered ( Table S1 and Table S2 ) , there is a clear enrichment for molecular functions related to phosphatidylinositol biosynthetic processes and the related processes of autophagosome and vacuole biogenesis ( Figure 3B ) indicating that they are a convergent target of selection across nitrogen-poor environments . Functional enrichment analysis of mutations in populations and among clones also identified several additional molecular processes related to nitrogen metabolism ( Figure 3B ) . Thus , the molecular basis of adaptive evolution in nitrogen-limited environments exhibits convergence at both the level of individual genes , and at the level of modules , defined by functionally related genes . It is possible that some adaptive alleles recovered in our experiments are not specifically related to nitrogen utilization , but underlie adaptation to the requirement of continuous growth in nutrient-limited conditions . To identify such loci we compared the loci associated with adaptive evolution in nitrogen-limited environments with those identified in previous studies of adaptation to glucose- , phosphate- and sulfur-limited environments [9] , [12]–[14] ( Figure 3A ) . Several loci mutated in both glucose- and nitrogen-limited chemostats encode components of signaling pathways that regulate cell growth in response to the nutritional state of the environment . At least two of these genes ( RIM15 and WHI2 ) regulate entry into a quiescent ( G0 ) state . Loss of the ability to enter G0 may be beneficial in the chemostat , as even transient entry into G0 will prolong the cell division cycle leading to cells being outcompeted . Selection for this class of mutations may be analogous to the recurrent loss of function mutations found in the stress response sigma factor , rpoS , in experimental evolution of E . coli in chemostats [33] . No mutated loci were shared with phosphate and sulfur-limited selections . The population adapted to ammonium-limitation was the only population in which we did not detect evidence of CNVs in either clones or the entire population ( Figure 2B , Table S1 and Table S2 ) . However , clones from this population displayed the greatest divergence in nitrogen catabolite repression ( NCR ) gene expression among all clones analyzed ( Figure 4A and Figure S9 ) and had large fitness increases ( Figure 1A ) suggesting that they had undergone significant adaptive evolution . We found that these two clones , and a third that was not analyzed for gene expression , contain mutations in the DNA binding domain of the zinc finger transcription factor GAT1 ( Figure 4B ) , which encodes a positive regulator of NCR expression [15] . A subset of NCR genes is increased in expression in these clones including those encoding the high affinity ( MEP2 ) and low affinity ( MEP1 and MEP3 ) ammonium permease genes ( Figure 4A ) . Interestingly , several NCR transcripts are also decreased in expression suggesting that the GAT1 mutations may have differential effects on its transcriptional targets . In addition to mutations in GAT1 , we found that the three clones from the ammonium-limitation selection contained one of two different mutations in the identical codon of a predicted transmembrane domain of the high affinity ammonium transporter MEP2 , a transcriptional target of GAT1 [34] ( Figure 4C ) . Furthermore , two of these clones contained mutations in LST4 , which encodes a protein required for efficient sorting of permeases from the Golgi to plasma membrane [35] . The acquired mutations in LST4 are unlikely to render it non-functional based on drug sensitivity assays ( Figure S10 ) . The three genes , GAT1 , MEP2 and LST4 that comprise this recurrently selected multilocus genotype encode functionally related gene products ( Figure 4D ) consistent with adaptive evolution proceeding via the sequential accumulation of variation in genetic networks within lineages . We aimed to determine the temporal dynamics with which the mutations in GAT1 , MEP2 and LST4 occurred and were selected . Population sequencing of the ammonium-limitation adapted population after 250 generations of selection identified 10 SNPs with detectable allele frequencies ( >5% ) ( Table S1 and Table S2 ) . Allele frequencies in the population are informative about the order in which mutations were acquired in each asexually reproducing lineage; however , the timing of mutational events cannot be deduced on the basis of allele frequencies . To reconstruct the evolutionary history of the lineages we determined allele frequencies throughout the evolution experiments using Sanger sequencing [9] ( methods; Figure S11 ) . The resulting trajectories ( Figure 5A ) show that within a single population the same two locus genotype ( gat1 , mep2 ) was independently generated and selected three times ( lineages A1 , B1 , and B3 ) and the three locus genotype ( gat1 , mep2 , lst4 ) was generated at least twice ( lineages A1 and B3 ) . Interestingly , in both lineages , mutations in GAT1 and LST4 occurred in rapid succession and subsequently increased in frequency ( i . e . lineage A0 and lineage B3 in Figure 5A ) , which is suggestive of a synergistic interaction between LST4 and GAT1 . Although we detect dramatic changes in allele frequencies during the selection no individual genotype swept to complete fixation ( i . e . a “hard sweep” ) . Rather , competition ( i . e . clonal interference ) between lineages bearing different alleles in the identical multi-locus genotype resulted in alternating “soft sweeps” . As functionally related genes are enriched for genetic interactions [36] , we hypothesized that epistatic interactions might exist between GAT1 , MEP2 and , LST4 . To test this hypothesis we constructed strains containing the eight possible combinations of the gat1-2 , lst4-2 and mep2-2 alleles identified in clone 3 ( methods ) . The mutations in MEP2 and GAT1 are individually beneficial; however , the mutation in LST4 does not confer a selective advantage on its own ( Figure 5B ) . The double mutation genotypes comprised of either mep2-2 and lst4-4 or gat1-2 and lst4-2 are more fit than expected by summation of their individual fitness effect providing evidence for positive epistasis . However , we found that the combined effect of the gat1-2/lst4-2/mep2-2 alleles does not result in significantly increased fitness compared with the gat1-2/lst4-2 or mep2-2/lst4-2 double mutant genotypes consistent with negative epistasis . To more accurately compare fitness effects of different genotypes we directly competed double mutant genotypes directly with the gat1-2/lst4-2/mep2-2 genotype . Consistent with our initial observations we find that the gat1-2/lst4-2/mep2-2 triple mutant genotype is not significantly fitter than the gat1-2/lst4-2 or lst4-2/mep2-2 double mutant genotypes and is in fact significantly less fit than the gat1-2/mep2-2 genotype . Thus , an LST4 mutation is beneficial only in the background of an individual mutation in GAT1 or MEP2 whereas it is detrimental in the background of the GAT1/MEP2 double mutant ( Figure 5C ) . This sign epistatic interaction is consistent with the order of mutation acquisition in the three lineages in the population: an LST4 mutation is observed after the occurrence of a GAT1 mutation ( lineage A0 ) or a MEP2 mutation ( lineage B3 ) , but not in the lineage that contains a mutation in both GAT1 and MEP2 ( lineage B1 ) .
In a chemostat , the rate of cell growth is constrained by the concentration of a single nutrient that is essential for growth [41] . Thus , there is intense selective pressure for adaptive strategies that improve the import or metabolism of the growth-limiting nutrient . In our study , we initially provided a single source of nitrogen at a growth-limiting concentration . We observed massively increased fitness of in selected lineages following 250 generations of selection when fitness was assessed in the same environment as that in which the selection was performed . In the majority of cases , analysis of individual lineages identified CNVs that include a transporter gene that specifically transports the molecular form of nitrogen provided in the environment . Thus , in addition to the amplification of the GAP1 locus in glutamine- and glutamate-limited conditions [10] , we find DUR3 amplification alleles in urea-limited environments , DAL4 amplification alleles in allantoin-limited environments and PUT4 amplification alleles in proline-limited environments . The fact that these CNVs are detected in DNA samples of entire populations indicates that they are at high frequency in these populations , most likely as a result of selection . Transcriptome analysis indicates that these alleles result in increased gene expression , which likely results in increased protein production . Our new results are consistent with previous studies in budding yeast that have identified amplification of the HXT6/7 locus in populations adapted to glucose-limitation [9] , [12] , [17] and amplification of the SUL1 locus , encoding the high affinity sulfur-permease , in populations adapted to sulfur limitation [9] . The large fitness increases attributable to these specific CNV alleles means that they dominate the evolutionary dynamics of adapting populations thereby limiting the genetic diversity in nutrient-limited environments . CNV alleles have been reported to underlie increased fitness in a diversity of selective environments and organisms , including humans , suggesting that they are a class of genetic variation that are of general importance for adaptive evolution . Increased fitness associated with nutrient transporter amplification is specific to nutrient-poor environments . Using competitive growth rate assays in nitrogen-rich environment we find that evolved clones tend to have decreased fitness . Similar fitness trade-offs in carbon-rich environments have been reported for lineages adapted to glucose-limited chemostats [14] . Amplified transporter alleles may be an underlying source of this antagonistic pleiotropy . Previously , we have shown that inactivating mutations in GAP1 are selected in chemostats containing limiting concentrations of non-amino acid nitrogen sources [10] . In the current study we identified a PUT4 inactivating mutation in a lineage evolved under arginine limitation ( Figure 3A ) . In environments in which the limiting nutrient is present in a predominant molecular form , loss of some transporter genes may be beneficial either through reduction in the energetic cost of their unnecessary production or as a result of a function that is deleterious in the particular environment . Future work will be required to rigorously test the hypothesis that CNV alleles are a molecular basis of antagonistic pleiotropy . In addition to selection of specific transporter amplification alleles in different nitrogen-limited environments , we find evidence for convergent routes to increased fitness across different nitrogen-limited environments . The most striking evidence comes from the multiple inactivating and nonsynonymous mutations that we identified in VAC14 . We found at least one , and as many as three , independent alleles within the 2 . 6 kb coding region of VAC14 in eight of the eleven populations that we studied ( Figure 3A ) . VAC14 encodes a scaffold component of the protein complex regulating inter-conversion of phosphatidylinositide-3-phosphate ( PI3P ) to phosphatidylinositide-3 , 5-bisphosphate ( PI ( 3 , 5 ) P2 ) [32] . In addition , we found mutations in FAB1 , which encodes a PI3P 5-kinase and VAC7 , a regulator of FAB1 , in different nitrogen-limited populations , albeit , much less frequently than VAC14 mutations ( Table S1 ) . Control of PI ( 3 , 5 ) P2 levels by VAC14 , VAC7 and FAB1 is important for several cellular processes including protein trafficking and maintenance of vacuole size and acidity [42] , [43] . Loss of function of VAC14 results in decreased PI ( 3 , 5 ) P2 levels leading to enlarged vacuoles due to defective vacuolar fission [44] . Enlarged vacuoles may be beneficial in nitrogen-limited conditions as vacuoles function as a reserve for nitrogen stores as well as being the compartment for recycling of cytosolic proteins through autophagy [45] . Non-synonymous mutations in the VAC7 and FAB1 may have similar consequences on PI ( 3 , 5 ) P2 levels and vacuole biogenesis as VAC14 loss of function mutations . Although identifying the precise mechanistic basis by which mutations in these functionally related genes contribute to increased fitness in nitrogen-limited environments requires additional study , their selection in different nitrogen-limited environments , and their absence in the mutational spectra identified in other nutrient-limited conditions reported to date , suggests that novel alleles at these loci underlie a generalist strategy specific to nitrogen-limited conditions . By integration of our results with previous studies in other nutrient-limited environments , we find evidence for adaptive strategies involving remodeling of the TORC1 and Ras/PKA signaling pathways that may be general to nutrient limitation . These signaling pathways control cellular growth rate in response to nutrient availability by regulating diverse cellular processes [46] , [47] . In particular , mutations in the regulator of cell cycle exit and entry into G0 , RIM15 are found in different glucose- and nitrogen-limitation selections ( Figure 3A ) . RIM15 is known to have an important role in integrating signals from multiple nutrient responsive signaling pathways including TORC1 and Ras/PKA [48] , [49] . A reduced capacity to enter a G0 state could be beneficial in a variety of nutrient-limitations in chemostats . Consistent with this hypothesis , additional genes that are mutant in both nitrogen- and glucose-limited chemostats include WHI2 , a negative regulator of G1 cyclin expression , IRA1 and GPB2 , both of which are negative regulators of the Ras/PKA pathway , and NGR1 , an RNA-binding protein involved in regulation of cell growth control . Selection for this class of mutations in different nutrient limitations is consistent with the argument that recurrent selection for loss of rpoS in E . coli populations evolved in glucose- , nitrogen- [50] and phosphorous-limited [51] chemostats underlies a tradeoff between the cellular response to nutrient starvation and maintenance of stress resistance . Although transporter amplifications dominate the majority of our adaptive evolution experiments , we did not identify transporter amplification alleles in two of our populations ( ammonium and arginine limitation ) ; the population that underwent adaptive evolution in an ammonium-limited environment was the only population in which we did not identify any CNVs or large-scale chromosomal events . Nutrient transport is still a primary target of selection in this population as we found two independently acquired non-synonymous SNPs that result in amino acid substitutions at the same amino acid residue in MEP2 ( G352A and G352S ) . The mutated site is in a predicted trans-membrane domain ( Figure 4C ) making it likely that these mutations alter the affinity of MEP2 for ammonium either directly or indirectly . Fitness tests of one of a strain containing one of these mutations ( G352A ) show that this variant confers a fitness increase exceeding 10% ( Figure 5B ) . Interestingly , we find evidence that independently generated alleles containing this precise variant may have been selected in natural yeast populations . Although our ancestral strain , which is isogenic to S288c , encodes a glycine at residue 352 in MEP2 , this site is polymorphic among S . cerevisiae strains with 19/26 strains in the SGD database ( http://www . yeastgenome . org ) encoding an alanine at residue 352 . Moreover , the reference genomes of Saccharomyces sensu stricto species , including S . uvarum , S . mikatae , and S . paradoxus , all contain an alanine at residue 352 in MEP2 homologues . It is interesting to note that a recent study reported recurrent selection of MEP2 fusion alleles when a hybrid S . cerevisiae/S . uvarum strain was evolved in ammonium-limited chemostats [52] . S . cerevisiae and S . uvarum differ at 17 residues in the MEP2 protein , one of which is the 352nd amino acid . Consistent with the importance of the 352A allele under conditions of ammonium-limitation , all independently selected S . cerevisiae/S . uvarum MEP2 fusion alleles retained the carboxy terminus-encoding portion of the S . uvarum MEP2 allele , which codes for an alanine at codon 352 . Collectively , these observations suggest that the selection that we imposed in the laboratory bears some resemblance to selection experienced by yeast cells in the natural world with a strikingly convergent response to selection at the molecular level . The population adapted to ammonium-limitation provides evidence that accumulation of variation in functionally related genes underlies adaptive evolution in nutrient-limited environments . Two lineages within the population that contain mutations in MEP2 also contained mutations in GAT1 , which encodes a transcriptional activator of MEP2 ( in addition to other NCR genes ) as well as mutations in LST4 , which encodes a protein that functions in protein sorting to plasma membranes [53] . Analysis of the dynamics with which these mutations were selected demonstrates that their sequential acquisition underlies clonal interference dynamics in this population . Clonal interference due to multiple independent mutations at the same locus has been documented in a variety of experimental evolution studies ( e . g . [54] ) . Our current results show that competing lineages in the same population can accumulate mutations at multiple , common loci as has been observed in E . coli [29] . Interestingly , unlike the recurrently selected three locus genotype identified in [29] comprising variants in spoT , rbs and nadR , which encode functionally unrelated gene products the three loci that define the recurrently selected genotype identified in our study , GAT1 , MEP2 and LST4 , comprise a functionally related gene network ( Figure 4D ) . The order in which mutations at these three loci are acquired appears to be constrained by epistatic interactions . By studying all possible allelic combinations at these three loci we determined that the lst4-2 allele exhibits positive epistasis with the mep2-2 and gat1-2 alleles individually . However , the two locus gat1-2/mep2-2 genotype is more fit than the three locus gat1-2/mep2-2/lst4-2 genotype ( Figure 5C ) . This negative epistatic interaction is consistent with the observation that an LST4 mutation occurs in the background of a GAT1 mutation ( lineage A0 ) or a MEP2 mutation ( lineage B3 ) , but does not occur in the lineage in which both a GAT1 and MEP2 mutation has already occurred ( lineages B1 and B2 ) ( Figure 5A ) . It is also interesting to note that the double mutant genotypes ( gat1-2/lst4-2 and lst4-2/mep2-2 ) and the triple mutant genotype ( gat1-2/lst4-2/mep2-2 ) do not differ significantly in their fitness ( Figure 5C ) , suggesting that they will coexist in an evolving population . Consistent with this expectation , the lineages A0 and A1 , which differ only at LST4 and the lineages B1 and B3 , which differ at LST4 and two additional loci , co-exist that for around 100 generations ( Figure 5A ) . Increasingly , resolution of the multigenic basis of quantitative trait variation to nucleotide variants demonstrates that allelic variants in functionally related genes underlies adaptive evolution [55] , [56] . As the multi locus genotype that we have identified is 1 ) comprised of functionally related gene products that 2 ) interact epistatically with one another , we propose that it comprises a gene network polymorphism ( GNP ) similar to that reported for the galactose-utilization regulon segregating in diverged Saccharomyces kudriavzevii populations [57] . Given a sufficiently large population size , we show that nearly identical GNPs can be recurrently generated and selected within a population resulting in “soft sweeps” in which the GNPs are maintained at intermediate frequencies . The rapid generation of a GNP in a particular niche may lead to balanced unlinked GNPs ( buGNPs ) segregating in the larger population as observed in the Saccharomyces kudriavzevii population [57] . Our study provides new insight into the functional basis of adaptive evolution in nutrient-limited environments . Consistent with the low concentration of a single growth-limiting substrate representing the dominant selective pressure in a chemostat we find evidence for strong selection of alleles that enhance transport of the specific molecular form of the limiting nutrient . In addition , we have identified a mechanism underlying adaptive evolution that appears to be shared among different nitrogen-limited environments , involving phospholipid metabolism and vacuole biogenesis , and a mechanism shared between nitrogen- and carbon-limited environments , entailing nutrient-responsive growth regulating pathways . The identification of a finite number of adaptive strategies in nutrient-limited environments suggests that adaptive evolution of large populations in nutrient-limited environments proceeds along a limited number of paths . Thus , the combination of precise knowledge of the selective environment experienced by a population of organisms and the molecular mechanisms that underlie growth and survival in that environment is likely to greatly enhance the predictability of adaptive evolution .
For all adaptive evolution experiments we founded populations with a haploid derivative ( FY4 ) of the S288c reference strain . For competition assays , we integrated constitutively expressed mCherry or mCitrine-labeled constructs , marked with the kanMX4 cassette , at the HO locus using the high efficiency yeast transformation protocol [58] . All nitrogen-limiting media contained 800 µM nitrogen regardless of the molecular form of the nitrogen and 1 g/L CaCl2-2H2O , 1 g/L of NaCl , 5 g/L of MgSO4-7H2O , 10 g/L KH2PO4 , 2% glucose and trace metals and vitamins as previously described [59] . We founded populations with FY4 in 200 mL of nitrogen-limited media . Chemostat cultures were maintained using Sixfors fermentors ( Infors ) at 30°C , constantly stirred at 400 rpm in aerobic conditions and diluted at a rate of 0 . 12 hr−1 ( population doubling time 5 . 8 hr ) . Each steady-state population of ∼1010 cells was maintained in continuous mode for 250 generations ( ∼2 months ) . A 2 mL population sample was obtained every 20 generations and archived at −80°C in 15% glycerol . Following 250 generations of selection we randomly plated cells onto rich media ( YPD ) , and selected an unbiased sample of 94 clones . We grew all clones from each population in 96 well plates containing the same nitrogen source as that used in the selection experiment and recorded optical densities at 600 nm every 0 . 5 hr over 24 hours using a 96-well Tecan plate reader . Each plate included the ancestral strain ( FY4 ) and a blank well . We estimated the growth rate and the saturation density of all strains using the ‘grofit’ package [60] in R and selected three clones from each population for further analysis . We determined the DNA content of evolved clones by staining with Sytox green and analyzing at least 10 , 000 cells using flow cytometry . FY4 and an isogenic diploid ( FY4/FY5 ) were used for calibration . In addition , each evolved clone was mated with an isogenic strain ( FY5 ) of the opposite mating type ( MATα ) . The resulting strain was sporulated and at least 20 tetrads were dissected using a micromanipulator . Spore viability was determined after three days growth on YPD at 30°C . Each mutant was competed in a chemostat against the ancestral strain ( FY4 ) or a mutant bearing gat1-2 , mep2-2 , and lst4-2 mutations , engineered to constitutively express either mCherry or mCitrine , in the same nitrogen-limited condition used in the selection experiment . We inoculated the unlabeled evolved clone and labeled reference strain in separate chemostat vessels and obtained steady-state cultures of 200 mL . We then mixed the evolved clone with the labeled reference strain to a final ratio of 1∶5 . We obtained 2 mL samples every 2–3 generations over a total of ∼20 generations . Samples were stored at 4°C in phosphate buffered saline ( PBS ) containing 0 . 01% Tween 20 . The relative ratio of the fluorescently labeled reference strain and the unlabeled evolved clone was measured by counting at least 100 , 000 cells from each sample using flow cytometry . We used linear regression of the log transformed ( ln ) ratio of evolved/reference strain abundance against time ( in generations ) to estimate the selection coefficient ( s , the slope of the fit linear line ) and associated standard error ( s . e ) using the ‘lm’ function in R . We calculated the 95% confidence interval of the regression coefficient in R . The relative fitness , normalized to wild type , is 1+s . Competition assays in batch culture were performed using synthetic deficient ( SD ) media containing 5 g/L ammonium sulfate and were performed using analogous methods by first growing evolved and fluorescently-labeled ancestral strains in isolation to log phase and then mixing them at a 1∶1 ratio . Cultures were maintained in log phase growth for 24 hours ( less than 12 generations ) and sampled 5–6 times . The relative abundance of the two strains and fitness coefficients were determined using the same flow cytometry and analytical methods used for chemostat competitions . RNA samples were obtained from evolved clones grown in chemostats limited for the same nitrogen source in which they had been selected . In addition , we obtained RNA samples of the ancestral strain ( FY4 ) grown in each of the nitrogen-limited conditions . Gene expression profiling was performed using Agilent 60-mer DNA microarrays as previously described [9] , [24] . We used a common reference for all expression analysis , obtained from a sample of the ancestral strain grown in an ammonium sulfate-limited chemostat growing at a dilution rate of 0 . 12 hr−1 . We identified gene expression variation specific to evolved clones by normalizing each mRNA abundance measurement with the expression level of that transcript in the ancestral strain grown in the same environment . Array Comparative Genomic Hybridization ( aCGH ) was performed using Agilent 60mer DNA microarrays as previously described [9] , [24] . Genomic DNA ( gDNA ) from evolved clones and entire populations was prepared using the QIAGEN genomic DNA extraction kit , labeled with Cy3 and co-hybridized with Cy5-labeled DNA from the ancestral strain . The resulting log2 transformed ratio was segmented using the ‘DNAcopy’ package [61] in R . We obtained gDNA from each evolved clone and the ancestral strain ( FY4 ) from 10 mL overnight cultures using the QIAGEN genomic DNA extraction kit . For population samples , gDNA was extracted from 10 mL samples taken directly from the adapting population . 1 µg of gDNA sample was then sonicated in a Covaris AFA to obtain fragments of 300–500 bp . To blunt the ends of fragmented gDNA we incubated with PNK ( 10 Unit ) and T4 DNA polymerase ( 12 unit ) at 20°C for 30 min , and then purified using QIAGEN Min-Elute Columns . Adenosine overhangs were added to the blunted DNA using Exo ( - ) Klenow ( 15 Unit ) incubated at 37°C for 20 minutes , followed by purification using QIAGEN Min-Elute Column and elution in 19 µL EB buffer . To multiplex genome sequencing we ligated one of six unique 120 bp adapters ( BIOO ) using Quick ligase at 23°C for 20 minutes . The ligated samples were purified , and adaptor dimers removed , using AMPure XP beads ( Agencourt ) . The purified samples were loaded on a 2% agarose gel with TAE buffer , run at 100 V for 60 min and then stained with SYBR gold . We excised a region of the gel corresponding to 300 to 500 bp and then recovered DNA using a QIAquick Gel Extraction kit . The ligated DNA was PCR amplified using adapter-specific primers and High-Fidelity DNA polymerase in 25 µL reaction volume for 12 cycles to minimize amplification . The concentrations of libraries were determined by qPCR using the Kapa SYBR qPCR Master mix kit and the PhiX library sample as a control . The final samples were diluted in 10 mM Tris-HCl , pH 8 . 0 and 0 . 05% Tween 20 and 2 nM of each DNA library was loaded onto a flow cell . DNA libraries were sequenced using either single end ( 36 bp and 77 bp ) or paired end ( 2×100 bp or 2×50 bp ) protocols on a Illumina HiSeq 2000 . Standard metrics were used to assess data quality . We used the Saccharomyces cerevisiae S288C reference genome , obtained from the SGD database on Feb 03 , 2011 to align reads using BWA 0 . 5 . 9 [62] . We trimmed bases with base quality less than 20 from the 3′ end of each read . We removed reads with mapping quality less than 20 . In addition , PCR duplicates were removed using Picard 1 . 57 ( http://picard . sourceforge . net ) . We generated BAM files from all remaining reads using samtools 0 . 1 . 18 [63] . The average read depth of all sequenced strains is ∼160 X as shown in the Table S4 . To identify SNPs we used samtool 0 . 1 . 18 and bcftools 0 . 1 . 17 with the Bayesian inference option . We determined an empirical quality score cutoff of 160 using bcftools . For paired end sequencing data we excluded all anomalous read pairs . As clonal individuals are haploid we required SNP alleles to have call frequencies close to 1 . 0 . In duplicated genomic regions or diploidized clones , which may contain heterozygous SNPs , we lowered this requirement to a call frequency near 0 . 5 . In addition , we excluded all SNP calls that were also identified in the ancestral strain . To identify small insertions and deletions ( indels ) we used the DINDEL package [64] . We first generated candidate variants from BAM files using DINDEL , and then realigned each of them to the reference sequence in order to minimize false positive calls that are frequent in repetitive regions . Indels detected by DINDEL package are therefore defined as those that are shorter than the sequence read length ( 50 bp or 100 bp depending on sequencing mode ) . We developed a heuristic threshold to identify low frequency SNPs in population sequencing data . First , we used two different BQ cutoffs , of 20 and 30 , to identify SNPs using SNVer [65] . By comparing different population sequencing data to each other and to the ancestor , we identified SNPs in populations as ones that ( 1 ) are not found in the sequencing data from the ancestor and ( 2 ) exist uniquely in sequencing data from one population using both the high ( 30 ) and low ( 20 ) BQ cutoff options . We empirically found that optimal p-value cutoff of SNP calls generated using SNVer was 1×10−8 , and the minimum total number of read counts covering the SNP location should be 50% of the average read counts in each population sequencing data . Using these heuristics we were able to detect SNPs with frequencies of at least 5% in population sequencing data . The allele frequency of each SNP in a population was determined by dividing the number of reads containing the alternative base by the total number of bases mapping to that position . We collected all GO terms from ‘GO . db’ and ‘org . Sc . sgd . db’ packages in R , resulting in 6 , 366 ORFs assigned to 4 , 583 GO terms . We excluded any GO terms for which the number of assigned genes is less than 2 or more than 100 . For a tested set of mutated genes we excluded ones without any GO annotation , incremented the count for each additional mutation identified in loci with multiple independent alleles and included both genes neighboring an intergenic SNPs . We then counted how many mutated loci are assigned to each term . We computed the p-value for each GO term using a one-tailed Fisher exact test . We used a Bonferroni correction to correct for multiple hypothesis testing . We tested clones for sensitivity to 10 mM D-histidine ( D-His ) and 500 µM azetidine-2-carboxylate ( ADCB ) , which are imported by nitrogen catabolite repression ( NCR ) regulated transporters [66] . We aimed to test drug sensitivities in both NCR-repressing and NCR-activating conditions . Therefore we used plates that containued either ammonium , which represses NCR-regulated genes or proline , which results in derepression of NCR-regulated genes [67] . Each mutant was first grown in liquid cultures containing YPD or SD plus ammonium sulfate ( SD-AS ) . We then spotted normalized cell concentrations at ten-fold dilutions on solid agar containing SD-AS or SD plus 5 g/L proline ( SD-Pro ) with or without the drug . Sensitivity to drugs was determined following 2 days growth at 30°C . We prepared gDNA from population samples taken at 7 intermediate time point in addition to the final generation ( i . e . 24 , 61 , 102 , 137 , 173 , 213 , and 250 generations ) using a rapid gDNA extraction protocol [68] . We amplified 200–500 bp length amplicons that contain the SNP at a central position . All amplicons were sequenced using Sanger sequencing and the resulting electropherogram analyzed using PeakPicker to estimate allele frequencies as described [9] , [69] . Vectors of allele frequencies were clustered and averaged if the Pearson correlation coefficient of two mutations was greater than 0 . 97 and the difference in allele frequencies in the final generation ( based on deep sequencing ) was less than 4% . As allele frequency estimates from Sanger sequencing are less accurate than those obtained from deep sequencing data we excluded a small number of allele frequency estimates derived from Sanger sequencing that were inconsistent with our deep sequencing results . All steps in this procedure are summarized in Figure S11 . We backcrossed clone 3 , recovered from the ammonium-limited condition to the ancestral strain of opposite mating type ( FY5; MATα ) , sporulated the hybrid diploid and dissected tetrads . All segregants were tested for mating type using halo assays [70] . We obtained more than one hundred backcrossed strains bearing different combinations of the 5 mutations acquired by clone 3 . Genomic DNA for each strain was prepared using a rapid DNA extraction protocol [68] . Genotyping was performed using allele specific PCR ( the list of allele specific primers is presented in Table S5 ) . Eight strains identified by this process contained all possible combinations of the three mutations of interest – gat1-2 , mep2-2 and lst4-2 – and the ancestral alleles of the two additional loci ( RIM15 and FAB1 ) that were not studied . Each strain was individually competed against the mCitrine-labeled reference strains as described . All DNA sequencing data are available from the NCBI Sequence Read Archive with accession number SRP032757 . DNA microarray data are available through the NCBI Gene expression Omnibus with accession number GSE52787 . | We studied adaptive evolution in different nitrogen-limited environments using long-term selection of asexually reproducing Saccharomyces cerevisiae populations in chemostats . Using next generation sequencing and DNA microarrays , we identified all acquired genetic variation associated with increased fitness , in both individual lineages and entire populations . We find that amplification alleles that include nutrient transporter genes specific to the molecular form of the nitrogen present in the environment are a common mechanism underlying increased fitness . In addition , we identified a general strategy for adaptation to nitrogen-limited environments that entails remodeling of phospholipid biogenesis required for producing important cellular components including vacuoles and autophagosomes . More general strategies for adaptation to nutrient-limited environments point to a role for re-wiring of signaling pathways that coordinate cell growth with nutrient availability . We reconstructed the evolutionary dynamics of a population evolving in ammonium-limited conditions and find that a multi-locus genotype is repeatedly selected within the population and constrained by epistasis . We propose that this genotype constitutes a “gene network polymorphism ( GNP ) , ” which may be a common outcome of adaptive evolution . Our study suggests that when the selective pressure is understood the molecular basis of adaptive evolution in large microbial populations may be predicted with reasonable precision . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbial",
"mutation",
"model",
"organisms",
"microbial",
"evolution",
"genetics",
"yeast",
"and",
"fungal",
"models",
"biology",
"microbiology",
"evolutionary",
"biology",
"genomic",
"evolution",
"gene",
"networks",
"evolutionary",
"genetics",
"saccharomyces",
"cerevis... | 2014 | Molecular Specificity, Convergence and Constraint Shape Adaptive Evolution in Nutrient-Poor Environments |
Koolen-de Vries syndrome ( KdVS ) is a multi-system disorder characterized by intellectual disability , friendly behavior , and congenital malformations . The syndrome is caused either by microdeletions in the 17q21 . 31 chromosomal region or by variants in the KANSL1 gene . The reciprocal 17q21 . 31 microduplication syndrome is associated with psychomotor delay , and reduced social interaction . To investigate the pathophysiology of 17q21 . 31 microdeletion and microduplication syndromes , we generated three mouse models: 1 ) the deletion ( Del/+ ) ; or 2 ) the reciprocal duplication ( Dup/+ ) of the 17q21 . 31 syntenic region; and 3 ) a heterozygous Kansl1 ( Kans1+/- ) model . We found altered weight , general activity , social behaviors , object recognition , and fear conditioning memory associated with craniofacial and brain structural changes observed in both Del/+ and Dup/+ animals . By investigating hippocampus function , we showed synaptic transmission defects in Del/+ and Dup/+ mice . Mutant mice with a heterozygous loss-of-function mutation in Kansl1 displayed similar behavioral and anatomical phenotypes compared to Del/+ mice with the exception of sociability phenotypes . Genes controlling chromatin organization , synaptic transmission and neurogenesis were upregulated in the hippocampus of Del/+ and Kansl1+/- animals . Our results demonstrate the implication of KANSL1 in the manifestation of KdVS phenotypes and extend substantially our knowledge about biological processes affected by these mutations . Clear differences in social behavior and gene expression profiles between Del/+ and Kansl1+/- mice suggested potential roles of other genes affected by the 17q21 . 31 deletion . Together , these novel mouse models provide new genetic tools valuable for the development of therapeutic approaches .
The Koolen-de Vries syndrome ( KdVS ) has a prevalence estimated at 1/55 , 000 based upon CNV studies[1–3] and was primary described as a consequence of the 17q21 . 31 microdeletion . Patients with KdVS present characteristic facial dysmorphisms[4] and clinical features including intellectual disability , friendly behavior , hypotonia , and several brain anomalies[5–7] . Microdeletions and microduplications of genomic fragments in the 17q21 . 3 region ranging from 400 to 800kb have been found in individuals with intellectual disability[6 , 8]; these genomic fragments include five protein-coding genes: CRHR1 , SPPL2C , MAPT , STH , and KANSL1 . The reciprocal duplication is much more rare than the deletion . To our knowledge , only eight patients have been described in the literature[8–12] . The symptoms are heterogeneous and include craniofacial malformations , microcephaly , psychomotor delay , poor verbal and motor skills , and reduced social interaction[8] . Two cases out of eight have been diagnosed with autism spectrum disorder ( ASD ) . Loss-of-function mutations and atypical deletions restricted to the KANSL1 gene , encoding the KAT8 Regulatory NSL Complex Subunit 1 , have been found in several KdVS patients . Interestingly , phenotypic comparison of both the 17q21 . 31 microdeletion and KANSL1 heterozygous mutation patients show similar clinical severity , implicating that haploinsufficiency of KANSL1 is sufficient to cause the full manifestation of KdVS phenotype[3 , 13–15] . KANSL1 is a member of the evolutionarily conserved nonspecific lethal ( NSL ) complex that controls various cellular functions , including transcription regulation and stem cell identity maintenance and reprogramming[16 , 17] . The NSL complex contains the histone acetyltransferase MOF ( males absent on the first ) encoded by KAT8 which acetylates histone H4 on lysine 16 ( H4K16 ) and with lower efficiency on lysines 5 and 8 ( H4K5 and H4K8 , respectively ) to facilitate transcriptional activation[18 , 19] . Recent studies in flies have shown that KANSL1 acts as a scaffold protein interacting with four NSL subunits including WDR5 which plays a critical role in assembling distinct histone-modifying complexes with different epigenetic regulatory functions[20] . Genes within the human 17q21 . 31 region are highly conserved on mouse chromosome 11E1 . Crhr1 , Sppl2c , Mapt and Kansl1 orthologs have all been found in the same orientation as in the human H1 haplotype . To investigate the pathophysiology of KdVS and microduplication syndrome , we generated first a mutant mice bearing deletion ( Del/+ ) , and duplication ( Dup/+ ) of the 17q21 . 31-homologous Arf2-Kansl1 genetic interval and looked for phenotypes related to the human condition . We studied behavior , cognition , craniofacial and brain morphology of single Deletion carried compared to wild type and pseudo-disomic ( Del/Dup ) controls . Then we compared these data to results obtained with mutant mice for Kansl1 and extended our analysis to gene expression . We found a large phenotypic overlap with altered molecular mechanisms controlling the hippocampus synaptic response .
Del/+ and Dup/+ mice were generated on the C57BL/6N ( B6N ) genetic background ( see supplementary information; Fig 1 ) . In comparison with wild-type ( wt ) littermates , he Del allele frequency was reduced significantly while the transmission of the Dup allele and the Del/Dup carriers were not affected ( Table 1 ) , thus demonstrating that lethality is associated with the deletion on the B6N background . We generated and characterized a compound Del-Dup cohort with littermates carrying four genotypes: Del/+ , wt , Del/Dup , and Dup/+ . First , we followed general parameters . Compared to wt mice , Dup/+ mice were underweight , whereas Del/+ and Del/Dup mice were not ( Fig 1d ) . At 20 weeks of age . Del/+ , Dup/+ , and Del/Dup animals showed significantly reduced body length compared to wt littermates ( Fig 1e ) . In comparison with wt , Del/+ littermates showed lower adiposity levels ( Fig 1f ) at the same age . Nevertheless , we did not detect any notable differences in feeding behavior between mutant and wt animals during a circadian activity test ( S1 Fig ) . Patients with 17q21 . 31 CNVs have impaired intellectual and adaptive functioning[4 , 5] . As a primary experiment , we looked at the activity and the Del/+ and Dup/+ mice displayed a normal circadian pattern ( S1 Fig ) . However , in comparison with wt , Del/+ mice showed reduced spontaneous locomotor activity ( Fig 2a ) as well as reduced rearing behavior during the light phase ( Fig 2a ) . In the open field , no differences in exploration , locomotor activity , rearing behavior , or time spent in the center of the area were observed between mutant mice and wt littermates ( Fig 2b; S1 Table; S2a–S2d Fig ) . In the elevated plus maze , Del/+ and Dup/+ mice explored the same number of arms and spent similar periods of time in open arms as those observed for wt ( S1 Table ) . We evaluated motor coordination and learning using the rotarod test but no differences were observed between mutant Del/+ and Dup/+ mice , and wt mice ( S2a–S2c Fig ) . Similarly , for the grip test , no differences in the muscular strength were observed between mutant mice and wt littermates ( S2d Fig ) . Spatial learning and memory was assessed in the Morris water maze , but similar acquisition and retention were observed in wt and mutant animals ( S3 Fig ) . In the Y maze test , we found similar working memory parameters in wt and mutant animals ( S2 Table ) . Next , we evaluated our models in the novel object recognition test . During the acquisition session , mice of all genotypes spent an equal amount of time exploring the sample object ( S2 Table ) . After 3h of retention delay , Del/+ mice are able to differentiate the novel versus the familiar object but show object discrimination deficits compared to wt whereas Dup/+ and Del/Dup mice showed similar memory capacities ( Fig 2c ) compared to control . Associative memory was evaluated with the fear conditioning test . No significant differences in the baseline and post-shock freezing levels were observed between mice of all genotypes ( S2 Table ) . In the context session , Dup/+ mice displayed a higher level of freezing with significant differences in the last 2 min of the test ( F ( 3 , 56 ) = 6 . 399 , P < 0 . 001; Dup/+ vs wt: P = 0 . 027; Fig 2d ) . The cued session was performed 5 h after the context session . During the presentation of the conditioning cue , all genotypes demonstrated higher freezing incidence ( Fig 2d ) . During the second 2-min long cue , Del/+ and Dup/+ animals showed lower and higher levels of freezing , respectively , in comparison with wt littermates ( H ( 3 , 56 ) = 20 . 609 , P < 0 . 001; Del/+ vs wt: P = 0 . 002 , Dup/+ vs wt: P = 0 . 036 ) . To challenge cued fear extinction in those animals , we used another fear conditioning/extinction protocol with reinforced conditioned stimuli ( CS ) and long-term follow-up; we separated cohorts for the Del/+ and for Dup/+ with their own littermate controls . This revealed opposite effects on the capacity of Del/+ and Dup/+ animals to extinguish the fear response ( Fig 2e ) . Dup/+ animals presented with higher levels of freezing as compared to their wt littermates suggesting that the fear trace persists in those animals . In contrast , the freezing levels of Del/+ animals were globally lower compared to those of their wt littermates , indicating that the fear memory trace is less stable in those animals . To determine whether sociability traits were evident in our mouse models , we first used the three-chamber sociability test . Familiar and unfamiliar animals were of similar sex and genetic background than experimental animals but were of younger age in order to avoid aggressiveness . In the first phase of the test , the social interest session , Del/+ mice spent relatively more time than wt littermates exploring the unfamiliar mouse ( F ( 3 , 51 ) = 3 . 447 , P = 0 . 023; Del/+ vs wt: P = 0 . 017; Fig 2f , Session 1; S4c Fig ) . In the second phase , the social discrimination session , Del/+ mice interacted with familiar and novel congeners for longer periods compared to wt littermates as observed in the first session ( F ( 3 , 51 ) = 6 . 034 , P = 0 . 001; Del/+ vs wt: P = 0 . 002; Fig 2f , Session 2; S4d Fig ) . Nevertheless no differences in the percentage of time spent to explore the novel congener versus the familiar congener were observed between mutant mice and wt littermates ( Fig 2e ) . We studied the influence of 17q21 . 31-homologous CNVs on the mouse craniofacial structure . We analysed computed tomography ( CT ) cranial scans of animal heads combined with 3D reconstruction of skull images using 39 cranial landmarks ( S5a Fig ) . Separate cohorts of Del/+ and Dup/+ females were used for the Euclidean distance matrix analysis[21] and MorphoJ analysis[22] . Skull size in Del/+ ( T = 0 . 115; S5b Fig ) and Dup/+ animals ( T = 0 . 115; S5d Fig ) was similar to that of wt littermates . The skull shape measurements were nominally altered in Del/+ ( Z = 0 . 077; S5c Fig ) and Dup/+ animals ( Z = 0 . 052; S5e Fig ) compared to those in wt littermates . Principal component analysis helped to identify change in the skull shape in Del/+ and the Dup/+ versus control littermates with the three main components ( PC1 and PC2; S5f Fig ) accounting for 59 . 2% of the total variance . Differences are more pronounced in the skull shape of the Del/+ mice than in wt controls with a predominantly shorter nasal bone and a broadening of the face at the level of the zygomatic spine and squamosal junction . In addition to neuropsychiatric features , over 50% of patients with 17q21 . 31 microdeletion also present with various brain structure changes[4 , 5 , 15] . Furthermore , 50% of patients with the 17q21 . 31 microdeletion present with microcephaly[8] . To identify potential morphological alterations of brain regions , we analyzed the brain structure of 8 Del/+ , 10 wt , 11 Del/Dup , and 8 Dup/+ mice using magnetic resonance imaging ( MRI ) . Overall , we found significant differences in total brain volume between the genotypes ( F ( 3 , 33 ) = 14 . 14 , p < 0 . 001; Del/+: 458±23 mm3 , wt: 448±10 mm3 , Del/Dup: 446±12 mm3 , Dup/+: 412±13 mm3 , brain volumes given as mean±sd ) . Dup/+ animals showed a globally reduced brain volume in comparison with that of the other genotypes . Using a segmented atlas that divides the brain into 159 separate brain regions[23–25] , we examined the 83 structures of at least 1 mm3 in size . A reduction of the whole brain volume was noticed for Dup/+ animals ( Fig 3a ) . Brain structures significantly affected after a correction for multiple testing included the hippocampus , amygdala , nucleus accumbens , cingulate complex , entorhinal cortex , frontal region , and perirhinal cortex . Notably , for the majority of these structures , we observed opposite absolute volume changes in Del/+ and Dup/+ animals in comparison with values determined in wt littermates ( Fig 3b ) . Relative volumes of the discussed regions are represented in the supplementary information ( S3 Table ) . To explore if genes from the Arf2-Kansl1 region regulate electrophysiological parameters in mouse neurons as suggested by changes in ChIP-seq profiles , we assessed basal synaptic transmission and synaptic plasticity by measuring field excitatory postsynaptic potentials ( fEPSPs ) in acute hippocampal slices from Del/+ and Dup/+ mice . In Del/+ mutants , we observed decreased fEPSP slopes in mutant slices , especially in response to higher stimulus strengths ( S6a Fig ) . Mean slopes of fEPSPs invoked by the maximum stimulus strength ( 4 . 2 V ) were significantly smaller in slices from Del/+ mice ( 1 . 46 ± 0 . 09 mV/ms ) than wt littermates ( 1 . 87 ± 0 . 09 mV; F ( 1 , 13 . 34 ) = 8 . 31; P = 0 . 025; two-way nested ANOVA , genotype effect ) . The mean paired-pulse ratio of slopes of fEPSPs evoked at a 50 ms interpulse interval was also significantly lower in mutant slices ( S6c Fig; F ( 1 , 11 . 04 ) = 6 . 506; P = 0 . 027 . No significant changes in LTP elicited by theta-burst stimulation were noted in slices from Del/+ mice ( S6 Fig ) . Basal synaptic strength was slightly enhanced in slices from Dup/+ mice , as fEPSPmax mean slope was nominally higher in slices from Dup/+ mice ( 2 . 12 ± 0 . 09 mV/ms ) than in slices from wt littermates ( 1 . 89 ± 0 . 09 mV/ms; S6b Fig ) . However , the effect did not reach statistical significance ( F ( 1 , 8 . 67 ) = 3 . 09; P = 0 . 114; two-way nested ANOVA , genotype effect ) . Likewise , paired-pulse facilitation ( S6d Fig ) and LTP were not significantly different in slices from Dupl/+ and litter-matched wt mice ( S6f Fig ) . We performed similar examinations of mice with heterozygous ablation of Kansl1 ( Kansl1+/- mice ) in the same B6N genetic background and compared the outcome with the Del/+ phenotypes . ) . In comparison with wt mice , Kansl1+/- adult animals were underweight ( two-way ANOVA genotype effect F ( 1 , 30 ) = 11 . 729 , P = 0 . 004; Fig 4a ) Kansl1+/- mice had a significantly smaller body size ( F ( 1 , 15 ) = 11 . 516 , P = 0 . 004 ) and lower adiposity level ( F ( 1 , 15 ) = 6 . 813 , P = 0 . 020; Fig 4a ) than wt littermates in 20 week old animals . In aggregate , these data indicate many similarities in basic traits between the Kansl1+/- and the Del/+ carriers . Next , we examined the behavior of Kansl1+/- mice . In a circadian activity test , Kansl1+/- mice displayed normal patterns of activity ( S7 Fig ) . However their baseline locomotor activity levels differed from those of wt littermates during the dark phase ( F ( 1 , 16 ) = 8 . 482 , P = 0 . 010 ) and the light phase ( F ( 1 , 16 ) = 8 . 573 , P = 0 . 010; Fig 4b ) . In the novel open field arena , Kansl1+/- mice demonstrated an increased level of rearing behavior ( F ( 1 , 15 ) = 4 . 846 , P = 0 . 044: Fig 4c ) . To investigate further this hyperactivity , we performed visual observations of animals in odorless home-cages ( Fig 4d ) . Mutant mice displayed more intensive rearing behavior ( F ( 1 , 15 ) = 7 . 207 , p = 0 . 017 ) and also showed a decreased level of digging behavior ( F ( 1 , 15 ) = 12 . 268 , p = 0 . 003 ) in comparison with wt littermates . These results indicate a global alteration of Kansl1+/- activity characterized , in particular , by locomotor hypoactivity and vertical hyperactivity . During the learning phase of the rotarod test , Kansl1+/- mice displayed higher levels of motor coordination and learning than wt mice ( two-way ANOVA genotype effect F ( 1 , 30 ) = 115 . 867 , P < 0 . 001; Fig 4e ) . In the test phase , Kansl1+/- mice showed improvements for speed higher than 10rpm ( Fig 4f ) , a phenotype not observed in the deletion due to lower power of the tests . Recognition memory was assessed in mice by using the novel object recognition task with a retention delay of 3 h . While no difference was observed in the acquisition session ( S4 Table ) , Kansl1+/- mice displayed a significant memory impairment compared to wt during the choice session ( F ( 1 , 15 ) = 22 . 566 , P < 0 . 001; Fig 5a ) . Then , we evaluated associative memory with the fear conditioning test . No differences in the baseline and post-shock freezing levels were detected between Kansl1+/- mice and wt littermates in the conditioning session ( Fig 5b; S4 Table ) . In the context session , Kansl1+/- mice displayed a lower incidence of freezing than wt littermates ( H ( 1 , 16 ) = 6 . 419 , P = 0 . 011 ) . In the cue session , a decreased freezing level was detected in Kansl1+/- mice during the second 2-min cue period ( F ( 1 , 16 ) = 16 . 748 , P < 0 . 001 ) . Finally , we evaluated animal social behaviors with the three-chamber sociability test and the social interaction test ( Fig 5c and 5d; S8 Fig ) . In both tests , no differences were observed between Kansl1+/- mice and wt littermates . To identify potential gene expression differences in KdVS models , we carried out epigenetic profiling in the hippocampus , a brain region implicated in learning and memory processes , isolated from 3 Del/+ , 3 Kansl1+/- and 6 wt ( 3+3 matched littermates ) . We performed ChIP-Seq of H3K4me3 which is a histone mark located in actively expressed genes . As expected , H3K4me3 marks were lower in the Arf2-Kansl1 region with half peak height values for Arf2 , Crhr1 , Mapt and Kansl1 in the Del/+ samples compared to those observed in wt samples and for the neighboring genes ( Fig 6a ) . Analysis of H3K4me3 tracks ( DESeq2 , p<0 . 01 ) revealed 788 and 751 misregulated promoters in Del/+ and Kansl1+/- , respectively ( Fig 6b ) . Clustering of Pearson correlations , an unbiased method to measure the degree of similarity between large data sets , showed clear segregation between conditions and high concordance of biological replicates ( Fig 6c; data are available in S5 to S10 Tables ) . Cell-type marker analysis ( Fig 6d; see Methods ) [26] revealed that up-regulated promoters observed in Del/+ mostly corresponded to genes expressed in neuronal populations ( pyramidal neurons and interneurons ) . Furthermore , Gene Ontology ( GO ) analysis revealed that they present significant enrichments for the synapse ( p<7 . 9e-20 ) , and dendrite compartments ( p<3e-14 ) as well as synaptic transmission ( p<9 . 1e-24 ) or neurogenesis processes ( p<1 . 3e-22; S12 and S20 Tables ) . In contrast , the term oxidoreductase and mitochondrion were found enriched respectively in Del/+ and Kansl1/+ down-regulated promoters using DAVID [27] but no other GO significant enrichments were found for Del/+ ( S11 and S19 Tables ) . Of the 470 promoters up-regulated in Del/+ , 36% ( 172 ) were also up-regulated in Kansl1+/- , whereas the two genotypes shared 67% ( 211 ) of genes that were down-regulated ( Fig 6e ) . Among all the promoters up-regulated in either of the genetic conditions , we observed 4 distinct clusters of genes ( Fig 6f ) . For cluster 1 , 160 genes enriched in non-neuronal populations ( astrocytes , ependymocytes , choroid plexus , mural cells , Fig 6g ) are up-regulated in the hippocampus of Kansl1 heterozygotes but not of Del/+ mice . Cluster 2 encompassed 214 genes up-regulated in both genetic conditions and enriched in markers of CA1/CA2 pyramidal neurons and astrocytes , while Cluster 3 contained 212 genes enriched in neuronal markers and whose expression was up-regulated to a greater extent in Del/+ . Finally , we noted that cluster 4 comprised 162 genes upregulated specifically in Del/+ and expressed in CA1/CA2 pyramidal neurons and astrocytes . Each cluster showed a specific GO enrichment profile ( Fig 6h ) . Several neuronal processes , including synaptic transmission and neurogenesis were overrepresented GO terms in genes from clusters 2 , 3 and 4 . Cluster 2 genes ( up in both models ) were also enriched in DNA-packaging and nucleosomes ( Fig 6h ) . In the Del/+ hippocampi , 8 genes involved in social behavior were up-regulated and only 2 of these , Tbx1 and Nr2e1 , were also dysregulated in Kansl1+/- mice ( Fig 6i ) . These results suggest a dominance of Del/+ with respect to Kansl1 for determining social behavior . To explore if genes from the Arf2-Kansl1 region regulate electrophysiological parameters in mouse neurons as suggested by changes in ChIP-seq profiles , we assessed basal synaptic transmission and synaptic plasticity by measuring field excitatory postsynaptic potentials ( fEPSPs ) in acute hippocampal slices from Del/+ and Dup/+ mice . In Del/+ mutants , we observed decreased fEPSP slopes in mutant slices , especially in response to higher stimulus strengths ( S6 Fig ) . Mean slopes of fEPSPs invoked by the maximum stimulus strength ( 4 . 2 V ) were significantly smaller in slices from Del/+ mice ( 1 . 46 ± 0 . 09 mV/ms ) than wt littermates ( 1 . 87 ± 0 . 09 mV; F ( 1 , 13 . 34 ) = 8 . 31; P = 0 . 025; two-way nested ANOVA , genotype effect ) . The mean paired-pulse ratio of slopes of fEPSPs evoked at a 50 ms interpulse interval was also significantly lower in mutant slices ( S6 Fig; F ( 1 , 11 . 04 ) = 6 . 506; P = 0 . 027 . No significant changes in LTP elicited by theta-burst stimulation were noted in slices from Del/+ mice ( S6c–S6e Fig ) . Basal synaptic strength was slightly enhanced in slices from Dup/+ mice , as fEPSPmax mean slope was nominally higher in slices from Dup/+ mice ( 2 . 12 ± 0 . 09 mV/ms ) than in slices from wt littermates ( 1 . 89 ± 0 . 09 mV/ms ) . However , the effect did not reach statistical significance ( F ( 1 , 8 . 67 ) = 3 . 09; P = 0 . 114; two-way nested ANOVA , genotype effect ) . Likewise , paired-pulse facilitation ( S6 Fig ) and LTP were not significantly different in slices from Dupl/+ and litter-matched wt mice ( S6 Fig ) .
In this study , we described the first mouse models of Koolen-de Vries syndrome ( KdVS ) and 17q21 . 31 microduplication syndrome . Del/+ mice showed similar phenotypes observed in KdVS patients: higher level of social interaction , lower level of recognition memory , associative learning and memory and brain malformations [3 , 5 , 15 , 28] . We found a single phenotypic similarity between patients carrying the 17q21 . 31 microduplication and Dup/+ mice , which is microcephaly that has been reported in 50% of the human individuals with this microduplication [8] . Several SNPs associated with risk for Alzheimer’s disease ( AD ) were identified near MAPT and KANSL1 in humans and they appeared to be correlated with an overexpression of both genes in different brain regions [29] . This observation was further supported by the description of a familial form of late onset AD due to the microduplication of the 17q21 . 31 region [30] . Thus it would be important now to follow cognition in ageing cohorts carrying the Dup/+ allele generated here as young individuals analysed in the present study do not shown any cognitive impairment , but rather more some improvement in associative memory . Overall , the phenotypic comparison observed in the Del/+ and Kansl1+/- models confirms the critical importance of KANSL1 in KdVS[3 , 15] . Nevertheless , increased social interaction was not found to be affected in Kansl1 haploinsufficient mice whereas it is predominant in humans with KANSL1 mutations [3 , 15] . This discrepancy could reflect either different dosage threshold levels in the mouse and human that governs proper social interaction , with the mice needing more change to induce such friendly phenotype . Indeed we have found more genes associated with social behavior misregulated in the Del/+ brain compared to the Kansl1 heterozygotes . An educated guess would be that the haploinsufficiency of another gene ( s ) from the Arf2-Kansl1 region would contribute to this phenotype in mouse . The hippocampal epigenetic analysis of 17q21 . 31 models unraveled several features . Only the mitochondrion term was enriched in the Kansl1/+ down-regulated genes , a situation partially similar to a recent study where KANSL1 and its partner MOF were observed in mitochondria regulating expression of genes involved in oxidative phosphorylation [31] . Interestingly the oxidoreductase term was enriched in Del/+ down-regulated genes . Thus it would be interested to follow mitochondrial activity in the mouse models . Another common set of dysregulated promoters , were largely affecting CA1 neuronal populations and neuronal functioning and includes many genes with long introns . Common up-regulated genes also appear to be implicated in DNA-packaging and nucleosomes , possibly reflecting the outcome of the misregulated KANSL1 activity . Interestingly , several genes ( Adcyap1 , Cntnap2 , Grid1 , Nrxn1 , Nrxn3 , Ucn , Tbx1 , and Nr2e1 ) , that are associated with disorders [32 , 33] , stress response [34] , social behaviors [35–37] or autism spectrum disorders [38–42] , are up-regulated specifically in the hippocampus of Del/+ mice . Deregulation of these genes may be a molecular underpinning of the friendly/amiable affect of 17q21 . 31 deletion patients . Expression of the majority of those genes ( except for Txb1 and Nr2e1 ) was not altered in Kansl1+/- mice . We also emphasize that two overexpressed genes , Ucn and Adcyap1 , and one underexpressed gene , Chd1l , found deregulated in Del/+ mice are linked to corticotropin release and are associated with stress response . Those genes may be relevant for the overly friendly social phenotypes observed in 17q21 . 31 deletion carriers . Electrophysiological experiments confirmed that dosage of one or several genes within the 17q21 . 31 syntenic region affects basal synaptic transmission and short-term plasticity of excitatory synapses in the hippocampus . Noted disturbances in the expression level of several genes could contribute to this impairment . For example , dysregulation of Cntnap2 could affect migration of interneurons [37] and inhibitory synaptic function [43] , which could , in turn , alter excitatory synaptic responses . Other gene affected by 17q21 . 31 mutations is Nrxn1 that shapes the balance between excitatory and inhibitory synaptic activity [44 , 45] . Such a defect at the expression level may account for the change in synaptic strength and impaired learning and memory . In conclusion , this study confirms a previously hypothesized role of KANSL1 in the manifestation of KdVS phenotypes and extends substantially our knowledge about biological processes affected by these mutations . With these new genetic tools , we can explore the function of these genes and dissect further the pathophysiological mechanisms to eventually inform potential therapeutic avenues .
The 17p21 . 31 mutant mice carrying the deletion of the Arf2–Kansl1 region ( noted Del/+ ) , or the reciproqual duplication ( noted Dup/+ ) , were generated on the C57BL/6N genetic background ( see Supplementary information ) . The Kansl1+/- mutant mice were derived in a C57BL/6N genetic background from the unique IKMP ES cell clone HEPD0766_8_G02 . Kansl1tm1b ( EUCOMM ) Hmgu[46] animals were obtained by breeding Kansl1tm1a ( EUCOMM ) Hmgu/+ mice with animals expressing the Cre recombinase[47] to generate the Kansl1tm1b ( EUCOMM ) Hmgu/+ ( noted here Kansl1+/- ) . The local ethics committee , Com’Eth ( n°17 ) , approved all mouse experimental procedures , under the accreditation number 2012–069 . Behavioral studies were conducted in 12-20-week old animals . All assessments were scored blind to the genotype as recommended by the ARRIVE guidelines[48 , 49] . All the experimental procedures for behavioral assessments have been described[50 , 51] and are detailed in the supplementary information . Craniofacial phenotyping is described in the supplementary data . Magnetic resonance imaging ( MRI ) was used to identify alterations of brain regions in 17q21 . 31-homologous CNV mice ( 8 Del/+ , 10 wt , 11 Del/Dup , and 8 Dup/+ mice ) . MRI scans were acquired from 41 male mice at 43 weeks of age with specimens prepared as described[52] and detailed with the image processing in the supplementary information . Acute hippocampal slices were used to record field excitatory post synaptic potentials ( fEPSPs ) , by using an electrophysiological suite of 8 MEA60 set-ups consisting of a MEA1060-BC pre-amplifier and a filter amplifier ( gain 550× ) ( Multi Channel Systems , Reutlingen , Germany ) as described[50 , 53] . All experiments were performed using two-pathway stimulation of the Schaffer collateral/commissural fibers in the CA1 area of 350-μm thick hippocampal slices ( see supplementary information ) . Adult hippocampi from Del/+ , Kansl1+/- and wt mice were dissected and snap-frozen in liquid nitrogen . Tissue samples were ground in a liquid-nitrogen chilled mortar and the resulting powder was used for ChIP . Chipping for H3K4me3 ( Diagenode A5051-001P ) was performed as in ( www . blueprint-epigenome . eu/UserFiles/file/Protocols/Histone_ChIP_July2014 . pdf ) . Libraries were synthetized with KAPA Hyper prep kit ( KK8504 ) following the manufacturer’s instructions . The libraries were pooled ( 4/lane ) and sequenced on the illumina HiSeq . Libraries were mapped with BWA ( 0 . 6 . 2 ) . Peaks were called with a custom C++ script and DEseq2 ( R+ ) was used to perform statistical comparisons . Data are deposited in GEO under accession GSE80311 . All enrichment analyses are made from standard hypergeometric tests with Benjamini or Bonferroni correction . GO annotations are updated to 25/6/2015 . ChIP-seq data and ChIP-seq supplementary tables were deposited in GEO and available at the link https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? token=cnavacaedzgztgz&acc=GSE80311 . Cell-types enrichments are based on the single-cell RNAseq data from Amit Zeisel et al . , 2015 , “cell types in the mouse cortex and hippocampus revealed by single-cell rna-seq” [26] . In this work , single-cell RNAseq was used to measure trascriptomes of >3000 single cells , allowing to define the markers of 11 different cell types in adult ( P21-P30 ) hippocampus and somatosensory cortex . Given that also our ChIP-seq is done on adult ( P30 ) hippocampus , single-cell RNAseq data from Zeisel et al . becomes a highly valuable resource to gain insight at the cell type level . Here we performed standard hypergeometric tests with bonferroni correction against the cell-type markers derived from data of Zeisel et al . in order to evaluate the abundance of specific markers in deregulated gene sets . A significant enrichment ( p<0 . 01 ) means that a high amount of markers of a specific cell type is found in Kansl1+/- or Del/+ deregulated genes , suggesting that the latter cell type should be particularly affected . The complete statistical data and the lists of markers found in de-regulated genes are fully available in S5 to S20 Tables . All acquired behavioral data were analyzed using a one-way ANOVA analysis with a post-hoc Tukey’s test , when applicable , non-parametric Kruskal-Wallis test or Mann-Whitney U test using the Sigma Plot software ( Ritme , France ) . The Pearson’s chi-squared test was used for mutant allele transmission . Data are represented as the mean ± s . e . m . and differences were considered to be significant if P < 0 . 05 . When comparing freezing levels between wt , Del/+ , and Dup/+ animals over different time points during extinction we used the two-Way ANOVA Repeated Measures statistical test followed by Holm-Sidak post-hoc tests to evaluate for interactions between the groups . Otherwise , when comparing wt data with those obtained from respective Del/+ or Dup/+ animals for a single time point , we used Student’s t-test . When data did not follow a normal distribution , we used the Mann-Whitney rank-based statistical test . In electrophysiological experiments input-output relationships were compared initially by a mixed model repeated-measures ANOVA and Bonferroni post hoc test implemented in Prism 5 ( GraphPad Software , San Diego , USA ) using individual slice data as independent observations . Since several slices were routinely recorded from every mouse , fEPSPmax , PPF and LTP values between wt and mutant mice were compared using two-way nested ANOVA design with genotype ( group ) and mice ( sub-group ) as fixed and random factors respectively ( STATISTICA v . 10 , StatSoft , USA ) . DF error was computed using the Satterthwaite’s method and main genotype effect was considered significant if P < 0 . 05 . Graph plots and normalization were performed using OriginPro 8 . 5 ( OriginLab , Northampton , USA ) . Electrophysiological data are presented as the mean ± s . e . m . with n and N indicating number of slices and mice respectively . | The 17q21 . 31 deletion syndrome , also named Koolen-de Vries syndrome ( KdVS ) , is a rare copy number variants associated in humans with intellectual disability , friendly behavior , congenital malformations . The syndrome is caused either by microdeletions in the 17q21 . 31 region or by variants in the KANSL1 gene in human . The reciprocal 17q21 . 31 microduplication syndrome is not so well characterized . To investigate the pathophysiology of the syndromes , we studied the deletion , the duplication of the syntenic region and a heterozygous Kansl1 mutant in the mouse . We found affected morphology and cognition , similar to human condition , with genes controlling chromatin organization , synaptic transmission and neurogenesis dysregulated in the hippocampus of KdVS models . In addition we found that synaptic transmission was altered in KdVS mice . Our results demonstrate the implication of KANSL1 in the manifestation of KdVS and extend substantially our knowledge about altered biological processes . Nevertheless , phenotypic differences between deletion and Kansl1+/- models suggested roles of other genes affected by the 17q21 . 31 deletion . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"sociology",
"brain",
"vertebrates",
"social",
"sciences",
"mice",
"neuroscience",
"animals",
"mammals",
"biological",
"locomotion",
"animal",
"models",
"model",
"organisms",
"animal",
"behavior",
"experimental",
"organism",
"sys... | 2017 | Mouse models of 17q21.31 microdeletion and microduplication syndromes highlight the importance of Kansl1 for cognition |
Myocyte enhancer factor 2 protein ( Mef2 ) is an evolutionarily conserved activator of transcription that is critical to induce and control complex processes in myogenesis and neurogenesis in vertebrates and insects , and osteogenesis in vertebrates . In Drosophila , Mef2 null mutants are unable to produce differentiated muscle cells , and in vertebrates , Mef2 mutants are embryonic lethal . Schistosome worms are responsible for over 200 million cases of schistosomiasis globally , but little is known about early development of schistosome parasites after infecting a vertebrate host . Understanding basic schistosome development could be crucial to delineating potential drug targets . Here , we identify and characterize Mef2 from the schistosome worm Schistosoma mansoni ( SmMef2 ) . We initially identified SmMef2 as a homolog to the yeast Mef2 homolog , Resistance to Lethality of MKK1P386 overexpression ( Rlm1 ) , and we show that SmMef2 is homologous to conserved Mef2 family proteins . Using a genetics approach , we demonstrate that SmMef2 is a transactivator that can induce transcription of four separate heterologous reporter genes by yeast one-hybrid analysis . We also show that Mef2 is expressed during several stages of schistosome development by quantitative PCR and that it can bind to conserved Mef2 DNA consensus binding sequences .
Schistosomiasis is a parasitic disease infecting over 200 million people worldwide with an at risk population of over 750 million [1] . The disease is caused by blood fluke worms of the genus Schistosoma , primarily by the species S . mansoni , S . haematobium , and S . japonicum . Ninety-three percent of worldwide cases of schistosomiasis ( 192 million ) occur in sub-Saharan Africa , with the highest prevalence of disease occurring in school aged children , adolescents , and young adults , where morbidity is causative of a variety of symptoms ranging from impaired physical growth to more than 280 , 000 deaths per year [2] , [3] , [4] . Due to the primary use of a single drug for treatment , praziquantel , concerns have arisen regarding the possible development of drug resistance , for which some reports suggest may be selected for in a laboratory setting [5] . Because of the profound global impact and the implications of schistosome caused disease , a complete understanding of the biology of these organisms is of importance to the research and medical community . Schistosomes have a complex parasitic life cycle requiring molluscan and mammalian hosts and an intricate process of morphological and functional changes . Free-swimming cercariae infect the mammalian host via direct penetration of the skin . Upon host skin invasion , the small 90–215 micrometer long schistosomulum , develops a digestive tract , exchanges its glycocalyx for a tegument and must grow into a 7–20 millimeter long , muscular adult worm that is uniquely adapted to its host [6] , [7] , [8] , [9] , [10] , [11] , [12] . Following migration to the venules of the hepatic portal system , male and female worms pair and produce eggs , which are excreted from the host . Eggs hatch into miracidia , another morphologically distinct free-swimming stage that will infect the molluscan host . Within the snail , sporocysts develop and produce cercariae , which exit to restart the life cycle . Recent advancements have been made in schistosome research , including RNAi gene silencing studies [13] , [14] , microarray analyses [15] , [16] , proteomic analyses [17] , [18] , laser micro-dissection microscopy of tissues [19] , and cell-specific labeling techniques allowing for in-depth tissue visualization [20] . While promising , there is still little known about the function and expression of genes within the schistosome parasite throughout its complex life cycle . Myocyte enhancer factor ( Mef2 ) proteins are members of the evolutionarily conserved MADS-box family , named after the four initially discovered members ( Mini Chromosome Maintenance 1 , AGAMOUS , Deficiens , and Serum Response Factor ) , found in yeast , Arabidopsis thaliana , Antirrhinum majus , and humans , respectively [21] , [22] , [23] , [24] . The function and expression of Mef2 proteins have been investigated in many organisms , where they function as transcription factors that regulate cellular differentiation , morphogenesis , proliferation , T-cell selection , and survival [25] , [26] , [27] . Mef2 proteins also have a conserved Mef2 protein domain , which functions with the MADS domain in protein dimerization , DNA binding , and co-factor interactions; these domains are coupled with a highly variable , usually C-terminal domain , that functions in transcriptional activation [26] , [28] , [29] . Mef2 proteins bind as both homo- and heterodimers to a ten base pair Mef2 consensus sequence CTAWWWWTAG that is evolutionarily conserved across multiple species , although some variations to this consensus have been observed [30] , [31] , [32] , [33] . The single Mef2 gene in Drosophila ( D-Mef2 ) is perhaps the most studied Mef2 homolog [28] , and is required for embryonic myogenesis [34] . Knockout studies have shown that the loss of D-Mef2 results in a complete loss of all muscle tissues [35] and loss of differentiation in all muscle cell lineages: somatic , cardiac , and visceral [36] . Chromatin immunoprecipitation ( ChIP ) and DNA microarrays have been used to elicit over 200 directly targeted genes in Drosophila and more than 650 regions in the genome bound by the activator [36] , [37] . One such target , Actin57B , is directly regulated by D-Mef2 for differential expression in cardiac , skeletal , and muscle cell lineages , where it binds to the target consensus sequence CTATTTTTAG contained in the Actin57B promoter [38] . The vertebrate Mef2 family has four alternate spliceforms , Mef2A-D , characterized by highly varied C terminal activation domains [26] and overlapping , yet distinct patterns of expression [28] . In mammals , Mef2 requires interaction with other myogenic basic Helix Loop Helix transcriptional activators to direct myogenic differentiation [39] . The four homologs play a significant role in vertebrate heart development , where they act as regulators of other important cardiac transcription factors [28] , [40] . In addition to its role in myogenesis , vertebrate Mef2 acts as a regulator of neural crest and craniofacial development in both zebrafish [41] and mammals [42] , and aids in activation of bone development [43] , neuronal differentiation [44] , [45] , muscle regeneration [46] , and T-Cell development [28] . Mef2 proteins have conserved DNA binding sites . The vertebrate Mef2A , -C , and –D have similar DNA binding ability , although Mef2B exhibits a reduced binding efficacy [26] . Mef2A and Mef2D bind the most common Mef2 binding consensus sequence , CTAAAAATAG [33] , [47] . The Mef2 homolog Rlm1 , from the budding yeast , Saccharomyces cerevisiae , also binds this consensus sequence and can heterodimerize with mammalian Mef2A . Rlm1 functions in the mitogen activated protein kinase pathway as an important mediator in cell wall biosynthesis [48] . Recently , a Mef2 homolog in the liverwort plant species Marchantia polymorpha ( M . polymorpha ) was shown to play a role in gametophytic generation with affinity as a homodimer for the sequences CTATTTTTAG and CTATATATAG [49] , showing the evolutionary conservation of Mef2 proteins and Mef2 binding properties . This evidence suggests that Mef2 is a pivotal and highly conserved transcriptional activator , making it a prime target of interest in the functional genetics of S . mansoni . Here , we identify and characterize the Schistosoma mansoni Mef2 ( SmMef2 ) . We demonstrate that it is a functional transcriptional activator , that it is expressed during different stages of schistosome development , and that it recognizes Mef2 specific target sequences . Finally , we present several genes that may be potential downstream targets of SmMef2 .
The protein sequence of the yeast Mef2 homolog Rlm1 was used for a Basic Local Alignment Tool for proteins ( BLASTp ) analysis against the S . mansoni genome databases GeneDB [50] and SchistoDB [51] . Identified sequences were then queried using Washington University protein Blast ( WU-BLASTp ) against the S . cerevisiae and H . sapiens genomes to compare MADS-box and Mef2 regions of homology . BLASTp analysis was done using the identified Mef2 homolog in S . mansoni against all genomes using NCBI BLAST ( http://blast . ncbi . nlm . nih . gov ) to identify other homologs across a variety of species . Identified protein sequences were downloaded and compared to S . mansoni Mef2 utilizing ClustalW2 ( http://www . ebi . ac . uk/Tools/msa/clustalw2/ ) [52] , [53] . The phylogram phylogenetic tree was drawn from a ClustalW-generated multiple sequence alignment of Mef2 homologs using the neighbor-joining method , and a Gonnet protein weight matrix with the gap open set at 10 , the gap extension set at 0 . 2 and the gap distances set at 5 . Total RNA was purified from six-week old adult worms , four- hour schistosomula , cercariae , and sporocysts using Trizol Reagent purification ( Invitrogen , Carlsbad , CA ) and Purelink columns ( Invitrogen , Carlsbad , CA ) . RNA was quantitated using a NANODROP 8000 spectrophotometer ( Thermo Scientific , Waltham , MA ) and gel analysis and RNA quality was assessed by visualization on a 2% agarose gel . Snails containing the Puerto Rican strain of Schistosoma mansoni originated from stocks maintained by the NIAID Schistosome Resource Center at the Biomedical Resource Institute ( Rockville , MD ) . Transformation of cercariae and culturing of schistosomula were performed as previously described [54] , [55] . DNA Primers using the InFusion Cloning System ( Clontech , Mountainview , CA ) were designed based on the identified Smp_129430 ( SmMef2 ) spliced gene sequence and the pGBKT7 vector ( Clontech , Mountainview , CA ) following manufacturer's recommendations . Primers were ordered from Integrated DNA Technologies ( IDT , Coralville , IA ) . Forward primer oAT007 ( 5′-GAA TTC CCG GGG ATC CGT CGA CTT ATG GGT CGC AAA AAA ATA CTC ATC AAG AAG-3′ ) and reverse primer oAT008 ( 5′-ATG CGG CCG CTG CAG GTC GAC TCA AAG GTG GCG CAC ACG TTT AAG AGG GTT-3′ ) were used to clone SmMef2 by the one-step RT-PCR SuperScriptIII/PlatinumTaq system ( Invitrogen , Carlsbad , CA ) using mixed total RNA from sporocyst , cercariae , and adult worms . The cDNA product was subcloned into vector pGBKT7 ( Clontech , Mountainview , CA ) at the Sal I site and this plasmid was used to transform chemically competent One Shot TOP10 cells ( Invitrogen , Carlsbad , CA ) . Colonies were selected and grown in LB containing kanamycin liquid media . Plasmid DNA was purified using the Nucleospin Plasmid miniprep kit ( Clontech , Mountainview , CA ) and verified by restriction analysis and DNA sequencing to make plasmid pEJ1108 . SmMef2 from plasmid pEJ1108 was amplified by PCR and subcloned between Not I and Sal I sites of the expression vector pMAL-c5x ( New England Biolabs , Ipswitch , MA ) using InFusion ( Clontech , Mountainview , CA ) with forward primer oAT019 ( 5′-TCC ATG GGC GGC CGC ATG GGT CGC AAA AAA ATA CTC ATC AAG ) and reverse primer oAT020 ( 5′-TTC GGA TCC GTC GAC TCA AAG GTG GCG CAC ACG TTT AAG AGG ) to make plasmid pEJ1114 . The product was analyzed by restriction analysis and sequenced . Plasmid pEJ1108 was used to transform yeast strain AH109 ( ordered from Clontech , Mountainview , CA; genotype MATa , trp1-901 , leu2-3 , 112 , ura3-52 , his3-200 , gal4Δ , gal80Δ , LYS2::GAL1UAS-GAL1TATA-HIS3 , GAL2UASGAL2TATA-ADE2 , URA3::MEL1UAS-MEL1TATA-LacZ , MEL1 ) and grown on SD -Trp plates . Transcriptional activity was tested using four different reporter genes present in the AH109 strain . Each reporter gene ( HIS3 , ADE2 , lacZ , MEL1; encoding histidine 3 , adenine 2 , beta-galactosidase , and alpha galactosidase , respectively ) is under control of a Galactose 4 protein ( Gal4 ) dependent promoter and grown on selective synthetic media ( SD ) . AH109 transformed with the pGBKT7 vector alone served as a negative control , and positive controls used AH109 transformed with pEJ780 , a pGBKT7-based plasmid containing the full Galactose 4 transcript ( GAL4 ) . Expression of Histidine and Adenine auxotrophy was also tested by spot test analysis using four , 10-fold serial dilutions of liquid synthetic media without the amino acid tryptophan ( SD –Trp ) , then grown on selective media missing either adenine , histidine or tryptophan ( SD –Ade , SD –His , SD –Trp , respectively ) . O . D . 600 values of the positive control , negative control , and experimental cultures were matched before plating using the Nanodrop 8000 . Growth was indicative of a positive result . Expression of MEL1 was tested by an α-galactosidase assay on SD -Trp where yeast cells were screened for blue color . Expression of lacZ was tested by a β-galactosidase assay and screened for blue color . RNA reverse transcription reactions were carried out using SuperScript III RT , RNAse OUT , and oligo ( dT ) 12–18 ( Invitrogen , Carlsbad , CA ) using 1 µg of total RNA extracted from sporocysts , cercariae , schistosomula , and adult worms , as per manufacturer's recommendations . The reaction was carried out for one hour at 50°C , and treated with 10 U RNase H ( New England Biolabs , Ipswitch , MA ) and placed at 37°C for 20 minutes to remove any hybridized mRNA . Quantitative Polymerase Chain Reaction ( quantitative PCR ) primers were designed using primer 3 software [56] ( Table S1 ) . Primers were checked for specificity using NCBI Primer-BLAST ( http://www . ncbi . nlm . nih . gov/tools/primer-blast/ ) and for hetero- and homo-dimers and hairpin structures using Oligo Analyzer ( http://www . idtdna . com/analyzer/applications/oligoanalyzer/ ) . Quantitative PCR was performed using SYBR Green master mix on a StepOnePlus Real-Time PCR system with StepOne Version 2 . 0 software ( Applied Biosystems , Carlsbad , CA ) , using 2 µL of the Reverse Transcriptase ( RT ) reactions described above . The following quantitative PCR conditions were used: 95°C for 10 minutes , 40 cycles of 95°C for 15 seconds and 60°C for 1 minute , followed by melt curve analyses . Replicates were manually screened , and those containing significant multiple melting peaks or bad passive reference signals were removed from analysis . Threshold cycle values and standard curve parameters were determined as described above . Cycle threshold ( CT ) values of Mef2 and potential downstream targets were evaluated by 2−ΔΔCT methods using cyclophilin as a reference gene with sporocysts as an endogenous control . CT values for cyclophilin were consistent across all stages tested . Bar graphs were generated using StepOne software ( Applied Biosystems ) . Error bars represent standard error [57] . Vectors pEJ1114 and pMAL-c5x ( New England Biolabs , Ipswich , MA ) that express either a Maltose binding protein fused to SmMef2 ( MBP-SmMef2 ) or Maltose Binding Protein ( MBP ) alone were used to transform BL21 DE3 cells ( Invitrogen , Carlsbad , CA ) . Cultures were shaken at 37°C to an O . D . 600 of 0 . 9 and protein expression was induced with IPTG at a final concentration of 2 mM . The culture was shaken for 19 . 5 hr at 20°C . Cell extracts were prepared as outlined in the pMAL Protein Fusion and Purification System manual ( New England Biolabs , Ipswitch , MA ) . Briefly , cells were resuspended in column buffer with PMSF and Halt protease inhibitor cocktail ( Thermo Scientific , Waltham , MA ) , lysed with lysozyme and pulse sonification , and cleared by centrifugation at 12 , 500×g for 30 min . Protein was purified by diluting cell extract 1∶6 and running through a gravity column consisting of 10 mL Amylose High Flow Resin ( New England Biolabs , Ipswitch , MA ) . Protein was eluted with maltose and then transferred to storage buffer by utilizing 100 , 000 MWCO Amicon Ultra 2 ml centrifugal filters . Total protein concentration was quantitated by Bradford assay . DNA binding oligonucleotide sequences were designed based on the Actin57B promoter region from Drosophila melanogaster , specifically , the region from −218 bp to −188 bp of the translation start codon ( 5′-GCTGAAGGAT{ctatttttag}GCGGATCGGC-3′ ) , which contains a Drosophila Mef2 binding site ( internal brackets ) [38] . Three double-stranded oligonucleotide pairs AT11 , AT12 , and AT13 ( labeled as F for forward and R for reverse complement ) were designed that make three different versions of the Mef2 binding consensus . Oligonucleotide oAT11F ( 5′-GCTGAAGGAT{ctatttttag}GCGGATCGGC-3′ ) , the forward sequence for the double stranded oligo-pair AT11 , and reverse complement oligonucleotide oAT11R ( 5′- GCCGATCCGCCTAAAAATAGATCCTTCAGC-3′ ) were used . For subsequent oligos , the internal Mef2 site was modified to test other Mef2 binding sites . Double-stranded oligonucleotide ( ds-oligo ) AT12 contains one of the most common consensus binding sequences [47] , [58] , made up of the forward and reverse complement oligonucleotides oAT12F ( 5′-GCTGAAGGAT{ctaaaaatag}GCGGATCGGC-3′ ) and oligonucleotide oAT12R ( 5′-GCCGATCCGCCTATTTTTAGATCCTTCAGC-3′ ) . Ds-oligo AT13 contains an M . polymorpha Mef2 binding site , generated with oligonucleotide oAT13F ( 5′-GCTGAAGGAT{ctatatatag}GCGGATCGGC-3′ ) and oAT13R ( 5′- GCCGATCCGCCTATATATAGATCCTTCAGC-3′ ) [49] . Ds-oligo AT14 contains the DNA binding site for the S . cerevisiae gene NDT80 as a negative control , made with oAT14F ( 5′-GCTGAAGGAT{gtcacaaaat}GCGGATCGGC-3′ ) and oAT14R ( 5′- GCCGATCCGCATTTTGTGACATCCTTCAGC-3′ ) [59] , [60] . For all double-stranded “hot” oligonucleotides , the forward oligonucleotides were designed with a 5′ biotin label , while “cold” double-stranded oligonucleotides were designed with no biotin label; all oligonucleotides were ordered from Integrated DNA Technologies ( IDT , Coralville , IA ) . Oligonucleotides were annealed by mixing equimolar amounts of forward and reverse oligonucleotides in . 05 M NaCl , . 01 M Tris pH8 . 0 , and 1 mM EDTA . The oligonucleotides were boiled at 95°C for 10 min , and then cooled 1°C per 60 sec down to 23°C on a Multigene Thermal Cycler ( Labnet Technologies , Edison , NJ ) . Electrophoretic Mobility Shift Assay ( EMSA ) reactions were performed using the LightShift Chemiluminescent EMSA kit ( Thermo Scientific , Waltham , MA ) as described by the manufacturer . Briefly , binding reactions contained 180 fmol hot dsDNA , 36 pmol ( 200× ) cold dsDNA ( where applicable ) , 1 . 4 µg of Amylose-purified protein extract ( MBP-SmMef2 or MBP ) , and 1× supplied binding buffer , 2 . 5% glycerol , 5 mM MgCl2 , 50 ng/µL Poly ( dI-dC ) Inhibitor DNA , and . 05% NP-40 . Binding reactions were prepared on ice and placed at room temperature for 20 min . Reactions were mixed with supplied 5× loading buffer , and loaded onto a prerun 10 cm×10 cm 5% polyacrylamide/0 . 5× TBE native gel , and run in 0 . 5× TBE for 70 min at a constant voltage of 150V . The gel was placed in a blotting apparatus with a Biodyne B Pre-Cut Modified Nylon Membrane ( size 0 . 45 µm , Thermo Scientific , Waltham , MA ) and transferred in 0 . 5× TBE at constant current of 380 mA for 1 hr . Following transfer , the membrane was crosslinked on a CL-1000 Ultra Violet crosslinker ( UVP , Upland , CA ) on automatic settings for 120 mJ/cm2 . The membrane was developed according to kit protocol , and visualized with a Charged Coupled Device ( CCD ) camera set to an exposure time of 1–2 min . Potential downstream targets of SmMef2 were screened for Mef2 binding sites within 5000 bp upstream of the expected translation start codon . Fifty-six S . mansoni genes were screened and their upstream sequences were obtained from GeneDB [50] . Putative targets were selected for screening based on suspected gene function or high levels of homology as revealed by BLASTp analysis against known Mef2 targets in D . melanogaster [36] . Two types of binding sites were used for screening; strong consensus binding sites were defined by the sequence CTAWWWWTAG , the Mef2 consensus sequence [30] , [32] , [33] , while weak consensus binding sites were defined by either CTTWWWWTAG or CTAWWWWTAA . These sites differ from the common consensus by a single nucleotide in either the third or the last base pair position . The two “weak” sequences were selected based on a screening of over 200 Mef2 binding sites , which determined these two sequences to be the next most frequently occurring binding sites for Mef2 after the consensus sequence [33] . Binding sequences and distance upstream of the translation start codon were noted .
BLASTp analysis with the yeast Mef2 homolog Rlm1 against the Schistosoma mansoni genome database identified a putative myocyte enhancer factor 2 ( Smp_129430 ) , which contains conserved MADS-box and Mef2 domains ( Figure S1 ) . A WU-BLASTp with the putative SmMef2 against the S . cerevisiae yeast genome showed that the schistosome Mef2 and the yeast Mef2 homolog , Rlm1 are 49% identical and 68% similar across 88 amino acids at the N-terminus , which encodes the MADS Box and Mef2 domains of putative SmMef2 and Rlm1 ( Figure S1A ) . BLASTP against the human Mef2A protein showed greater similarity ( Figure S1B ) . We observed that putative SmMef2 and human Mef2A , Mef2B , Mef2C and Mef2D proteins are 78–80% identical and 90% similar across amino acids 1–88 at the N-terminus ( data not shown ) . These data agree with known homology between other Mef2 homologs , where the Mef2 and MADS-box domains , particularly the MADS-box , are highly conserved across species [26] , [28] . When the remaining C-terminus ( amino acids 89–661 ) of putative SmMef2 was used to search for homologs by BLASTp in NCBI , excluding putative SmMef2 , we found no protein , nor protein domains with any significant homology . Although the sequence similarity of the DNA binding domains of putative SmMef2 and human Mef2 proteins A-D are highly conserved , the lack of conservation among the activation domain of SmMef2 is not unsubstantiated as the activation domain of Mef2 proteins characteristically displays a high degree of variation , even across homologs within the same species [26] , [33] . To test whether putative SmMef2 is actively expressed , we extracted RNA from sporocysts , cercariae , and adult worms . Using a 1∶1∶1 mixture of RNA from these three developmental stages for reverse transcriptase PCR ( RNA from mixed developmental stages permits amplification in fewer reactions when the expression profile of a gene is unknown ) , we cloned a 2 , 193 base pair sequence encoding 731 amino acids . The cloned sequence is 67 amino acids larger than predicted in the schistosome database for the putative Mef2 ( Smp_129430 ) . The extra 201 nucleotides are located in predicted intron 4 . There is also a nine-nucleotide deletion AACAATAAT after nucleotide 1908 of SmMef2 ( Figure 1 ) . The corresponding nucleotide and protein sequences are found in Figure S2 . The BLASTp analysis described above was repeated with the sequenced version of putative SmMef2 with similar results , which will be referred to as SmMef2 , to distinguish it from sequence Smp_129430 . The SmMef2 DNA sequence has been submitted to the Genbank database under accession number JN900476 . Mef2 proteins are proposed to have arisen out of a duplication of the ancestral MADS box domain before the divergence of plants and animals [61] . In plants , the Mef2-like sequences are characterized as type II MADS box proteins . To further characterize SmMef2 relative to other Mef2 proteins , we constructed a Mef2 phylogenetic tree using ClustalW that includes sequences from yeast , plants , amphibians , insects , and humans ( Figure S3 ) . The phylogenetic tree demonstrates that SmMef2 is linked to other Mef2 proteins . We find it intriguing that SmMef2 is classified within the same lineage as mammalian Mef2B , suggesting that it might be a Mef2B protein , although at this point we are not convinced there are enough data to support that argument . In concurrence with other studies , this phylogenetic tree suggests that Mef2B is divergent from other Mef2 proteins in mammals [62] . Many Mef2 proteins encode for activators of transcription . If SmMef2 is a Mef2 protein , then it should potentially be able to function as a transcriptional activator ( TA ) . To test whether SmMef2 can function as a TA , we took a yeast 1-hybrid genetics approach , utilizing the yeast expression plasmid pGBKT7 ( Clontech , Mountainview , CA ) and yeast strain AH109 ( Clontech , Mountainview , CA ) . A fusion protein was generated combining the DNA binding domain of Gal4 ( Gal4-DBD ) , from the expression plasmid pGBKT7 , with the full-length SmMef2 . The AH109 strain used for the 1-hybrid analysis contains yeast GAL1 and GAL2 promoters that controls expression of 4 different reporter genes , Histidine 3 ( HIS3 ) , Adenine 2 ( ADE2 ) , beta galactosidase ( lacZ ) , and alpha galactosidase ( MEL1 ) . The Gal4-DBD alone cannot activate the reporter genes and was used as a negative control ( Figure 2A ) , whereas the complete Gal4 protein with it own activation domain can activate the reporter genes and was used as a positive control ( Figure 2B ) . If SmMef2 is a TA , then when fused to the Gal4 DBD , it will drive transcription by utilization of the transactivation domain from SmMef2 ( Figure 2C ) . To test the levels of reporter activity for ADE2 and HIS3 reporter genes , we performed standard spot tests at comparable dilutions on test plates with nutritional markers . All cells were grown to log phase , then diluted to equal cell counts measured to within 0 . 01 optical density 600 nm ( OD600 ) . These were then serially diluted and grown for 3 days at 30°C . We found that SmMef2 displayed transcriptional activity in all reporters tested ( Figure 3 ) . Serial dilutions of SmMef2 on SD -Ade and SD -His plates grew exceptionally well and did not show a reduction in spot size until 104-fold dilution , whereas the positive control had a reduction of growth at a dilution factor of 102 ( Figure 3A , 3B ) . The negative control ( Gal4-DBD alone ) did not elicit auxotrophy for adenine or histidine ( Figure 3A , 3B ) . Gal4 is a strong transcriptional activator in yeast , but we observed that cells expressing the positive control ( full length Gal4 protein ) did not grow as well as cells expressing Mef2 on SD –Ade or SD –His plates [63] , [64] . This is not surprising as there is sufficient evidence suggesting that one transcriptional coactivator can interfere with the function of another . This has been reported when using overexpression of Gal4 , where the activator is so strong that it deleteriously affects general transcription , resulting in reduced growth rate [65] . To directly test whether slower growth is due to a reduced ability of the positive control to activate the reporter genes relative to SmMef2 , or whether it is a reduction in growth rate due to overexpression of the full length Gal4 protein , we tested growth on synthetic media without tryptophan . The plasmid , pGBKT7 carries a gene that encodes for tryptophan auxotropy , and was used to clone the Gal4-DBD/SmMef2 fusion protein and the full length Gal4 protein . Growth on synthetic media without tryptophan ( SD –Trp ) does not select for reporter activity but selects for the presence of the pGBKT7 plasmid; therefore , all cells should grow equally well . On media without tryptophan , the positive control expressing full length Gal4 protein grows slower than either Mef2 or the negative control , which has the plasmid expressing the Gal4-DBD alone ( Figure 3D ) . The growth on SD –Trp correlates with the growth rate seen on SD –Ade and SD –His , and corroborates that the slower growth of the positive control is not due to an inability to activate the reporter genes . SmMef2 was also tested for positive transcriptional activity using an alpha galactosidase assay . In this screen , cells expressing the MEL1 reporter gene turn blue in the presence of X-alpha-gal ( Clontech , Mountainview , CA ) , whereas colonies not expressing the MEL1 reporter will remain white . Twenty-six transformants were picked , patched , and screened for reporter activity . All 26 colonies tested were blue ( Figure 3C ) . The positive control was blue and the negative control was white , as expected ( Figure 3C ) . Similar results were also observed when SmMef2 was tested using the beta galactosidase assay ( data not shown ) . These data demonstrate that SmMef2 is a functional activator . The Gal4 system is a common method for testing transcriptional activity in eukaryotes via heterologous in vivo expression in yeast [64] . Our demonstration of this approach using a schistosome genes suggests that this approach could be a viable strategy to elucidate functional TAs in S . mansoni , as described previously in budding yeast [66] . Mef2 proteins bind DNA directly to regulate diverse developmental programs . Mef2 proteins recognize the DNA consensus CTAWWWWTAG and bind either as a homo or heterodimer [28] , [31] , [67] , [68] . We proposed that if SmMef2 is a Mef2 protein , that it should be able to recognize a version of the Mef2 DNA consensus . To address binding capabilities of the SmMef2 , we made a protein hybrid of SmMef2 that is N-terminally fused to the maltose binding protein ( MBP-SmMef2 ) . The MBP-SmMef2 protein hybrid was expressed and purified from bacteria and tested by electrophoretic mobility shift assay ( EMSA ) to determine whether it could recognize Mef2 consensus sequences ( Figure 4A ) . Forward ( F ) and Reverse ( R ) complementary DNA oligonucleotides were designed based on a 30-basepair region of the Drosophila melanogaster Actin 57B promoter which contains a centralized Mef2 DNA-binding site . We cloned three different variations of the centralized core Mef2 sequence CTAWWWWTAG ( Figure 4A ) . Double-stranded oligonucleotide ( ds-oligo ) AT11 has the sequence CTATTTTTAG , the wildtype sequence found in the Drosophila Actin 57B promoter . Ds-oligo AT12 is the more commonly observed Mef2 consensus , CTAAAAATAG , and is recognized by human Mef2A and Mef2C proteins . The Mef2 consensus in ds-oligo AT13 ( CTATATATAG ) , was identified in the liverwort M . polymorpha and is a perfect palindrome . Finally , as a negative control , ds-oligo AT14 GTCACAAAA , does not contain a Mef2 binding consensus; it is replaced with the middle sporulation element ( MSE ) that is recognized by the yeast Ndt80 protein , a meiosis-specific transcriptional activator [59] , [60] , and should not be recognized by SmMef2 . The EMSA data produce 3 bands; the lower band is single-stranded , biotin-labeled DNA and is consistent across all samples; the mid-lower band is double-stranded labeled DNA; and the upper band is DNA shifted due to protein binding by SmMef2 . We find that SmMef2 can bind to all three Mef2 consensus ds-oligo sequences , AT11 ( CTATTTTTAG , data not shown ) , AT12 ( CTAAAAATAG ) , and AT13 ( CTATATATAG ) ; Figures 4B and 4C , Lanes 3 ) . To address whether SmMef2 preferentially binds either consensus sequence , we tested whether unlabeled “cold” AT11 , AT12 , or AT13 could compete for binding against labeled probe AT12 or AT13 . We initially predicted that SmMef2 would most likely prefer the mammalian consensus , but under these conditions , it preferred the consensus sequence from the liverwort M . polymorpha , and followed by ( in order of preference ) AT12 and AT11 ( Figure 4B and 4C , Lanes 3–6 ) . Ds-oligo AT14 , the negative control , did not compete against labeled probe for binding ( Figure 4B and 4C , Lanes 7 ) . MBP protein alone was not able to bind AT11 ( not shown ) , AT12 or AT13 ( Figure 4B and 4C , Lanes 2 ) , demonstrating that SmMef2 was solely responsible for the shift . These data , in combination with sequence conservation and the ability for the SmMef2 to act as a TA , provide strong evidence that SmMef2 encodes a Mef2 homolog . Therefore , we name this gene SmMef2 for Schistosoma mansoni myocyte enhancer factor 2 . Mef2 proteins play significant roles in development , particularly in early myogenesis and neurogenesis . We asked when SmMef2 is expressed during schistosome development . To address this question , we extracted RNA from sporocysts , cercariae , 4-hour schistosomula , and adult worms and measured transcript levels by quantitative PCR . There is little literature focused on myogenesis or neurogenesis in schistosomes , although several electrophysiological studies on muscle and nerve cells or physiological descriptions have been published [69] , [70] , [71] , [72] , [73] . We made a simple prediction that these developmental stages might undergo myogenesis or neurogenesis for the following basic functions: 1 ) sporocysts- to produce fully developed and swimming cercariae that must exit the molluscan and follow chemical and visual cues to find and invade a mammalian host , 2 ) cercariae- for the reasons just mentioned , 3 ) schistosomula- to produce new muscle and neurons needed for motility , elongation , growth and development into adult worms after transformation , and 4 ) adult worms- for muscle motility , and muscle neuronal maintenance . There is some discussion that myogenesis in newly transformed schistosomula and adult worms might occur at the site of cercarial tail separation , proposed due to the characterization of cercarial and adult musculature patterns and the speculation that development of adult musculature occurs directly from larval musculature or undifferentiated stem cells derived from larvae [72] . This could suggest that if Mef2 is involved in myogenesis , its expression should be correlative with stages during which muscle development occurs . Quantitative PCR was used to generate an expression profile of Mef2 , with cyclophilin used as a reference gene and the sporocyst stage used as a calibrator ( Figure 5 ) . Our data show a significant enrichment of SmMef2 beginning at the schistosomula stage , with decreased levels in the adult stage . The 6-fold increase in expression from sporocysts to schistosomula is especially striking given that levels drop 5-fold in cercariae; this means that Mef2 expression is upregulated over 30-fold during the transition from cercariae to schistosomula . If SmMef2 is involved in myogenesis , then this observation is consistent with the proposal by Mair et al suggesting that the origins of myogenesis initiate after larval transformation and continue in the adult worm [72] , and this increase correlates with significant changes in the organism's morphology and musculature during its transition to an adult . Since schistosome parasites express Mef2 , we asked whether we could bioinformatically identify potential targets of SmMef2 . We screened for possible schistosome homologs of 56 genes regulated by Mef2 in Drosophila [36] using a simple BLASTp analysis . Twenty-six genes had either weak or strong Mef2 binding sequences within 5 , 000 bp upstream from the predicted translation start site , while 8 of these genes contained the core Mef2 consensus sequence CTAWWWWTAG ( Table 1 ) . Three of these genes ( DNA replication licensing factor MCM7 , actin , and four and a half LIM domains ) contain a sequence tested by our EMSA analysis ( Table 1 ) . Sixteen genes are putatively involved in muscle development , three in neuronal development , and one gene in mini chromosome maintenance . Several proteins identified are homologous to those in Drosophila . These include: three actin genes homologous to Actin57B ; four tubulin beta chains homologous to the beta tubulin gene 60D ( Btub60D ) ; BMP antagonist noggin and Transforming Growth Factor beta ( TGF-ß ) family homologs to Drosophila TGFß homolog Decapentaplegic ( Dpp ) ; two Wnt-related proteins ( Wnt proteins are involved in neurogenesis , patterning and development ) [74] , [75]; crp1/csrp1/crip1 and two , four and a half LIM domains proteins with homology to Muscle LIM protein at 60A ( Mlp60A ) ; and a Genomic Screened Homeobox ( Gsx ) family homeobox protein with homology to Mesenchyme homeobox 2 protein ( Mox2 ) , that is important for early mesoderm patterning [76] and limb muscle development [77] in mice , and which is correlative with SmMef2 being involved in myogenesis in schistosome parasites . Each of these genes has a Mef2 binding site , suggesting that they may be potential targets of SmMef2 . Using quantitative PCR , we analyzed the transcriptional profile of three genes ( Netrin- Smp_146840 , tropomyosin- Smp_022170 , and tubulin e-chain- Smp_028360 ) in sporocysts , cercariae , 4 h schistosomula , and adults . Each of these genes has potential Mef2 binding sites located within 1500 base pairs upstream of the translation start sites . Unfortunately , the initial results were inconclusive , although expression level seemed to increase in adults consistently ( data not shown ) . One explanation for this result may be that the Mef2 gene is transcribed , but Mef2 protein has not yet activated its targets . This can be addressed by looking at transcript levels of potential SmMef2 targets at later time points . Here we show that schistosome parasites express a myocyte enhancer factor 2 , which we name SmMef2 for Schistosoma mansoni myocyte enhancer factor 2 . We identified SmMef2 in a search for schistosome homologs to yeast transcriptional activators . SmMef2 has the conserved MADS box and Mef2 domains found in Mef2 activators . Mef2 proteins play a role in transcriptional activation . Using the yeast 1-hybrid and EMSA , we show that SmMef2 is an activator of transcription and that it specifically recognizes Mef2 consensus sequences in vitro . Furthermore , quantitative PCR data show a developmentally regulated pattern of expression with comparatively high transcript levels in four-hour schistosomula and adult worms . This leads us to propose that SmMef2 may play a significant role in the development from schistosomula to adult worms . Finally , we describe potential targets of Mef2 genes that are found in schistosomes and contain consensus sequences in their promoter regions . Taken together , these data provide the first description of a Mef2 activator in a helminth . Mef2 resides at an early point in the evolution of animals . Understanding how myogenesis works in schistosomes could provide insights into the evolution of mammalian myogenesis . In addition , although we know that the MADS box and Mef2 domains are highly conserved , we observed that outside of this region , SmMef2 varies in sequence dramatically from its mammalian and insect homologs . This difference might provide an opportunity for its exploitation as a potential drug target . | Schistosome parasites infect more than 200 million people worldwide and cause human schistosomiasis . Free-swimming schistosome larvae are highly mobile and invade and penetrate the host's skin to perpetuate their lifecycle in their human host , growing from 90–215 micrometers in length as a schistosomulum to a 7–20 millimeter long adult worm . Few molecular pathways have been identified in schistosome worms that are important for parasite early development . The myocyte enhancer factor protein 2 is a major regulator of muscle and nerve development in mammals and insects and is highly conserved from bread yeast to vertebrates . Here we identify and characterize the Mef2 activator from parasitic schistosome worms , the first described in any parasitic worm , and delineation of its function may be important to further understanding the basic biology of schistosome early development . Additionally , since schistosomes developed early evolutionarily , an investigation of schistosome Mef2 regulatory mechanisms could lead to a greater understanding of the development of early muscle and neurogenic development in animals . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"biotechnology",
"medicine",
"biochemistry",
"developmental",
"biology",
"infectious",
"diseases",
"zoology",
"global",
"health",
"genetics",
"biology",
"genomics",
"microbiology",
"genetics",
"and",
"genomics",
"parasitology"
] | 2012 | Identification and Characterization of a Mef2 Transcriptional Activator in Schistosome Parasites |
The stress-activated protein kinase Gcn2 regulates protein synthesis by phosphorylation of translation initiation factor eIF2α . Gcn2 is activated in amino acid-deprived cells by binding of uncharged tRNA to the regulatory domain related to histidyl-tRNA synthetase , but the molecular mechanism of activation is unclear . We used a genetic approach to identify a key regulatory surface in Gcn2 that is proximal to the predicted active site of the HisRS domain and likely remodeled by tRNA binding . Mutations leading to amino acid substitutions on this surface were identified that activate Gcn2 at low levels of tRNA binding ( Gcd- phenotype ) , while other substitutions block kinase activation ( Gcn- phenotype ) , in some cases without altering tRNA binding by Gcn2 in vitro . Remarkably , the Gcn- substitutions increase affinity of the HisRS domain for the C-terminal domain ( CTD ) , previously implicated as a kinase autoinhibitory segment , in a manner dampened by HisRS domain Gcd- substitutions and by amino acid starvation in vivo . Moreover , tRNA specifically antagonizes HisRS/CTD association in vitro . These findings support a model wherein HisRS-CTD interaction facilitates the autoinhibitory function of the CTD in nonstarvation conditions , with tRNA binding eliciting kinase activation by weakening HisRS-CTD association with attendant disruption of the autoinhibitory KD-CTD interaction .
Eukaryotic cells harbor stress-activated protein kinases that allow cells to reduce bulk protein synthesis while simultaneously activating the transcription of genes encoding stress management proteins . The target of these kinases is Ser-51 of the α-subunit of translation initiation factor 2 ( eIF2α ) . Phosphorylation of eIF2α reduces the function of eIF2 in recruiting methionyl initiator tRNA to the 40S ribosomal subunit by impairing the recycling of eIF2-GDP to eIF2-GTP by guanine exchange factor eIF2B and thereby reducing the cellular concentration of eIF2·GTP·Met-tRNAiMet ternary complexes . The inhibition of ternary complex assembly diminishes the rate of general translation but enables translation preinitiation complexes to bypass multiple “decoy” AUG start codons in the mRNA leader of GCN4 mRNA and translate the coding sequences for Gcn4 , a transcriptional activator of amino acid and vitamin biosynthetic genes in budding yeast ( reviewed in [1] ) . A similar mechanism up-regulates translation of mammalian ATF4 and ATF5 mRNAs when eIF2α is phosphorylated by Gcn2 or one of the other mammalian eIF2α kinases PKR , PERK and HRI [2] [3] . PKR is a key component of the innate immune response , PERK is crucial for responding to ER stress , and HRI couples globin synthesis to heme availability in reticulocytes [4] . Interestingly , rodent Gcn2 mediates the animal’s aversion to amino acid-deficient diets [5] , dampens protein synthesis in muscle during leucine starvation [6] , and functions in lipid homeostasis [7] and in learning and memory formation [8] . Mammalian Gcn2 has also been implicated in tumor cell survival , innate and T-cell mediated immune responses , and DNA repair ( reviewed in [9] ) ; and recently mutations in human Gcn2 were linked to pulmonary hypertension [10] . Hence , elucidating the molecular mechanism of Gcn2 regulation is of importance to multiple aspects of human development and physiology . Because eIF2α kinases act by inhibiting translation , their functions must be tightly regulated to limit eIF2α phosphorylation to the appropriate stress conditions . The Gcn2 kinase domain ( KD ) is intrinsically inert and depends on interactions with four other domains within Gcn2 to achieve an active conformation ( Fig . 1A ) [11] . Latency of Gcn2 KD activity depends on a rigid hinge connecting the N- and C-lobes of the KD , promoting a partially closed active site cleft and occluded ATP-binding pocket , and also on a non-productive orientation of helix αC in the N-lobe that blocks the proper positioning of ATP phosphates for catalysis ( Fig . 1B ) [12 , 13] . Binding of uncharged tRNA to the region immediately C-terminal to the KD , related in sequence to histidyl-tRNA synthetase ( HisRS ) is required to activate Gcn2 in amino acid-starved cells ( Fig . 1A ) [14 , 15 , 16 , 17] . Authentic HisRS is the enzyme that aminoacylates tRNAHis for protein synthesis . Consistent with the fact that Gcn2 is activated by starvation for any amino acid [15] , the Gcn2 HisRS-related domain ( henceforth , just HisRS domain ) is not specific for binding histidyl tRNA [17] . An N-terminal segment in the HisRS domain that interacts with a portion of the KD containing the hinge is required for kinase activation [18] , suggesting that tRNA binding might alter the HisRS-KD interface to evoke an active conformation of the KD . A pseudokinase domain ( YKD ) , incapable of binding ATP or Mg+2 in vitro [19] , is located just N-terminal to the KD and is also required for kinase activation ( Fig . 1A ) [16 , 20] . Our recent work established that the YKD must interact directly with the KD for kinase activation and identified likely KD-YKD contact sites that can be altered to either impair or constitutively activate Gcn2 kinase function in vivo [21] . The extreme C-terminal domain ( CTD ) of Gcn2 plays multiple roles in kinase regulation , both positive and negative , including dimerization , ribosome binding , and autoinhibition of the KD ( Fig . 1A ) [1] . Activation of Gcn2 is dependent on KD dimerization [22] in a back-to-back , parallel orientation ( Fig . 1B ) , as described for the active dimer of PKR [23] . While the KD , HisRS region , and CTD are all capable of self-interaction as isolated domains , only the CTD is essential for dimerization and attendant activation of full-length Gcn2 [24 , 25] . Gcn2 likely dimerizes constitutively through CTD self-interaction [24] , and it is possible that the mode of KD dimerization switches from the antiparallel orientation seen in the crystal structure of the inactive conformation of the Gcn2 KD [12] to the parallel , PKR-like mode of dimerization required for kinase function ( Fig . 1B ) [22] [23] . Recent work elucidated the three-dimensional structure of the CTD dimer , which is evident with some differences in both yeast and mammalian Gcn2 [26] . In addition to dimerization , the CTD mediates ribosome association of Gcn2 ( Fig . 1A ) [27] , which depends on conserved basic residues that mediate RNA binding by the CTD in vitro and are crucial for activation of Gcn2 in vivo [28] . Gcn2 activation also requires trans-acting factors Gcn1 and Gcn20 , which form a complex that must interact with both the N-terminal “RWD” domain of Gcn2 and translating ribosomes for Gcn2 activation in starved yeast cells [29 , 30 , 31 , 32 , 33] . These and other findings [34] , support the model that Gcn2 is activated by uncharged tRNA that binds first to the decoding center of a translating ribosome and is subsequently transferred to the HisRS domain in Gcn2 , with Gcn1/Gcn20 enhancing one or both of these reactions involving uncharged tRNA [33] . However , stable association of mammalian Gcn2 with ribosomes was not observed [26] , and it was proposed that the RNA binding activity of the CTD supports tRNA binding by mammalian Gcn2 , in the manner described previously for yeast Gcn2 [17] . The yeast Gcn2 CTD also appears to interact with the KD in a manner that impedes kinase activation ( Fig . 1A ) [17 , 18] , as a mutation that constitutively activates Gcn2 kinase function , GCN2c-E803V ( substitutes Glu-803 for Val in the KD ) weakens interaction between the isolated KD and CTD and also increases tRNA binding by Gcn2 in vitro [17 , 18] . Consistent with an autoinhibitory function , eliminating the CTD from mouse Gcn2 activates eIF2α phosphorylation and abrogates stimulation by uncharged tRNA in vitro [35] . The finding that tRNA competed for interaction between the isolated KD and a HisRS-CTD segment of yeast Gcn2 in vitro [17] led to the proposal that the HisRS and CTD domains both dissociate from the KD on tRNA binding . However , complete dissociation of the HisRS domain seems incompatible with the subsequent finding that an N-terminal segment of the HisRS region interacts with the KD and is crucial for Gcn2 activation at a step following tRNA binding , suggesting that this portion of the HisRS domain remains engaged with the KD in the activated state [18] . In addition to the autoinhibitory CTD-KD interaction , the CTD mediates an inhibitory interaction with translation elongation factor eEF1A that can be overcome by uncharged tRNA [36] . While it is clear that tRNA binding to the HisRS domain is required for activation , and a stimulatory interaction of the HisRS-N region with the KD seems likely , it was unclear how tRNA binding might antagonize the autoinhibitory KD-CTD interaction and simultaneously promote stimulatory association of the YKD with the KD . In an effort to answer this question , we identified substitutions in the HisRS region that restore kinase activation by the gcn2-m2 variant , which harbors substitutions in conserved residues of the predicted HisRS active site cleft that impair tRNA binding in vitro and kinase activation in vivo [15 , 16 , 17] . We reasoned that such m2 suppressors could alter a regulatory surface in the HisRS whose interactions with another domain are modulated by tRNA binding in a manner mimicking the tRNA-bound state of WT Gcn2 . Interestingly , the locations of these suppressors led us to identify a regulatory patch predicted to be surface-exposed and proximal to the region in the HisRS domain corresponding to the active site of authentic HisRS , below dubbed the “pseudo-active site” , which can be altered to either activate or impede Gcn2 function . Our finding that certain of the ( Gcn- ) inactivating substitutions strengthen HisRS-CTD interaction without affecting tRNA binding in vitro implies that one stimulatory consequence of tRNA binding is to weaken HisRS-CTD association . This inference leads to an appealing model for how tRNA binding releases the autoinhibitory KD-CTD interaction , promotes YKD-KD association , and thereby activates Gcn2 .
In an effort to identify residues in the Gcn2 HisRS domain involved in regulation of kinase function by uncharged tRNA , we randomly mutagenized the coding sequence for the HisRS domain in a plasmid-borne gcn2-m2 allele and selected for suppressors of the sensitivity to 3-aminotriazole ( 3-AT ) conferred by this allele in yeast cells . The m2 mutation substitutes two residues in highly conserved motif 2 in the pseudo-active site of the HisRS domain , impairing tRNA binding by Gcn2 in vitro and abolishing activation of Gcn2 kinase function in vivo [15 , 17] . Defective activation of Gcn2 confers sensitivity to 3-AT , an inhibitor of histidine biosynthesis , by preventing Gcn2-dependent induction of GCN4 translation and attendant derepression of histidine biosynthetic enzymes under Gcn4 control . Thus , transformants of a gcn2∆ strain harboring WT GCN2 or GCN2c-M788V , whose product is constitutively activated [37] , grow well on 3-AT medium , whereas gcn2-m2 transformants do not ( Fig . 2A , rows 1–3 ) . Interestingly , we identified 3 mutations in the HisRS domain that suppress the 3-ATS phenotype of the m2 mutation , with the strongest growth on 3-AT displayed by the m2 , T1328S mutant ( Fig . 2A , 3-AT , rows 4–6 vs . 2 ) . As expected , a mutant allele combining all three suppressors with m2 also confers a strong 3-ATR phenotype ( Fig . 2 , row 7 ) . Comparison of the single and triple suppressor alleles at elevated temperature ( 37° ) , which exacerbates sensitivity to 3-AT , reveals that combining the suppressor mutations in one allele confers greater resistance to 3-AT than that given by any of the single suppressors ( S1 Fig . ) . The allele combining all three suppressors with m2 additionally conferred resistance to a combination of tryptophan analog 5-fluorotryptophan and histidine analog triazolealanine ( Fig . 2A , 5-FT/TRA , row 7 ) . The 5FTR/TRAR phenotype signifies constitutive , Gcn4-mediated derepression of tryptophan and histidine biosynthetic enzymes , known as the Gcd- phenotype [37] . Accordingly , the GCN2c-M788V allele confers growth on 5-FT/TRA medium , whereas GCN2+ cells , and Gcn- strains like gcn2-m2 , are sensitive to these analogs ( Fig . 2A , rows 1–3 ) . GCN2c-M788V alters the ATP binding pocket of the KD to elevate kinase activity at low levels of uncharged tRNA [12 , 37] . Thus , it appears that combining all three m2 suppressors confers constitutive activation of Gcn2 , even in the presence of the m2 mutation . In accordance with their suppression of the 3-ATS phenotype of m2 , all three suppressor mutations also restored Gcn2 kinase function under starvation conditions . Western blot analysis of whole cell extracts ( WCEs ) revealed that 3-AT evokes the expected increase in eIF2α phosphorylated on Ser-51 ( eIF2α-P ) relative to total eIF2α in GCN2 cells , whereas m2 cells have no detectable eIF2α-P; and M788V cells display high-level eIF2α with and without 3-AT treatment ( Fig . 2B , lanes 1–6 ) . Importantly , each of the suppressor alleles restored 3AT induction of eIF2α-P in m2 cells , without increasing Gcn2 abundance ( Fig . 2B , lanes 7–14 ) . In agreement with its 5FTR/TRAR phenotype , the m2 mutant harboring all three suppressors also displayed a greater than WT level of eIF2α-P in nonstarvation conditions ( Fig . 2B , lane 13 vs . 1 ) , indicating constitutive activation of Gcn2 . Consistent with these findings , the m2 suppressors increase expression of a Gcn4-dependent HIS4-lacZ reporter [37] . HIS4-lacZ expression in 3-AT-starved cells is ~8-fold lower in m2 versus WT cells , and each of the mutants containing one or more suppressor mutations displays substantially higher reporter expression in 3-AT treated cells compared to that seen in the m2 single mutant , although only a slight increase was observed for the A1353V suppressor ( Fig . 2C ) . The particularly large increases in HIS4-lacZ expression observed for the m2 strains harboring T1328S or the triple suppressor mutation are consistent with their marked 3-AT-resistant phenotypes ( Fig . 2A ) . It is noteworthy that all of the suppressor strains display an induction of eIF2α-P in response to 3-AT treatment ( Fig . 2B , lanes 7–14 , 3-AT + vs . – ) , implying that their Gcn2 variants can be activated by uncharged tRNA accumulating in histidine-deprived cells . As demonstrated below , the m2 mutation reduces , but does not abolish tRNA binding by Gcn2 in vitro . It was possible , therefore , that the suppressor mutations overcome the activation defect of m2 simply by restoring robust tRNA binding by the HisRS domain . Alternatively , they could increase the ability of low-levels of tRNA bound by the m2 variant of the HisRS domain to activate Gcn2 . If the latter was true , we reasoned that the suppressor mutations should elevate eIF2α phosphorylation when separated from the m2 mutation in nonstarvation conditions by enabling Gcn2 activation by the low , basal level of uncharged tRNA present in amino acid-replete cells . Consistent with this last prediction , when separated from m2 , the Y1092C and triple suppressor mutations each evoked strong resistance to 5-FT/TRA ( Fig . 2A , rows 10 , 11 , 14 ) . They also conferred marked increases in eIF2α-P ( Fig . 2D , lanes 5 , 7 , 13 ) and HIS4-lacZ expression ( Fig . 2E , lanes 3 , 4 , 7 ) under nonstarvation conditions , comparable in degree to that given by GCN2c-M788V . The A1353V single mutation conferred smaller increases in eIF2α-P accumulation and derepression of HIS4-lacZ ( Fig . 2D , lane 1 vs . 11; Fig . 2E , column 1 vs . 6 ) . Thus , it appears that the triple suppressor mutation , Y1092C , and to a lesser extent A1353V increase the ability of low-level uncharged tRNA present in nonstarved cells to stimulate Gcn2 kinase function . The triple mutant was chosen as the exemplar Gcd- variant for subsequent biochemical studies described below . Surprisingly , despite being the most effective suppressor of m2 , the T1328S single mutation produced only a slight increase in eIF2α-P ( Fig . 2D ) and no increase in resistance to 5-FT/TRA or HIS4-lacZ expression ( Fig . 2A & E ) . Thus , although T1328S restores robust activation of the m2 variant in starved cells , it does not appreciably activate otherwise WT Gcn2 in nonstarvation conditions . To evaluate the locations of the m2 suppressors in the predicted structure of the HisRS domain , we constructed a multiple sequence alignment of Gcn2 HisRS domain sequences from various fungal species ( S2 Fig . ) , and also an alignment of a subset of these HisRS domain sequences with authentic HisRS enzymes from diverse eukaryotic species ( S3 Fig . ) . The latter reveals regions of considerable sequence similarity between authentic HisRS and the Gcn2 HisRS domains spanning the region extending from motifs 1 and 2 , conserved in all class II aminoacyl tRNA synthetases , portions of the insertion domain between motifs 2 and 3 and the HisA and HisB motifs unique to HisRS enzymes , and the N-terminal half of class II motif 3 . Interestingly , the three m2 suppressors Y1092C , T1328S , and A1353V alter residues in the vicinity of motif 2 , HisB , and within motif 3 , respectively ( Fig . 3A and S3 Fig . ) . Conserved residues of these motifs include active site residues that directly contact different moieties of the intermediate HAM formed in the first step of tRNA aminoacylation ( Fig . 3A and S3 Fig . , residues labeled with H ( histidyl ) , P ( phosphate ) , S ( sugar ) , or A ( adenine ) ) . These critical residues are color-coded in the “ribbons” depiction of the crystal structure of the T . cruzi HisRS-HAM complex shown in Fig . 3C ( salmon ( H ) , cyan ( P ) , orange ( S ) , or dark gray ( A ) ) . Interestingly , six GCN2c mutations described previously [37] also alter residues located in or nearby the conserved HisRS motifs in the primary sequence , including F1134L and D1138N ( motif 2 ) , A1197G ( insertion domain ) , N1295D and H1308Y ( near HisA ) , and G1338D ( motif 3 ) ( Fig . 3A and S3 Fig . ) . Because the structure of the Gcn2 HisRS domain is unknown , we used the sequence alignment between Gcn2 and authentic HisRSs ( S3 Fig . ) and the crystal structure of T . cruzi authentic HisRS to predict the locations of m2 suppressors and GCN2c substitutions in the three-dimensional structure of the Gcn2 HisRS domain ( Fig . 3B ) . It is striking that all three m2 suppressors and 5 of the 6 previously identified GCN2c mutations alter residues within , or in proximity to , the pseudo-active site of the HisRS domain ( green residues: m2 suppressors; blue residues previously known GCN2c mutations ) . In fact , several mutations alter residues corresponding to amino acids in HisRS that make direct contacts with the adenine ( F1134L ) , histidyl ( D1138N ) , or ribose ( A1353V ) moiety , while others are located only one or two residues away in the polypeptide chain from amino acids contacting the histidyl ( Y1092C and T1328S ) or phosphate ( N1295 ) moiety of HAM ( Fig . 3A , S3 Fig . ; cf . Fig . 3B-C ) . In the cases of T1328S , A1197V and G1338D , these residues are predicted to be surface-exposed and ( at least for T1328S and G1338D ) in proximity to one another ( Fig . 3D-E ) at the “top” of the predicted pseudo-active site cleft of the Gcn2 HisRS domain ( Fig . 3B ) . The predicted locations of these last three substitutions led us to consider a model in which this surface of the HisRS domain interacts with another region in Gcn2 to regulate kinase function in a manner that is modulated by binding of uncharged tRNA to the pseudo-active site . In this view , the putative regulatory interaction involving this patch of the HisRS domain would be altered by the Gcd- m2 suppressors and GCN2c substitutions mapping in the HisRS domain in a way that mimics the effect of uncharged tRNA binding to the pseudo-active site of the WT Gcn2 HisRS domain . We reasoned that if the foregoing hypothesis is correct , then it should be possible to isolate Gcn- substitutions affecting the same exposed surface of the HisRS domain altered by the m2 suppressor T1328S and Gcd- substitution G1338D , but with the opposite effect on its putative regulatory interactions with other Gcn2 domains . To test this idea , we first determined the degree of sequence conservation of residues on this face of the HisRS domain by projecting the sequence conservation scores obtained from the alignment of fungal Gcn2 HisRS domains ( S2 Fig . ) onto a surface representation of the crystal structure of T . cruzi HisRS ( Fig . 3E ) . We then determined the phenotypes conferred by substituting two highly conserved residues , Arg-1325 and Asp-1327 , which are surface exposed and located in proximity both to one another and the residues altered by T1328S and G1338D ( Fig . 3D-E ) . Strikingly , substitutions of Arg-1325 with Ala or Glu ( R1325A , R1325E ) and substitutions of Asp-1327 with Ala or Lys ( D1327A , D1327K ) completely abrogate Gcn2 function . Thus , all four substitutions confer strong sensitivity to 3-AT ( Fig . 4A ) , eliminate detectable eIF2α-P in both nonstarvation and starvation conditions ( Fig . 4B ) , and evoke low basal expression of the HIS4-lacZ reporter at levels comparable to , or even below , that given by the m2 mutation ( Fig . 4C ) ; and all of these Gcn- phenotypes occur without any reduction in the level of Gcn2 itself ( Fig . 4B ) . These findings are consistent with the possibility that highly conserved residues Arg-1325 and Asp-1327 are critical constituents of a regulatory patch exposed on the surface of the HisRS domain near the pseudo-active site cleft ( Fig . 3E ) . Interestingly , a mutant combining the strong Gcn- mutation D1327K with the Gcd- triple m2 suppressor Y1092C/T1328S/A1353V described above exhibits a 3-AT-sensitive phenotype ( Fig . 4A ) and a defect in derepression of HIS4-lacZ expression ( Fig . 4C ) nearly indistinguishable from those seen for the D1327K single mutant , indicating that the strong activation defect conferred by D1327K is epistatic to the constitutively activating phenotype of the Gcd- suppressor substitutions . We proposed above that the Gcd- substitutions identified as m2 suppressors restore Gcn2 kinase function to the m2 variant by altering a regulatory interaction of the HisRS domain in a way that mimics the effect of uncharged tRNA and allows for Gcn2 activation at low levels of bound tRNA . To bolster this view and eliminate the alternative possibility that they simply overcome the effect of m2 of impairing tRNA binding , we purified the gcn2-m2 product and the Gcn2 variant harboring the m2 substitutions in combination with all three suppressor substitutions , and compared them to WT Gcn2 for binding [32P]-labeled total tRNA using a gel mobility shift assay to detect Gcn2-tRNA complexes . In accordance with previous results [15 , 17] , the m2 product displayed an obvious defect in tRNA binding compared to WT Gcn2; however , unlike the results of our previous studies , it retained appreciable tRNA binding activity ( Fig . 5A ) . ( This disparity in results might be attributable to the fact that , unlike the gcn2-m2 protein examined here , this variant was unstable and subject to degradation when purified from a different yeast strain used in previous studies [17] . ) The fact that m2 does not abolish tRNA binding in vitro but completely impairs activation of Gcn2 in vivo might indicate that the m2 substitutions in the pseudo-active site cleft impair a regulatory interaction of the HisRS domain with another region in Gcn2 in addition to reducing tRNA binding . Importantly , the presence of all three suppressors in a quadruple mutant harboring the m2 substitutions did not increase the tRNA binding activity compared to that measured for gcn2-m2 ( Fig . 5A ) . These findings are consistent with our conclusion that the suppressor substitutions restore eIF2α-P formation by enhancing kinase function at the low tRNA occupancy permitted by the m2 substitutions , rather than restoring high-level tRNA binding to the HisRS domain . We also examined whether Gcn- substitutions affecting the conserved , surface-exposed residues Arg-1325 and Asp-1327 proximal to the pseudo-active site affect tRNA binding . Importantly , we saw little or no effect on tRNA binding by the Gcn- substitutions D1327A and D1327K ( Fig . 5B ) , implying that they impair activation of Gcn2 by disrupting the ability of bound tRNA to trigger activation of kinase function rather than reducing the amount of bound tRNA . By contrast , the Gcn- substitution of Arg-1325 , R1325A , abolished tRNA binding by Gcn2 ( Fig . 5C ) , making it likely that substitutions of this residue impair Gcn2 activation by reducing the level of bound tRNA , although they could also disrupt the proposed regulatory interactions involving the HisRS pseudo-active site . In accordance with previous findings , deletion of HisRS residues 1048–1071 evokes a greater than WT level of tRNA binding by Gcn2 ( Fig . 5C ) , supporting our previous conclusion that removing this N-terminal segment of the HisRS domain impairs Gcn2 activation by disrupting a stimulatory interaction of the tRNA-bound HisRS domain with the KD rather than impairing tRNA binding by Gcn2 [18] . Based on its reduced electrophoretic mobility , the gcn2-∆1048–1071 variant might also exhibit a less compact conformation compared to WT Gcn2 . We wished to confirm that the key regulatory mutations of interest , the Gcn- substitutions D1327A and D1327K , and the Gcd- triple substitution Y1092C/T1328S/A1353V , alter Gcn2 kinase activity in vitro in the manner predicted by their phenotypes in vivo . To this end , we conducted in vitro kinase assays with the relevant purified Gcn2 proteins using [γ-32P]-ATP and a truncated form of recombinant eIF2α as substrates , and employed SDS-PAGE/autoradiography to detect the reaction products . It was shown previously that WT Gcn2 displays similar kinase activity whether purified from starved or nonstarved cells , but that the m2 mutation reduces kinase activity in vitro , indicating that WT Gcn2 becomes activated in vitro by deacylated tRNA in cell lysates prior to purification [16] . Thus , although yeast Gcn2 cannot be activated further by adding tRNA to kinase assays , the activity levels of Gcn2 variants with HisRS domain substitutions should reflect their abilities to be activated by tRNA during purification . Consistent with their Gcn- phenotypes , the D1327A and D1327K variants also exhibit substantially reduced autophosphorylation and eIF2α substrate phosphorylation activities in vitro ( Fig . 5D ) . Moreover , the Gcd- variant Y1092C/T1328S/A1353V exhibits an obvious increase in kinase activity relative to WT Gcn2 ( Fig . 5D ) . Our finding that Gcn- variants D1327A/D1327K are completely defective for Gcn2 activation ( Fig . 4A ) but retain robust tRNA binding activity ( Fig . 5B ) made them good candidates for mutations that alter a regulatory interaction of the HisRS region that mediates allosteric activation of kinase function by uncharged tRNA . Previously , we demonstrated that distinct segments of the isolated HisRS domain interact with the isolated KD or CTD of Gcn2 [18] . As noted above , the N-terminal HisRS segment ( HisRS-N ) interacts with the KD and the Δ1048–1071 deletion in this region impairs Gcn2 activation without reducing tRNA binding , thus identifying a stimulatory HisRS-N/KD interaction [18] . Moreover , the C-terminal HisRS segment ( HisRS-C ) was shown to interact with the CTD [18] , and as it encompasses Asp-1327 , we hypothesized that the Gcn- D1327A/D1327K substitutions impair Gcn2 function by altering the HisRS-CTD interaction . We obtained evidence supporting this hypothesis using the yeast two-hybrid assay . In agreement with previous results [24] , a LexA-fusion to the WT Gcn2 HisRS domain shows little interaction with a fusion of the B42 activation domain to the CTD . Remarkably , introducing Gcn- substitutions D1327A or D1327K into the lexA-HisRS fusion greatly enhanced this two-hybrid interaction ( Fig . 6A ) . By contrast , the Gcd- triple substitution Y1092C/T1328S/A1353V had no significant effect on the HisRS/CTD interaction when introduced into otherwise WT lexA-HisRS . Interestingly , however , these Gcd- substitutions diminished the enhanced two-hybrid interaction conferred by the D1327K Gcn- substitution ( Fig . 6A ) . In an effort to confirm the two-hybrid findings , we examined in vitro interaction of a LexA-CTD fusion expressed in yeast cells with immobilized GST fusions containing mutant or WT HisRS segments purified in yeast . As both the HisRS and CTD segments have RNA binding activity [17 , 28] , the reactants were treated with micrococcal nuclease to eliminate indirect association between these segments bridged by RNA . Paralleling the two-hybrid results , the D1327K substitution greatly increased binding of LexA-CTD to GST-HisRS compared to the low-level binding observed for both WT GST-HisRS and the variant harboring the Gcd- triple substitution Y1092C/T1328S/A1353V; and introducing the Gcd- triple substitution reduced binding by the D1327K variant ( Fig . 6B ) . It could be argued that the truncated species of the GST-HisRS-D1327K fusion that are not observed for WT GST-HisRS ( lower blot , lane 3 vs . 2 ) mediate the relatively greater binding of LexA-CTD by GST-HisRS-D1327K; however , this interpretation is inconsistent with the fact that the GST-HisRS fusion harboring substitutions D1327K/Y1092C/T1328S/A1353V contains even greater levels of the truncated species but binds relatively smaller amounts of LexA-CTD compared to GST-HisRS-D1327K ( Fig . 6B , lane 5 vs . 3 ) . Our finding that Y1092C/T1328S/A1353V does not reduce the HisRS/CTD interaction when introduced into the otherwise WT HisRS segment might be explained by proposing that a physiologically relevant interaction between the WT HisRS and CTD domains cannot be captured outside of the context of full-length Gcn2 in two-hybrid or pull-down assays unless the HisRS segment contains the Gcn- substitutions D1327A/D1327K that stabilize HisRS/CTD association . Together , the findings in Fig . 6A-B suggest that the D1327A/D1327K substitutions impair activation of Gcn2 by strengthening the HisRS-CTD domain interaction , while the Gcd- substitution Y1092C/T1328S/A1353V activates Gcn2 by weakening HisRS/CTD association . A corollary of this last conclusion is that the HisRS-CTD domain interaction in WT Gcn2 stabilizes the inactive conformation of Gcn2 , which would persist constitutively in the Gcn- mutants D1327A/D1327K with attendant impairment of Gcn2 activation . If so , then binding of uncharged tRNA to the HisRS region might be expected to weaken the HisRS-CTD domain interaction as one means of activating Gcn2 . Supporting this possibility , we found that the enhanced HisRS-CTD two-hybrid interactions conferred by D1327A or D1327K in vivo were abolished by starving the cells for isoleucine/valine by treatment with sulfometuron methyl ( SM ) ( Fig . 6C ) , an inhibitor of the ILV2-encoded biosynthetic enzyme , which is known to increase the level of uncharged tRNAIle and tRNAVal and activate Gcn2 in vivo [15 , 38] . By contrast , the previously demonstrated two-hybrid interaction between LexA-KD and B42-CTD fusion proteins was unaffected by SM treatment ( Fig . 6D ) , as was expression of the two-hybrid reporter conferred by the LexA-B42 activator . These findings are consistent with the idea that accumulation of uncharged tRNAIle and tRNAVal and their attendant increased binding to the LexA-HisRS-D1327A and LexA-HisRS-D1327K fusions weakens the ability of these LexA-HisRS proteins to form complexes with the B42-CTD fusion in vivo . They also support the idea that the gcn2-D1327A and gcn2-D1327K variants are defective for a regulatory interaction with the CTD that is normally disrupted by uncharged tRNA binding to the HisRS domain . To provide additional evidence that tRNA binding to the HisRS domain reduces its ability to interact with the CTD , we examined the effect of tRNA on this interaction in vitro . Consistent with previous results [18 , 24] , [35S]-methionine labeled HisRS fragment can be pulled down with immobilized GST fusions containing the Gcn2 KD , CTD or HisRS region itself , with the last interaction reflecting dimerization of the HisRS domain [18] ( Fig . 6E ) . Addition of increasing amounts of purified yeast tRNAPhe reduced interaction of the [35S]-HisRS fragment with all three GST fusions; however , the magnitude of the reduction was larger for GST-CTD ( ~5 . 4-fold ) compared to GST-KD ( ~1 . 8-fold ) or GST-HisRS ( ~2 . 1-fold ) . Moreover , interaction of [35S]-HisRS with GST-CTD was unaffected by an equivalent concentration of an unstructured model mRNA [39] ( Fig . 6F ) , suggesting specificity for tRNA in weakening the HisRS-CTD domain interaction . Together , these findings provide evidence that a tight interaction between the HisRS and CTD domains favors the inactive conformation of Gcn2 and that tRNA binding to the HisRS domain activates Gcn2 at least partly by weakening the HisRS-CTD interaction . As discussed below , based on previous findings indicating that direct interaction of the CTD with the KD contributes to the latency of Gcn2 kinase function [18] , we propose that the HisRS-CTD interaction helps to stabilize this inhibitory CTD-KD interaction in a manner that is diminished by uncharged tRNA binding to the HisRS domain in amino acid-starved cells . The model alluded to above envisions that the non-activated state of Gcn2 is characterized by domain interactions between the HisRS-C and CTD , and between the CTD and KD , which are destabilized by tRNA binding to the HisRS domain to evoke the activated state . We reasoned that the Y1092C/T1328S/A1353V Gcd- substitutions , which destabilize the HisRS/CTD interaction and confer constitutive activation of Gcn2 , would evoke a conformational change in full-length Gcn2 that mimics the activated , tRNA-bound state of WT Gcn2 . Supporting this possibility , we found that the Gcd- triple substitution increases the sensitivity of full-length purified Gcn2 to digestion by elastase , reducing the amount of full-length protein remaining after a fixed time of incubation compared to WT Gcn2 or the Gcn- variant D1327K ( S4 Fig . ) . Trypsin digestion also reduced the amounts of the largest intermediates in addition to the full-length protein for the Y1092C/T1328S/A1353V variant compared to the WT and D1327 proteins ( S4 Fig . ) . Judging by the amount of full-length Gcn2 remaining after partial digestion , the Gcn- variant D1327K appears to be somewhat less sensitive than WT Gcn2 to protease digestion , consistent with the tighter HisRS/CTD domain interaction conferred by D1327K ( Fig . 6A-C ) ; although this difference is less pronounced than that between WT and the Y1092C/T1328S/A1353V variant ( S4 Fig . ) . These results support the idea that activation of Gcn2 by the Y1092C/T1328S/A1353V substitutions involves the elimination of inhibitory domain interactions , which favors a less compact conformation of Gcn2 . It should be noted that in separate experiments we observed a decrease in protease sensitivity of WT Gcn2 on addition of excess tRNAPhe . While this result is ostensibly at odds with the notion that tRNA binding evokes a more extended , protease-sensitive conformation of Gcn2 , it seems possible that contacts between tRNA and the HisRS or CTD domains would reduce protease access to these Gcn2 segments and compensate for loss of protein-domain interactions in the tRNA-free state of Gcn2 .
In this study , we used a genetic approach to identify a novel regulatory surface in the HisRS domain of Gcn2 , juxtaposed to the pseudo-active site cleft where tRNA binds , which participates in the activation of kinase function in amino acid starved cells through its association with the Gcn2 CTD . One of the residues belonging to this regulatory surface , Thr-1328 , was identified by isolating suppressors of the m2 lesion in motif 2 of the HisRS domain , a mutation that reduces tRNA binding to Gcn2 and abolishes activation of kinase function in starved cells . Two other m2 suppressors alter residues Tyr-1092 and Ala-1353 located within the pseudo-active site cleft . The suppressor substitution Y1092C , as well as the combination of all three suppressor substitutions in the same protein , confer constitutive activation of Gcn2 function in the absence of the m2 substitutions—the Gcd- phenotype—and we showed that the triple substitution does not suppress the tRNA binding defect evoked by m2 . Hence , rather than influencing the level of tRNA binding , we propose that these mutations evoke a conformational change in the HisRS domain that mimics the consequences of tRNA binding to the WT HisRS region , which then mediates activation of the adjacent KD . In this view , the m2 suppressors allow rearrangement of Gcn2 to the active conformation at a lower occupancy of tRNA in the HisRS pseudo-active site . This alteration would compensate for the reduced affinity for tRNA of the m2 variant , and in the cases of Y1092C and A1353V allow for activation of otherwise WT Gcn2 by the basal level of uncharged tRNA in non-starved cells to produce the Gcd- phenotype . Our conclusion above that T1328S corrects the activation defect conferred by m2 but does not appreciably activate otherwise WT Gcn2 , ie . T1328S is not Gcd- , indicates that replacing Thr with Ser at this position promotes tRNA binding only in the context of the m2 alterations of the HisRS pseudo-active site . This restricted efficacy of T1328S is consistent with the fact that Thr-1328 is not evolutionarily conserved in Gcn2 HisRS domains , and is even substituted with Ser in some species ( S2 Fig . ) . Given that the m2 lesion abolishes Gcn2 activation in vivo but only reduces tRNA binding in vitro , the m2 substitutions might also impair regulatory interactions of the HisRS domain that can be compensated by the m2 suppressors . A second line of evidence supporting this model is that all 6 previously identified GCN2c mutations affecting the HisRS domain [37] involve substitutions mapping within , or proximal to , the pseudo-active site cleft . These mutations were identified by screening randomly mutagenized GCN2 alleles for the Gcd- phenotype , rather than selecting for m2 suppressors . The striking clustering of these 6 GCN2c substitutions in the predicted structure of the HisRS domain suggests that the pseudo-active site cleft is the key regulatory hub in this domain . The GCN2c mutations D1138N and F1134L alter residues in proximity to those substituted by the m2 suppressors Y1092C and A1353V within the pseudo-active site ( Fig . 3B ) and thus , according to our model , would evoke a rearrangement of the active site that mimics the effect of tRNA binding . The GCN2c mutations G1338D and A1197G introduce substitutions proximal to the active site , but located on a distinct surface , with Gly-1338 nearly adjacent on that surface to Thr-1328 ( altered by the m2 suppressor T1328S ) . We envision that this surface patch in the WT HisRS domain communicates with the pseudo-active site and is remodeled by tRNA binding in a manner mimicked by the Gcd- substitutions G1338D , A1197G , and m2 suppressor T1328S . A third line of genetic evidence supporting this model came from making targeted alanine substitutions of two HisRS domain residues that are invariant among Gcn2 homologs and exposed on the putative regulatory surface that circumscribes the Gcd- substitutions G1338D , A1197G , and m2 suppressor T1328S . Ala or Lys substitutions of the highly conserved residue D1327 completely abolish Gcn2 function in vivo while retaining robust tRNA-binding activity in vitro . It is remarkable that Gcn- and Gcd- substitutions of nearby residues belonging to this patch of the HisRS surface have opposite effects on Gcn2 activation . We envision that the Gcn- substitutions D1327K/D1327A either impede the proposed conformational remodeling of this surface patch induced by tRNA binding or alter the affinity of the remodeled surface for its binding partner within Gcn2 . The latter possibility is supported by our finding that Gcn- substitutions D1327K/D1327A enhance interaction between the HisRS and CTD domains , whereas the Gcd- triple substitution Y1092C/T1328S/A1353V partially reverses this effect in the quadruple mutant also containing D1327K . These findings imply that tight binding between the HisRS regulatory patch identified here and the CTD stabilizes the inactive conformation of Gcn2 . Consistent with this , the increased yeast two hybrid interactions between the HisRS and CTD domains evoked by Gcn- substitutions D1327K/D1327A are eliminated under conditions of isoleucine/valine starvation , in which the uncharged cognate tRNAs accumulate and Gcn2 is activated . Furthermore , interaction between the WT HisRS and CTD domains was antagonized in vitro by tRNA , but not by an equal concentration of unstructured mRNA . These findings support the idea that one aspect of Gcn2 activation by uncharged tRNA involves the ability of tRNA bound to the HisRS domain to weaken HisRS/CTD interaction . As noted above , we previously identified an autoinhibitory CTD/KD interaction that appears to be disrupted by tRNA binding to the HisRS domain [17 , 18] . More recently , we obtained strong evidence that the YKD domain stimulates Gcn2 activity by directly interacting with the KD , and proposed that the inhibitory CTD/KD interaction would compete with this stimulatory YKD/KD interaction , and that tRNA binding to the HisRS domain would shift the balance towards the stimulatory YKD/KD interaction [21] . Integrating our current findings with these previous results suggests the attractive possibility that tRNA binding to the HisRS domain antagonizes the HisRS/CTD interaction to promote a more open conformation of Gcn2 in which the CTD is less tightly bound to the KD . This would allow the YKD to compete more effectively with the CTD for binding to the KD , thereby eliminating autoinhibition by the CTD and correcting structural impediments to kinase activity inherent in the Gcn2 KD ( Fig . 7A ) . Thus , the ability of tRNA binding to weaken the HisRS/CTD interaction would provide a mechanism that serves to replace the inhibitory KD/CTD interaction with the stimulatory YKD/KD interaction . Consistent with the idea that activation of Gcn2 involves rearrangement to a more open conformation lacking domain interactions between the CTD and both the HisRS-C and KD , we found that the activating Gcd- substitution Y1092C/T1328S/A1353V increases the sensitivity of purified WT Gcn2 to digestion by elastase and trypsin . However , high-resolution structural analyses of full-length Gcn2 in the presence and absence of tRNA are clearly required for a rigorous test our model in Fig . 7A . A distinctive feature of authentic HisRS enzymes is that substrate binding involves an induced-fit mechanism in which histidine binding evokes movement of the “insertion domain” and HisA loop in a way that properly orients a key catalytic arginine residue ( Arg-259/Arg-314 of E . coli/T . cruzi HisRS ) for the formation of histidyl adenylate ( HAM ) . Binding of ATP evokes additional motion of the m2 loop , which moves yet again on ejection of pyrophosphate following HAM formation [40] [41] . The presence of a bound HAM analogue increased the affinity of E . coli HisRS for tRNAHis [42] , which might indicate that conformational changes induced by HAM binding also evoke a rearrangement of the active site that optimizes contacts with the acceptor stem of tRNAHis . Interestingly , it appears that the propensity of HisRS for histidine-induced rearrangement of the active site has been exploited to enable another HisRS-related protein , HisZ , to regulate the catalytic subunit of the octameric subfamily of ATP-phosphoribosyltransferase , HisG , an enzyme of histidine biosynthesis . HisZ contains the allosteric binding site for feedback-inhibition of HisG by histidine , and it is thought that conformational changes in the HisZ pseudo-active site evoked by histidine-binding evoke a remodeling of the HisG dimer interface to stabilize the inactive conformation [43] . Thus , in contrast to Gcn2 , the binding of histidine rather than tRNA to the HisRS subunit ( HisZ ) allosterically regulates the catalytic activity of the binding partner ( HisG ) , and the allosteric molecule ( histidine ) inhibits rather than stimulates the associated enzyme activity . Nevertheless , it seems plausible to propose that the pseudo-active site in the Gcn2 HisRS domain has evolved to evoke a conformational rearrangement of proximal , surface-exposed residues in response to binding of tRNA ( rather than histidine ) in the manner envisioned by our model . It is intriguing that the HisA loop , highly conserved among Gcn2 homologs ( S2C Fig . ) , is predicted to be juxtaposed between the 3’ end of tRNA and the HisRS regulatory surface identified here ( S5 Fig . ) , and thus could provide a path for transducing the aminoacylation status of the 3’ end of bound tRNA to the HisRS-CTD regulatory interface . As noted above , we previously identified a positive regulatory interaction between the HisRS-N segment and the KD and mapped the KD-interacting region between residues 1028–1120 [18] , which encompasses the N-terminal dimerization determinant we identified in the Gcn2 HisRS domain [18] ( Fig . 7B , see orange and brown surfaces on the two protomers of the T . cruzi HisRS dimer ) . Interestingly , this region is contiguous with that corresponding to the portion of the HisRS-C segment that interacts with the CTD [18] ( Fig . 7B , light and dark cyan surfaces that harbors the surface-exposed residues altered by the regulatory substitutions D1327A/D1327K and T1328S identified here ( red residues in Fig . 7B ) . It is tempting to propose that the contiguity of the HisRS-N and HisR-C segments will juxtapose their respective interaction partners , the KD and CTD , and enable cooperativity in KD/CTD interaction ( Fig . 7B-C ) . This model also seems compatible with the antiparallel mode of KD dimerization observed in the crystal structure of the inactive state of the Gcn2 KD [12] ( Fig . 7C , red arrow connecting KDs in the two protomers ) . Eliminating the HisRS-C/CTD interaction on tRNA binding , as we proposed above , would eliminate the proposed cooperativity and destabilize CTD binding to the KD , allowing the YKD access to the KD instead ( Fig . 7A ) . Release of the inhibitory HisRS-C/CTD interaction could also facilitate isomerization of the KDs to the parallel mode of dimerization required for their activation , and this alternative mode of dimerization could be further stabilized by the stimulatory YKD-KD interaction ( Fig . 7A ) .
Multiple sequence alignments were generated using MUSCLE at http://www . ebi . ac . uk/Tools/msa/muscle/ . ConSurf [44] and PyMOL [45] were used to obtain sequence conservation scores and project the surface representation of sequence conservation on the crystal structure of the Trypanosoma cruzi authentic HisRS ( PDB:3HRK , Fig . 3E ) . To obtain a hypothetical model of the Gcn2 HisRS-uncharged tRNA complex , the co-crystal structure of S . cerevisiae AspRS-tRNAAsp complex ( PDB: 1ASZ[Ref: PubMed: 8313877] ) was aligned with T . cruzi HisRS ( PDB: 3HRK ) by superimposing the highly conserved catalytic core domain ( 327 residues ) using Dali pairwise comparison with default parameters [Ref: PubMed: 19481444] . This alignment produced a robust Z score of 12 . 9 , a RMSD of 3 . 0 Å , and minimal clashes between tRNAAsp and HisRS . A similar alignment procedure was previously used to model HisRS interaction with tRNAHis . [Ref: PubMed 7556055; PubMed 11329259] The locations of Gcn2 residues involved in this study were then projected onto the T . cruzi HisRS crystal structure based on the sequence alignment between Gcn2 and authentic HisRSs ( S3 Fig . ) . Plasmids employed are listed in Table 1 . For Gcd- mutations identified by random mutagenesis , p2201 was subjected to error-prone PCR mutagenesis using the GeneMorph II kit ( Stratagene ) by using primer pairs PS-3 ( 5’-TCTATTTGATAACTCAGTTCCAAC-3’ ) and PS-4 ( 5’- TCAGGAATATGTATAAGAAAGGTGAC-3’ ) . The KpnI-NheI 1 . 8-kb GCN2 fragment encoding the HisRS-CTD was isolated from plasmid DNA prepared from a pool of E . coli transformants harboring mutagenized plasmids and subcloned into p2201 . Plasmid DNA prepared from a pool of the resulting E . coli transformants was introduced into yeast strain H1149 and transformants were selected on SC-Ura medium containing 15 mM 3-AT . Resident plasmids were isolated from colony-purified transformants and subjected to DNA sequence analysis to identify the relevant mutations . As multiple mutations generally occurred , QuikChange® site-directed mutagenesis ( Stratagene ) was used to produce plasmids pSL501 , pSL502 and pSL503 , containing only single mutations in GCN2 . Site-directed mutagenesis was also used to generate the novel derivatives ( listed in parenthesis ) of the following previously constructed plasmids: p2201 ( pSL501-pSL507 ) , p722 ( pSL508-pSL525 ) , pHQ430 ( pSL535-pSL538 ) , pHQ601 ( pSL539-pSL541 ) . Plasmids pSL526 , pSL527 , pSL529 , pSL530 and pSL542 were generated by replacing the 3 . 0-kb BspEI-NheI fragment in pSL101 or pSL102 with the corresponding fragment from p722 derivatives harboring the appropriate GCN2 mutations . Transformants of H2684 bearing plasmids pSL101 , pSL102 , pSL526 , pSL527 , pSL529 , pSL530 and pSL542 were grown to saturation in SC-Ura medium , diluted to A600 = 0 . 2 in SC-Ura containing 10% galactose as carbon source and grown to A600 of ∼2 . 5 . Cells were harvested ( ∼25 g ) , washed with cold distilled water containing EDTA-free protease inhibitor cocktail ( PIC ) ( Boehringer Mannheim ) and 0 . 5 mM PMSF , resuspended in ice-cold binding buffer ( BB ) ( 100 mM sodium phosphate [pH 7 . 4] , 500 mM NaCl , 0 . 1% Triton X-100 , EDTA-free PIC , 1 μg/ml leupeptin , and 1 mM PMSF ) and disrupted using SPEX freezer mill ( model 6870 ) . Lysates were clarified by centrifugation at 39 , 000 × g for 2 h at 4°C and mixed with 1 ml of M2-FLAG affinity resin ( Sigma ) overnight at 4°C . The resin was washed three times with 10 vol of BB and Gcn2 was eluted with 100 units of AcTEV protease in 500 μl of 1X TEV buffer ( 50mM Tris-HCl [pH 8 . 0] , 0 . 5 mM EDTA , 1mM DTT ) . The eluates were concentrated with an Amicon Centricon filter ( exclusion limit of Mr 10 , 000 ) and dialyzed against 10 mM Tris-HCl [pH 7 . 4] , 50 mM NaCl , 20% glycerol and stored at −800 C . The eIF2α−ΔC protein was purified from E . coli as previously described [16] . Preparation of GST and GST fusion proteins of GCN2 were carried out as described previously [24] . β-galactosidase assays of HIS4-lacZ expression were conducted on WCEs prepared from cultures grown in SD medium containing only the required supplements . For non-starvation conditions , saturated cultures were diluted 1:50 and harvested in mid-logarithmic phase after 6 h of growth . For starvation conditions , cultures were grown for 2 h under repressing conditions and then for 6 h after the addition of 3-AT to 10 mM or sulfometuron methyl ( SM ) to 0 . 5 μg/ml . β-galactosidase activity was assayed as described previously [46] and expressed as nanomoles of o-nitrophenyl-β-D-galactopyranoside hydrolyzed per min per mg of protein . For Western analysis , WCEs were prepared by trichloroacetic acid extraction , as described previously [47] , and immunoblot analysis was conducted as described [24] using phosphospecific antibodies against eIF2α-P ( Biosource International ) and polyclonal antibodies against eIF2α [48] or Gcn2 [49] . Assays of Gcn2 autophosphorylation were conducted as described previously [11] . Binding of tRNA by Gcn2 was measured using a gel mobility shift assay as described previously [21] . Pull-downs of LexA-CTD in yeast WCEs with GST-HisRS fusion proteins were conducted as follows . Immobilization of GST fusion proteins on glutathione-Sepharose 4B beads was carried out by incubating the purified fusion proteins at 0 . 5 μg/μL of beads ( bed volume ) in buffer A ( 20mM Tris/HCl pH7 . 5 , 100mM NaCl , 0 . 2mM EDTA , 1mM DTT ) containing 0 . 1% Triton X-100 at room temperature for 30 min with rocking . The beads were washed and resuspended in the same buffer . Five hundred μg of WCE prepared from pHQ311 transformants of HQY132 was treated with 12 , 000 units of micrococcal nuclease in the presence of 2mM CaCl2 for 10 min at 37°C . Nuclease-treated WCE was then added to beads ( 10-μL bed volume ) containing 5 μg of bound GST fusion proteins and the volume was increased to 200 μL with buffer A . The mixtures were incubated at 4°C for 2 h with rocking . The beads were collected by brief centrifugation in a microcentrifuge , washed three times with 500 μL of buffer A , resuspended in 40 μL of Tris-Glycine SDS Sample Buffer ( Novex ) , and fractionated by SDS-PAGE , transferred to nitrocellulose membranes , and probed with antibodies against GST or LexA . The immune complexes were visualized by enhanced chemiluminescence ( ECL; GE Healthcare Life Science ) according to the vendor’s instructions . Pull-downs of [35S]-HisRS domain fragments were conducted as follows . In vitro transcription/translation with [35S]-methionine was conducted using the TNT T7 Coupled Reticulocyte Lysate System ( Promega ) according to the vendor’s instructions . The resulting [35S]-HisRS domain fragments were partially purified by ammonium sulfate precipitation as described previously [50] and resuspended in 50 μL of buffer A ( described above ) containing 12 . 5% glycerol . Immobilization of GST fusion proteins on glutathione-Sepharose 4B beads was carried out by incubating the purified fusion proteins at 0 . 5 μg/μL of beads ( bed volume ) in buffer A containing 0 . 1% Triton X-100 at room temperature for 30 min with rocking . The beads were washed and resuspended in the same buffer . Five microliters of [35S]-HisRS domain fragments were added to beads ( 10-μL bed volume ) containing 5 μg of bound GST fusion proteins along with the indicated amount of tRNAPhe ( Sigma-Aldrich , # R4018 ) , or synthetic mRNA ( GGAAUCUCUCUCUCUCUCUCUGCUCUCUCUCUCUCUCUCUCUC ) synthesized by T7 polymerase as described in [39] , and the volume was increased to 200 μL with buffer A . The mixtures were incubated at 4°C for 2 h with rocking . The beads were collected by brief centrifugation in a microcentrifuge , washed three times with 500 μL of buffer A , resuspended in 40 μL of SDS sample buffer , and fractionated by SDS-PAGE . For detecting the [35S]-HisRS domain fragments , the gels were fixed with a solution of isopropanol:water:acetic acid ( 25:65:10 ) , treated with Amplify ( GE Healthcare Life Science ) , dried , and subjected to fluorography at −80°C . Plasmids encoding the appropriate LexA- and B42-Gcn2 fusions were cotransformed into yeast strain HQY132 . The transformants were selected on synthetic complete medium lacking uracil , histidine , and tryptophan ( SC−Ura−His−Trp ) . Two-hybrid interactions were indicated by β-galactosidase activities in cell extracts of three or more independent transformants . For these assays , cells were grown for 38 h to saturation in SC−Ura−His−Trp and were diluted 1:50 into the same medium containing galactose ( 2% ) and raffinose ( 1% ) as carbon sources ( SC/Gal/Raf−Ura−His−Trp ) . When indicated , sulfometuron was added to the medium at a final concentration of 0 . 5μg/mL . Cells were harvested in the mid-logarithmic phase after 6 h of growth . β-Galactosidase assays were carried out as described above . Aliquots of 8 μg of purified Gcn2 were incubated with 0 . 001 units of elastase ( Sigma-Aldrich ) or 2 pg of trypsin ( Sigma-Aldrich ) in 10 mM Tris-HCl [pH 7 . 4] , 50 mM NaCl , 20% glycerol for 5 min at room temperature and reactions were quenched by adding SDS sample buffer to a final concentration of 1X followed by heat inactivation at 95°C for 5 min . Digested samples were separated by SDS/PAGE and stained with Coomassie brilliant blue . | The survival of all living organisms depends on their capacity to adapt their gene expression program to variations in the environment . When subjected to various stresses , eukaryotic cells modulate general and gene-specific protein synthesis by phosphorylating the α-subunit of eukaryotic translation initiation factor 2 ( eIF2α ) . The yeast Saccharomyces cerevisiae has a single eIF2α kinase , Gcn2 , activated by uncharged tRNAs that accumulate in amino acid starved cells , which bind to a regulatory domain homologous to histidyl-tRNA synthetase ( HisRS ) . Gcn2 also contains a C-terminal domain implicated in autoinhibition of Gcn2 . Our findings identify a direct interaction between the CTD and a novel regulatory surface in the HisRS domain that is required for inhibition of Gcn2 function in non-starved cells , which is down-regulated by uncharged tRNA . The results further suggest that tRNA binding to the pseudo-active site in the HisRS domain remodels its proximal CTD-binding surface to weaken HisRS/CTD interaction and thereby release the autoinhibitory function of the CTD to activate kinase function . This study provides new molecular insights into how tRNA binding can modulate regulatory interactions among the HisRS , CTD , and kinase domains of Gcn2 to elicit kinase activation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Interaction between the tRNA-Binding and C-Terminal Domains of Yeast Gcn2 Regulates Kinase Activity In Vivo |
Many pathogens are equipped with factors providing resistance against the bactericidal action of complement . Yersinia enterocolitica , a Gram-negative enteric pathogen with invasive properties , efficiently resists the deleterious action of human complement . The major Y . enterocolitica serum resistance determinants include outer membrane proteins YadA and Ail . Lipopolysaccharide ( LPS ) O-antigen ( O-ag ) and outer core ( OC ) do not contribute directly to complement resistance . The aim of this study was to analyze a possible mechanism whereby Y . enterocolitica could inhibit the antibody-mediated classical pathway of complement activation . We show that Y . enterocolitica serotypes O:3 , O:8 , and O:9 bind C4b-binding protein ( C4bp ) , an inhibitor of both the classical and lectin pathways of complement . To identify the C4bp receptors on Y . enterocolitica serotype O:3 surface , a set of mutants expressing YadA , Ail , O-ag , and OC in different combinations was tested for the ability to bind C4bp . The studies showed that both YadA and Ail acted as C4bp receptors . Ail-mediated C4bp binding , however , was blocked by the O-ag and OC , and could be observed only with mutants lacking these LPS structures . C4bp bound to Y . enterocolitica was functionally active and participated in the factor I-mediated degradation of C4b . These findings show that Y . enterocolitica uses two proteins , YadA and Ail , to bind C4bp . Binding of C4bp could help Y . enterocolitica to evade complement-mediated clearance in the human host .
Yersinia enterocolitica is a food-borne enteric pathogen of humans and animals . The invasive strains predominantly belong to serotypes O:3 , O:9 , O:5 , 27 and O:8 . Successful colonization of the intestinal tract is the prerequisite for Y . enterocolitica infection . Bacteria pass through the acidic content of the stomach and reach the small intestine . Soon after invading the M cells , bacteria enter the lamina propria and encounter innate and adaptive immune responses [1] . Y . enterocolitica can cause enterocolitis , mesenteric lymphadenitis and , as a post-infectious complication , reactive arthritis [2] . Typically , infections lead to specific antibody responses . Mechanisms whereby Y . enterocolitica can evade the immune system killing and why it causes reactive arthritis are not fully understood . The complement system is a crucial constituent of the innate immunity . Its activation via the classical ( CP ) , lectin ( LP ) or alternative pathway ( AP ) may lead to the killing of microbes by direct lysis or by complement opsonin-enhanced phagocytosis [3] . Many pathogens , however , can resist complement attack [4] . Under microbe-free circumstances activation of the complement system must be effectively controlled as excessive activity could cause complement consumption , host cell damage or inflammation . C4b-binding protein ( C4bp ) down-regulates complement activity by acting as a fluid-phase inhibitor of the CP or LP [5] . The predominant form of this large ( 570 kDa ) octopus-shaped glycoprotein [6] consists of seven identical α-chains and one β-chain [7] , [8] . The α- and β-chains are composed of eight and three complement control protein ( CCP ) domains each , respectively . The chains are bundled together by disulphide bonds at their most C-terminal parts [9] . C4bp inhibits CP and LP activation at steps that involve C4b . The N-terminal domains of the C4bp α-chains bind C4b to prevent the assembly of the CP C3-convertase ( C4b2a ) , accelerate its natural decay and render C4b susceptible for factor I ( FI ) -mediated cleavage and inactivation [10] . Distinct pathogenic microorganisms have been demonstrated to be able to bind host complement regulators , such as factor H ( FH ) or C4bp , to exploit their protective properties and prevent complement activation on the microbial surfaces . These pathogens include e . g . Streptococcus pyogenes , S . pneumoniae , Neisseria gonorrhoeae , N . meningitidis , Borrelia burgdorferi , Escherichia coli K1 , Moraxella catarrhalis , Bordetella pertussis , Haemophilus influenzae and Candida albicans [11]–[25] . Y . enterocolitica resists efficiently complement-mediated killing [26]–[29] . This resistance depends mainly on two outer membrane proteins YadA and Ail [29]–[33] , both expressed exclusively at 37°C . YadA , encoded by the virulence plasmid ( pYV ) , is a trimeric ( monomer 43–45 kDa ) , lollipop-shaped protein composed of the head , neck , coiled-coil stalk and membrane anchor domains . The trimer projects 30 nm out from the outer membrane to form a fibrillar matrix covering the bacterial surface [34] , [35] . Ail is a 17 kDa protein encoded chromosomally . It is predicted to comprise eight membrane spanning β-strands and four extracellular loops located close to the cell membrane [36] , [37] . Ail seems to be masked to some extent by the distal parts , O-antigen ( O-ag ) and the outer core ( OC ) , of lipopolysaccharide ( LPS ) [30] . Unlike in many other Gram-negative bacteria , in Y . enterocolitica serotype O:3 , the O-ag homopolymer and the OC hexasaccharide are linked to the inner core forming a branched structure . Though both O-ag and OC are needed for colonization of the gut [38] , [39] their role in serum resistance appears to be indirect [30] . Similar to other pathogens also Y . enterocolitica binds the AP inhibitor FH [40] and we recently demonstrated that Y . enterocolitica serotype O:3 bacteria in fact use both YadA and Ail to bind the AP regulator FH ( Biedzka-Sarek et al . , submitted for publication ) but no studies on possible regulation of the CP or C4bp binding to the bacterium have been reported . Since Y . enterocolitica efficiently escapes all the complement activation pathways we examined in this study whether Y . enterocolitica also interacts with the major CP and LP regulator , C4bp . We show that C4bp binding to Y . enterocolitica is also mediated by YadA and Ail . Since the proteins are located at different layers on the bacterial surface , Ail is masked by O-ag and OC while YadA is well surface-exposed . In consequence , Ail binds C4bp when not blocked by O-ag and OC , while YadA-mediated C4bp-binding occurs regardless of the LPS expression status . As an end result the Y . enterocolitica –bound C4bp retains its function thereby modulating CP and LP activation on the bacterial surface .
Human serum samples devoid of anti-Yersinia antibodies were collected from healthy human donors and stored at −70°C . Serum was heat-inactivated ( HIS ) by incubation for 30 min at 56°C . C4bp with protein S was purified from pooled human plasma as described previously [41] . C4b , factor I and factor H were supplied by Calbiochem . C4bp was labeled with 125I ( NEN , Boston , MA ) using the Iodogen method [42] . Triton X-114 ( Tx-114 ) and heparin sodium salt were purchased from Sigma Chemicals . Phosphate-buffered saline ( PBS ) , Veronal-buffered saline ( VBS , 1 . 8 mM Na-barbital , 3 . 3 mM barbituric acid , 147 mM NaCl , pH 7 . 5 ) or Tris-based solutions were used as assay buffers . 0 . 1% gelatin – VBS ( GVBS ) or hypotonic 1/3 GVBS was used in 125I-C4bp binding assays . The bacterial strains and plasmids used in this study are listed in Table 1 . For the C4bp-binding and inhibition assays as well as for the serum adsorption assay , bacteria were grown in the RPMI 1640 medium at 37°C . This medium increases YadA expression . Prior to use , exponential-phase bacteria were washed with VBS or PBS . When appropriate , antibiotics were added to the growth medium at the following concentrations: kanamycin ( Km ) , 100 µg/ml in agar plates and 20 µg/ml in broth , chloramphenicol ( Clm ) , 20 µg/ml , and ampicillin , 50 µg/ml . To generate a YadA-negative strain of Y . enterocolitica O:8 the yadA -gene of pYV8081 was cloned in a 4 . 4 kb XbaI-PvuI fragment ( nt 45398–49838 of pYV8081 , accession number NC_008791 ) between the EcoRV and XbaI sites of pTM100 to generate plasmid pYMS3221x . The kanamycin resistance GenBlock ( KmGB ) removed by AccI digestion from pUC4K was cloned into the ClaI site within the yadA gene of pYMS3221x to obtain pYMS3223 . The expression of YadA was abolished by the KmGB-insertion ( data not shown ) . pYMS3223 was transformed into E . coli S17-1 and mobilized into the wild type Y . enterocolitica O:8 strain 8081 and recombinants , which had lost the vector plasmid due to double crossing-over between pYV8081 and pYMS3223 , were screened for by looking for KmR ClmS strains and one such strain was named as YeO8-116 . YeO8-116 did not express YadA as verified by SDS-PAGE and autoagglutination tests ( data not shown ) . Restriction digestions and Southern hybridizations of the isolated virulence plasmid of YeO8-116 showed that it carried the KmGB and an inactivated yadA -gene . In addition , YeO8-116 was calcium dependent , and produced the Yop proteins as released proteins identical to the wild type strain ( data not shown ) . The strain Ye08-116 has been used also in earlier studies [43]–[45] . Plasmid pTM100 was electroporated into the E . coli JM109 strain . Triparental conjugation was used to mobilize the plasmids pTM100 and pTM100-ail ( from JM109/pTM100 and JM109/pTM100-ail ) with a help of E . coli HB101/pRK2013 to YeO3-c-Ail-OCR . The matings were performed as described elsewhere [30] . Bacteria grown to mid-logarithmic phase were suspended in hypotonic 1/3 GVBS . A 50 µl ( 3×108 ) aliquot of the bacterial suspension was incubated with 50 µl of radiolabeled C4bp ( 5 , 000–20 , 000 cpm ) for 30 minutes at 37°C . The assays using BSA ( 0–300 nM/assay ) , unlabeled C4bp ( 0–300 nM/assay ) or factor H ( 0–300 nM/assay ) as competitors were performed with 40 µl of bacteria ( 3×107 ) in reaction mixtures containing 125I-C4bp , or 125I-BSA as a control . The effects of heparin ( 0–1000 µg/ml ) and NaCl ( 50–650 mM ) on binding were assayed in reaction mixtures containing 125I-C4bp and 40 µl of bacteria ( 3×108 ) . After incubation the mixtures were centrifuged through 20% ( w/v ) sucrose in 1/3 GVBS to pellet the bacteria with the bound radiolabeled protein . Tubes were frozen , the bottoms of the tubes were cut out and radioactivities in the pellets and supernatants were measured with a γ-counter ( Wallac , Finland ) . The ratios of bound to total activities were calculated . Bacteria ( 3×108 ) were incubated in 5% heat-inactivated serum at 37°C for 30 min . Thereafter , the bacteria were washed 5–6 times with 400 µl of PBS or 1/3 PBS and bound proteins were eluted with 0 . 1 M glycine-HCl ( pH 2 . 7 ) or PBS , respectively . Supernatants were collected and those acidified due to elution with glycine were additionally neutralized with 1 M Tris-HCl ( pH 7 . 5 ) . Samples of the last wash and elution fractions were subjected to a non-reducing 8% SDS-PAGE gel electrophoresis and subsequently transferred onto nitrocellulose membranes . The membranes were blocked and incubated with 1∶7 , 500 diluted sheep anti-human C4bp antiserum ( The Binding Site , Birmingham , UK ) and further with 1∶10 , 000 diluted peroxidase-conjugated donkey anti-sheep antiserum ( Jackson Immunoresearch ) . The proteins were detected by enhanced chemiluminescence . The cofactor assay was performed to analyze the effect of C4bp on FI-mediated cleavage of C4b . Bacteria ( 4×109 ) were incubated with C4bp ( final concentration 50 µg/ml ) in 40 µl of PBS for 30 min at 37°C with shaking . After washing for four times with PBS , bacteria ( 109 ) were pelleted and resuspended in 30 µl of PBS containing FI ( final concentration 50 µg/ml ) and C4b ( final concentration 35 µg/ml ) . The reactions were incubated for 45 min at 37°C with shaking . A reaction where C4b and FI were incubated for 45 min at 37°C with 50 µg/ml of C4bp was used a positive control . In addition , a negative control comprising of C4b and FI incubated without C4bp , was included . After incubation the samples were centrifuged , supernatants were collected , mixed with Laemmli buffer and subjected to 12 . 5% SDS-PAGE and immunoblotting using rabbit anti-human C4c antiserum ( DAKO , 1∶5 , 000 ) . The bound anti-human C4c antibodies were detected using HRP-conjugated swine anti-rabbit IgG ( DAKO; 1∶5 , 000 ) . Supernatants were also analyzed by immunoblotting with sheep anti-human C4bp antiserum to exclude unbound C4bp . The extracts Tx-YadA ( from E . coli JM103/pYMS4450 ) , Tx-Ail ( from E . coli JM109/pTM100-ail ) and vector control extracts from E . coli strains JM103/pL2 . 1 and JM109/pTM100 were prepared as described previously [46] with slight modifications . Briefly , bacteria were grown overnight at 37°C in 400 ml of Luria broth supplemented with appropriate antibiotics . Bacteria were centrifuged ( 3 , 000×g , 15 min ) and incubated on ice for 1 h in 20 ml of lysis buffer ( 10 mM EDTA , 50 mM glucose , 25 mM Tris-HCl [pH 8 . 0] , 5 mg/ml of lysozyme ) . Triton X-114 , prepared as described previously [47] , was then added to the lysate to a final concentration of 5% . The extraction was carried out by incubating the mixture at 4°C for 24 h with slow rocking . Subsequently , the mixture was incubated overnight at 37°C to separate the water and Tx-114 phases followed by centrifugation ( 4 , 000×g , 10 min ) to clear the phases . The Tx-114 phase was recovered and stored at 4°C . The OG-Ail and vector control extracts from E . coli JM109/pTM100-ail and JM109/pTM100 were prepared using OG as described elsewhere [37] . Samples from Tx-YadA , Tx-Ail and OG-Ail as well as from the three control extracts ( see above ) were run into a 7 . 5–17 . 5% SDS-PAGE gel . Proteins were electrotransferred onto a nitrocellulose membrane . Membranes were blocked with 5% skimmed milk and incubated with the radioactively labeled C4bp ( 106 cpm/assay ) or 5% NHS . Binding of the protein was detected by autoradiography or immunoblotting with rabbit anti-human C4bp ( 5 µg/ml ) as described above . YadA on the membrane was detected by immunoblotting with the monoclonal antibody 3G12 [48] . The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/ ) accession number for the gene sequence of Ye O:8 YadA is X13881 . For the protein sequences the accession numbers are CAE53849 for Ail ( Ye serotype O:3 ) and CAA32086 for YadA ( Ye 6471/76 serotype O:3 ) .
In order to study whether Y . enterocolitica binds the classical and lectin pathway complement inhibitor C4bp serotype O:3 , O:8 , O:9 and biotype 1A strains were incubated with 125I-C4bp and phase-separated from unbound 125I-C4bp by centrifugation through 20% sucrose . As shown in Fig . 1 , all strains belonging to pathogenic pYV-positive serotypes O:3 , O:8 and O:9 bound 125I-C4bp with a binding capacity of ∼20–40% . The less virulent pYV-negative biotype 1A Y . enterocolitica strain ( 27675 ) bound 125I-C4bp significantly less or not at all . The binding percentages by strains that did not bind C4bp ( including E . coli K12 C600 ) ranged between 1–8% under the conditions used ( not shown ) . To examine cofactor activity of the Y . enterocolitica–bound C4bp , serotype O:3 wild type bacteria and Ail-expressing YeO3-028-OCR strain were incubated with purified C4bp as described in Experimental procedures . A clinical isolate , less virulent Ye biotype 1A strain was included as a control because the strain bound C4bp clearly less than the other two strains used . Unbound C4bp was removed by extensive washing with PBS . The bacteria with surface-bound C4bp were subsequently incubated with FI and C4b . Following the incubation , the C4bp-cofactor activity was verified by immunodetection of C4b-cleavage products in the supernatants using anti-C4c antibodies . As a polyclonal antibody the anti-C4c antibody detects the α′-chain , the β-chain and cleavage fragments of the α′-chain . As shown in Fig . 2 , Y . enterocolitica-bound C4bp displayed cofactor activity for FI-mediated cleavage of C4b with both strains YeO3 and Ye-028-OCR . This is indicated by the appearance of the C4b α′-chain cleavage fragments ( Fig . 2 ) . The molecular weight of the cleavage fragment generated by Y . enterocolitica-bound C4bp corresponds to the fragment generated in the absence of bacteria by FI incubated with C4bp and C4b ( Fig . 2 , positive control lane ) . Clearly less cleavage products were seen with the biotype 1A strain , as expected based on the low binding of radiolabeled C4bp . No cleavage was observed when bacteria were incubated with FI and C4b suggesting that C4bp bound to bacterial surface is essential for the observed C4b cleavage ( data not shown ) . Serum resistance of Y . enterocolitica depends greatly on the expression of the pYV-encoded YadA protein [29]–[31] , [49] , [50] . Thus , we first examined the role of YadA in the acquisition of serum C4bp . To this end we incubated the wild type O:3 and O:8 strains ( YeO3 and 8081 , Table 1 ) and their YadA-negative pYV-positive derivatives ( YeO3-028 and YeO8-116 , respectively ) in heat-inactivated serum ( HIS ) . After extensive washings , the bacteria-bound serum proteins were eluted and subjected to immunoblotting using anti-C4bp antiserum . Material eluted from the wild type O:3 and O:8 strains contained C4bp while strains lacking YadA displayed significantly less C4bp in the eluted fractions ( Fig . 3 ) . This suggested that YadA is involved in C4bp binding . Residual C4bp-binding by YadA-negative strains , however , suggested a role for other Y . enterocolitica factor ( s ) in C4bp-binding . To identify C4bp receptors on Y . enterocolitica O:3 surface a set of 12 strains expressing YadA , Ail , LPS O-ag and OC in different combinations was tested for the ability to bind 125I-C4bp in 1/3 GVBS buffer containing 50 mM NaCl . YadA was indispensable for the maximal C4bp binding . Both pYV- and YadA-negative strains ( YeO3-c and YeO3-028 , respectively ) displayed equally low levels of bound C4bp suggesting that the main factor responsible for the binding is pYV-encoded YadA ( Fig . 4 ) . Almost all YadA-negative strains bound much less C4bp than the wild type strain . The two sole exceptions were strains expressing Ail in the absence of both O-ag and OC ( Fig . 4 , YeO3-c-OCR and YeO3-028-OCR ) . Both Ail-expressing strains were found to bind 125I-C4bp . The fact that the removal of either O-ag ( Fig . 4 , YeO3-028-R , YeO3-R1 ) or OC ( Fig . 5 , YeO3-028-OC , YeO3-c-OC ) was not sufficient to promote Ail-C4bp interaction suggests that either of them can block Ail-mediated C4bp-binding to the bacterial surface . To confirm Ail-mediated C4bp binding , we complemented in trans the YeO3-c-Ail-OCR strain ( YadA− , Ail− , O-ag− , OC− ) with pTM100-ail carrying the cloned ail gene . The resulting strain YeO3-c-Ail-OCR/pTM100-ail restored C4bp-binding ability , while the vector control YeO3-c-Ail-OCR/pTM100 strain was unable to bind the CP-regulator ( Fig . 5 ) . C4bp binding was shown to be specific since no binding of 125I-labeled control protein , BSA , to any of the tested strains was detected ( Fig . 5 ) . In addition , the strain with trans-complemented ail ( YeO3-c-Ail-OCR/pTM100-ail ) displayed about four-fold higher C4bp binding capacity ( 80% ) than the YeO3-c-OCR strain expressing Ail from the chromosomally-located ail gene ( Figs . 4 and 5 ) . This difference , however , can be explained by the overexpression of Ail by pTM100-ail due to a copy-number effect ( data not shown ) . Neither O-ag nor OC contributed to C4bp binding since YadA- Ail- negative strains expressing full LPS ( YeO3-c-Ail ) or its rough ( YeO3-c-Ail-R ) or OC-less ( YeO3-c-Ail-OC ) derivatives bound only negligible amounts of C4bp ( Fig . 4 ) . To see a direct interaction of YadA and Ail with C4bp we tested the binding of serum C4bp to TritonX-114 extracted YadA and Ail , Tx-YadA and Tx-Ail , respectively , in affinity blotting . C4bp binding to YadA trimer was observed ( Fig . 6 ) . Binding to Tx-Ail , however , could not be detected ( data not shown ) . Also β-octylglucoside-extracted Ail ( OG-Ail ) failed to bind C4bp when tested in a ligand blotting assay ( data not shown ) . This suggests that Ail loses its appropriate conformational structure upon extraction and/or processing for SDS-PAGE . Electrostatic forces and ion pairings have been suggested to be essential for C4bp binding to C4b [51] . The nature of C4bp binding to YadA and Ail was examined using the wild type strain YeO3 that expresses both YadA and Ail ( but the latter is blocked by O-ag and OC as demonstrated above ) , mutant strain YeO3-028-OCR that expresses unblocked Ail , and YeO3-c-Ail-OCR/pTM100-ail expressing Ail in trans . Bacteria were incubated with 125I-C4bp in 1/3 GVBS alone or supplemented with NaCl to create a salt-concentration gradient ranging from 50 to 650 mM . After centrifugation through 20% sucrose bound 125I-C4bp was measured using a gamma-counter . In general , at low salt concentrations both YadA- and Ail-mediated C4bp-binding was the highest and showed a tendency to decrease with increasing salt concentrations ( Fig . 7A ) . The C4bp binding to YeO3-c-Ail-OCR/pTM100-ail was not affected by an increase in salt concentration . YadA-mediated C4bp-binding , however , was more salt sensitive than Ail-mediated binding . A two-fold decrease in C4bp binding to YadA ( strain YeO3 ) was observed already at NaCl concentration of 100 mM while Ail-mediated C4bp-binding at this salt concentration was not affected . Thus , C4bp binding to YadA depends on ionic interactions between the proteins . The fact that C4bp-binding to YadA , similarly as that to C4b [51] , was almost completely abolished in the presence of 250 mM salt ( Fig . 7A ) , suggests that YadA and C4b have affinity for C4bp at site ( s ) with similar properties . Heparin binds to the N-termini of C4bp α-chains , i . e . to CCP1-3 [52]–[54] . We tested whether heparin inhibits the binding of C4bp to YadA- or Ail-expressing strains ( Fig . 7B ) . Bacteria were incubated with 125I-C4bp in the presence of heparin ( 0–1000 µg/ml ) . As shown in Fig . 7B heparin efficiently and dose-dependently inhibited 125I-C4bp binding to Ail , while significant reduction of 125I-C4bp binding to YadA could only be observed at the highest heparin concentration of 1000 µg/ml . Delayed response to heparin was observed with the strain over-expressing Ail ( YeO3-c-Ail-OCR/pTM100-ail ) . The relative binding of 125I-C4bp to Ail-expressing strain Ye-028-OCR could be dose-dependently reduced to 55% by adding 0–300 nM factor H . C4bp binding to YeO8 wt strain was not affected by the addition of 0–300 nM factor H ( data not shown ) . To verify whether YadA- and Ail-mediated 125I-C4bp-binding was specific , wild type and YeO3-028-OCR strains were incubated with 125I-C4bp in the presence of unlabeled C4bp or BSA ( 0–300 nM ) . In this assay , ten-fold less bacteria ( 3×107 ) were used when compared to the heparin and salt inhibition experiments . Under these conditions 125I-C4bp binding to both , YadA ( wild type ) and Ail ( YeO3-028-OCR and YeO3-c-Ail-OCR/pTM100-ail ) , was significantly inhibited in the presence of unlabeled C4bp ( Fig . 7C ) . YadA-mediated 125I-C4bp-binding was inhibited to a greater extent than that of Ail . The presence of BSA did not affect the 125I-C4bp binding neither to YadA nor to Ail ( Fig . 7D ) . With ten-fold higher numbers of bacteria , the inhibition with unlabeled C4bp could not be observed possibly because of excessive amounts of C4bp-binding surface proteins , YadA and Ail ( data not shown ) .
The complement system is an essential part of host defense against many microorganisms . A number of pathogens , however , have evolved mechanisms to subvert complement activation at different steps of the cascade . To survive and establish infection in the gut and surrounding tissues , Y . enterocolitica must resist complement-mediated opsonization and lysis . It is thus equipped with surface factors that confer resistance to serum , such as the outer membrane proteins , YadA and Ail [29]–[33] , [55] . Although it has been shown that Ail promotes resistance to complement killing , the mechanism of Ail-mediated serum resistance has remained unknown . YadA , in turn , has been shown to be the major serum resistance determinant of Y . enterocolitica [29] , [30] , [49] . Thus , not surprisingly , mechanisms underlying YadA-mediated resistance have for long been of interest . It has been speculated that the formation of YadA-composed velvet-like coat on the bacterial surface could by itself act as a shield protecting against complement [56] , [57] . There is evidence , however , for YadA-mediated binding of the alternative pathway regulator FH and inhibition of the complement cascade at both C3 and C9 levels [40] , [58] . This manifests as a reduced binding of C3b and as a failure of the membrane attack complex to incorporate into the outer membrane of Y . enterocolitica [49] . Lipopolysaccharide O-ag and OC are involved in complement-resistance indirectly [30] . They block outer membrane proteins , such as small-sized Ail , thereby having a negative influence on bacterial resistance to serum [30] . This study demonstrated a novel immune evasion mechanism of Y . enterocolitica , C4bp binding by YadA and Ail proteins . All serotypes tested , O:3 , O:8 and O:9 , were shown to bind the host complement regulator C4bp to avoid opsonophagocytosis and bactericidal action of serum ( Fig . 1 ) . The bacteria also acquired this CP-inhibitor from serum , as demonstrated by serum adsorption assays ( Fig . 3 ) . In addition , binding of purified radiolabeled C4bp to Y . enterocolitica could be observed ( Figs . 1 and 4 ) . This shows that the binding is direct and does not involve other serum proteins . Importantly , FI cofactor assay showed that C4bp bound to Y . enterocolitica surface was functionally active ( Fig . 2 ) . By binding C4bp Y . enterocolitica can thus inhibit antibody-mediated CP , and the LP . C4bp receptors on Y . enterocolitica surface were identified using a set of serotype O:3 mutants expressing YadA , Ail , O-ag and OC in different combinations ( Fig . 4 ) . Analyses of 125I-labeled C4bp binding to Y . enterocolitica O:3 strains showed that YadA was crucial for capturing this CP regulator . Therefore , YadA- or pYV-negative mutants bound only marginal amounts of C4bp , exceptions being the strains expressing Ail in the absence of O-ag and OC ( Fig . 4 ) . Thus , Ail could bind C4bp solely when accessible on the outer membrane . This was additionally confirmed by trans-complementing ail in a strain missing all four factors ( YeO3-c-Ail-OCR ) . Since this strain lacks O-ag and OC , C4bp binding to Ail was strongly favored ( Fig . 5 ) . The demonstration of C4bp receptors as YadA and Ail correlated with the previously published serum resistance results of these Ye O:3 strains [30] . These results revealed that the major serum resistance determinant of Y . enterocolitica was YadA and that the removal of LPS O-ag and OC potentiated Ail-mediated complement resistance of YadA-negative strains [30] . It is possible that during infection the production of LPS , O-ag , and OC is suppressed . Similar phenomenon has been observed for Salmonella [59] , [60] . In addition , LPS was shown not to contribute to serum resistance directly . Accordingly , in the present work we observed binding of C4bp neither to O-ag nor to OC ( Fig . 4 ) . YadA is a member of a large family of surface proteins of Gram-negative bacteria . These trimeric autotransporter proteins exert many functions and are required for full virulence of pathogenic species . Some of these proteins , such as Actinobacillus actinomyctemcomitans Omp100 , E . coli EibD , Haemophilus ducreyi DsrA and Moraxella catarrhalis UspA1 and UspA2 , confer resistance to serum [14] , [61]–[66] . Interestingly , DsrA , UspA1 and UspA2 have been shown to capture C4bp [14] , [63] . Apparently , C4bp-binding is a mechanism shared by multiple members of this family of autotransporters . The interaction between YadA and C4bp appeared to be ionic strength-dependent ( Fig . 7A ) . Interestingly , salt inhibited YadA-C4bp interaction similarly to that between C4bp and C4b [51] . The fact that the positively charged cluster of amino acids between CCP1 and CCP2 is involved in C4b binding [54] would suggest that these CCPs are needed also for the YadA-C4bp interaction . Heparin inhibition assay , however , showed discordant results ( Fig . 7B ) . The electronegative polysaccharide , heparin , alike C4b , binds to CCP1-2 , and thus partially competes with C4b for C4bp binding [52]–[54] . Heparin inhibition data showed an initial increase in C4bp binding to YadA at low heparin concentrations ( Fig . 7B ) . This could be theoretically explained by the binding of C4bp oligomers , formed in the presence of heparin , to YadA . Higher doses of heparin inhibited rather weakly C4bp binding to YadA , and 50% inhibition of the binding could only be observed at the highest heparin concentration of 1 mg/ml ( Fig . 7B ) . Thus , the C4bp binding sites for C4b and YadA do not seem to be identical , but most likely are overlapping . Electrostatic forces thus mediate the YadA-C4bp complex formation . Ail belongs to a family of β-barrel outer membrane proteins that include Salmonella enterica serovar Typhimurium PagC and Rck , and Enterobacter cloacae OmpX [67]–[71] . These proteins , though highly similar in structure , do not appear to share many of their functions . The only protein sharing serum resistance phenotype with Ail is Rck [72] . Here we provided evidence for Ail-mediated C4bp binding . The mechanism of C4bp binding to Ail appeared to be different from that of C4b and YadA , as heparin efficiently and dose-dependently inhibited the binding of C4bp to Ail ( Fig . 7B ) . This observation suggested that Ail-binding involved the CCP1-3 domains of C4bp α-chain . The Ail-C4bp interaction was also less sensitive to salt when compared to that of YadA-C4bp ( Fig . 7A ) or C4b-C4bp interactions [51] . Thus , other than electrostatic forces , e . g . hydrophobicity , could also be involved in this interaction . Ail binding sites on C4bp , however , are most likely not fully equivalent , though possibly overlapping , with those involved in C4b-C4bp or YadA-C4bp interactions . Y . pestis , the causative agent of plague , has been shown to be resistant against complement-mediated killing and to bind C4bp [25] . In contrast to Y . enterocolitica , the surface protein Ail ( also called OmpX ) seems to be solely responsible for the serum resistance property in Y . pestis [73] , [74] . Y . pestis Ail protein shares about 70% sequence identity with Y . enterocolitica Ail . As a result of several gene mutations , however , Y . pestis does not express YadA or O-ag [75] . Here we have shown that in YadA-negative Y . enterocolitica strain , removing the Ail-masking O-ag and OC greatly enhances the binding of C4bp ( Fig . 4 ) . Based on these previous findings and our current results about Ail binding C4bp , it is probable that constitutively expressed Ail binds C4bp also on the surface of Y . pestis and is , at least partly , responsible for the high complement resistance . In summary , this study provides the first evidence that Y . enterocolitica acquires the CP regulator C4bp in a functionally active form able to promote degradation of C4b . The binding depends on the two outer membrane proteins YadA and Ail , the latter binding C4bp only when well surface-exposed , i . e . , not blocked by O-ag or OC ( Fig . 8 ) . Y . enterocolitica is thus able to take advantage of the captured C4bp and is likely to be able to prevent both C4b-mediated opsonization and formation of the CP C3-convertase ( C4bC2a ) . Consequently , the bacteria can avoid complement-mediated lysis and increase their chances to survive in the human host . It is also remarkable that Y . enterocolitica uses both YadA and Ail to recruit both the AP and CP regulators FH and C4bp , respectively . This way the pathogen can ascertain that it will be protected from complement activation during different phases of infection . | To cause disease in humans , pathogenic bacteria have to evade the versatile immune system of the host . An important part of innate immunity is the complement system that is composed of over 30 proteins on host cells and in blood able to detect and destroy foreign material . To survive , bacteria can bind complement regulator proteins onto their surfaces and thus inhibit the activation of complement . Previously , it has been shown that food-borne diarrhoea-causing Yersinia enterocolitica can survive in human serum because of two bacterial surface proteins , YadA and Ail . These proteins have been shown to bind a complement alternative pathway regulator , factor H . Here , we show that both proteins also bind the classical and lectin pathway inhibitor , C4b-binding protein . These results together explain the serum resistance of Y . enterocolitica . The ability to evade complement attack is apparently important for the pathogenicity of Yersinia enterocolitica . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"microbiology/immunity",
"to",
"infections",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections",
"infectious",
"diseases/gastrointestinal",
"infections"
] | 2008 | Yersinia enterocolitica Serum Resistance Proteins YadA and Ail Bind the Complement Regulator C4b-Binding Protein |
The non-virulent Wolbachia strain wMel and the life-shortening strain wMelPop-CLA , both originally from Drosophila melanogaster , have been stably introduced into the mosquito vector of dengue fever , Aedes aegypti . Each of these Wolbachia strains interferes with viral pathogenicity and/or dissemination in both their natural Drosophila host and in their new mosquito host , and it has been suggested that this virus interference may be due to host immune priming by Wolbachia . In order to identify aspects of the mosquito immune response that might underpin virus interference , we used whole-genome microarrays to analyse the transcriptional response of A . aegypti to the wMel and wMelPop-CLA Wolbachia strains . While wMel affected the transcription of far fewer host genes than wMelPop-CLA , both strains activated the expression of some immune genes including anti-microbial peptides , Toll pathway genes and genes involved in melanization . Because the induction of these immune genes might be associated with the very recent introduction of Wolbachia into the mosquito , we also examined the same Wolbachia strains in their original host D . melanogaster . First we demonstrated that when dengue viruses were injected into D . melanogaster , virus accumulation was significantly reduced in the presence of Wolbachia , just as in A . aegypti . Second , when we carried out transcriptional analyses of the same immune genes up-regulated in the new heterologous mosquito host in response to Wolbachia we found no over-expression of these genes in D . melanogaster , infected with either wMel or wMelPop . These results reinforce the idea that the fundamental mechanism involved in viral interference in Drosophila and Aedes is not dependent on the up-regulation of the immune effectors examined , although it cannot be excluded that immune priming in the heterologous mosquito host might enhance the virus interference trait .
Wolbachia is a vertically transmitted endosymbiont that infects up to 70% of all insect species . The association is usually not obligatory for the insect and many Wolbachia strains assure their maintenance in populations by manipulating the reproduction of their host [1] . Interestingly , some strains interfere only weakly with host reproduction but still spread and are maintained in insect populations [2] . Their success may be explained by an additional positive selective advantage associated with Wolbachia infection . One possible advantage is the recently described pathogen blocking that the bacterium confers upon its host . This phenotype was first demonstrated in Drosophila , where Wolbachia induces resistance to different types of RNA viruses by reducing viral titer and/or making the host resistant to virus pathogenicity [3]–[5] . The extent and nature of blocking vary according to the virus and the Wolbachia strains tested . For example , Wolbachia reduces the titer of the closely related DCV and Nora viruses in Drosophila melanogaster and D . simulans [4] , [5] and as a consequence , the pathology associated with those two viruses is less intense in Wolbachia-infected flies [3]–[5] . In contrast , the bacterium does not affect FHV titer in Drosophila but still reduces the pathogenicity of the virus [3]–[5] . In D . simulans , the wAu Wolbachia strain has a strong effect against DCV pathogenicity , whereas the strains wHa and wNo do not [5] . This observation is thought to be related to the low infection density of wHa and wNo in Drosophila compared to that of the wAu strain [5] . Wolbachia does not naturally infect the main mosquito vector of dengue viruses , Aedes aegypti . However , two Wolbachia strains originally isolated from D . melanogaster ( wMelPop-CLA and wMel ) and one strain originally from A . albopictus ( wAlbB ) have been successfully trans-infected into A . aegypti and subsequently stably maintained [6]–[8] . All of these strains express cytoplasmic incompatibility in A . aegypti as they do in their original hosts , D . melanogaster and A . albopictus [6]–[8] . In addition , the virulent wMelPop-CLA strain that lacks normal replication control and reduces lifespan in D . melanogaster also does so in A . aegypti [7] . As observed in Drosophila , Wolbachia-infected A . aegypti are more resistant to RNA virus infection , including dengue and chikungunya [8] , [9] , as well as bacteria , nematodes and Plasmodium [9] , [10] . Transient somatic infection of the main African vector of human malaria , Anopheles gambiae , by wMelPop also significantly decreased Plasmodium infection intensity [11] . The molecular mechanisms involved in Wolbachia-mediated pathogen protection are still not clear . One plausible hypothesis is that Wolbachia interferes with pathogens by pre-activating the immune response of its host . The virulent strain wMelPop-CLA activates a wide range of immune processes in A . aegypti , including the Toll and Imd signaling pathways , anti-microbial peptide synthesis , melanization , RNA interference and opsonisation [9] , [10] and the somatic infection of An . gambiae by wMelPop caused an increase in expression of opsonisation genes [11] . Evidence for the role of opsonisation in protection against Plasmodium in this host was demonstrated by knocking down expression of the TEP1 gene [11] . Transcriptional analyses of A . aegypti immunity genes showed that wAlbB increases expression of genes in the Toll pathway and in particular the anti-microbial peptide gene , defensin [12] . Activation of the Toll pathway has been shown previously to suppress dengue infection in mosquitoes [13] . Each of these previous studies was limited in that they examined Wolbachia strains that were either virulent and/or recently introduced into naturally uninfected host species . To our knowledge , only two previous studies have examined expression of innate immune genes in insect species naturally infected by Wolbachia , including D . simulans , D . melanogaster and A . albopictus . In these cases no differences in gene regulation were observed between Wolbachia-infected insects and their uninfected counterparts [14] , [15] . Since all previous studies that have shown evidence of immune activation have been based on recently established heterologous infections , it is unclear how generalizable the Wolbachia activation of the mosquito immune system is for all insects . To determine whether immune up-regulation by the bacterium is a general mechanism underlying Wolbachia-induced dengue interference , we performed transcriptional analyses on the two heterologous associations , wMel and wMelPop-CLA infected A . aegypti , and the two native associations , wMel and wMelPop infected D . melanogaster . We also tested if the non-virulent strain wMel blocks dengue replication in Drosophila as it does in mosquitoes . If the same strain of Wolbachia blocks the replication of the same virus in different hosts , we can make the parsimonious assumption that virus interference is likely to have a common mechanistic basis across different hosts . This cross-comparison with the two Wolbachia strains and dengue virus in both native and heterologous hosts allows us to remove extraneous effects , such as recent transfer to a heterologous host or virulence associated with the wMelPop infection , that might confound an understanding of the underlying mechanistic basis of Wolbachia-induced viral interference . This study also contributes to our understanding of the physiological impact of wMel infection on A . aegypti . This is of particular relevance because wMel-infected A . aegypti have been released in north Queensland , Australia , in a field trial using Wolbachia as a biocontrol mechanism for dengue [16] . In the near future , this biological tool is also likely to be applied in dengue-endemic areas of Vietnam and Indonesia [17] .
We examined the global transcriptional response of mosquitoes to Wolbachia infection using microarrays . We compared the responses of 8 day old , non blood-fed A . aegypti females stably transinfected with wMelPop-CLA ( line PGYP1 ) or wMel ( line MGYP2 ) to those of the corresponding tetracycline-cured lines PGYP1 . tet and MGYP2 . tet . The design of the microarray included 12 , 336 transcripts , which represented 12 , 270 of the 15 , 988 genes present in the A . aegypti genome . We considered a gene to be up- or down-regulated by wMelPop-CLA or wMel infection if the fold change in transcription relative to non-infected mosquitoes was significantly different from 1 . 0 and greater than 1 . 5 . Because the Drosophila genome is better characterized , we identified Drosophila orthologs of each A . aegypti gene where possible to obtain additional functional annotations . The wMelPop-CLA infection affected the transcription of far more genes ( 2723 ) than the wMel infection ( 327 ) ( Figure 1 ) . This is likely related to wMelPop-CLA's higher density in its host , broader cellular tropism and pathogenicity [8] , [9] , [18] . Based on Gene Ontology ( GO ) annotations , wMelPop-CLA has an impact on a broader range of A . aegypti biological and molecular functions than wMel ( Table 1 , 2 ) . Many of the changes in gene regulation observed in mosquitoes infected with the virulent strain wMelPop-CLA are likely to be responses to the high physiological cost imposed by that strain . To identify mechanisms more likely to be involved in pathogen interference , we decided to focus on the 210 gene transcripts that showed significant changes in expression in both PGYP1 and MGYP2 compared to uninfected mosquitoes ( Figure 1 ) . Among those genes , 138 gene transcripts had functional annotations ( Table S1 ) . Most of the 210 transcripts were either up-regulated in both PGYP1 and MGYP2 or down-regulated in both lines ( Table S1 ) . However , the magnitude of response was typically greater to wMelPop-CLA infection ( Table S1 ) . One of the few genes differentially expressed between PGYP1 and MGYP2 is AAEL002487 , which is up-regulated in MGYP2 and down-regulated in PGYP1 . This gene encodes the protein P53 regulated pa26 nuclear protein sestrin ( dSesn in Drosophila ) ( Table S1 ) . This protein is involved in the regulation of the target of rapamycin ( TOR ) , a key protein in age-related pathologies like life-shortening or muscle degeneration [19] , two phenotypes exclusively associated with wMelPop-CLA pathogenicity in A . aegypti [7] , [20] . Among the 210 genes , most of the genes showing the greatest up-regulation are immune genes ( Table S1 ) . Gene Ontology ( GO ) annotations also revealed enrichment in genes related to immunity and proteolysis for MGYP2 and PGYP1 ( Table 1 , 2 ) . The results obtained for PGYP1 are in accordance with a previous study of A . aegypti infected by wMelPop-CLA [10] . The virulent strain wMelPop-CLA significantly affected regulation of many characterized immune genes in the mosquito ( Table S2 , [10] ) . By comparison , many fewer of these genes were activated by wMel ( Table 3 , S1 , S3 ) . Those included genes encoding anti-microbial peptides , four cecropins ( CECE , CECF , CECN , CECD ) , one defensin ( DEFC ) and one diptericin ( DPT1 ) . The magnitude of change in expression was substantial for some of these genes . The activation of these peptides is regulated by both Toll and Imd pathways , but we found up-regulation only of some Toll pathway genes , including the peptidoglycan recognition protein PGRP-SA and the Gram-negative binding proteins GNBPB4 and GNBPA1 ( GNBP1 Drosophila homologs , Table 3 ) . The Toll pathway effector defensin was the most highly up-regulated immune gene in A . aegypti infected by wMel ( Table 3 ) . This is consistent with the results of Bian et al [12] , who examined immune gene expression in heterologous wAlbB infection in A . aegypti and found that among the immune genes tested defensin was also the most up-regulated . Excluding anti-microbial peptides and the Toll pathway , the only other immune response activated by both wMel and wMelPop-CLA in A . aegypti was melanization . Four genes in this pathway were up-regulated: one pro-phenoloxidase ( PPO4 ) , one dopachrome-conversion enzyme ( DCE ) that converts dopachrome into 5 , 6-dihydroxyindole just before melanin production by phenoloxidase [21] , one putative protease inducer sp7 and one protease inhibitor Srpn4 ( Table 3 ) . The activation of these genes suggests that production of melanin is induced in Wolbachia-infected mosquitoes . Since a comparative approach between Drosophila and Aedes to examine the effect of immune activation on virus interference is predicated on an assumption that dengue virus interference also occurs in Wolbachia-infected Drosophila , we tested the ability of dengue virus serotype 2 ( DENV-2 ) to grow in Drosophila carrying the wMel Wolbachia strain . For both dengue virus strains , 92T and ET300 , the total number of flies infected by dengue was lower in the presence of wMel , with only 40% of flies detected positive for the 92T strain compared with 93% for the Wolbachia-uninfected control . Similarly for the ET300 strain , 73% of Wolbachia-infected flies were positive for dengue compared to 93% for the Wolbachia-uninfected control ( Figure 2 ) . In addition , for the flies that did become infected with dengue the amount of DENV-2 RNA present was significantly reduced in the presence of wMel ( Figure 2 ) . It was unsurprising to note that dengue grew to higher levels when injected into its natural mosquito host compared to Drosophila but regardless of absolute virus levels significant Wolbachia interference effects were detected in both insect species . Dengue injection in flies did not have an effect on insect life span nor increased mortality compared to controls ( data not shown ) . Considering that the Wolbachia strains wMelPop [22] and wMel in their original host interfere with natural Drosophila RNA viruses and also with dengue virus replication , we then investigated the possibility that both Wolbachia strains boost Drosophila immunity as seen in the heterologous mosquito host . We examined by quantitative real time PCR the expression of the Drosophila homologs of the mosquito immune genes identified through microarray analysis to be up-regulated in the presence of Wolbachia . There have been multiple gene losses and gene duplications in immune gene families in both flies and mosquitoes [23] , and we were therefore unable to reliably identify all orthologs for our anti-microbial peptide genes and pro-phenoloxidase genes of interest . Thus , we targeted all the cecropin , diptericin and pro-phenoloxidase genes present in the genome of D . melanogaster . In total 13 immune genes were analyzed: seven anti-microbial peptide genes , two Toll pathway genes and four melanization genes ( Table 4 ) . No significant changes in the expression of anti-microbial peptide genes were observed for w1118wMelPop or w1118wMel , except for cecropin A1 ( Table 4 ) . The expression of cecropin A1 was two-fold higher in the presence of wMelPop , whereas no change was observed in the presence of wMel ( Table 4 ) . No gene expression was detected for the cecropins B and C for either of the Drosophila lines tested . No significant changes in diptericin transcription were observed in Wolbachia-infected flies , which suggests that the Imd signaling pathway is not stimulated by Wolbachia in Drosophila . The expression patterns of two major genes in the Toll pathway , PGRP-SA and GNBP1 , differed between flies infected by wMel and wMelPop . A slight inhibition of PGRP-SA was observed in flies infected by wMelPop , while in wMel-infected flies there was no effect . For GNBP1 , a minor but significant difference , 1 . 29-fold change , was observed for w1118wMel but not for w1118wMelPop ( Table 4 ) . The expression of only a single melanization gene was affected by wMel infection: proPO-A1 was down-regulated . In contrast , in flies infected with wMelPop , proPO-A1 was significantly up-regulated and another melanization gene , CG42640 , was down-regulated ( Table 4 ) . An enrichment of gene transcripts encoding the iron binding proteins transferrin and ferritin was detected in the data obtained from the A . aegypti transcriptome analysis in response to wMel and wMelPop-CLA infections ( Table 1 , 2 , S1 ) . These proteins have multiple functions in insects , including iron homeostasis and immunity [24] , two potential mechanisms that could be involved in Wolbachia-mediated pathogen protection . The expression of the genes encoding transferrin 1 ( Tsf1 ) and the light chain of ferritin ( Fer2lch ) was evaluated in w1118wMel and w1118wMelPop compared to w1118tet . However , no induction was found in Wolbachia-infected flies ( Table 4 ) and wMelPop infection even significantly reduced the expression of transferrin . The expression of immune genes was also tested in the same fly lines ( w1118wMel and w1118tet ) infected with DENV-2 , strain 92T . Even in the presence of dengue , wMel infection did not increase the expression of anti-microbial peptides and pro-phenoloxidases ( Figure S1 ) . No correlation was found between the amount of dengue detected and the level of expression for each of the anti-microbial peptide and pro-phenoloxidases genes tested in each fly line ( Figure S2 ) .
Host immune priming by Wolbachia offers an appealing mechanistic explanation for pathogen blocking as it is conceivable that this single effect could lead to protection against a diversity of pathogens . The objective of this study was to compare the effect of two closely-related strains of Wolbachia on the immune system of hosts where the age of the Wolbachia association differs . By comparing wMelPop-CLA and wMel we could exclude any potential immune activation that may simply be due to the virulence of the wMelPop-CLA infection . By examining both D . melanogaster and A . aegypti , we were able to dissect aspects of the immune response that may be attributed solely to a host's response to a recently acquired Wolbachia infection . This analysis depends on an assumption that the mechanism of virus interference is similar in the two insect hosts . Considering that Wolbachia infection in Drosophila interferes with dengue replication , as it does in A . aegypti , the assumption of a similar mechanism seems parsimonious . Moreover the success of maintaining dengue in Drosophila , even if viral replication is not as strong as in A . aegypti , provides a tractable genetic model for future studies into the mechanistic basis of Wolbachia-mediated dengue interference . A previous analysis of A . aegypti whole genome transcription in response to wMelPop-CLA revealed strong immune induction by the bacterium [10] . In this present study , a similar approach was taken to analyze the impact of the non-virulent wMel strain on the immune system of A . aegypti , in comparison with the wMelPop-CLA strain . The results obtained revealed that wMel induces the activation of far fewer immunity genes in the mosquito . The comparative analysis between the different lines identified common responses only for genes encoding anti-microbial peptides , the Toll pathway and melanization-associated proteins . Recent studies have provided important insights into A . aegypti immune response to dengue virus , showing that the Toll pathway and anti-microbial peptides are important for the mosquito's defense against dengue infection [13] , [25] . Melanization is also a prominent immune response in insects against parasites like malaria and nematodes [26] but as far as we know it has never been demonstrated for dengue . The main anti-viral pathway , RNA interference [27] , seems to be activated exclusively by wMelPop-CLA . Several pieces of evidence also indicate that RNAi cannot explain virus blocking . First , Glaser et al [28] showed that even in Ago2 ( a key gene in the RNAi pathway ) mutant flies , Wolbachia infection increases resistance to viruses . Second , Frentiu et al [29] demonstrated that wMelPop-CLA induces complete inhibition of dengue virus replication in the C6/36 cell line that has been shown to be defective in the RNAi pathway [30] . This comparative analysis between wMel and wMelPop-CLA infection within A . aegypti supports the potential implication of anti-microbial peptides and Toll pathway activation in dengue virus interference by the bacterium . If we assume that the fundamental mechanism involved in Wolbachia-mediated dengue interference is the same in mosquitoes and flies , and this mechanism is immune-based , then the same constitutive immune induction should also be observed in D . melanogaster infected by wMel or wMelPop . We tested for transcriptional changes of the same immune genes identified through microarray analysis in D . melanogaster in response to Wolbachia infection , and identified a number of statistically significant changes . However , in no case were these changes consistent between wMel and wMelPop infection . Furthermore , if we employed the same threshold for biological significance we used for our microarray data , that a gene is significantly up-regulated by Wolbachia infection only when its level is changed at least 1 . 5-fold compared with non-infected flies , we would conclude that wMel did not constitutively prime any of the different immune genes tested in its natural host D . melanogaster . Those results are in accordance with previous data showing no pre-activation of different immune genes in D . melanogaster , D . simulans and A . albopictus by Wolbachia [14] , [15] . In summary , the only immune genes up-regulated by wMelPop-CLA and wMel in A . aegypti are anti-microbial peptides , Toll pathway and melanization genes . However , the same Wolbachia strains did not up-regulate these genes in Drosophila , and yet dengue interference occurs in this host . This indicates that the up-regulation of these immune effector genes is not required to interfere with dengue virus replication , although it is likely that the immune up-regulation that occurs in mosquitoes , presumably due to the recent association with Wolbachia , might enhance this effect .
All the mosquito strains used in this study were laboratory lines of A . aegypti infected with wMel ( MGYP2 ) or wMelPop-CLA ( PGYP1 ) , and their tetracycline-treated uninfected counterparts , MGYP2 . tet and PGYP1 . tet [7] , [8] . Adult mosquitoes were kept on 10% sucrose solution at 25°C and 60% humidity with a 12-h light/dark cycle . Larvae were maintained with fish food pellets ( Tetramin , Tetra ) . The fly experiments were performed with w1118 fly lines stably infected with wMel ( w1118wMel ) [31] and wMelPop ( w1118wMelPop ) [18] compared to the tetracycline-cured lines derived by the addition of tetracycline ( 0 . 3 mg/ml ) to the adult diet for two generations . Those lines were confirmed to be free of Wolbachia by PCR , using primers specific for the wMel and wMelPop IS5 repeat [22] . Females were kept under controlled conditions , low-density ( 30 females per vial ) , at 25°C with 60% relative humidity and a 12-h light/dark cycle . Three replicate pools of 20 female mosquitoes , 8 days post-eclosion were collected from each of the four lines ( PGYP1 , MGYP2 , PGYP1 . tet and MGYP2 . tet ) , snap frozen in liquid nitrogen and extracted for total RNA using Trizol ( Invitrogen ) . RNA was then purified using RNeasy kits ( Qiagen ) according to manufacturer's instructions . Whole-genome microarrays were then used to compare gene expression in the Wolbachia-infected lines relative to uninfected controls , using a dual-color reference design . All sample preparations and hybridizations were then carried out by the IMB Microarray Facility at the University of Queensland . Briefly , sample quality was examined using the Agilent 2100 Bioanalyzer ( Agilent Technologies ) and fluorescent cDNA was synthesized using Agilent Low RNA Input Linear Amplification Kit with Cy3 or Cy5 . Each infected line and respective paired tetracycline-treated line was represented by 3 biological replicates ( 3 pools above ) . A total of 6 hybridizations were then carried out for each biological replicate , 3 labeled with cy3 and three with cy5 ( dye swaps ) . Microarrays were of the 4×44 K format ( Agilent ) each containing standard control features and 3 technical replicates of each 60 mer feature randomly distributed across the layout . The A . aegypti genomic sequence ( Vectorbase genome build 1 . 1 ) was used for construction of oligonucleotide microarrays using eArray Version 5 . 0 ( Agilent Technologies ) . After removing probes that cross hybridized , a total of 12 , 336 transcripts that represented 12 , 270 genes were spotted onto each microarray [32] . For each transcript , raw data was extracted and analyzed using Genespring v . 9 . 0 ( Agilent Technologies ) . An intensity dependent ( Lowess ) normalization ( Per Spot and Per Chip ) was used to correct for non-linear rates of dye incorporation as well as irregularities in the relative fluorescence intensity between the dyes . Hybridizations from each mosquito line were used as replicate data to test for significance of expression changes using the cross-gene error model . The occurrence of false positives was corrected using the q-value [33] , [34] . All array data have been deposited in ArrayExpress ( http://www . ebi . ac . uk/microarray-as/ae/ ) under the accession number E-MEXP-2931 . Functional annotations of A . aegypti genes were retrieved from Biomart [35] in Vectorbase [36] and analyzed using the Ontologizer software with the parent child intersection method [37] , [38] . The over-expression of particular GO categories in the microarray data set was tested against the distribution of GO categories for the A . aegypti genome . Dengue virus serotype 2 ( DENV-2 ) , strains 92T [9] and ET300 were isolated from human serum collected from patients from Townsville , Australia , in 1992 and East Timor in 2000 , respectively . DENV-2 ( strains 92T and ET300 ) was propagated and quantified as described by Frentiu et al [29] . For virus injection , 8 day old D . melanogaster females ( w1118wMel and w1118tet ) and A . aegypti females ( MGYP2 and MGYP2tet ) were briefly anesthetized with CO2 and injected under a dissecting scope into their thorax with a pulled glass capillary and a handheld microinjector ( Nanoject II , Drummond Sci . ) . 69 µl of virus stock ( 107 pfu/ml ) or sterile PBS 1X were injected . After injection flies and mosquitoes were maintained under identical controlled conditions , low-density ( 10 females per vial or cup ) , at 25°C with 60% relative humidity and 12-h light/dark cycle . Insects were collected 8 days post-injection and kept at −80°C for RNA extraction . RNA extraction was done on 15 individual 16 day old females per condition using Trizol ( Invitrogen ) . 1 µg of total RNA was kept to quantify DENV-2 while the rest was used for immune gene expression analysis as described below . Accumulation of genomic ( +RNA ) RNA strands was assessed by quantitative real time PCR using hydrolysis probes specific to the 3′ UTR region of the four dengue serotypes [39] with modifications ( A . T . Pyke , unpublished data ) . The sequences of the primers were FWD: 5′-AAGGACTAGAGGTTAGAGGAGACCC-3′ and RWD: 5′-CGTTCTGTGCCTGGAATGATG-3′ and the sequence of the probe was 5′- AACAGCATATTGACGCTGGGAGAGACCAGA-3′ . 1 µg of total RNA for each sample was mixed with 0 . 625 µM of the reverse primer plus 0 . 2 mM dNTPs . Samples were incubated at 86°C for 15 minutes and 5 minutes on ice , then 5X first strand buffer and 100 U of Superscript III ( Invitrogen ) was added to a total volume of 20 µl . Samples were incubated at 25°C for 10 minutes , followed by 42°C for 50 minutes and 10 minutes at 95°C to inactivate the transcriptase . The qPCR reaction consisted of 2 µl of the synthesized cDNAs , 5 µl of 2X LightCycler 480 Probes Master ( Roche ) , 0 . 5 µM of each primer ( see above ) and 0 . 5 µM of the probe ( see above ) in 10 µl total volume . Reactions were performed in duplicate in a LightCycler 480 Instrument ( Roche ) with the following conditions: 95°C for 5 minutes , and 45 cycles of 95°C for 10 s , 60°C for 15 s , 72°C for 1 s . A standard curve was created by cloning the DENV-2 3′UTR region fragment into pGEM® T-Easy ( Promega ) . After linearization with Pst I the plasmid was serially diluted into known concentrations and run in parallel , in order to determine the absolute number of DENV-2 copies in each 1 µg of total RNA . First , percentages of individuals infected with dengue were calculated for each treatment . Then only individuals with dengue infection ( non zero quantification ) were used to examine the effect of wMel on dengue titer using Mann-Whitney U tests ( Graph Pad Prism 5 ) . RNA extraction from flies was done using between 10 to 15 individual 8 day old females per condition using Trizol reagent ( Invitrogen ) . To eliminate any contamination by DNA , samples were treated with DNase I recombinant ( Roche ) , in accordance with the manufacturer's instructions . cDNAs were synthesized from 1 µg of total RNA , using oligodT primers and the SuperScript III enzyme ( Invitrogen ) , in accordance with manufacturer's instructions . For each sample qRT-PCR was performed in triplicate on a 10 times dilution of the cDNAs using Platinum SYBR Green ( Invitrogen ) according to the manufacturer's protocol . Primers are listed in Table S4 . The temperature profile of the qPCR was 50°C for 2 minutes ( UDG incubation ) , 95°C for 2 minutes , 45 cycles of 95°C for 5 s , 60°C for 5 s , 72°C for 10 s with fluorescence acquisition of 78°C at the end of each cycle , then a melting curve analysis after the final cycle . The housekeeping gene rpS17 was used to normalize expression . Target gene to housekeeping gene ratios were obtained for each biological replicate using Q-Gene software [40] . Raw data were graphed as median ± interquartile range ( IQR ) and outliers beyond 1 . 5 IQR excluded . Treatment effects on expression ratios were then examined using the Mann-Whitney U tests ( Graph Pad Prism 5 ) . The occurrence of false positives was corrected using the q-value [33] , [34] . | Wolbachia pipientis is an inherited intracellular bacterium that is widespread in insects . Because of its ability to interfere with various pathogens such as dengue viruses , nematodes and Plasmodium in insects , it has been proposed as a possible tool to control insect-transmitted disease . Recently , two strains of Wolbachia that interfere with RNA viruses in their natural host , Drosophila melanogaster , were introduced into the naturally uninfected mosquito vector of dengue fever , Aedes aegypti . As in their natural host , those two strains block the replication and the dissemination of viruses in the mosquito . Some studies suggest that pathogen blocking is due to Wolbachia priming the insect innate immune system . Here , we show that Wolbachia induces transcription of some immunity related genes only in its new host A . aegypti , and not in its natural host D . melanogaster , while Wolbachia reduces dengue replication in both hosts . These results suggest that immune priming by Wolbachia might not be the only mechanism responsible for viral interference . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology"
] | 2012 | The Relative Importance of Innate Immune Priming in Wolbachia-Mediated Dengue Interference |
State transitions allow for the balancing of the light excitation energy between photosystem I and photosystem II and for optimal photosynthetic activity when photosynthetic organisms are subjected to changing light conditions . This process is regulated by the redox state of the plastoquinone pool through the Stt7/STN7 protein kinase required for phosphorylation of the light-harvesting complex LHCII and for the reversible displacement of the mobile LHCII between the photosystems . We show that Stt7 is associated with photosynthetic complexes including LHCII , photosystem I , and the cytochrome b6f complex . Our data reveal that Stt7 acts in catalytic amounts . We also provide evidence that Stt7 contains a transmembrane region that separates its catalytic kinase domain on the stromal side from its N-terminal end in the thylakoid lumen with two conserved Cys that are critical for its activity and state transitions . On the basis of these data , we propose that the activity of Stt7 is regulated through its transmembrane domain and that a disulfide bond between the two lumen Cys is essential for its activity . The high-light–induced reduction of this bond may occur through a transthylakoid thiol–reducing pathway driven by the ferredoxin-thioredoxin system which is also required for cytochrome b6f assembly and heme biogenesis .
Photosynthetic organisms are constantly subjected to changes in light conditions . These organisms have developed different mechanisms to rapidly acclimate to this changing environment . At one extreme , when the absorbed light excitation energy vastly exceeds the assimilation capacity of the photosynthetic apparatus , these organisms need to protect themselves . Excess light energy is dissipated as heat through nonphotochemical quenching , which involves conformational changes in the light-harvesting system of photosystem II [1] . In contrast , under low light , photosynthetic organisms optimize the absorption capacity of their antenna systems . This is especially true when changes in light quality occur that lead to the preferential stimulation of either photosystem II ( PSII ) or photosystem I ( PSI ) , which are linked through the photosynthetic electron transport chain . Under these conditions , balancing of the light excitation energy between the antenna systems of PSII and PSI occurs through a process called state transitions [2–4] . Upon preferential excitation of PSII , the plastoquinone pool is reduced , a process that favors binding of plastoquinol to the Qo site of the cytochrome b6f complex and leads to the activation of a thylakoid protein kinase required for the phosphorylation of the light-harvesting system of PSII ( LHCII ) [5 , 6] . In the green alga Chlamydomonas reinhardtii , the LHCII protein set consists of Type I ( Lhcbm3 , Lhcbm4 , Lhcbm6 , Lhcbm8 , and Lhcbm9 ) , Type II ( Lhcbm5 ) , Type III ( Lhcbm2 and Lhcbm7 ) , and Type IV ( Lhcbm1 ) proteins , and of CP26 and CP29 [7] . Because of their nearly identical sequences and size , several of these Lhcbm proteins cannot be distinguished by SDS-PAGE . Most of them fractionate into two major bands called P11/P13 ( Type I ) and P17 ( Type III ) . CP29 , Lhcbm5 , P11 , P13 , and P17 are phosphorylated during a state 1 to state 2 transition [7–9] . Although CP29 and Lhcbm5 are mobile during state transitions , it is not yet clear which among the other LHCII proteins of C . reinhardtii are mobile [10 , 11] . The phosphorylation of LHCII is followed by a displacement of LHCII from PSII to PSI , thus increasing the size of the PSI antenna at the expense of the PSII antenna and rebalancing the excitation energy between both photosystems . Binding of the mobile LHCII to PSI requires the PsaH subunit [12] . This state corresponds to state 2 . The process is reversible as preferential excitation of PSI leads to the dephosphorylation of LHCII by unknown phosphatases and its return to PSII ( state 1 ) . State transitions can be induced , not only by changes in light conditions , but also through changes in cellular metabolism . Thus , in C . reinhardtii , the process can be triggered when the level of ATP is low or when the cells are grown in the dark under anaerobiosis . These conditions lead to the influx of reducing equivalents into the plastoquinone pool and to the activation of the LHCII kinase [13] . Moreover , transition from state 1 to state 2 in C . reinhardtii is associated with a switch from linear to cyclic electron transfer [14 , 15] . In this alga , the major role of state transitions appears to be ATP homeostasis . In land plants , the LHCII kinase can also be activated in the dark by the addition of sugar compounds [16] . Recent studies further confirm that state transitions are not limited to the balancing of excitation energy between the photosystems but that they also play a major role in the adjustment of the light reactions with carbon metabolism [17] . In C . reinhardtii , transition from state 1 to state 2 causes the displacement of 80% of LHCII from PSII to PSI , as deduced from the measurements of the quantum yield of PSI and PSII charge separation [18] . In contrast in land plants , only 15% of LHCII is mobile during state transitions , although this process is associated with considerable structural rearrangements of the thylakoid membranes [19] . The large size of the mobile LHCII antenna leads to significant changes in fluorescence yield during state transitions in C . reinhardtii , a feature that has been exploited for the screening of mutants deficient in this process [9 , 20] . Such a screen has revealed the existence of the thylakoid Ser-Thr protein kinase Stt7 ( AA063768 ) [21] . Mutants deficient in this kinase are deficient in LHCII phosphorylation and fail to undergo a transition from state 1 to state 2 . The Stt7 protein kinase is associated with the thylakoid membrane and contains a potential transmembrane domain upstream of the catalytic kinase domain . In Arabidopsis thaliana , the ortholog STN7 ( NP_564946 ) is also specifically involved in LHCII phosphorylation and state transitions [21 , 22] . At this time , it is not yet clear whether Stt7 and STN7 act directly on LHCII or whether they act further upstream as part of a kinase cascade . The cytochrome b6f complex plays a key role in the activation of the kinase [4] . It is thus very likely that Stt7/STN7 interacts directly with this complex . If LHCII is the substrate of Stt7/STN7 , an interaction between the two is expected . Here , we have used coimmunoprecipitations and pull-down experiments to show that interactions of this type do indeed occur . Our data reveal that Stt7 acts in catalytic amounts . We also show that Stt7 contains a transmembrane region with the catalytic domain on the stromal side of the thylakoid membrane and the N-terminal region in the lumen . This domain appears to play a key role in the regulation of the kinase activity .
To understand how the Stt7 kinase functions , we first estimated its abundance , in particular its molar ratio compared with the cytochrome b6f complex under state 2 conditions . An antibody was raised against Stt7 and the amount of Stt7 was estimated using recombinant Stt7 protein for calibration . A similar calibration was performed with the cytochrome b6f complex ( see Figure S1 ) . This analysis revealed that the molar ratio between Stt7 and the cytochrome b6f complex is 1:20 , clearly indicating that Stt7 is present at substoichiometric levels compared to the photosynthetic complexes . The Stt7/STN7 protein kinase has been shown to be associated with the thylakoid membrane [21] . However , it is not known whether Stt7/STN7 acts singly or whether it is associated with other proteins in a larger complex whose composition might change during state transitions . To test this possibility , thylakoid membranes from the stt7 mutant complemented with Stt7 containing a haemagglutinin ( HA ) -tag at its C-terminal end ( Stt7-HA ) were isolated . Membranes were prepared from cells in state 1 obtained by illumination in low light ( 6 μmol m−2 s−1 ) in the presence of DCMU ( 3- ( 3 , 4-dichlorophenyl ) -1 , 1-dimethylurea ) for 30 min and in state 2 by incubating the cells under anaerobic conditions for 30 min in the dark . The occurrence of state transitions was verified by measuring the change in maximum fluorescence ( Fmax ) . The thylakoid membranes were solubilized with n-dodecyl-β-maltoside and fractionated by sucrose density gradient centrifugation . Individual fractions of the two gradients from Stt7-HA thylakoid membranes were separated by PAGE and then tested by immunoblot analysis using antibodies directed against HA , Cytf , PsaA , D1 , CP26 , CP29 , and Lhcbm5 ( Figure 1A ) . Under both state 1 and state 2 conditions , the Stt7 protein kinase was associated with a large complex that partly overlaps with the high molecular weight fractions of PSI and the cytochrome b6f complex but not with PSII ( Figure 1A ) . No major changes in the distribution of the thylakoid complexes were observed under the state 1 and state 2 conditions used . A similar distribution of the complexes was found in the stt7 mutant , confirming that the high molecular weight LHC complexes are not only formed under state 2 , but also under state 1 conditions ( Figure 1B ) . Although the level of Stt7-HA was between 25%–50% compared to Stt7 in wild-type cells ( Figure S2 ) , state transitions proceeded to the same extent as in the wild type ( unpublished data ) . Moreover , the amount of photosynthetic complexes was the same in Stt7-HA , stt7 , and wild-type cells ( Figure S2 ) . As expected , immunoblots with an anti–P-Thr antiserum revealed increased phosphorylation of several proteins in state 2 , notably the major LHCII proteins P11 , P13 , and P17 in the wild-type strain ( Figure 1A ) . Moreover , a weak phosphorylation signal corresponding to Lhcbm5 was detected under state 2 conditions in the same high molecular weight fractions containing PSI . Although an increase of phosphorylation was also observed for the PSII core proteins CP43 and D2 under state 2 conditions , in other experiments , no significant increase of phosphorylation of these proteins was detected between state 1 and state 2 . The immunoblots in Figure 1A indicate that the levels of Stt7 are significantly higher in state 2 than in state 1 . To examine this further , cells containing Stt7-HA were grown for 2 h under state 2 conditions , and growth was continued for 4 h either under state 2 or state 1 conditions ( Figure 2 ) . At different time points , aliquots of cells were processed for immunoblot analysis with HA and Cytf antibodies . Whereas the level of Stt7 remained the same under continuous state 2 conditions , its level decreased gradually 1 h after the shift to state 1 conditions . It decreased 4-fold after 2 h and more than 20-fold after 4 h ( Figure 2C ) . After shifting the cells to state 2 conditions , the level of Stt7 increased 2-fold after 2 h but did not reach its initial value ( unpublished data ) . Addition of cycloheximide did not affect the decline of Stt7 , indicating that no newly synthesized protease is involved in this process ( Figure 2D ) . However , addition of a protease inhibitor mixture to the cells abolished the degradation of Stt7 under state 1 conditions ( Figure 2E ) . The degradation of Stt7-HA under prolonged state 2 conditions monitored by immunoblotting with the HA antiserum could be due to the removal of the HA tag from Stt7 . To test this possibility , we repeated this experiment with the Stt7-HA strain by using both Stt7 and HA antiserum . In both cases , a decline of the Stt7 kinase was confirmed under state 2 conditions ( Figure S3A ) . We checked that this decrease also occurs with untagged Stt7 ( Figure S3B ) . To identify the type of proteases involved , different protease inhibitors were tested under the same conditions as above: ACA ( ε-aminocaproic acid ) ( Sigma ) ( 50 mM ) , AEBSF ( 4- ( 2-aminoethyl ) -benzenesulfonyl fluoride hydrochloride ) ( Roche ) ( 5 mM ) , NEM ( N-ethylmaleimide ) ( Sigma ) ( 10 mM ) , phenylmethanesulfonyl fluoride ( Sigma ) ( 5mM ) , and EDTA ( 50 mM ) . Leupeptin and NEM were also used at different concentrations ( see Figure S4 ) . Samples were taken and analyzed at different time points . Whereas NEM completely prevented the breakdown of Stt7 , the serine protease inhibitors ACA and AEBSF had no effect , and EDTA enhanced this process ( Figure S4A and S4B ) . Other inhibitors of cysteine proteases besides NEM , such as E64 and leupeptin , prevented Stt7 degradation under prolonged state 1 conditions ( Figure S4C , S4D , and S4E ) . Although no convincing proof of chloroplast Cys proteases has been reported , their existence cannot be excluded . A bioinformatic search for Cys proteases in plastids of A . thaliana did not reveal any convincing candidate ( see Text S1 ) . As high-light treatment is known to lead to the inactivation of the LHCII protein kinase [23 , 24] , we tested whether Stt7 was stable under these conditions . Cells adapted to state 2 were subjected to high-light treatment ( 900 μmol m−2 s−1 ) , and the Stt7 levels were determined by immunoblotting at different times . Under these conditions , a steady decrease of Stt7 was observed ( Figure 2F ) , which could be fully prevented by addition of leupeptin ( Figure 2G ) . Under the same light regime , the level of PSII and of PSI were nearly unaffected ( Figure 2F ) . However , measurements of the ratio of variable ( Fv ) over maximal florescence ( Fv/Fmax ) revealed that this ratio decreased from 0 . 7 to 0 . 2 , indicating photodamage to PSII without apparent decrease of D1 protein level . We took advantage of the decrease in Stt7 under prolonged state 1 conditions to test whether the cells are still able to switch from state 1 to state 2 under these conditions . Cells were first grown under state 1 conditions as shown in Figure 2 . After different time periods , cells were collected and assayed for state transitions using the uncoupler FCCP ( carbonyl cyanide p-fluoromethoxyphenylhydrazone ) , a known inducer of transition to state 2 [13] . The fluorescence emission spectra at low temperature and the level of Stt7 were measured under state 1 and state 2 conditions . Figure 2H shows that transition to state 2 occurred readily in these cells collected after 4 h under state 1 conditions , although the amount of Stt7 protein kinase was decreased 20-fold ( Figure 2I ) . This indicates that the Stt7 kinase acts in catalytic amounts . Activation of the LHCII kinase depends critically on the cytochrome b6f complex [4] . Moreover , this kinase is likely to interact with its putative substrate LHCII . To test whether the Stt7 kinase interacts with the cytochrome b6f complex and/or LHCII , thylakoid membranes from the Stt7-HA strain in state 1 or state 2 were solubilized with dodecyl maltoside and immunoprecipitated with HA antiserum . The immunoprecipitates were fractionated by SDS-PAGE and immunoblotted with antibodies against several known thylakoid proteins . Figure 3A shows that the LHCII proteins P13 , P11 , P17 , CP29 , CP26 , and Lhcbm5 were coimmunoprecipitated with the Stt7 kinase . The signal obtained with CP29 was only detectable under state 2 conditions . Signals were also observed with Cytf and Rieske protein as well as with the PSI subunit PsaA . In contrast , no interaction was observed between Stt7 and D1 from PSII and with subunit α of ATP synthase ( Figure 3A ) , indicating that the interactions detected between Stt7-HA and the other photosynthetic proteins are specific . Furthermore , no signal was observed with the untagged wild-type strain ( Figure 3A ) . Reciprocal immunoprecipitations with PsaA , Cytf , and P11 antibodies confirmed the interaction of these proteins with Stt7 ( Figure 3B ) . To further test the interaction of Stt7 with the cytochrome b6f complex , the Stt7-HA strain was transformed with petA containing a His tag at its 3′-end [25] . After testing the homoplasmic state of the transformed strain for petA-His , thylakoid membranes were isolated , solubilized , and the cytochrome b6f complex was purified on a Ni-NTA column . After washing the column , immunoblotting of the eluted fraction revealed that a small portion of Stt7 kinase was associated with the tagged cytochrome b6f complex , but not with the untagged strain ( Figure 4A ) . To determine which of the subunits of the cytochrome b6f complex interacts with Stt7 , a pull-down assay with GST-Stt7 and solubilized purified cytochrome b6f complex was performed . The results ( Figure 4B ) show that the Rieske protein , but not Cytf , could be eluted from the GST-Stt7 column , indicating that Stt7 interacts with the Rieske protein . Although a putative transmembrane domain within Stt7 is predicted by several algorithms [21] , the presence of four Pro residues within this domain raises some questions , and two models need to be considered . In the first , the N-terminal end of Stt7 would be separated from the large catalytic domain of Stt7 by a transmembrane domain . In the second model , Stt7 could be localized entirely on one side of the thylakoid membrane . To distinguish between these possibilities , the Stt7 protein was tagged with FLAG and HA at its N-and C-terminal ends , respectively . It should be noted that FLAG-Stt7-HA is not functional but that it stably accumulates in the thylakoid membranes ( see below ) . Because the presumed substrates of the Stt7 kinase are localized on the stromal side of the thylakoid membrane , it is expected that the kinase domain of Stt7 is also located on the stromal side . This was tested by isolating intact thylakoid membranes from the strains containing Stt7-HA or FLAG-Stt7-HA and by subjecting the thylakoid membranes to mild digestion with protease V8 . The resulting protein extracts were then examined by PAGE and immunoblotting using antibodies directed against HA , FLAG , PsaD , and OEE2 . PsaD is known to be partially exposed to the stromal side , whereas OEE2 is entirely located on the lumenal side of the thylakoid membrane . OEE2 and the main body of PsaD are thus expected to be protected from any external protease . Under conditions ( 75 μg/ml V8 ) in which proteolysis mildly affected the OEE2 protein and led to partial digestion of PsaD , the level of Stt7 was significantly decreased as measured with HA antibodies , confirming that the kinase domain is located on the stromal side of the membrane ( Figure 5A ) . Similar results were obtained with Stt7-HA ( unpublished data ) . In contrast when antibodies against FLAG were used , products of smaller size were detected , indicating that the N-terminal end of Stt7 is localized on the lumenal side of the thylakoid membrane ( Figure 5A ) . Sonication of the thylakoid membranes followed by V8 protease treatment revealed that the levels of protected fragments detected with the FLAG antibodies were significantly reduced as also observed with the lumenal protein OEE2 ( Figure 5A ) . Sonication also led to enhanced degradation of the HA-tag of Stt7 , presumably because the domain of the kinase exposed to the stroma was more accessible to the protease under these conditions and/or the thylakoid membrane was damaged . In contrast , no difference was observed for PsaD with and without sonication . The orientation of Stt7 was further tested with the yeast split-ubiquitin system [26] . The C-terminal fragment of ubiquitin ( Cub ) was fused to either the N- or C-terminal end of Stt7 and expressed in yeast together with the N-terminal end of ubiquitin ( Nub ) fused to the C-terminal end of the endoplasmic reticulum ( ER ) protein Alg5 , which places Nub on the cytoplasmic side of the membrane . Ubiquitin was reconstituted only with the construct in which Cub was fused to the C-terminal end of Stt7 but not when it was fused to the N-terminal end ( Figure 5B ) . Because membrane proteins usually insert with their lumen domain in the periplasmic space of yeast , the two-hybrid results are fully compatible with the topology of Stt7 derived from the protease protection studies . The transmembrane domain of Stt7 near its N-terminal end is preceded by a region that contains two conserved Cys separated by four residues that are also conserved in the orthologous STN7 kinase of Arabidopsis [21] . In land plants , it has been shown that under high light , the LHCII kinase is inactivated through the ferredoxin-thioredoxin system [23 , 24] . One possibility is that these conserved Cys are targets of this redox system . To test the role of these residues , the two Cys were changed individually to Ser or Ala by transforming the stt7 mutant with the Stt7-HA-C68S/A and Stt7-HA-C73S/A constructs . In all four cases , the mutant kinase accumulated as in Stt7-HA cells ( Figure 6A ) . However , the low-temperature fluorescence emission spectra measured under conditions inducing state 1 or state 2 were nearly identical in these mutants , indicating that they are deficient in state transitions ( Figure 6A ) . As expected , an increase of PSI fluorescence at 715 nm , which is characteristic for state 2 , was detected in the rescued Stt7-HA strain . Moreover , the Cys mutants failed to phosphorylate LHCII under state 2 conditions ( Figure 6B ) . Thus , the Cys residues are critical for kinase activity , although they are separated from the catalytic kinase domain by the transmembrane region . The insertion of a FLAG-tag near the N-terminal end of mature Stt7 abolished state transitions and kinase activity but did not affect the stable accumulation of the protein ( Figure 6A and 6B ) . Moreover , the presence of the FLAG tag specifically prevented the coimmunoprecipitation of Stt7 with the Rieske protein , but not with Cytf ( Figure 6C ) . At first view , this appears to contradict the results of the pull-down experiment , which indicate that the Rieske protein , but not Cytf , interacts with Stt7 . In the case of the pull-down experiment , recombinant Stt7 protein was used in which the transmembrane may not be correctly folded . This domain could be responsible for the binding to Cytf . In the case of the immunoprecipitation , solubilized thylakoid membranes were used in which the interaction between Stt7 and Cytf is preserved . The addition of the FLAG epitope appears to prevent proper interaction of Stt7 with the Rieske protein . In contrast , coimmunoprecipitations of Stt7-C68S occurred both with Cytf and the Rieske protein ( Figure 6C ) . Taken together , these results indicate that the N-terminal end of Stt7 plays a crucial role in the activation of its kinase activity and that this domain may be involved in the interaction with the Rieske protein .
State transitions lead to a considerable reorganization of the antenna systems of PSII and PSI in C . reinhardtii . Moreover , they are accompanied by large changes in the PSI-LHCI supercomplex [11 , 27] . Fractionation of solubilized thylakoid membranes by sucrose density gradient centrifugation revealed no major changes in the distribution of Stt7 and the LHCII proteins between PSII and PSI under state 1 and state 2 conditions . Interestingly , Stt7 is associated with a large molecular weight complex which cofractionates with the high molecular weight fractions of the cytochrome b6f complex and PSI , but clearly not with PSII ( Figure 1 ) . These partial cofractionations of Stt7 with the cytochrome b6f complex and PSI are compatible with the coimmunoprecipitation experiments ( Figure 3 ) . The exact composition of these complexes remains to be determined . Our results on the fractionation of the complexes differ slightly from those of Takahashi et al . [11] . Whereas these authors found a clear shift of CP26 , CP29 , and Lhcbm5 towards the PSI region upon a transition from state 1 to state 2 , which they attribute to the formation of a large PSI-LHCII supercomplex , we only detected a small increase in the ratio between high and low molecular weight fractions of the sucrose gradient in state 2 for CP26 and Lhcbm5 ( Figure 1A ) . In the case of CP29 , there was no significant difference in its distribution in state 1 and state 2 . These differences may be due to the fact that state 2 and state 1 were induced through different means in our study and that of Takahashi et al . They used FCCP plus NaF and DCMU plus staurosporine for inducing state 2 and state 1 , respectively , whereas we used anaerobiosis and vigorous aeration in the presence of DCMU . We further confirmed the presence of the LHCII proteins detected in the high molecular weight fractions under state 1 conditions in the stt7 mutant . Although we could not detect major changes in the distribution of the LHC complexes in the sucrose density gradient under state 1 and state 2 conditions , there were changes in the phosphorylation patterns . A phosphorylated form of Lhcbm5 was detectable in the high molecular weight fractions under state 2 , but not state 1 conditions ( Figure 1A ) . Takahashi et al . [11] observed phosphorylated forms of CP29 and Lhcbm5 in this fraction . A phosphorylated form of CP29 in the high molecular weight fraction was also reported by Kargul et al . [10] . The differences between these studies might be partly accounted for by the different specificities of the anti–P-Thr antibodies used . A striking feature is that the level of Stt7 is significantly lower in state 1 than in state 2 . This was further investigated by a state 2–state 1 time course experiment ( Figure 2 ) . Within 2 h after shifting from state 2 to state 1 conditions , the level of Stt7 decreased to one fourth of state 2 levels . The level of Stt7 decreased further to 2%–5% of state 2 levels after 4 h in state 1 , indicating that Stt7 is unstable under prolonged state 1 conditions . Thus , Stt7 is more abundant and stable under conditions where it is active . The decrease of Stt7 under state 1 conditions could be prevented by addition of inhibitors of Cys proteases to intact cells . It is therefore possible that under state 2 conditions , Stt7 is more protected from proteases either because of a posttranslational modification , e . g . , phosphorylation , or of its association with other proteins . Although experimental evidence for the presence of cysteine proteases in chloroplasts is weak , their existence and role in Stt7 turnover cannot be excluded . It is possible that other proteases are involved in this process , such as the FtsH and Deg proteases , which are known to degrade thylakoid membrane proteins [28] . Some Deg proteases are indeed sensitive to NEM [29] . We note that the decrease in Stt7 occurs only after a prolonged period in state 1 , whereas state transition is a short-term acclimation response that occurs within minutes . It is therefore possible that the control of the Stt7 level is part of a long-term acclimation response . There is indeed evidence for a role of the ortholog STN7 of A . thaliana in such a process [30] . The level of Stt7 decreases under high light . Under the same conditions , PSII levels remained constant . In land plants , the LHCII kinase is known to be inactivated through the ferredoxin-thioredoxin system under high light [31] . It is thus conceivable that the kinase is less stable when it is maintained for a prolonged period in its inactive state , induced either by state 1 conditions or high light . Although the redox state of the plastoquinone pool is critical for activation of the kinase , it does not solely determine the level of Stt7 . Under state 1 conditions , the plastoquinone pool is oxidized , whereas it is expected to be reduced under high light . A close interaction between Stt7 and the cytochrome b6f complex is apparent from the coimmunoprecipitation results . This complex is known to play a critical role for the activation of the kinase during a state 1 to state 2 transition [32] . Mutants deficient in cytochrome b6f complex fail to phosphorylate LHCII and are blocked in state 1 [32] . The original state transition model postulates that upon activation of the Stt7 kinase through the cytochrome b6f complex , the kinase is released from the complex to phosphorylate LHCII . However , we find that the association of the Stt7 kinase with the cytochrome b6f complex does not markedly change between state 1 and state 2 . One possibility is that another downstream kinase is phosphorylated by Stt7 , which in turn phosphorylates LHCII . Alternatively , the Stt7 kinase may be part of a large supercomplex that includes the cytochrome b6f complex and the PSII-LHCII complex . We have not been able to detect complexes of this kind , although it is possible that they are only formed transiently . The PetO subunit of the C . reinhardtii b6f complex is known to be phosphorylated during state transitions [33] . Despite several attempts , we were unable to detect any interaction between Stt7 and PetO based on coimmunoprecipitations or yeast two-hybrid screens . It remains to be seen whether the phosphorylation of PetO depends on Stt7 . To identify which subunit of the cytochrome b6f complex interacts with Stt7 , pull-down experiments were performed . The Rieske protein was identified as an interactant ( Figure 4B ) . Based on structural studies of the mitochondrial bc1 and of the chloroplast b6f complexes , electron transfer between plastoquinol at the Qo site and cytochrome f ( Cytf ) is mediated by the Rieske protein which moves from a proximal position when the Qo site is occupied by plastoquinol to a distal position when the Qo site is unoccupied [25 , 34 , 35] . It has been suggested that this dynamic behavior of the Rieske protein could be coupled to the activation of the Stt7 kinase [36–38] . Such a dynamic model is compatible with the low abundance of the Stt7 kinase with a molar ratio of 1:20 relative to the cytochrome b6f complex found in this study . Assuming one cytochrome b6f complex for one PSII core complex and an average of ten LHCII proteins per PSII reaction center [39] , the molar ratio of Stt7 kinase to LHCII protein can be estimated at 1:200 . At first sight , the kinase would have to phosphorylate several LHCII substrates and may undergo multiple rounds of activation . It is possible that the Stt7 kinase acts first on the LHCII located in the edges of the grana and that this phosphorylation induces extensive remodeling of the thylakoid membrane as reported recently during state transitions in A . thaliana [19] . These changes may facilitate the access of Stt7 to LHCII in the grana core . Alternatively , the observed strong LHCII phosphorylation under state 2 conditions could be due to very flexible movements of PSII-LHCII supercomplexes in the grana core and grana margins . An intriguing component of the cytochrome b6f complex is its single chlorophyll a molecule whose chlorine ring lies between helices F and G of the PetD subunit , whereas the phytyl chain protrudes near the Qo site [25] . Interestingly , mutants affected in the binding site of chlorophyll a , besides having reduced cytochrome b6f turnover , also display a decreased rate of transition from state 1 to state 2 [36] . This region may thus either be an interaction site for the N-terminal region of Stt7 or it could act as a sensor for the presence of plastoquinol at the Qo site and initiate a signaling pathway through the chlorophyll a molecule towards the catalytic domain of Stt7 on the stromal side of the thylakoid membrane . The coimmunoprecipitation experiments indicate that Stt7 is associated with LHCII in most cases under both state 1 and state 2 conditions . LHCII could be the direct substrate of Stt7 or , alternatively , Stt7 could be part of a multikinase complex , which ultimately phosphorylates LHCII . The only marked difference in coimmunoprecipitation between state 1 and state 2 was observed for CP29 ( Figure 3 ) . CP29 is particularly interesting . First , this monomeric LHCII together with CP26 and Lhcbm5 has been proposed to act as linker between the LHCII trimers and the dimeric PSII reaction center for the transfer of excitation energy [11 , 39–41] . Second , during transition from state 1 to state 2 , CP29 undergoes hyperphosphorylation: in addition to Thr6 and Thr32 , which are phosphorylated in state 1 , Thr16 and Ser102 are phosphorylated in state 2 [10 , 42] . Third , electron microscopy ( EM ) analysis of PSII-LHCII complexes lacking CP29 could not distinguish between C2S2 and PSII monomeric complexes [39] . Fourth , in maize bundle sheath cells , which carry out mostly cyclic electron flow , the few remaining PSII complexes are monomeric [43] . Taken together , these results raise the possibility that the phosphorylation of CP29 in state 2 may act as a switch for cyclic electron flow with the detachment of CP29 from PSII and its monomerization . The coimmunoprecipitation experiments also reveal an interaction of Stt7 with the PSI complex . This finding is surprising , as one would expect that the kinase acts on LHCII bound to PSII and that it would not interact with PSI . However , movement of PSI-LHCI complexes from the stromal lamellae to the grana margins following LHCII phosphorylation occurs in land plants , and it was proposed that the PSI absorption cross section is increased in this region through interaction of PSI-LHCI with the phosphorylated LHCII originating from the grana [44] . It is possible that the grana margins constitute a platform where the observed interactions of Stt7 with PSI could occur . It is not known whether the kinase is active or inactive when it is bound to PSI , and the role of this association remains to be determined . Analysis of the Stt7 amino acid sequence by bioinformatic means predicts the presence of a single transmembrane domain . However , this putative transmembrane domain contains four Pro residues that may prevent the formation of an α helix . It was therefore important to test the topology of Stt7 by experimental means using Stt7 tagged at its N-terminal end with FLAG and at its C-terminal end with HA . Using intact thylakoid membranes , we showed that whereas the C-terminal end of Stt7 is susceptible to protease digestion , the N-terminal end is protected , indicating that Stt7 indeed contains a transmembrane domain with the kinase domain on the stromal side and the N-terminal end in the lumen . The N-terminal region contains Cys68 and Cys73 , which are conserved in the ortholog of Stt7 in land plants [21] . Among the seven Cys residues in Stt7 , these are the only conserved Cys residues between Stt7 and STN7 . In land plants , the LHCII kinase is inactivated by high-light treatment through the ferredoxin-thioredoxin system [31] . The two conserved Cys could therefore be the targets of this redox system and/or play a major role in the activation of the kinase . By changing either of the two Cys to Ala or Ser , the kinase was inactivated . There was no phosphorylation of LHCII under state 2 conditions , and the mutants were blocked in state 1 . Possibly a disulfide bridge between these two Cys is required for kinase activity , or redox changes of this disulfide bridge are critical for its activity ( Figure 7 ) . The question arises how the redox state of these two Cys in the lumen is regulated through the redox state of the stromal compartment . Recently , at least two components of a transthylakoid thiol–reducing pathway have been identified in chloroplasts . The CcdA thiol disulfide transporter is a polytopic thylakoid protein with two highly conserved Cys in membrane domains , which is able to convey reducing power from the stroma to the lumen [45] . The second component is the Hcf164 thioredoxin-like protein , which acts as a thiol disulfide oxidoreductase [46 , 47] . This system , which is also conserved in bacteria , is thought to be involved in cytochrome c and cytochrome b6f assembly but could also have additional roles . In this respect , given the tight physical association of Stt7 with the cytochrome b6f complex and the requirement of this active complex for the activation of the kinase , it is tempting to propose that the redox state of the Cys68 and Cys73 couple is controlled through the same thioreduction system that operates in cytochrome b6f assembly . These changes in redox state of Stt7 could in turn induce conformational changes of the kinase and affect both its activity and stability .
Chlamydomonas reinhardtii wild-type and mutant cells were grown as described [48] . The stt7 mutant and stt7 complemented with Stt7-HA were used [21] . The HA-tag consists of six copies of the HA peptide YPYDVPDYA inserted at the C-terminal end of Stt7 . In some experiments , the double-tagged FLAG-Stt7-HA was used with FLAG inserted after the Stt7 transit peptide and the HA-tag at the C-terminal end of Stt7 . Strains were maintained on Tris-acetate-phosphate ( TAP ) medium at 25 °C in dim light ( 10 μmol m−2 s−1 ) . The stt7 mutant strain complemented with Stt7-HA was also transformed with the ph6FA1 plasmid [25] in order to obtain a strain expressing Stt7-HA and cytochrome f tagged with a His-tag . Homoplasmicity of this strain was checked by PCR . Stt7-HA consists of six copies of the HA epitope YPYDVPDYA inserted at the C-terminal end of Stt7 . The molecular mass of this HA tag is 8 . 4 kDa and that of Stt7-HA is 85 kDa . Compared with Stt7-HA , FLAG-Stt7-HA contains in addition the FLAG tag RDYKDHDGDYKDHDIDYKDDDDKS with a molecular mass of 3 kDa inserted after the 41 amino acid transit peptide of Stt7 . Cultures were grown in TAP medium to a density of 2 × 106 cells/ml . Cells were subsequently concentrated 10-fold in HSM medium . State 1 was induced by incubating cells in 10−5 M DCMU ( 3- ( 3 , 4-dichlorophenyl ) -1 , 1-dimethylurea ) in dim light ( 10 μmol m−2 s−1 ) under strong aeration , and state 2 was obtained by incubating cells under anaerobic conditions in the dark . Cells were harvested at 6 , 000g for 10 min and resuspended in buffer at a density of 108 cells/ml and broken in a French press at 1 , 200 psi . Thylakoid membranes ( 0 . 8 mg/ml ) were prepared as described [11] and solubilized with 0 . 9% n-dodecyl-β-maltoside for 30 min in ice , then 0 . 5 ml were layered on a sucrose density gradient ( 0 . 1–1 . 3 M sucrose in 5 mM Tricine-NaOH [pH 8 . 0] , 0 . 05% n-dodecyl-β-maltoside ) and centrifuged at 280 , 000g for 16 h in a SW40 Beckman rotor . After centrifugation , the gradient was divided into 18 fractions that were analyzed by SDS/PAGE and immunoblotting . The antibodies used were against HA , FLAG , Cytf , PsaA , D1 , CP26 , CP29 , Lhcbm5 , and P-Thr . Cultures were grown in TAP medium to a density of 2 × 106 cells/ml . After a 10-fold concentration , cells were incubated in HSM medium under anaerobic conditions in the dark for 2 h . Cells were then maintained under state 2 conditions or cultured under dim light under strong aeration ( state 1 conditions ) . In some cases , cycloheximide ( 10 μg/ml ) or protease inhibitor cocktail ( Roche ) ( 2× concentration recommended by the manufacturer ) were added prior to the onset of state 1 conditions . Chlorophyll fluorescence emission spectra were recorded with a Jasco FP-750 spectrofluorimeter using intact cells at a concentration of 106 cells/ml frozen in liquid nitrogen . The excitation light had a wavelength of 435 nm , and emission was detected from 650 to 800 nm . Fv and Fmax measurements at room temperature were performed with a Hansatech PAM fluorimeter . Proteins from thylakoid membranes isolated from cells in either state 1 or state 2 were separated on a 15% SDS-polyacrylamide 6 M urea gel and transferred to nitrocellulose membranes . The membranes were blocked with bovine serum albumin and incubated with rabbit anti-phosphothreonine antibody ( Cell Signaling Technology ) . Thylakoid membranes ( 0 . 8 mg/ml ) were solubilized with n-dodecyl-β-maltoside for 30 min in ice , and nonsolubilized material was removed by centrifugation at 12 , 000g for 10 min at 4 °C . Fifty microliters of anti–HA-affinity matrix was added to 0 . 3 mg of chlorophyll of solubilized membranes and incubated overnight at 4 °C . The beads were washed five times in TBS-BSA ( 100 mM Tris/HCl [pH 7 . 5] , 150 mM NaCl , 0 . 05% BSA ) , and the bound proteins were eluted in 40 μl of 2× SDS loading buffer ( 100 mM Tris/HCl [pH 6 . 8] , 4% SDS , 0 . 2% bromophenol blue , 20% glycerol ) for 30 min at room temperature . The immunoprecipitated proteins were analyzed by immunoblotting with antibodies against subunits of the photosynthetic complexes . The full-length Stt7 cDNA was cloned in the pGEX-4T-1 expression vector ( Amersham Pharmacia Biotech ) . Proteins were expressed in Escherichia coli as GST fusion and purified with the GST fusion system kit ( Amersham Pharmacia Biotech ) . Two micrograms of GST fusion protein were immobilized on glutathione Sepharose 4B beads ( incubation for 2 h at 4 °C ) and mixed with 50 μg of solubilized thylakoid extract in 0 . 5 ml of PBS . After overnight incubation at 4 °C on a rotary shaker , beads were washed four times with the same buffer , resuspended in 30 μl of SDS gel-loading buffer ( 100 mM DTT , 2% SDS ) , and the eluate was fractionated by SDS-PAGE . Thylakoids were washed three time in 50 mM Hepes ( pH 7 . 4 ) , 0 . 3 M sucrose [49] without protease inhibitors and resuspended in NH4HCO3 25 mM . Samples were sonicated in a waterbath sonicator with 1-min sonication and 30-s cooling five times . Endoproteinase GLU-C ( Sigma ) was added at indicated concentrations to thylakoid membranes at a chlorophyll concentration of 0 . 3 mg/ml at room temperature for 15 min . The digestion was arrested by TCA precipitation and samples were resuspended in 1× loading buffer containing 8 M Urea . Further analyses of the samples were performed by SDS-PAGE and immunoblotting . All steps were performed at 4 °C with 1 mM AEBSF ( Roche ) as protease inhibitor in all buffers . After solubilization with 0 . 9% n-dodecyl-β-maltoside , 640 μg of thylakoid membranes ( chlorophyll equivalent ) were incubated 2 h with Ni-NTA matrix ( Qiagen ) in the presence of 200 mM NaCl and 10 mM imidazole . The matrix was washed five times in washing buffer ( TBS , 200 mM NaCl , 20 mM imidazole ) , and the bound proteins were eluted in 2× 1 ml of elution buffer ( TBS , 250 mM NaCl , 300 mM imidazole ) . The eluted proteins were analyzed by immunoblotting with antibodies against subunits of the photosynthetic complexes . The split-ubiquitin experiments were performed using the DUALmembrane kit 3 from Dualsystems Biotech ( http://www . dualsystems . com ) . Cub fused to the artificial transcription factor LexA-VP16 was fused to either the N-terminal end of Stt7 ( minus the chloroplast signal peptide ) or to its C-terminal end ( including STE leader sequence ) according to the manufacturer's protocol using SfiI restriction sites . These constructs were then tested for their ability to release LexA-VP16 when coexpressed with an integral membrane protein ( Alg5 ) fused at its C-terminal to the N-terminal half of ubiquitin ( Nub ) . If both Cub and Nub are located in the cytoplasm , spontaneous reassociation will occur , and ubiquitin-specific proteases will be recruited , releasing LexA-VP16 , which will in turn activate the reporter genes ( Ade and His ) . | To grow optimally , photosynthetic organisms need to constantly adjust to changing light conditions . One of these adjustments , called state transitions , allows light energy to be redistributed between the two photosynthetic reaction center complexes in a cell's chloroplasts . These complexes act in concert with other components of the photosynthetic machinery to turn light energy into cellular energy . A key component in the regulation of state transitions is the chloroplast protein Stt7 ( also known as STN7 ) , which can modify other proteins by adding a phosphate group . When light levels change , the oxidation level of a pool of another chloroplast component , plastoquinone , changes , which in turn activates Stt7 , inducing it to phosphorylate specific proteins of the light-harvesting complex of one reaction center . As a result , a portion of this light-harvesting complex is transferred from one photosynthetic reaction center to the other , thereby optimizing photosynthetic efficiency . Here , we have addressed the configuration of Stt7 within the thylakoid membrane of the chloroplast and the molecular mechanisms underlying its activation . Our data reveal that the level of Stt7 protein changes drastically under specific environmental conditions , that the protein does not need to be present in a one-to-one ratio with its targets for activity , and that it associates directly with a number of components of the photosynthetic machinery . The protein-modifying domain of Stt7 is exposed to the outer side of the thylakoid membrane , whereas the domain critical for regulation of its activity lies on the inner side of the thylakoid membrane . These results shed light on the molecular mechanisms that allow photosynthetic organisms to adjust to fluctuations in light levels . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"biology"
] | 2009 | Analysis of the Chloroplast Protein Kinase Stt7 during State Transitions |
The corneal micropocket angiogenesis assay is an experimental protocol for studying vessel network formation , or neovascularization , in vivo . The assay is attractive due to the ease with which the developing vessel network can be observed in the same animal over time . Measurements from the assay have been used in combination with mathematical modeling to gain insights into the mechanisms of angiogenesis . While previous modeling studies have adopted planar domains to represent the assay , the hemispherical shape of the cornea and asymmetric positioning of the angiogenic source can be seen to affect vascular patterning in experimental images . As such , we aim to better understand: i ) how the geometry of the assay influences vessel network formation and ii ) how to relate observations from planar domains to those in the hemispherical cornea . To do so , we develop a three-dimensional , off-lattice mathematical model of neovascularization in the cornea , using a spatially resolved representation of the assay for the first time . Relative to the detailed model , we predict that the adoption of planar geometries has a noticeable impact on vascular patterning , leading to increased vessel ‘merging’ , or anastomosis , in particular when circular geometries are adopted . Significant differences in the dynamics of diffusible aniogenesis simulators are also predicted between different domains . In terms of comparing predictions across domains , the ‘distance of the vascular front to the limbus’ metric is found to have low sensitivity to domain choice , while metrics such as densities of tip cells and vessels and ‘vascularized fraction’ are sensitive to domain choice . Given the widespread adoption and attractive simplicity of planar tissue domains , both in silico and in vitro , the differences identified in the present study should prove useful in relating the results of previous and future theoretical studies of neovascularization to in vivo observations in the cornea .
Neovascularization , or new blood vessel formation , is an important process in development , wound healing , cancer and other diseases . The corneal micropocket angiogenesis assay , shown in Fig 1 , is widely used for studying neovascularization in vivo [1–3] . The assay involves the implantation of a pellet containing pro-angiogenic compounds into the cornea of a small rodent and observation of the resulting vessel formation over time . Although the cornea is normally avascular , the assay remains popular due to the relative ease with which neovascularization can be observed in the same animal at multiple time-points . In rodents , the cornea is hemispherical , with a thickness on the order of 10 to 20 vessel diameters [4] . The pellet is not usually located on the pole of the cornea; it is shifted slightly toward the base or limbus . It is evident from experimental images that the geometrical configuration of the cornea-pellet system influences neovascularization patterns [1] . As shown schematically in Fig 1C , new vessel formation is often focused in the region where the distance between the pellet and limbus is smallest , with vessels tending to grow toward the pellet . The simplicity of the micropocket assay has made it an attractive candidate for comparing vessel network formation with mathematical ( in silico ) and in vitro models . As reviewed in Jackson and Zheng [6] , many mathematical models of neovascularization have been motivated by the micropocket assay , providing valuable insights into the process [5 , 7–14] . To date , these models have exclusively adopted either one-dimensional ( 1D ) or two-dimensional ( 2D ) representations of the cornea-pellet system , with the former allowing efficient , continuum modeling of the developing vessel network by the solution of systems of partial differential equations ( PDEs ) [5 , 10 , 11] . While most modeling studies are based on qualitative analyses of the assay , some have performed more direct , and even quantitative , comparisons with experimental observations . For example , in a series of studies , Tong and Yuan [3 , 7 , 13] developed a model of the assay using a 2D circular domain , as shown in Fig 1E , based on earlier discrete modeling approaches by Stokes and Lauffenburger [9] . The authors compared predicted patterns of vascularization with their own experimental observations , using a range of metrics such as vessel length , migration distance and projected width of the vascularized region . The authors used their theoretical model to better understand the interplay between diffusible growth factors , growth factor binding to endothelial cells and endothelial cell density , based on observations of vascularization as pellet loading was increased . Harrington et al . [14] used a similar modeling approach to study inhibitor loading and positioning in the cornea , with qualitative comparisons of vascular patterning with experiment . Jackson and Zheng [6] developed a detailed , discrete , model of endothelial cell proliferation and migration in a 2D circular domain . The authors performed qualitative and quantitative comparisons of vascular patterning with experimental results from Sholley et al . [2] . More recently , Vilanova et al . [15] developed a phase-field model of individual vessels and simulated the assay in a 2D circular domain as an element of a more detailed study . The authors performed qualitative comparisons of vascular patterning and front velocity with the previous studies of Tong and Yuan [3 , 7] . In terms of continuum models , Connor et al . [5] used a classical 1D modeling approach to perform detailed quantitative comparisons of predicted vessel densities with their own experimental measurements . Given: i ) the widespread use of 1D and 2D models of the assay , ii ) the use of both qualitative and quantitative comparisons between predicted patterning and experiment , and iii ) the observation that the geometrical configuration of the cornea-pellet system influences neovascularization patterns in experimental images , it is important to understand how the three-dimensional ( 3D ) geometry of the assay affects vessel network formation relative to the planar tissue domains typically used in mathematical models . While there are many 3D mathematical models of sprouting angiogenesis [16–20] , none have focused on the particular geometry of the cornea-pellet system , making it difficult to predict the influence of the assay geometry without a dedicated study . Such a study brings the additional challenges of needing to use a relatively large simulation domain and accounting for the interaction of vessels with curved tissue boundaries at the epithelial and endothelial surfaces of the cornea . In the present study , we develop a discrete , 3D , off-lattice mathematical model of neovascularization in the cornea-pellet system , focusing on emulating the in vivo configuration . We use a simplified treatment of the underlying biology , focusing instead on how the adoption of different geometries , including planar 2D and 3D cultures in rectangular or circular configurations , affects vessel network formation relative to the in vivo case . The primary strengths of the study are: i ) the simulation of neovascularization in large , 3D tissue domains and ii ) the ability to compare predictions across several different geometries and with different biophysical processes activated and de-activated . As part of this comparison we also identify metrics of neovascularization with high and low sensitivity to the choice of tissue domain . As a result , we can predict which in vitro and in silico tissue domains most closely resemble the conditions of the in vivo experiment and which metrics of neovascularization are most suitable for performing comparisons . Given the widespread adoption and attractive simplicity of planar tissue domains , the differences identified here should prove useful in relating and translating the results of previous and future in silico and in vitro studies of neovascularization to in vivo observations in the cornea .
Fig 2 shows 3D renderings of each of the studied domains , along with the naming convention used when presenting results . The pellet radius rp = 200 μm is based on data from Connor et al . [5] , but reduced from their value of 300 μm to facilitate placement in the cornea . In 3D simulations , pellets are assumed to have a thickness of Tp = 40 μm and are situated mid-way between the epithelial and endothelial sides of the cornea . In 2D simulations , the cornea thickness is neglected , while in 3D a value of T = 100 μm is used [4] . The cornea radius is fixed at R = 1300 μm , which is a suitable value for mouse [4] . The ‘Hemisphere’ geometry is formed by a 360° revolution of a circular arc of radius R and angle 90° about the polar axis , followed by an extrusion through a distance T along the inward normal to the revolved surface , giving a 3D volume . The cylindrical pellet is placed inside this volume and is completely enclosed by it . For the Hemisphere , the pellet height h above the limbus is the distance as projected into 2D , as would be typically measured in experimental images , rather than the distance along the geodesic from the limbus to the pellet . All simulations begin with a single blood vessel positioned a small height ε = 100 μm above the base of the cornea and mid-way between the epithelial and endothelial sides of the cornea . The vessel occupies the entire width ( or circumference ) of the domain at that position . Vessels are represented as collections of infinitesimally thin , straight-line segments joined at point locations , denoted ‘nodes’ , shown schematically in Fig 3 . Nodes can be connected to one or more segments . They are assigned numerical or boolean attributes , such as ‘Radius’ and ‘Migrating’ respectively as needed . In the present study line segments can be thought of as corresponding to vessel centrelines . Nodes do not necessarily correspond to individual biological cells , rather a constant number of endothelial cells per unit vessel length E L = 1 20 [23] is assumed on each line segment , based on 5 μm radius capillaries . During angiogenesis the vessel network is updated at discrete time intervals Δt = 1 . 0 h following a sequence of migration , sprouting and anastomosis stages . In a single time step the following stages occur in order: tips migrate , ‘nearby’ tips anastomose , new tips form due to sprouting and any remaining ‘nearby’ tips anastomose . A simple , phenomenological model of sprouting angiogenesis is used . The average rate of sprout formation at a node located at x is [21]: P ( x , t ) = P m a x L ¯ s E L c ( x , t ) c ( x , t ) + c 50 ( 1 ) where Pmax is the rate of sprouting per cell , c ( x , t ) is the VEGF concentration at location x and time t , c50 is the VEGF concentration at which the rate of sprouting is half-maximal and L ¯ s is the averaged length of the two line segments joined to the node . Concentrations at sampled locations are calculated by interpolation from nodal values on finite element meshes using linear triangular or tetrahedral shape functions . A simple description of lateral inhibition is used , with P = 0 within a distance 1 E L of a node that has already been selected for sprouting . Simulations are discretized in time using a fixed step of Δt . In each time step a random number z ∈ [0 , 1] is chosen from a uniform distribution at each node and a sprout forms if z < PΔt . A different random number is generated at each node . Sprouts form in the network by creating a new node at the sprout location and offsetting it by the tip speed s times the time increment Δt in a random direction , normal to the parent line segment . A new line segment is created between the new node and the original sprout location and the new node is marked as ‘Migrating’ . Migrating tips ( nodes marked as ‘Migrating’ ) , illustrated in Fig 3 , are assumed to move at constant speed s = 10 μm h−1 . This speed is chosen so that the average time for the vascular front to reach the pellet is on the order of 4 days , which is consistent with experimental observations [1] . A persistent , off-lattice random walk is used to describe the migration of tips through the extracellular matrix of the stroma [6 , 7 , 14] . The migration direction m is given by: m = m p + χ ∇ c | ∇ c | 1 + χ , ( 2 ) where χ is a dimensionless weighting parameter controlling chemotactic sensitivity , mp is a unit vector in the persistence direction and ∇c is the gradient of the VEGF concentration , calculated using a centered difference approach from nodal solutions on a finite element mesh and sampled at the tip . The random persistence direction mp is obtained by rotating the unit tangent vector along the vessel τ an angle θtip away from its original direction in an arbitrary plane , as shown in Fig 3 , following similar approaches to account for extracellular matrix interactions in [3 , 14] . The angle θtip is chosen from a normal distribution with zero mean and standard deviation σ . The approach for modeling chemotaxis is similarly based on those of previous studies [9 , 13] . An important distinction from previous studies is that the finite extents of the cornea are accounted for: migrating tips are not permitted to leave the simulation domain . A ‘soft contact’ model is adopted so that tips approaching the boundary of the domain are gradually deflected along the tangent to the bounding surface . Biologically this represents tip cells failing to penetrate the stiffer tissue present on the epithelial and endothelial cornea surfaces , while ‘soft’ contact is chosen for more robust numerics . The strength of the repulsion from the boundary increases as the boundary is approached according to: ϕ ( x ) = ϕ m a x d c r i t - d ( x ) d c r i t + d ( x ) ( 3 ) where d ( x ) is the minimum distance to the domain boundaries , dcrit is the distance to the boundary at which repulsion is activated and ϕmax is the dimensionless maximum repulsion strength . The repulsion biases the motion along the tangent to the bounding surface according to: r = ϕ ∇ d ^ ( 4 ) where the · ^ operator produces a random unit tangent . The position of migrating nodes which have not just sprouted is updated from x ( t ) to x ( t + Δt ) at each time step: x ( t + Δ t ) = x ( t ) + s m + r | m + r | Δ t . ( 5 ) where s is the tip velocity and Δt is the time step size in hours . Fixed values of the repulsion strength ϕmax = 5 and critical repulsion distance dcrit = 25 μm are used for all simulations . These values are chosen to ensure a gradual deflection away from the surface , without overly influencing the migration of tips that are far away from the boundaries . An illustrative example application of the boundary repulsion model on a contrived circular domain is shown in Fig 3B . Endothelial tip cells are known to find and merge with other endothelial tips and immature blood vessels during migration , in a process known as anastomosis [24] . There is still uncertainty about the mechanisms by which they meet , but mechanical and chemical guidance are known to contribute [24] . When moving from 2D to 3D models of sprouting angiogenesis it is necessary to define a region within which tips will merge with vessels and each other to allow for the identification of intersections . In this study , a relatively small radius of rana = 5 μm is used , which is on the order of the vessel radius . During simulations , sprouting and migration events occur during discrete time intervals Δt . Anastomosis is implemented by identifying the nearest line segment to nodes marked as ‘Migrating’ after each migration or sprouting event . If the distance from the node to the line segment ( point to line distance ) is less than the radius rana an anastomosis event occurs . An anastomosis event can be either a ‘tip-to-tip’ interaction , in which the migrating node is moved to be coincident with its neighbor and both are de-activated , or a ‘tip-to-vessel’ interaction in which the migrating node is moved onto the line segment , de-activated and a new branch is formed . Given the small anastomosis radius used in the present study it is assumed that only biological cells in direct physical contact will anastomose , which may be overly restrictive if mechanical guidance ultimately plays a strong role in the process . VEGF dynamics are treated in two different ways in this study . In the first case , the pellet dynamics are ignored and a time independent , spatially varying VEGF concentration field is imposed in the tissue domain . In the second case , the dynamics of VEGF release from a nylon pellet are explicitly modeled . The motivation for the former model is that it allows the effects of cornea geometry on angiogenesis to be observed independently of the pellet representation . For the first case , a constant VEGF gradient between the limbus and the pellet is applied , given by: d c d x = c p h + ε , ( 6 ) where x is a positional coordinate along the geodesic between the limbus and pellet , ε is a small offset from the base , corresponding to the position of the initial vessel , and cp is the VEGF concentration in the pellet , which is assumed to be constant in time in this case . By specifying: c ( ε ) = c p ε h + ε ( 7 ) the same concentration and concentration gradient magnitudes are maintained at the limbus in all representations . Aside from the Hemisphere , the concentration at the base is 0 nM and at height h + ε ( the pellet location ) it is cp . For the second case , the transport and decay of the VEGF in the pellet are considered , meaning that cp can now change over time . The dynamics of VEGF in the pellet are not known in detail , but as per Tong and Yuan [7] a high rate of reversible binding to the nylon pellet constituents is assumed . Under this assumption it is possible to derive the following relationship between the total cp and free cf amounts of VEGF in the pellet [5]: c f = c p θ ( 8 ) where θ ≥ 1 is a dimensionless binding parameter . Free VEGF can decay in the pellet at a rate λp or leak through the cornea-pellet interface , which has an effective permeability κp . It is assumed that the VEGF concentration is spatially uniform within the pellet , which has volume Ωp . Balancing mass leads to the following differential equation describing the time rate of change of VEGF in the pellet: d c p d t = - λ p c f ︸ VEGFDecay - κ p Ω p ∫ ∂ Ω ( c f - c ) d Ω ︸ LeakageThroughInterface , ( 9 ) where the integral is over the pellet surface ∂Ω and c is the concentration of VEGF in the cornea , at the interface . The initial concentration cp ( t = 0 ) can be determined from the implanted VEGF mass m = 300 ng [5] by: c p ( t = 0 ) = 1 1000 m MWVEGF Ω p ( 10 ) where the VEGF molecular weight MWVEGF is 45 kDa or 45 000 g mol−1 and the factor of 1 1000 converts from mol m−3 to M . It is assumed that VEGF diffuses isotropically in the cornea , with a diffusion coefficient D = 2 . 52 × 10−7 m2 h−1 [25 , 26] and decays naturally at a rate λ = 0 . 8 h−1 [27] . It is also assumed to enter perfused vessels ( and be washed away ) and bind to endothelial cells . Combining these processes , we deduce that the dynamics of VEGF in the cornea can be described by the following reaction-diffusion equation: ∂ c ∂ t = D ∇ 2 c ︸ Diffusion - λ c ︸ Decay - 2 π R v ρ κ v ( c - c b ) ︸ LeakageIntoVessels - n k e c c c + c 50 ︸ BindingtoECs . ( 11 ) Here κv = 3 × 10−4 m h−1 [28 , 29] is the permeability of vessels to VEGF , Rv = 5 μm [30 , 31] is the assumed vessel radius , cb = 0 M is the amount of VEGF in the blood , assuming it is quickly removed , and ρ and n are respective vessel line and tip densities . The parameter kec is the rate of VEGF binding per endothelial cell and c50 is the VEGF concentration at which the rate of binding is half maximal . Continuum reconstructions of the vessel line and tip densities are calculated from the discrete network representation by summing the total vessel length ( or number of ‘migrating nodes’ ) per finite element and dividing by the element volume . These quantities are then used in the calculation of source and sink rates on an element-by-element basis in the finite element solution of the PDEs . Although widely used [6 , 16] , this approach can lead to a PDE and angiogenesis model solution dependence on mesh size . In the present study the ratio of element length to vessel diameter is approximately 3 . The extent to which VEGF will pass through the outer cornea layers is not clear , nor whether it will pass through the epithelial layer and into the aqueous humor or the collagen-rich limbus . It is assumed that the rate of such leakage is low , and no flux boundary conditions are imposed on all outer surfaces of the cornea ∂Ωcornea , that is: D n · ∇ c | ∂ Ω cornea = 0 ( 12 ) where n is the inward surface normal . On the cornea-pellet interface the following mass balance is assumed: - D n · ∇ c | ∂ Ω = - κ p ( c f - c ) , ( 13 ) where c f = c p θ is the amount of free VEGF in the pellet . Eqs ( 9 ) – ( 13 ) are solved numerically ( detailed below ) , subject to the initial condition of no VEGF in the cornea . In the 2D geometries , the cornea-pellet interface ∂Ω is a line of length w for the planar case or 2πrp for the circle . In the Planar3D geometries it is a rectangle of height T or Tp , depending on whether a finite sized pellet is assumed , and width w . In the remaining geometries , the interface is the entire outer surface of the spatially resolved pellet . Table 1 summarizes the parameter values adopted in this study . Parameter values with sources denoted as ‘This Study’ are discussed in this section unless previously introduced . A pellet thickness of Tp = 40 μm is used in this study , which is less than the Tp = 60 μm value reported in Connor et al . [5] . This is to facilitate placement of the pellet in the simulated Hemisphere geometry . The chemotactic sensitivity range χ ∈ [0 , 0 . 5] is chosen to cover extreme cases where the resulting vessel network is not directed towards the pellet and highly directed towards it . The range of the deviation in persistence angle σ ∈ [0 , 20] degrees covers cases where straight vessels form , through to cases with vessels with tortuosity similar to that observed in experimental images of the assay . The global time step Δt = 1 h is chosen to give average segment lengths of 10 μm , which leads to a physically realistic vessel tortuosity . The initial offset from the limbus of ε = 100 μm is in agreement with experimental images [5] . The amount of growth factor in implanted pellets is usually known by mass , with a value of 300 ng for VEGF reported in Connor et al . [5] . For our pellet volume of 0 . 0075 mm3 and VEGF molecular weight of 45 kDa , this corresponds to a pellet concentration of approximately cp = 1330 μm , which is adopted for the dynamic VEGF model . For the time-independent VEGF model lower pellet concentrations of cp = 1 to 100 nM are used , which give similar concentrations at the limbus to the dynamic model in the early stages of the simulation . The bound fraction of VEGF in the dynamic model θ = 30 is chosen to give a time of VEGF depletion in the pellet of approximately 4 days . The VEGF PDE is solved using the finite element method with linear basis functions . A simple forward-Euler time-stepping scheme is adopted , with suitable time steps identified by convergence studies . The maximum PDE solution time step is 0 . 05 h and a typical grid side length is 30 μm . PDE solutions are updated to the end of the global time step Δt before solutions are sampled for use in the sprouting and migrations rules . Simulation results are presented in terms of the ‘vessel line density’ ρ ( x , t ) , which is defined as the vessel length per unit volume , and ‘tip density’ n ( x , t ) , which is the number of migrating tips per unit volume . Densities are calculated on structured grids , with values averaged over grid cells that are equidistant from the limbus . To reduce noise caused by the sampling of discrete vessels and tips onto the grids , two Gaussian smoothing passes are applied to ρ and n before further processing [39] . CPU times for 90 simulated hours on a single processor range from 15 seconds for the Planar2D case , with a fixed VEGF field , to 30 minutes for the Hemisphere and dynamic VEGF model . The most computationally expensive elements of the simulation are the PDE solution times and spatial searches for anastomosis events .
Fig 4A shows simulated vessel networks after 85 hours ( 3 . 5 days ) for the case with a fixed VEGF concentration field . Anastomosis is found to be more prevalent in the 2D domains , leading to a reduced number of tips and greater confinement of tips toward the advancing front . In 3D , the presence of multiple vessels through the cornea thickness is evident . In the Circle2D case , there is a tendency for tips to move together as the center is approached , this focusing effect being due to the domain geometry . Fig 4B quantifies the maximum tip and vessel line densities in each domain after 85 hours , with and without anastomosis . Without anastomosis , the circular domains have tip densities higher than the planar domains and Hemisphere by a factor of 1 . 7 due to geometrical effects . Despite the extra vessel length and volume available for sprouting in 3D , line and tip densities are similar to the 2D cases . When anastomosis is active , the tip and line densities in the 2D cases decrease by greater amounts than in 3D . In the planar domain , the tip density decreases by a factor of 5 . 3 for the Planar2D case but only 1 . 8 for the Planar3D case . Similarly , it decreases by a factor of 6 . 4 for the Circle2D case , but only 2 . 1 for the Circle3D . These results show that the 2D domains lead to greater anastomosis , with the highest tendency for anastomosis in the circular domains . This effect becomes increasingly apparent as the initial pellet concentration cp is varied from 1 through to 100 nM ( shown in S1 Fig ) . As shown in Fig 4C and the full tip and line density profiles in S2 Fig , the differences between 2D and 3D domains become more pronounced with time , as the capacity for sprouting in 2D is reduced due to anastomosis and lateral inhibition . The increasing density in the circular domains with time is again due to geometric effects . Fig 5A shows predicted VEGF concentrations in each domain after 1 hour for the dynamic VEGF model . VEGF distributions are quite different across each domain , showing the importance of choosing a suitable representation of the pellet . The Planar2D and Planar3D domains have a relatively high VEGF concentration at the cornea-pellet interface along the entire domain width W . When pellets of finite width are used the region of higher concentration is localized to a line of length w on the interface . The Planar3D case , with finite pellet width , has a noticeably lower VEGF concentration than the 2D case , due to the pellet thickness Tp being smaller than that of the cornea T . In the circular domains , the situation is reversed , with higher VEGF concentrations in the 3D domain due to a greater surface density at the cornea-pellet interface . A higher concentration is observed in the Hemisphere for the same reason . Over time the VEGF in the pellet depletes , with the decay term in Eq ( 9 ) being dominant . This leads to a similar rate of decay across all domains , as shown in Fig 5B . The VEGF has largely decayed after 4 days , with a 95% reduction from the maximum value in the tissue at this time . The variation in VEGF concentrations shown in Fig 5A , combined with the greater tendency for anastomosis in 2D and circular domains shown in Fig 4 , leads to a variety of predicted maximum tip and line densities across the studied domains for the dynamic VEGF model , as shown in Fig 5C . In this case , higher densities are predicted in the planar geometries with extended pellets , while geometries with the finite pellet width have a lower density , more comparable with the Hemisphere , due to focusing of the vascularized region . When the pellet is moved closer to the limbus the general trend is for an increase in the maximum tip density . In relative terms , the tip density increases most in the Hemisphere and Planar3D geometry with finite pellet , both by a factor of 1 . 7 , although in absolute terms the density in the Planar3DFinite geometry is approximately half that of the Hemisphere . In contrast , in the circles , the tip density is reduced by a factor of 1 . 25 as the pellet is moved towards the limbus . This is a geometric effect , due to the breaking of symmetry as the pellet is moved away from the center of the circle . The dynamics of the maximum densities are similar to those shown in Fig 4C , with the rate of density increase being greatest in the circular domains . Fig 6 summarizes the locations of the vascular front across all of the studied domains at 85 hours , with activation and de-activation of various biophysical features of the angiogenesis model . The ‘distance of the vascular front to the limbus’ , d in Fig 1 , is found to be insensitive to domain choice , for the dynamic VEGF model shown in Fig 6B , the variation from the Hemisphere value across all domains for the predicted vascular front location is at most 3 . 7 percent . The lack of sensitivity to domain choice is likely due to: i ) the closest tips to the pellet always being near the line of symmetry of all domains , ii ) the statistical effect of the metric always accounting for the ‘fastest’ moving vessels amongst the population and iii ) the assumption of constant migration speed in the adopted model of tip migration . Increased sensitivity to the domain geometry would be expected if the migration speed depended on VEGF concentration . The location of the maximum tip density and half-maximum tip density are useful additional metrics in cases where they can be measured . As shown in Fig 6C , these metrics are more sensitive to changes in the biophysical mechanisms of network formation than the front location . For example , the tendency for tip cells to be positioned closer to the moving front in 2D domains ( shown in Fig 4A ) is captured in the bottom glyph in Fig 6C . The effect of strong chemotaxis is similar , as shown in the top glyph . When chemotaxis is relatively weak , or the degree of persistence in the random walk is low , the location of the maximum tip density is moved closer to the limbus . These tendencies are captured across all domains using the maximum tip density and half-maximum tip density location metrics , although they are more sensitive to domain choice than the front location . In all cases the front velocity is approximately constant in time , which is in agreement with experimental observations [7 , 40] , and is similar across all domains . Fig 7A shows the different predicted vessel network patterns in a selection of domains as the pellet is moved closer to the limbus for the dynamic VEGF model . The clear differences in vessel network patterning are not well captured by the ‘distance to limbus’ metric in this case . Although the maximum tip and line density metrics used in Fig 5 are useful in the context of modeling , they can be difficult to measure experimentally . This is because endothelial tip cells are not obvious at typical imaging resolution , while line density measurements are subject to potential errors as multiple vessels may overlap through the cornea thickness [7] . In contrast , it is easier to estimate the ‘vascularized fraction’ or volume of the domain with vessels divided by total domain volume directly from images . In the present study this metric is calculated by accumulating the volume of the cells in the structured grid used in the calculation of densities that are occupied by vessels and dividing by the volume of all cells in the grid . As shown in Fig 7B the ‘vascularized fraction’ metric is sensitive to the differences in vascularization between domains shown in Fig 5B , and also captures the trend for increased vascularization when the pellet is moved closer to the limbus in all domains , except for the circles . As such , the metric is predicted to be useful for differentiating neovascularization patterns and translating observations across geometries .
In this study we developed a 3D , off-lattice mathematical model to predict neovascularization in a spatially resolved representation of the corneal micropocket assay for the first time . We used the model to study how: i ) the geometry of the cornea-pellet system in the micropocket assay affects vessel network formation and ii ) which metrics of neovascularization are most sensitive to geometrical differences between typical in silico , in vitro and in vivo tissue domains . We predict that: All raw data and software used in this study are available on the Zenodo public archive at https://doi . org/10 . 5281/zenodo . 995720 . Instructions for reproducing the study figures are included in the archive . Source code is available under a BSD-3 Clause license and other data under a Creative Commons CC-BY-4 license . | Neovascularization , or the formation of new blood vessels , is an important process in development , wound healing and cancer . The corneal micropocket assay is used to better understand the process and , in the case of cancer , how it can be controlled with drug therapies for improved patient outcomes . In the assay , the hemispherical shape of the cornea can influence the way the vessel network forms . This makes it difficult to directly compare results from experiments with the predictions of mathematical models or cell culture experiments , which are typically performed on flat substrates or planar matrices . In this study , we use mathematical modeling to investigate how the hemispherical shape of the cornea affects vessel formation and to identify how sensitive different measurements of neovascularization are to geometry . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"cardiovascular",
"physiology",
"endothelial",
"cells",
"applied",
"mathematics",
"ocular",
"anatomy",
"radii",
"mathematical",
"models",
"geometry",
"finite",
"element",
"analysis",
"epithelial",
"cells",
"angiogenesis",
"developm... | 2018 | The importance of geometry in the corneal micropocket angiogenesis assay |
We present a novel surrogate modeling method that can be used to accelerate the solution of uncertainty quantification ( UQ ) problems arising in nonlinear and non-smooth models of biological systems . In particular , we focus on dynamic flux balance analysis ( DFBA ) models that couple intracellular fluxes , found from the solution of a constrained metabolic network model of the cellular metabolism , to the time-varying nature of the extracellular substrate and product concentrations . DFBA models are generally computationally expensive and present unique challenges to UQ , as they entail dynamic simulations with discrete events that correspond to switches in the active set of the solution of the constrained intracellular model . The proposed non-smooth polynomial chaos expansion ( nsPCE ) method is an extension of traditional PCE that can effectively capture singularities in the DFBA model response due to the occurrence of these discrete events . The key idea in nsPCE is to use a model of the singularity time to partition the parameter space into two elements on which the model response behaves smoothly . Separate PCE models are then fit in both elements using a basis-adaptive sparse regression approach that is known to scale well with respect to the number of uncertain parameters . We demonstrate the effectiveness of nsPCE on a DFBA model of an E . coli monoculture that consists of 1075 reactions and 761 metabolites . We first illustrate how traditional PCE is unable to handle problems of this level of complexity . We demonstrate that over 800-fold savings in computational cost of uncertainty propagation and Bayesian estimation of parameters in the substrate uptake kinetics can be achieved by using the nsPCE surrogates in place of the full DFBA model simulations . We then investigate the scalability of the nsPCE method by utilizing it for global sensitivity analysis and maximum a posteriori estimation in a synthetic metabolic network problem with a larger number of parameters related to both intracellular and extracellular quantities .
The utility of mathematical modeling in biology is on the rise due to computational advancements and the increasing availability of data provided by high-throughput experimental techniques [1] . Flux balance analysis ( FBA ) is widely used for modeling cellular metabolism in a large range of metabolic and biochemical engineering problems [2 , 3] . Given a constrained metabolic network , FBA assumes the intracellular fluxes are regulated by the cell to optimize a predefined cellular objective function ( e . g . , maximizing the biomass growth rate [4] ) subject to mass balances of the intracellular metabolites and other feasibility constraints ( e . g . , bounds on the substrate uptake and product secretion rates ) . However , FBA only identifies metabolic flux distributions at steady-state and , thus , provides no information on metabolite concentrations or the dynamic behavior of the fluxes . A dynamic extension to FBA , commonly referred to as dynamic FBA ( DFBA ) , was originally developed in [5] and has been subsequently applied in several applications [6–9] . In DFBA models , the intracellular fluxes are given by the solution of a FBA model , which is coupled to a set of dynamic equations that describes the time-varying nature of the extracellular substrate and product concentrations as a function of the extracellular environment [10] . The key assumption in DFBA is that the intracellular fluxes equilibrate instantaneously . This “quasi steady-state” assumption is valid as long as the intracellular dynamics are significantly faster than the extracellular dynamics . Generally , the prediction of the behavior of biological systems such as those described by DFBA models can be subject to various sources of uncertainty including unknown model parameters , unknown model structure , and experimental uncertainty such as measurement error [11] . Accurate quantification of these uncertainties , as well as their impact on the quality of model predictions , is vital when applying these models in decision-support or optimization tasks such as parameter estimation or optimal experiment design . The task of uncertainty quantification ( UQ ) can be divided into two major problems: forward uncertainty propagation and inverse uncertainty estimation . The forward problem focuses on propagating all uncertainties through the model to predict the overall uncertainty in the outputs , whereas the inverse problem aims to calibrate the model with experimental data [12–14] . However , the most commonly used UQ methods are intractable for expensive-to-evaluate computational models [15] , which has severely limited their application to DFBA models . An overview of the various challenges in DFBA simulations can be found in [16] . Surrogate modeling techniques are being increasingly adopted to enable complex UQ analyses that would otherwise be impossible . Of the available surrogate modeling approaches , polynomial chaos expansions ( PCEs ) are one of the most commonly used methods for UQ , which have been shown to yield accurate representations of model outputs using limited computational resources in various engineering systems [17–20] as well as biological systems [21–23] . However , an important underlying assumption in PCE is that the model response is a smooth function of the uncertain parameters such that the response can be accurately approximated by a collection of polynomial functions . For non-smooth models , PCE has been shown to either converge very slowly or even fail to converge altogether depending on the type of non-smoothness [24 , 25] . This is a critical challenge in DFBA models because they are known to become singular ( i . e . , lose differentiability ) at certain time points due to the underlying quasi steady-state assumption [10 , 26 , 27] , meaning that even state-of-the-art PCE methods are not directly applicable to DFBA models . In this work , we propose an extension to PCE , referred to as non-smooth PCE ( nsPCE ) , that can adequately capture the non-smooth behavior exhibited by DFBA models . The underlying concept behind the proposed nsPCE framework is that the time of occurrence of any singularity in a DFBA model is a smooth function of the parameters , which can be effectively modeled with a PCE . Thus , for any given time of interest , the PCE model of the singularity time can be used to partition the parameter space into two non-overlapping regions ( or elements ) that represent the collection of parameters for which the singularity has and has not occurred . Separate PCEs can then be constructed over each of these elements , leading to a piecewise polynomial approximation of the overall model response . We adopt a non-intrusive , regression-based approach for PCE construction from a limited number of expensive DFBA simulations . In particular , we take advantage of state-of-the-art sparse regression methods to systematically locate the terms that have the greatest impact on the model response out of a very large candidate set of terms . By exploiting sparsity , we can mitigate the curse-of-dimensionality that can plague traditional PCE , allowing the application of the proposed nsPCE approach to problems with reasonably large number of uncertain parameters . To demonstrate the effectiveness of the nsPCE method , it is applied to accelerate Bayesian estimation of parameters in the substrate uptake kinetic expressions of diauxic growth of a batch monoculture of Escherichia coli on a glucose and xylose mixed media . The metabolic network reconstruction used for E . coli is iJ904 , which is a genome-scale model that contains 1075 reactions and 761 metabolites [28] . Parameter estimation is performed using measurements of the concentrations of extracellular metabolites and biomass that are taken at certain time points throughout the batch . We selected this particular system due to the fact that reported parameter estimates were determined from experimental data using a trial-and-error procedure [8] . This was likely due to the computational complexity of the genome-scale DFBA model in conjunction with the limited data set that may not enable unique estimation of parameters . In addition , we demonstrate how nsPCE can be applied to vastly speedup forward UQ analyses including global sensitivity analysis and estimation of the probability distribution of the model response . To demonstrate the scalability of nsPCE , it is used for maximum a posteriori parameter estimation in a synthetic metabolic network problem with twenty unknown parameters related to quantities in both the intracellular reaction network and the extracellular environment . The codes that implement the proposed nsPCE method for generic DFBA models are provided at the repository [29] .
We focus on modeling a microbial cultivation process using dynamic flux balance analysis ( DFBA ) , in which the bioreactor is viewed as a combination of the fluid medium ( extracellular environment ) and the microorganisms ( intracellular environment ) . Cell walls act as physical boundaries between these two phases , through which certain chemical metabolites are exchanged . The DFBA model can be mathematically formulated as [26] s˙ ( t ) =f ( t , s ( t ) , v ( s ( t ) ) ) , s ( t0 ) =s0 , ( 1 ) with v ( s ( t ) ) being an element of the solution set of the flux balance model v ( s ) ∈argmaxvh ( v , s ) subjectto:Av=0 , vLB ( s ) ≤v≤vUB ( s ) , ( 2 ) where s denotes the state variables describing the extracellular environment ( e . g . , concentrations of substrates , biomass , and products ) with time derivative s˙ and initial conditions s0; v denotes the metabolic fluxes that include both intracellular fluxes and exchange rates; A is the stoichiometric matrix of the metabolic network; and vLB ( s ) and vUB ( s ) are the lower and upper bounds on the fluxes , respectively , which are functions of the extracellular concentrations . The vector function f , specified by the set of mass balances in the extracellular medium , defines the rate of change of each component of s and must be integrated to determine the time evolution of extracellular concentrations . The scalar function h is the cellular objective that is maximized by the cells . Whenever more than one microbial species are present in the culture , then multiple flux balance models of the form ( 2 ) must be incorporated into ( 1 ) [10] . DFBA models can be classified as ordinary differential equations with embedded optimization wherein the lower-level FBA optimization can either be a linear or nonlinear program [30] . A variety of methods have been developed for integrating DFBA models , which are summarized in S1 Text . We focus on the direct approach for integrating DFBA models in this work due to its ability to ensure accurate solutions through the use of error-controlled integration schemes . Another advantage of the direct approach is that a unique solution set to the FBA ( 2 ) can be obtained using lexicographic optimization [10 , 27] , which may help overcome numerical challenges that can occur when using alternative DFBA simulators ( e . g . , see [31 , Chapter 3] ) . Since the direct approach requires continuous monitoring and identification of any active set changes in ( 2 ) , it constitutes a dynamic simulation with discrete events ( i . e . , a hybrid system ) . In the next section , we present the proposed nsPCE method that is capable of directly accounting for the hybrid nature of DFBA models . The PCE method is guaranteed to converge as both the number of model evaluations N and number of terms in the expansion P increase; however , the rate of convergence can be very slow whenever M exhibits any singularities [24] . This is a primary challenge in DFBA models since they can lose differentiability when a switch in the active set of the FBA problem ( 2 ) occurs . Inspired by [25] , we look to take advantage of the following multi-element representation of PCE as it is capable of capturing non-smooth behavior Y=M ( X ) =∑k=1Ne∑α∈NMak , αΨk , α ( X ) ISk ( X ) , ( 11 ) where Ne denotes the number of elements; Sk , ak , α , and Ψk , α denote the local support , coefficient , and orthogonal polynomials in element k , respectively; S=⋃k=1NeSk; and ISk ( X ) are indicator random variables defined by ISk ( X ) ={1ifX∈Sk0otherwise . k=1 , … , Ne ( 12 ) The indicator random variables can be used to define the following conditional random variables Xk=X| ( ISk ( X ) =1 ) with PDF fXk ( xk ) =fX ( xk ) Pr ( ISk ( X ) =1 ) =fX ( xk ) ∫SkfX ( x ) dx . ( 13 ) The local polynomials in ( 11 ) are orthogonal with respect to Xk while the coefficients are similarly defined as in ( 7 ) but now in terms of Xk . This implies that the same strategies discussed above for building the polynomials , estimating the coefficients using regularized least squares , truncating the expansion , and sequentially populating the ED can be utilized locally within each element . The remaining unanswered question is how to design the elements {Sk}k=1Ne to limit the growth in the number of model evaluations since N will scale approximately linearly with Ne . The best decomposition should ensure that the model response behaves smoothly in every element . The proposed nsPCE method decomposes the support into two elements S1 and S2 that denote , respectively , the set of parameters for which the singularity has not and has occurred . This idea is best illustrated through a simple example . Consider the following non-smooth ODE system y˙=−x if y > 0 and y˙=0 otherwise with initial condition y0 > 0 , whose solution is given by y ( t , x ) ={y0−tx , ify0>tx , 0 , otherwise . ( 14 ) This function is not differentiable at the time point ts ( x ) = y0/x , which can be thought of as the “singularity manifold” in the parameter support space , i . e . , ts is the boundary function that separates S1 and S2 . At any given time of interest t , the two elements can be defined in terms of ts ( x ) as follows S1 ( t ) ={x∈S:ts ( x ) >t} , S2 ( t ) ={x∈S:ts ( x ) ≤t} . ( 15 ) Let us briefly analyze the behavior of these elements . The elements are continuous functions of time , meaning that every time of interest t requires a different decomposition . Whenever t is outside of the support of ts ( X ) , then one of these sets is empty and we revert back to traditional PCE that covers the full support S . In light of this , we can easily generalize the idea to the case of multiple ns > 1 sequential singularities as long as the random variables {tsi ( X ) }i=1ns do not have overlapping supports . When multiple non-overlapping singularities are present , we must simply find the support in which t lies and define the two elements using that corresponding boundary function . The case of overlapping supports is more challenging due to the fact that more elements would need to be created based on the intersection of S1 and S2 for all active singularities . For the simple scalar example in ( 14 ) , we can analytically derive the boundary function; however , this is not generally possible in DFBA models . Based on the observation that the singularity boundary depends smoothly on the parameters , we instead propose to construct a sparse PCE model to approximate the boundary in multiple dimensions , i . e . , ts≈t^sPCE . The nsPCE method thus creates a surrogate model with the following structure for any x ∈ S M^nsPCE ( x ) =∑k=12a^k⊤Ψk ( x ) ISk ( x ) ={a^1⊤Ψ1 ( x ) , ifx∈S1 , a^2⊤Ψ2 ( x ) , ifx∈S2 , ( 16 ) where the coefficients a^k are estimated from the sparse regression problem a^k=argmina˜k1Nk‖Yk−Aka˜k‖22+λk‖a˜k‖1 , k∈{1 , 2} ( 17 ) based on the local ED Xk={xk ( 1 ) , … , xk ( Nk ) } and Yk={yk ( 1 ) , … , yk ( Nk ) } in terms of Nk samples . Notice that the full DFBA model must be integrated when constructing t^sPCE . Instead of discarding this information , it can be reused by storing the list of state and time points generated when integrating the DFBA model and then interpolating these points when calculating the model response function . Thus , we can use this approach to initialize the ED X , model response data Y , and singularity time data Ts . Using Ts along with the set definitions in ( 15 ) , we can easily partition X and Y into the required local EDs . The sequential ED strategy is then applied in each element to ensure that the target error is achieved . A flowchart that summarizes the main steps of the nsPCE method is shown in Fig 1 . By evaluating the nsPCE surrogate in ( 16 ) , which is much cheaper to evaluate than the full model M , on a collection of Monte Carlo samples of the parameters , we can directly approximate statistical properties of Y including moments , parametric sensitivities , or even its full distribution . The complete set of Matlab scripts that implement the nsPCE method is available at [29] . All of the modifiable parameters in the algorithm are defined in the “User inputs” section of the main_pce . m script , which automatically executes the steps summarized in Fig 1 . It is important to note that the scripts require the installation of two additional packages that integrate the DFBA model and construct sparse PCE models . The nsPCE scripts are written to be compatible with readily available DFBA and PCE toolboxes to provide flexibility . The simulation of DFBA models can be done with any non-smooth integration code including COBRA [44] , ORCA [45] , and DFBAlab [10] . All files needed by the DFBA integrator should be placed in the dfba_model folder . We opt for DFBAlab in this work due to certain numerical advantages that it exhibits over the available alternatives ( see [27 , 31] for more details ) . The sparse PCE operations are carried out using UQLab [43] , which implements the hybrid LAR method as well as the required calculations to determine the cross-validation error εLOO . The syntax in main_pce . m is heavily based on UQLab . Hence , some modifications to the source code may be needed to perform the same operations with other toolboxes .
This case study is based on a DFBA model of a batch fermentation reactor consisting of an E . coli monoculture , which has been investigated for the production of valuable chemicals such as ethanol . Here , we focus on the initial phase of batch operation of the E . coli fermentation reactor under aerobic growth in a glucose and xylose mixed media [8] . No ethanol production is observed under aerobic conditions ( i . e . , this phase is mainly used to increase the biomass ) , such that the concentration of ethanol can be omitted from the dynamics . This case study is commonly used as a benchmark for comparing DFBA solvers ( see , e . g . , [16 , 27 , 31] ) , as it exhibits stiff dynamics and multiple singularities . The dynamic mass balance equations of the form ( 1 ) for the extracelluar environment can be summarized as follows b˙ ( t ) =μ ( t ) b ( t ) , g˙ ( t ) =−ug ( t ) b ( t ) , z˙ ( t ) =−uz ( t ) b ( t ) , ( 18 ) where b ( t ) , g ( t ) , and z ( t ) denote the biomass , glucose , and xlyose concentrations at time t , respectively . The uptake kinetics for glucose , xylose , and oxygen are given by Michaelis-Menten kinetics ug ( t ) =ug , maxg ( t ) Kg+g ( t ) , uz ( t ) =uz , maxz ( t ) Kz+z ( t ) 11+g ( t ) Kig , uo ( t ) =uo , maxo ( t ) Ko+o ( t ) , ( 19 ) where parameters ug , max , uz , max , uo , max , Kg , Kz , Ko , and Kig correspond to the maximum substrate uptake rates , saturation constants , and inhibition constants . It is assumed that the reactor oxygen concentration , o ( t ) , is controlled and is therefore constant . The growth rate μ ( t ) , on the other hand , is determined from the metabolic network model of wild-type E . coli . The chosen metabolic network reconstruction was iJR904 [28] , which contains 1075 reactions and 761 metabolites . The cells are assumed to maximize growth , implying ( 2 ) is an LP of the form μ ( t ) =minvc⊤v , s . t . Av=0 , vgext=ug ( t ) , vzext=uz ( t ) , voext=uo ( t ) , vLB≤v≤vUB , ( 20 ) where c is a vector of weights that represent the contribution of each flux to biomass formation while vgext , vzext , and voext are , respectively , the exchange fluxes for glucose , xylose , and oxygen ( i . e . , elements of the flux vector v ) . Thus , the metabolic network interacts with the extracellular environment through the exchange fluxes in ( 19 ) . The initial conditions of the batch are assumed to be fixed at 0 . 03 g/L of inoculum , 15 . 5 g/L of glucose , and 8 g/L of xylose; the oxygen concentration is kept constant at 0 . 24 mmol/L; and A , c , vLB , and vUB are specified by the iJR904 model . However , the parameters in the substrate uptake rates ( 19 ) should be fit to experimental data since they cannot be easily predicted from first principles . This problem of identifying the model parameters was partially tackled in [8] , where most of the parameters were fixed according to estimates provided in the literature while uz , max and Kig were adjusted by trial-and-error to match transient measurements of biomass , glucose , and xylose . The reported parameter estimates are given in S1 Table . Since o ( t ) is fixed , uo , max and Ko can be lumped into a single parameter uo . These six parameters are unknown and here are modeled as a random vector whose elements are independent and uniformly distributed around ±10% of the nominal values . We selected this range to reflect a reasonable level of confidence in the reported literature values . In the following , we demonstrate how the proposed nsPCE surrogate modeling method can facilitate UQ tasks that are otherwise computationally intractable with respect to the full DFBA model . All reported computations are performed in MATLAB R2016a on a MacBook Pro with 8 GB of RAM and a 2 . 6 GHz Intel i5 processor . The DFBA model is simulated using DFBAlab with default options for integration and LP optimization tolerances . CPLEX was used as the LP solver and MATLAB ode15s was used as the integrator . This case study is based on a synthetic metabolic network originally introduced in [31 , Chapter 8] . The goal of this case study is to show that the proposed nsPCE method can be applied to problems with a larger number of parameters as well as alternative UQ approaches . The synthetic metabolic network consumes a carbon source C , a nitrogen source N , and an oxygen source O to produce lipids L , ethanol E , biomass X , ATP , and some oxidation product COX . Although used for illustrative purposes , this network is meant to mimic the behavior of living organisms in the sense that: ( i ) E can only be produced in the absence of O , ( ii ) L can only be accumulated in the absence of N , ( iii ) there is a minimum ATP requirement , and ( iv ) the aerobic oxidation of C produces more energy than fermentation of C . The set of reactions can be summarized as vC:Cex→C , vN:Nex→N , vO:Oex→O , vOX:C+O→SATP , OXATP+SOX , OXCOXex , vFerm:4C→SATP , FermATP+Eex+SOX , FermCOXex , vL:4C+SATP , LATP→L , vX:SC , XC+SN , XN+SATP , XATP→X , vATP:ATP→ATPmaintenance , ( 25 ) where the subscript ex denotes extracellular metabolites and all of the reactions are assumed to be unidirectional . The unknown stoichiometric coefficients are denoted by Si , j , where i represents the metabolite name and j represents the reaction name . The dynamic mass balance equations for the extracellular environment are given by X˙ ( t ) =vX ( t ) X ( t ) , X ( 0 ) =X0 , C˙ ( t ) =−vC ( t ) X ( t ) , C ( 0 ) =C0 , N˙ ( t ) =−vN ( t ) X ( t ) , N ( 0 ) =N0 , O˙ ( t ) =−vO ( t ) X ( t ) , O ( 0 ) =O0 , L˙ ( t ) =vL ( t ) X ( t ) , L ( 0 ) =0 , E˙ ( t ) =vFerm ( t ) X ( t ) , E ( 0 ) =0 , COX˙ ( t ) = ( SOX , OXvOX ( t ) +SOX , FermvFerm ( t ) ) X ( t ) , COX ( 0 ) =0 , α˙ ( t ) =γ ( t ) , α ( 0 ) =0 , ( 26 ) where α is a penalty state that remains equal to zero until the state trajectories become infeasible ( e . g . , when all of the metabolites are depleted ) . A detailed discussion on how to determine the instantaneous penalty value γ is provided in [10] , which is automatically computed in DFBAlab . We assume that the uptake kinetics are given by the following expressions vCUB ( s ) =max ( 0 , vmax , CCKC+C11+EKiE ) , vNUB ( s ) =max ( 0 , vmax , NNKN+N ) , vOUB ( s ) =max ( 0 , vmax , OOKO+O ) , ( 27 ) where s = ( X , C , N , O , L , E , COX , α ) is the vector of extracellular species . A hierarchical set of objectives is used in the FBA problem ( 2 ) to ensure that unique reaction fluxes are obtained ( see S3 Table ) . A total of twenty parameters in this DFBA model , appearing in both intracellular and extracellular quantities , are assumed to be uniformly distributed between upper and lower bounds summarized in S4 Table .
The nsPCE method is specifically constructed to take advantage of the hybrid LAR method for sparse regression , which was originally developed in [19] . As such , nsPCE directly inherits the beneficial scalability properties of hybrid LAR that introduces two sources of sparsity into the expansions: ( i ) low-rank truncation that discards basis terms that lead to high-order interaction of the parameters that are irrelevant in most engineering problems and ( ii ) regularized least squares is used to systematically add basis terms that are strongly correlated to the model response . Additionally , the risk of over-fitting the surrogate model to the available data set can be reduced even further by making the approach basis-adaptive , i . e . , separate PCE models are fit for varying maximum degrees and the one with the lowest error is selected . The basis-adaptive hybrid LAR approach has been successfully applied to a wide-variety of problems and has consistently shown the ability to greatly mitigate the curse-of-dimensionality that is inherent in traditional PCE methods ( see , for example , [19 , 59 , 60] ) . To the best of our knowledge , [60] tackled the largest problem to-date , which is a hydrogeological model with 78 parameters ( 68 identified to be sensitive ) that can be accurately represented using a sparse PCE trained using only 2000 model evaluations . Although uncertainty in high-dimensional DFBA models has not been explored in the literature , these promising results and those shown in the synthetic case study give some confidence that nsPCE may be able to scale to the sizes needed to solve these challenging problems . Note that very recent work has shown that sparse PCEs can be applied to ultrahigh-dimensional problems ( on the order of 104 parameters ) by incorporating a dimensionality reduction step before training the surrogate model [61] . It may be possible to use similar approaches to incorporate uncertainty in the complete set of intracellular model parameters into the nsPCE surrogate models . These are interesting and important challenges that deserve further investigation . In this work , the surrogate models are trained using experimental designs ( EDs ) populated with random samples of the parameters . Recent work has demonstrated that the number of ED points needed to achieve a desired accuracy level can be further reduced by maximizing the information content of the sample locations . Multiple approaches have been developed to tackle this challenging problem , including coherence-optimal sampling [62] and numerical “moment-matching” optimization [34 , 37] . The optimal placement of samples in arbitrary domain shapes in a sequential fashion remains largely unexplored in the literature . Additionally , the current implementation of nsPCE involves only two elements; however , it is unclear if the convergence rate can be improved even more by further decomposing these elements . An adaptive approach for decomposing the random parameter space that uses sensitivity information to decide which elements to split was proposed in [25] . A similar concept could be potentially utilized within nsPCE , though the method would likely benefit from the incorporation of more advanced geometries than simple boxes . Many of the difficulties encountered during parameter estimation are related to poor identifiability of model parameters . Performing parameter identifiability tests can help mitigate these difficulties by ensuring the parameter estimation problem is well-posed , which is especially important when dealing with limited experimental data and/or considering a large number of model parameters . It is common to distinguish between structural and practical identifiability . Structural identifiability is a theoretical property of the model structure that depends only on the observation function and the manipulated input function . Since a structurally non-identifiable parameter is independent of the accuracy of available experimental data , it cannot be resolved by a refinement of existing measurements . The only remedy is a qualitatively new measurement or experiment that alters the structure of the mapping between the parameters and the data . In contrast , practical identifiability also takes into account the amount and quality of the measured data , meaning that it can in principle be resolved by improving the quality of the measurements or increasing the number of measured time points . A thorough treatment of these issues in the context of biological models can be found in , e . g . , [63–65] . To the best of our knowledge , structural and practical identifiability analysis has not been demonstrated on DFBA models , which is an interesting area for future work . It is important to note that , although many methods exist for detecting non-identifiable parameters , they often have restrictions on the class of functions so that they are not directly applicable to DFBA models . Although not observed here , sequential Monte Carlo ( SMC ) can suffer from degeneracy wherein fewer and fewer particles retain significant weight . This is especially prevalent in high-dimensional problems including those with a large number of parameters or a large time horizon [66] . In [67] , it is shown that the degeneracy phenomenon occurs unless the sample size is chosen to be exponential in the dimension , which indicates some type of curse-of-dimensionality . This sample degeneracy can be protected against by adding a rejuvenation step that “moves” the resampled particles according to a Markov chain Monte Carlo ( MCMC ) transition kernel [51] . This operation does not change the target distribution , but does reduce impoverishment since identical replicates of a single particle are replaced with new values . The most challenging part of the MCMC step is ensuring that the samples obtained realistically represent the desired distribution . It is known that convergence of the Markov chain fails for posteriors that are not proper , which can happen whenever the prior is improper ( e . g . , uniform density with infinite bounds ) or non-identifiable parameters exist in the model [68] . In these situations , neither the prior assumptions nor the likelihood that represents the experimental data sufficiently constrain the posterior distribution . As such , the convergence properties of SMC and MCMC methods may improve considerably by resolving parameter identifiability issues before running the algorithm [69] . The selection of optimal conditions for conducting experiments ( e . g . , measurement times , initial conditions , and time-varying input profiles ) is important for ensuring maximum information is extracted from the observations , especially when the experiments are expensive and time-consuming to perform . For example , it may be useful to change the feed rate or the measurement times in the considered case study so that the data ensures tight parameter estimates are obtained . Optimal experiment design ( OED ) has been extensively studied in the classical framework wherein the design criteria are defined as some scalar function of the Fisher information matrix ( FIM ) [70 , 71] . More recently , OED has been tackled from a fully Bayesian perspective that replaces the approximated classical design criteria with an expected utility function that is rigorously chosen from a decision-theoretic point of view [72–74] . The nsPCE surrogate models could be used to efficiently evaluate classical or Bayesian design criteria at any fixed experimental condition . However , the parameter space decomposition depends strongly on the experiment , such that separate surrogates need to be constructed for all experiments of interest . This is not a major challenge when only a small number of experiments are considered , but may become intractable for continuous design spaces . Developing efficient procedures for both classical and Bayesian OED in genome-scale DFBA models is an important area for future research . One possible direction is to treat the experiment design variables as parameters when constructing the surrogate model , as suggested in [75] for global PCE . It would be interesting to see how well nsPCE can handle these additional dimensions , since the model responses would likely be highly sensitive to the design variables . | Construction and validation of mathematical models in biological systems involving genome-scale biomolecular networks is a challenging problem . This article presents a novel surrogate modeling method that can accelerate parameter inference from experimental data and the quantification of uncertainty in the predictions of complex dynamic biological models , with a particular emphasis on nonlinear models with non-smooth behavior . The method is applied to infer extracellular kinetic parameters in a batch fermentation reactor consisting of diauxic growth of E . coli on a glucose/xylose mixed media as well as a larger synthetic metabolic network problem . The proposed approach enables rigorous quantification of parameter uncertainty to determine whether or not available data is sufficient for estimation of all unknown model parameters . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"carbohydrate",
"metabolism",
"chemical",
"compounds",
"metabolic",
"networks",
"experimental",
"design",
"carbohydrates",
"glucose",
"metabolism",
"organic",
"compounds",
"glucose",
"simulation",
"and",
"modeling",
"xylose",
"research",
"design",
"mathematics",
"algebra",
... | 2019 | Fast uncertainty quantification for dynamic flux balance analysis using non-smooth polynomial chaos expansions |
DNA viruses , retroviruses and hepadnaviruses , such as hepatitis B virus ( HBV ) , are vulnerable to genetic editing of single stranded DNA by host cell APOBEC3 ( A3 ) cytidine deaminases . At least three A3 genes are up regulated by interferon-α in human hepatocytes while ectopic expression of activation induced deaminase ( AICDA ) , an A3 paralog , has been noted in a variety of chronic inflammatory syndromes including hepatitis C virus infection . Yet virtually all studies of HBV editing have confined themselves to analyses of virions from culture supernatants or serum where the frequency of edited genomes is generally low ( ≤10−2 ) . We decided to look at the nature and frequency of HBV editing in cirrhotic samples taken during removal of a primary hepatocellular carcinoma . Forty-one cirrhotic tissue samples ( 10 alcoholic , 10 HBV+ , 11 HBV+HCV+ and 10 HCV+ ) as well as 4 normal livers were studied . Compared to normal liver , 5/7 APOBEC3 genes were significantly up regulated in the order: HCV±HBV>HBV>alcoholic cirrhosis . A3C and A3D were up regulated for all groups while the interferon inducible A3G was over expressed in virus associated cirrhosis , as was AICDA in ∼50% of these HBV/HCV samples . While AICDA can indeed edit HBV DNA ex vivo , A3G is the dominant deaminase in vivo with up to 35% of HBV genomes being edited . Despite these highly deleterious mutant spectra , a small fraction of genomes survive and contribute to loss of HBeAg antigenemia and possibly HBsAg immune escape . In conclusion , the cytokine storm associated with chronic inflammatory responses to HBV and HCV clearly up regulates a number of A3 genes with A3G clearly being a major restriction factor for HBV . Although the mutant spectrum resulting from A3 editing is highly deleterious , a very small part , notably the lightly edited genomes , might help the virus evolve and even escape immune responses .
The human genome harbours a group of 11 genes encoding cytidine deaminases , the majority having substrate specificity for single stranded DNA ( ssDNA ) [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] . These include the prototypical enzyme APOBEC1 ( A1 ) and activation induced deaminase ( AICDA ) . The large seven gene APOBEC3 cluster spans ∼150kb at ch22q13 . 1 [11] . Two additional genes , APOBEC2 and APOBEC4 , show homology to the above , although no editing activity has been described so far for either . Many of the human APOBEC3 ( A3 ) enzymes can edit the cDNA of numerous retroviruses , retrovirus elements and hepadnaviruses in tissue culture experiments [3] , [4] , [6] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] , [21] . Yet in vivo only the lentiviruses , of which human immunodeficiency virus ( HIV ) is the most notorious , hepatitis B virus ( HBV ) , human papillomaviruses ( HPV ) and TTV genomes have proven to be edited [21] , [22] , [23] , [24] , [25] , [26] , [27] . The outcome of cytidine deamination is oxidation of the C4 amino group yielding uridine - in short cytidine deamination is mutagenic . The degree of editing can be as little as a few bases per kilobase or up to 90% of all cytidine residues , APOBEC3A deamination of hepatitis B virus DNA in tissue culture being a case in point [16] . In virology , mutations are generally related to the plus strand . Hence , so called G→A hypermutants reflect cytidine deamination of ( − ) stand DNA while C→T hypermutants reflect editing of the viral ( + ) strand . Due to the extent of editing , hypermutation can be seen as part of an innate anti-viral response . This theme is echoed by the fact that some APOBEC3 genes are up regulated by interferon-α and –γ in a wide variety of cells including primary human hepatocytes [28] , [29] , [30] , [31] , [32] , [33] , [34] , [35] . While AICDA expression is chiefly expressed in germinal centre centroblasts [36] , ectopic expression of human AICDA has been shown in at least four settings , all involving chronic inflammation – HCV associated chronic hepatitis , Helicobacter pylori associated stomach cancer , human colitis and chronic inflammatory bile duct disease [37] , [38] , [39] , [40] , [41] . Transgenic mice bearing the human AICDA or A1 genes invariably induce cancers , the organ specificity being dependent on the promoter [42] , [43] , [44] . As there is a long historical link between chronic inflammation and cancer [45] , [46] , the ensemble suggested a link between aberrant AICDA expression and , by inference , expression of other human APOBEC genes and editing of the nuclear genome [41] , [44] , [47] . Cancer frequently emerges from a background of cellular dysplasia . For the liver , cirrhosis is seen as a polyclonal proliferative disease , a prodrome that generally precedes the HCC . Yet virtually all APOBEC editing studies of the HBV genome have confined themselves to analyses of virions from culture supernatants or serum where the frequency of edited genomes is generally low ( ≤10−2 , [21] , [25] ) . In view of the above observations , we decided to look at the nature and frequency of HBV editing in cirrhotic liver samples taken during removal of a primary HCC . Given the finding of ectopic AICDA expression in HCV associated chronic hepatitis [39] , [48] , we chose to work with cirrhotic tissue from HBV mono- and HBV plus HCV double infections . It is shown here that while human AICDA can indeed hyperedit HBV genomes in an in vitro setting , editing by AICDA is a rare event in vivo . By contrast , up to 35% of HBV genomes are edited in vivo by several A3 enzymes including A3G indicating that human A3 deaminases represent major restriction factors for HBV replication . Yet through HBV editing , A3 deaminases generate a mutant spectrum upon which selection can occur . It is suggested that APOBEC3 editing may contribute towards loss of HBeAg antigenemia and immune escape .
A succinct description of the 41 cirrhotic samples is given in Table S1 . Ten DNA samples from patients with alcoholic cirrhosis , 10 HBV+ , 11 HBV+HCV+ and 10 from HCV+ patients were analysed . Complementary DNA was made to total RNA extracted in parallel with the DNA samples and hybridized to a custom made TaqMan PCR chip comprising all 11 human cytidine deaminases related genes ( A1 , A2 , A3A–H , A4 and AICDA ) , a number of pro-inflammatory , apoptotic and mismatch repair genes as well as 3 reference genes ( TRIM44 , HMBS and LMF2 ) that have been validated for analyses of liver tissue . Compared to the 4 normal livers , there was significant up regulation of 2–5 A3 genes in the order: HCV±HBV>HBV>alcoholic cirrhosis . A3C and A3D were up regulated for all groups , while the interferon inducible A3G was over expressed in virus associated , as opposed to alcoholic cirrhosis ( Figure 1 , Figure S1 & Table S2 ) . While there was a trend towards increased A3A and A3F expression in all four groups , it never reached statistical significance . AICDA was up regulated in 15/31 HBV and/or HCV samples , indicating that ectopic expression of AICDA is also a feature of HBV liver disease ( Figure 1 inset ) . By contrast expression was undetectable in the normal liver , which is why the data are not normalized as per A3 data ( Figure S1 ) . APOBEC1 was phenomenally expressed in one sample ( #146 ) yet again normalization wasn't possible as it was weakly expressed in just one normal liver sample ( Figure 1 inset ) . APOBEC2 transcripts were relatively absent while APOBEC4 was undetectable throughout . A number of cellular genes associated with inflammation were significantly up regulated , notably FAS , FASLG , BCL2 , IFNγ and LTA ( Table S2 ) . An infectious molecular clone of HBV was transfected into the hepatocyte derived Huh7 cell line along with human AICDA construct and A3G as positive control . HBsAg secretion in the supernatant was used as readout for editing at a macroscopic level . Figure 2A shows that both AICDA and A3G expression reduced HBsAg to levels ≤50% of control . Given that transfection frequencies were ∼30–40% , this suggests that the majority of genomes were edited . At 72 hours , total DNA was recovered from the supernatants and 3DPCR performed on a segment of the X gene [16] , [21] , [49] , [50] . 3DPCR products were recovered as low as 85 . 2°C for the AICDA cotransfection compared to 86 . 6°C for the A3G control ( Figure 2B ) . The 88 . 7°C 3DPCR products were cloned and sequenced . As can be seen from the mutation matrices ( Figure 2C ) , AICDA ( mean cytidine deamination frequency , fc = 42% , range 16–61% ) was just as good as A3G at hyperediting HBV DNA ( fc = 33% , range 7–98% ) . As expected from extant data , AICDA editing of the HBV target was concentrated in GpC and ApC sites [1] , [51] unlike its A3 counterparts ( Figure 2D ) . Combined with previous data we now have a reference set for the HBV X region edited by 8/11 human cytidine deaminases [16] . Four deaminases ( A3G , A3H , A1 and AICDA ) showed polarized editing biases ( Figure 2D ) that can be used as hallmarks for specific deaminase activity in vivo . The singularity of AICDA and A3G editing with respect to other A3 enzymes becomes apparent when plotting the number of edited cytidine residues in selected dinucleotide contexts for single sequences ( Figures 2E ) . With these metrics in hand , 3DPCR was used to identify hyperedited HBV genomes from the cirrhotic samples . First round PCR DNA was performed on ∼0 . 5µg of total DNA using the X gene specific primers . In a second round , 3DPCR was performed at the restricting temperature of 88 . 7°C ( Figure 3A ) . Fifteen of 17 DNA samples ( 88% ) yielded robust amplification at the restrictive temperature . Five DNA samples – 2 HBV and 3 HBV+HCV - indicated by asterisks were cloned and sequenced . Hyperediting of the HBV genome was essentially confined to the minus DNA strand with only a handful of plus strand hypermutants , a finding in keeping with previous reports [21] . The mean cytidine editing frequencies were somewhat higher in the case of HCV co-infections ( fc = 38 . 8% , 41 . 7% and 46 . 9% ) compared to HBV monoinfection ( fc = 23 . 3% and 31 . 6% , Figure 3B ) . The dinucleotide editing context for these hyperedited genomes was remarkably uniform with a strong preference for CpC and an aversion for ApC , which fits rather well with the profile for A3G ( not shown ) . However , it is possible that such averaging could mask the occasional AICDA edited genome . Accordingly , a clonal analysis was used to highlight editing by distinct enzymes ( Figure 3C ) . Very few patient sequences mapped to the area characteristic of AICDA ( Figure 3C . vs . Figure 2E ) , indicating that it is not a major editor in vivo . By contrast , between 57–71% of patient sequences fell within the area covered by A3G ( Figure 3D vs . Figure 2E ) . The remaining sequences mapped to regions where there was considerable overlap between A3 deaminases ( Figure 2E ) . As A3C was significantly up regulated in these liver samples and expressed at greater levels than any other A3 gene ( Figure 1 ) , editing by A3C alone could explain the remainder . In order to determine accurately the overall hyperediting frequencies in vivo , we performed deep sequencing on cloned first round PCR DNA ( 95°C ) . To our surprise for 4/5 samples a sizeable proportion ( 10–35% ) of X gene segments showed unmistakable signs of hyperediting . The frequency distribution of editing is shown in Figure 4A . Some sequences harboured up to 50/58 ( 86% ) edited cytidines ( Figure 4A inset ) , comparable to those recovered from a transfection experiment . Figure 4B shows the frequency distribution of hyperedited genomes from patients recovered by 3DPCR . That the two frequency distributions are distinct reflects selection against lightly edited genomes during 3DPCR , as previously noted [21] . For the 95°C derived sequences , dinucleotide and clonal analysis indicated that the edited bases showed the same hallmarks as those recovered by 3DPCR , that is a penchant for CpC typical of A3G editing with at least one other A3 enzyme explaining the remainder ( Figure 4C–E ) . Hence , Figure 4B represents a subset of a more general distribution typified by Figure 4A . Assuming that the highly edited part of the distribution ( n≥18 , Figure 4B ) is unaffected by suboptimal amplification , then the expected number of edited genomes can be calculated to be 300×79/7 = 3386 ( see legend to Figure 4 ) . In other words ∼12 fold more genomes are edited than suggested by 3DPCR . Yet the actual number is probably higher as the X gene segment analyzed represents only ∼5% of the HBV genome . A genome with a wild type X gene sequence could be edited elsewhere . Hence , the real frequencies of slightly edited genomes are almost certainly higher . A sizeable proportion of neo-synthesized HBV DNA returns to the nucleus to augment the pool of cccDNA , the viral replication template . DNA bearing multiple dU residues might be degraded by the UNG-APE1 pathway or , if copied on the opposite strand , the resulting dU ( − ) ∶dA ( + ) base pair might be corrected to dT∶dA . To explore this issue we performed first round and 3DPCR on sample #326 using Pfu polymerase . Like all archaeal DNA polymerases , Pfu is unable to amplify DNA templates bearing dU [52] . As can be seen , Pfu amplification recovered much less hyperedited DNA than Taq polymerase ( Figure 4F ) . There were enough 3DPCR products from the 88 . 7°C amplification to allow cloning and sequencing . The genomes were exclusively edited on the minus strand ( mean fc = 25% , Figure 4G ) indicating that multiple dU∶dA pairs can be repaired to dT∶dA . The corresponding amplification using Taq ( Figure 4G ) yielded a comparable frequency ( fc = 32% ) . These observations indicate that substantial repair does occur for a small number of hyperedited genomes . The HBV core orf allows translation from the first , suboptimal AUG giving rise to the HBeAg precursor . Serum HBeAg is a marker for high viremia . Initiation from the second , optimal AUG , leads to synthesis of the capsid monomer , HBcAg . Over the course of a chronic infection , G→A mutations arise in the precore region , particularly at residue G1896 and to a lesser extent G1897 resulting in the loss of HBeAg synthesis and HBeAg seroconversion [53] , [54] . These mutations result in a W28Stop change . There is no consensus as to whether this is due to A3 editing [20] , [25] , [55] despite the fact that G→A transitions in a run of four Gs ( 4Cs on the edited minus strand ) are reminiscent of a hot spot for A3 deamination [21] . To explore this issue , a region spanning the precore region was analyzed by PCR and 3DPCR from a HBV/A3G transfection and patient #326 . As can be seen from Figure 5A , G1896 was indeed a hot spot for A3 editing ( 70–80% editing on the minus strand ) both in vivo and for the A3G transfection experiment . Editing of the plus strand was also evident . This led us to analyze the core region for samples #203 , #318 and #372 at 85 . 5°C ( 3DPCR ) and 95°C ( PCR ) . The data are summarized in Figure 5B . A gradient of editing is apparent with G1896 and G1897 apparently hot spots for A3 editing . Focussing on mutations identified at 95°C , which can be considered as relatively high frequency events compared to those recovered by 3DPCR , they overlap nicely with those reported in a number of previous studies ( Figure 5C ) suggesting that A3 editing may well explain their origin [54] , [55] . The major viral surface antigen , HBsAg , is the basis of the highly successful subunit vaccine . Vaccine escape mutants are known , with a glycine to arginine ( G145R ) being the most frequent substitution mapping to the common double loop “a” determinant ( residues 110–149 , Figure 5D ) [56] , [57] . A G145E substitution is also known , as are a few other loop changes . The G145R mutant also emerges among HBV-immunoglobulin ( HBIg ) treated chronic carriers [58] , [59] and goes undetected by standard diagnostic kits [60] . The local dinucleotide context associated with the G145 substitutions ( Figure 5D ) are again reminiscent of A3 editing . DNA spanning the “a” determinant was analyzed using the two-step PCR/3DPCR approach . Not surprisingly 3DPCR recovered hypermutants from all 4 samples tested at a restricting temperature of 85 . 5°C . The G145 codon was not a hotspot in vivo . As expected , the degree of editing was lower for genomes recovered at 95°C . Among sequences harbouring substitutions in codon 145 most were hypermutants with the G145E substitution while the G145R substitution was accompanied by other mutations ( Figure 5E ) .
The transcriptome data shows that numerous A3 genes are up regulated in virus associated cirrhosis , notably A3B , A3C , A3G , A3H and AICDA accompanied by a number of genes frequently up regulated in inflammatory tissues . By contrast , for alcoholic cirrhosis only two A3 genes ( A3C and A3D ) were significantly up regulated . Using the dinucleotide context as a hallmark to identify a posteriori the deaminase in vivo , A3G clearly emerges as a major restriction factor for HBV replication in cirrhosis along with at least one other A3 deaminase , which could be A3C ( Figure 3C , D ) . The degree of editing is every bit as extensive as in co-transfection experiments using the powerful CMV-IE promoter in tissue culture [16] , [21] . The up regulation of AICDA doesn't impact HBV editing in vivo despite the fact that AICDA shows itself to be a potent editor of replicating HBV genomes ( Figure 2 ) . This apparent dichotomy could be explained by the neogenesis of lymphoid follicles harbouring AICDA+ centroblasts , a feature associated with chronic inflammation [61] , [62] , [63] , or circulating AICDA+ B cells as has been described for chronic HCV infection [64] , [65] , [66] . That A3G is the dominant A3 enzyme fits nicely with the fact that of all the A3 genes , A3G is the most strongly up regulated by interferon-α in primary hepatocytes [29] . Other reports show that the gene is also sensitive to induction by interferon–γ [28] , [30] , [34] . Pegylated interferon–α is used to treat a proportion of HBV infected individuals and the present data may explain part of that effect [67] . A recent report made a link between IFN-α treatment and HBeAg seroconversion , although they concluded that the link was tenuous given the low levels of editing in sera [25] . Although they didn't sequence precore DNA they used a 3DPCR approach . As shown here , the technique tends to underestimate levels by ∼10 fold ( Figure 4 ) . Accordingly hyperedited mutant frequencies reported in that study [25] can be revised upwards to ∼2–33% , in excellent agreement with the present findings ( <2–35% ) . Given the strong impact of A3 deaminases on HBV replication in late stage disease , where does the virus replicate , especially as it is not known to encode an IFN or A3 antagonist ? Some simple possibilities could be interferon resistant or A3low hepatocytes . On a background of HBV+HCV double infection , HBV titres are generally lower [68] , [69] , as though HCV infection rendered the liver an even more hostile environment for HBV . It might be that the strong pro-inflammatory responses associated with HCV immune responses induce A3 genes with their detrimental effect on HBV . That the mean HBV cytidine deamination frequency was higher from the double infection compared to the monoinfection is the result expected if this hypothesis is correct ( Figure 3B ) . As there has been debate as to the importance of A3G alleles in HIV disease ( H186R [70] ) , it is possible that polymorphisms in the A3G gene impact the outcome of HBV infection . Certainly the most striking of all A3 polymorphisms , ΔA3B−/− , doesn't impact HBV disease [22] , [71] . Given their ability to hypermutate DNA , A3 enzymes generate complex mutant spectra , the vast majority probably being deleterious . Even so , lightly A3 edited genomes predominated ( Figure 4A ) while a small fraction resists degradation and are repaired to standard DNA as the Pfu/Taq comparison shows ( Figure 4F ) . Thereafter selection will operate on the remaining genomes . It would seem that IFN induced A3 editing may indeed lead to the occasional emergence of variants , of which the precore C1896T and/or C1897T and HBsAg G145R , E mutants are tangible examples . In this respect , there are remarkable parallels between some RNA viruses and IFN-α induction of the dsRNA adenosine deaminase , ADAR-1L . Editing by this enzyme of A1012 in the hepatitis D virus genome is essential to complete replication [72] , [73] . Similarly , a handful of edited adenosine residues allows escape of respiratory syncytial viruses from monoclonal antibodies [74] , [75] . In conclusion , the cytokine storm associated with chronic inflammatory responses to HBV and HCV clearly up regulates a number of A3 genes with A3G clearly being a major restriction factor for HBV . Could this also be a feature of other chronic inflammatory syndromes or even autoimmune diseases ? Although the mutant spectrum resulting from A3 editing is highly deleterious , a very small part , notably the lightly edited genomes might help the virus evolve and even escape immune responses .
Patients were predominantly males , the mean ages being 60 years ( HBV ) , 63 years ( HBV+HCV ) , 64 years ( HCV ) and 67 years ( alcoholic ) . All were negative for HIV . No patient was on interferon therapy in the months prior surgery . The study was approved by an Institutional Human Research Review Board ( RBM 2005–019 ) and by Institut Pasteur . Written informed consent was obtained for each patient . Forty-one cirrhotic samples as well as 4 normal livers were dissected and directly frozen in liquid nitrogen after surgical removal . Healthy samples represent tissue surrounding benign tumours such as angioma or focal nodular hyperplasia . Total RNA extraction was performed by a phenol-based method ( Euromedex , Souffelweyersheim , France ) . DNase treatment was performed on 10 µg of total RNA using a DNA-free kit ( Ambion ) . The RNA concentration and integrity were assessed using 100 ng of each RNA isolate to perform a capillary gel electrophoresis analysis ( RNA 6000 Nano chip kit , Agilent Technologies , Palo Alto , CA ) to establish an RNA integrity index ( RIN ) . All the samples had acceptable quality , 86% with 7 . 0<RIN<10 . 0 and 14% , 4 . 1<RIN<6 . 9 . Reverse transcription of 1 µg RNA was performed in a final volume of 20 µl ( High-Capacity cDNA Archive kit , Applied Biosystems ) . Messenger RNA were quantified by the TaqMan Low Density Array ( TLDA ) technology ( Applied Biosystems , Courtaboeuf , France ) . Pre-designed hydrolysis probe and primer sets for target genes were factory loaded into the 384 wells of TLDA configured to contain duplicates per target gene ( Table S2 ) . Quantitative PCR was performed using cDNA samples corresponding to 400 ng of starting RNA and TaqMan Universal PCR Master Mix ( Applied Biosystems ) for 48 target genes in duplicate . QPCR conditions were one step of 94 . 5°C for 10 min . followed by 40 cycles at 97°C for 30 sec . and 59 . 7°C for 1 min . on a 7900HT Micro Fluidic Card instrument ( Applied Biosystems ) . For data analysis , gene expression values were determined using the calculation of the relative quantitation ( RQ ) of target genes normalized to a calibrator corresponding to 4 normal livers . RQ calculation was performed using the {Delta}{Delta}CT method using the geometric mean of three reference genes ( TRIM44 , HMBS and LMF2 ) . The three references genes were selected among 12 constant genes arising from a previous array analysis of 70 HCC samples and 9 normal livers for which we applied the algorithms described [76] in the geNorm manual available on the web site http://medgen . ugent . be/~jvdesomp/genorm The study was performed according to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments ( MIQE ) and redaction of the manuscript according to the RDML ( Real-Time PCR Data Markup Language ) data standard ( http://www . rdml . org ) . The pCayw plasmid and all APOBEC and AICDA expression plasmids have been previously described as have the cell lines and transfection protocols [16] , [21] . Transfections were performed independently in triplicate on Huh7 cells and supernatant HBsAg was measured every day with the Monalisa HBsAg PLUS Kit ( Bio-Rad ) . 3DPCR [50] was performed on a Eppendorf gradient Mastercycler S programmed to generate a 4–12°C gradient in the denaturation temperature . A fragment of the X region was amplified by employing a nested procedure . The first-round primers were: 5′Xout: 5′CGCAAATATACATCGTATCCAT and 3′Xout: AAGAGTYYTYTTATGTAAGACYTT , where Y is T , C and R is A , G . First PCR , conditions were: 5 min 95°C then ( 30 sec , 95°C; 30 sec , 60°C; 1 min , 72°C ) ×35 . Nested PCR was performed with 1/50 of the first round , primers were: 5′Xin: ATGGCTGCTARGCTGTGCTGCCAA and 3′Xin: AAGTGCACACGGTYYGGCAGAT , amplification conditions were: 5 min ( 82–93°C ) , then ( 1 min , 82–93°C; 30 sec , 60°C; 1 min , 72°C ) ×35 then at 10 min 72°C . The precore amplification was performed as following , first PCR , conditions were: 5 min 95°C then ( 30 sec , 95°C; 30 sec , 55°C; 1 min , 72°C ) ×35 with primers 5′PreCout: GTACTAGGAGGCTGTAGGCATA and 3′PreCout: 5′ AGAGCTGAGGCGGTATCTAGAA . Nested PCR was performed with 1/50 of the first round , conditions were: 5 min ( 82–93°C ) , then ( 1 min , 82–93°C; 30 sec , 55°C; 1 min , 72°C ) ×35 , then 10 min at 72°C and primers were: 5′PreCin: TAAATTGGTCTGCGCACCAGCA and 3′PreCin: GATCTCGTACTGAAGGAAAGAA . Amplification of the HBsAg was performed with a nested procedure , first PCR conditions were: 5 min 95°C then ( 30 sec , 95°C; 30 sec , 60°C; 1 min , 72°C ) ×35 , primers were: 5′HBsout: CGGCGTTTTATCATCTTCCTCTTCAT and 3′HBsout: CATCCATATAACTGAAAGCCAAACAGT . Nested PCR was performed with 1/50 of the first round , conditions were 5 min ( 82–93°C ) , then ( 1 min , 82–93°C; 30 sec , 60°C; 1 min , 72°C ) ×35 then 10 min at 72°C with primers , 5′HBsin: TCTTCATCCTGCTGCTATGCCTCAT and 3′HBsin: AAAGCCCTACGAACCACTGAACAAAT . PCR and 3DPCR products were purified from agarose gels ( Qiaex II kit , Qiagen , France ) and ligated into the TOPO TA cloning vector ( Invitrogen , France ) . The 88 . 7°C 3DPCR products obtained from X gene amplification were chosen as experience shows they provide a wide range of edited HBV genomes . While products analyzed at lower temperatures are more edited , they proved to be more homogeneous . Sequencing was outsourced to Cogenics . All mutations were verified on the chromatogram . | Retroviruses and hepadnaviruses such as hepatitis B virus ( HBV ) are vulnerable to mutation by host cell single stranded DNA cytidine deaminases . The result is hypermutated viral peppered with uracil residues . While there are potentially 11 such human enzymes , the major players belong to the 7 gene APOBEC3 cluster on chromosome 22 , some of which can be activated by anti-viral interferons . We investigated the nature and frequency of HBV editing in 41 cirrhotic samples following surgical removal of primary hepatocellular carcinoma . Numerous APOBEC3 genes were activated in the decreasing order HCV±HBV>HBV>alcoholic cirrhosis . We observed that APOBEC3G was the dominant restricting factor in vivo with up to 35% of HBV edited genomes . Among the HBV mutants generated by APOBEC3 editing , we found a small fraction of lightly APOBEC3G edited genomes that can impact HBV replication in vivo and possibly immune escape . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/viral",
"infections",
"pathology/histopathology",
"virology/viruses",
"and",
"cancer",
"virology"
] | 2010 | Massive APOBEC3 Editing of Hepatitis B Viral DNA in Cirrhosis |
Protein kinase B ( PKB/Akt ) is a pivotal regulator of diverse metabolic , phenotypic , and antiapoptotic cellular controls and has been shown to be a key player in cancer progression . Here , using fluorescent reporters , we shown in cells that , contrary to in vitro analyses , 3-phosphoinositide–dependent protein kinase 1 ( PDK1 ) is complexed to its substrate , PKB . The use of Förster resonance energy transfer detected by both frequency domain and two-photon time domain fluorescence lifetime imaging microscopy has lead to novel in vivo findings . The preactivation complex of PKB and PDK1 is maintained in an inactive state through a PKB intramolecular interaction between its pleckstrin homology ( PH ) and kinase domains , in a “PH-in” conformer . This domain–domain interaction prevents the PKB activation loop from being phosphorylated by PDK1 . The interactive regions for this intramolecular PKB interaction were predicted through molecular modeling and tested through mutagenesis , supporting the derived model . Physiologically , agonist-induced phosphorylation of PKB by PDK1 occurs coincident to plasma membrane recruitment , and we further shown here that this process is associated with a conformational change in PKB at the membrane , producing a “PH-out” conformer and enabling PDK1 access the activation loop . The active , phosphorylated , “PH-out” conformer can dissociate from the membrane and retain this conformation to phosphorylate substrates distal to the membrane . These in vivo studies provide a new model for the mechanism of activation of PKB . This study takes a crucial widely studied regulator ( physiology and pathology ) and addresses the fundamental question of the dynamic in vivo behaviour of PKB with a detailed molecular mechanism . This has important implications not only in extending our understanding of this oncogenic protein kinase but also in opening up distinct opportunities for therapeutic intervention .
A key downstream relay in various growth factors and hormones is the activation of the serine/threonine protein kinase ( PKB/Akt ) . PKB activates a plethora of proteins that are involved in metabolism , proliferation , growth , and survival [1–3] . Several lines of evidence indicate that the PKB pathway is involved in human cancer , and in particular , its overexpression induces malignant transformation and chemoresistance [2 , 4 , 5] . Its activation is thought to proceed through the recruitment of the protein to membranes via interaction of its PH domain with the phosphoinositides produced by phosphoinositide-3-kinase [specifically PtdIns ( 3 , 4 , 5 ) P3 and PtdIns ( 3 , 4 ) P2] [6 , 7] . The lipid-bound PKB is then phosphorylated by 3-phosphoinositide–dependent protein kinase 1 ( PDK1 ) , which is also recruited through its PH domain binding to PtdIns ( 3 , 4 , 5 ) P3 . The PDK1 phosphorylation , critical for activation , occurs at Thr308 in the activation T-loop of PKBα [8] . A second phosphorylation within a C-terminal hydrophobic motif at Ser473 acts in synergy to fully activate the protein kinase . It is believed that this phosphorylation occurs via the mTor:rictor pathway [9] . Our understanding of the mechanism of activation of PKB remains limited as , unlike other AGC kinases that are substrates for PDK1 [10] , direct in vitro or in vivo interaction of PKB with PDK1 has not been observed . The current models of PKB activation only speculate , based on the regulation of other AGC kinases , how PKB may change its conformation to interact with PtdIns at the plasma membrane and thereafter be activated by PDK1 . The functionally independent behaviour of the PH and kinase domains of PKB is well characterised; the structure of both has been solved separately [6 , 7 , 11–13] . In isolation , the PH domain retains intrinsic lipid binding properties and , similarly , the kinase domain retains function [14 , 15] . The kinase domain is well understood structurally with respect to its inactive and active conformers [12 , 13] . However , the more complex interactions between the PH and kinase domains have not succumbed to direct structural analysis and remain to be elucidated . The recent advent of an in vivo probe to monitor PKB conformational changes in conjunction with molecular modeling has permitted analysis of the mechanism by which these two domains interact and how PKB changes conformation in relation to its activation by PDK1 . We have addressed the possible molecular mechanisms involved in the conformational dynamics of PKB and dissected its activation in intact cells . This has important implications not only in extending our understanding of this critical regulator but also in opening distinct opportunities for therapeutic intervention .
It has been assumed that upon activation of various growth factor pathways , PDK1 and PKB uniquely colocalise at the plasma membrane , permitting PDK1 to phosphorylate PKB . To assess in single cells the dynamic relationship between PKBα and PDK1 , Förster resonance energy transfer ( FRET ) was exploited . Two fusion proteins were generated by genetically encoding enhanced green fluorescent protein ( EGFP ) ( donor ) at the N terminus of PDK1 and monomeric RFP ( acceptor ) at the N terminus of PKBα . NIH3T3 cells were transfected with the expression vectors for GFP-PDK1 and RFP-PKB , and the expressed proteins behaviours were followed upon stimulation of the platelet-derived growth factor ( PDGF ) receptor pathway ( note that the levels of expression were the minimum required for data acquisition ) . For optimum spatial resolution , FRET was monitored by two-photon fluorescence lifetime imaging microscopy ( FLIM ) . Resonance energy transfer was detected by the decrease in the donor lifetime in the presence of the acceptor . The variations in lifetime are presented by lifetime distributions curves and mean FRET efficiency . Coexpression of RFP-PKB with GFP-PDK1 in NIH3T3 cells ( Figure 1A ) induced a decrease in the lifetime of GFP-PDK1 . The decrease in the donor lifetime is quantitatively presented by the lifetime distribution curves . The blue curve is the GFP-PDK1 when expressed alone , and the green curve is when it is expressed with RFP-PKB ( acceptor ) . From this reduction in lifetime , we deduced that , at steady state , GFP-PDK1 and RFP-PKB interact in the cytoplasm . The cytoplasmic mean FRET efficiency presented as box and whiskers plots is 2 . 5 ( Figure 1C ) . Upon PDGF stimulation , both PKB and PDK1 translocated to the plasma membrane , and a further decrease in the average lifetime was detected ( orange curve ) . The translocation of RFP-PKB to the plasma membrane is clearly shown in Figure S6 . The separation of the lifetime distributions at the plasma membrane versus the cytoplasm ( red curve: plasma membrane; light green curve: cytoplasm ) showed that at the plasma membrane , the mean FRET efficiency is 7 . 1 , and in the cytoplasm , it is 4 . 9 ( Figure 1C ) . These results indicated that PDK1 and PKB interacted under basal conditions and that the stimulation of the PDGF receptor pathway promoted an enrichment of the PKB–PDK1 complex at the plasma membrane . This intermolecular interaction was also demonstrated by frequency domain FLIM ( Figure S1 ) . This behaviour indicates that in the cytoplasm , the PKB–PDK1 complex is at equilibrium determined by the bulk phase concentrations . Upon stimulation , the local concentration of both proteins is enhanced at the plasma membrane , whereby the equilibrium is shifted toward the formation of the PKB–PDK1 complex . To verify that the PKB–PDK1 interaction was due to docking and not purely concentration effects in the cytoplasm due to transfection , GFP-PKB alone or together with RFP-PKB was expressed in NIH3T3 cells ( Figure 1B ) . The lifetimes from cells with expression levels of donor ( GFP-PKB ) and acceptor ( RFP-PKB ) , comparable to those of GFP-PDK1 and RFP-PKB , were identified . Despite the colocalisation of GFP-PKB and RFP-PKB in the cytoplasm and the corecruitment of both PKBs at the plasma membrane , upon activation , the FRET efficiency did not change . This is clearly illustrated by the complete lack of lifetime variation ( Figure 1B and 1C ) . These data indicated that PDK1 and PKB were docking in the cytoplasm and at the plasma membrane under steady state and stimulated conditions . To further investigate whether the docking was PtdIns ( 3 , 4 , 5 ) P3 dependent , cells were pretreated with the phosphatidylinositol 3-kinase inhibitor LY294002 prior to PDGF stimulation . Figure 1D shows that with LY294002 , the translocation of PKB and PDK1 is prevented . However , a cytoplasmic FRET efficiency of 3 . 4 , similar to basal conditions , was still observed ( Figure 1E ) . The PDGF-induced phosphorylation of Thr308 and Ser473 was reduced on pretreatment with LY294002 ( Figure S2E ) . This indicated that the recruitment of the PKB–PDK1 complex to the plasma membrane was dependent on PtdIns ( 3 , 4 , 5 ) P3 production , while the docking of PKB to PDK1 in the cytoplasm was PtdIns ( 3 , 4 , 5 ) P3 independent . To assess the role of PtdIns ( 3 , 4 , 5 ) P3 in an inhibitor-independent manner , PKB and PDK1 PH domain mutants ( mutants that do not efficiently bind to phosphoinositide lipids: PDK1 RRR and PKB R25C [16 , 17] ) were used to determine their influence on interaction with their wild-type binding partners . Cells were cotransfected with GFP-PDK1/RFP-PKB R25C or GFP-PDK1 RRR/RFP-PKB . Under basal conditions , both wild-type proteins interacted with their mutant partner ( Figure S3A and S3B ) . However , upon activation , the PH mutants did not translocate to the plasma membrane and a change in FRET efficiency was not detected ( Figure S3C ) . It is concluded that complex formation in the cytosol is independent of lipid binding . PDK1 docks with phosphorylated/acidic hydrophobic motifs at the C termini of some AGC kinases . This is affected through a PDK1 interacting fragment domain ( PIF pocket ) and additionally through a phosphate-binding site within the upper lobe of the PDK1 kinase domain . PDK1 mutants defective in these interactions , namely the L155E mutant in the PIF pocket and the R131A mutant in the phosphate-binding pocket , still interacted with PKB in the cytoplasm or on PDGF stimulation at the plasma membrane ( Figure S4A and S4B ) . Moreover , the PKB mutant that is not phosphorylated in response to agonists , T308A/S473A ( Figure S4C ) , also retained binding capacity , confirming independently that PKB Ser473 phosphorylation was not required for docking into PDK1′s phosphate-binding pocket . Therefore , the interaction of PKB and PDK1 is independent of the PIF and the phosphate-binding pockets and does not require phosphorylation of Thr308 or Ser473 ( Figure S4D ) . In line with this observation , it has also been shown by using knock-in mutants of PDK1 that the PIF pocket and the phosphate-binding pocket of PDK1 were not implicated in the activation of PKB [18 , 19] . Therefore , the interaction in the cytoplasm occurs via mechanisms that are distinct from those of other AGC kinases . Since these kinases interacted in the cytoplasm , it was of interest to determine why PKB was not constitutively phosphorylated by the associated PDK1 in the basal state . One of the mechanisms that regulate the basal levels of Ser473 is the dephosphorylation of this residue by a serine/threonine protein phosphatase , PP2A , implicated in the downregulation of PKB activation [20] . It was postulated that dephosphorylation may dominate the cytoplasmic phosphorylation of PKB . To test this hypothesis , PP2A activity was blocked by treating cells with okadaic acid . The time-dependent increase of RFP-PKB Ser473 phosphorylation upon treatment indicated that in the absence of PP2A , phosphorylation of Ser473 was enhanced , showing that in NIH3T3 cells PP2A regulates Ser473 phosphorylation in the cytoplasm ( Figure S2A ) . The okadaic acid–insensitive phosphatase PHLPP [21] cannot therefore regulate the basal phosphorylation process . Unlike Ser473 , the okadaic acid stimulation of Thr308 was modest . It was possible that the “basal state” phosphorylation induced by okadaic acid occurred via a transient PtdIns ( 3 , 4 , 5 ) P3–associated form . To investigate this , we exploited the nonbinding PH domain mutant of PKB ( RFP-PKB R25C ) . The okadaic acid–induced phosphorylation of Ser473 and Thr308 did not differ significantly from that of wild-type PKB , whereas the PDGF-induced phosphorylation of RFP-PKB R25C was significantly reduced ( Figure S2B ) . These data indicated that the okadaic acid–induced phosphorylations occurred in the cytoplasm . In RFP-PKB–transfected cells , the okadaic acid–induced Ser473 phosphorylation was at the same level as the Ser473 phosphorylation induced by PDGF ( Figure S2A , densitometry ) . However , in the case of phospho-Thr308 , the okadaic acid signal was reduced compared to phosphorylation triggered by PDGF stimulation ( Figure S2A , densitometry ) . The reduced phosphorylation of Thr308 led to the hypothesis that Thr308 was inaccessible in the cytoplasmic PDK1 complex due to PH domain steric hindrance . If this hypothesis is correct , then the removal of the PH domain should alter the kinetics of Thr308 phosphorylation in response to okadaic acid . Thr308 phosphorylation of the deletion mutant RFP-ΔPH PKB was more extensive than the wild-type PKB ( Figure S2C ) . This provided evidence that , in the cytoplasm , Thr308 was not fully accessible to PDK1 and hence could not be phosphorylated . To verify that in the RFP-ΔPH-PKB mutant Thr308 phosphorylation was due to PDK1 , NIH3T3 cells were cotransfected with RFP-ΔPH PKB and GFP-PDK1 wild-type or the PH domain mutant GFP-PDK1 RRR ( Figure S2D ) . In all cases , a prominent phosphorylation of Thr308 independent of PDGF was detected . Therefore , it is established that when the PKB PH domain is absent , Thr308 could be phosphorylated by PDK1 within the cytosolic complex . Thus , for maintenance of PKB in a Thr308-dephosphorylated state under basal conditions , its PH and kinase domains are predicted to interact to obscure Thr308 accessibility . To test the above hypothesis and determine the possible interactive sites of the PH and kinase domains , molecular modeling was exploited . The calculations of lipophilicity potentials of the PH and kinase domains resulted in finding two isolated hydrophobic patches that were located on the kinase domain ( Figure 2A , in blue ) . On the PH domain , four small hydrophobic patches were found ( Figure 2A , in cyan ) and were located on the same side of the protein . To test the validity of this molecular model , the electrostatic potential maps of each domain were calculated . The complementarity observed between the positive ( red ) and negative ( blue ) lobes for the two domains ( Figure 2B ) was almost perfect . The upper positive lobe of the PH domain corresponded to the upper negative lobe of the kinase domain , and reciprocally , the negative lobe on the PH domain corresponded to a positive lobe on the kinase domain . Figure 2C shows the complementarity between the acidic and the basic residues in the kinase and PH domains of PKB . The yellow arrows indicate the pair of complementary residues on each domain . The fit of the hydrophobic interactions predicts that the Trp80 on the PH domain , at the extremity of the variable loop 3 , inserts inside a deep cleft in the kinase domain around residues Lys297 , Glu298 , and Glu314 ( Figure 2D ) . Furthermore , the crystal structure of the isolated PKB PH domain showed that Trp80 may have an important role in the interaction of the two domains of PKB [7] . From the complementarity of the basic and acidic residues located around Trp80 in the PH and kinase domains , it was envisaged that mutations of these regions would disrupt the docking of these domains . The region where Trp80 interacts in the kinase domain is indicated by an orange arrow ( Figure 2C ) . The two polar residues Gln79 and Thr82 , surrounding Trp80 , were mutated to Glu . These mutations were predicted to result in disruption of the two domains since the acidic residues repulse one another instead of the polar residues of the PH domain interacting with acidic residues of the kinase domain . To monitor and validate in situ our hypothesis on the docking sites of PKB-PH and kinase domains and specifically the Gln79Glu/Thr82Glu mutant ( PKB-EE mutant ) , a PKB reporter was constructed . Full-length PKBα with EGFP on the N terminus and monomeric red fluorescent protein ( mRFP ) on the C terminus ( GFP-PKB-RFP ) was genetically encoded . Variations in PKB conformation were monitored by changes in FRET efficiency . NIH3T3 cells were transfected with PKB tagged only with GFP to use as the reference lifetime ( GFP-PKB ) or with the double-tagged GFP-PKB-RFP reporter ( Figure 2E ) . Under basal conditions , a change in GFP lifetime and FRET efficiency in the GFP-PKB-RFP reporter construct was monitored ( Figure 2E , box and whiskers plots ) . The dynamics of the double-tagged reporter is shown in Figure S5 . It is of note that double-tagged PKB can be phosphorylated on both Thr308 and Ser473 ( Figure 2F ) . The FRET signal under basal conditions was indicative of the close proximity of the two domains in the inactive conformer of PKB . Upon the expression of the Gln79Glu/Thr82Glu mutant ( GFP-PKB EE-RFP ) , a further decrease in the lifetime of the reporter ( red lifetime distribution curve ) was detected and the FRET efficiency significantly increased to 5 . 2 ( Figure 2E , box and whiskers plots ) . As predicted from molecular model , the double-Glu mutation perturbed the juxtaposition of the PH and kinase domains as reflected in this altered N-terminal to C-terminal FRET ( Figure 2E ) . The molecular model implied that a close interaction between the PH and kinase domains was formed as predicted from the complementary surfaces and shown by the extremely significant variations in FRET efficiency between wild-type and mutant PKB . This inactive conformer will be referred to as the “PH-in” conformation . In order for PDK1 to access Thr308 , it was predicted that the PH domain alters its association with the kinase domain on binding to lipids , i . e . , a “PH-out” conformation . To test this hypothesis , the GFP-PKB-RFP reporter was expressed in NIH3T3 cells and its behaviour was monitored upon stimulation with PDGF . Figure 3A illustrates that PDGF induced a change in the conformation of PKB concomitant with its translocation to the plasma membrane . The lifetime distribution curves demonstrated the changes in lifetime ( Figure 3B ) , as did the FRET efficiencies ( Figure 3C ) . By separating the plasma membrane pixels from those of the cytoplasm , the mean FRET efficiency of each compartment was calculated to be 7 . 0 and 4 . 8 , respectively ( Figure 3C , p = 0 . 0006 ) . The cytoplasmic FRET efficiency in PDGF-treated cells in this acute response was similar to basal level FRET efficiency ( Figure 3C ) , indicating that the “PH-in” conformer , unless recruited to the plasma membrane , remained unchanged . Pretreatment with LY294002 inhibited the conformational change in wild-type PKB as seen by the overlap of the distribution curves ( Figure 3B , green and dashed green lines ) . Similarly , the nonbinding PH domain mutant of PKB ( RFP-PKB R25C-RFP ) was not recruited to the plasma membrane and did not change conformation upon stimulation with PDGF ( unpublished data ) . In conclusion , the association of the PH domain with phosphoinositides at the plasma membrane induced a change in PKB conformation by stabilising the “PH-out” conformer and so permitting Thr308 phosphorylation . It has been suggested that PKB in its active state dissociates from the membrane and can phosphorylate soluble targets [22] . However , the “PH-in” conformer may be obscuring its substrate-binding pocket so that PKB's downstream targets will not have access to it . Thus , we speculated that upon dissociation from the plasma membrane , phosphorylated PKB would remain in a “PH-out” conformer where it can phosphorylate cytosolic substrates . To follow the dynamics of the “PH-out” conformer on dissociation from the plasma membrane , the behaviour of the PKB conformation reporter ( GFP-PKB-RFP ) was assessed in live cells ( Figure 4 ) . Cells were transfected with the reporter , and upon stimulation with PDGF , images were acquired every 54 s for up to 20 min ( Figure 4A ) . The intensity images and the quantification of fluorescence intensity illustrated a replenishment of the cytoplasmic fluorescence , which coincided with a decrease in the fluorescence intensity at the plasma membrane ( Figure 4B ) upon recruitment of PKB . The lifetime maps show that the translocated PKB changed conformation , and over time , the populations of the shorter lifetime pixels were augmented in the cytoplasm and the nucleus ( Figure 4A ) . The quantification of these lifetime variations in six different experiments demonstrated that the mean FRET efficiency , upon stimulation , increased gradually over time ( Figure 4C ) . This indicated that during the 20 min of filming , the de novo PDGF-activated “PH-out” PKB conformer dissociated from the membrane and became distributed in the cytoplasm and the nucleus .
Assembling these observations produces a new molecular model for PKB activation dynamics ( Figure 4D ) . Prior to stimulation , PKB and PDK1 form a complex in the cytoplasm that is in constant equilibrium between the associated and dissociated forms . Under basal conditions , PKB is maintained in its inactive form by the interaction of its PH and kinase domains ( “PH-in” ) . The “PH-in” conformation is responsible for preventing the phosphorylation of Thr308 by the associated PDK1 . The cytoplasmic and membrane interactions of PKB and PDK1 are similar in nature . These proteins are in dynamic equilibrium , and it is their corecruitment that concentrates them at the plasma membrane to shift the equilibrium state and form more complex at this location . PKB PH domain interaction with phosphoinositides and its concomitant change in conformation ( “PH-out” ) form the critical step that allows the associated , corecruited PDK1 to phosphorylate Thr308 . In live cells , the dynamics of PKB activation shows that it remains in the “PH-out” conformer upon dissociation from the plasma membrane . Active PKB accumulates in both the cytoplasm and the nucleus . This observation suggests that the phosphorylated kinase domain has a lower affinity for the PH domain , thereby sustaining the “PH-out” conformer , whereas the dephosphorylated kinase domain has a higher affinity for the PH domain . This higher affinity for the PH domain , in the inactive form of PKB , maintains the “PH-in” conformer . The “PH-out” conformer probably sustains its activity until phosphatases render it inactive; once inactivated by dephosphorylation , the “PH-out” conformer returns to the “PH-in” conformation . Aside from providing a rationale for observations related to the mechanism of PKB activation , this detailed model makes interesting predictions regarding the potential sites for influencing PKB and its activation . Specifically , it is predicted that compounds influencing the “PH-in” versus “PH-out” conformers will influence the activation state of PKB . This may have implications both in cancer where PKB is aberrantly activated and in diabetes where transient PKB activation might be insulinomimetic .
Mowiol 4–88 , LY294002 in solution , okadaic acid , and sodium salt were from Calbiochem ( Merck KGaA , http://www . merck . de ) . Human PDGF was from R&D Systems ( http://www . rndsystems . com ) . Cyan 3 dye was from Amersham Biosciences UK ( http://www . amersham . com ) . Phospho-Akt Thr308 and Ser473 as well as pan Akt were from Cell Signaling ( New England BioLabs UK , http://www . neb . com ) . Anti-HA and -GFP antibodies were in-house monoclonal antibodies . Sodium borohydride ( NaBH4 ) and other chemicals were from Sigma-Aldrich Company Ltd . ( http://www . sigmaaldrich . com ) . Poly-l-lysine–coated glass coverslips were from Marathon Laboratory Supplies ( London , United Kingdom ) . pRSET B-mRFP was kindly provided by Dr Roger Tsien ( University of California ) . The construct pCMV5-EGFP-PDK1 ( human PDK1 ) was described previously [23] . To create mRFP-PKB , mRFP was PCR amplified from pRSET B-mRFP with the following oligos: mRFP sense: agagaattcgccgccaccatggcctcctccgaggacgtcatcaaggagttcatgcgc , and mRFP antisense: atcgaattctgcaccggtggagtggcggccctcggcgcgctcgtactgttcc . PCR-amplified mRFP was then subcloned into the murine AKT1 ( described elsewhere [24] ) in an EcoRI site in N terminus of HA-PKB . An extra EcoRI site ( in the vector ) C terminus of PKB was eliminated by mutagenesis prior to the subcloning of RFP using the following primers: EcoRI less-sense: gccacgctgtcctccgcaattcgagctctagaggatcccgg , and EcoRI less-antisense: ccgggatcctctagagctc gaattgcggag gacagcgtggc . mRFP-ΔPH-PKB was made by PCR-amplifying mRFP and put in instead of EGFP using the flanking restriction sites NheI-BglII in pEGFP-ΔPH-PKB vector ( previously described [25] ) . The oligo used for the PCR amplification of mRFP , kindly provided by Dr Johanna Durgan ( CRUK , London , England ) , are NheI sense: aatatagctagcatggcctcctccgaggacg , and BglII antisense: atataaagatcttgcgccggtggagtggcggccctc . The different mutants of PKB and PDK1 were obtained by direct site mutagenesis using the Quick Change mutagenesis kit from Stratagene ( Stratagene Europe , http://www . stratagene . com ) . The oligonucleotides sequence used for each mutant is as follows: PKB-R25C sense: caagacctggcggccatgctacttcctcctcaagaatgatggcacc , PKB-R25C antisense: ggtgccatcattcttgaggaggaagtagcatggccgccaggtcttg , PKB-S473A sense: cgcaggccccacttcccccagttcgcctactcggccagcagcacg , PKB-S473A antisense: cgtgctgctggccgagtaggcgaactgggggaagtggggcctgcg , PKB-T308A sense: gacggtgccaccatgaaggccttttgcggcacacctgagtacctg , PKB-T308A antisense: caggtactcaggtgtgccgcaaaaggccttcatggtggcaccgtc , PDK1-RRR-LLL sense: gcggaagggtttatttgcactactactacagctgttgctcacagaagg , PDK1-RRR-LLL antisense: ccttctgtgagcaacagctgtagtagtagtgcaaataaacccttccgc , PKB-EE sense: cctttatcatccgctgcctggagtggaccgaagtcattgagcgcaccttc , PKB-EE antisense: gaaggtgcgctcaatgacttcggtccactcc aggcagcggatgataaagg , PDK1-L155E sense: catttcaggacgacgagaaggagtatttcggccttagttatgcc , PDK1-L155E antisense: ggcataactaaggccgaaatactccttctcgtcgtcctgaaatg , PDK1-R131A sense: ggtcccctatgtaaccagagaggcggatgtcatgtcgcgcctgg , and PDK1-R131A antisense: ccaggcgcgacatgacatccgcctctctggttacataggggacc . pEGFP-HA-Akt construct was described previously ( murine AKT1 [14]; to simplify , we called this construct GFP-PKB ) . GFP-PKB-RFP was created by putting PCR-amplified mRFP ( EcoRI-XbaI ) instead of YFP in the vector pCDNA3-GFP-PKB-YFP ( described previously [26] ) , using the following oligos: mRFP-Ct sense: agagaattcgcctcctccgaggacgtcatcaaggagttcatgcgcttcaagg , and mRFP-Ct antisense: atctctagattatgcaccggtggagtggcggccctcggcgcgctcg tactgttcc . Note that an extra XbaI site had to be removed by QuickChange in the pCDNA3-GFP-PKB-YFP construct before the subcloning of mRFP ( EcoRI-XbaI ) . NIH3T3 cells from ATCC ( http://www . atcc . com ) were maintained in DMEM 10% donor calf serum and seeded at 150 , 000 in a well of a six-well plate or in a MatTek dish . The transfection was done with 2 μg of DNA of the different constructs ( 2 + 2 μg for cotransfections ) using LipofectAMINE/PLUS reagent ( GIBCO BRL , http://www . invitrogen . com ) in OptiMEM medium ( GIBCO BRL ) as recommended by the manufacturer . The cells were left for 3 h in the transfection mix and then the medium was removed and replaced with DMEM 10% donor calf serum . The experiments were performed 24 or 48 h after transfection . After stimulation or treatment as indicated , the cells were lysed for 15 min on ice in lysis buffer ( 20 mM Tris-HCl [pH 7 . 4] , 150 mM NaCl , 100 mM NaF , 10 mM Na4P2O7 , and 10 mM EDTA supplemented with Complete protease inhibitor cocktail tablet [Roche , http://www . roche . com ) ] . To terminate the reaction , 4× SDS sample buffer ( 125 mM Tris-HCl [pH 6 . 8] , 6% SDS , 20% glycerol , 0 . 02% bromophenol blue supplemented with 10% β-mercaptoethanol ) was added , and the samples were boiled for 5 min . The proteins were separated on a 10% SDS-PAGE gel . The gels were then transferred onto PVDF membrane ( Immobilon P; Millipore , http://www . millipore . com ) , incubated in blocking buffer TBS-T ( 10 mM Tris-HCl [pH 7 . 4] , 150 mM NaCl , 0 . 05% Tween-20 ) supplemented with 3% BSA for 1 h , washed in TBS-T/1% milk for 1 h , and incubated with the different antibodies phospho-Akt ( Thr308 ) antibody ( rabbit polyclonal from Cell Signaling , http://www . cellsignal . com ) or phospho-Akt ( Ser473 ) ( rabbit polyclonal antibody raised against the C-terminus phosphopeptide HFPQFpSYSASS of Akt1 ) at 1:1 , 000 for 2 to 3 h in TBS-T/3% BSA . Anti-GFP ( mouse monoclonal antibody , CR-UK ) was used at 1:5 , 000 for 1 h . The secondary HRP-antibodies were used at 1:5 , 000 in TBS-0 . 2% Tween-20 in 5% milk for 1 h . Western blots were revealed by incubation with ECL ( Amersham Biosciences ) . Density analysis of bands was done with NIH ImageJ 1 . 33u ( National Institutes of Health , http://rsb . info . nih . gov/ij ) . The analysis was performed by substracting the background of the autoradiography . The intensity of the bands was normalised to the total amount of protein , and the relative variations were plotted . NIH3T3 cells were seeded at 150 , 000 on 30-mm glass-bottom tissue culture dishes ( MatTek ) and transfected as described above . At 24 h after transfection , the cells were treated in DMEM containing 10% donor calf serum . The cells were washed twice with PBS and then fixed in 4% paraformaldehyde in PBS for 10 min . The dishes were washed twice with PBS and then 2 ml of PBS supplemented with 2 . 5% ( w/v ) 1 , 4-diazabicyclo-[2 . 2 . 2]-octane ( DABCO ) as an antifade was added to the dishes . The images were acquired straightaway or stored at 4 °C . For live experiments , the cells were seeded and transfected as described above . They were treated with 60 ng/ml PDGF in the same medium , and the lifetime was measured every 54 s with two-photon FLIM for 20 min . The live experiments were repeated six times to verify reproducibility . Details about the method to detect FRET by time domain FLIM can be found elsewhere [27] . All the images were acquired on a modified TE 2000-E inverted microscope ( Nikon Ltd . , http://www . nikon . co . uk ) . The fluorescence lifetime measurements were obtained using an SPC 830 time-correlated single photon counting ( TCSPC ) electronic card ( Becker and Hickl , http://www . Becker-Hickl . de ) in the reverse stop-start mode . A mode-locked tuneable Ti:sapphire laser ( Mira 900; Coherent , http://www . coherentinc . com ) pumped by a solid-state diode laser ( Verdi; Coherent ) was used . For two-photon excitation of EGFP , the laser was tuned at 890 nm and pumped at 6 W . The Ti:sapphire laser generates 125-fs pulses with a repetition rate of 76 . 26 MHz and an average power output of 450 mW . The laser beam was focused with a ×40 oil immersion objective lens ( 1 . 3 NA , Plan-Fluor; Nikon Ltd . ) . The fluorescence was detected through the same objective in a descanned configuration and detected with a fast photomultiplier ( Hamamatsu 7400 , Hamamatsu , http://www . hamamatsu . com ) after filtering with a bandpass filter ( 510 ± 10 nm , Chroma Technology Corp , http://www . chroma . com ) . The acquisition parameters were adjusted to avoid photobleaching [28] and to detect enough photons such that the signal-to-noise ratio was high enough for data analysis . The power of the laser beam was decreased to 9 mW with a neutral optical density filter . The acquisition time was maintained at 300 s for less-intense cells but could be reduced to 30 s for brighter ones . A dwelling time of 17 μs per pixel with a resolution of 256 × 256 pixels was used . Photobleaching was evaluated by comparing the photon count upon time , maintaining the photobleached fraction of EGFP inferior to 2% in all the acquisitions . Depending on the fluorescence intensity of the cells , the maximum photon count of the intensity decay was obtained from 100 to 1 , 000 counts after a spatial binning of 11 × 11 pixels . The binning was maintained constant throughout all experiments . The count rate frequency was below 105 photons/s to avoid any pulse pile-up . Epifluorescence intensity images of both EGFP and mRFP were acquired with the mercury lamp source of the TE 2000-E microscope and the fluorescence detected by a cooled CCD camera ( Hamamatsu ORCA-ER ) . The cubes set in the TE 2000-E microscope turret were FITC ( Nikon Ltd . ) for EGFP and G-2A ( Nikon Ltd . ) for mRFP . No bleed-through was detected . The analyses were performed using in-house software developed by the Advanced Technology Development Group at the Gray Cancer Institute . FRET measurements were based on the reduction of the fluorescence lifetime of the donor upon specific quenching with the acceptor . The fluorescence lifetime of the donor ( τD ) was obtained by fitting the intensity decay with a monoexponential fit . When cells were transfected with both EGFP-PDK1 and mRFP-PKB , two distinct populations can be considered in first approximation . The first one consists of the noninteracting EGFP-PDK1 , with a fluorescence lifetime τD . The second consists of the interacting EGFP-PDK1 with mRFP-PKB . The formation of the complex EGFP-PDK1/mRFP-PKB leads to a transfer of energy from the donor to the acceptor , inducing a decrease in the fluorescence lifetime of EGFP-PDK1 . When measuring the fluorescence intensity decay of EGFP-PDK1 within the cell , photons from both interacting and noninteracting EGFP-PDK1 were collected . Similarly , when transfected with EGFP-PKB-mRFP , the conformational change of the fusion protein leads to a variation of the FRET efficiency due to a variation of distance between the two fluorescent proteins . In both types of experiments , the intensity decays should follow a multiexponential law . However , biexponential fits necessitate a high signal-to-noise ratio . A minimum photon count of 1 , 000 at the maximum of the intensity decay is required . Considering the limitations inherent to experiments in cells , such as photobleaching and low concentration of EGFP , the number of photons was insufficient to accurately proceed with a multiexponential analysis . Moreover , to accurately identify the contribution of two lifetimes , it is preferable that they differ by at least a factor of 2 , which is generally not the case with GFP-like fluorescent proteins . Therefore , to detect FRET , we measured the relative variation of lifetime obtained from monoexponential fit [29] . In each experimental condition , a minimum of ten cells were acquired and analysed . The experiments were repeated at least three times for reproducibility . The intensity images were converted to fluorescence lifetime maps by fitting the fluorescence intensity decay at each pixel . The fluorescence lifetime was represented in pseudo-colour with the same scale so that colours were comparable in each experiment . Fluorescence lifetime distributions were represented by histograms . The histograms were normalised to the total number of pixels of the cell . We used a nonparametric Mann-Whitney test to compare the medians of the two data sets for the two-photon FLIM data using GraphPad InStat software ( version 3 . 0 for Mac-2001; http://www . graphpad . com ) . To interpret the distribution of data , box and whiskers plots were used . The box and whiskers plot is a histogram-like method for displaying upper and lower quartiles , and maximum and minimum values in addition to median [30] . An unpaired t-test with Welch correction was used for frequency FLIM data to determine the significance ( p-values at 95% confidence interval ) of variation in lifetime of GFP-PDK1 in the presence or not of HA-PKB ( Cyan 3 dye [Cy3]-HA antibody ) . Fifty-six or more cells were analysed per experiment; the number of experiments were at least ten . Enrichment in short lifetime pixels over time was quantified from the lifetime histograms . The lifetime histogram for each time point was normalised by the total number of pixels Fi = fi/Ni , where fi and Ni are , respectively , the lifetime histogram and the total number of pixels at time point t = i . The normalised histogram was subtracted from one of the first images ΔFi = FI − F0 , where Fi and F0 are , respectively , the normalised histograms at t = i and t = 0 . The difference was integrated over τ and the enrichment in short lifetime pixels obtained by taking its maximum ηi = Max ∫ΔFidτ . The enrichment ηi was plotted as a function of time . Cells were selected when translocation could be observed . Lifetime measurements were acquired every 54 s . The enrichment in short lifetime pixels was calculated as described above as a function of time . Some cells presented transient periods during which no significant lifetime variation was measurable . For those cells , the enrichment in short lifetime pixels was represented only after this transition period . NIH3T3 cells were seeded at 30 , 000 on poly-l-lysine–coated glass coverslips in 24-well plates and transfected as described above but with a maximum of 0 . 4 μg of DNA . After treatment , the cells were washed twice in PBS and fixed in 4% paraformaldehyde ( PFA ) in PBS for 10 min . The cells were washed in PBS and permeabilised with 0 . 2% Triton X-100 in PBS for 5 min before being incubated with 1 mg/ml NaBH4 in PBS for 5 min . Following another wash with PBS and 10-min incubation in the blocking buffer PBS/1% BSA , the cells were incubated for 1 h with 0 . 22-μm filtered anti-HA antibody labeled with Cy3 in the same buffer . The coverslips were washed twice with PBS and once with water and mounted on slides with the mounting medium Mowiol 4–88 containing 2 . 5% ( w/v ) DABCO as an antifade . The slides were then analysed in the frequency domain FLIM . A detailed description of the FRET monitored by frequency domain FLIM can be found elsewhere [31] . We have monitored lifetime detection in the frequency ( phase ) domain . Phase methods provide an average lifetime where sinusoidally modulated light is used to excite the sample . The lag in the emitted fluorescence signal permits measurement of phase ( τp ) and modulation depth ( τm ) of the fluorescence . The lifetime , <τ> , is the average of phase shift and relative modulation depth ( τm + τp ) /2 of the emitted fluorescence signal . We monitor the decrease in lifetime of donor ( EGFP ) in presence of the acceptor ( mRFP or Cy3 dye ) . All images were taken using a Zeiss Plan-APOCHROMAT ×100/1 . 4 numerical aperture , phase 3 oil objective , with images recorded at a modulation frequency of 80 . 218 MHz . GFP-PDK1 was excited using the 488-nm line of an argon/krypton laser , and the resultant fluorescence was separated using a combination of dichroic beamsplitter ( Q505 LP; Chroma Technology Corp . ) and narrow band emission filter ( BP 514/10; Lys & Optik , Lyngby , Denmark ) . For the PH domain , we used the x-ray structure with Protein Data Bank ( http://www . rcsb . org/pdb ) code 1UNP [7] . In this structure , only the two first amino acids were lacking and were not rebuilt ( Figure 5 ) . There is no available structure for the kinase domain of human Akt-1 ( PKBα ) , but there are several structures for the homolog human Akt-2 kinase domain ( PKBβ ) . Among them , one of the most complete is 1GZN with a resolution of 2 . 5 Å [11] . In this side chain , the following amino acids are lacking: 146–147 , 198–206 , 253 , 296 , and 442 . These were rebuilt prior to any experiment using the Homology module of Insight II ( Accelrys , http://www . accelrys . com ) . Because the two proteins ( Akt-1 and Akt-2 ) have a very high homology rate ( 98 . 8% with 82 . 3% of strict identity ) , building a model from this homology sequence was straightforward . Moreover , in the chosen alignment ( Figure 5 ) , there is a difference of only one amino acid between the two sequences that results from a two–amino acid deletion after position 114 and a one–amino acid insertion at position 268 ( Figure 5 , arrows ) . Alignment of sequences was performed using the Homology multiple-sequence aligner software via the PAM ( 120 or 250 ) matrices or the ClustalW software ( http://www . ebi . ac . uk/clustalw ) as implemented within Homology module . In these latter cases , alignments were performed using the Blosum and Gonnet matrices . The crude model of each domain , after full protonation at pH 7 . 4 , was refined by several cycles of splice repairs on the loops and relaxation on the SCR's side by 100 steps of Steepest Descent and 500 cycles of conjugate gradient . Small 5-ps molecular dynamics runs were performed on each loop in order to relax the strains . Finally , the entire backbone was fixed and the side chains were fully minimised with conjugate gradient method until a root-mean-square of 0 . 1 kcal/Å/mol was obtained . The minimisation process was resumed by applying a tethering constraint to the backbone . This constraint was slowly lowered ( starting with a 100 kcal/Å/mol ) force constant ) in a stepwise manner so that at each step the energy of the root-mean-square gradient was less than 0 . 1 0 . 1 kcal/Å/mol ( steps are 100 , 50 , 25 , 10 , and 0 0 . 1 kcal/Å/mol ) . Lipophilicity potentials were calculated using a previously described software [32] that calculated on each node of a 1-Å cubic grid the sum of the lipophilicity contribution of each nonhydrogen atom as a function of distance . On the kinase domain , a unique patch of hydrophobic potential is located on the same side as the ATP active site with the following amino acids: Leu316 , Leu321 , Lys356 ( alkyl chain only ) , Phe358 , Leu362 , and Met363 . For the PH domain , four small hydrophobic patches are also located on the same side of the protein and on a “plateau” with the following amino acids: Val4 , Tyr18 , Ile19 , Leu52 , and Trp80 . The Molecular Electrostatic Potential ( MEP ) maps were drawn within the software PYMOL ( version 0 . 97 ) after calculations performed using the implemented APBS plug-in . This last package enabled us to efficiently evaluate electrostatic properties by solving the Poisson-Boltzmann equation [33] . All possible hydrophobic and electrostatic interactions were properly fitted . The Trp80 at the extremity of a loop inserts precisely inside a deep cleft in the kinase domain with Lys297 , Glu298 , and Glu314 located around it . Moreover , in this configuration , the C terminus of the PH domain is on the same side as the N terminus of the kinase domain . In this model , the PH domain completely blocks the entry of the ATP active site and Thr308 will not be accessible .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession numbers for the genes and gene products discussed in this paper are mRFP ( AF_506027 ) , murine AKT1 ( NM_009652 ) , and construct pCMV5-EGFP-PDK1 ( human PDK1 , AF_017995 ) . The Swiss-Prot ( http://www . ebi . ac . uk/swissprot ) database code for the homolog human Akt-2 kinase domain ( PKBβ ) is P31751 . | Regulation of intracellular signaling depends on the precise operation of molecular switches such as kinases and phosphatases . Disruption of their activities leads to inappropriate cellular proliferation , growth , and survival . Protein kinase B ( PKB ) is a critical kinase that regulates events downstream of growth factor receptors . It is also involved in human cancer , where its overexpression induces malignant transformation . We studied the molecular mechanisms of PKB's interaction with its upstream regulator , 3-phosphoinositide–dependent protein kinase 1 ( PDK1 ) . By using a fluorescent probe , we monitored the conformational changes of PKB in cells using Förster resonance energy transfer detected by fluorescence lifetime imaging microscopy . Applying this approach , we show that PKB and PDK1 are found as complexes in the cytoplasm . Despite this proximity to its regulator , we show that PKB remains inactive through an intramolecular interaction of its pleckstrin homology ( PH ) domain and kinase domain . We refer to this inactive state as the “PH-in” conformer . Following growth factor activation , PKB changes conformation to the “PH-out” conformer and is phosphorylated by PDK1 . The active “PH-out” conformer dissociates from the plasma membrane to phosphorylate downstream proteins . Our in vivo studies provide a new model for the mechanism of PKB . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"biochemistry",
"cell",
"biology",
"computational",
"biology",
"biophysics",
"homo",
"(human)"
] | 2007 | Intramolecular and Intermolecular Interactions of Protein Kinase B Define Its Activation In Vivo |
The intraerythrocytic parasite Plasmodium—the causative agent of malaria—produces an inorganic crystal called hemozoin ( Hz ) during the heme detoxification process , which is released into the circulation during erythrocyte lysis . Hz is rapidly ingested by phagocytes and induces the production of several pro-inflammatory mediators such as interleukin-1β ( IL-1β ) . However , the mechanism regulating Hz recognition and IL-1β maturation has not been identified . Here , we show that Hz induces IL-1β production . Using knockout mice , we showed that Hz-induced IL-1β and inflammation are dependent on NOD-like receptor containing pyrin domain 3 ( NLRP3 ) , ASC and caspase-1 , but not NLRC4 ( NLR containing CARD domain ) . Furthermore , the absence of NLRP3 or IL-1β augmented survival to malaria caused by P . chabaudi adami DS . Although much has been discovered regarding the NLRP3 inflammasome induction , the mechanism whereby this intracellular multimolecular complex is activated remains unclear . We further demonstrate , using pharmacological and genetic intervention , that the tyrosine kinases Syk and Lyn play a critical role in activation of this inflammasome . These findings not only identify one way by which the immune system is alerted to malarial infection but also are one of the first to suggest a role for tyrosine kinase signaling pathways in regulation of the NLRP3 inflammasome .
Malaria is a widespread infectious disease that affect up to 300 million individuals in the tropical and sub-tropical regions of the world , and is responsible for 2–3 million deaths annually [1] . Malaria is caused by parasites of the Plasmodium genus and is characterized by episodic fevers , anemia , headache and organ failure . Plasmodium parasites feed on erythrocyte hemoglobin and uses a heme detoxification mechanism that results in the formation of an insoluble , inert , dark-brown crystalline metabolic waste called hemozoin ( Hz ) [1] , [2] . Hz is involved in the fever observed during the malaria process as intravenous injection of Hz caused thermal deregulation and was associated with the induction of pyrogenic cytokines [3] . In addition , the release of both Plasmodium-derived Hz and merozoites during the erythrocyte burst phase of the disease coincides with the massive induction of pro-inflammatory cytokines , such as IL-1β and TNF , and with the periodic fevers characteristic of malaria [3] , [4] . IL-1β secretion is controlled by the recently described inflammasome , a signaling platform scaffold composed of NLR family members such as NLRC4 ( NOD-like receptor containing CARD domain or IPAF ) and members of the NLRP ( NOD-like receptor containing pyrin domain ) family including NLRP1 and NLRP3 ( also known as NALP3 and cryopyrin ) . In addition , the NLRP3 inflammasome is composed of the adaptor molecule ASC ( Apoptosis-Associated Speck-Like Protein ) and the effector molecule caspase-1 , the latter which is responsible for the cleavage of pro-IL-1β into its active form [5] , [6] . TNF is induced by a wide variety of innate receptors but in particular by many members of the Toll-like receptors ( TLR ) . It was previously reported that Hz can induce IL-1β secretion in vitro and in vivo [7] , [8] , however , TLRs are not required for the Hz-induced inflammatory response [9] . Given the clear association of IL-1β with the induction of fever and recent studies demonstrating that the NLRP3 inflammasome senses inorganic materials , such as monosodium urate ( MSU , a gout-associated uric-acid crystals ) , silica , asbestos and aluminum hydroxide by producing IL-1β [6] , we tested whether Hz can activate the NLRP3 inflammasome . In addition , while NLRP3 ligands have been well identified , little is known about the upstream mechanisms that regulate its activation . Some mechanisms that have been proposed include efflux of potassium , increased intracellular calcium , reactive oxygen species ( ROS ) generation and lysosome disruption [6] , [10] . However , having previously reported that both MSU and Hz can trigger production of inflammatory mediators via the activation of signaling cascades involving MAP kinase family members and various transcription factors , we have herein addressed the role of upstream signaling in the activation of the inflammasome that results in IL-1β production in response to the malarial pigment Hz .
In these studies we utilized a chemically synthesized Hz to prevent contamination that could result from native Hz purification; the synthetic Hz is morphologically and chemically similar to native Plasmodium-isolated Hz ( Fig . S1 ) . Previously , we reported that both synthetic and native Hz induce similar expression profiles of chemokines and pro-inflammatory cytokines [7] . In addition , the synthetic Hz was subjected to elemental analysis to assess its purity . Theoretical calculated values of the molecular formula of Hz ( C68H62N8O8Fe2 ) give 66 . 35% of carbon ( C ) , 5 . 08% of hydrogen ( H ) and 9 . 10% of nitrogen ( N ) . We have obtained elemental values from our synthetic Hz preparation very close with the theoretical one ( C: 66 . 5%; H: 5 . 3%; N: 8 . 9% ) . To further show the purity of Hz , we performed an agarose gel with 200 µg of Hz and we did not detect any trace DNA or RNA contamination ( Fig . S2A ) and treatment with DNase or RNase did not interfere with Hz-induced IL-1β production ( Fig . S2B ) . These data indicate that our synthetic Hz preparation is high purity and free of contaminant . To evaluate whether Hz activates the inflammasome , we measured IL-1β secretion by PMA-differentiated human monocytic cell line ( THP-1 ) stimulated with increasing concentrations of Hz or MSU . Hz- and MSU-induced IL-1β production was found to be comparable ( Fig . 1A ) . In accordance with previous studies showing that HSP-90 stability [11] modulates inflammasome assembly , we found that Hz-induced IL-1β secretion was reduced in the presence of the HSP-90 inhibitor geldanamycin D ( Fig . 1B ) . Inhibition of caspase-1 activity using a specific competitor ( Y-VAD-FMK ) [12] or a broad caspase inhibitor ( Z-VAD-CHO ) also blocked Hz-induced IL-1β ( Fig . 1C ) . To confirm the activation of caspase-1 we used the bone-marrow-derived macrophages ( BMDM ) , since detection of the active form of caspase-1 in THP-1 cells is difficult as reported by others [13] , [14] . Here , we show that Hz induced cleavage of caspase-1 to its enzymatically active ( p10 subunit ) form . BMDM were pre-stimulated with LPS in order to prime the induction of pro-IL-1β . As shown in Figure 1D , Hz and MSU , but not the pre-treatment with LPS , induced cleavage of caspase-1 and mature IL-1β production , which was completely abolished in BMDM from caspase-1 deficient mice . These results suggest a role for the inflammasome in Hz-induced IL-1β production . To further establish which intracellular receptors and/or adaptor proteins are activated by Hz , we used BMDM from mice deficient in NLRP3 , ASC or another NLR , NLRC4 ( NLR containing CARD domain , also known as IPAF ) . We found that Hz- and MSU-induced caspase-1 activation and IL-1β maturation were dependent on NLRP3 and ASC but not NLRC4 ( Fig . 2A ) . On the other hand , macrophages from NLRC4 mice failed to respond to Salmonella typhimurium infection ( Fig . S3 ) . To evaluate whether activation of the NLRP3 inflammasome is involved in Hz-induced inflammatory responses in vivo , mice were injected intraperitoneally with Hz and then neutrophil recruitment to the site of injection was examined . Hz induced significant recruitment of neutrophils to the peritoneal cavity in wild type , but not in ASC-deficient ( Fig . 2B ) or in NLRP3-deficient mice ( Fig . 2C ) . As expected , NLRC4 was not involved in the inflammatory response induced by Hz ( Fig . 2C ) . We further investigated whether IL-1β directly contributed to the recruitment of neutrophils . As expected , IL-1β deficient mice showed a significant decrease in the number of neutrophils elicited by Hz stimulation ( Fig . 2D ) . However , we did not observe a complete abrogation of neutrophil influx as previously seen with IL-1 receptor-deficient mice stimulated with other inflammasome ligands [15] . These results suggest that a portion of the Hz-induced inflammatory response in vivo may results from other ligands of the IL-1 receptors and/or other cytokines and chemokines known to be induced by Hz [3] , [7] , [8] . Thus far , we have shown that Hz-induced IL-1β production is dependent on the NLRP3 inflammasome , in addition , it is known that IL-1β is involved in malarial fever [4] . To evaluate the role of IL-1β and the NLRP3 inflammasome during malarial disease we infected IL-1β- and NLRP3-deficient mice with Plasmodium chabaudi adami DS , which is a mouse virulent strain . Of interest , both IL-1β- and NLRP3 mice presented a slight but significant lower body temperature ( Fig . 3A and 3B ) and parasitemia ( Fig . 3C and 3D ) in the early phase of infection . These knockout mice also showed a significantly prolonged survival compared with wild type mice , but ultimately succumbed to the infection ( Fig . 3E and 3F ) . Finally , in the late phase of infection , the level of IL-1β was significantly lower in NLRP3-deficient mouse in comparison with wild type mice ( Fig . 3G ) and was not detectable in IL-1β-deficient mouse ( data not shown ) . These results indicate that IL-1β is an important factor in the pathophysiology during malaria infection . Hz is rapidly engulfed by phagocytes , both in infectious and experimental conditions [2] . Therefore , to test the importance of phagocytosis on Hz-induced IL-1β production , cells were treated with cytochalasin D - a powerful actin polymerization inhibitor - prior to the addition of the crystals . Consistent with other crystals that induce inflammasome activation [15] , [16] , [17] , we found that Hz-induced IL-1β seems to be dependent on its internalization ( Fig . 4A ) . Furthermore , under certain conditions phagocytosis requires cholesterol-rich lipid domains [18] and as expected , cholesterol depletion by MβCD inhibited HZ-induced IL-1β ( Fig . S5A ) , which was due to the disruption of lipid rafts ( Fig . S5C ) . Further characterization of Hz phagocytosis by confocal immunofluorescence microscopy revealed that Hz was internalized in a vacuole that acquired lysosomal features , as shown by the presence of Lamp-1 surrounding the engulfed Hz phagosomes ( Fig . 4B ) . Phagocytosis is generally accompanied by the generation of reactive oxygen species ( ROS ) , which modulates inflammasome activation by crystals such as silica [19] , MSU [15] and asbestos [20] . Since Hz induces ROS production [7] its requirement in Hz-induced IL-1β production was evaluated . The ROS scavenger , N-acetylcysteine ( NAC ) inhibited both Hz- and MSU-induced IL-1β production ( Fig . 4C ) , which suggests a potential upstream role for ROS in inflammasome activation by Hz . Cellular potassium efflux is another critical step in inflammasome activation induced by all known NLRP3 activators [21] , [22] . As shown in the Figure 4D , inhibition of potassium efflux by high concentrations of extracellular potassium decreased IL-1β production induced by Hz . The above results suggest that Hz shares a common mechanistic pathway in the activation of the NRLP3 inflammasome with classical triggers such as ATP and others insoluble crystals [21] , [23] . Recently , lysosomal destabilization has been proposed as one mechanism whereby inorganic materials such as silica and aluminum hydroxide activate the inflammasome [17] . To assess lysosomal morphology in the context of Hz stimulation , we performed a confocal analysis of PMA-matured THP-1 cells loaded with a self-quenched conjugate of ovalbumin ( DQ-OVA ) that fluoresces only upon proteolytic degradation . We found that Hz did not affect the shape of lysosomes in comparison to untreated cells . In contrast , silica-treated cells contained swollen lysosomes ( Fig . 4E ) , suggesting that Hz may activate the inflammasome through distinct , but related pathway . Indeed , inhibition of the lysosomal cysteine protease ( cathepsin B ) by the specific inhibitor CA-074 abrogated IL-1β induced by Hz and silica ( Fig . 4F ) [17] . However , it is still unclear how this enzyme is involved in inflammasome activation and indeed , many of the proximal signaling events in NLRP3 and NLR activation remain unknown . Whereas we obtained clear evidence that Hz can induce IL-1β production in an inflammasome-dependent manner that required active cathepsin B , we did not find evidence of Hz-induced lysosomal rupture as previously reported with silica [17] . Release of cathepsin B without lysosomal rupture has been observed in monocytes treated with the potassium ionophore nigericin [24] . In addition , the widely expressed Spleen Tyrosine Kinase ( Syk ) was shown to be required for cathepsin B release into the cytosol in a model of B cell receptor-mediated apoptosis [25] . We therefore screened Hz-activated macrophages for changes in their tyrosine phosphorylation profiles . Consistent with the possible involvement of Syk , we observed a band with an apparent molecular weight of 72 kDa that was phosphorylated in response to Hz , but not MSU ( Fig . 5A ) . We then carried out anti-Syk immunoprecipitation , followed by anti phospho-tyrosine analysis and found that Syk was phosphorylated in response to Hz , but not MSU stimulation ( Fig . 5B ) . Even by extending the time-course of stimulation , MSU did not induce Syk phosphorylation ( Fig S4A ) . Syk is typically activated via receptors or adaptor proteins containing immunoreceptor tyrosine-based activation motifs ( ITAMs ) or ITAM-like domains phosphorylated by Scr family kinases following receptor clustering [26] , [27] . The Src kinase inhibitor PP2 decreased the Hz-induced Syk phosphorylation in a dose dependent manner ( Fig . 5C ) . Syk activation can be mediated by the Scr family kinase member Lyn [28] . Lyn is typically found in lipid raft signaling platforms and disruption of these rafts by MβCD ( Fig . S5C ) indeed blocked , in dose-dependent manner , Syk phosphorylation in Hz-stimulated monocytes ( Fig . S5B ) . Using BMDM from Lyn-deficient mice , we found that Hz-induced Syk phosphorylation required Lyn , and further confirmed that MSU does not utilize this signaling pathway in either murine or human macrophages ( Fig . 5 ) . Next we evaluated the role of Lyn and Syk in Hz-induced IL-1β production . IL-1β secretion stimulated by Hz was inhibited in macrophages treated with the Syk inhibitor piceatannol ( Fig . 6A ) , the Scr kinase inhibitor PP2 ( Fig . 6B ) , and more specifically using Lyn-deficient BMDM ( Fig . 6C ) . Importantly , in this last experiment , Hz-induced IL-1β production was only partially inhibited , which suggest that another member of the Src kinase family could play the same role of Lyn , since these kinases are known to be functionally redundant [28] . Of note , MSU-induced IL-1β production was not affected in Lyn-deficient BMDM pre-treated with LPS . To evaluate the relative roles of LPS and Hz in the induction of this signaling pathway , we treated BMDM with LPS and we observed that LPS by itself did not induce phospho-Syk , and indeed pre-treatment with LPS reduced Hz-induced Syk phosphorylation ( Fig . S4B ) . Furthermore , Hz-induced Syk activation is not affected by the absence of the MyD88 adaptor protein ( Fig . S4C ) . However , MyD88-deficient cells show a delay in the phosphorylation of c-jun N-terminal kinase ( JNK ) stimulated by LPS ( Fig . S4C ) , similar as previously reported [29] . These results rule out a possible effect of LPS on Syk phosphorylation . Consistent with the involvement of this kinase in a pathway upstream of the inflammasome , NLRP3- , ASC- and NLRC4-deficient macrophages exhibited normal Syk phosphorylation upon Hz stimulation ( Fig . 6D ) . Syk activates various downstream signaling pathways , including phosphoinositide 3-kinase ( PI3K ) [30] and extracellular signal-regulated kinase ( ERK ) . To test whether the PI3K pathway is required for propagation of the Syk signaling pathway following Hz exposure , the PI3K inhibitor wortmannin was used prior to Hz stimulation . Inhibition of PI3K indeed abrogated IL-1β maturation ( Fig . 7A ) . We have previously identified MAPK activation upon Hz stimulation of macrophages [31] . We therefore attempted to isolate which pathways might be required for Hz-induced IL-1β production using known p38 and ERK kinase inhibitors . Whereas p38 phosphorylation can be observed following Hz stimulation , inhibition of p38 with SB203580 failed to block Hz-induced IL-1β production ( Fig . 7B–D ) . On the other hand , inhibition of ERK with Apigenin abrogated Hz-induced IL-1β secretion ( Fig . 7E ) . Altogether , these results reveal that Lyn/Syk activation following Hz exposure initiates the PI3K and ERK signaling pathways and these pathways appear to regulate the production of mature IL-1β . While a number of stimuli are known to activate the NLRP3 inflammasome , there is no evidence that NLRP3 directly recognizes these ligands . Therefore an indirect pathway of NLRP3 activation is likely , however the identity of the direct molecular switch of NLRP3 has not been identified . Our studies provide the first evidence for a role of tyrosine kinase signaling molecules in NLRP3 activation . To examine whether Syk can modulate the inflammasome by directly interacting with its components , we immunoprecipitated Syk and then immunoblotted for potential partners associated with Syk by silver staining and western blotting ( Fig . 8A ) . Selected differential bands were analyzed by LC-tandem mass spectrometry . Interestingly , two to three different peptides covering 11–23% of the Pyrin domain ( Pyd ) [32] were identified . Pyrin domains are known to mediate protein-protein interactions and are crucial in many of the NLR inflammasome complexes , and in particular , mediate the NLRP3 and ASC interaction [6] . We therefore confirmed by western blotting whether NLRP3 or ASC can be co-immunoprecipitated ( co-IP ) with Syk . Whereas NLRP3 was shown to weakly interact with Syk , ASC was found to strongly associate with this kinase upon Hz stimulation ( Fig . 8B ) . These findings suggest that Syk , and possibly other unidentified signaling kinases , can associated with the ASC/NRLP3 inflammasome . Another possible mechanism is that Syk could be controlling the NLRP3 inflammasome by regulating cathepsin B activation . First , we tested if Hz can induce release of the active form of cathepsin B in the supernatant and as showed in the Figure 9A , Hz did not induce cathepsin B release into supernatant as has been observed with MSU and silica . However , using a cathepsin B substrate that emits red fluorescence upon cleavage we demonstrated that Hz induces rapid ( 30 min ) and transient ( maximum 1 . 5 h ) intra-compartmental cathepsin B activation that was dependent on Syk activation ( Fig . 9B ) . These results indicate that Syk not only can associate with the inflammasome component but it can also modulate cathepsin B activation .
It has been described that NLRP3 senses many crystalline materials that are involved in inflammatory diseases , such as MSU [15] , silica [19] , and asbestos [20] . Here we provide the first demonstration that the malaria pigment hemozoin ( Hz ) can also activate the NLRP3 inflammasome . Importantly , the Hz concentration shown to activate the NLRP3 inflammasome in vitro is similar in range to the concentration of Hz in the blood of patients with moderate parasitemia [8] , [33] . Moreover , it was never shown in the previous studies that direct contact between a crystal and NLRP3 is necessary to induce activation . Similarly , we found that Hz does not translocate from the phagosome/lysosome compartment to the cytoplasm , as it is located within LAMP-1-positive compartments , suggesting that Hz activated the NLRP3 inflammasome in an indirect manner . It has been proposed that the NLRP3 inflammasome senses not only pathogen-associated molecular patterns but also danger signals such as stress-related molecules [5] . In agreement , here we show that Hz-induced IL-1β production was dependent on ROS generation and potassium efflux into the cytoplasm . In addition to previous studies on the inflammasome , we further identified an upstream signaling pathway involving the Src kinase Lyn , the tyrosine kinase Syk and Syk-downstream kinases such as PI3K and ERK that collectively appear to be involved in the regulation of Hz-induced IL-1β production . Simultaneously to us , it has been recently reported that Syk kinase is involved in upstream signaling of NLR inflammasome triggered by fungi [34] . Whether these findings represent a general regulatory mechanism of this intracellular innate immune response will need further investigation . The Lyn/Syk pathway appears to be uniquely activated in the innate response to Hz crystals , as opposed to other NLRP3-activating crystals such as MSU . In our hands , MSU did not induce Syk or Lyn phosphorylation in PMA-differentiated THP-1 cells nor in BMDM . However , MSU was previously reported to trigger Syk phosphorylation in dendritic cells [35] and human neutrophils [36] , as well as Lyn phosphorylation in neutrophils [37] . An intriguing question is how this signaling cascade may modulate the inflammasome/IL-1β production . For instance , we found some indication that Syk can interact with ASC , but not NLRP3 . ASC , as it is well known , interacts with NLRP3 . These results suggest that Syk may modify ASC . In support of this finding , there is evidence that the ASC pyrin domain can be phosphorylated [38] . Moreover , hyperphosphorylated PSTPIP1 ( proline serine threonine phosphatase-interacting protein ) was shown to interact with the pyrin protein [39] , resulting in its conformational change and further its interaction with ASC [40] . Another possible mechanism whereby kinases can modulate IL-1β production is by modulating intracellular calcium concentration or cathepsin B activation . Syk is involved in the activation of intracellular calcium mobilization in other models [41] . In fact , increased calcium concentrations have been found to modulate inflammasome activation by different stimuli such as MSU and UV radiation [22] , [42] . Finally , Syk was found to control the activation of cathepsin B and Hz-induced IL-1β production was dependent on cathepsin B activation , similar to other inflammasome activators such as silica , MSU [17] or nigericin [24] . We showed that specific inhibition of Syk blocked the Hz-induced cathepsin B activation . Collectively , it is clear that different steps in the Hz-induced IL-1β production can be regulated by intracellular signaling . However , further study will be necessary to better characterize these regulatory events in regards to the different inorganic crystals that can trigger NLRP3 inflammasome activation . Another interesting observation is that Hz-activated cathepsin B occurred in the intracellular compartment and is rapidly quenched ( 1–3 hours ) , suggesting either a transient activation or cathepsin B release into the cytosol . The idea of transient activation of cathepsin B by Hz is supported by the absence of cathepsin B in the supernatant of cells stimulated with Hz and the absence of lysosomal damage upon Hz treatment . The mechanism utilized by Hz-activated cathepsin B to modulate the inflammasome remains unclear . However , a possible mechanism is that cathepsin B can activate directly caspase-1 as it has been shown in previous works [17] , [24] . Of interest , both caspase-1 and cathepsin B , in addition to inflammasome components and IL-1β are found in multivesicular bodies surrounded by LAMP-1 [43] . It is known that Syk and Syk-activated downstream kinases such as PI3K regulate the trafficking of intracellular vesicles [44] . In this way , Hz-induced Syk might be controlling not only the inflammasome cascade but also the trafficking of multivesicles . The Lyn/Syk activation finding raises the intriguing possibility that an as yet unidentified receptor or adaptor protein containing an ITAM or ITAM-like domain , such as Dectin-1 , TREM family members , Siglec or DAP12 [26] , [27] , might be activated upon Hz stimulation to trigger the signaling cascade involved in inflammasome activation . However , a recent work with dendritic cells demonstrated that MSU did not require a surface receptor - instead the crystals interact with surface lipid rafts and this was enough to trigger Syk/PI3K pathway [35] . In our study , we have demonstrated that lipid rafts are involved in the Hz-induced signaling pathway and IL-1β production . Other potential receptors that could mediate Hz-triggered signaling are the Toll-like receptors ( TLR ) . However , we have recently demonstrated in collaboration with Parroche and colleagues [9] that Hz alone fails to activate TLRs except when Hz is coated with parasitic DNA and consequently activating TLR9 . Similarly , we also observed that HEK293 cells transfected with different TLRs were not activated by Hz although these cells were able to induce NF-κB activation following specific ligand stimulations ( Jaramillo and Olivier , unpublished data ) . We also showed that the MyD88 signaling pathway is not involved in the Hz-induced Syk phosphorylation . Experiments to identify surface receptors or lipids that recognize Hz are currently underway . In the present work we further supported the role of NLRP3-mediated IL-1β production in Hz-mediated inflammatory cell recruitment using IL-1β deficient mice . Apart from its inflammatory role , IL-1β is a pyrogenic cytokine that in small concentrations induces the production of other cytokines such as IL-6 and can cause hypertension and fever [45] . In fact , we showed that NLRP3- and IL-1β-deficient mice exhibited lower body temperature during the early phase of P . chabaudi Adami infection . Hz-induced IL-1β can be the mediator of the up-regulation of chemokines and cytokines during malaria infection , which is independent of TLRs but dependent on MyD88 [46] . This suggests that another MyD88 dependent receptor such as IL-1R is involved and supports a role for IL-1β in malaria-related pathology . Corroborating this hypothesis , we showed that IL-1β- and NLRP3- deficient mice showed a better survival than wild type mice in murine experimental model of malaria . Not surprisingly , it was not sufficient to provide full protection likely due to the complexity of malarial disease , which is under the regulation of many different receptors , cytokines , signaling events and physiological features . Collectively , our study provides the first demonstration that a malarial-derived metabolic product , namely hemozoin , can induce NLRP3 inflammasome activation and IL-1β production though the involvement of the Src kinase Lyn and the tyrosine kinase Syk . However , excessive IL-1β secretion can be deleterious to the host; in fact , we observed that higher production of IL-1β correlates with early death in murine experimental malaria . Therefore these findings strongly support the fact that Hz is critical in malaria pathology . A better understanding of the molecular and cellular events regulating malaria inflammatory-related pathologies may provide new insights into the design of treatments aimed at reducing the exaggerated inflammatory disorders and debilitating sequelae .
With the subheading Ethics Statement , all protocols used in this study were approved by the Institutional Animal Care and Use Committees at the McGill University or Yale University . IL-1β- and Lyn-deficient mice were provided by Dr . G . Sébire and Dr . K . W . Harder ( University of Sherbrooke , Quebec and University of British Columbia , Vancouver , Canada ) , respectively . The generation of IL-1β- , Lyn- , NLRP3- , ASC- , caspase-1- , and NLRC4-deficient mice has been described previously [47] , [48] , [49] , [50] , [51] . Caspase-1- , ASC- , and NLRP3-deficient mice were backcrossed onto the C57BL/6 genetic background for at least nine generations . NLRC4-deficient mice were backcrossed onto the C57BL/6 genetic background for at least six generations . Age- and sex-matched C57BL/6 mice purchased from the National Cancer Institute or Charles River were used as WT controls . Hemin ( >99% of purity ) was purchased from Fluka; RPMI-1640 medium , Penicillin-Streptomycin-Glutamine ( PSG ) from Wisent , fetal bovine serum ( FBS ) , Alpha MEM medium from Gibco; CV-Cathepsin B detection kit , PP2 , piceatannol , geldanamycin , cytochalasin D , Y-VAD-FMK and Z-VAD-CHO from Biomol; MSU , anti-human NLRP3 and ASC from Alexis Biochemical; inhibitor protease cocktail from Roche; CHAPs from Fisher; A/G-coupled agarose beads , anti-human pro-IL-1β , anti-human or murine caspase-1 and anti-Syk from Santa Cruz; True Blot anti-rabbit Ig , anti-phosphoY/HRP from eBioscience; PVDF from Bio-rad; anti-LAMP-1 Ab from Developmental Studies Hybridoma Bank at the University of Iowa; anti-human mature IL-1β , anti-pp38 and anti-p38 from Cell signal; anti-pSyk and anti-pY ( 4G10 ) from Upstate; rat or goat anti-murine IL-1β and recombinant IL-1β from R&D system; DQ-OVA from Invitrogen; anti-rat AlexaFluor 568 , cholera toxin B-AlexaFluor 568 from Molecular Probes; DRAQ5 from Biostatus; Fluoromount-G from Southern Biotechnology; all others unlisted or not indicated reagents were purchased from Sigma . L929 and THP-1 cell line from ATCC . MyD88 KO BMDM was generated from MyD88-deficient mice and kindly supplied by Dr . Danuta Radzioch ( McGill University , Montreal , Canada ) . Native and Synthetic Hz have been obtained as previously described [8] , [31] . We have modified synthetic Hz preparation , using high purity chemical reagents ( >99% of purity ) , as follows: 0 . 8 mmol Hemin was dissolved in degassed NaOH ( 0 . 1 M ) for 30 minutes with mild stirring . pH 4 . 0 was adjusted adding drop-wise propionic acid . The mixture was allowed to anneal at 70°C for 18 hours . Then washed three times with NaHCO3 ( 0 . 1 M ) for three hours and the last wash with MeOH . All washes were alternated with distilled H2O . Finally , the sample was then dried in a vacuum oven overnight over phosphorous pentoxyde . All synthetic hemozoin samples were analyzed by X-ray powder diffraction , field emission gun scanning electron microscopy , and infra-red spectroscopy to characterize the crystalline state of Hz . Hz purity was assessed by elemental analysis [52] . THP-1 cells ( ATCC ) were cultured with RPMI-1640 medium supplemented with 10% FBS , 1% PSG , 50 µM of 2-β-mercaptoetanol , Glucose 4 . 5 g/L and 1 mM sodium pyruvate . THP-1 differentiation: ( 1 . 5×106 cells/mL ) were incubated with 0 . 5 µM of PMA , after three hours cells were washed and plated at 0 . 75×106 cells/mL or 0 . 2×106 cell/0 . 5 mL in 12 well plates ( IL-1β ) or 24 well plates containing coverslips ( confocal ) and incubated for 20–24 hours . This treatment increases the phagocytic properties of the cells and induces a constitutive production of pro-IL-1β . Prior to stimulation , cells were washed and 500 µL of Alpha MEM medium without FBS was replaced . Cells were pre-treated with different drugs for 1 hour and stimulate with Hz , MSU or silica as indicated in figure legends . Gender and age matched wild type ( WT ) , NLRP3- or IL-1β-deficient mice were injected i . p . with 5×104 Plasmodium chabaudi adami DS infected red blood cells obtained from syngeneic infected mice . Parasitemia was assessed at day 5 , 7 and then every day by examination of Giemsa stained blood smears and was expressed as mean parasitemia . Body temperature was measured using an infrared thermometer ( La Crosse Technology ) . Survival of mice was monitored and blood serum was collected when the temperature dropped down to 26°C . IL-1β was measured by ELISA with rat monoclonal and goat anti-mouse IL-1β . The detection limit was 6 . 25 pg/mL of IL-1β . Bone marrow cells were obtained by flushing the femurs and tibias from mice . Cells were used from fresh or from frozen marrows . Erythrocytes were lysed with 2 mL of NH4Cl ( 155 mM ) in Tris/HCl ( 10 mM ) , pH 7 . 2 ( 9∶1 solution ) /mouse . Bone marrow cells were adjusted to 7×106 cells/10 mL and plated in 100 mm dishes with RPMI-1640 medium supplemented with 1% of PSG , 10% FBS and 30% ( v/v ) L929 cell culture supernatant . The supernatants of bone marrow cells were changed every two days in order to renew the cytokines and nutrients . After 7 days , the culture dishes were washed with PBS and replaced by ice cold PBS , incubated on ice for 15 min and cells were vigorously detached . BMDM were adjusted to 1 . 5×106/2 mL or 0 . 2×106 cells/0 . 5 mL in RPMI medium supplemented with 5% FBS ( Gibco ) and 1% of PSG and plated in 6 well plates ( IL-1β ) or 24 wells plate ( confocal ) . The next day , cells were washed with warm PBS ( 37°C ) and replaced by 500 µL of Alpha MEM medium without FBS . Cells were , as indicated in figure legends , stimulated with Hz , MSU or infected with Salmonella typhimurium as described by Franchi et al . [53] . Supernatant and cell extract analysis: After designated incubation time , supernatants were collected and protein was precipitated with trichloroacetic acid at 10% final concentration . Precipitates were then dissolved in Tris/HCl 0 . 1 mM pH 8 . 0 and Laemmli sample load buffer . Cell extracts were obtained by lysing cells with Igepal 1% ( for signaling , in 1× PBS , 20% Glycerol , 1× inhibitor protease cocktail , 2 mM Na3VO4 and 1 mM NaF ) or triton 1% ( for caspase-1 , in TNE buffer: 10 mM Tris/HCl pH 7 . 5 , 150 mM NaCl , 5 mM EDTA and 1 . 5× inhibitor protease cocktail ) . Whole supernatant protein and equal amount of protein or cell lysate were subjected to SDS-PAGE and immunoblot analysis . IP: Cells lysates were extracted with lysis buffer ( 1% CHAPs detergent in TNE buffer , 1× inhibitor cocktail , 2 mM Na3VO4 and 1 mM NaF ) . Cells lysates were pre-incubated for two hours at 4°C with protein A/G-coupled agarose beads and 1 µg of unspecific matched isotype control antibody ( Ab ) . Equal amount of protein were immunoprecipitated with protein A/G-coupled agarose beads or True Blot anti-rabbit Ig and 2 µg of specific or unspecific matched isotype control Ab overnight . Beads were spun down 3 times with lysis buffer and proteins were denatured in Laemmli load buffer . SDS-PAGE/Immunoblot: Samples from supernatants , cell extracts or IP were subjected to 10% ( signaling ) or 15% ( IL-1β and caspase-1 ) acrylamide gel ( all reagents from Laboratoire Mat . Inc . , Montreal , Qc , Canada ) or 4–12% NuPAGE® gel ( for p10 caspase-1 and IP , Invitrogen ) . After transfer onto PVDF membranes , they were subjected to immunoblot analysis with the indicated Ab and matched secondary HRP-conjugated Ab . In some experiments , optical density was determined using AlphaDigiDoc 1000 v3 . 2 software ( Alpha Innotech corporation ) . OVA uptake: THP-1 cells ( 0 . 2×106 cells/coverslip 12 mm from Fisher ) were treated with 10 µg of DQ-OVA in the absence or presence of Hz ( 200 µg/mL ) or Silica ( 400 µg/mL ) for 30 min , washed and incubated up to three hours . Laser settings were adjusted on DQ-OVA fluorescence emission that is stronger than hemozoin or silica . Phagosome: BMDM were fixed , permeabilized using 0 . 1% Triton X-100 , and non-specific surface Fcγ-receptor binding were blocked as described [54] . For immunofluorescence experiments , cells were labelled with the rat anti-LAMP-1 Ab and an anti-rat AlexaFluor 568 . DRAQ5 was used to visualize DNA . Cathepsin B activity: THP-1 cells ( 0 . 2×106 cells/coverslip 12 mm from Fisher ) were pre-treated for 30 min with 5 µM of piceatannol and stimulated or not with Hz ( 200 µg/mL ) . A cathepsin B substract ( Arg-Arg ) 2 linked with cresyl violet were given 30 min before the end of incubation time and cleaved substract generated a red fluorescence . All coverslips ( THP-1/OVA or BMDM ) were mounted on slides with Fluoromount-G . Detailed analysis of protein localization on the phagosome was performed by using an oil immersion Nikon Plan Apo 100 ( N . A . 1 . 4 ) objective mounted on a Nikon Eclipse E800 microscope equipped with a Bio-Rad Radiance 2000 confocal imaging system ( Bio-Rad Laboratories , Hercules , CA ) . WT , IL-1β- , NLRP3- , ASC- , caspase-1- and NLRC4-deficient mice were injected intraperitoneally with 800 µg of hemozoin in 1 ml of endotoxin-free PBS . Control groups were injected with 1 mL of PBS . After six hours , the mice were euthanized and the peritoneal cavity was washed with 10 mL of PBS . Cells recovered from the peritoneum were counted and the percentage of neutrophils was determined from an H&E stain ( DiffQuick; Dade Behring , Inc . ) of a cytospun sample . Unpaired Student's t-test was used when comparing two groups and ANOVA/Bonferroni test when comparing more than two groups . The differences were considered significant when p<0 . 05 . Survival curves for infected and control mice were compared using the Mantel-Haenszel test . Statistical analysis was performed using Prism 5 . 00 software ( GraphPad , San Diego , Calif . ) . | Malaria is widespread in the tropical and sub-tropical regions of the world , and is responsible for 2–3 million deaths annually . This disease is caused by parasites of the Plasmodium genus . The parasite feeds on the hemoglobin of red blood cells and generates a metabolic waste called hemozoin ( Hz ) . Hz is released into the blood circulation during the rupture of red blood cells , which coincides with the production of many cytokines such as interleukin-1β ( IL-1β ) , responsible in part for the periodic fever that is characteristic of the malaria disease . Here , we investigated how Hz activates macrophages ( cells that engulf foreign material ) to produce IL-1β . We found that Hz is taken up by macrophages initiating signals such as the tyrosine kinases Syk and Lyn that communicate to intracellular receptors . We also showed that Hz-induced IL-1β production is dependent on activation of the intracellular receptor NLRP3 , the adaptor protein ASC and a protease called caspase-1 that cleaves IL-1β , therefore allowing it to be released from the cells . These findings not only identify one way in which the immune system is alerted to malarial infection but also dissect some of the signaling events triggered by Hz in the NLRP3 inflammasome pathway . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/cell",
"signaling",
"infectious",
"diseases/protozoal",
"infections",
"immunology/leukocyte",
"signaling",
"and",
"gene",
"expression",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"immunology/leukocyte",
"activation"
] | 2009 | Malarial Hemozoin Activates the NLRP3 Inflammasome through Lyn and Syk Kinases |
The emergent human pathogen Mycoplasma genitalium , with one of the smallest genomes among cells capable of growing in axenic cultures , presents a flask-shaped morphology due to a protrusion of the cell membrane , known as the terminal organelle , that is involved in cell adhesion and motility and is an important virulence factor of this microorganism . The terminal organelle is supported by a cytoskeleton complex of about 300 nm in length that includes three substructures: the terminal button , the rod and the wheel complex . The crystal structure of the MG491 protein , a proposed component of the wheel complex , has been determined at ~3 Å resolution . MG491 subunits are composed of a 60-residue N-terminus , a central three-helix-bundle spanning about 150 residues and a C-terminal region that appears to be quite flexible and contains the region that interacts with MG200 , another key protein of the terminal organelle . The MG491 molecule is a tetramer presenting a unique organization as a dimer of asymmetric pairs of subunits . The asymmetric arrangement results in two very different intersubunit interfaces between the central three-helix-bundle domains , which correlates with the formation of only ~50% of the intersubunit disulfide bridges of the single cysteine residue found in MG491 ( Cys87 ) . Moreover , M . genitalium cells with a point mutation in the MG491 gene causing the change of Cys87 to Ser present a drastic reduction in motility ( as determined by microcinematography ) and important alterations in morphology ( as determined by electron microscopy ) , while preserving normal levels of the terminal organelle proteins . Other variants of MG491 , designed also according to the structural information , altered significantly the motility and/or the cell morphology . Together , these results indicate that MG491 plays a key role in the functioning , organization and stabilization of the terminal organelle .
Mycoplasmas are microorganisms belonging to the class of Mollicutes ( ‘soft skin’ ) that evolved from Gram-positive bacteria by genome reduction and are characterized by the absence of a cell wall , by their small cell sizes and by their reduced biosynthetic machinery . Consequently , these microorganisms live in nature as obligate parasites depending on the uptake of essential nutrients from their hosts . Mycoplasma species vary in form and many are able to move by gliding motility [1 , 2] . In particular , Mycoplasma genitalium , one of the smallest autoreplicative microorganisms known , is a motile species belonging to the pneumoniae cluster of mycoplasmas . Due to its small genome , of only ~480 protein-coding genes , M . genitalium has been used as a model of minimal cell [3 , 4] and is the subject of intense work in systems biology research [5–7] . M . genitalium is an emergent and prevalent sexually transmitted pathogen involved in urogenital infections in humans , including non-gonococcal and non-chlamydial urethritis and inflammatory reproductive tract diseases in women . A review article on this issue described the need for an early diagnostic of the infection , which increases the risk of HIV transmission when persistent [8 , 9] . In addition , an intense search for novel therapeutic agents against M . genitalium has been launched as several studies revealed the existence of isolates resistant to treatments with azithromycin [10–12] , indicating that there is a continuous need to search for potential drug and vaccine targets in this microorganism [11 , 13] . M . genitalium has a flask-shaped morphology that consists of a cell body with a protrusion of the cell membrane , called the terminal organelle , which is the scaffold for cell adhesion , division and motility , processes deeply related to infectivity . The terminal organelle is supported by a complex cytoskeleton that is formed by three main substructures: the terminal button , the rod and the wheel complex located , respectively , at the tip , the center and the rear with respect to the cell body [2 , 14–16] . Moreover , adhesins P110 and P140 , which are very abundant at the surface of the terminal organelle , are essential for attachment to host cells , together with the accessory proteins MG218 , MG312 , MG317 and MG491 [17–20] A model for gliding motility in mycoplasmas from the pneumoniae cluster proposed a cyclic process where the rod , anchored to the wheel complex , has a central role [14 , 15] . According to this model in a first step , the tip of the terminal organelle binds to the substrate with the rod fully extended and , in a second step , the rod contracts , dragging the cell forward . However , we have very recently demonstrated that this model is no longer valid since M . genitalium cells remain motile in the absence of the rod element [20] . The same work also highlighted the role of P110 and P140 adhesins and of the P32 protein on promoting cell movement as previously proposed [1 , 21] . The wheel complex might also be involved in chromosome segregation by attaching the mycoplasma chromosome to the terminal organelle [22] . M . genitalium MG200 and MG491 proteins have been proposed as components of the terminal organelle wheel complex ultrastructure [23 , 24] and in agreement with this , it has recently been found that MG200 and MG491 interact with each other specifically influencing cell motility [25] . In this work , the crystal structure of MG491 was determined and found to present a unique tetrameric organization as a dimer of asymmetric pairs of subunits . The structural information guided the design of MG491 variants , which presented striking alterations in cell motility and in cell morphology demonstrating the key role played by MG491 in the organization and functioning of the terminal organelle of M . genitalium .
Crystals of the same type were obtained from both the full length MG491 protein ( residues 1 to 346 ) and from a construct of the protein N-terminal region , MG491-Nt ( residues 1 to 308 ) , though times required for crystallization changed from a few months to a few weeks , respectively . Initial phases were derived from Single-wavelength Anomalous Diffraction data [3] collected at the selenium absorption edge of a MG491-Nt variant where three isoleucine residues ( Ile36 , Ile168 , Ile205 ) had been replaced by seleno methionines ( see Material and Methods ) ( Fig 1A ) . Four selenium sites were located within the crystal asymmetric unit , with two sites related to the other two by a Non-Crystallographic Symmetry ( NCS ) two-fold axis . Structure determination was then achieved by density modification , averaging between both the less isomorphous crystals ( see Materials and Methods ) and using the NCS two-fold axis ( S1 Fig ) . The final refined structure , with four subunits in the crystal asymmetric unit ( residues 65–204 , 67–203 , 66–203 and 62–205 for subunits A , B , C and D , respectively; Figs 1B , 2A and 2B ) , has agreement Rwork and Rfree factors of 22 . 17% and 24 . 93% , for a seleno methionine MG491-Nt data set at 3 . 0 Å resolution ( Table 1 , S2 Fig , PDB entry code 4XNG ) . The unexpected presence of four subunits in the crystal asymmetric unit had two major implications: i ) extensive proteolysis had to have happened in the C-terminal region of the protein ( not visible in the determined structures ) during crystallization . Four subunits each with 308 residues would give an unacceptably low crystal solvent content of 4% . ii ) Four subunits cannot be symmetrically related by the only two-fold symmetry found . The structure determined for MG491 subunits consists of an antiparallel three-helix-bundle , with helices α1 ( residues 70–102 ) , α2 ( 111–145 ) and α3 ( 167–203 ) connected by loops L1 ( 103–110 ) and L2 ( 146–166 ) , respectively ( Fig 1A and 1B ) . Helix α1 is kinked in its central part , due to the insertion of a π-helix turn starting in residue Cys87 ( Fig 1B ) . Only a small and variable number of residues , from three in subunit B to eight in subunit D , could be traced preceding helix α1 . Therefore , about sixty residues in the N-terminal region of MG491 appear to be flexible with respect to the subunit three-helix-bundle domain . Differences between the four subunits present an averaged root mean square deviation ( r . m . s . d . ) for Cα atoms of only 0 . 32 Å , which increases to 0 . 70 Å between subunits not related by the NCS two-fold axis , with the largest deviations corresponding to loop L1 and to the central part of loop L2 ( residues 152–160 ) . The four MG491 subunits found in the crystal asymmetric unit present three different types of intersubunit interfaces that were named symmetric , tight and loose ( Figs 2B and 3A–3C ) . The symmetric interface , with a total buried area of ~650 Å2 , corresponds to interactions across the two-fold symmetry axis and involves only subunits A and C , while subunits B and D do not contact with each other ( Fig 3A ) . The tight interfaces , with a total buried area of ~2*1200 Å2 , correspond to the two interfaces between subunits in the pairs A/D and C/B ( within each pair subunits are related by ~72° rotation ) ( Fig 3B ) . The loose interfaces , with a total buried area of ~2*450 Å2 , correspond to the two interfaces between subunits in the pairs A/B and C/D ( within each pair subunits are related by ~108° rotation ) ( Fig 3C ) . Therefore , the four subunits found in the crystal asymmetric unit present a network of ( extensive ) intersubunit interactions strongly suggesting that the MG491 molecule can form tetramers , in agreement with studies by gel filtration , crosslinking with glutaraldehyde and nano-ElectroSpray Ionization Mass Spectrometry ( S3 Fig ) . Despite the fact that the four subunits in the tetramer are structurally similar , as reflected by the low r . m . s . d . values , they are placed in two different environments . Therefore , two types of subunits can be identified according to the residues that participate in the tight and loose interfaces of each subunit . The organization of the MG491 tetramer , with only a two-fold symmetry , can be defined as a dimer ( C2 molecular symmetry ) of asymmetric pairs of subunits . Quantification of the deviation from an accurate four-fold molecular symmetry gives an average for the Cα atoms of all the residues of 7 . 3 Å ( Fig 3D ) . Attempts to form a regular ( symmetric ) oligomer using only the interactions corresponding to the tight interface would result in a helical aggregate with four subunits at most due to steric clashes ( S4A Fig ) . In turn , oligomerization using only the loose interface would result in helical aggregates with three subunits at most ( S4B Fig ) . Loop L2 mediates interactions in both the tight and the loose interfaces presenting a different conformation in each interface . In the loose interface the backbone in the central part of loop L2 moves towards the neighbor subunit by ~2 Å , with the side chain of Phe158 flipping to a different conformation about 7 . 5 Å away ( Fig 4A and 4B ) . Interestingly , the conformational changes observed for loop L2 result in similar intersubunit interactions for the two interfaces when analyzed with LigPlot+ [28] . In particular , interaction of Phe157 with Gly91 in the tight interface is mirrored as interaction of Phe157 with Gly80 in the loose interface ( S5 Fig ) . In the MG491 tetramer , the four Cys87 residues ( Cys87 is the only cysteine in the whole MG491 sequence ) are located close to each other ( Fig 4C ) , in particular across the symmetric interface , suggesting the possible formation of disulfide bonds between subunits . However , in the structure determined disulfide bonds are absent , which might be due to the presence of reducing agents required for crystallization . Following an established oxidation protocol to form disulfide bonds in vitro [29] , the protein was diluted in 1x PBS ( pH 7 . 5 ) to a final concentration of 1 mg/ml and incubated for 2–4 h in 1% ( v/v ) DMSO , resulting in the formation of intersubunit disulfide bridges but only between ~50% of the subunits ( Fig 4C ) . This result supports a departure of symmetry in the molecular organization of MG491 that would agree with disulfide bridges being formed only between the two subunits at the symmetric interface of the MG491 tetramer . The sequence alignment ( http://espript . ibcp . fr ) [30] between MG491 and its M . pneumoniae homolog , P41 , gives an overall identity of 53% mainly due to the high identity found for the proteins N-terminal regions ( until about residue 203 in MG491 , Fig 1A ) . In particular , the most conserved regions are for MG491 helices α1 and α3 as well as for loops L1 and L2 . Accordingly , the MG491 structure is expected to be well preserved in the M . pneumoniae protein P41 . To study the biological relevance of the unique molecular organization of MG491 and to further investigate the function of this protein , three structure-guided M . genitalium mutant strains were engineered with mutations in residues directly involved in the interactions between subunits . These three protein variants were: i ) Cys87 replaced by a serine; ii ) Phe157 and Phe158 substituted both by alanines and iii ) the peptide from Asn155 to Lys160 , corresponding to a large fragment of loop L2 , deleted ( Fig 1A ) . An additional mutant strain lacking the N-terminal region of MG491 ( residues 1–61 ) was also engineered to gain insight into the function of this region . Mutant alleles mg491C87S , mg491F157A-F158A , mg491ΔloopL2 and mg491ΔNt were introduced in pMTncat plasmids [23] under the control of the MG438 promoter ( Fig 5A ) . These mini-transposons were electroporated into cells from M . genitalium Δmg491 null mutant strain lacking MG_491 [20] . One colony from each transformation experiment was selected for the different alleles and named mg491-C87S , mg491-F157A-F158A , mg491-ΔloopL2 and mg491-ΔNt , respectively . Transposon insertion sites were investigated by direct genome sequencing and all the selected transformants showed transposon insertion sites in genes other than those involved in the terminal organelle architecture and/or gliding motility functioning ( Table 2 ) . Upon introduction of the wild type allele in Δmg491 cells , steady-state levels of the MG491 protein were restored in the Δmg491-mg491cat strain ( Fig 5B ) . Normal levels of MG491 were also observed in the mg491-F157A-F158A , mg491-ΔloopL2 and mg491-ΔNt mutant strains ( Fig 5B ) . However , a lower amount of MG491-Cys87Ser was detected in Δmg491-C87S cells , suggesting that Cys87 might play an important role in protein stability . The apparent molecular weight of the deletion variant protein MG491ΔNt was ~40 kDa , in agreement with the expected value . Lower levels of MG491 have already been shown to correlate well with the existence of several downstream events in terminal organelle related proteins [20] . Therefore , it was not surprising to observe in Δmg491 cells , a drastic decrease in the amount of adhesion proteins P110 and P140 , and of most of the cytadherence accessory proteins ( Fig 6A and 6B ) . However , these cells exhibited normal amounts of the cytadherence accessory proteins MG200 and MG219 . The adhesin and cytadherence accessory proteins levels were also restored upon reintroduction of the wild type MG_491 allele in the M . genitalium Δmg491 strain and a similar effect was observed in the transformants containing the mutant alleles mg491-C87S , mg491-F157A-F158A and mg491-ΔloopL2 . In contrast , the levels of adhesins and cytadherence accessory proteins were not restored after the introduction of the mutant allele in mg491-ΔNt , indicating that the N-terminal region of MG491 has an important role in the formation and stabilization of the terminal organelle ( Fig 6A and 6B ) . Cells from the Δmg491 strain showed a filamentous morphology when observed by scanning electron microscopy ( Fig 7A ) . The characteristic flask-shaped morphology typically observed in wild type cells was restored in the Δmg491-mg491cat strain ( Fig 7B ) . The gliding properties were also restored in these cells , showing no significant differences when compared to those exhibited by G37 wild type cells ( Table 3 , S1 and S2 Movies , S6 Fig ) . Cells from the mg491-C87S strain showed normal terminal organelles ( Fig 7C ) but this strain also presented a high frequency of cells bearing multiple terminal organelles , which correlated with a reduced number of motile cells and a slower mean velocity as measured by time lapse microcinematography ( Table 3 , S3 Movie ) . When examining microcinematographies of G37 wild type cells , 18% of the motile cells show one or more resting periods . These resting periods are short and seem not to be related to cell division . Remarkably , the frequency of motile cells showing resting periods in mg491-C87S strain was 49% , indicating that the high frequency of non-motile cells might be a consequence of these resting periods . Likewise , a large amount of cells bearing multiple terminal organelles was also observed when examining the mg491-ΔloopL2 strain ( Fig 7E ) but these cells showed , in addition , a drastic decrease in different gliding motility parameters ( Table 3 , S4 Movie ) and a low hemadsorption activity ( S6 Fig ) . Moreover , both strains showed normal levels of all known proteins involved in gliding motility ( Fig 6A and 6B ) and exhibited no significant changes in the overall terminal organelle architecture ( Fig 8 ) . These data suggest that gliding motility impairments detected in these strains are a direct consequence of the mutations introduced in MG_491 . In contrast , the gliding properties and the frequency and architecture of terminal organelles in mg491-F157A-F158A cells were similar to those of wild type cells ( S5 Movie and Figs 7D and 8C ) . However , this variant shows a lower hemadsorption activity than G37 wild type ( Table 3 , S6 Fig ) and a large amount of minute cells smaller than 0 . 35 μm in size as revealed by electron microscopy ( Fig 7D and Table 3 ) . Cells from this strain were stained with Hoechst 33342 , examined by time lapse microcinematography and finally visualized by epifluorescence microscopy . Most of the minute cells analyzed ( 93 . 3% ) showed no detectable fluorescence after staining with Hoechst indicating that these cells did not contain detectable amounts of DNA . Among these non-fluorescent cells , 53 of them ( 54 . 1% ) were found motile during the examination period ( S7 Fig ) , indicating that these minute cells were consequence of terminal organelle detachments . Such minute cells are rarely observed in M . genitalium G37 wild type strain . In contrast , cell detachments are frequently observed when the terminal organelle is not properly anchored to the cell body . Minute cells were previously described to be the result of terminal organelle detachments from the main cell body in M . genitalium cells lacking the C-terminal region of MG491 [22] and also in M . pneumoniae cells with a disrupted MPN311 gene , which codes for the P41 protein ( S1 Table ) [30] . Thus , the presence of minute cells in the mg491-F157A-F158A strain suggests that the intersubunits interactions promoted by Phe157 and Phe158 are required for the proper assembly of MG491 , possibly playing an important role in the stabilization of the protein quaternary structure . However , oligomerization of protein variant Phe157Ala-Phe158Ala appears similar to the wild type protein presenting , surprisingly , even a slightly increased stability ( S8 Fig ) . Finally , electron microscopy analysis of the mg491-ΔNt strain revealed the presence of a large amount of cells with filamentous morphology and the absence of rods inside these filaments ( Fig 8E ) , similar to what was observed when examining the parental Δmg491 strain [20] . Moreover , no motile cells were observed for this strain ( S6 Movie ) , suggesting that MG491 is involved in the assembly of the terminal organelle and in its stabilization through the protein N-terminal region .
Human pathogen M . genitalium , from the pneumoniae cluster of mycoplasmas , presents a flask-shaped morphology conferred by a polar structure , known as terminal organelle , which neither structurally nor functionally is yet well understood . MG491 protein from M . genitalium shares a high sequence identity with the M . pneumoniae protein P41 , which is known to be an important component of the terminal organelle in M . pneumoniae and has been located at the base of the electron-dense core [31] . The location of MG491 in the terminal organelle of M . genitalium was also supported by the finding that a 25-residue region interacts specifically with MG200 [25] , a protein that had been shown to be involved in gliding motility [23 , 24] . The structural characterization of MG491 in this work , indicates that MG491 subunits are composed of three distinct regions with a 60-residue N-terminus , a central three-helix-bundle spanning about 150 residues and a C-terminal region that contains the residues that interact with MG200 and appears to be mostly unstructured ( Fig 1A ) . Only the central helix-bundle is well defined in the electron density maps of crystals from several constructs of MG491 ( Fig 1B ) . All the solved crystal structures contain four crystallographically independent subunits , which are interwoven by a network of interactions with each other ( Fig 2A and 2B ) . Surprisingly , each one of this tetrameric ensembles is organized with only one two-fold symmetry axis that relates pairs of subunits ( Fig 3A ) . These pairs can be defined in two alternative ways referred as loose or tight according to the extension of the interacting interface between the two subunits in the pair ( Fig 3B and 3C ) . Steric clashes make it impossible to model regular oligomers containing only one kind of these interacting surfaces ( S4 Fig ) . The biological relevance of this unique organization has been assessed by the characterization of M . genitalium mutant strains with alterations in residues involved in intersubunit interactions . The MG491 variant Cys87Ser preserves normal levels of all the other terminal organelle proteins but presents a very significant reduction in motility ( Table 3 ) , comparable to the effects observed in deletion mutants of whole proteins involved in gliding motility [24 , 32] . Terminal organelle development is synchronized with cell division and cytokinesis appears to be highly coordinated with gliding motility , which is also essential for segregation of the terminal organelles to the opposite cell poles . In this way , alterations in motility often result in the presence of cells bearing multiple terminal organelles [32–34] . In contrast , cells from mg491-C87S mutant strain show only a very modest increase ( 7 . 6% ) in the frequency of cells with multiple terminal organelles . Moreover , the gliding velocity of these cells is not significantly lower than that exhibited by wild type cells ( Table 3 ) . The large number of non-motile cells in the mg491-C87S mutant is strongly correlated with an increased frequency of cells showing resting periods , rather than with the presence of cells stalled in the cytokinesis process , as observed in other gliding mutants [24 , 32] . The higher frequency of resting periods in the mg491-C87S mutant strain suggests that the Cys87 residue of MG491 might have an important role in the regulation of gliding motility . Interestingly , the frequency of resting periods was also found increased in M . genitalium cells lacking the EAGR box from MG200 [16] , reinforcing the relevance of the interplay between MG200 and MG491 in the regulation of gliding motility [22] . The strikingly severe effects of a single point mutation on the only cysteine residue suggests a major role for this cysteine that is likely related with the asymmetric formation of intersubunit disulfide bridges observed in vitro ( Fig 4D ) . Only the two subunits in the tetramer of MG491 that interact across the molecular two-fold symmetry axis ( subunits A and C in Figs 3 and 4 ) are expected to participate in this interaction , while the cysteine residues of the other two subunits would remain reduced or available for different interactions . MG491 variants designed to alter the tight and loose interfaces by deleting the central part of loop L2 ( ΔloopL2 ) or replacing two of the loop residues ( Phe157Ala-Phe158Ala ) also resulted in significant changes in cell motility and cell morphology ( Table 3 and Figs 7E and 7F and 8C and 8D ) . As expected , alterations in the deletion variant ΔloopL2 , which also includes residues Phe157 and Phe158 , are stronger than those observed for the Phe157Ala-Phe158Ala variant and , accordingly , the frequency of cells with multiple terminal organelles is higher in ΔloopL2 cells ( Table 3 ) . In contrast , the MG491 double mutation variant Phe157Ala-Phe158Ala showed a significant increase in the amount of minute cells or terminal organelles detached from the main cell body despite the fact that no clear changes were observed in vitro for the oligomerization of the variant ( S8 Fig ) . The increased frequency of minute cells strongly supports that Phe157 and Phe158 residues have a main role in the stability of the wheel complex or in the interactions of the wheel complex with the rod . Despite the complexity of the terminal organelle of mycoplasmas , here we show that this structure can be a reachable target for a thorough characterization zooming out in resolution from atomic to cellular levels . In this work , the structural information obtained from the crystal structure of MG491 has guided the preparation of several M . genitalium mutant strains of this protein . As a result , the motility and morphology of M . genitalium cells have been importantly affected , providing , for the first time , information on how the structure of a protein relates with the organization , stabilization and functioning of the terminal organelle . Motile mycoplasmas with spreading deficiencies are associated to a reduced infectivity [17 , 35] , which emphasizes the relevance of MG491 in the virulence of M . genitalium .
The E . coli XL1-Blue strain was used to amplify the plasmids used in this study and was grown on LB agar plates or liquid LB media overnight . Ampicillin was added at 0 . 1 mg/ml . M . genitalium G37 wild type and mutant strains were grown in SP-4 broth at 37°C under 5% ( v/v ) CO2 in tissue culture flasks ( from TPP , Switzerland ) until mid-log phase of growth . Transformant colonies were isolated on SP-4 agar plates supplemented with 2 μg/ml tetracyclin and 34 μg/ml chloramphenicol . The coding sequence of the MG_491 gene was amplified from M . genitalium G37 wild type genomic DNA with oligonucleotides 5MG491 and 3MG491 and ligated into a pBE plasmid [36] . The triplet coding for Trp232 from the MG491 protein was changed from TGA to TGG by amplification of this plasmid with oligonucleotides MutMG491PA and MutMG491PB and circularization of the amplicon with T4 DNA ligase . Afterwards , the sequence coding for this mutated version of the full length MG491 protein was cloned between NdeI and XhoI restriction sites of a pET21d expression vector ( Novagen , Madison , WI , USA ) , which also codes for a C-terminus hexa-histidine tag . The resulting vector was transformed into E . coli BL21 ( DE3 ) cells and the transformant cells were plated on LB/agar plates supplemented with ampicillin . After checking the correctness of the DNA sequence , the transformant cells were cultivated in 1 l LB medium containing 0 . 1 mg/ml ampicillin and induced overnight with 1 mM IPTG at 20°C with constant shaking after reaching an OD600 of ~0 . 6 . Subsequently , the cells were harvested by centrifugation at 4500 xg for 15 min at 4°C . The pellet was resuspended in lysis buffer ( 0 . 02 M Tris-HCl ( pH 8 . 0 ) , 0 . 5 M NaCl , 0 . 02 M imidazole , complete EDTA free protease inhibitor ( Roche Diagnostics , Mannheim , Germany ) ) and the cells disrupted by sonication . The total lysate was then centrifuged twice for 20 min at 45000 xg to remove cells debris and filtered through a 0 . 22 μm filter . The his-tagged MG_491 gene product present in the resulting supernatant was firstly purified through a 5 ml HisTrap HP column ( GE Healthcare Life Sciences , Uppsala , Sweden ) previously equilibrated in 0 . 05 M Tris-HCl ( pH 8 . 0 ) buffer containing 0 . 5 M NaCl and 0 . 02 M imidazole , concentrated to a suitable volume and then loaded on a Superdex 200 16/60 gel filtration column ( GE Healthcare Life Sciences , Uppsala , Sweden ) equilibrated in 0 . 05 M Tris-HCl ( pH 8 . 0 ) containing 0 . 15 M NaCl . To obtain the phases for the X-ray structure determination several methionine residues ( absent in the MG491 sequence ) were introduced based on secondary structure element predictions , in positions corresponding to Ile36 , Ile168 , Ile205 and Ile313 . A new expression vector was prepared ( pET21d-MG491-B ) using pET21d-MG491 as template and the oligonucleotide primers containing the appropriate target substitutions ( see S1 Table ) . Limited proteolysis experiments performed with Trypsin on a MG491 sample generated a fragment of about 30–35 kDa with an intact N-terminal ( revealed by Edman sequencing ) , which suggested that the C-terminal region is more accessible and thus more susceptible to proteolysis . Given this , and using the pET21dMG491-B vector as template and the appropriate primers ( see S1 Table ) , a shorter variant of the protein was designed spanning MG491 residues 1 to 308 ( MG491Δ308 ) . The resulting PCR fragment was finally cloned into a pOPINE expression vector [37] , which encodes for an extra lysine and a hexa-histidine tag at the C-terminal end of the construct . This new vector was then transformed into E . coli BL21 ( DE3 ) cells and the MG491-Nt protein was expressed and purified following the same protocol used to prepare the full length protein . Additionally , the seleno methionine-labeled MG491-Nt protein was produced by growing a 0 . 1 l pre-culture overnight at 37°C in presence of 400 μl L-methionine at 10 mg/ml , 2 ml of 50% ( w/v ) glucose ( freshly prepared and filtered through a 0 . 22 μm filter ) and the appropriate antibiotic . Cells were then recovered by centrifugation at 4500 xg for 15 min , washed three times with 1x PBS , to remove the L-methionine that has not been incorporated by the cells , and finally resuspended in 2 ml 1x PBS . This cell pellet was then used to inoculate 1 l of SelenoMet media ( Molecular Dimensions Ltd . , Newmarket , UK ) in presence of 9 ml L-seleno methionine at 10 mg/ml and supplemented with OnEx solutions 1 , 2 and 3 from the Overnight Express Autoinduction Systems 1 ( Novagen , Madison , WI , USA ) . Cells were grown for 6 h at 37°C , then the temperature was lowered to 25°C and growth was continued for 20 h with constant shaking before harvesting . The seleno methionine-labeled MG491-Nt protein was finally purified following the same protocol used for the full length MG491 and MG491-Nt proteins . Under these conditions , the proteins eluted as single peaks consistent with tetramers of ~200 kDa , respectively . The propensity of MG491 to form tetramers was also assessed and confirmed by crosslinking with glutaraldehyde [38] and by nano-ElectroSpray Ionization Mass Spectrometry ( S3 Fig ) . Crystals of the full length MG491 , MG491-Nt and seleno methionine-labeled MG491-Nt ( respectively at concentrations of 10 mg/ml , 8 mg/ml and 15 mg/ml ) , were grown at 20°C by the vapour-diffusion method over a reservoir containing 0 . 2 M lithium sulphate monohydrate , 25% ( w/v ) PEG 3350 and 0 . 1 M Bis-Tris ( pH 6 . 5 ) or 0 . 1 M HEPES ( pH 7 . 5 ) or 0 . 1 M Tris-HCl ( pH 8 . 5 ) . Before data collection crystals were transferred to a drop of reservoir solution containing 15% ( v/v ) propylene glycol as cryoprotectant and flash-cooled in liquid nitrogen . Crystals of MG491-Nt , soaked for 10 to 60 sec in a drop of mother liquor containing 12 . 5–100 mM of 5-amino-2 , 4 , 6-triiodoisophthalic acid ( I3C , Sigma ) , were then rapidly back-soaked [39] in a drop of mother liquor containing 15% ( v/v ) propylene glycol as cryoprotectant and flash-cooled in liquid nitrogen . X-ray diffraction data was collected at 100 K on beamlines ID23-1 [40] and ID29 [41] ( ESRF , Grenoble , France ) for crystals of the full length MG491 and seleno methionine-labeled MG491-Nt proteins and on beamline PROXIMA1 ( SOLEIL , Gif-sur-Yvette , France ) for crystals derivatized with the I3C compound . All beamlines used were equipped with PILATUS 6M-F detectors [42] . For an optimal measurement of the anomalous differences on the seleno methionine-labeled MG491-Nt crystals , a MiniKappa goniometer mounted on beamline ID29 ( ESRF , Grenoble , France ) was used to re-orient the investigated crystal before data collection , aligning a crystallographic axis along the rotation axis such that Bijvoet mates were on the same image [43] . Data were integrated with XDS [44 , 45] , the output unmerged XDS ASCII file reflection . HKL was then converted to MTZ format by COMBAT and a list of free reflections generated ( CCP4 Program Suite v6 . 4 . 0 ) . The resulting reflection files were finally scaled with SCALA [46 , 47] . Phasing statistics for each data set containing anomalous differences were assessed with the processing software XDS , SCALA , XPREP ( Bruker AXS Inc . , Madison , Wisconsin , USA . ) or SHELXC from the SHELX suite [48 , 49] . All crystals from the different protein constructs belonged to the orthorhombic space group P21212 , with unit cell parameters in the range of a = 96–98 Å , b = 107–112 Å and c = 62–70 Å , indicating an important non-isomorphism not only between native and derivative crystals but also between different derivative crystals ( Table 1 and S1 Fig ) . The HKL2Map GUI interface [50] was used to run the SHELX triad . Initial maps , obtained from the seleno methionine-labeled MG491-Nt data set with the highest anomalous signal , were improved by extensive density modification procedures including averaging between the less isomorphous crystals with programs DM and DMMULTI [51 , 52] . The command-line utility phenix . get_cc_mtz_mtz , from Phenix suite [53] , which uses RESOLVE [54] , was used to facilitate comparisons between density maps with origin shifts compatible with the space group symmetry . The model was completed and refined in rounds of manual rebuilding and restrained refinement with REFMAC [55] , using TLS and isotropic B-factors only in the final stages of refinement . The quality of the final model was validated using MolProbity [56] and PROCHECK [57] ( Table 1 ) . Interacting surfaces were analyzed with Pymol ( The PyMOL Molecular Graphics System , Version 1 . 5 . 0 . 4 Schrödinger , LLC ) and the electrostatic representation was generated with the APBS plug-in . Deviations from perfect C2 and C4 cyclic symmetry were calculated for the Cα atoms as the interatomic distances differences ( null when the symmetry is perfect ) between pairs of subunits [58] . The pMTnMG491cat plasmid containing a mini-transposon bearing the coding sequence of MG_491 under the control of the MG438 promoter was amplified using the phosphorylated oligonucleotide P-C87SMG491/5 and the oligonucleotide C87SMG491/3 . The resulting PCR fragment was circularized by ligation of the blunt ends to obtain the pMTnMG491C87S plasmid . Similarly , pMTnMG491cat plasmid was amplified using the oligonucleotide FFAAMG491/5 and the phosphorylated oligonucleotide P-loop/3 . The amplicon was circularized by ligation to obtain the pMTnMG491FA plasmid . The pMTnMG491cat plasmid was also amplified using the oligonucleotide loop/5 and the phosphorylated oligonucleotide loop/3 . The PCR product was circularized by ligation to obtain pMTnMG491loop plasmid . Finally , the 855 bp 3’ coding sequence of MG_491 was amplified using oligonucleotides MG491pr438ct/5 and MG491/3 . The PCR fragment was excised with ApaI and XhoI restriction enzymes and ligated into a pMTncat plasmid [23] to obtain the pMTnMG491ΔNt plasmid . The four constructed plasmids were electroporated into Δmg491 cells and the transformants were isolated in SP-4 agar plates supplemented with tetracyclin and chloramphenicol . Transposon insertions were considered to disrupt a gene sequence when they fell within the 5'-most 80% of the ORF and were located after at least three codons from the start of the protein-coding region [3] . Total protein extracts of mycoplasma strains were electrophoresed in standard SDS-PAGE gels and stained with Coomassie Brilliant Blue or transferred electrophoretically to PVDF membranes following standard procedures [59] . PVDF membranes were probed with anti-MG217 at 1:500 dilution [60] , anti-HMW3 at 1:5 000 dilution [61] , anti-P41 at 1:1 000 dilution [62] , anti-P32 at 1:2 000 dilution , anti-MG200 at 1:5 000 dilution [24] and anti-MG219 at 1:1 000 dilution . The hemadsorption activity of M . genitalium G37 wild type and MG491 mutant strains were quantitatively determined by flow cytometry as previously described [63] using a FACSCalibur ( Becton Dickinson ) . The fraction of non-attached mycoplasma cells was plotted vs the concentration of red blood cells and fitted to inverse Langmuir Isotherm curves by iteration using the KaleidaGraph software ( Synergy ) . The Wald test was used to find statistically significant differences in the dissociation constant ( KD ) of the different strains with the G37 wild type strain . Samples of mid-log phase cultures of G37 wild type strain and Δmg491-mg491cat , Δmg491-mg491C87S , Δmg491-mg491F157A-F158A and Δmg491-mg491loopL2 were 200x diluted and grown overnight on 8-well μ-slides ibiTreat ( IBIDI ) . A Δmg491-mg491ΔNt undiluted sample was also grown overnight on 8-well μ-slides ibiTreat . Culture medium was replaced with fresh pre-warmed SP-4 before observations . Cell motility was examined at 37°C and 5% ( v/v ) CO2 using a Nikon Eclipse TE 2000-E inverted microscope . Images were captured at 2 sec intervals for 2 min . The percentage of motile cells in each strain was measured from 200 single cells and the differences were considered significant when the P value <0 . 05 using a standard χ² test . The mean velocity was measured from 25 motile cells of each strain and a significant difference was considered to be a P value <0 . 05 using a standard T-test . A sample of mid-log phase culture of mg491-F157A-F158A strain was diluted 200x in SP-4 and grown overnight on 8-well μ-slides ibiTreat ( IBIDI ) . Just before visualizing cells , culture medium was replaced with fresh pre-warmed SP-4 containing Hoechst 33342 0 . 01 mg/ml . Cells were observed by phase contrast and epifluorescence in a Nikon Eclipse TE 2000-E inverted microscope . Phase contrast and DAPI ( excitation 387/11 nm , emission 447/60 nm ) epifluorescence pictures were captured with a Digital Sight DS-SMC Nikon camera controlled by NIS-Elements BR software . Samples of M . genitalium G37 wild type and mutant strains were diluted as previously described and grown overnight in SP-4 medium over coverslips at 37°C and 5% ( v/v ) CO2 . Then , coverslips were dehydrated and metalized as previously described [17] and were visualized in a Merlin scanning electron microscope ( Zeiss ) . The percentage of single cells with more than one terminal organelle and the percentage of cells with a size smaller than 0 . 35 μm were measured from 200 single cells . A significant difference was considered to be a P value <0 . 05 using a χ² test . Samples of M . genitalium G37 wild type and mutant strains were diluted as previously described and grown overnight in SP-4 medium over holey carbon-coated grids at 37°C and 5% ( v/v ) CO2 . Each grid was washed with 1x PBS supplemented with 0 . 9 mM CaCl2 and 0 . 49 mM MgCl2 ( PBSCM , Sigma ) , blotted to remove the liquid excess and immediately plunged into liquid ethane in a Leica EM CPC cryo-workstation ( Leica Microsystems ) . The grids were transferred to liquid nitrogen and kept at -179°C during image capturing in a 626 Gatan cryoholder ( Gatan ) . The grids were examined on a JEOL 2011 transmission electron microscope operating at an accelerating voltage of 200 kV . Micrographs were recorded using a Gatan USC1000 camera under low electron dose conditions to minimize damage by electron beam radiation . A moderate underfocus between -30 μm and -15 μm was used to increase the contrast of the micrographs . | Mycoplasma genitalium is one of the smallest bacteria known and a common human pathogen . M . genitalium cells present a flask-shaped morphology due to the presence of a characteristic protrusion , known as the terminal organelle , that has several key biological roles . The terminal organelle allows mycoplasmas to move on solid surfaces by a distinctive type of cellular motility that is also related to pathogenicity . In the present study we have determined the first crystal structure of a terminal organelle protein ( MG491 ) . Using the structural information , we have designed and prepared several variants of this protein . M . genitalium cells expressing the variant proteins showed striking differences with respect to the unmodified mycoplasma cells . In one of the variants , we replaced a cysteine residue by a serine , which implies the exchange of just one sulfur atom by oxygen , resulting in a drastic reduction of motility and important alterations in cell morphology . Other variants changed the speed or the frequency of the movements . These results demonstrate that MG491 plays a key role in the functioning , organization and stabilization of the terminal organelle and are also a clear example of the interplay between atomic resolution details and the highest levels of cellular organization . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"condensed",
"matter",
"physics",
"microbiology",
"mutation",
"crystals",
"materials",
"science",
"mycoplasma",
"mycoplasma",
... | 2016 | Structure-Guided Mutations in the Terminal Organelle Protein MG491 Cause Major Motility and Morphologic Alterations on Mycoplasma genitalium |
The readily releasable pool ( RRP ) of vesicles is a core concept in studies of presynaptic function . However , operating principles lack consensus definition and the utility for quantitative analysis has been questioned . Here we confirm that RRPs at calyces of Held from 14 to 21 day old mice have a fixed capacity for storing vesicles that is not modulated by Ca2+ . Discrepancies with previous studies are explained by a dynamic flow-through pool , established during heavy use , containing vesicles that are released with low probability despite being immediately releasable . Quantitative analysis ruled out a posteriori explanations for the vesicles with low release probability , such as Ca2+-channel inactivation , and established unexpected boundary conditions for remaining alternatives . Vesicles in the flow-through pool could be incompletely primed , in which case the full sequence of priming steps downstream of recruitment to the RRP would have an average unitary rate of at least 9/s during heavy use . Alternatively , vesicles with low and high release probability could be recruited to distinct types of release sites; in this case the timing of recruitment would be similar at the two types , and the downstream transition from recruited to fully primed would be much faster . In either case , further analysis showed that activity accelerates the upstream step where vesicles are initially recruited to the RRP . Overall , our results show that the RRP can be well defined in the mathematical sense , and support the concept that the defining mechanism is a stable group of autonomous release sites .
The readily releasable pool ( RRP ) of vesicles is a reference concept for studies of presynaptic function . The concept was originally proposed to explain quantitative relationships between the frequency of presynaptic action potentials and short-term depression at neuromuscular junctions [1] , but has since been used as a framework for a wide variety of central synapses . The current idea is that only a few per cent of vesicles in typical presynaptic terminals are ready to release at any given time and that at least some readily releasable vesicles are morphologically docked to the active zone and primed for release [2] . Such an organization suggests that presynaptic function might be determined by the aggregate behavior of a fixed population of stable , autonomous release sites [3–6] . The concept of a fixed population of release sites was never proven , but fits well with a wide assortment of results from excitatory hippocampal synapses [7–11] . However , the molecular biology of synaptic vesicle trafficking seems to be complicated , and at least one attempt at a comprehensive model of short-term plasticity has questioned the utility of the RRP as a useful premise [12] . More concretely , the idea that the RRP has a fixed capacity for storing vesicles is fundamental to the concept as originally envisioned [1 , 4] . And yet , estimates of RRP size at calyx of Held synapses in the medial nucleus of the trapezoid body ( MNTB ) in the brain stem vary at least 5-fold between studies , and experimental details that should be irrelevant , such as the level of extracellular Ca2+ , seem to play a key role [13–16] . On the other hand , the RRP seems to have a well-defined size at hippocampal synapses; the Ca2+-dependence of transmitter release at hippocampal synapses is instead wholly because Ca2+ controls the efficiency of the coupling between action potentials and transmitter release [7 , 9 , 10] . The reasons for differences between calyces of Held and hippocampal synapses are not clear . The extracellular Ca2+ level seems to be most relevant when RRP size is estimated from the post synaptic responses evoked by trains of presynaptic action potentials , but less relevant—or not relevant—when neurotransmitter release is driven by briefly voltage clamping the presynaptic terminal at depolarized potentials [17] . Voltage clamp depolarization depletes the RRP in 10’s of ms , and potentially involves washing out endogenous solutes that might be important for second messenger signaling . In contrast , trains of action potentials require 100’s of ms , but can be evoked without altering the intracellular milieu . Thus , sustained activity might trigger an expansion of RRP capacity at the calyx of Held via second messenger mechanisms not present at hippocampal synapses [11 , 18] . On the other hand , currently available information about calyces of Held was extracted using experimental techniques that may not be directly comparable to the techniques used at hippocampal synapses , and it is possible that operating principles are more similar at the level of basic mechanisms than presently thought . For example , an alternative explanation for the apparent Ca2+ dependence of RRP size might be that action potential trains thought to be maximal were not sufficient to completely exhaust the RRP at the lower Ca2+ levels [19] . Here we use a variety of fiber stimulation protocols at calyces of Held from 14–21 days postnatal mice to show that , indeed , the RRP capacity for storing vesicles is not influenced by extracellular Ca2+ , even during trains of action potentials lasting 100’s of ms . Discrepancies with previous studies are explained by the presence during 100Hz stimulation of a standing flow-through pool of vesicles that are reluctant to release because of low release probability , but that are nevertheless immediately releasable . A quantitative analysis of the results demonstrated that the RRP concept can be well-defined , in the mathematical sense , in a way that is largely compatible with the original ideas in [1 , 3] . Vesicles with low release probability were not envisioned in the original conceptualization of the RRP , but could be explained by several competing hypotheses that have been proposed more recently . Either vesicle priming could be sequential , in which case vesicles with low release probability would be in an immature state of priming [20–22] . Or , vesicles with low and high release probability could be recruited in parallel to separate types of release sites [23] . Our results do not determine which explanation is correct , but do yield unexpectedly fast limits on sequential priming models , and show that parallel models are mathematically more parsimonious .
A primary aim was to monitor the changes in the rate of release that occur during train stimulation , including changes in so-called asynchronous release which is not tightly synchronized to individual action potentials [19] . To achieve this , recordings were acquired and digitized without gaps , and baselines were calculated from the 100ms interval preceding stimulation trains ( Fig 1B , upper left trace ) instead of the more usual method of calculating the baseline separately for individual responses . Stimulus artifacts were removed over windows lasting 1ms ( Fig 1B , upper right ) , which could have been problematic if the artifacts included long-lasting tails that extended outside of the window that was removed . To control for this , we calculated difference traces by subtracting traces recorded in 25μM CNQX from the traces in 1mM KYN . The procedure yields pure AMPA-type glutamate receptor responses without stimulus artifacts because CNQX is an AMPA receptor antagonist ( Fig 1A ) . In practice , a small residual component of the stimulus artifact remained in the difference traces ( Fig 1A , red trace ) , but could be eliminated by removing a narrower window of 0 . 5ms or less ( Fig 1B , red trace ) , and even when not eliminated did not contribute to measurements of the current integral because the positive and negative components canceled each other . Comparisons between responses from the difference traces and traces in 1mM KYN showed that any distortions caused by stimulus artifacts were not significant if present at all ( Fig 1C and 1D ) . This control experiment additionally ruled out non-linear contributions of glutamate uptake currents and ephaptic transmission; glutamate uptake currents were already shown to be absent from principal neurons of the MNTB at similar developmental stages in rats [26] . EPSCs during 300Hz stimulation depressed to a low steady state size within 100ms ( 30 action potentials; Fig 1C ) . The timing of depression was much slower than RRP depletion driven more directly by photolytic Ca2+ uncaging within patch-clamped calyces , suggesting that transmitter release was not rate-limited by molecular constraints on the release machinery but by the frequency of action potentials [24] . RRP size at hippocampal synapses does not seem to be influenced by extracellular Ca2+ levels , but the situation might be different for the calyx of Held . More specifically: the amount of neurotransmitter released by individual action potentials is well-known to increase when extracellular Ca2+ is increased at every synapse type , but—at least at hippocampal synapses—the total amount released by procedures that exhaust the RRP is constant . In contrast , some procedures thought to exhaust the RRP or an immediately releasable subdivision at the calyx of Held released more neurotransmitter when extracellular Ca2+ was higher , suggesting that elevating Ca2+ increases the capacity for storing vesicles [13 , 15] . Such results might indicate a qualitative difference between calyces and hippocampal synapses . However , an alternative explanation with some already published support would be that the 100Hz stimulation used to elicit release in the previous studies was not sufficient to exhaust the RRP completely at standard Ca2+ levels [17 , 19] . To determine if estimates of RRP capacity continue to depend on extracellular Ca2+ when stimulation was 3-fold faster , we compared the sum of postsynaptic responses during 300Hz stimulation ( 45 presynaptic action potentials in 150ms ) in 2mM , 4mM , and again in 2mM Ca2+ ( Fig 2A ) ; KYN was 2mM throughout . Trials were conducted in sets of 3 identical repetitions , with interleaved rest intervals of 1min , and the digitized traces for each set were averaged together before further analysis . Stimulus artifacts were removed as above and data were only accepted for further analysis if the sum of EPSCs recovered to within 5%; 2 of 7 preparations were discarded because reversal was not achieved . For analysis , a global baseline was subtracted as for Fig 1C and 1D and traces were divided into 45 sequential 3 . 33ms segments corresponding to the interval between pulses of stimulation . Responses were then quantified by calculating the integral of each segment ( Fig 2B ) . The response to the first pulse of stimulation was 2 . 0 ± 0 . 1-fold larger ( n = 5 ) in 4mM Ca2+ and subsequent depression occurred more rapidly so that the responses to pulses 15–30 were slightly smaller ( Fig 2B , magenta ) . However , the cumulative response was not different in 4mM Ca2+compared to in 2mM ( i . e . , nominally 1 . 02 ± 0 . 02-fold larger; Fig 2C and 2D ) . This result confirms that 300Hz stimulation is sufficient to exhaust the entire RRP when extracellular Ca2+ is 2mM , and suggests that the capacity of the RRP for storing vesicles is constant; we emphasize that the RRP measured here includes both slow- and fast-releasing subdivisions . We estimated that the RRP at the start of stimulation contained a mean of 2553 ± 343 ( n = 18 ) synaptic vesicles; the coefficient of variation across preparations was 55% . The estimate was calculated by dividing the cumulative response during 300Hz stimulation by the average charge transfer of spontaneous miniature EPSCs ( mEPSCs; 31 . 4 ± 2 . 1fC ) . The cumulative response was first corrected for the recruitment of new vesicles to the RRP during ongoing stimulation using Eqs 1 and 2 introduced below as part of a more detailed analysis; estimates generated using the back extrapolation method developed in [13] produced slightly lower values , but the more detailed analysis explains why using Eqs 1 and 2 is likely more accurate . The mEPSCs were measured over 10–30s of continuous recording before adding KYN and starting the experiment , and were therefore scaled by 0 . 13 for experiments conducted in 1mM KYN or 0 . 06 for experiments in 2mM KYN; preparations where the smallest mEPSCs could not be distinguished easily from noise were excluded . The mean quantal content and variation were similar to previous estimates from patch-clamped calyces where release was elicited ∼10-fold more quickly by step depolarizations that allowed massive Ca2+ influx via voltage gated ion channels [17] . The results of the quantal content analysis thus fit well with the idea that the RRP measured with trains of action potentials at 300Hz is the same quantity released by presynaptic step depolarizations . The studies using step depolarizations showed that once initiated , the rate of release does not decay away with a single exponential time course as would be expected if all readily-releasable vesicles undergo exocytosis with the same probability of release ( hereafter denoted by pv for probability of release per available vesicle within the RRP ) . Instead , the time course has multiple phases , which motivated the current concept that the RRP is made up of distinct slow-releasing and fast-releasing subdivisions; slow- and fast-releasing subdivisions have previously been termed SRP for Slow Releasing Pool and FRP for Fast Releasing Pool [28] . We therefore reasoned that the greater amount of release in elevated vs standard Ca2+ seen previously when action potential trains were 100Hz , and confirmed below , would be consistent with the results in Fig 2 if 100Hz was not intense enough to completely empty a slow-releasing subdivision of the RRP at the standard Ca2+ level . To explore this possibility and related alternatives , we performed frequency jump experiments where the frequency of stimulation was abruptly increased to 300Hz after inducing a steady state level of depression at 100Hz . Frequency jump experiments have been conducted previously at the calyx of Held , but at lower frequencies for a different purpose [29]; however , see [9 , 11] for frequency jump experiments conducted for the same purpose , but at hippocampal synapses . Sequential data processing is shown in Fig 3A–3C for interleaved trials where frequency jumps were initiated after both 500ms ( blue ) and 750ms ( magenta ) of 100Hz stimulation; we additionally interleaved trials where stimulation was 300Hz for 200ms for later comparisons ( Fig 3A and 3B , black ) . Values plotted in Fig 3B were obtained by integrating over sequential segments of 3 . 33ms duration . Only every third segment contained synchronous responses during 100Hz stimulation because the inter-stimulus interval was 10ms; the smaller values making up the lower of the double horizontal lines are measures of the asynchronous component of responses occurring more than 3 . 33ms after the individual pulses of stimulation . Plotted this way , individual responses can be seen to depress to a first plateau during 100Hz stimulation , and then to a second plateau that is lower during subsequent 300Hz stimulation . The result indicates that the quantity of neurotransmitter release elicited by individual action potentials was less at 300Hz , in-line with the expectation that rate-limiting steps in recruitment of new vesicles to the RRP played a role after 500ms of 100Hz stimulation [13] . The values in Fig 3C were obtained by integrating over segments of 10ms instead of 3 . 33ms , which provides a more direct comparison of release as a function of time for stimulation at 100 vs 300Hz . Direct comparisons are valid because stimulus artifacts were eliminated by removing 1ms windows every 3 . 33ms , even for the baseline and segments during 100Hz stimulation . The values of the segment integrals were larger after switching to 300Hz stimulation than during 100Hz stimulation—even though the release per individual action potential was less—because each segment contained responses to 3 action potentials . To quantify the increase without making assumptions about mechanism , we calculated an index by dividing the sum of values from the first 150ms after increasing the stimulation frequency ( blue points in dashed box in Fig 3C ) by the sum of values from the first 150ms of the trials where stimulation was 300Hz throughout ( box in Fig 3B; leftmost solid bar in Fig 3D; n = 20 ) . A baseline value for the index was calculated by dividing the sum of the matching values from trials where the stimulation was maintained at 100Hz ( magenta points in dashed box in Fig 3C ) by the sum of points in the box in Fig 3B . The baseline value ( leftmost open bar in Fig 3D ) was significantly less than for frequency jumps , confirming that increasing the stimulation frequency to 300Hz increased the rate of release . Stimulus artifacts did not play a role because an identical analysis of difference traces calculated as in Fig 1 produced a similar result ( third and fourth bars in Fig 3D ) . Difference traces were only available for a subset of preparations; for these , experimental trials were followed with matched trials in the presence of either 4mM KYN ( n = 3 ) or 25μM CNQX ( n = 4 ) . We additionally conducted analogous frequency jump experiments in 4mM Ca2+ ( Fig 3E and 3F ) . The idea was that increasing extracellular Ca2+ would increase the fraction of the RRP released by individual action potentials , which would lead to more RRP depletion . As predicted , the increase in release elicited by frequency jumps was less ( compare Fig 3F to Fig 3C ) ; the index of increase was midway between the baseline value and the value in 2mM Ca2+ ( Fig 3D , compare bars 1 and 5 ) . The indices are directly comparable because of the result , above , that the time-integrated response during the first 150ms of 300Hz trials was the same in 2mM and 4mM Ca2+ ( see Fig 2C and 2D ) . This result confirms that the increase in release elicited by frequency jumps is caused by release of transmitter from a readily-releasable supply that was not released during 100Hz stimulation . Most of the increase in the rate of release seen at both 2 and 4mM Ca2+ was transient , confirming that 100Hz stimulation leaves a residual supply of readily releasable vesicles that can be induced to undergo exocytosis by increasing the frequency to 300Hz . Further analysis using Eqs 1 and 2 introduced below indicated that 100Hz stimulation depleted the RRP: 87% ± 3% when Ca2+ was 4mM; 79% ± 3% when Ca2+ was 2mM; and 61% ± 8% ( n = 4 ) when Ca2+ was 1 . 2mM , which is at or below the level in vivo[30] and the lower level used in [15] ( see S1 Fig ) . The rate of release during the frequency jump experiments did reach a new steady state after 120ms of 300Hz stimulation that was elevated compared to the steady state during 100Hz stimulation . The elevated steady state suggests that bulk recruitment of new vesicles was faster during 300Hz stimulation ( Fig 3C , green line ) . The elevation is in line with multiple mechanisms , including a likely increase in the number of vacancies within the RRP and possible activity-dependent acceleration of the mechanism underlying vesicle recruitment ( see Lemma 7: Second order corrections , which is introduced below as part of the more detailed analysis ) . The amount of increase was similar when the frequency jump was initiated 250ms later , after 750ms of 100Hz stimulation ( 96 ± 3% when Ca2+ was 2mM; Fig 3G ) . For this comparison , increases were calculated after subtracting the responses during matched trials where the stimulation was maintained at 100Hz for the full 950ms , which was necessary for a high precision analysis because of slowly-developing fatigue in recruitment of new vesicles , documented below . 100Hz trials matching frequency jumps after both 500 and 750ms were available for n = 7 preparations , all in 2mM Ca2+; for these , trials of 100Hz stimulation lasting 950ms were interleaved with the two types of frequency jumps and 300Hz trials . Thus , the presence of the unreleased supply was not simply because 50 action potentials at 100Hz were too few to exhaust the slow-releasing subdivision of the RRP . Instead , the RRP was maintained at a steady-state level of fullness . This could either be because readily releasable vesicles constituted a flow-through pool where recruitment of new vesicles is fast enough to balance the quantity undergoing exocytosis . Or , the steady state supply could be completely immobile when stimulation is 100Hz , and only accessed for release when the frequency is increased to 300Hz . However , correlations presented below between paired pulse facilitation/depression and the size of the steady-state supply seem to argue against the hypothesis that vesicles remaining within the RRP during steady state 100Hz stimulation are immobile , and therefore support the concept of a flow-through pool . We found that the mean value for pv for vesicles in the steady state supply was lower than the mean value when the RRP was full . To demonstrate the difference in a way that does not depend on assumptions about ongoing vesicle recruitment , we estimated a lower bound for the mean pv for all of the vesicles in the RRP at the start of stimulation ( Fig 4A , reciprocal of y-axis intercept of brown line ) that was higher than an upper bound for the mean value for vesicles in the steady-state supply during 100Hz stimulation ( Fig 4B , reciprocal of y-axis intercept of green line; see Lemma 1 in the Methods for details ) . Indeed , the time course of decay in response size seen after the frequency jumps was clearly slower than the decay during trials where 300Hz stimulation was initiated from rest ( Fig 4C ) , as expected if pv was lower for vesicles remaining in the RRP after partial depletion with 100Hz stimulation; this result is in-line with previous frequency jump experiments at the calyx of Held conducted at lower frequencies [29] . The result is not compatible with the simplest models where all RRP vesicles always have the same pv . This was expected because the simplest models were already strongly questioned by the previous evidence that the RRP is subdivided into slow- and fast-releasing subdivisions when synapses are fully rested . Moreover , the result is consistent with the current concept that vesicle priming is sequential , whereby vesicles that are newly recruited to the RRP initially have a low pv[20]; sequential priming models include the concept of positional priming where pv for newly recruited vesicles increases over time as vesicles that are docked and molecularly primed are translocated to areas of high Ca2+-channel density [21] . However , the result is additionally consistent with the fundamentally different alternative where readily releasable vesicles dock to separate sets of release sites with intrinsically low and high pv[9 , 23] . We refer to the second possibility as a parallel model because vesicles with low and high pv would be recruited in parallel . Finally , pv for a homogeneous population of releasable vesicles might have decreased a posteriori , after the start of stimulation , owing to use-dependent fatigue of the release machinery or even inactivation of Ca2+-channels [31] . Regardless of mechanism , we use the terms “reluctant” and “reluctantly-releasing” to describe readily releasable vesicles with low pv[23 , 32] . We do not assume that pv must be the same for all reluctant vesicles , and in fact we leave open the possibility that newly recruited vesicles go through multiple stages of priming with ascending values for pv , possibly starting from pv = 0 . To maintain terminology that is consistent with previous reports we reserve the term “slow-releasing” to describe the vesicles that are found within the slow-releasing subdivision of the RRP after long periods of rest . Below we show that sequential models predict that the mechanism for low pv vesicles in the flow-through pool is different than the mechanism for low pv vesicles in the slow-releasing subdivision of the RRP , whereas parallel models predict that the mechanism is the same ( i . e . , see Lemma 7 ) . A current concept is that slow- or reluctantly-releasing vesicles are released asynchronously , with a delay or slow kinetics after the triggering action potential , possibly owing to a final priming step taking tens of ms[19 , 33 , 34] . However , we found that the increased release elicited by the frequency jumps was tightly synchronized to action potentials ( Fig 5A and 5B ) . To quantify the amount of synchronous vs asynchronous release , we calculated the integrals of 3 . 33ms segments after removing the asynchronous component , and divided them by the full integral calculated beforehand . The asynchronous component was removed by subtracting a baseline measured between 2 . 8 and 3 . 3ms after each pulse of stimulation ( illustrated in Fig 5C ) . Only difference traces were analyzed because the stimulus artifacts in the raw data occluded the baseline window ( n = 7 , as noted above ) . The measurement would underestimate the synchronous fraction if the time courses of individual responses did not run to completion before the subsequent action potential , which seems likely during 300Hz stimulation . Even so , the synchronous fraction declined only a small amount , at most from 84% ± 3% during 100Hz stimulation to 77% ± 4% over the first 50ms after increasing the stimulation frequency to 300Hz ( Fig 5B ) . Meanwhile , the summed response was 2 . 7 ± 0 . 2-fold larger ( Fig 5C and 5D ) . Thus , the frequency jumps transiently increased the synchronous component of release approximately 2 . 5-fold , and more than 75% of the vesicles with low pvwere released synchronously . The results in Fig 2 argue against the idea that the capacity of the RRP for storing vesicles is a dynamic function of Ca2+ . In contrast , the results are in-line with the idea that the capacity is constant , which would occur if the RRP were made up of a fixed number of autonomous release sites that could be depleted of vesicles during heavy use ( Fig 6A ) . The sum of responses during the frequency jump trials was 1 . 64 ± 0 . 04-fold larger compared to when stimulation was 300Hz throughout , even though both protocols exhausted both fast- and slow-releasing subdivisions of the RRP . A difference of some amount would be predicted by many models that have been proposed , including ones where the RRP has a constant capacity , because some neurotransmitter would have been released from vesicles that were recruited to the RRP during ongoing stimulation , and there was more time for recruitment of new vesicles during the frequency jump trials ( 700ms vs 200ms ) [16 , 35] . However , we show below that the standard models employed for estimating the capacity of the RRP and vesicle recruitment rates make predictions that are not quantitatively in line with the result . It was therefore important to test the feasibility of models where release sites are autonomous by comparing the relative amounts of release before and after the frequency jumps and when 300Hz stimulation was initiated from rest . To describe the analysis , we first distinguish between unitary and bulk concepts of recruitment and release . The unitary rate of vesicle recruitment is the fraction of vacant space within the RRP that is replenished in a given amount of time . That is , if the RRP is made up of autonomous release sites , the unitary rate would be the rate at which a release site recruits a vesicle and consequently becomes full . In contrast , bulk recruitment is the rate at which vesicles are recruited to the RRP as a whole , and is fast at the calyx of Held in part because the RRP is large—i . e . , with thousands of release sites—and in part because the unitary rate is faster than at standard synapse types , which is shown below . By definition , the unitary recruitment rate equals the bulk rate divided by the capacity for storing vesicles when the RRP is empty; for Fig 6A the capacity would be the number of release sites . However , the bulk rate is only guaranteed to be linearly related to the unitary rate when the RRP is empty because a key consequence of models such as in Fig 6A is that new vesicle recruitment would be blocked from release sites that were already full . Likewise , the unitary rate of release is the fraction of the vesicles within the RRP that are released in a given unit of time , which is equivalent to pv multiplied by the stimulation frequency . Meanwhile , the bulk release rate is the aggregate rate of release from the entire RRP and is not necessarily related to the unitary release rate in a straightforward way; e . g . , at hippocampal synapses , at least , the bulk rate can be depressed owing to RRP depletion at the same time when the unitary rate is enhanced by residual Ca2+-dependent mechanisms [11] . A single release site is depicted by the Markov chain in Fig 6B where the site switches between two states , either filled with a vesicle ( F-state ) , or empty ( E-state ) . The unitary recruitment rate is depicted as αi , t ( where i is an index that identifies each release site and t is time ) . βi , tis the unitary rate of release . Although the diagram is simple , the model incorporates enough flexibility to reproduce the behavior of all models , sequential or parallel , that are based on the premise that readily releasable vesicles are docked and primed at a fixed number of autonomous release sites . That is , allowing βi , t to vary in time provided enough flexibility to account for sequential transitions from low to higher pv stages of vesicle priming and other mechanisms that affect pv in both positive and negative directions such as paired-pulse facilitation and Ca2+-channel inactivation [11 , 36–38] . Allowing βi , t to vary across release sites was necessary for parallel models to account for the slow and fast-releasing subdivisions of the RRP and for sequential models to suppress the assumption that vesicle recruitment and subsequent sequential priming transitions are synchronized across release sites . Meanwhile , allowing αi , t to vary in time and across release sites allowed for possible activity-dependent acceleration of the vesicle recruitment mechanism [39] , and possible heterogeneity across release sites [23] . In any case , merely allowing the value of αi , t to change in time and to vary across release sites would not exclude the special cases where αi , t is constant in time or across release sites ( but see below ) . We refer to general models , such as in Fig 6B , that can reproduce the behavior of entire categories of specific models , as sparse models . Specific models make predictions that are more precise , but often depend on assumptions that have not been verified . In contrast , sparse models can be used to elucidate boundary conditions that must then apply to a wide range of specific models . In the present case , release and new vesicle recruitment to the RRP could be modeled as: d R R P t d t = α ^ t · ( R R P 0 - R R P t ) - β ^ t · R R P t ( 1 ) where RRP0 is the number of release sites , RRPt is the number that are occupied at time t , α ^ t is the mean value of αi , t for all release sites that are empty , and β ^ t is the mean βi , t for all sites that are full . Eq 1 was derived from the model in Fig 6B , and therefore , any boundary conditions established by the equation would be inherited by all models that satisfy the initial premises; see Lemma 2 in the Methods for the derivation and confirmation that sequential models are covered even when priming occurs through discrete states , such as in Fig 6C . The initial goal was to use Eq 1 to divide cumulative responses during train stimulation into two fractions: the response generated by releasing all the transmitter in RRP0; and the response generated by releasing ∫ α ^ t · ( R R P 0 - R R P t ) , which is the transmitter recruited during the stimulus train and is referred to below as the cumulative recruitment . The idea is that the response generated by releasing RRP0 would be the same for all stimulation protocols , and therefore , mismatching estimates for RRP0 from frequency-jump trials compared to when 300Hz stimulation was initiated from rest would rule the model out . We did not attempt to estimate specific values for α ^ t or β ^ t a priori . Instead , we started with the special case where α ^ t is constrained to be some constant , referred to below as α ^ f i x e d; the value of α ^ f i x e d was not specified a priori , but a unique value was determined by the data ( see below ) . Although recognized beforehand as a potential oversimplification , the special case was a convenient starting point because we showed previously that it can be used to extract unique values for α ^ f i x e d , β ^ t , and RRP0 from experiments where the RRP is exhausted , or at least when driven into a partially empty steady state [9] . The change from α , β , N , and n in the previous study to α ^ f i x e d , β ^ t , RRP0 , and RRPt here is purely notational and does not alter the mathematical relationships that were derived previously . The analysis involved finding the unique value for α ^ f i x e d where R R P 0 = R s s · ν α ^ f i x e d when Rss is the release per action potential at steady state and ν is the frequency of action potentials . To accomplish this , we used computer simulations to calculate the predicted amount of bulk recruitment for a range of values for α ^ f i x e d ( see Methods ) . We then chose the value where the difference between the cumulative release and cumulative recruitment over the entire train was equal to R s s · ν α ^ f i x e d . The mathematically more precise procedure used in [9] produced the same results , but simulations were used in the present study because they allowed flexibility needed below for modeling activity-dependent acceleration of the vesicle recruitment mechanism . The analysis produced values for α ^ f i x e d of 4 . 65/s for the 300Hz trains and 4 . 70/s for the frequency jump trials , which are in-line with previous estimates at the calyx of Held [40] . The portion of responses generated by release of newly recruited vesicles for the data plotted in Fig 7A are demarcated by blue lines . We emphasize that the values for α ^ f i x e d pertain specifically to vesicle recruitment to the RRP , not to any subsequent transitions to higher pv stages of vesicle priming , which would be downstream . Instead , the timing of downstream priming transitions would influence β ^ t , and some information could have been extracted at this point in the analysis ( e . g . , see Figure 2A of [41] ) . However , interpreting the information in the context of sequential priming would have required making additional assumptions , which we sought to avoid . In any case , the value of RRP0 estimated from frequency jump trials was 14 ± 3% less than from trials where 300Hz stimulation was initiated from rest ( blue bars in Fig 7B ) , ruling out models covered by the general framework where the unitary rate of vesicle recruitment to the RRP is constant . For comparison to previous studies , the green lines in Fig 7A are from a back extrapolation analysis , which is a standard in the field , that implicitly assumes that the bulk rate of recruitment—as opposed to the unitary rate—is constant throughout [13]; Fig 7B shows that the mismatch was similar for this method ( 17 ± 3% , green bars ) , ruling out models where the bulk rate is constant . Finally , ignoring the contribution of recruitment until after achieving a steady state ( brown lines in Fig 7A ) results in an overestimation of 32 ± 4% ( brown bars in Fig 7B ) , ruling out models where vesicle recruitment does not occur until the RRP is completely empty . The mismatches between the two estimates for RRP0 could be eliminated by allowing α ^ t to accelerate from an initially low value as hypothesized in [39] . This works because acceleration lessens estimates of bulk recruitment during the early part of trains ( magenta lines in Fig 7A ) —and consequently increases estimates of RRP0—and the effect is disproportionately larger when the acceleration is induced more slowly , as would be expected during the 100Hz trains at the beginning of the frequency jump trials compared to when stimulation was 300Hz from the outset . As proof of the concept that incorporating recruitment rate acceleration could resolve the discrepancy , we modeled the acceleration as the single rising exponential: α ^ t = α ^ m a x · 1 - e - t τ ( 2 ) where τ is a free parameter that could be manipulated to model the time course of engagement of the acceleration mechanism , and α ^ m a x is a maximum value that is determined by the data . The idea was that each action potential would increase α ^ t by a constant fraction up to α ^ m a x; α ^ t was already known to approach some maximum because response sizes during 300Hz stimulation settle to a steady state and do not increase after the RRP has been exhausted ( e . g . , Fig 1C ) , as would occur otherwise [9] . We found that estimates of RRP0 from the frequency jump trials and trials where 300Hz stimulation was initiated from rest were equal when τ = 10 ν , where ν is the frequency of stimulation ( magenta bars , left panel of Fig 7B ) . Lower values for τ were not sufficient and higher values produced an overcorrection . More complicated models of the time course of acceleration could resolve the discrepancy equally well . However , a key point is that the acceleration mechanism must increase the rate at which vesicles are recruited to the RRP . No amount of acceleration of downstream transitions to priming stages with higher pv could substitute . This is notable because some other experimental paradigms and analysis , documented previously , were not able to distinguish clearly between acceleration at the step of recruitment and downstream effects [42] . The distinction is possible in the present case because acceleration of vesicle recruitment is fundamentally different from acceleration of downstream priming steps for models covered by Eq 1 where only vacant release sites recruit new vesicles; release sites occupied by vesicles occlude recruitment , even when pv = 0 . An analogous analysis of the experiments conducted in 4mM Ca2+ also required recruitment rate acceleration , but in this case τ = 5 ν , whereas τ = 10 ν produced an overcorrection ( magenta bars , right panel of Fig 7B ) . The lower value for τ indicates that fewer action potentials would be required to accelerate the recruitment mechanism when extracellular Ca2+ is higher , which is in-line with the previous studies indicating that recruitment rate acceleration is mediated by Ca2+ influx for at least one other synapse type [43] . Unexpectedly , when acceleration was incorporated , α ^ m a x was fixed by the data to a lesser value for the frequency jump trials than for the trials where 300Hz stimulation was initiated from rest ( 3 . 9/s vs 4 . 3/s; the numbers correspond to experiments where Ca2+ was 2mM ) . If the modeling framework is correct , the discrepancy would indicate that vesicles were recruited ∼10% more slowly during 300Hz stimulation at the end of the frequency jump trials than during the trials where 300Hz stimulation was initiated from rest . In other words , the recruitment mechanism must have fatigued a small amount during 500ms of 100Hz stimulation . To verify this , we first compared the steady state response size during 300Hz stimulation after frequency jumps to the steady state when 300Hz stimulation was initiated from rest . The size was 9 . 9 ± 2 . 8% lesser for the frequency jump trials , which was predicted by the model because the size of the responses that continue to be evoked when the RRP is exhausted would be proportional to the rate at which new vesicles are recruited [9] . Additional experiments showed that the fatigue persisted long enough to slow RRP replenishment during subsequent rest intervals . We stimulated calyces with pairs of 300Hz trains separated by 1s-long rest intervals before ( black in Fig 7C ) and immediately following ( red ) 100Hz conditioning trains lasting 5s ( see diagram atop Fig 7C ) . An index of the replenishment occurring over the rest intervals was calculated by dividing the sum of responses during the second 300Hz train of each pair by the sum of responses during the first 300Hz train of the pair that was initiated without prior 100Hz conditioning . The index was 52 ± 5% after 100Hz conditioning compared to 76 ± 1% ( n = 4 , p < 0 . 02 ) without conditioning , which is a decrease of 31 ± 7% . The decrease was more than the 10% measured above likely because 100Hz conditioning was 5s instead of 500ms . Taken together , the identification of fatigue lends support to the modeling framework defined by Eq 1 because the predictions were confirmed with two orthogonal types of experiments that could both be interpreted independently of any model . A related phenomenon may have been identified previously in rats for an earlier developmental stage using a different experimental technique [44] . The fatigue in vesicle recruitment could be incorporated into the framework defined by Eq 1 in a variety of ways , but could not be manipulated to eliminate the prediction that activity accelerates the unitary recruitment rate , α ^ , or even to substantially alter the estimate of τ = 10 ν when acceleration was modeled with Eq 2 ( Fig 7D ) . Although previous studies identified mechanisms that accelerate the recruitment of vesicles from the slow- to fast-releasing subdivisions of the RRP , α ^ pertains to the upstream step , where vesicles are initially recruited to the RRP as a whole . To our knowledge , acceleration specifically at the upstream step had not been demonstrated previously for the calyx of Held; the already identified acceleration would instead influence β ^ t in Eq 1 . Therefore , to verify the prediction that activity accelerates the rate at which vesicles are recruited to the RRP as a whole , we measured the time course of RRP replenishment during rest intervals that followed action potential trains . We used pairs of 300Hz trains separated by experimentally varied rest intervals as diagrammed in Fig 8A and 8B . Each train was 150ms at 300Hz ( 45 action potentials ) to ensure that both fast and slow-releasing subdivisions of the RRP were exhausted . The fractional amount of RRP replenishment during each rest interval was calculated by dividing the response integral during the entire second train by the integral during the first . Recruitment of vesicles during ongoing stimulation was factored out by including interleaved trials where the rest interval between trains was nominally zero ( i . e . , 3 . 33ms ) . We reasoned that a mechanism that accelerates vesicle recruitment during trains of action potentials would disengage during subsequent rest intervals , causing recruitment to slow down . In contrast , if acceleration did not occur , the unitary recruitment rate would be maintained at a constant during rest intervals , and the RRP would replenish more than 98% during the first 1s alone; that is , R R P t = 1 - e - α ^ m a x · t where RRPt is the fractional fullness of the RRP at time t and α ^ m a x = 4 . 3 / s from above . However , full replenishment took much longer than 1s ( Fig 8A–8C ) , supporting the prediction from the general model defined by Eq 1 that the recruitment rate was accelerated by the preceding activity . At a more quantitative level , the time course of RRP replenishment could not be approximated with any single exponential function ( i . e . , one with a constant unitary rate ) ; this is in-line with previous measurements [28 , 45 , 46] . However , the full RRP replenishment time course was closely approximated by: R R P t = 1 - e - ∫ α ^ t ( 3 ) where α ^ t is the decaying double exponential: α ^ t = ( α ^ m a x - α ^ ∞ ) · [ w · ( e - t τ f ) + ( 1 - w ) · ( e - t τ s ) ] + α ^ ∞ ( 4 ) and w = 0 . 95 , τf = 50ms , τs = 7s , α ^ m a x = 4 . 3 / s , and α ^ ∞ = 1 12 / s ( magenta dashed line in Fig 8C ) . Eq 3 is relevant because it was derived from Eq 1 by assuming that: β ^ t = 0 , as expected during rest intervals; RRP0 = 0 because the RRP was empty at the beginning of the rest interval; and RRPt → ∞ = 1 . 0 because of the way the replenishment values in Fig 8C were normalized [9]; α ^ ∞ would be the baseline unitary recruitment rate expected in the absence of activity . In this case , Eq 4 would describe the time course over which the acceleration mechanism disengages during rest intervals . If so , disengagement at the calyx of Held would be much faster than at excitatory hippocampal synapses . Nevertheless , the τf = 50ms value was in-line with expectations because disengagement at excitatory hippocampal synapses followed the clearance of residual Ca2+ , which is likely much faster at the calyx of Held [47]; the time course of disengagement at hippocampal synapses was measured in [43] and Ca2+ clearance in [11] . In sum , the measurements of RRP replenishment during rest intervals and the results from the frequency jump experiments are both in-line with the prediction that activity accelerates the recruitment of vesicles to the RRP [39] . The logic is based on the assumption that the general model defined by Eq 1 is accurate , but further analysis did not yield any alternatives where the requirement for acceleration of the recruitment mechanism could be avoided . Notably , one set of alternatives where vesicles in the slow-releasing subdivision of the RRP are occluded from docking at release sites by vesicles already in the fast-releasing subdivision [48] , could account for the results from the frequency jump experiments with a mechanism that accelerates the downstream transition from the slow- to fast-releasing subdivision with no need for accelerating the mechanism that recruits vesicles to the slow-releasing subdivision . However , the results from the RRP replenishment experiments in Fig 8C are not compatible with these alternatives . That is , further analysis produced model-independent lower and upper bounds for the unitary rate of vesicle recruitment to the RRP of 3 . 21/s ± 0 . 15/s and 4 . 91/s ± 0 . 37/s during 300Hz stimulation ( brown and green lines in Figs 4A and 7A , respectively ) . Even the lower bound would predict that the slow-releasing subdivision would replenish more than 96% during the first 1s in the absence of a disengaging acceleration mechanism . Meanwhile , we reasoned that the fast-releasing subdivision would have to remain nearly completely empty during the 1s interval to account for the <40% RRP replenishment overall because the slow- and fast-releasing subdivisions each make up approximately half of the total [17] . But , an empty fast-releasing subdivision is not compatible with the observations that responses to isolated action potentials recovered as quickly as the RRP as a whole because isolated responses would primarily track replenishment of the fast-releasing subdivision ( compare squares to circles in Fig 8C ) . We did observe a small increase in the paired pulse ratio after short rest intervals ( Fig 8D ) , which is consistent with a transient decrease in the mean value for pv , but even this effect was no longer detectable after 6s of rest when RRP replenishment was still far from complete . Taken together , the results in Fig 8 support the general model because acceleration of the recruitment mechanism is an unavoidable prediction when the general model is applied to the results of the frequency jump experiments , which are orthogonal experiments , but not when alternatives to the general model are applied . Notably , the mechanism that causes this type of acceleration is likely distinct from the calmodulin-dependent and actin-dependent mechanisms that are thought to accelerate downstream steps in sequential priming because blockers largely abolished synchronous release but seem to have a relatively minor impact on the overall rate of recruitment and subsequent release during maximal stimulation [28 , 34] . Both sequential and parallel models predict that the mean dwell time for readily releasable vesicles would be ∼115ms during 100Hz trains of action potentials; a longer dwell-time would produce a higher steady state level than was seen ( Lemma 5 ) . In the context of sequential models , this implies that the unitary rate for traversing the complete set of sequential transitions from the initially low value of pv to pv ≫ 6% would be ∼9/s . The value is 35-fold faster than estimated previously during rest intervals , but consistent with the evidence that sequential transitions are accelerated by activity via mechanisms involving calmodulin , and actin [22 , 34] . On the other hand , parallel models are also fully compatible with the results above . In this case , activity-dependent enhancement mechanisms such as facilitation , augmentation , and post-tetanic potentiation would increase pv , much like sequential priming except the transitions would reverse during rest intervals [3 , 11 , 37 , 38 , 49] . The requirement for activity-dependent acceleration of α ^ t established above would continue to apply; in particular , the requirement could not be explained by disproportionately faster recruitment to release sites with low pv ( Lemma 6 ) . Indeed , parallel models were more parsimonious because of a hard requirement from two orthogonal sets of experiments that low and high pv release sites would be present in near equal proportions both when the RRP is nearly exhausted and when full . That is , the combination of the size of the steady state supply during 100Hz stimulation and the timing of recruitment of vesicles to the RRP extracted from the frequency jump experiments in the present study forces parallel models to predict that release sites with low and high pv are present in near equal proportions ( Lemma 7 ) ; this would be the average for all calyces , see below for evidence of variation between preparations . The prediction matches direct measurements of the slow- vs fast-releasing subdivisions of the RRP when the RRP is full [17] . The accurate prediction based solely on results from the frequency-jump experiments is remarkable because it is extrapolated from the steady state fullness of the RRP when ∼80% depleted , and is independent of the details of short-term depression induced by action potential trains that were initiated when the RRP was full . An incorrect prediction would have ruled out parallel models , whereas sequential models must have at least one additional degree of freedom which could be tweaked to maintain compatibility with a broad range of possible outcomes . Independently of whether vesicles with high pv are primed sequentially from a state with low pv , or in parallel at a separate type of release site , the framework defined by Eq 1 could be used to derive the mean value for pv when the level of RRP fullness was in a steady state during 100Hz trains ( Lemma 3 ) . This value could then be combined with information about α ^ t to confirm the simultaneous presence of vesicles with a variety of values for pv ( Lemma 8 ) . The concept of heterogeneity in pv is already widely accepted , but the evidence that the heterogeneity occurs at a specific point in time is new and notable because it argues against the special case of parallel models where the slow component of release seen during maximal stimulation results from fatigue in the release machinery instead of from depletion of a slow-releasing subdivision of the RRP; these have been termed a posteriori models in [31] . Finally , results were highly reproducible when trials were repeated on the same preparation , but we observed striking variation between preparations in the size of the steady state supply during 100Hz trains compared to the size of the RRP when full . And , the paired pulse ratio at the beginning of trains was greater for preparations which later had larger steady state supplies ( Fig 9A , top ) , and the induction of depression was slower ( Fig 9A , bottom ) . The correlations pertained equally to 100Hz and 300Hz trains of action potentials ( 2 of 4 plots are not shown ) , and were not dependent on age within the 14–21 day range used here ( Fig 9 ) . Variation among calyces in the induction of depression has been reported previously [50] , but the correlation with the size of the steady state supply is new . The correlation suggests that the steady state level is mechanistically related to the size of the slowly-releasing subdivision when the RRP is completely full . Meanwhile , a frequent assumption is that low pv correlates with paired-pulse facilitation and high pv with paired-pulse depression [11 , 50–52] , which suggests that the slow-releasing vesicles found in resting RRPs are immediately releasable ( see Discussion ) . Sequential and parallel models produce fundamentally different explanations for the variation in the steady state level . For sequential models , the steady state would be a function of the timing of transitions from low to high pv ( Lemma 5 ) . In contrast , for parallel models , the steady state would be a function of the number of release sites with intrinsically low pv ( L p D 0 R R P 0 in Lemma 7 ) . For both parallel and sequential models , the steady state level would additionally be a function of the precise value of pv for reluctantly-releasing vesicles ( Lemma 3 ) and the unitary rate of recruitment to the RRP; the relevant parameter would be α ^ S S 100 in Lemma 4 for sequential models and α ^ L p D , S S 100 in Lemma 7 for parallel models . Intriguingly , for parallel models , the correlations in Fig 9A translated to correlations between L p D 0 R R P 0 and the same measurements of short-term plasticity ( Fig 9B ) . In contrast , no equivalent correlations emerged for any of the three parameters noted above when the deconvolution was performed assuming sequential priming; the information was likely distributed between α ^ S S 100 and the timing of the sequential transitions . The result is an additional indication that parallel models are more parsimonious , and suggests that long-term modulation of the numbers of release sites with intrinsic low vs high pv properties may be a significant determinant of short-term plasticity under a wide range of situations .
The current concept is that newly recruited RRP vesicles either are not releasable initially , or are only releasable with low pv; pv is then thought to increase over time as the state of priming matures . These are termed sequential priming models , which include: models where the molecular machinery for catalyzing exocytosis assembles gradually over time; and positional priming models where the rate limiting step is instead translocation to Ca2+-channels [20 , 21] . However , parallel models where vesicles with low and high pv are recruited to distinct types of release sites remain possible . Notably , as long as the fixed capacity principle is retained , all parallel and most sequential models are covered by the general model , including positional priming models where the capacity of a fast-releasing subdivision of the RRP is determined by limited availability of domains with high numbers of Ca2+ channels [21] . In contrast , some sequential models where fast-releasing vesicles occlude the transition of slow-releasing vesicles to a dedicated fast-releasing subdivision of the RRP are technically not covered , although in most cases these nevertheless behave like models that are covered when action potential frequency is high enough to exhaust the fast-releasing subdivision . Sequential and parallel models both predict that the RRP will transform into a flow-through pool containing vesicles with low pv during submaximal stimulation , which is in-line with the steady state supply identified here during 100Hz stimulation . Parallel models are more parsimonious because they reference a single mechanism to explain the presence of slow-releasing vesicles when the RRP is full and vesicles with low pv when the RRP is ∼80% depleted . In contrast , sequential models require an additional mechanism that could be spontaneous reverse transitions from high back to low pv states , as in [47] , or a limiting number of Ca2+ channels [21] . Whatever the identity , variation between preparations in the mechanism that establishes a slow-releasing subdivision within the RRP during rest intervals would have to correlate with variation in the mechanisms that determine the size of the flow-through pool to account for the results in Fig 9A . There is evidence suggesting that at least some of the heterogeneity between vesicles with low and high pv arises from intrinsic differences in release sites rather than a variety of stages in sequential priming , which is in-line with parallel models [54–56] . Nevertheless , the evidence for sequential priming is also intriguing [21 , 22 , 34] . Merging parallel and sequential ideas is possible , but only in a way that would make the sequential transition from not releasable to pv ≫ 6% faster than 9/s , which is already 35-fold faster than estimates from resting calyces of Held [22] . Models based on [57] that have been used widely to implement short-term depression in neural network simulations would be covered by the general model , but are not compatible with the present results because pv is forced to be the same for all readily releasable vesicles; this limitation was already present in the earliest models for neuromuscular junctions [1 , 4] . Other similar computational models , such as in [58] , are not strictly compatible with the premise of a constant capacity RRP because depletion does not occur more quickly after the induction of short-term enhancement , but subsequent studies have suggested that enhancement mechanisms increase pv and therefore speed depletion [8 , 37 , 38] . Activity and Ca2+-dependent acceleration of synaptic vesicle trafficking has been investigated previously at a broad range of synapse types [43 , 59 , 60] , including the calyx of Held [39] where some of the molecular biology and pharmacology is already known [22 , 28 , 34 , 39 , 46] . However , the standard experimental techniques do not automatically distinguish between acceleration of the mechanism by which vesicles are initially recruited to the RRP as a whole and acceleration of downstream mechanisms , such as sequential transitions from a slow- to fast-releasing subdivision [11 , 16 , 42] . Indeed , the calmodulin and actin-dependent mechanisms that have received most of the attention at the calyx of Held seem to be downstream mechanisms [22 , 28 , 34] . However , to our knowledge , information about whether or not activity additionally accelerates the upstream step where vesicles are initially recruited to the RRP was only previously available for excitatory hippocampal synapses [42 , 43] . Nevertheless , we show here that the initial recruitment step is accelerated many fold by activity as originally predicted [39]; the molecular mechanism remains to be identified , but presumably does not involve calmodulin because inhibitors of calmodulin dramatically alter the kinetics of release but do not effect much recruitment to the slow-releasing RRP subdivision [34] . The emerging conceptual similarities between the calyx of Held and Schaffer collateral synapses are remarkable , but we did find notable differences in parameter values . ( 1 ) Recruitment of new vesicles to the RRP during maximal use was ∼15-fold faster at the calyx of Held ( 4 . 3/s here vs 0 . 24/s at matching temperature in [9] ) . ( 2 ) Disengagement of the acceleration mechanism during rest intervals was two orders of magnitude faster ( 50ms here vs 10s in [43] ) . And ( 3 ) mean pv for all vesicles in the RRP when rested was higher ( 11 . 8 ± 1 . 2%—n = 20—vs 4 . 4% in [9] ) . In fact , mean pv for the vesicles with low pv was similar to the mean for all vesicles in rested RRPs at Schaffer collateral synapses ( see Lemma 3 ) , whereas pv for the vesicles with high pv was ∼5-fold higher ( Lemma 9 ) . The conceptual similarities make sense because the molecules are similar [2] . Indeed , the much faster disengagement of the acceleration mechanism at the calyx of Held could simply reflect much faster clearance of residual Ca2+ ( compare [47] to [11] ) . In contrast , the faster recruitment of vesicles to the RRP during maximal use likely indicates a difference in the molecular mechanism itself . Intriguingly , the time course of RRP replenishment during rest intervals was not largely different at the calyx of Held compared to Schaffer collateral synapses [9] , consistent with the possibility that the difference at the level of molecules is a single player involved in implementing the acceleration of vesicle recruitment to the RRP . In any case , the absence of fast vesicle recruitment during heavy use at Schaffer collateral synapses , and during rest intervals at both types of synapse , does not appear to be an intrinsic limitation of the biological material , suggesting instead a physiological role that remains to be elucidated [40 , 61–63] . There is no contradiction between the premise that the RRP has a fixed capacity and the concept that only a small subdivision might be immediately or effectively releasable at any given time [15 , 52] . However , the concept implies that some readily releasable vesicles are not immediately or effectively releasable , which is contrary to the original idea that RRP vesicles are ready to release . And , correlations between RRP size and responses to isolated action potentials [51 , 53] , and the correlations between the paired pulse ratio , time course of induction of depression , and size of the steady state supply of vesicles during 100Hz trains of action potentials in Fig 9 , all suggest that even vesicles within the slow-releasing subdivision of the RRP are immediately releasable , although with a low value for pv . In any case , even if some vesicles within the RRP are not immediately releasable , the transition to a releasable state can occur rapidly , in less than 10ms , when Ca2+ influx is massive [17] . Vesicles with high pv can be thought of as low pass frequency filters of information encoded by spike trains , whereas vesicles with low pv are high pass filters [23] . This potential source of computational power is often neglected in models of neural networks , partly because the principles of operation have not been clear . One key issue is whether vesicles with low and high pv can co-exist in RRPs of synapses with single active zones . That is , the release sites in mathematical models are likely the functional correlate of morphological docking sites in active zones , and it is possible that all vesicles docked at any given active zone all have similar values for pv; e . g . , owing to the density of Ca2+-channels [64 , 65] . This is an important topic because the synaptic connection between pairs of neurons in hippocampus and in other brain regions is often via a single synapse containing only one active zone [66] . Thus , determining if synapses with single active zones can simultaneously process vesicles with low and high pv—allowing multiple types of frequency filtering—will be important for understanding computational principles in circuits throughout the brain .
Animal protocols were approved by the University of Nevada and Universidad de Navarra and conformed to the guidelines of the National Institutes of Health and Spanish Royal Decree 1201/2005 . Tissue from n = 32 animals of both sexes was used for this in vitro study . Animals were rapidly decapitated , without anesthesia . Transverse slices ( 200μm ) containing the MNTB were prepared from C57/Bl6 mice ( 14–21 days old ) in ice cold dissection solution as previously described [67] , except the dissection solution was ( in mM ) : 85 NaCl , 2 . 5 KCl , 25 NaHCO3 , 1 . 25 NaH2PO4 , 25 glucose , 75 sucrose , 7 MgCl2 , 0 . 5 CaCl2 , 0 . 4 L-ascorbic acid , 3 myo-inositol , and 2 Na-pyruvate . After cutting , slices were incubated in artificial cerebrospinal fluid ( ACSF ) for 45–60min at 35C , and subsequently stored at room temperature ( 22–24C ) for up to 6 hours . ACSF was ( in mM ) : 125 NaCl , 2 . 5 KCl , 2 CaCl2 , 2 MgCl2 , 25 glucose , 25 NaHCO3 , 1 . 25 NaH2PO4 , 0 . 4 ascorbic acid , 3 myo-inositol , and 2 Na-pyruvate . Both solutions were continuously oxygenated with a gas mixture of 95% O2/5% CO2 . Slices were bathed in a <1ml recording chamber with ACSF that was continuously refreshed at approximately 2ml/min . When used , KYN was added in powder form to already oxygenated ACSF and stirred vigorously for at least 0 . 5 hours prior to use . Neurons were visualized with infrared differential interference contrast microscopy ( BX51WI , Olympus , Japan ) via a 40x water-immersion objective . Axons from the ventral cochlear nucleus were stimulated with a bipolar tungsten electrode spanning the ventral stria at the mid-line , and excitatory post synaptic current ( EPSCs ) responses were recorded from principal MNTB neurons , voltage clamped at −70mV with a PC-505B amplifier ( Warner Instruments , USA ) , or a Multiclamp 700B ( Axon ) . Stimulus intensity was set ∼25% above threshold and was 2–3 . 5V for 50–100μs . AMPA-receptor mediated EPSCs were isolated with 100μM DL-APV , 50μM picrotoxin , and 0 . 5μM strychnine . Intracellular recording solution was ( in mM ) : 130 Cs-gluconate , 10 CsCl , 5 Na2 phosphocreatine , 10 HEPES , 5 EGTA , 10 TEA-Cl , 4 MgATP , and 0 . 3 GTP , with pH adjusted to 7 . 2 . Recording pipettes were pulled from thick-walled borosilicate glass ( GC150F −10 , Harvard Apparatus , USA ) with a Sutter P-97 electrode puller to open tip resistances of 1 . 6–2 . 5MΩ . Series resistance in whole-cell recording configuration was <10MΩ , and was compensated 80–92% . All recordings were at room temperature . For most recordings , amplifier , stimulation , and data acquisition were controlled by a computer running in-house software on top of a Debian Linux operating system patched for real-time functionality with the RealTime Application Interface for Linux ( www . rtai . org ) ; the data for experiments in S1 Fig were recorded using PClamp . It was often possible to repeat several trials of each experiment in individual preparations and the digitized recordings of identical trials were averaged before further analysis . Preparations were always allowed at least 60s of rest before each experiment was initiated to allow synapses to recover completely between trials . For the experiments with a single experimental variable , the experimental and control trials were alternated . For time courses , the order of trials was shuffled . A minimum of 5min was allowed for solution exchange when adding drugs or changing the Ca2+ concentration . Analysis was accomplished using in-house software written in C++ and MATLAB . Aggregate data in figures and text are summarized with mean ± s . e . m . Statistical significance from pairwise comparisons was assessed with the Kolmogorov-Smirnov test: * = p < 0 . 05; ** = p < 0 . 01; *** = p < 0 . 001 . function RecruitVals = SimulateRecruit ( Resps , AlphaLUT , SegLen ) Vacancy ( length ( Resps ) ) = 0; RecruitVals ( length ( Resps ) ) = 0; for i = 2: ( length ( Resps ) ) Vacancy ( i ) = Vacancy ( i − 1 ) + Resps ( i − 1 ) − RecruitVals ( i − 1 ) ; % Vacancy is RRP ( 0 ) − RRP ( t ) RecruitVals ( i ) = Vacancy ( i ) *LookUp ( i − 1 , SegLen , AlphaLUT ) *SegLen; end; function alpha = LookUp ( RespIndex , SegLen , LUTable ) time = ( RespIndex—0 . 5 ) * SegLen; [ ˜ , ClosestIndex] = min ( abs ( LUTable ( : , 1 ) − time ) ) ; alpha = LUTable ( ClosestIndex ( 1 ) , 2 ) ; | Short-term plasticity has a dramatic impact on the connection strength of almost every type of synapse during normal use . Some synapses enhance , some depress , and many enhance or depress depending on the recent history of use . A better understanding is needed for modeling information processing in biological circuits and for studying the molecular biology of neurotransmission . Here we show that first principles at the calyx of Held , such as whether or not a readily-releasable pool of vesicles in the presynaptic terminal has a fixed capacity for storing vesicles , are unexpectedly similar to synapse types that are used at much lower frequencies . Our study establishes new methods for studying the function of presynaptic molecules , and the results suggest that a tractable general model of short-term plasticity can capture the full computational power of dynamic synaptic modulation across a large range of synapse types and situations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"plant",
"anatomy",
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"vesicles",
"nervous",
"system",
"membrane",
"potential",
"cell",
"processes",
"electrophysiology",
"neuroscience",
"plant",
"science",
"synaptic",
"plasticity",
"nerve",
"fibers",
"cel... | 2016 | A Well-Defined Readily Releasable Pool with Fixed Capacity for Storing Vesicles at Calyx of Held |
Global deworming programs aim to reach 75% of at-risk preschool-age children ( pre-SAC ) by 2020 . The 2013 global pre-SAC deworming coverage initially published by the World Health Organization ( WHO ) was 23 . 9% , but this estimate inadequately captured deworming delivered through Child Health Day ( CHD ) platforms . To update global and regional coverage estimates of pre-SAC deworming in 2013 by supplementing data from the WHO Preventive Chemotherapy and Transmission Control ( PCT ) databank with national CHD data . UNICEF country offices ( n = 82 ) were mailed a questionnaire in July 2014 to report on official national biannual CHD deworming coverage as part of the global vitamin A supplementation coverage reporting mechanism . Coverage data obtained were validated and considered for inclusion in the PCT databank in a collaboration between UNICEF and WHO . Descriptive statistical analyses were conducted to update the number of pre-SAC reached and the number of treatments delivered . Of the 47 countries that responded to the UNICEF pre-SAC deworming questionnaire , 73 data points from 39 countries were considered for inclusion into the WHO PCT databank . Of these , 21 new data points were from 12 countries were newly integrated into the WHO database . With this integration , deworming coverage among pre-SAC increased to 49 . 1% , representing an increase in the number of children reached and treatments administered from 63 . 7 million to 130 . 7 million and 94 . 7 million to 234 . 8 million , respectively . The updated databank comprised 98 mass deworming activities conducted in 55 countries , in which 80 . 4% of the global pre-SAC population requiring deworming reside . In all , 57 countries requiring deworming were not yet represented in the database . With the inclusion of CHD data , global deworming programs are on track to achieving global pre-SAC coverage targets . However , further efforts are needed to improve pre-SAC coverage reporting as well as to sustain and expand deworming delivery through CHDs and other platforms .
Soil-transmitted helminthiases ( STH ) are a group of parasitic diseases caused by intestinal worms that are transmitted to humans through fecally-contaminated soil [1] . Although STH is classified as a neglected tropical disease ( NTD ) , it is estimated that over 3 . 5 billion people are at risk of infection , 10%-15% of whom are children of preschool age ( pre-SAC ) [2] . STH-related morbidity increases with higher worm load and includes iron deficiency , protein malnutrition and poor cognitive development [3 , 4] . The regular deworming of pre-SAC with anthelminthic agents is an efficient and effective method of keeping STH worm loads below the levels associated with morbidity [1 , 5] . In its strategic plan on “Eliminating Soil-Transmitted Helminths as a Public Health Problem in Children” , the World Health Organization ( WHO ) outlined strategies and milestones towards “eliminating [STH-related] morbidity in all children by 2020” [6] . The plan aims to increase global pre-SAC deworming coverage to 50% by 2015 and 75% by 2020 [6] . The Preventive Chemotherapy and Transmission Control ( PCT ) databank was instituted to track progress to global coverage targets by combining coverage information from different sources [7] . According to the databank , global pre-SAC deworming coverage has progressively declined from 37 . 1% in 2010 to 30 . 6% in 2011 and 24 . 7% in 2012 [8–10] . The initial coverage estimate for 2013 was 23 . 9% [11] . The initial 2013 coverage estimates were generated from 43 countries representing 54 . 0% of the global pre-SAC population requiring deworming [11 , 12] . WHO and its partners recognize that not all countries requiring deworming report data to the databank and that not all mass deworming administrations ( MDAs ) are captured among reporting countries [13] . There has been no systematic attempt to collect data on deworming administered through Child Health Days , which are common delivery mechanisms for deworming alongside vitamin A supplementation ( VAS ) and other high-impact interventions among pre-SAC [14] . This paper describes the first systematic attempt to collect CHD deworming coverage in an effort to update the global 2013 pre-SAC coverage figures in the global WHO PCT databank .
UNICEF developed a spreadsheet-based questionnaire for deworming coverage data that was developed based on the VAS CHD reporting form used for official global VAS reporting [15] . The form includes questions on deworming coverage by age group , frequency , type of anthelminthic used and distribution mechanism employed . In line with the VAS reporting methodology , data on MDA were collected separately for semester 1 ( January–June , 2013 ) and semester 2 ( July–December , 2013 ) . Where possible , the PCT databank has country entries separated by rounds , we refer to each entry in the PCT databank as a data point [7] . In July 2014 , UNICEF Headquarters ( HQ ) sent these forms to its 82 country programs prioritized for VAS programs [16] . In these countries , UNICEF staff were requested to complete these forms with implementing partners and to obtain approval from relevant government ministries . Upon receipt , UNICEF HQ reviewed the data for completeness and consistency . Preliminary data or data that had not been officially approved by the government were excluded from further analysis . The WHO PCT Databank is the official global tracking tool for global pre-SAC deworming coverage [7] . Started in 2003 , it contains country-specific data provided by national NTD focal points , the Schistosomiasis Control Initiative , and development partners [7 , 12] . Each record in the database presents the number of pre-SAC that require deworming , the program target population , the program coverage ( the proportion of individuals treated as per program target set ) and the national coverage ( the proportion of the population requiring deworming in the country that has been treated ) . This database is continuously updated . WHO initially published the 2013 global pre-SAC coverage in January 15th 2015 [11] . These estimates did not include the 2013 UNICEF CHD data . The coverage estimates collected by UNICEF were considered for updating the initial 2013 global coverage data in the PCT databank . As a first step , the semester-specific CHD events ( i . e . , data points ) were compared with the data points already included in PCT database . All data points were then categorized into three groups: data points only present in the UNICEF deworming database , data points only present in the PCT databank and data points present in both databases . Data points only present in the UNICEF deworming database were validated based on number of pre-SAC treated , number of pre-SAC targeted and reported program coverage . The sources used as the standard for comparison were the latest available census and 2013 United Nations Population Division ( UNPD ) estimates [17] . Data points present in both databases were assessed for consistency and validation . They were considered to be identical if reported coverage was within 5% , in which case the WHO PCT data point was selected . In cases where data points were not considered identical , the data point containing the higher number of pre-SAC treated was chosen in line with WHO procedures . Any new or revised data points were uploaded into the WHO PCT databank and published [12] . The updated PCT databank was used to generate new global , regional and national level pre-SAC deworming coverage estimates . Regional estimates were calculated using WHO regional classifications [18] . These coverage estimates were defined as the proportion of the pre-SAC population requiring deworming that received treatment [10] . The number of pre-SAC that require deworming are determined by WHO in partnership with Ministries of Health ( MoHs ) after taking into account demographic , epidemiologic and sanitation data [19] . In addition , data reporting gaps were calculated by comparing the proportion of pre-SAC population requiring deworming for the responding countries against the global pre-SAC population requiring deworming over for the years 2006–2013 using both the initial and updated datasets . Standard descriptive statistical analyses were used to calculate the number of treatments provided by anthelminthic agent and to determine the number of MDAs that surpassed the program and national coverage targets of 75% . All analyses were conducted using Microsoft Excel ( Microsoft , Redmond , WA ) .
In the initial analysis of the PCT databank , 63 . 9 million pre-SAC that required deworming received the intervention in 2013 ( Table 2 ) . The global coverage estimate in this version of the PCT databank was 23 . 9% . South-East Asia region ( 40 . 2% ) had the highest coverage followed by the Americas ( 33 . 3% ) and Western Pacific Region ( 22 . 9% ) . In the updated databank , the number of pre-SAC requiring deworming who were treated increased to 130 . 7 million , thus more than doubling the global coverage from 23 . 9% to 49 . 1% ( Fig 2 ) . The highest increases in pre-SAC treated were observed in Africa and South-East Asia with an additional 41 . 4 and 19 . 4 million additional pre-SAC treated , respectively . With the update of the PCT databank , the number of treatments administered increased from 94 . 7 million to 234 . 8 million ( Table 3 ) . Of the treatments delivered in 2013 according to the updated PCT databank , albendazole was used exclusively in 39 rounds ( 56% of all treatments ) , mebendazole was used exclusively in 26 rounds ( 25% of all treatments ) , and a combination of anthelminthics ( always including albendazole ) in the remaining rounds ( Fig 3 ) .
With the current CHD deworming coverage assessment , a total of 35 new data points were added to the WHO PCT databank . These new data points captured an additional 140 . 1 million deworming dosages , thus increasing the global number of pre-SAC deworming dosages administered in 2013 to 234 . 8 million . With this inclusion , the global 2013 pre-SAC deworming coverage increased from 23 . 9% to 49 . 1% . This new global pre-SAC deworming coverage estimate thwarts apparent global coverage decreases observed from 2010 to 2012 [8–10] . In fact , the estimate puts global deworming programs back on track to attain the 50% coverage by 2015 [13] . The update also aligns with the 2012 London Declaration on Neglected Tropical Diseases ( NTDs ) goals of fostering collaboration and coordination on NTDs and providing regular updates on the progress in reaching 2020 goals and remaining gaps [20] . Nevertheless , given that 57 countries requiring require deworming are not represented in the PCT databank , there is insufficient progress towards demonstrating that all countries requiring deworming start programs by 2015 [6] . CHDs delivered nearly half of all global pre-SAC treatments in 2013 , thus illustrating the strategic importance of this delivery mechanism for attaining global pre-SAC coverage goals . CHDs may take the form of integrated special immunization activities or special events designed to deliver VAS and other high-impact interventions among to pre-SAC . The design of CHDs and the package of interventions offered can be tailored to the local contexts [21] . In fragile health systems , CHDs serve as a major delivery platform for high-impact interventions targeted to pre-SAC [22 , 23] . In settings with stronger health systems , they can be increasingly integrated into decentralized , routine primary health care , such as through the local budgeting and management of the events by health districts [21] . Nevertheless , CHDs may divert resources away from the delivery of routine health services [14] . In CHDs , deworming is often co-delivered alongside VAS owing to logistical and epidemiological considerations [2 , 14] . Deworming may increase the acceptability of other high-impact interventions delivered in CHDs because caretakers consider the excretion of worms in the child’s feces observed shortly after drug administration as a sign of improved child health [24] . Given the frequent co-delivery of VAS and deworming , the global , annual two-dose VAS coverage of 65% may serve as an indication of pre-SAC deworming coverage achievable through co-delivery with VAS [25] . Prior to this coverage assessment , global pre-SAC deworming was inadequately covered in the official 2013 PCT database . This data gap is due to incomplete reporting of CHD activities to WHO Geneva . Similar reporting gaps have been observed previously for deworming medications delivered by non-governmental organizations [26] and for deworming drugs delivered outside of national STH control programs [27] . Ideally , national Ministry of Health NTD focal persons would collect and report deworming coverage to WHO Geneva . As links with NTD focal persons at national levels are strengthened , UNICEF will continue the systematic coverage assessment of deworming delivered to pre-SAC through CHDs . The global deworming coverage assessment identified 24 data points from integrated poliomyelitis campaigns that also delivered deworming but that could not be included in the PCT databank because the number of pre-SAC reached and targeted exceeded WHO-defined levels deemed to be plausible . If these data points had been accepted , the global number of pre-SAC treated and the total number of treatments delivered would have increased by an additional 27 . 7 and 99 . 2 million , respectively . This inclusion would have increased the global pre-SAC coverage to 59 . 1% . In subsequent global coverage reporting exercises , harmonization of data rules employed by the STH- and poliomyelitis communities is warranted . As progress is maintained towards the eradication of poliomyelitis , the number of supplementary immunization activities delivering deworming and other high-impact interventions will decrease substantially [28] . As a result , national health planners need to ensure that STH control is reflected in the poliomyelitis transition plans and explore other delivery mechanism for VAS , deworming , and other child interventions currently delivered through vertical CHD-type events . This coverage assessment is limited by a number of factors . First , the CHD deworming coverage data collected is generally based on tally sheets completed during the supplementation events . Numerators generated from tally sheets may suffer from discounting , double counting or summation errors during their aggregation [29] . However , this system has been a long-standing data source for the global tracking of VAS coverage trends , and is deemed to be useful for the tracking deworming coverage trends as well [15] . Second , the district-disaggregated demographic and epidemiologic data used to generate estimates of children requiring deworming have limited accuracy , especially when sanitation data are used as a proxy for epidemiologic data [30] . Third , the current assessment focused on the 82 global VAS priority countries , and thus did not cover 35 countries requiring deworming . The scope of the coverage assessment will be extended to all global deworming priority countries in the future and to increase response rates among all countries contacted . Fourth , the link between pre-SAC deworming coverage data as presented here needs to be more rigorously linked to assessments demonstrating health impact analyses , and lastly , more information is needed on the quality of deworming formulations used in MDA programs [2] . In sum , this CHD coverage assessment showed that global pre-SAC deworming programs achieved more than double the coverage previously reported , and are therefore on track to achieve 2015 global pre-SAC coverage targets . With the increased attention on the role of NTDs for equitable and sustainable development [31 , 32] , deworming coverage reporting and program delivery should be further strengthened in an effort to reach pre-SAC in greatest need and to achieve the global 75% coverage target for 2020 . | Soil-transmitted helminthiases are a group of parasitic diseases caused by intestinal worms that are linked to poor physical and cognitive development among preschool aged children . The administration of deworming drugs designed to reduce the intensity of the worm infection in the child is effective and efficient intervention to control the disease and has set the goal of deworming 75% of at-risk children by 2020 . However , global WHO-reported coverage decreased from 37 . 1% to 24 . 7% from 2010 to 2012 . In 2013 , the first coverage estimate released was 23 . 9% , but as in previous years , this estimate did not adequately capture coverage achieved through Child Health Days , which are integrated campaign-style events where deworming is often co-delivered alongside vitamin A supplementation and other high impact child interventions . In this paper , we mailed a questionnaire to UNICEF country offices requesting data pertaining to preschool age deworming conducted through Child Health Days . After reviewing submissions and integrating data into the global databank , we report that the global coverage now stands at 49 . 4% putting us on track to achieve the global goal by 2020 . The sharp increase in coverage illustrates the importance of Child Health Days for attaining global pre-SAC coverage goals . | [
"Abstract",
"Introduction",
"Method",
"Results",
"Discussion"
] | [] | 2015 | The Role of Child Health Days in the Attainment of Global Deworming Coverage Targets among Preschool-Age Children |
The biological pacemaker approach is an alternative to cardiac electronic pacemakers . Its main objective is to create pacemaking activity from added or modified distribution of spontaneous cells in the myocardium . This paper aims to assess how automaticity strength of pacemaker cells ( i . e . their ability to maintain robust spontaneous activity with fast rate and to drive neighboring quiescent cells ) and structural linear anisotropy , combined with density and spatial distribution of pacemaker cells , may affect the macroscopic behavior of the biological pacemaker . A stochastic algorithm was used to randomly distribute pacemaker cells , with various densities and spatial distributions , in a semi-continuous mathematical model . Simulations of the model showed that stronger automaticity allows onset of spontaneous activity for lower densities and more homogeneous spatial distributions , displayed more central foci , less variability in cycle lengths and synchronization of electrical activation for similar spatial patterns , but more variability in those same variables for dissimilar spatial patterns . Compared to their isotropic counterparts , in silico anisotropic monolayers had less central foci and displayed more variability in cycle lengths and synchronization of electrical activation for both similar and dissimilar spatial patterns . The present study established a link between microscopic structure and macroscopic behavior of the biological pacemaker , and may provide crucial information for optimized biological pacemaker therapies .
Oscillating , autonomous or spontaneous electrical activity is the basis of normal heart physiology [1] , as well as some impaired rhythms triggered by ectopic activity [2] . Two oscillating mechanisms or clocks , the membrane and calcium clocks , are hypothesized to control the sinoatrial node ( SAN ) isolated cellular rate [3–5] . Membrane clock refers to the synergy of transmembrane ionic currents [6 , 7] , and calcium clock to the oscillations of intracellular calcium concentration [8] . Developmental variations may change magnitudes of the respective clock components [9] . Interplay between these two strongly coupled mechanisms may be responsible for spontaneous activity and temporal fluctuation in heart rate [10] . At the cellular level , the clocks basically create an ionic imbalance during the diastolic period , leading to a net inward flux of ionic current that slowly increases membrane potential until the threshold ( ~ −40 mV ) to fire an action potential is reached . Inducing this net inward flux of ionic current during the diastole can actually generate automaticity in otherwise quiescent cardiomyocytes ( CMs ) . This principle has been exploited in the design of biological pacemakers ( BPs ) , a therapeutic alternative to overcome the shortcomings of cardiac electronic pacemakers [11] in the treatment of bradycardia . Different procedures have been proposed , including injection-based gene [12] and cell therapy [13] , that locally modify cardiomyocyte phenotype or bring differentiated cells in the myocardium . These concepts are limited by the lack of control on the spatial distribution and phenotype of pacemaker ( PM ) cells within the resting but excitable cellular network of the myocardium . We have shown that density and spatial distribution of PM cells can alter significantly the emergence and characteristics of multicellular spontaneous activity [14] . In fact , density and spatial distribution of PM cells , a priori unknown in BPs , may lead to a non-negligible intrinsic variability in the spontaneous activity of the overall network . Intrinsic variability is defined as behavioral discrepancies among BP samples that had undergone the exact same protocol . This phenomenon could eventually compromise the success of BP implantation in patients , and is observed even in in vitro BP models like monolayer cultures of neonatal rat ventricular myocytes ( NRVMs ) , which are also heterogeneous network of autonomous and quiescent cardiomyocytes [15] . In the present simulation study , besides density and spatial distribution of PM cells , we introduce two additional variables: ( a ) automaticity strength and ( b ) anisotropy . Automaticity strength is defined as the ability of a pacemaker cell to maintain robust spontaneous activity with fast rate and to drive neighboring quiescent cells . It is strongly related to the amplitude of the net inward ionic current into the PM cell during the late diastolic period and the rising phase of the action potential ( AP ) . For example , adding fetal bovine serum to monolayer cultures of NRVMs “strengthens” automaticity , i . e . favors higher firing rate , by upregulating inward long-lasting activation calcium current ICaL [16] . Early versions of engineered BPs have been created from quiescent monolayer cultures or quiescent in vivo CMs via the use of different techniques to upregulate inward pacemaking current If [17] . The second newly introduced variable , anisotropy , can be created in cultures of NRVMs via several methods , notably by patterning the culture substrate [18 , 19] or directly seeding the cell into a thin slice of decellularized cardiac tissue [20] . These methods usually lead to functional cardiac network with elongated cell and faster propagation in the longitudinal direction [20] . It has been proposed that linear anisotropy could facilitate BP function [21] . However , the underlying mechanism remains unclear since most studies do not assess specific effects of anisotropy on spontaneous activity but instead focus on contractile function [20] , electrical activation [22] , or orientation-related response to stretch [23] . This study aims to assess modulation effects of automaticity strength and anisotropy on the spontaneous activity of cardiac monolayers with various densities and spatial distributions . The non-linear relationship between those two variables and automaticity will be characterized with simulation methods and discussed in details .
Semi-discrete microstructure models are more suitable than continuous models when individual cell sizes , shapes and orientations are variables under investigation [24 , 25] . For this reason , a previously described semi-discrete microstructure model [26] was used to simulate two 2D network geometries corresponding to isotropic and anisotropic monolayers . The two network geometries were identical in all aspects , except: ( a ) aspect ratio of cells ( AR , length divided by width of the cell ) , and ( b ) distribution of gap junctions . As summarized in Table 1 , a grid of 920 x 920 nodes was created with 6 μm resolution , and assigned to 42 , 642 CMs to create a 5 . 5 mm x 5 . 5 mm monolayer . Each cell included ~20 nodes . CMs for anisotropic geometry had an average AR of 3 compared to 1 for isotropic geometry . Longitudinal and transverse intercellular conductivities were adjusted to fit experimental conduction velocities found in NRVMs monolayer cultures [18] . The experimental isotropic conduction velocity was reported to be 16 . 8 ± 2 . 1 cm/s in all directions; and for anisotropic monolayer cultures , the longitudinal and transverse conduction velocities were 20 . 8 ± 3 . 2 cm/s and 10 . 9 ± 2 . 9 cm/s respectively . Intercellular coupling was set to 0 . 04 nS per 6 μm border length for CMs in isotropic network , and 0 . 062 nS and 0 . 034 nS per 6 μm border length respectively along longitudinal and transverse borders of CMs in anisotropic network . The Luo-Rudy Phase 1 ( LR1 ) mathematical model of ventricular cell [27] was used to represent the CMs , with the application of a constant inward bias current ( Ibias ) to generate spontaneous activity [28 , 29] . The LR1 model is simulated at every node of each cell , and examples of APs and total ionic currents obtained in a single cell with Ibias = 2 . 6 μA/cm2 and Ibias = 3 . 5 μA/cm2 are illustrated in Fig 1A and 1B respectively . This cell model was chosen because its bifurcation structure related to oscillatory behavior has been fully characterized . Indeed , bifurcation analysis undertaken with AUTO continuation software [30] is displayed in Fig 1C . The S-shape curve of fixed points has a lower and upper branch connected by an intermediate branch of unstable fixed points . Both the lower and upper branches change stability through subcritical Hopf Bifurcations ( H1 and H2 , magenta square ) . Stable cycle exist in between the two Cycle Saddle nodes bifurcation ( blue lines from SNC1 and SNC2 ) . At high Ibias , the branch of unstable cycles created at SNC2 ( green line ) connect the Hopf bifurcation H2 . The branch of unstable cycles created at SNC2 exist only on a small interval of Ibias and ends through a Homoclinic bifurcation with the intermediate branch of fixed point . Similarly , the branch of unstable cycles created at H1 exist on a tiny interval of Ibias and also disappears through a Homoclinic bifurcation with the intermediate branch of fixed point . The cycle length of stable spontaneous activity decreased with Ibias and ranged from 1989 ms to 516 ms ( Fig 1D ) . The 2D cardiac network was assumed to contain two populations of cells: PM ( Ibias = 2 . 6 μA/cm2 or Ibias = 3 . 5 μA/cm2 ) and quiescent ( Ibias = 0 μA/cm2 ) excitable cells . The density of pacemaker cells ( Daut ∈ [0 , 1] ) was defined as the percentage of PM cells within the network . The spatial distribution was dependent on a variable ( pthr ∈ [0 , 1] ) determining how homogeneous PM cells were spread in the network . Fig 2 provides intuitive disambiguation between density and spatial distribution . Density described the total number of PM cells in the network , regardless of their scattering within the network . Spatial distribution described the scattering of PM cells , and hence completed the spatial information from the overall density . Both variables were required to fully specify a spatial pattern . Two networks might have the same density of PM cells but different spatial distributions , or conversely they might share similar spatial distribution of PM cells but had different densities . Higher values of Daut and pthr correlate with higher percentage and more homogeneous spatial distribution of PM cells respectively . A previously described stochastic algorithm [14] was modified and implemented to randomly attribute positions to PM cells in the isotropic and anisotropic networks . Aggregation occurred when a PM cell was placed in the immediate neighborhood of another PM cell . Otherwise nucleation was said to occur , i . e . the PM cell occupied a position where only quiescent CMs were in its immediate neighborhood . Any random process of the algorithm followed continuous uniform probability distribution . For each pair ( Daut , pthr ) , the algorithm worked as stated below and as illustrated in Fig 3: Default Ibias value for all cells was 0 , i . e . all cells are quiescent unless set otherwise . Placing a PM cell consisted in setting Ibias for all nodes of a cell to a single non-zero value inside the interval [2 . 6–4 . 7] μA/cm2 . The same procedure was used to distribute spontaneous cells in the isotropic and anisotropic cell layouts . In Fig 4 are presented the first 50 x 50 nodes of two geometries . Cells in network with isotropic geometry demonstrated no preferential orientation compared to cells in network with anisotropic geometry which are clearly oriented along the longitudinal axis . Within the cardiac 2D network , a PM cluster was a subgroup of interconnected PM cells . Two PM cells were considered interconnected if they shared gap junctions . The size of a PM cluster was the number of PM cells in that cluster , and the maximum PM cluster size Scluster was the size of the network’s biggest PM cluster . Porosity was the fraction of quiescent cells in a PM cluster . In fact , any subgroup of interconnected PM cells was a PM cluster , but quiescent cells might also be enclosed within the cluster , i . e . totally surrounded by the cluster’s PM cells . Given S¯Tcluster , the average of maximum PM cluster size including both PM and quiescent cells , porosity was defined as follows: Porosity=1−S¯clusterS¯Tcluster where Porosity: average fraction of quiescent cells in the largest PM cluster S¯cluster: average of maximum PM cluster size , counting only PM cells S¯Tcluster: average of maximum PM cluster size , counting both PM and quiescent cells As previously described [26] , a 2D monodomain approximation with fine discretization was used to formulate the microstructure model . No-flux boundary conditions were applied to the four sides of the network . Initial conditions for all cells corresponded to the resting state of quiescent cells . The total simulation duration was 10 s , and the steady-state behaviors were reached rapidly within two to three autonomous period for most spontaneous cases , the longest transient behaviors found at the transition between non-autonomous to autonomous multicellular activity . Analysis was done on simulations after removing the first action potential thus including the time from the 2nd action potential up until 10 s . Simulations were performed to study the effect of Daut and pthr on the spontaneous activity of 4 groups: 400 pairs ( Daut , pthr ) were drawn from 20 values of Daut ∈ {0 . 05 , 0 . 1 , … , 0 . 95} and 20 values of scaled pthr1/4 ∈ {0 . 05 , 0 . 1 , … , 1} . We used non-regular spacing for pthr due to its nonlinear effect on cell distribution . Eight random realizations of the networks were generated for each pair ( Daut , pthr ) and for each group , generating 4 groups x 8 networks x 400 ( Daut , pthr ) = 12 , 800 simulations , as detailed below: Post-simulation analysis was performed in Matlab ( The Mathworks , Natick , MA ) . The network was said to have spontaneous activity if two complete activations or more were detected during 10 s of simulation . Conversely , the simulation was labeled as non-automatic if a single AP or no activation was detected . The activation time of an action potential ( AP ) was defined as the time when the transmembrane voltage depolarizes beyond -40 mV . For the ith AP of a given simulation , the activation map ( Mtact , i in ms ) was a matrix constructed from detected activation times for all nodes . A first set of measures was computed for each 10 s simulation with spontaneous activity . The spontaneous cycle length value ( Δtact , i in ms ) for that map was: Δtact , i=median[ΔMtact , i] ( 3 ) And the average spontaneous cycle length Δtact , i in ms ) for the series of N APs was: Δtact¯=1N−1∑i=2NΔtact , i ( 4 ) A second set of measures were defined for each pair ( Daut , pthr ) to assess how spontaneous activity behaves for similar spatial patterns and hence the variability between realizations of the same random process of pattern generation . In fact , as previously stated , eight monolayers have been produced for every pairs ( Daut , pthr ) . The monolayers produced with the same ( Daut , pthr ) had similar spatial patterns , compared to monolayers produced with different ( Daut , pthr ) which had dissimilar spatial patterns . ΔT¯act: average cycle length over n simulations for a pair ( Daut , pthr ) σΔTact: standard deviation of cycle length over n simulations for a pair ( Daut , pthr ) Δtact , j¯: cycle length for the jth simulation , as stated in Eq ( 4 ) T¯sync: average synchronization time over n simulations for a pair ( Daut , pthr ) σTsync: standard deviation of synchronization time over n simulations for a pair ( Daut , pthr ) τsync , j: synchronization time for the jth simulation , as stated in Eq ( 5 ) The same process is used to calculate T¯sync , x , T¯sync , y , longitudinal and transverse components of Tsync . The third set of measures were defined for dissimilar spatial patterns , i . e . monolayers with different values of ( Daut , pthr ) . ΔT¯act: average cycle length over n simulations for a pair ( Daut , pthr ) , as stated in Eq ( 6 ) T¯sync: average synchronization time over n simulations for a pair ( Daut , pthr ) , as stated in Eq ( 8 ) Furthermore , each monolayer was divided in two areas separated by a square of 2 . 75 mm side ( half of the side of the monolayer ) to distinguish between border and central foci ( first initiation site of AP ) . Border foci behavior was assessed by anisotropy ratio . The border foci anisotropy ratio ( r ) was defined as: r=ηLηT ( 12 ) ηL: number of border foci in longitudinal x-direction ( anisotropy direction ) ηT: number of border foci in transverse y-direction
S¯cluster , average of Scluster over 8 monolayers for each pair ( Daut , pthr ) , increased with Daut , independently of pthr1/4 , with a transition from below 10 , 000 PM cells to over 30 , 000 PM cells around Daut = 0 . 5 ( Fig 5A and 5D ) . The extent of the transition phase , i . e . the number of pairs ( Daut , pthr ) with S¯cluster between 10 , 000 and 30 , 000 PM cells , correlated with increased standard deviation of S¯cluster ( Std . Scluster ) . In fact , Std . Scluster ( Fig 5B and 5E ) was below 1 , 000 PM cells for all pairs ( Daut , pthr ) except for the transition phase where Std . Scluster > 2 , 000 PM cells . As shown in Fig 5C and 5F , the number of clusters ( Nclusters ) increased with Daut as long as Daut ≤ Daut , max ( white solid line ) , and then decreased for Daut > Daut , max . For each pthr1/4 , Daut , max was the maximum Daut beyond which M2 became empty , i . e . there was no more available site to perform cluster nucleation during the creation of the spatial distribution of PM cells . Thus , once Daut , max has been reached , only aggregation was possible for increasing Daut . In Fig 6 , a relationship was established between the transition in Scluster and Daut , max . The maximum of the derivative of S¯cluster as a function of Daut , i . e . max[S¯cluster′] = max[ΔScluster/ΔDaut] , representing the sharpness of the transition of Scluster from below 10 , 000 PM cells to over 30 , 000 PM cells , was calculated for all pthr1/4 and plotted against Daut , max . The sharpness of the transition was inversely proportional to Daut , max ( Fig 6D ) . Differences between isotropic and anisotropic geometries were unsurprisingly negligible for S¯cluster , Nclusters , and Daut , max , since the number of neighbors was approximately the same for both geometries . As such , spatial distribution of spontaneous cells created by our aggregation and nucleation process are not affected by cell preferential orientation . In general , 2D cardiac networks with isotropic geometry demonstrated circular-shaped electrical activation , as illustrated in Fig 7B for the pattern shown in panel a . Networks with the anisotropic geometry typically had ellipse-shape electrical activation ( Fig 7D for the pattern in panel c ) . For a specific pair ( Daut , pthr ) , given n the number of simulations with automaticity: ( a ) [n = 8] meant automaticity occurred for all 8 simulations , ( b ) [0<n<8] for 1 to 7 simulations , and ( c ) [n = 0] for no simulation with automaticity . As shown in Fig 8 , autonomous activity occurs more often in ISO-3 . 5 and ANISO-3 . 5 compared to ISO-2 . 6 and ANISO-2 . 6 over all pairs ( Daut , pthr ) . For example , 51 . 5% of pairs ( Daut , pthr ) demonstrated [n = 8] in ISO-3 . 5 versus 13 . 3% in ISO-2 . 6 . Interestingly , proportions of pairs ( Daut , pthr ) with [0<n<8] were very similar for all four groups ( ~9% ) . Automaticity was thus more likely to be observed for higher values of Daut and lower values of pthr1/4 , and no difference in occurrence of autonomous activity was found between isotropic and anisotropic geometries . A transition curve from [n = 0] to [0<n<8] was the line drawn by the minimum Daut required for each pthr1/4 to transition from [n = 0] to [0<n<8] . Similarly , the transition curve from [0<n<8] to [n = 8] was the line drawn by the minimum Daut required for each pthr1/4 to transition from [0<n<8] to [n = 8] . In Fig 9A–9D , transition curves from [n = 0] to [0<n<8] ( solid line ) and from [0<n<8] to [n = 8] ( dashed line ) were superimposed to the color scale map of Scluster and porosity , for isotropic and anisotropic networks . S¯cluster combined with porosity offered crucial insights on the morphology of the transition curves . Typically , automaticity did not appear when S¯cluster was below 5 , 000 PM cells and when porosity was over 0 . 35 . The higher pthr1/4 was , the greater Scluster had to be to generate automaticity . In all groups , larger S¯cluster was required to reach [n = 8] versus [0<n<8] . Networks with strong automaticity displayed spontaneous activity at smaller values of S¯cluster and higher porosity . For all 4 groups , transition curves to [0<n<8] were subtracted from transition curves to [n = 8] , and the differences were displayed in Fig 9E and 9F . Peak differences were higher for ISO-2 . 6 ( 0 . 50 ) and ANISO-2 . 6 ( 0 . 45 ) compared to ISO-3 . 5 ( 0 . 35 ) and ANISO-3 . 5 ( 0 . 40 ) respectively . Furthermore peak differences occurred for lower pthr1/4 for ISO-2 . 6 and ANISO-2 . 6 ( ~0 . 2 ) compared to ISO-3 . 5 and ANISO-3 . 5 ( ~0 . 40 ) . In Fig 10A–10D , ΔT¯act was calculated for each pair ( Daut , pthr ) with [n = 8] and displayed as color scale map vs . Daut and pthr1/4 . Independently of groups and for any given pthr1/4 , ΔT¯act decreased with increasing Daut . And ranges , i . e . intrinsic variability for dissimilar patterns , of ΔT¯act for ISO-2 . 6 ( 16 . 19% ) and ANISO-2 . 6 ( 14 . 56% ) were much smaller than ISO-3 . 5 ( 98 . 86% ) and ANISO-3 . 5 ( 100 . 96% ) . In Fig 10E is presented mean[ΔT¯act] , calculated as the average ΔT¯act over all pairs ( Daut , pthr ) with [n = 8] shown in Fig 10A–10D . Detailed values can be found in Table 2 . Mean[ΔT¯act] for ISO-2 . 6 and ANISO-2 . 6 were respectively 110 . 26% and 110 . 96% higher compared to those of ISO-3 . 5 and ANISO-3 . 5 . This behavior was not surprising , considering the difference observed in single cell simulations ( 1428 ms for Ibias = 2 . 6 μA/cm2 compared to 599 ms for Ibias = 3 . 5 μA/cm2 , in Fig 1D ) . Besides , it is important to notice that , for a given Ibias value , single pacemaker cells always display lower cycle length than monolayers . No difference in mean[ΔT¯act] was found between isotropic and anisotropic geometries for [n = 8] , but focusing on the set with [0<n<8] yielded interesting results . Pooling together all pairs ( Daut , pthr ) satisfying the condition [0<n<8] ( Table 2 ) , there was a higher mean[ΔT¯act] ANISO-2 . 6 versus ISO-2 . 6 ( 3 . 08% ) , and ANISO-3 . 5 versus ISO-3 . 5 ( 5 . 50% ) . Similarly to ΔT¯act , σΔTact was calculated for each pair ( Daut , pthr ) with [n = 8] . Mean[σΔTact] , the average value of σΔTact over all pairs ( Daut , pthr ) with [n = 8] is also shown in Fig 10F and Table 2 for all groups . Mean[σΔTact] was a measure of intrinsic variability between monolayers with similar spatial patterns of PM cells . Mean[σΔTact] for ISO-2 . 6 and ANISO-2 . 6 were respectively 78 . 52% and 116 . 68% higher compared ISO-3 . 5 and ANISO-3 . 5 . A difference between isotropic and anisotropic geometries was found where mean[σΔTact] in ANISO-2 . 6 was 30% higher vs . ISO-2 . 6 , while a more limited increase of 8% was found for ANISO-3 . 5 compared to ISO-3 . 5 . Difference in spatial characteristics of spontaneous activity between isotropic and anisotropic geometry was evaluated . The position of foci ( i . e . first initiation sites of electrical activation ) was estimated to determine if differences existed between the two network geometries . Independently of geometry , focal activation was highly stable in time for a given spatial pattern of spontaneous cells , demonstrating no beat-to-beat variability . For direct comparison between geometries , the focal position of the last simulated spontaneous beat was selected for all simulations with automaticity ( i . e . [n>0] ) . Pooled positions were plotted in Fig 11A–11D . The proportion of central foci over all foci is shown in Fig 11E and 11F and Table 2 for pairs of ( Daut , pthr ) with [n = 8] and pairs ( Daut , pthr ) with [0<n<8] . Anisotropic geometries demonstrated fewer central foci , independently of Ibias or n values . For pairs ( Daut , pthr ) with [n = 8] , proportions of central foci decreased by 69 . 11% from ISO-2 . 6 to ANISO-2 . 6 , and by 40 . 84% from ISO-3 . 5 to ANISO-3 . 5 . The drop was less important for pairs of ( Daut , pthr ) with [0<n<8] case , where proportions of central foci fell by 45 . 35% from ISO-2 . 6 to ANISO-2 . 6 , and by 22 . 05% from ISO-3 . 5 to ANISO-3 . 5 . It was also interesting to observe that cases with [0<n<8] demonstrated fewer central foci than cases with [n = 8] , independently of Ibias value or geometry . In fact , in ISO-2 . 6 and ANISO-2 . 6 , proportions of central foci respectively fell by 83% and 69 . 9% from [n = 8] to [0<n<8] . The difference between pairs of ( Daut , pthr ) with [n = 8] and [0<n<8] was less important when Ibias = 3 . 5 μA/cm2 . As a matter of fact , the drop of central foci proportions from [n = 8] to [0<n<8] was 25 . 58% in ISO-3 . 5 and 1 . 94% in ANISO-3 . 5 . The border foci in the longitudinal x-direction were the foci located outside the red box , exclusively to the left and to the right . The border foci in the transverse y-direction were the foci located outside the red box , exclusively at the top and the bottom . Non-exclusive border foci at the corners , i . e . foci that are common to longitudinal and transverse directions ( blue areas in Fig 11A–11D ) , were not considered in the calculations . Ratio with values greater than one suggested that there were more border foci in the longitudinal direction compared to the transverse direction . Values of r were calculated for [n = 8] and [0<n<8] and are presented in Table 2 . Interestingly , r in anisotropic geometry are consistently greater than one and always higher compared to the value obtained in isotropic geometry . In fact , for [n = 8] case , r raised by 27% from ISO-2 . 6 to ANISO-2 . 6 , and by 204% from ISO-3 . 5 to ANISO-3 . 5 . For [0<n<8] , r raised by 297% from ISO-2 . 6 to ANISO-2 . 6 , and by 119% from ISO-3 . 5 to ANISO-3 . 5 . No clear preference in border foci position was found for the isotropic network . Synchronization of electrical activation is an important marker of spontaneously beating multicellular monolayer , either in silico or in vitro . T¯sync was displayed as color map vs . Daut and pthr1/4 for each pair ( Daut , pthr ) with [n = 8] in Fig 12A–12D . No particular tendencies were observed in T¯sync map for ISO-2 . 6 and ANISO-2 . 6 . For ISO-3 . 5 and ANISO-3 . 5 , and for Daut ≥ 0 . 8 , T¯sync increased and then decreased as a function of pthr1/4 . Ibias = 3 . 5 μA/cm2 led to much higher ranges compared to Ibias = 2 . 6 μA/cm2 . Indeed , range[T¯sync] for ISO-2 . 6 ( 6 . 5% ) and ANISO-2 . 6 ( 16 . 9% ) were much smaller than ISO-3 . 5 ( 414% ) and ANISO-3 . 5 ( 485% ) . Moreover , the increase in ranges could be noticed in anisotropic geometries versus isotropic geometries: 160% increase from ISO-2 . 6 to ANISO-2 . 6 and more moderate 17 . 15% increase from ISO-3 . 5 to ANISO-3 . 5 . Anisotropy consistently yielded higher synchronization times . Mean[T¯sync] rose by 31 . 87% from ISO-2 . 6 to ANISO-2 . 6 and by 26 . 64% from ISO-3 . 5 to ANISO-3 . 5 . Mean[T¯sync , y] obviously played an important role in that increase . In fact , the unsurprising decrease of synchronization times in the direction of anisotropy mean[T¯sync , x] was associated with a dramatic increase of mean[T¯sync , y] , leading to a higher resultant mean[T¯sync] .
In summary , a pure change from isotropic to anisotropic substrate modelled by an elongated cell shape and anisotropic intercellular conductivity without modifications of ion channel expression nor spatial distribution has limited effects on spontaneous activity . However , increasing the intrinsic rate of autonomous cells has a much stronger effects . Although the two were studied together and independently , it is of importance to note that there is a strong possibility that both changes ( anisotropy and autonomous strength ) could be coupled [44] . Further work is thus needed to uncover this importance of the interaction ( and by how much methods to induce cellular anisotropy can increase the cellular automaticity strength ) and to elucidate how it could favor the BP development . | Implantation of electronic pacemakers is a standard treatment to pathologically slow heart rhythm . Despite improving quality of life , those devices display many shortcomings . Bioengineered tissue pacemakers may be a therapeutic alternative , but associated design methods usually lack control of the way cells with spontaneous activity are scattered throughout the tissue . Our study is the first to use a mathematical model to rigorously define and thoroughly characterize how pacemaker cells scattering at the microscopic level may affect macroscopic behaviors of the bioengineered tissue pacemaker . Automaticity strength ( ability of pacemaker cell to drive its non-pacemaker neighbors ) and anisotropy ( preferential orientation of cell shape ) are also implemented and give unparalleled insights on how effects of uncontrollable scattered pacemaker cells may be modulated by available experimental techniques . Our model is a powerful tool to aid in optimized bioengineered pacemaker therapies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"microstructure",
"cardiovascular",
"anatomy",
"electricity",
"condensed",
"matter",
"physics",
"anisotropy",
"materials",
"science",
"pacemakers",
"materials",
"physics",
"ionic",
"current",
"bioenergetics",
"myocar... | 2018 | In silico study of multicellular automaticity of heterogeneous cardiac cell monolayers: Effects of automaticity strength and structural linear anisotropy |
Among the several challenges faced by bloodsucking arthropods , the vertebrate hemostatic response against blood loss represents an important barrier to efficient blood feeding . Here we report the first inhibitor of collagen-induced platelet aggregation derived from the salivary glands of a black fly ( Simulium nigrimanum ) , named Simplagrin . Simplagrin was expressed in mammalian cells and purified by affinity-and size-exclusion chromatography . Light-scattering studies showed that Simplagrin has an elongated monomeric form with a hydrodynamic radius of 5 . 6 nm . Simplagrin binds to collagen ( type I-VI ) with high affinity ( 2–15 nM ) , and this interaction does not involve any significant conformational change as determined by circular dichroism spectroscopy . Simplagrin-collagen interaction is both entropically and enthalpically driven with a large negative ΔG , indicating that this interaction is favorable and occurs spontaneously . Simplagrin specifically inhibits von Willebrand factor interaction with collagen type III and completely blocks platelet adhesion to collagen under flow conditions at high shear rates; however , Simplagrin failed to block glycoprotein VI and Iα2β1 interaction to collagen . Simplagrin binds to RGQOGVMGF peptide with an affinity ( KD 11 nM ) similar to that of Simplagrin for collagen . Furthermore , Simplagrin prevents laser-induced carotid thrombus formation in vivo without significant bleeding in mice and could be useful as an antithrombotic agent in thrombosis related disease . Our results support the orthology of the Aegyptin clade in bloodsucking Nematocera and the hypothesis of a faster evolutionary rate of salivary function of proteins from blood feeding arthropods .
Salivary glands ( SGs ) of blood feeding arthropods have been studied for their roles in blood feeding and pathogen transmission to vertebrate hosts . As in other bloodsucking Nematocera , black flies require a blood meal for egg development . To acquire a blood meal , the mandibles of the fly cut into the skin with rapid scissor-like movements , causing blood to pool from which it will feed , with blood feeding usually taking four to five minutes [1] . This feeding behavior triggers the hemostatic response of the vertebrate host against blood loss , which represents a formidable barrier to efficient blood feeding [2] . The first step in the hemostatic cascade is platelet interaction with the exposed extracellular matrix at sites of injury . Collagen is recognized as the most thrombogenic component of the subendothelial matrix . Endothelial damage—such as that caused by blood feeding arthropods—can lead to exposure of collagen to circulating blood , in particular to platelets , leading to thrombogenesis . Multiple collagen receptors have been identified on the platelet surface including immunoglobulin superfamily member GPVI , GPIb and integrin α2β1 , among others ( reviewed in [3] ) . These individual receptors likely play specific roles to mediate collagen-induced platelet adhesion , activation , and consolidation [3] , [4] . Absence of any of these components can lead to serious physiologic consequences . For example , von Willebrand disease caused by quantitative or qualitative defects of vWF can cause excessive mucocutaneous bleeding after even minor tissue damage [5] . To counteract the hemostatic system of the host , saliva of blood feeding arthropods contains a complex array of pharmacologically active compounds that act as anticlotting , antiplatelet , vasodilator , anti-inflammatory , and immunomodulatory compounds . Some functional and biochemical characterizations from black fly SGs have previously been reported [6]–[9] . Among the salivary platelet aggregation inhibitors in mosquitoes , it was recently discovered that Anopheles stephensi ( AAPP ) and Aedes aegypti ( Aegyptin ) express a collagen-binding protein that inhibits collagen-induced platelet aggregation by blocking its interaction with three major ligands , namely , GPVI , von Willebrand factor ( vWF ) , and integrin α2β1 [10]–[12] . These mosquito proteins have a low complexity and acidic amino terminus region rich in glycine/aspartate/glutamate and a relatively more conserved and complex carboxyterminus . Proteins with these characteristics were found in black flies [13]–[15]; however , their overall identity was only 25% when aligned to mosquito proteins [2] . Black flies and mosquitoes share a common blood feeding ancestor at ∼250 million years ago ( MYA ) [16] , giving ample time for diversification of this protein family , although the biophysical , biochemical , and pharmacologic characterization of this protein family in black flies remains to be elucidated . To the extent that they are similar to those of mosquitoes , a point could be made for their orthologous relationship , despite accelerated evolution , probably driven by their hosts' immune pressure over millions of years [17] . Here we report the first collagen-induced platelet aggregation inhibitor from Simulium nigrimanum SGs ( Simulium platelet aggregation inhibitor , Simplagrin ) . Simplagrin specifically inhibits vWF interaction with collagen under static conditions and completely blocks platelet adhesion to collagen under flow conditions at high shear rates . Simplagrin binds to the vWF-recognition peptide ( RGQOGVMGF ) with an affinity ( KD 11 . 1±0 . 59 nM ) similar to that of Simplagrin collagen I and III ( 5 . 6±0 . 52 nM and 2 . 1±0 . 35 nM , respectively ) . Furthermore , Simplagrin prevents laser-induced carotid thrombus formation in mice in vivo without significant bleeding . From an evolutionary viewpoint , our results support the orthology of the Aegyptin clade in bloodsucking Nematocera and the hypothesis of a faster evolutionary rate of salivary function of distantly related proteins , and the central role of platelet aggregation inhibition in blood feeding arthropods .
Adenosine diphosphate ( ADP ) and phorbol myristate acetate were obtained from Sigma ( St . Louis , MO , USA ) . Ristocetin and arachidonic acid ( Ara ) were from Chrono-Log Corp . ( Haverton , PA , USA ) . 9 , 11 dideoxy 9α , 11α methanoepoxy prostaglandin F2α ( U46619 ) was purchased from Cayman Chemicals ( Ann Arbor , MI , USA ) ; thrombin receptor-activating peptide ( TRAP ) was from EMD Biosciences ( La Jolla , CA , USA ) . Thrombin and PPACK ( D-phenylalanyl-L-propyl-L-arginine chloromethyl ketone ) were from Haematologic Technologies ( Essex Junction , VT , USA ) . Convulxin was from Santa Cruz Biotechnology , Inc , ( Santa Cruz , CA , USA ) and GPVI-His from R&D Systems ( Minneapolis , MN , USA ) . Soluble human collagen type I , III , IV , V and VI were also from R&D Systems . Synthesis and preparation of collagen derived peptides were as described in [11] . Public Health Service Animal Welfare Assurance #A4149-01 guidelines were followed according to the NIAID and the National Heart , Lung , and Blood Institute ( NIH ) Office of Animal Care and Use ( OACU ) . These studies were carried out according to the NIAID and NHLBI animal study protocol ( ASP ) approved by the OACU Committee , with approval IDs ASP-LMVR3 and ASP-2-CB-38 ( R ) Amendment 18 . Adult female mice weighing 15–25 g were housed under controlled conditions of temperature ( 24±1°C ) and light ( 12 hours light starting at 7:00 am ) , and all experiments were conducted in accordance with standards of animal care defined by the Institutional Committees ( NIH ) . Simplagrin expression construct was based on Sim-50 sequence ( ACZ28269 ) and engineered to contain the mature protein and a 6X-His tag before the stop codon . The synthetic , codon optimized gene was produced by BioBasic Inc . ( Markham , ON , Canada ) and subcloned into VR2001 TOPO vector ( modified version of the VR1020 vector; Vical Incorporated [San Diego , CA , USA] ) . About 1 mg of plasmid DNA ( VR2001-Simplagrin construct ) was obtained using GeneElute HP endotoxin free plasmid MEGA prep kit ( Sigma ) . The plasmid was purified through a 0 . 22 µm filter , and the recombinant protein was produced by transfecting FreeStyle293 F cells ( Invitrogen , San Diego , CA , USA ) . After 72 hours , transfected cell culture was harvested and the supernatant containing the secreted recombinant protein was centrifuged ( 100×g , 15 minutes ) , frozen , and shipped to our lab until purification . HEK293 cell supernatant containing the recombinant protein was loaded onto a Ni2+ column ( 5 mL bed volume; GE Healthcare , Piscataway , NJ , USA ) following the manufacturer's directions . Fractions were step eluted with 5 , 20 , 300 and 1000 mM imidazole ( in 10 mM Tris , 500 mM NaCl , pH 8 . 0 ) and then loaded onto a size exclusion column ( Superdex 200 HR10/30; GE Healthcare ) using the AKTA purifier system . Proteins were eluted isocratically at a flow rate of 0 . 5 mL/minute in 25 mM Tris , 150 mM NaCl , pH 8 . 0 . Purified recombinant protein was submitted to automated Edman degradation for N-terminal sequencing . To detect purity of Simplagrin , 10 µg of purified protein was loaded in a 4–12% NuPAGE gel ( Life Technologies , Gaithersburg , MD , USA ) and the gel stained with Coomassie blue . FITC labeling of Simplagrin was carried out using the FITC protein labeling kit according to the manufacture's recommendations ( Life Technologies ) . Briefly , labeling reaction was incubated at room temperature ( RT ) for two hours , and unreacted dye was removed by size exclusion chromatography as described above . The final concentration and degree of labeling were determined by measuring the absorbance spectrum of the labeled Simplagrin as indicated below . Concentration of purified Simplagrin was estimated by its absorbance at 280 nm using an ND1000 spectrophotometer ( NanoDrop Technologies , Wilmington , DE , USA ) and corrected according to the calculated molar extinction coefficient ε280 nm of 10 , 930 M−1cm−1 . The purity and solution state of purified Simplagrin were analyzed using size exclusion chromatography with online multiangle light scattering ( SEC-MALS-QELS-HPLC ) , refractive index , and ultraviolet ( UV ) detection . The instrument , a Waters Corporation ( Milford , MA , USA ) HPLC model 2695 and photodiode array detector model 2996 operated by Waters Corporation Empower software , connected in series to a Dawn EOS light scattering detector and Optilab DSP refractive index detector ( Wyatt Technology , Santa Barbara , CA , USA ) , was used as directed by the manufacturer . Wyatt Technology's Astra V software suite was used for data analysis and processing . For separation , a TSK gel G3000PWxl column ( 7 . 8 mm×30 cm; 6 µm particle size ) ( Tosoh Bioscience , King of Prussia , PA , USA ) was used with a TSK gel Guard PWxl column ( 6 . 0 mm×4 . 0 cm , 12 µm particle size ) . The column was equilibrated in mobile phase ( 1 . 04 mM KH2PO4 , 2 . 97 mM Na2HPO4 • 7H2O , 308 mM NaCl , 0 . 5 M urea , pH 7 . 4 , 0 . 02% sodium azide ) for at least 60 minutes at 0 . 5 mL/min prior to sample injection . SEC-MALS-HPLC analysis was performed on the Simplagrin using an isocratic elution at 0 . 5 mL/min in mobile phase . Bio-Rad gel filtration standards were run for size comparisons . To evaluate whether recombinant Simplagrin underwent glycosylation by HEK293 cells , an enzymatic deglycosylation assay was carried out . Fifteen µg of Simplagrin or Lundep ( positive control [18] ) were denatured at 95°C for 5 minutes and then treated with a deglycosylase mix containing PNGase F , sialidase , β-galactosidase , glucosaminidase , and O-glycosidase ( Enzymatic DeGlycoMx Kit; QA-Bio , Palm Desert , CA , USA ) . After three hours of enzymatic deglycosylation , the samples were submitted to electrophoresis . Electrophoretic mobility of Simplagrin before and after enzymatic deglycosylation was compared in a Coomassie blue stained NuPAGE ( Life Technologies ) . CD spectra were measured by a Jasco J-715 spectropolarimeter with the solutions in a 0 . 1-cm path length quartz cuvette in a cell holder thermostated by a Neslab RTE-111 circulating water bath . Spectra were scanned four times from 260 to 200 nm and averaged ( speed 50 nm/min , time constant one second ) at 25°C . After baseline correction , CD spectra were converted into mean residue ellipticity values using the formula described in [11] . All SPR experiments were carried out in a T100 instrument ( GE Healthcare ) following the manufacturer's instructions . Sensor CM5 , amine coupling reagents , and buffers were also purchased from GE Healthcare . HBS-P ( 10 mM Hepes , pH 7 . 4 , 150 mM NaCl , and 0 . 005% ( v/v ) P20 surfactant was used as the running buffer for all SPR experiments . All SPR experiments were analyzed using the Biacore Evaluation software v2 . 0 . 3 provided by GE Healthcare . All SPR experiments were carried out three times . Collagen type I or type III ( 20 µg/mL ) in acetate buffer ( pH 4 . 5 ) was immobilized over a CM5 sensor via amine coupling . The immobilization target was aimed to 800 resonance units ( RU ) , resulting in a final immobilization of 796 . 6 RU for collagen type I and 818 . 6 RU for collagen type III . Blank flow cells were used to subtract the buffer effect on sensograms . Alternatively , Simplagrin ( 50 µg/mL in acetate buffer; pH 4 . 5 ) was immobilized on a CM5 at a surface density of 500 RU . Kinetic experiments were carried out with a contact time of 180 seconds at a flow rate of 30 µL/minute at 25°C . Simplagrin-collagen complex dissociation was monitored for 1 , 800 seconds , and the sensor surface was regenerated by a pulse of 20 seconds of 10 mM HCl at 40 µL/minute . Sensograms were fitted using the 1∶1 Langmuir interaction model . For steady-state affinity calculations , Simplagrin was immobilized as described above and different concentrations of soluble collagen ( type I–VI ) were used as analyte . Steady-state affinity was calculated by plotting the equilibrium response ( Req ) levels against the analyte concentration . The affinity constant ( KD ) is reported as the analyte concentration at 50% of saturation . Thermodynamic parameters for Simplagrin collagen type I and III interaction were obtained from independent kinetic experiments using the Thermo Wizard assay program . Briefly , six different concentrations of recombinant Simplagrin ( 10 , 25 , 50 , 100 , 250 , and 500 nM ) were injected over immobilized collagen type I or III at 15 , 20 , 25 , 30 , 35 , and 40°C . Contact time , dissociation time , and regeneration of the sensor surface were done as described above . Resulting sensograms were fitted to the 1∶1 interaction model with global Rmax . The association ( Ka ) and dissociation ( Kd ) rate constants , as well as the affinity constant ( KD ) , were obtained and fitted to a linear form of the van't Hoff equation to estimate the ΔH and ΔS . This experiment was essentially carried out as described by Calvo et al . [10] . Briefly , Simplagrin was immobilized on a CM5 sensor chip ( GE Healthcare ) , and different analytes were injected over the sensor for 120 seconds at a flow rate of 20 µL/minute . Complex dissociation was monitored for 400 seconds , and the sensor chip surface was regenerated with a 10 second pulse of 10 mM HCl at 30 µL/minute . Human soluble collagen type I , III , IV , V , or rat tail type I ( 25 µg/mL in PBS; pH 7 . 4 ) were immobilized overnight at 4°C . Wells were washed with PBS and blocked with BSA ( 2% v/v , in PBS ) for two hours . Simplagrin ( 0–3 µM ) diluted in PBS-T ( PBS , 1% BSA , 0 . 05% Tween , pH 7 . 4 ) was added in quadruplicates . After two hours , wells were washed in PBS-T and incubated with anti-Simplagrin ( 1 µg/mL ) in the same buffer . After one hour , wells were washed 3× and incubated with alkaline phosphatase-coupled anti-rabbit IgG ( 1∶5000 , in PBS-T ) . After one hour incubation at 37°C , the wells were washed four times with PBS-T and stabilized p-nitrophenyl phosphate liquid substrate ( Sigma ) was added . Colorimetric analysis was performed by measuring absorbance values at 405 nm . Platelet aggregation was measured as described previously using an aggregometer [10] . Briefly , platelet rich plasma ( PRP ) was prepared from medication free donors by plateletpheresis ( Department of Transfusion Medicine , NIH Clinical Center ) . Diluted PRP ( 1∶3 in Tyrode's buffer ) in the presence or absence of Simplagrin ( 10 µL ) or PBS ( control ) were pre stirred in the aggregometer for three minutes to monitor preaggregation effects . Aggregation was then induced using the following agonists: fibrillar collagen ( 1 or 10 µg/mL ) , phorbol 12-myristate 13-acetate ( 0 . 5 µM ) , ADP ( 10 µM ) , TRAP ( 5 µM ) , U46619 ( 0 . 7 µM ) , arachidonic acid ( 1 mM ) , epinephrine ( 50 µM ) , ristocetin ( 1 mg/mL ) , convulxin ( 100 pM ) , and thrombin ( 0 . 1 U/mL ) . All experiments were carried out in triplicate using PRP from three different healthy donors . Polystyrene plates were coated with 100 µL of collagen type III ( 0 . 03 µg/mL ) , RGQOGVMGF peptide ( 30 µg/mL ) , or a 2% ( w/v ) solution of bovine serum albumin ( BSA ) diluted in PBS for two hours at 37°C . After washing twice with PBS to remove unbound protein , residual binding sites were blocked by adding 5 mg/mL denatured BSA overnight at 4°C . After washing 3× with 50 mM Tris HCl , 150 mM NaCl , and 0 . 05% ( v/v ) Tween 20 , pH 7 . 4 ( TBS-T ) , increasing concentrations of recombinant Simplagrin ( ranging from 0 . 05 to 3 µM ) were added to the wells and incubated at 37°C for one hour . Wells were washed again and incubated with 3 nM of vWF factor VIII free ( Haematologic Technologies Inc . ) in TBS-T supplemented with 2% ( w/v ) BSA . After one hour at 37°C , wells were washed 3× with TBS-T , and a polyclonal rabbit anti human vWF ( DakoCytomation , Glostrup , Denmark ) was added ( 1∶500 in TBS-T ) and incubated for one hour at 37°C . After three washes with TBS-T , an alkaline phosphatase conjugate anti-rabbit IgG ( whole molecule; Sigma ) was added ( 1∶10 , 000 ) and incubated at 37°C for 45 minutes . Before adding stabilized p-nitrophenyl phosphate liquid substrate ( Sigma ) , wells were washed 6× with TBS-T . After 30 minutes of substrate conversion , the reaction was stopped with 3 N NaOH and the absorbance read at 405 nm using a Thermomax microplate reader ( Molecular Devices , Sunnyvale , CA , USA ) . Net specific binding was obtained by subtracting optical density values from wells coated only with BSA from the total binding measured as described above . All experiments were performed in triplicate . Polyclonal antibodies against Simplagrin ( wild type , full length ) were raised in rabbits by Spring Valley Laboratories , Inc . ( Woodbine , MD , USA ) using a standard protocol . Briefly , rabbits were immunized 3× with 125 µg of Simplagrin every 21 days and the serum collected at day 72 . A 10-mL aliquot of rabbit serum ( immunized or naïve ) was diluted to 50 mL in phosphate buffer , pH 6 . 5 , and loaded onto a 5 mL HiTrap protein A HP column ( GE Healthcare ) and the IgG eluted with a linear gradient of citric acid ( 100 mM , pH 3 . 4 ) using an Akta purifier system ( GE Healthcare ) . Fractions containing purified IgG were pooled and dialyzed against 1× PBS for 16 hours at 4°C . IgG quantification was based on 1 absorbance unit at 280 nm = 0 . 7 mg/mL . For western blot analysis , Simplagrin ( 5 µg ) was electrophoresed in a 4–12% NuPAGE in MES buffer ( Invitrogen ) . After electrophoresis , samples were electrotransferred onto nitrocellulose membrane using an iBlot gel transfer system ( Invitrogen ) . The membrane was incubated overnight at 4°C with TBS ( 25 mM Tris , 150 mM NaCl , pH 7 . 4 ) containing 5% ( w/v ) powdered nonfat milk ( blocking buffer ) , followed by incubation for 90 minutes at RT with purified anti-Simplagrin rabbit IgG diluted 1∶1000 in blocking buffer . The membrane was washed 4× with TBS-T and incubated with goat anti-rabbit alkaline phosphatase conjugated ( Sigma ) diluted 1∶10 , 000 in blocking buffer . The immunoblot was developed by addition of 1 mL of Western Blue stabilized substrate for alkaline phosphatase ( Promega , Madison , WI , USA ) . Experiments were performed to detect whether Simplagrin blocks collagen interaction with GPVI , vWF , or integrin α2β1 . Recombinant GPVI ( 25 µg/mL ) in acetate pH 4 . 5 buffer was immobilized on a CM5 sensor with a final surface density of 1 , 753 . 2 RU . A blank flow cell was used to subtract any effect of buffer in the refractory index change . Then different concentrations ( 3 . 175 , 6 . 125 , 12 . 5 , 25 , and 50 µg/mL ) of collagen I alone ( control ) or previously incubated ( 15 minutes at RT ) with 500 nM of Simplagrin in HBS-P buffer was injected over immobilized GPVI for 120 seconds at 20 µL/minute . Complex dissociation was monitored for 400 seconds . Sensor surface was regenerated between runs by a 30 second pulse of glycine solution , pH 1 . 5 . To verify that immobilized GPVI was still active after all the injection and regeneration cycles , 50 µg/mL of collagen I was injected for 120 seconds at a flow rate of 20 µL/minute and the resulting sensogram compared with the one obtained before . A control experiment was carried out using convulxin at different concentrations ( 2 . 5 , 5 , and 10 nM ) incubated with buffer or saturating concentrations of Simplagrin ( 500 nM ) followed by injection of the mixture over immobilized GPVI , as described above . Alternatively , immobilized collagen type III was allowed to interact with Simplagrin alone or preincubated with saturating concentrations of RGQOGVMGF peptide in HBS-P . Contact time , dissociation and regeneration of the sensor chip were carried out as described above . Polystyrene plates were coated with 100 µL of collagen type III ( 0 . 3 µg/mL ) or a 2% ( w/v ) solution of BSA diluted in PBS for two hours at 37°C . After washing twice with PBS to remove unbound protein , residual binding sites were blocked by adding 5 mg/mL denatured BSA overnight at 4°C . After washing 3× with 50 mM Tris HCl , 150 mM NaCl , and 0 . 05% ( v/v ) , TBS-T , pH 7 . 4 , increasing concentrations of recombinant Simplagrin ( ranging from 0 . 0015 to 1 . 5 µM ) were added to the wells and incubated at 37°C for one hour . Wells were washed again and incubated with 3 nM of vWF factor VIII free ( Haematologic Technologies Inc . ) in TBS-T supplemented with 2% ( w/v ) BSA . After one hour at 37°C , wells were washed 3× with TBS-T , and a polyclonal rabbit anti-human vWF ( DakoCytomation ) was added ( 1∶500 in TBS-T ) and incubated for one hour at 37°C . After three washes with TBS-T , an alkaline phosphatase conjugate anti-rabbit IgG ( whole molecule; Sigma ) was added ( 1∶10000 ) and incubated at 37°C for 45 minutes . Before adding the stabilized p-nitrophenyl phosphate liquid substrate ( Sigma ) , wells were washed 6× with TBS-T . After 30 minutes of substrate conversion , the reaction was stopped with 3 N NaOH and absorbance read at 405 nm using a Thermomax microplate reader ( Molecular Devices ) . Wells coated only with BSA were used as assay blanks . All experiments were performed in triplicate . Coverslips ( 22×22 mm , no . 0 ) were treated with H2SO4: H2O2 ( 4∶1 ) for 20 minutes to remove contaminants , followed by ultrasonic washing with deionized water and ultraviolet cleaning . Coverslips were coated with fibrillar ( 100 µg/mL; Chronolog ) or soluble collagen type III ( 100 µg/mL ) for ten minutes , rinsed in deionized water , and incubated overnight with denaturated BSA ( 4 mg/mL ) . Coverslips were treated with 150 µL of Simplagrin ( 0–5 µM ) for 15 minutes and excess removed by inversion . A coated coverslip formed the bottom of the parallel plate flow chamber ( Glycotech ) , and a silicone rubber gasket determined the flow path height of 254 µm . Anticoagulated blood ( 50 µM PPACK ) was mixed with Simplagrin and aspirated using an infusion/withdrawal pump ( Model 940; Harvard Apparatus ) through the flow chamber at a flow rate of 0 . 6 mL/minute , producing a shear rate of 1 , 500 s−1 as described before [10] . Blood was perfused for 240 seconds , followed by washing with Tyrode's buffer to remove unbound platelets and red blood cells . Platelet adhesion under flow conditions was recorded using a Leica DMI6000 microscope ( Leica Microsystems , Inc . , Bannockburn , IL , USA ) using 100x objective with NA = 1 . 30 , and an ORCA ER digital camera ( Hamamatsu Photonic Systems , Bridgewater , NJ , USA ) . Image acquisition and the digital camera were controlled by ImagePro 5 . 1 software ( Media Cybernetics , Silver Spring , MD , USA ) . Female mice ( C57BL/6 , 20–25 g weight ) were anesthetized with intramuscular xylazin ( 16 mg/kg ) followed by ketamine ( 100 mg/kg ) . The right common carotid artery was isolated through a midline cervical incision , and blood flow was continuously monitored using a PRB Doppler flow probe coupled to a TS420 flow meter ( Transonic Systems , Ithaca , NY , USA ) . Fifteen minutes before induction of thrombosis , animals were injected in the cava vein with Simplagrin ( 20 or 100 µg/kg ) or PBS ( control ) . Thrombosis was induced by slow injection ( over 2 minutes ) of 90 mg/kg body weight of rose Bengal dye ( Fisher Scientific , Pittsburgh , PA , USA ) into the cava vein at a concentration of 60 mg/mL . Just before injection , a 1 . 5 milliwatt ( mW ) , 540 nm green-light laser ( Melles Griot , Carlsbad , CA , USA ) was applied to the desired site of injury from a distance of 3 cm . Mean carotid artery blood flow was monitored for 80 minutes or until stable occlusion occurred , at which time the experiment was terminated . Following in vivo thrombus formation measurements in mice , injured carotid arteries were excised and fixed in 2 . 5% glutaraldehyde in 0 . 1 M cacodylate buffer , pH 7 . 2 , and shipped to Histoserv Inc . ( Germantown , MD , USA ) for histologic sectioning and staining with hematoxylin and eosin . Samples were examined by light microscopy in a Zeiss-LSM 510 microscope . BALB/c mice ( female , 15–17 g body weight ) were anesthetized as above with a combination of xylazine and ketamine ( 16 and 100 mg/kg , respectively ) and Simplagrin ( 20 or 100 µg/kg ) or PBS in a final volume of 100 µL were administered intravenously via tail vein . After 15 minutes , the mouse tail was cut within 2 mm of diameter and carefully immersed in 40 mL distilled water at RT . The hemoglobin content in water solution ( absorbance at 540 nm ) was used to evaluate blood loss . Results are expressed as means ± SEM . Statistical significance was determined using Student's t-test or analysis of variance ( Bonferroni post-test comparison ) using GraphPad v6 . 0 ( GraphPad Software Inc . , San Diego , CA , USA ) . Significance was set at P<0 . 05 .
Previous black fly sialotranscriptome analysis [14] revealed the presence of expressed sequence tags ( ESTs ) distantly related to Aegyptin , a mosquito platelet aggregation inhibitor [10] . Cluster Sim-50 ( Simplagrin ) is the fourth most abundant transcript in the SGs of S . nigrimanum and accounts for approximately 3% of all sequenced ESTs ( 36 of 1215 ) . The cDNA of Simplagrin has an open reading frame of 861 bp coding for a protein of 286 amino acids ( aa ) . It has a predicted signal peptide [19] of 20 aa , indicative of secretion . The predicted mature Simplagrin has a calculated molecular mass of 28 , 828 . 88 Da , with an isoelectric point of 4 . 02 with six potential N-glycosylation sites . Alignment of Simplagrin with Aegyptin ( Figure 1A ) shows an overall amino acid identity of 25% . Although having small amino acid conservation , these black fly proteins recognize mosquito proteins including Aegyptin by psiblast , and have a non-promiscuous motif G-x ( 27 , 30 ) -L-x-S-x ( 5 ) -L-Q-x ( 16 , 17 ) -S-x-I-x ( 2 ) -C-F-x ( 20 ) -C-x ( 3 , 9 ) -C at their carboxyterminus that is common to mosquito and black fly proteins . A more detailed phylogenetic analysis of Aegyptin-like proteins from Nematocera shows robust clades for Anopheline , Culicine , Simulium , and Phlebotomus [2] . Three-dimensional modeling of Simplagrin shows that this protein has a low complexity/disorganized aminoterminal region with a relatively more complex carboxyterminus domain ( Figure 1B and C ) . Recombinant Simplagrin was expressed in HEK293 cells and purified by affinity and size exclusion chromatography ( Figure 2A ) as described by Calvo et al . [10] . Identity and purity of the recombinant protein was verified by N-terminal sequencing and liquid chromatography-mass spectrometry analysis ( not shown ) . Protein A-purified polyclonal antibodies raised against Simplagrin in rabbits recognized the recombinant protein in western blot ( Figure 2A , inset ) . Circular dichroism spectroscopy ( CDS ) analysis of Simplagrin shows that it comprises primarily an α-helix ( 56% ) followed by 29% of remainder/disorganized , 15% of β-sheet , and 4% β-turn structures ( Figure 2B ) . This distribution of secondary structures is in agreement to the modeled 3D structure of Simplagrin ( Figure 1B and C ) . Although the predicted molecular mass of Simplagrin is 30 . 25 kDa ( including the 6xHis-tag ) , it is eluted at a higher apparent molecular mass of 160 kDa when loaded on a size-exclusion column ( Figure 2C ) . This abnormal chromatographic pattern suggests that Simplagrin could be oligomeric or non-globular in nature . To further investigate this feature , we used dynamic light-scatter plotting to analyze the hydrodynamic radius of Simplagrin . Our results show that Simplagrin has an elongated monomeric form with a hydrodynamic radius of 5 . 6 nm and a calculated mass of 32 kDa ( Figure 2D ) . This difference between the expected and calculated molecular mass of 30 . 25 kDa may be due to incomplete separation of the aggregate peak ( 7% ) and the monomer peak ( 93% ) . No glycosylation was found when Simplagrin was treated with an enzymatic deglycosylase mix ( Figure 2E ) . A similar result was reported for recombinant Aegyptin [10] . Because Simplagrin appears to belong to the Aegyptin family of salivary proteins , we carried out a collagen-binding assay using SPR , solid-phase experiments , and fluorescence microscopy . For SPR experiments , Simplagrin was immobilized on a CM5 sensor chip and different analytes were flowed over the sensor's surface . Despite the low similarity between Aegyptin and Simplagrin at the amino acid level , SPR analysis identified collagen ( human types I , III , IV , V , and VI , and type I from rat tail ) as the molecular partner of Simplagrin ( Figure 3A ) . No detectable binding was observed to vitronectin , fibronectin , fibrinogen , vWF , recombinant platelet receptors GPVI and integrin α2β1 , and coagulation factors IIa and Xa ( Figure 3A ) . No effect of Simplagrin on blood clotting , vasodilation , inflammation , or murine splenocyte proliferation was found ( not shown ) . We also verified the collagen binding activity of Simplagrin using a solid phase . For this experiment , collagen-coated wells were incubated with different concentrations of Simplagrin and the binding activity measure by an ELISA based assay utilizing anti-Simplagrin antibodies . Figure 3B shows results for the solid phase assay similar to those found with SPR analysis . Alternatively , FITC labeled Simplagrin was used to visualize its binding to fibrillar collagen . For this experiment , FITC labeled Simplagrin ( 1 µM ) was allowed to interact with fibrillar collagen ( immobilized on a glass coverslip ) . After 15 minutes of incubation at RT , the coverslips were washed five times with TBS and analyzed under bright field and fluorescence microscope . Figure 3C ( upper panel ) shows that FITC labeled Simplagrin binds to the collagen fibrils immobilized on the cover slip . Collagen incubated with buffer alone did not show autofluorescence under the same conditions ( Figure 3C , lower panel ) . Due to the collagen binding activity displayed by Simplagrin , a more detailed analysis of Simplagrin-collagen interaction was carried out to calculate the kinetic and thermodynamic constants of this interaction . Immobilized collagen type I and III were used as a ligand , and Simplagrin at concentrations ranging from 0 . 015–1 µM ( serial dilutions ) was flowed over the sensor for 180 seconds . Simplagrin-collagen complex dissociation was monitored for 30 minutes , and the resulting sensograms were fitted using a 1∶1 binding model . Sensograms of a typical kinetic experiment are shown in Figure 4A and B . Using SPR , we calculated the KD values for Simplagrin collagen I and III , of 5 . 51±0 . 52 nM and 2 . 10±0 . 35 nM , respectively ( Table 1 ) . Affinity of Simplagrin for human collagen type I , III , IV , V , VI and type I of rat tail was also calculated using Simplagrin as ligand , and the resulting sensograms fitted with a steady-state model ( Figure 5 ) . The KD values for these collagens are in the same order of magnitude as those of Simplagrin collagen I and III ( Table 2 ) . For thermodynamic analysis of Simplagrin-collagen interaction , different concentrations of Simplagrin ranging from 25 nM to 500 nM were injected over immobilized collagen type I or III at different temperatures ( 15 , 20 , 25 , 30 , 35 , and 40°C ) . The association ( ka ) and dissociation ( kd ) rate constants , as well as the affinity constant ( KD ) , were obtained and fitted to a linear form of the van't Hoff equation to estimate the ΔH , ΔS , and ΔG ( Figure 4C and D ) . Thermodynamic analysis of the Simplagrin-collagen interaction shows that this interaction is both entropically and enthalpically driven with a larger negative ΔG , indicating that this interaction is favorable and occurs spontaneously ( 3 ) . CDS analysis of Simplagrin-collagen interaction shows that no conformation change occurs during complex formation ( Figure 6 ) . Because collagen is recognized as the most thrombogenic component of the subendothelial matrix and the collagen-binding activity displayed by Simplagrin , we investigated whether Simplagrin has any effect in collagen-induced platelet aggregation . As shown in Figure 7A , Simplagrin ( 1 µM ) inhibited human platelet-rich plasma ( PRP ) aggregation induced by a lower concentration of collagen ( 1 µg/mL ) , with 0 . 16 µM causing only a small delay in the shape change of platelets; however , Simplagrin ( 1 µM ) failed to inhibit platelet aggregation when a higher dose of collagen ( 10 µg/mL ) or collagen-related peptide ( CRP ) ( 5 µg/mL ) was utilized as platelet aggregation agonist . Simplagrin ( 1 µM ) shows no effect in platelet aggregation induced by other agonists , including the thromboxane A2 analog U46619 and CRP ( Figure 7B ) . Although the current evidence suggests that the primary signaling of collagen induced platelet aggregation occurs via GPVI , other studies have shown that platelet response to low collagen concentrations is highly aspirin sensitive and therefore thromboxane A2 mediated [20] . Our results suggest that Simplagrin inhibition of collagen-induced platelet aggregation does not involve direct blocking of GPVI-collagen interaction . Rather , the inhibition might be due to steric hindrance , assuming that Simplagrin binds to a collagen region other than that to which GPVI binds . Multiple collagen receptors have been identified on the platelet surface , namely GPVI , integrin α2β1 , and GPIbα through collagen bound vWF complex [21]–[23] . In an attempt to identify the collagen region targeted by Simplagrin , SPR- and ELISA-based assays were designed . For SPR analysis , Simplagrin was immobilized on a CM5 sensor chip , and synthetic collagen-derived peptides were flowed over the sensor for 90 seconds at 20 µL/minute . As a positive control , collagen type I was also flowed over immobilized Simplagrin . Figure 8A shows that Simplagrin specifically binds to collagen ( positive control ) and the vWF collagen receptor peptide RGQOGVMGF; however , no detectable binding was observed when CRP [ ( GPO ) 10] or Iα2β1 ( GFOGER ) collagen-derived peptides were flowed over immobilized Simplagrin . For solid-phase analysis , a 96-well plate was coated with collagen type I ( positive control ) or RGQOGVMGF collagen-derived peptide . Wells were incubated with different concentrations of Simplagrin , and the binding capacity of Simplagrin to these two molecules was detected with purified rabbit anti-Simplagrin antibodies as described in Methods . Identical results were obtained using this method , confirming that Simplagrin specifically binds to RGQOGVMGF collagen-derived peptide ( Figure 8B ) . Next we calculated the Simplagrin RGQPOVMGF affinity constants using SPR . To compare affinity of Simplagrin RGQOGVMGF peptide to that of Simplagrin collagen , kinetic analysis was carried out . For this experiment , the ka , kd , and KD constants were calculated with Simplagrin as ligand and RGQOGVMGF peptide as analyte ( Table 4 ) . Figure 8C shows the sensograms of a typical kinetic experiment . The calculated KD value ( 11 . 1±0 . 59 nM ) for Simplagrin-RGQOGVMGF collagen-derived peptide interaction is in the same order of magnitude as that of Simplagrin collagens ( Table 4 ) . These results show that Simplagrin specifically binds to the RGQOGVMGF sequence in collagen . Because collagen-derived peptide RGQOGVMGF and collagen were reported to be specific ligands for vWF , we determined the ability of Simplagrin to compete with collagen for vWF binding . For this assay , vWF was immobilized on a CM5 sensor chip , and collagen and RGQOGVMGF peptide were allow to interact with immobilized vWF in the presence or absence of saturating concentrations of Simplagrin . Figure 9A shows that Simplagrin can indeed block the interaction of collagen or RGQOGVMGF to immobilized vWF . No direct binding of Simplagrin to vWF was detected . Alternatively , Simplagrin was preincubated with saturating concentrations of RGQOGVMGF peptide and flowed over immobilized collagen type III . Figure 9B shows that Simplagrin binding to collagen is abrogated in the presence of RGQOGVMGF peptide indicating that Simplagrin binding site in collagen is the RGQOGVMGF sequence . As an orthogonal method , an ELISA was carried out to investigate the vWF collagen-blocking capacity of Simplagrin . For this experiment collagen-coated wells were preincubated with different concentrations of Simplagrin before vWF was added to each well . The blocking assay was assessed by the reduction of bound anti-vWF antibodies . Figure 9C shows that Simplagrin blocks , in a dose-response manner , the interaction of vWF collagen , with a calculated IC50 of approximately 0 . 1 µM . Finally , SPR was used to investigate whether Simplagrin interferes with collagen-GPVI interaction . For this experiment , GPVI was immobilized on a CM5 sensor chip followed by injection of collagen I or CRP , preincubated with or without saturating concentrations of Simplagrin . Figure 9D shows that Simplagrin does not completely abrogate collagen binding to immobilized GPVI even at 1∶20 and 1∶40 collagen:Simplagrin molar ratios . Furthermore , Simplagrin failed to block CRP-GPVI interaction . On the other hand , convulxin GPVI interaction was not affected by presence of Simplagrin ( Figure 9E ) . No direct binding of Simplagrin to GPVI was detected . This result further supports that Simplagrin binds preferentially to the RGQOGVMGF sequence in collagen and also explains the lack of platelet aggregation inhibition at higher concentrations of fibrillar collagen ( 10 µg/mL ) . This competition assay reinforces the hypothesis of steric hindrance of platelet aggregation inhibition of Simplagrin at low concentrations of fibrillar collagen . The initial steps in the hemostatic cascade include platelet interaction with the exposed extracellular matrix at sites of injury as well as ADP released by damaged cells . On vascular injury at sites of high-shear rates , the platelet integrin GPIbα interacts with collagen-bound vWF to initiate the tethering of circulating platelets to the vessel wall . Tethered platelets subsequently roll on the damaged vessel wall , a process that is amplified by activation of platelet integrin GPIIb/IIIa [24] . Because Simplagrin blocks collagen vWF interaction , we investigated whether Simplagrin was able to reduce platelet adhesion to collagen under high wall shear stress . For this experiment , anticoagulated whole blood containing different concentrations of Simplagrin ( 0 , 0 . 18 , 0 . 37 , 0 . 75 , 1 . 5 and 3 µM ) was perfused for 180 seconds over fibrillar collagen coated coverslips at high shear rates ( 1500 s−1 ) followed by continuous perfusion of Tyrode's buffer for 180 seconds at the same shear rate . Figure 10 shows that adhesion of platelet to fibrillar collagen was reduced in a dose-dependent manner with an IC50 of approximately 0 . 75 µM , with complete abrogation of platelet tethering to collagen fibers at 3 µM . This result is in agreement with the SPR and solid-phase analyses showing that Simplagrin inhibits collagen vWF interaction under static conditions . We did not find any effect on platelet adhesion under static conditions using calcein-labeled platelets ( not shown ) . The antithrombotic effect of Simplagrin was investigated using a photochemically induced thrombosis model in C57BL/6 mice . Figure 11A shows that arterial blood flow of control mice treated with PBS alone stopped in 26±1 . 78 minutes ( 25–29 minutes ) . On the other hand , mice treated with Simplagrin at 20 and 100 µg/kg had their blood flow stopped at 30±2 . 09 ( 26–30 minutes ) and 46 . 8±4 . 4 min ( 42–55 minutes ) , respectively . The prolongation of time to arterial occlusion was statistically significant ( P<0 . 05 ) at 100 µg/kg . Considering that the most widely used anti-platelet agents ( e . g . , aspirin , heparin , and Clopidogrel , among others ) carry the risk of increased bleeding rate , we investigated whether Simplagrin can increase the bleeding rate in mice . Evaluating the effect of Simplagrin in tail-bleeding assays would provide an additional measure of its antihemostatic effect in vivo . Tail bleeding was determined in BALB/c mice after intravenous injection of PBS ( control ) or Simplagrin ( 20 and 100 µg/kg ) . No significant difference in tail bleeding was found in mice treated with Simplagrin or PBS ( Figure 11B ) . Our results demonstrate that Simplagrin can inhibit collagen-induced platelet aggregation and adhesion without affecting general hemostasis . Although platelet aggregation inhibition in general can cause prolongation of bleeding time , mouse tail bleeding is highly sensitive to levels of coagulation factors [25] .
For blood feeding arthropods , the vertebrate hemostatic responses represent an important barrier for acquiring a successful blood meal . For this reason , blood feeders have evolved salivary secretions rich in molecules that affect hemostasis , including vasodilators and inhibitors of blood clotting and platelet aggregation . Among the platelet inhibitors , antagonists of collagen induced platelet aggregation and adhesion have been found in SG of ticks and other hematophagous animals ( reviewed in [2] ) . It was recently reported that An . stephensi and Ae . aegypti each express a salivary collagen-binding protein that inhibits collagen-induced platelet aggregation and adhesion . Both molecules ( AAPP from An . stephensi and Aegyptin from Ae . aegypti ) were shown to block collagen interaction with its three major ligands , GPVI , vWF , and integrin α2β1 [10]–[12] . The Aegyptin family of salivary proteins has also been described in the Diptera suborder Nematocera , where anophelines and Simulium appear to have a single gene coding for this family , while culicines have multiple genes [2] . Comparisons between Anopheles , Culex , Aedes , and now Simulium sialomes ( from the Greek sialo = saliva ) showed that each species contained genus-specific salivary protein families and even subgenus-specific families , suggesting that the evolution of salivary proteins has occurred at a very fast pace , possibly caused by the immune pressure of their hosts [26] . The mechanism of action of Simplagrin can be explained by its binding to collagen , blocking vWF-collagen interaction . By blocking this interaction , Simplagrin ensures a delay in platelet activation and aggregation , especially in small arteries and arterioles where shear force is rather high ( e . g . , 1500s−1 ) and the interaction between GPIb complex and vWF is crucial . A possible explanation for the platelet aggregation-inhibition effect of Simplagrin at low concentrations of fibrillar collagen can be proposed on the basis that collagen has multiple binding sites for GPVI , vWF , and integrin α2β1 , all of them mapped in close proximity [23] . The complexity of fibrillar collagen organization may also facilitate interaction between receptors bound to adjacent or nearby collagen monomers within a fiber [23] . Taking into consideration that Simplagrin displays an elongated , non-globular conformation , steric hindrance seems to be responsible for its platelet aggregation inhibition at low collagen concentrations . Interestingly , the C-terminal domain of Simplagrin also binds to collagen but with significantly lower affinity for collagen type I and III ( KD 284 . 7±23 . 8 nM and 162 . 3±5 . 16 nM , respectively ) when compared to that of Simplagrin to collagen I and III ( KD 5 . 6±0 . 52 nM and 2 . 1±0 . 35 nM , respectively ) . As expected from this low affinity for collagens , the C-terminal domain was unable to inhibit collagen-induced platelet aggregation; however , it causes a delay in the shape change in platelets activated with collagen ( Figure S1 ) . These results are in agreement previously published work demonstrating that the C-terminal domain of Aegyptin and AAPP is responsible for collagen binding and platelet aggregation inhibition[11] , [27] . The overall difference in platelet aggregation and adhesion inhibition between Simplagrin and Aegyptin ( Table 5 ) could be explained by the conformation change occurring upon collagen-Aegyptin interaction ( Figure 7A and B ) . “In solution” analysis results showed that collagen undergoes unwinding caused by Aegyptin as determined by CDS; however , no significant conformational change was detected in Simplagrin-collagen complex . Fibrillar collagen can assemble stable triple helices , which in turn can form a complex 3D fibrous superstructure needed for platelet receptor binding to collagen and platelet activation ( reviewed in [22] , [28] ) . It was proposed that blood feeding success through inhibition of platelet aggregation is a vital salivary function in blood feeding arthropods [29] . These authors hypothesized that faster probing and feeding times would reduce the duration of vector-host contact and hence increase survival of the feeder . In this scenario , unwinding of collagen by Aegyptin may represent an evolutionary advantage over Simplagrin , making it a more efficient inhibitor and resulting in shorter probing and feeding time in Ae . aegypti mosquitoes [30]; however , the resulting antihemostatic effect of saliva—deriving from dozens of protein families , many still with unknown function—should not be attributed to a single molecule . It has also been proposed that Culicidae and Simuliidae families diverged in the middle Triassic , sharing a common blood- ( or insect hemolymph ) - feeding ancestor approximately 250 MYA . Accordingly , this common ancestor originated before the irradiation and expansion of birds in the Jurassic ( 200 MYA ) and of mammals 60 MYA . As the Culicomorpha evolved to produce mosquitoes ( Culicidae ) , frog-feeding flies ( Corethrellidae ) , biting midges ( Ceratopogonidae ) , and black flies ( Simuliidae ) , land vertebrates evolved to produce birds and mammals . Within this scenario , the blood feeding sialome evolved within each fly species in concert with their vertebrate “hemostome” and “immunome , ” the latter battering the salivary proteins by either neutralizing their function or creating inflammatory reactions such as pain or itching that , in concert with host behavioral defenses , could disrupt feeding . This scenario of concerted “birth and death” evolution of multigene families [31] could explain the evolutionary speed of the blood feeding sialome as well as the recruitment or exaptation of new gene families [15] . Perhaps a collagen-binding salivary protein was an evolutionary innovation present in an ancient dipteran ancestor that evolved into the current Aegyptin-like protein family commonly found in sialomes of Culicomorpha , excluding Corethrella appendiculata [32] . It might be possible that this salivary function has been substituted by a different gene family in C . appendiculata SGs . In conclusion , our results show that Simplagrin is a nonglobular , elongated collagen-binding protein that prevents platelet adhesion under high shear stress . Simplagrin binds directly to the von Willebrand binding sequence RGQOGVMGF in collagen , retaining the same biologic function as the mosquito Aegyptin-like proteins but with a distinct mechanism of inhibition . Despite the low similarity between Simplagrin and Aegyptin at the amino acid level , the possibility of convergent evolution of the Aegyptin family cannot be excluded since the proposed the mechanism ( s ) of action differ between both molecules . Due the overall similarity in the modelled 3D structure of Simplagrin and Aegyptin ( Figure S2 ) , it might be argued that even structures could have converged at this low level of amino acid identity . Our results support the common origin of hematophagy in mosquitoes and black flies as proposed by Grimaldi and Engel [16] , with the presence of a unique protein family with conserved primordial function found in black flies and mosquitoes . | Blood feeding arthropods—like mosquitoes and black flies—have evolved salivary secretions rich in molecules that affect hemostasis , including vasodilators and inhibitors of blood clotting and platelet aggregation . Among the platelet inhibitors , antagonists of collagen-induced platelet aggregation and adhesion have been found in salivary glands of blood feeders . Here we report the first collagen-binding protein from salivary glands of a black fly . This molecule prevents thrombosis in mice without causing significant bleeding , making it an attractive candidate as an antithrombotic agent . Because blackflies and mosquitoes shared a common blood feeding ancestor approximately 250 million years ago , it appears that collagen-binding activity in salivary glands was an evolutionary innovation present in an ancient dipteran ancestor . Our work highlights the central role of inhibition of platelet aggregation as a vital salivary function in blood feeding arthropods . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biochemistry",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"proteins",
"epidemiology",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"disease",
"vectors"
] | 2014 | Simplagrin, a Platelet Aggregation Inhibitor from Simulium nigrimanum Salivary Glands Specifically Binds to the Von Willebrand Factor Receptor in Collagen and Inhibits Carotid Thrombus Formation In Vivo |
Subjects typically choose to be presented with stimuli that predict the existence of future reinforcements . This so-called ‘observing behavior’ is evident in many species under various experimental conditions , including if the choice is expensive , or if there is nothing that subjects can do to improve their lot with the information gained . A recent study showed that the activities of putative midbrain dopamine neurons reflect this preference for observation in a way that appears to challenge the common prediction-error interpretation of these neurons . In this paper , we provide an alternative account according to which observing behavior arises from a small , possibly Pavlovian , bias associated with the operation of working memory .
Animal behavior all too rarely follows the precepts of simple theories such as normatively optimal choice . Prominent examples of this arise in the florid fancies of Breland & Breland's animal actors [1] , or in the complexities of negative automaintenance or omission schedules [2]–[4] . Such failures and irrationalities have been important sources of theory revision and refinement , for instance leading to suggestions about the competition and cooperation of multiple systems of control [5]–[7] , some instrumental and adaptive; others Pavlovian and hard-wired . In this paper , we study one apparent departure from optimality , namely a type of ‘observing behavior’ [8] , [9] , which has been the subject of a recent important electrophysiological study [10] . In brief , subjects are programmed to receive either a large or small reward , with its size being determined stochastically . When faced with the choice of finding out ( by being presented with a suitably distinctive cue ) sooner rather than later which of the two rewards they will ultimately receive , subjects prefer to know sooner . A lack of indifference despite the equality of the outcomes has been found to be widely true even if the knowledge cannot influence the outcome , and , at least in other experiments , even if this choice is expensive [8] , [9] , [11]–[13] . In economics , the same anomaly is referred to in terms of “temporal resolution of uncertainty” [14] , explained by such notions as savoring [15]–[17] , with subjects enjoying the anticipation of good things to come . The correct interpretation of this form of observing behavior has been the subject of substantial debate ( see , e . g . [9] ) . Superficially attractive theories , such as a desire to gain Shannon information [18] have been dealt fatal blows , for instance with animals preferring to observe more even when the number of bits they receive by doing so is less ( e . g . , as the probability of getting the large reward becomes smaller than , [12] ) . A recent study on observing behavior in macaques [10] has offered a new perspective on the problem . These authors recorded from putative dopamine neurons in the midbrain whilst monkeys chose to observe . According to a common theory , these neurons report a temporal difference error in predictions of future reward [19] , [20] as in reinforcement learning accounts of optimal instrumental choice [21] . Bromberg-Martin and Hikosaka [10] showed that: ( a ) the macaques did observe; and furthermore ( b ) the activity of dopamine neurons was associated with the choice they make . However , although the behavior and activity are mutually consistent , observing behavior offers no instrumental benefit and therefore it should also not be associated with any prediction errors . Bromberg-Martin and Hikosaka suggested that this means that the dopamine cells are reporting on some aspects of the benefit of information gathering in addition to aspects of reward . In this paper , we examine the extent to which this form of observing behavior can be explained by temporal difference learning , coupled with the same mechanism that provides an account of a wide range of departures from normative choice , namely a Pavlovian influence over instrumental actions [4] . In particular , we assume that subjects only make associative predictions when they are appropriately engaged in the task . If the level of this engagement is influenced by the size of the predictions ( the putatively Pavlovian effect ) , then stimuli predicting certain or deterministic large future rewards ( one outcome of an observing choice ) will lead to more engagement than stimuli that leave uncertain the magnitude of the future rewards . This idea can be seen as a realization of the suggestion made by Dinsmoor [9] that the predictions of future reward associated with stimuli influence the attention paid to them . We show that occasional failures of engagement , modeled as a breakdown in the working memory for the representational state , can lead directly to both the preference for observing and the apparently anomalous dopamine activity , without need for any reference to ‘information’ . We also examine the various factors that control the strength of observing in this model .
Bromberg-Martin and Hikosaka's experiment ( see Methods and Figure 1 ) involved the most precise conditions for establishing observing behavior . On each trial , thirsty subjects had a 50% chance of receiving a small or large volume of water directly into their mouths . There were three sorts of trials: forced-information , forced-random and free choice . On forced-information trials , the subjects were presented with a single target ( C; just an orange square in the figure ) and , after looking at it , would receive one of two cues ( S; an orange ‘+’ , or S; an orange ‘’ ) according to the volume they were to receive in a couple of seconds . On forced-random trials , looking at the single target ( C; green square ) led again to one of two cues ( S; green ‘*’ , or S; green ‘o’ ) . However , either of these could be followed by either small or large rewards; and thus they provided no discriminative information about the forthcoming reward . Finally , on free choice trials , both orange and green targets were provided , and the subjects could choose whether to receive the discriminative ( orange ) or non-discriminative ( green ) cues . Figures 2a;b show primary behavioral results from the study for two subjects – both gradually expressed a bias towards the discriminative ( orange ) option in the free-choice trials . As Bromberg-Martin and Hikosaka stressed , under a standard associative learning or temporal difference scheme , there is no difference between the expected reward for the discriminating and non-discriminating option , and so no reason to expect this strong and enduring preference . We built a model of this which , with one critical exception that we discuss below , involves a standard temporal difference learning algorithm [21] , [22] . Forced-choice and free-choice trials permit learning about the future expected rewards associated with the various targets and stimuli , training the values of the states . Then , on free-choice trials , the selection depends on the relative values , via a softmax function ( see methods ) . Figure 2c;d shows the results from simulations of our model , with parameters chosen to match Bromberg-Martin and Hikosaka's two subjects . The model closely matches qualitative features of the monkeys' performances . In standard models such as this , in which there is a delay between the presentation of cues and the rewards that they predict , an assumption has to be made about the way that the subjects maintain knowledge about their state in the task , and indeed keep time . Many different possibilities have been explored , from delay lines to complex patterns of activity evolving in dynamical recurrent networks ( e . g . , [23]–[28] ) . All of these amount to forms of working memory – and so present the minimal requirement that the subjects continue to be engaged in the task throughout the delay in sufficiently intense a manner as to maintain this ongoing memory . Thus the critical exception to conventional temporal difference learning in our model is to assume that this maintained engagement is influenced by the current predicted value . That is , if the value is high , then engagement is readily maintained; if the value is low , then engagement can be weakened or lost . Losing engagement is detrimental to the subject in the context of the present task; by analogy with a similarly detrimental effect in negative automaintenance , we consider it a form of Pavlovian misbehavior [4] . Pavlovian responses are typically elicited in an automatic manner based on appetitive or aversive predictions , and can exert benign or malign influences over the achievement of subjects' apparent goals . Normally , such responses are overt behaviors; here , along with several recent studies [29] , [30] , we consider internal responses , associated with the operation of working memory . Mechanistically , these could come , for instance , from the influence dopamine itself exerts on the processes concerned [31] . In the model , we consider engagement to be lost completely on some trials as a stochastic function of the evolving predicted value . Such losses have the effect of decreasing the subjective value of cues and states associated with lower values below their objective worth; in particular exerting a negative bias on the non-discriminative cues ( S; S ) compared with the discriminative cue associated with the large reward ( S ) , which will more rarely experience such losses . Figure 3 shows the effective probability of disengagement at different timepoints as well as showing the effect this has on the expected reward . Disengagement associated with S is benign , since the outcome on those trials is modelled as being close to in any case . Altogether , this creates a bias towards choosing the discriminative option on free-choice trials , as is evident in Figure 2c;d . The difference between the parameters for Figures 2c;d is in the parameter governing the strength of the competition in the softmax ( and for Figure 2c;d respectively ) . Monkey V's results are consistent with a larger value of than monkey Z; smaller leads to more stochasticity and a lower overall degree of preference . The asymptotic preference for observing is monotonic in . Bromberg-Martin and Hikosaka [10] also recorded the activity of putative midbrain dopaminergic cells during the performance of the task . Figure 4a shows the activity of an example neuron in the various conditions . The population response is similar ( Figure 4 of [10] ) albeit , as has often been seen , with an initial brief activation to the forced choice non-discriminative case , likely because of generalization [32] . Firing at the time of the discriminative or non-discriminative cues ( marked ‘cue’ ) and the delivery or non-delivery of reward ( ‘reward’ ) is just as expected from the standard interpretation of these neurons , i . e . , that they report the temporal difference prediction error in the delivery of future reward [19] , [20] . However , it is their activity at the time of the targets indicating the forced-informative or forced-random trials ( marked ‘target’ ) that is revealing about observing . The target indicating a forced-informative trial was associated with a small but significant phasic increase in activity; whereas that indicating the random cues was followed by a small decrease in the firing rate . Under the temporal difference interpretation of the neurons , this is consistent with the preference exhibited by the monkeys , but not with the objective value of the options . Figure 4b shows modelled dopamine activity in the variable engagement temporal difference model ( here , negative prediction errors have been compressed compared with positive ones , see methods; [33] , [34] ) . This shows exactly the same pattern shown in the monkey data . Note that , once the subject has learned the associations and learned the preference for choosing the discriminative option in the free choice trials , these trials will overall be more frequent than the forced-random trials , and so the negative prediction error associated with the latter will be larger than the positive prediction error associated with the former . Figure 5 decomposes the modelled responses in the cases that there is successful and failed engagement between cues and reward or non-reward . The most significant effect of the complete failure to engage given an non-discriminative cue , is that if the large reward is provided , then there is a greater response than expected from a 50% prediction . The possibility of using this to test the theory is discussed below . In a version of the task that involved choice between immediate or delayed information about upcoming rewards , Bromberg-Martin and Hikosaka [10] further showed that switching the colors of the cues without warning led to a slow reversal of the observing choice ( Figure 6a;b ) . Figure 6c;d shows the same for the model using identical softmax parameters to those in Figure 2c;d . The switch in preference evolves at a similarly glacial pace . Various other features of observing can be examined through the medium of the model . Figure 7a;b show the consequence of the reinforcing outcome being aversive ( e . g . , an electric shock ) rather than appetitive . One key question in this case is whether failure to engage is controlled more by salience or valence . Figure 7a shows the former case , for which a prediction of a large punishment also protects engagement ( symmetrically with reward; inset plot ) . In this case , subjects prefer the random to the discriminative cues , since disengagement leads to subjective preference . Such preference for random cues might also come from adding a fixed value to all the potential rewards , thus allowing the moderately large disengagement in S to have a subtractive value on its expected values ( Bromberg-Martin , personal communication , 2010 ) . However such an effect would likely be small . Figure 7b shows the case in which valence ( from appetitive to aversive ) determines disengagement , with predictions of punishments leading to more failures of engagement than small rewards . This again supports observing behavior . Unfortunately , experimental tests of the case involving punishment [35] have not enjoyed the precision of the paradigm adopted by Bromberg-Martin and Hikosaka , leaving open the question as to which of these patterns arises . Another important experimental manipulation has been to vary the probability of the larger versus the smaller reward . As decreases from 1 towards 0 . 5 there is an increase in the observing bias ( i . e . , a greater tendency to choose the discriminative option ) . Below this , the nature of the bias depends on the assumption about how the choices are generated . A choice rule that depends on the difference in expected values ( ) leads to a bias that ultimately decreases towards as these values themselves decrease towards . However , the bias is asymmetric about ( black curve in Figure 7c ) . If , instead , the choices are based on the ratio of the values ( ) , the choice bias can continue to increase as approaches ( red curve ) . Just such an increase in observing was shown by Roper and Zentall [12] as reward schedules thinned . While some studies have also manipulated the size of the reward [36]–[38] , our model does not make any direct predictions about this . It is possible that adaptation would scale the response to the overall sizes of available rewards ( as indeed found for phasic dopamine activity in [39] ) , and the metrics of this would have to be known in order to make predictions about disengagement . One extra factor that is important for analysing behavior is that the biases inherent in disengagement are small and develop over a long time-scale , consistent with the stately progress evident in Figure 2 . However , this means that the initial course of learning can be subject to significant influence from the initial values ascribed to the different options , leading to biases that are incommensurate with the final , long term , state . Figure 7d shows an example . For the blue curve , the initial values of all states are low ( ) , but the probability of a reward is high ( ) ; for the red curve , the initial values are high ( ) , but the probability of a reward is low ( ) . In the former case , there is substantial initial over-observation; in the latter , initial under-observation .
We have provided an account of ‘observing behavior’ that shows how it can arise from a small Pavlovian bias over instrumental behavior associated with disengagement from a task , rather than any aspect of information seeking . Pavlovian biases are rife in decision-making; and accommodating them does not necessitate any further change to the standard underlying theory of the activity of dopaminergic neurons that has not already been suggested to accommodate other data . What we have done here is specify the shape of such an interaction based on disengagement in the task . We intended specifically to capture [10] experiment on macaques . However our results do touch upon other , but emphatically not all , instances of observing in the literature . Experiments such as [10] into observing are designed to maximize the effects of what is a relatively small anomaly in decision making ( compared , for instance , with the more extreme misbehavior evident in negative automaintenance [2] or the schedule task [40] ) . Indeed , in this case , the subjects did not have to pay a penalty for observing . Thus , under standard decision-making conditions , we may expect the net effect of disengagement to be modest , leaving near-optimal behavior within the scope of the model . Dinsmoor [9] suggested an account of the phenomenon based on his observation of ‘selective observing’ , i . e . , that the subjects would preferentially focus on stimuli associated with higher probabilities of reward . This idea met some resistance ( some of which is contained in the commentary to [9] ) , partly based on experimental tests in which the subjects were not able to avoid the low value predictive cues . Our account can be seen as a form of selective observing , but involving internal actions associated with the allocation of engagement and attention , rather than external actions involving preferential looking . It might seem that these accounts are close to Mackintosh's [41] suggestion that attention is preferentially paid to stimuli that are strong predictors of affectively important outcomes . However , in Mackintosh's account , attention particularly influences the speed of learning ( the associability of the stimulus ) rather than the fact of it ( at least in the absence of competing predictors ) , and so would not have the asymptotic effect that is apparent in the experiments we have discussed . Another interesting account of observing is Daly and Daly's DMOD [42] , which learns predictions associated with frustration ( when reward is expected , but does not arrive ) , and courage ( when reward is actually delivered during a state of frustration ) . These extra predictions warp the net expected values associated with the different cases in observing , favoring observing responses . The theory underlying DMOD is the original Rescorla-Wagner [43] version of the delta rule [44] , whose substantial modification by Sutton and Barto [45] to account for secondary conditioning led to the original prediction error treatment of the activity of dopamine neurons in appetitive conditioning [19] . It would be necessary to extend DMOD in a similar way , and to make an assumption about which of its three prediction errors ( or other quantities ) are reflected in the activity of dopamine neurons , in order to determine its match to the neurophysiological data . The failure of TD models to capture behavioral aspects of frustration is , however , notable . To some tastes , the most theoretically appealing accounts of observing start from the notion that animals seek to acquire information about the world [46] . However , formal informational theories have difficulty with the results of reducing the probability of reward ( Figure 7c; [12] ) , which reduce the uncertainty and the information gained , but increase observing . More informal theories , such as that suggested by [10] require more precise specification to be tested against accounts such as the one here . The sloth of initial learning and reversal apparent in Figure 6 ( taking 1200–2400 choice trials , 3000–7000 trials overall ) might be considered suggestive evidence against an informational account , since it implies at the very least a nugatory value for the information . In terms of our account , there are various routes by which predicted values could influence persistent engagement . Failure to engage can be seen as the same sort of malign Pavlovian influence over behavior that is implicated in the poor performance of monkeys in tasks in which they know themselves to be several steps away from reward [40] , [47] . In that paradigm , it is an explicitly informative cue that the reward is disappointingly far away that leads to disengagement; this parallels the disappointment associated with the non-discriminative cue in observing . The most obvious mechanism associated with engagement is the influence of dopamine itself over working memory [31]; however , whether this is the phasic dopamine signal associated with prediction errors for reward [19] or a more tonic dopamine signal associated with a longer term average reward rate [48] , [49] is not clear . Alternatively , some theories suggest that working memory is controlled by a gating process [29] , [30] associated with the basal ganglia , treating internally- and externally directed action in a uniform manner . Dopamine certainly influences the vigor associated with external actions [48]–[50]; it is therefore reasonable to assume that it might also influence internal engagement . We specialized our description of the model to the particulars of the experiment conducted by Bromberg-Martin and Hikosaka [10] . The most important question for other cases concerns the conditions under which re-engagement occurs . Since disengagement is seemingly rather rare , it is hard to get many hints from this experiment , and we might assume that it is reward delivery itself that causes re-engagement . However in a more general setup ( e . g . without reward delivery at fixed time points ) , a mechanism for re-engagement is necessary . One possible way to do that would be by stochastically re-engaging based on either the reward prediction error or expected value . Such a mechanism of re-engagement could happen at any time point but would be extremely likely to happen at the delivery of reward , as well as for the initiation of a new trial . To be fully generalizable we also need to specify the case for disengagement at the time of an action selection . While in a disengaged state we envision the animal not performing an explicit choice , thus potentially not responding within an allocated time . If a choice is required to progress in the behavioral setup it would happen after an eventual re-engagement . The model raises some further questions . First , we assumed that the probability of disengagement is a function of the actual prediction . However , it is possible that this function scales with the overall magnitude or scale of possible rewards , making the degree of observing relative rather than absolute . There is a report that phasic dopamine itself scales in an adaptive manner [39] , , and this would be a natural substrate . A second issue is whether disengagement is occasioned by the change in predictions associated with the phasic dopamine activity , or the level of the prediction itself . If the former , then in tasks such as the one studied by Bromberg-Martin and Hikosaka [10] , where substantial prediction errors only happen with phasic targets and cues , the state could , for instance , just be poorly established in working memory at the outset , because of a weak dopamine signal , and this could lead to a subsequent chance of disengagement . We adopted the simpler scheme in which it is the ongoing predictive value that controls the chance of disengagement . One experiment that hints in the direction of change is that of Spetch et al . [52] ( for a more recent study see [53] ) . In this , pigeons were given the choice between a certain ( 100% ) or uncertain ( 50% , but observed ) reward . Surprisingly , the level of engagement to the latter ( measured by the number of pecks to the illuminated key ) was many times to that of the former , and the pigeons duly made the suboptimal choice . The model presented in this paper does tie engagement to choice in a similar way , but we would be unable to explain such a strong effect . A variant of the model for which engagement is governed by prediction errors rather than predictions would show some contrast effect that could favor the uncertain , but observed , reward . However , it would be hard to explain such a stark contrast . A third issue is whether disengagement is complete ( and stochastic ) , or partial ( and , at least possibly , deterministic ) . We considered the former case , and indeed , this leads to a straightforward prediction that the histogram of the dopamine response at the time of a delivered reward in the non-discriminative case might have two peaks; one associated with continuing engagement to the point of reward; the other , which would be roughly twice as high , associated with prior disengagement . However , it is also possible that less dramatic changes in engagement occur during the interval between cues and reward . If many individual neural elements are involved in the engagement ( for instance in working memory circuits devoted to timing ) , then some could disengage before others . This might even lead to a non-uniform behavior among different dopamine cells . Unfortunately , the low firing rates of these cells make it hard to discriminate between these various possibilities . Finally , the question arises as to the computational rationale for value-dependent disengagement . Other instances of Pavlovian misbehavior , such as withdrawal from cues associated with predictions of low values , can find plausible justifications in terms of evolutionary optimality . Disengagement might be seen in the same way , as a Pavlovian spur to exploration [54] in the face of poor expected returns . From the perspective of conditioned reinforcement , our account suggests that the issue that is often studied is not really the one that is critical . Various investigators ( see , for instance , the ample discussion in Lieberman et al . 1997 [55] about the differences between their findings and those of Fantino and Case 1983 [56] ) have considered whether stimuli like S are conditioned reinforcers because of their association with the reward . For us , S and S and S are all conditioned reinforcers . The key question for observing behavior is instead an apparent concavity: the average worth of two different stimuli associated deterministically with small and large rewards is greater than the worth of a single stimulus associated stochastically with the same outcome statistics ( see [57] ) . It is this non-linearity that demands explanation , and not merely the fact , for instance , of savoring or anticipation of the future reward , which could quite reasonably also be purely linear . Some accounts put the weight of the non-linearity onto the stimulus associated surely with the large reward . By comparison , our account places this emphasis onto the non-discriminative stimuli , suggesting that they are more likely to lead to disengagement . The same is true of other sources of non-linearity , for instance a mechanism that accumulates distress from the prolonged variance/uncertainty in the non-discriminative pathway . Various versions of the ‘observing task’ have also been tested on humans [55] , [56] , [58] . These studies have shown consistent observing behavior , but , partly because of the different reading of the issue of conditioned reinforcement to the one discussed above , have often focused on different questions and methods from those in Bromberg-Martin and Hikosaka [10] . For instance , one question has been whether subjects would observe if they only ever found out S and never S – the idea being that conditioned reinforcement could support observing of the latter but not the former . Unfortunately , the answers have been confusing [55] , perhaps partly because of issues about how cognitive effects ( e . g . , expectations of controllability ) influence the results . Note , in particular , that we have only modeled observing behavior associated with repeated experience and learning , and not the sort of single-instance decisions that are often used in human cases . In conclusion we have shown that the often observed effect of ‘observing’ , preferring a behaviorally irrelevant discriminating stimulus cue , can readily be explained by a bias caused by Pavlovian misbehavior , putting it in the same category as a range of other suboptimalities . Informational accounts , however seductive , are not necessary .
We model value learning using a modified version of a standard temporal difference model [21] , [22] . We assume the task can be specified as a Markov process , where the participant estimates the expected long run future reward ( value ) of each state as , updating it according to ( 1 ) where is the learning rate , and is the change in expected value given by: ( 2 ) where is the delivered reward , and is the state that follows . Learning proceeds for all three sorts of trials ( forced disc . , forced non-disc . and choice trials ) . The modelled dopamine signal for Figures 4 and 5 is . The only deviation from the standard TD model is in assuming that the correct updating of this system is dependent on maintaining engagement , for instance in working memory . We assume the probability of disengagement of the course of state to be ( 3 ) per unit of time ( in seconds ) . Hence , for a given state the probability of a correct updating is given by , where is the amount of time spent in the state ( see Figure 1 ) . and are fixed parameters . We assume the consequence of disengagement to be the transition to a specific fixed ( non-updating ) state of value and hence the updating signal for is ( 4 ) The system stays in this state , until a reward is delivered at the end of the trial . At this point the system is ‘re-engaged’ creating a TD error relative to the fixed state ( see Figure 5 ) . We assume that any potential disengagement in the intertrial interval is negated by the initiation of a new trial . Choice is only possible at one state , between progressing to either state and state . Given the learned values , we assume the subject performs choice based on the Softmax or Luce choice rule [59] ( 5 ) Note that it is straightforward to see that this version of softmax is dependent on the difference in values ( ) , whereas using the logarithm of the value ( as in Figure 7c ) causes the function to be dependent on the ratio of values ( ) . In the limit without any failures in updating the learned values would approach the true value , where the expectation is taken over states . However with a chance of failure dependent on the value , the iterative solution in Figure 7c can be given by solving ( 6 ) numerically . For all figures we assumed and . For Figs . 2 and 6 we used parameters , and . For the aversive stimuli in Figure 7a–b we assumed negative reward values . For Figure 7a the parameters were . For Figure 7b the parameters were . For Figure 7d the parameters were . To mimic the fact that dopamine neurons have less dynamic range for increases than decreases in firing rate , for Figure 4 we truncated the negative responses at −25 percent of the maximal positive response of the neuron . | The theory of Reinforcement Learning ( RL ) has been influential in explaining basic learning and behavior in humans and other animals , and in accounting for key features of the activity of dopamine neurons . However , perhaps due to this very success , paradigms that challenge RL are at a premium . One case concerns so-called ‘observing behavior’ , in which , at least in some versions , animals elect to observe cues that are predictive of future rewarding outcomes , although the observations themselves have no direct behavioral relevance . In a recent experiment on observing , the activity of monkey dopaminergic neurons was also found to be incompatible with classic RL . However , as is often the case , this was a task that allowed for potential interactions from a secondary behavioral system in which responses are directly triggered by values . In this paper we show that a model incorporating a next order of refinement associated with such Pavlovian interactions can explain this type of observing behavior . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"neuroscience/behavioral",
"neuroscience",
"neuroscience/animal",
"cognition",
"neuroscience/theoretical",
"neuroscience",
"neuroscience/cognitive",
"neuroscience"
] | 2010 | Pavlovian-Instrumental Interaction in ‘Observing Behavior’ |
Response to antidepressant treatment in major depressive disorder ( MDD ) cannot be predicted currently , leading to uncertainty in medication selection , increasing costs , and prolonged suffering for many patients . Despite tremendous efforts in identifying response-associated genes in large genome-wide association studies , the results have been fairly modest , underlining the need to establish conceptually novel strategies . For the identification of transcriptome signatures that can distinguish between treatment responders and nonresponders , we herein submit a novel animal experimental approach focusing on extreme phenotypes . We utilized the large variance in response to antidepressant treatment occurring in DBA/2J mice , enabling sample stratification into subpopulations of good and poor treatment responders to delineate response-associated signature transcript profiles in peripheral blood samples . As a proof of concept , we translated our murine data to the transcriptome data of a clinically relevant human cohort . A cluster of 259 differentially regulated genes was identified when peripheral transcriptome profiles of good and poor treatment responders were compared in the murine model . Differences in expression profiles from baseline to week 12 of the human orthologues selected on the basis of the murine transcript signature allowed prediction of response status with an accuracy of 76% in the patient population . Finally , we show that glucocorticoid receptor ( GR ) -regulated genes are significantly enriched in this cluster of antidepressant-response genes . Our findings point to the involvement of GR sensitivity as a potential key mechanism shaping response to antidepressant treatment and support the hypothesis that antidepressants could stimulate resilience-promoting molecular mechanisms . Our data highlight the suitability of an appropriate animal experimental approach for the discovery of treatment response-associated pathways across species .
A “one size fits all” approach is not effective or efficient in the treatment of major depressive disorder ( MDD ) . Although it would be ideal to tailor available treatments to individual patients [1] , patient-level antidepressant treatment outcomes are still highly unpredictable [2] . Identification of biomarkers predictive of individual treatment response or molecular biosignatures associated with response would dramatically improve the quality of care for MDD [3] . These biomarkers could also be expected to significantly reduce both treatment and loss-of-productivity costs . The latter become increasingly important because MDD has been shown to be the second leading cause of disability worldwide [4] . Finally , biomarkers could allow patient stratification and enable the selection of pathophysiologically distinct patient subgroups to allow optimized treatment choices based on biology . Such biomarkers could also inform the development of new interventions specifically targeting disease mechanisms in these subgroups . Conceivably , useful biomarkers for treatment response in depression could be developed through blood-based biomarkers , including genetic approaches , although psychophysiological and neuroimaging approaches are also promising [5] . However , despite considerable efforts , including large-scale hypothesis-free , genome-wide approaches during the past years [6 , 7] , no biological or genetic predictors of sufficient clinical utility have been identified for routine clinical use . Thus , the most effective treatment for each patient is currently identified through a trial and error process [2] . Among the potential barriers to the development of clinically useful biomarkers in depression , the following 3 have been identified as being most important . First , current symptom-based diagnoses likely group pathophysiologically distinct patients [8] , leading to considerable heterogeneity among patients diagnosed with MDD [9 , 10] . Second , there are a number of confounding environmental factors such as childhood maltreatment , previous life events , disease episodes , and different psychopharmacological treatment schedules that often remain unidentified and potentially reduce the power to detect true response biomarkers . Third , genetic background , age , and sex are all factors that significantly impact transcription profiles and other laboratory measurements , as well as treatment outcome [11] . In addition to the aforementioned problems , major psychiatric disorders , including MDD , are primarily viewed as brain disorders , so the question of whether peripheral measures can be informative for treatment response to centrally acting compounds such as antidepressants continues to be matter of debate [12] . During recent years , evidence has emerged that disease- and treatment-related changes may be reflected outside the central nervous system [13 , 14] , revealing a potential role for appropriate animal models to support biomarker discovery in MDD . To the best of our knowledge , neither an appropriate animal experimental approach nor a translational approach systematically addressing the potential of biosignatures predicting or tracking antidepressant treatment response has been published . To overcome some of the limitations of past approaches , we here present a conceptually novel approach that allows the selection of extreme phenotypes in an antidepressant-responsive mouse strain ( DBA/2J [15] ) and uses these extreme groups to identify peripheral blood biomarkers associated with behavioral treatment response , which are then tested in a human patient cohort . This strategy exploits the advantages of a murine approach for the purpose of biomarker discovery , i . e . , ( 1 ) to investigate a highly homogeneous group of animals in which differences in genetic background , age , and sex can be excluded , ( 2 ) to perform biomarker discovery under conditions in which interindividual confounding environmental influences , including drug plasma and brain levels , are reduced to a minimum and controlled for , and ( 3 ) to allow correlations of peripheral biomarkers with behavior but also with peripheral and central drug concentrations , and to test the overlap of blood and brain expression profiles . We hypothesize that these standardized conditions will facilitate the identification of valid peripheral biomarkers for antidepressant treatment response and allow translation to humans .
Experiments were carried out with male DBA/2J mice ( n = 140 ) from Charles River , France . On the day of arrival , the animals were 6–8 weeks old and from that day on were singly housed in standard cages under a 12L:12D cycle ( lights on at 0800 h ) and constant temperature ( 23 ± 2°C ) conditions . Food and water were provided ad libitum . Pharmacological treatment of all animals started at an age of 9–11 weeks . Behavioral testing was performed at an age of 11–13 weeks . The experiments were carried out in the animal facility of the Max Planck Institute of Psychiatry in Munich , Germany , and approved by the committee for the Care and Use of Laboratory Animals of the Government of Upper Bavaria , Germany . All experiments were carried out in accordance with the European Communities Council Directive 86/609/EEC . The sequential steps and experimental procedures are summarized in Fig 2 , indicating the number of animals for each experimental group . A large number of animals were treated twice a day with either paroxetine ( n = 90 ) , a commonly used selective serotonin reuptake inhibitor ( SSRI ) antidepressant or a vehicle ( n = 50 ) . On treatment day 15 , the animals received their last drug administration at 6 AM and were subjected to a FST 4 hours later . Directly after the FST , the animals were anesthetized with isoflurane and decapitated . Animals were anesthetized with isoflurane and killed immediately following the FST . Trunk blood was collected individually in 1 . 5-mL tubes . Brains were rapidly dissected and frozen at −80°C . Due to the complex character of the study , limitations in available specimens , stringent quality control ( QC ) , and exclusion of outlier data , we could not always achieve fully identical sample and group compositions throughout all data analysis levels . This also explains the sporadic appearance of nonconcordant group sizes , which we consider a minor but unavoidable drawback . Brain and plasma paroxetine concentrations were measured after extraction by high liquid chromatography and quantifications . Paroxetine plasma concentrations were considered as a covariate in the analysis of the microarray data . For details of the respective protocols , see [18] . For determination of brain tissue concentrations of paroxetine , tissue from the cerebellum was dissected and rapidly frozen on dry ice . The remaining trunk blood of each animal was collected in labeled 1 . 5-mL EDTA-coated microcentrifuge tubes ( Kabe Labortechnik , Nümbrecht , Germany ) . All blood samples were kept on ice until centrifugation at 8 , 000 rpm at 4°C for 15 min . After centrifugation , the blood plasma was transferred to new , labeled 1 . 5-mL microcentrifuge tubes . All plasma samples were stored frozen at −20°C until the determination of corticosterone by radioimmunoassay ( MP Biomedicals , Santa Ana , CA; sensitivity , 6 . 25 ng/mL ) . The data presented are shown as means + standard error of the mean , analyzed by the commercially available software SPSS 16 . 0 . For comparing 2 independent groups , data were analyzed with 2-tailed , independent samples Student t test in case of normal distribution of the data; otherwise , nonparametric comparisons were applied ( Mann–Whitney U test ) . For variables with more than 2 groups , 1-way ANOVA was performed followed by Bonferroni post hoc testing . Correlations were analyzed with a 2-tailed , bivariate Pearson’s correlation analysis . As nominal level of significance , p < 0 . 05 was accepted . Values outside the 95% confidence interval ( CI ) were defined as statistical outliers and excluded from the analyses . Part of the blood was processed according to the PAXgene blood miRNA Kit manufacturer’s instructions . Briefly , 350 μL of freshly collected trunk blood was immediately transferred into 1 . 5-mL tubes filled with 966 μL PAXgene solution ( RNA stabilizer reagent ) , gently inverted 10 times , incubated at room temperature ( RT ) for 2–24 hours , and then stored at −20°C before ribonucleic acid ( RNA ) isolation . Volume ratio of RNA stabilizer reagent to blood samples was kept at 2 . 76 , according to the manufacturer’s protocol . We assessed whether the observed gene expression profiles of good treatment responders and poor treatment responders were related to changes in blood cell proportions in the mice using CIBERSORT [19] . The input reference matrix of expression signature profiles of mouse tissue was obtained using ImmuCC [20] . These statistical tools infer proportions of 25 types of immune blood cell types . To assess the relevance of the gene expression transcripts for antidepressant response differences in humans , we tested their predictive ability to classify response status in a human sample . The sample ( n = 86 ) consisted of a subset of MDD patients treated with antidepressant drug treatment over 12 weeks from 2 samples recruited at Emory University School of Medicine ( N = 74 from [21] and N = 12 from [22] ) . In both studies , patients followed a similar protocol and were randomized to either antidepressant drug treatment or cognitive behavior therapy ( CBT ) , with the difference that patients were randomized to CBT , duloxetine , or escitalopram in PReDiCT [21] and to CBT or escitalopram in [22] . Only the subset of patients in the antidepressant treatment group with sufficient RNA quality at both time points was included in this study . Please see S2 Table for a brief synopsis of demographic and clinical parameters on the patients from clinical studies . Depression severity was assessed at baseline and week 12 using the Hamilton Depression Rating Scale ( 17 items , HDRS-17 ) . In both samples , blood was drawn at baseline and after 12 weeks of treatment into Tempus RNA tubes ( Applied Biosystems ) . RNA was isolated from peripheral blood in a 96-well format using the magnetic bead-based technology MagMAX for Stabilized Blood Tubes RNA Isolation Kit , compatible with Tempu Blood RNA Tubes ( Ambion/Life Technologies , Carlsbad , CA; cat# 4451893 ) . RNA was quantified using the Nanophotometer , and quality checks were performed on the Agilent Bioanalyzer ( Agilent Technologies , Santa Clara , CA ) . Only samples with RIN ≥ 6 with clear 18S and 28S peaks on the Bioanalyzer were used for amplification; the average RIN was 6 . 3 ( SD of 0 . 668 ) . RNA was further processed for generation of biotin-labeled amplified RNA using the Amplification Kit ( Ambion/Life Technologies , Carlsbad , CA; cat# 4393543 ) . cRNA was hybridized to Illumina Sentrix Arrays HT-12 v4 . 0 arrays using the Illumina TotalPrep-96 RNA ( Life technologies , Carlsbad , CA ) and incubated overnight for 16 hours at 55°C . Arrays were washed , stained with Cy3 labeled streptavidin , dried , and scanned on the Illumina BeadScan confocal laser scanner ( Illumina , San Diego , CA ) . QC was performed using the bead-array package in R for 86 samples and 47 , 282 probes . Probes with p-detection values of <0 . 01 in at least 10% of the samples in the whole data set were removed . Remaining probes were normalized and transformed using the vsn package in R . Not all samples were hybridized on the same batch , and thus we corrected for chip number using COMBAT . A total of 17 , 725 transcripts and 86 samples remained after QC . For the full drug-treated sample , 63 patients were classified as responders and 23 as nonresponders , according to percent changes in HDRS-17 scores from baseline to week 12 ( ≥50% or <50% change , respectively ) . Mouse gene expression transcripts ( n = 259 ) resulting from the microarray analysis and described in S1 Table were mapped to their human orthologue genes present in the Illumina HT-12 arrays ( n = 241 ) . Because some genes are represented by more than one probe , 288 probes were included in final analyses . Prediction models were built as soft margin support vector machines for classification using the e1071 packages in R with the parametrization gamma = 0 . 001; cost = 10 . Further analyses included only mouse transcripts at FDR of 5% ( n = 85 ) . These were also mapped to their human orthologue genes ( n = 77 ) ; 66 genes passed QC in the human study , which were represented by 92 probes . The sample was equally divided into training and test data sets for each of the analyses ( probes at q < 0 . 1 and q < 0 . 05 ) . Gene expression repeated measures from the patients at baseline and week 12 were available; we computed the absolute difference between the expression levels of the transcripts between those time points and tested whether these differences were able to predict response to antidepressant treatment in the test data set . We permuted the response-status labels 10 , 000 times in the training data set and predicted the response status in our test data . In addition , we compared the obtained prediction accuracy of our selected classification features against 1 , 000 classification models derived from randomly sampled features . Random feature sets also consisted of absolute difference in expression between baseline and week 12 of treatment and were size matched to the selected feature set . Those data were the input for soft margin support vector machine training and testing as indicated above . We assessed whether the observed gene expression changes in responders versus nonresponders were related to changes in cell proportions in the human samples using the Cell-type Computational Differential Estimation CellCODE R package [23] . Separate components for neutrophils , T cells , stimulated T cells , NK cells , dendrite cells , stimulated dendrite cells , monocytes , B cells , and plasma cells were extracted using markers from the IRIS reference data set provided by CellCODE . Two available tools have been used for pathway analyses: DAVID ( https://david . ncifcrf . gov/ ) and Pathway-Express [24] . Both tools were used with a list of gene symbols previously shown to be significantly regulated ( q-value < 0 . 1 ) with differential paroxetine response and interrogated with respect to a custom background that contained all microarray probes that have been used for computing inferential statistics . The background contained probes that passed our detection and variance filters . To determine the function overlap of differential paroxetine response with dex-regulated genes , we used data from a microarray experiment in male C57BL/6N mice at an age of 12 weeks ( mean body weight 26 . 8 ± 0 . 1 g ) , in which animals were treated with 0 . 1 mg/kg dexamethasone i . p . or vehicle ( N = 10 and 10 ) between 0900 and 1100 and sacrificed 4 hours later [25] . Trunk blood was collected into microcentrifuge tubes containing PAXGene RNA stabilizer solution and frozen at −20°C . RNA was then extracted using the PAXgene blood miRNA kit ( PreAnalytiX ) , amplified using the Illumina Total Prep 96-Amplification kit ( Life Technology ) , and then hybridized on Illumina MouseRef-8 v2 . 0 BeadChips . Analyses were performed using custom scripts in R . First , a common content for both microarray data sets was generated based on Illumina “Probe Ids . ” Within that common content , differentially expressed microarray probes were identified for both contrasts using an FDR threshold of q < 0 . 1 . For the differential paroxetine response , 179 probes passed that threshold . Then , the number of array probes overlapping with dex regulation by chance was determined using 100 , 000 random sampled gene sets of size N = 179 . For each trial , the overlap to the fixed dex-regulated gene list ( N = 1 , 882 ) was determined and all the results were finally compared to the overlap of paroxetine response genes with dex-regulated genes; this was done by counting the number of sampled sets that showed higher overlap ( >134 ) than the differential gene list . In addition , a 2 × 2 contingency table was computed for dex regulation and paroxetine response and these numbers were further used to perform a hypergeometric test . Calculation of statistical significance for a possible directionality of gene regulation was performed using a binomial test .
In order to detect the minimum effective dosage of paroxetine for the DBA/2J strain , 2 paroxetine concentrations ( 1 mg/kg body weight or 5 mg/kg body weight , twice daily ) were tested in a pilot study . The lower paroxetine concentration ( n = 29 ) failed to produce a significant behavioral treatment effect in the FST . The only parameter that was altered with the 1 mg/kg dosage was body weight ( T39 = −2 . 490 , p < 0 . 05 ) . Behavioral data , neuroendocrine measurements , and body weight are shown in S1 Fig . A dosage of 5 mg/kg evoked a significant antidepressant-like response in the FST ( Fig 3 ) . The following data were all collected from animals treated with 5 mg/kg paroxetine , which we considered to be the minimum effective dosage for the DBA/2J strain . There was no significant difference in plasma paroxetine concentrations between the good and poor treatment responder ( p = 0 . 19 ) . For paroxetine brain concentrations , a significant difference between good and poor responders could be detected ( p < 0 . 05 ) ( S3 Fig ) . Paroxetine brain and plasma concentrations were closely correlated ( r = 0 . 94; p < 0 . 0001 ) ( S3 Fig ) . Despite the lack of statistical association , we included plasma paroxetine concentrations as a covariate in further analyses on the transcriptome profiles in peripheral blood samples . Brain paroxetine concentrations were used as covariates in analyses of PFC samples . To identify signature gene expression profiles characteristic of the animals’ responder status , gene expression data sets of vehicle-treated animals , good responders , and poor responders were created by whole-genome gene expression microarray analysis on blood samples and analyzed ( n[vehicle] = 12 , n[good] = 12 , n[poor] = 12 ) . We evaluated both treatment effect and response status with respect to antidepressant treatment and with respect to paroxetine plasma concentrations . We also investigated whether paroxetine brain or plasma concentrations might have an effect on gene expression levels . Linear and quadratic regression analyses did not reveal any microarray probe that showed significant correlations with the related plasma paroxetine levels when controlling for multiple testing . No significant influence of paroxetine concentrations on gene expression profiles was observed . Nevertheless , identified technical batch effects in the data and measured paroxetine drug concentrations in blood were used as covariates in an ANOVA-based statistical model . Although no robust gene regulation was apparent when the treatment group ( independent of response ) was compared to the control group , there was a pronounced effect within the treatment group . We were able to detect a set of 259 transcripts that showed a significant difference in expression due to antidepressant response status at a false discovery controlled significance level of 10% ( q < 0 . 1 ) ( Fig 4; S1 Table ) , of which 85 had q < 0 . 05 ( S1 Table ) . We then aimed to see whether the observed gene regulation patterns in peripheral blood might overlap with effects observed in the PFC from the same animals . To test this , we first performed a cluster analysis on the difference in expression between the responder groups in the set of differentially regulated genes in blood . We then compared the results for these transcripts to the difference in expression between these 2 groups measured in PFC brain tissue in the same animals . The results are summarized in a heat map in Fig 4 and indicate that , within the selected gene set , there is no major common gene regulation pattern associated with response status between both tissues . No significant differences in immune cell subtypes between the different response groups were detected using CIBERSORT [19] and ImmuCC [20] ( see S3 Table ) . No significant change in immune cell subtypes using CellCODE [23] was associated with the response groups in the human sample ( see S4 Table ) . Therefore , none of the estimated cell proportions were included in further analyses . In the next step , we determined whether this transcriptional profile identified in the mouse model would also be relevant in the human data set . Therefore , we tested whether changes in the mRNA expression of the human orthologues of transcripts at FDR of 10% and at FDR of 5% , separately , are associated with response to antidepressant treatment . Differences in expression profiles from baseline to week 12 when using human orthologues of transcripts at FDR of 10% allowed prediction of response status ( at least 50% improvement in HDRS-17 from baseline to week 12 for responders ) with an accuracy of 76% , using all patients treated with antidepressant . The prediction persisted after we permuted the response-status labels 10 , 000 times ( pperm = 0 . 0328 ) . When a more stringent FDR of 5% cutoff was applied to the mouse transcripts , the corresponding human orthologues predicted response status with an accuracy of 81% in the human sample . The prediction persisted after we permuted the response-status labels 10 , 000 times ( pperm = 0 . 0018 ) . After showing that expression levels of the antidepressant response genes identified from mice are also informative for classification in a human sample , we further analysed the quality of the mouse-based feature selection in the human data set . For this , we compared the classification accuracy of our identified antidepressant-response features to classification accuracy given by randomly chosen and size-matched sets of gene expression probes in the human sample ( Fig 5 ) . In analogy to the previous classification approach , we used differences in gene expression from baseline to week 12 in 1 , 000 random sets of gene probes . Only 25 random gene probe sets showed higher or equal prediction accuracy than our feature panel selected from the animal model . This suggests that the information derived from the mouse experiments allowed the selection of transcripts for which the classification accuracy is better than for random gene expression background ( pperm = 0 . 026 ) . For functional annotations of the microarray results , we performed pathway analyses and included an overrepresentation analysis with DAVID , and we conducted a second analysis using Pathway-Express . The latter accounts for pathway topology and biological effect size . In both approaches , no significant results passing our threshold criteria were found . The top overrepresented categories in DAVID were entities associated with general gene transcription and did not reach significance levels . Although Pathway-Express showed formally significant results for a few specific Kyoto Encyclopedia of Genes and Genomes ( KEGG ) pathways , we excluded them because less than 2% of the pathway genes were regulated . We next integrated our results with another microarray data set that we had previously generated . Those data originated from mouse blood samples taken from animals that had been treated with the glucocorticoid receptor ( GR ) agonist dexamethasone ( dex [25] ) . To test whether GR activation responsive genes are overrepresented in our antidepressant response gene set , we used a permutation approach and computed the overlap of dex-regulated genes with the paroxetine response genes and compared it to matched random gene sets sampled from the paroxetine array results ( Fig 6 ) . Based on 2 , 852 array probes that constituted a common content for both independent data sets , 179 array probes of the 259 response associated probes could be used for this analysis . The overlap between the probes significantly regulated between the responder group and the ones regulated following dex administration was 134 out of 179 . Within 100 , 000 trials of drawing random gene sets of 179 probes , there were only 70 instances in which a higher overlap occurred . This reflects a permutation-based FDR of 7e-4 for enrichment of dex-regulated array probes in paroxetine response probes . A hypergeometric test yielded a p-value of 5 . 6e-4 , further supporting an enrichment of GR-responsive transcripts among response-associated genes . Standard enrichment analysis does not take into account the direction of gene regulation , and we were interested to see whether paroxetine response and dex regulation showed a directional overlap . Of the 134 array probes that are significantly regulated by dex and are , at the same time , between the paroxetine response groups , only 38 had a mismatch in the direction of the putative regulation . The majority of the regulated genes ( N = 96 ) are regulated in the same direction in both conditions , and based on a binomial distribution , such a result could not be observed if both outcomes ( same and opposite regulation ) had the same probability ( p = 7 . 2e-07 ) . Thus , we can conclude that there is a common direction of gene regulation for dex treatment and paroxetine response .
There are 2 obvious gaps of knowledge in depression treatment , namely ( 1 ) the lack of biosignatures predicting antidepressant response and ( 2 ) the lack of knowledge of the molecular mechanisms mediating the response to antidepressant pharmacotherapy . The latter is of particular importance for the eagerly awaited discovery of conceptually novel antidepressant treatment strategies , which can only be rationally realized with a deeper understanding of the molecular mechanisms underlying clinical response [26] . In recent years , the unbiased , i . e . , genome-wide , screening to identify genetic factors that could assist in the prediction of an individual’s drug response has been a major focus in depression research . Despite tremendous efforts , however , the results are fairly modest in identifying predictive genes in large genome-wide association studies [27–29] and even in a meta-analysis [6] . Instead , Tansey et al . [30] recently presented data implicating a highly polygenic architecture involving many common variants scattered across the genome , none of which have very large effects but cumulatively contribute to a substantial proportion of variation in antidepressant response . So far , only a few small studies provided first evidence that biochemical information ( e . g . , metabolomics ) could add to the panel of markers predicting response to a particular antidepressant in patients [31] , suggesting that alternative strategies need to be explored . However , studies to investigate the neurobiology of antidepressant treatment response have been hampered by the fact that no appropriate animal model addressing this issue had yet been described . Therefore , we embarked upon the development of an animal experimental approach modeling the heterogeneity in response to antidepressant treatment as closely as possible . In contrast to studies in patients , this model approach both enables an in-depth analysis of the neurobiological mechanisms shaping individual antidepressant response in the central nervous system and searches for peripheral biosignatures associated with treatment response . There are different approaches to model depression-like phenotypes ( i . e . , symptoms of depression ) in the mouse . While induction of depression-like symptoms following exposure to different types of stress , e . g . , chronic social defeat or chronic mild stress is one possible approach , the use of mouse strains with high innate anxiety- and depression-like behavior is also commonly accepted . The selection of the DBA/2J mouse strain , with its well-described high innate anxiety and responsiveness to antidepressant treatment [17] , enabled us to perform the pharmacological treatment under basal conditions , i . e . , without the need to subject the animals to an additional stress procedure that might have influenced the transcriptome data . A combination of stress exposure and antidepressant treatment within our approach would not allow us to identify the individual contribution of these 2 factors to the phenotype . Nonetheless , a comparison of stress-related and antidepressant response–related molecular events could enable the identification of shared molecular pathways . Oral treatment with the SSRI paroxetine significantly reduced—as expected—depression-like behavior . Remarkably , in addition to the overall antidepressant-like effect on promoting active coping strategies in the FTS , we detected a high variability in the behavioral outcome . Although the neurobiological mechanisms underlying antidepressant-induced behavioral changes in the FTS still are not fully understood [32] , we here used the FST as the laboratory animal equivalent of treatment response because it is the most commonly used test to screen for antidepressant efficacy in rodents [16] . Comparable approaches for stratification and extreme case sampling in animal models have been successfully introduced in the field of stress research [33] , and during recent years , they have enabled the identification of a number of key mechanisms shaping individual susceptibility to stress [34 , 35] . We considered plasma paroxetine concentration as a covariate on our microarray analyses , but we were not able identify a significant influence on the gene expression profile associated with treatment response . The selection of a rodent approach for biomarker discovery in psychiatric disorders has the advantage of minimizing potentially confounding variables , which , in clinical depression studies , so far have impeded biomarker discovery [12] . Due to the standardized experimental conditions , factors such as sex , age , and additional environmental factors , including pharmacological pretreatment , the time of day at which the blood sample is taken , physical exercise , food , and many others [36] , can be strictly controlled for , thus enabling the detection of true response biomarkers in a hypothesis-free approach . In a second step , these murine biomarkers can then be validated in the human population . Given the complexity of identifying true biomarker candidates in psychiatric disorders , the need to strengthen potential candidates by cross-species approaches [37] and to validate those in independent cohorts is considered crucial [38] . Aiming to enable a translational approach , we focused on the identification of transcriptome signatures in the periphery , because only those are relevant for clinical application . Several studies have investigated the use of human peripheral blood cells as surrogate material for different organs and tissues , including the central nervous system [39–41] . However , inconsistent results have been reported as to the overlap between transcriptome profiles in peripheral blood and brain [14] . To address issues of cross-tissue relevance , we compared peripheral transcriptome signatures with expression profiling data of the PFC of the same good- and poor-responding animals . We did not find any major common response status-associated gene regulation pattern between both tissues . We thus hypothesize that in depression treatment , blood cells might act as sentinels of treatment response but are not generally informative about central regulation processes , at least not in the PFC . In the next step and as a proof of concept , we sought to evaluate the relevance of the murine transcriptional signature associated with antidepressant treatment response in a human data set . Using a powerful within-participant approach investigating longitudinal transcription changes between baseline and week 12 of antidepressant treatment , we tested whether mRNA expression of the human orthologues of these transcripts changes with antidepressant treatment in peripheral blood in a subset of 2 human studies [21 , 22] . Differences in expression profiles from baseline to week 12 of the human orthologues selected on the basis of the murine transcript signature allowed prediction of response status ( percent change in HDRS-17 from baseline to week 12 ) with an accuracy of 76% in the human sample . Using a permutation strategy , we also showed that our set of transcripts was more likely to predict treatment outcome correctly than random sets of transcripts . We thus show the suitability of an appropriate animal experimental approach for the discovery of peripheral treatment response biomarkers . While promising , our findings certainly require validation in independent samples of patients with MDD . One aspect that needs more detailed investigation in future studies is the precise time course and stability of response-associated transcript changes , as we here integrated murine transcript data following 2 weeks of antidepressant treatment with patient data over a 12-week treatment course . The available evidence makes a compelling case implicating dysregulation of the stress hormone system , the so-called hypothalamus-pituitary-adrenocortical ( HPA ) system , in the pathogenesis of MDD [42 , 43] . Moreover , considerable evidence has accumulated suggesting that normalization of the HPA system might be the final step necessary for stable remission of the disease [44] , and it was further hypothesized that antidepressants may act through normalization of the HPA system function [45] . A recent study provided evidence that hormone-independent activation of the GR is involved in the therapeutic action of fluoxetine [46] , supporting the neurobiological link between GR signalling and antidepressant action . We could not detect any difference in corticosterone plasma concentrations between good and poor responders to paroxetine treatment directly after the FTS challenge , although assessment of plasma corticosterone concentrations at 1 time point , i . e . , 5 min after the FST , does not exclude potential dynamic changes in HPA system response ( i . e . , changes in the rise of corticosterone or HPA system feedback following initial activation ) . Evidence from measurements of HPA system activity in depressed patients , however , supports the notion that in vivo challenges such as the combined dexamethasone/corticotropin releasing hormone challenge test ( Dex-CRH test ) are superior to single baseline measurements of peripheral glucocorticoid concentrations in discriminating between depressed patients and healthy controls as well as treatment responders versus nonresponders . In addition , recent investigations have shown that dex-stimulated gene expression is a sensitive marker of GR-resistance in MDD [13] and that common genetic variants that modulate the initial transcriptional response to GR activation increase the risk for depression [25] . Therefore , we tested for an enrichment of GR-responsive genes in our antidepressant response gene set , a finding that could point to increased GR sensitivity in good- versus poor-responding animals . We demonstrated that ( 1 ) GR-regulated genes are significantly enriched in our cluster of antidepressant-response genes and ( 2 ) there is a common direction of gene regulation for dex treatment and paroxetine response . Our data are in line with a large body of previous evidence pointing to the normalization of GR resistance as an important feature of the clinical response to antidepressant treatment [43 , 47] and support the intriguing hypothesis that antidepressants could stimulate resilience-promoting molecular mechanisms [48] . Biomarkers or biosignatures , respectively , would not only allow monitoring of antidepressant treatment response in clinical practice but they also could assist in the evaluation of drug actions at an early stage in clinical trials of novel agents that are frequently marred by late attrition [49] . In particular , identifying biomarkers of response will be essential for assessing target engagement of novel mechanisms . We submit that our approach opens up the opportunity to generate a unique database for putative biosignatures predicting response to be assessed and validated in larger patients’ samples . In conclusion , we expect this translational approach to serve as a template for the discovery of improved and tailored treatment modalities for depression in the future . | Major depression is the second leading cause of disability worldwide . However , only one-third of patients with depression benefit from the first antidepressant compound they are prescribed . It is a fundamental problem that the outcomes of individual antidepressant treatments are still highly unpredictable . In clinical studies , discovery of biomarkers for antidepressant response is hampered by confounding factors such as the heterogeneity of the disease phenotype and additional environmental factors , e . g . , previous life events and different schedules of psychopharmacological treatment , which reduce the power to detect true response biomarkers . To overcome some of these limitations , we have established a conceptually novel approach that allows the selection of extreme phenotypes in an antidepressant-responsive mouse strain . In the first step , we identify signatures in the transcriptome of peripheral blood associated with responses following stratification into good and poor treatment responders . As proof of concept , we translate the murine data to a population of depressed patients . We show that differences in expression profiles from baseline to week 12 of the human orthologues predict response status in patients . We finally provide evidence that sensitivity of the glucocorticoid receptor could be a potential key mechanism shaping response to antidepressant treatment . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"antidepressants",
"gene",
"regulation",
"drugs",
"biomarkers",
"bioassays",
"and",
"physiological",
"analysis",
"pharmacology",
"mood",
"disorders",
"research",
"and",
"analysis",
"methods",
"gene",
"expression... | 2017 | Common genes associated with antidepressant response in mouse and man identify key role of glucocorticoid receptor sensitivity |
The ability to generalize over naturally occurring variation in cues indicating food or predation risk is highly useful for efficient decision-making in many animals . Honeybees have remarkable visual cognitive abilities , allowing them to classify visual patterns by common features despite having a relatively miniature brain . Here we ask the question whether generalization requires complex visual recognition or whether it can also be achieved with relatively simple neuronal mechanisms . We produced several simple models inspired by the known anatomical structures and neuronal responses within the bee brain and subsequently compared their ability to generalize achromatic patterns to the observed behavioural performance of honeybees on these cues . Neural networks with just eight large-field orientation-sensitive input neurons from the optic ganglia and a single layer of simple neuronal connectivity within the mushroom bodies ( learning centres ) show performances remarkably similar to a large proportion of the empirical results without requiring any form of learning , or fine-tuning of neuronal parameters to replicate these results . Indeed , a model simply combining sensory input from both eyes onto single mushroom body neurons returned correct discriminations even with partial occlusion of the patterns and an impressive invariance to the location of the test patterns on the eyes . This model also replicated surprising failures of bees to discriminate certain seemingly highly different patterns , providing novel and useful insights into the inner workings facilitating and limiting the utilisation of visual cues in honeybees . Our results reveal that reliable generalization of visual information can be achieved through simple neuronal circuitry that is biologically plausible and can easily be accommodated in a tiny insect brain .
Honeybees ( Apis mellifera ) display an impressive visual behavioural repertoire as well as astounding learning capabilities . Foragers rely on visual and olfactory cues identifying rewarding flowers . Being able to recognise informative cues displayed by flowers can be assumed to facilitate fast and efficient decision-making . Indeed , honeybees can be trained to discriminate by an impressive range of visual cues; symmetry [1–3] , arrangements of edges [4–6] , size [7 , 8] , pattern disruption [9] and edge orientation [10–12] . These abilities are all the more impressive since trained bees are able to apply these same learnt cues to patterns which may have little or no resemblance to the original training patterns , so long as they fall into the same class of e . g . plane of symmetry , or edge orientation . This rich visual behaviour despite a relatively tiny brain makes honeybees an ideal model species to explore how visual stimuli are processed and to determine if generalization requires a complex neuronal architecture . Using the published intracellular recordings of large-field optic ganglia neurons to achromatic stimuli [13 , 14] and the known anatomical morphologies of mushroom body ( learning centres ) class II ‘clawed’ Kenyon cells [15] we designed two simple , but biologically inspired models . These models were not created , or indeed in any way ‘tweaked’ to replicate performance at any particular visual task . Instead they attempt to explore how well , or poorly , the known neuronal types within the bee brain could solve real behaviourally relevant problems and how much neuronal complexity would be required to do so . The initial models presented here were therefore kept very basic with limited neuronal pathways and very simple synaptic connections from the optic lobes to the mushroom bodies . In addition , to comprehend how these optic lobe neuron responses alone may explain the bees’ discrimination abilities and behavioural performance , we did not employ any form of learning in these models . Since two of the optic ganglia ( medulla and lobula ) of bees extend a variety of axonal fibres to both the ipsilateral and the contralateral mushroom bodies and , as opposed to axons from different regions of the optic lobes that are distinctly layered within the mushroom bodes , there is no apparent segregation of the visual inputs from the individual corresponding left and right eye regions [16 , 17] , we tested the discrimination and generalization performance difference between retaining independent inputs from each eye and combining the neuronal input from both eyes within our simulated mushroom body models . These models allowed us to simulate achromatic pattern experiments and compare the simulation performances of our two different bee-brain models—henceforth called ‘simulated bees’ , to the performance of actual honeybees in these same specific experiments . We drew on twenty-four experiments from three published honeybee behaviour papers [18–20] providing results on both the discrimination abilities of free flying bees perceiving complex bar and spiral patterns from a distance , and their generalization abilities while fixating , slow hovering scans 1–5cm in front of presented patterns . The surprising ability of one of our extremely simple simulated bees to discriminate patterns correctly even with the partial occlusion of the test stimuli , its invariance to the location of the visual cues on the eyes , and generalization performances almost identical to real bees , provides new insights into the relationship between behaviour complexity and its neural circuitry underpinnings , significantly contributing to our understanding of the fundamental requirements needed for specific cognitive abilities .
The ability to discriminate between visual patterns is essential for honeybees allowing them to identify familiar flowers and landmarks while navigating on foraging trips and locating the correct hive entrance upon their return . Nonetheless even for these types of precisely defined visual stimuli , some form of location invariance of a stimulus on the retinae would undoubtedly be required , as it is unlikely bees would perfectly align the stimulus against their eyes on every single flight in order to make a discrimination decision . Indeed it would be an undesirable necessity that they should have to do so . To test our two models ( DISTINCT , MERGED ) for the effect of location of the stimuli within the visual field , we simulated the experiments of Zhang and Horridge [18] who explored the ability of freely flying honeybees to discriminate two large ( 24cm diameter ) vertically displayed patterns composed of multiple oriented bars . For these experiments , a bee’s pattern choice was recorded when it approached within 27cm of either pattern ( see [18] for apparatus description ) . Presuming that honeybees would learn the correct pattern features when feeding at , or being close to , the centre of a rewarding pattern , we first calculated our Kenyon cell responses to these same rewarding patterns . We next determined each of our simulated bees’ performance accuracies when any of the two given test stimuli patterns ( correct pattern was identical to the rewarding pattern , the incorrect pattern was a rotated or mirrored version of this rewarding pattern ) were offset horizontally between -200 pixels and +200 pixels in 25 pixel increments . A zero pixel offset would align the pattern perfectly in the centre of the field of view with half the pattern visible in each eye . Whereas a ±75pixel horizontal offset would remove the whole pattern from one eye’s visual field , and at ±200 pixels leave only a small portion of the pattern visible in just one eye ( Fig 2 ) . With zero offsets of the correct and incorrect test patterns , we found that the DISTINCT simulated bee was able to discriminate all of the presented pattern pairs . Indeed , despite its simplicity , the model design allowed it to outperformed real honeybees whose best result was 67% compared to DISTINCT simulated bee’s 78% accuracy for the same pattern pair ( Fig 2 ) . This model bee also discriminated the two pattern pairs that real honeybees failed to discriminate ( spiral patterns—bee: 53 . 7% p>0 . 7 n = 54 [18]–DISTINCT: 67% , octagonal patterns—bee: 56 . 4% p>0 . 2 n = 140 [18]–DISTINCT: 74% , see Fig 2 ) . The MERGED simulated bee results were far lower than the DISTINCT model’s discrimination accuracies but compared better to that of the experimental results . As with real honeybees’ behaviour , the MERGED simulated bee did not reliably discriminate the spiral and octagonal pattern , achieving simulation results of just 53% ( bee: 54% ) and 57% ( bee: 56% ) respectively . Out of the seven tested pattern pairs the only notable difference from the behavioural results was the MERGED bee’s inability to discriminate the two left / right reversed pattern pairs yielding only 49% and 52% respectively ( Fig 2 ) . Here honeybees achieved 62% and 65% in the behavioural experiments . Clearly the simpler model ( DISTINCT ) returned more accurate discrimination results and outperformed both the more derived model ( MERGED ) and the honeybees . Our results raise the interesting question why the honeybees performed so poorly on some of the patterns , when a very simple model ( DISTINCT ) was easily able to discriminate the patterns while using just eight large-field orientation-sensitive neuronal inputs . However , progressively offsetting the test patterns from the centre of the field of view revealed the lack of robustness of the DISTINCT model to cue variation . Here the simulation performances dropped much faster than that of the simulated bee using the MERGED model . In fact with as little as ±75 pixel offset ( where the whole pattern was still visible ) the performance of the DISTINCT simulated bee fell below 52% for all pattern pairs ( Fig 2 ) . With the MERGED model , all discriminable patterns ( >64% accuracy at 0 pixel offset ) ( still achieved accuracies above 60% when the patterns were offset by ±75 pixel . Even when these patterns were offset by as much as ±125 pixels rendering almost half of the patterns invisible the model’s lowest simulated performance for these experiments was 57%—i . e . markedly more than for the DISTINCT model . Beyond this offset distance , only the one pattern pair ( crosses , see Fig 2 ) was effectively discriminated , at a level of ≥59% accuracy during simulations even when only small portions of the patterns were still visible . Our results show that by simply combining inputs from both the left and right eyes onto mushroom body Kenyon cells , discrimination abilities are effectively freed of requiring perfect cue alignment on the retinae . Although this reduces the maximal discrimination accuracy , it allows for a much more robust and versatile employment of this cognitive tool in most realistic free flight navigation and resource locating scenarios . Experienced honeybee foragers may identify rewarding flowers based on those features that most reliably predict reward amongst the available flower species . Honeybees able to generalize to this limited feature set would reduce the need to learn all the exact features ( or indeed photographic templates ) of each individual flower type visited and subsequently having to best-match these numerous complex templates when foraging on novel or less frequented floral resources [10 , 20 , 23] . To explore these generalization abilities , Stach et al . [19 , 20] trained honeybees on two sets of six patterns where within each set there were similarly orientated bars in each quadrant of the patterns ( Fig 3 ) . They then tested the bees’ ability to generalize from these training patterns to novel variations of the patterns . Unlike the previous experiments , these bees were able to fixate a small distance from the pattern before their final choice selection was recorded when they actually touched either of the two test patterns . For our simulations we therefore presented all the patterns in the centre of the field of view with zero horizontal , or vertical , offset applied , assuming this would be where a honeybee would make its final decision . Fig 3 shows the experiments we simulated and the corresponding honeybee experimental results [19 , 20] . The overall average difference from the simulation performance of all 17 generalization experiments to the corresponding empirical results for the DISTINCT model’s simulated bee was -9 . 83% and just -7 . 77% for the MERGED model’s simulated bee . However , as a direct correlation comparison of the model performances and behavioural results is not appropriate ( see Methods ) , we followed the approach of the original studies [19 , 20] and compared the model results against the experimental performances within smaller batches of similar generalization type tasks . Our first batch of experiments , using patterns from Stach et al . 2004 [19] , tested simple generalization from the training sets of six patterns to three novel pattern pairs . The experimentally preferred test stimulus patterns had bars orientated in the same direction as the corresponding quadrants of the rewarding training patterns , versus the incorrect distractor patterns with a similar visual style to the matching correct test pattern but with bars orientated in different directions to those of the rewarding pattern in each quadrant . We found that simulations of both the DISTINCT and MERGED models produced simulated bee results almost identical to the honeybee behavioural results ( Fig 3 ) . Both the percentage of honeybee correct choice selections for correct test patterns and our simulated bees’ performances were all between 67% and 72% . Our second batch of experiments again followed the study of Stach et al . [19] , here the correct patterns had three quadrants with correctly orientated bars and the final quadrant did not , the incorrect test patterns had incorrectly oriented bars in all four quadrants . The DISTINCT model achieved ≥58% throughout but performed typically 5–10% below the honeybees ( Fig 3 ) . The simulated bee based on the MERGED model outperformed the simulated bee of the DISTINCT model on all test pattern pairs with simulation performances ranging from 61–72% , once again extremely similar to that of the honeybee behavioural result range of 65–74% . In our third batch of experiments utilizing the same Stach et al . dataset [19] , the correct and incorrect test stimuli were very similar , the correct patterns having correctly oriented bars in all four quadrants and the incorrect patterns had just one quadrant with incorrectly oriented bars . Simulations of the MERGED model failed to allow its bee to generalize to the correct pattern in three out of four experiments , with individual simulation trials failing to achieve a Kenyon cell similarity ratio of more than 0 . 5 ( Fig 3 ) . The DISTINCT simulated bee managed to correctly generalize all of these patterns but with low accuracy of just 56% to 60%; the corresponding honeybee results ranging from accuracies of 63% to 73% . Our fourth experiment set was compiled by taking test pattern pairs from the earlier work of Stach and Giurfa [20] . In this study , honeybees were presented with different combinations of either the original rewarding training pattern configuration , or the mirror image , or the left / right reversal of this layout . The DISTINCT model’s simulated bee was once again able to generalize correctly to all the experimental patterns ( Fig 3 ) . Although performing less well than real honeybees , the model showed similar lower generalization performances on the mirror image versus left-right patterns ( 56% ) compared to that of the original rewarding pattern versus the mirror image patterns ( 62% ) . The MERGED simulated bee typically achieved higher accuracies that were more similar to the honeybee results than that of the DISTINCT model’s bee , correct generalization performances ranged from +1% to -12% different to the empirical result . Of note , the bees achieved a surprising 82% correct choice accuracy on one of these test pattern pairs almost 10% higher than any other task , our models had high results on this experiment ( DISTINCT: 62% , MERGED: 66% ) but we did not see these particular simulations outperform all others . Only two of the eight test pattern pairs ( correct stimuli: original configuration , incorrect stimuli: left / right reversal ) failed to generalize correctly with a performance of just 51% ( individual simulation trial Kenyon cell similarity ratios ranging from 0 . 39 to 0 . 62 dependent on the particular pattern triplets presented ) compared to the honeybee correct choice selection of 69% . During simulations both the DISTINCT and MERGED simulated bees showed a preference for the left / right reversal configuration compared to the mirror image pattern , they also preferred the correct configuration to the mirror image layouts , as did real honeybees ( Fig 3 ) . The last of our experiment sets , again used patterns from Stach and Giurfa ( 2001 ) [20] . Fig 3 shows that both types of simulated bees were unable to generalize when presented with a chequerboard distractor pattern , with individual trial Kenyon cell similarity ratios as low as 0 . 4 ( i . e . ‘preferring’ the incorrect pattern ) . Conversely , honeybees always preferred left / right or mirror image versions of the rewarding pattern configuration to that of the chequerboard option with behavioural results of 65% and 74% respectively . Despite our models’ extreme simplicity , they largely predicted the honeybees’ generalization performances accurately for a majority of the tested pattern pairs . Our simulated bees did fail to generalize when the two test patterns were very similar ( Fig 3 ) . However , whereas honeybees were trained on both rewarding and unrewarding training patterns , our simulated bees only perceived the rewarding stimuli . This may account for some of the honeybees’ additional correct choice performance ( see Discussion ) . Nonetheless , these results indicate that seemingly ‘complex’ tasks do not require advanced cognition . Instead , our DISTINCT and MERGED models provide evidence that visual pattern recognition and classification may in fact be the emergent properties of connecting just a small number of large-field visual inputs .
Apparently sophisticated cognitive abilities are often seen as a result of an equally complex neuronal architecture . However , here , this view is fundamentally challenged . Despite honeybees having a tiny brain consisting of less than one million neurons ( as compared to eighty-six billion neurons in the human brain [24] ) , they still display an impressive range of cognitive abilities from learning to recognise pictures of human faces [25–27] to simple counting [28] . Using a modelling approach , we investigated how bees' ability to discriminate and generalize could be explained by simple neural networks . We have shown that for achromatic bar patterns , regularly used in honeybee behavioural experiments , bees may actually require very little sophistication in neuronal circuitry . The honeybee lobula orientation-sensitive neuron responses are thought [13 , 29] to be the result of the summation of smaller receptive field orientation-sensitive neurons in the bee lamina or medulla ( 1st , 2nd optic ganglia ) , similar to those found in other insect medullas [13 , 14 , 30–33] . This collation of smaller subunits allows the lobula orientation-sensitive neurons to encode a simplified summary of the oriented edges across the whole width of the bee eye . Although this means a bee cannot extract the exact retinotopic location or indeed orientation of individual edges through these neurons , our results show that , surprisingly , just eight of these large-field lobula neurons would be sufficient for the discrimination and generalization of the described patterns . Our models also demonstrate , despite their simplicity , that just a single layer of simple connections from the lobula orientation-sensitive neurons to the mushroom body Kenyon cells would suffice to reproduce the empirical generalization results between a given rewarding pattern and the two test patterns . In fact our models may have had a more difficult challenge than that of real bees . During training the honeybees were exposed to both the rewarding patterns with a sugar water reward but also an unrewarding ( water ) or even aversive solution ( quinine ) on the training distractor patterns , this differential training would allow the bees to learn both those features consistent with reward but also those pattern features that were to be avoided . There is empirical evidence to show that choice accuracy as well as the pattern features learnt by bees are affected by the training regime ( e . g . absolute conditioning ( no distractor pattern ) vs . differential conditioning [34 , 35] , and the penalty associated with a distractor [36–38] . Since it remains unclear how these different factors affect learning on the neuronal level , the theoretical models described here used very simple mathematics to calculate the similarity of the Kenyon cells responses to different stimuli , and from this produce theoretical simulated bee performances . Although this is very different to how learning would take place within the honeybee mushroom bodies , it did allow us to investigate how the lobula orientation-sensitive neuron responses alone may affect the honeybees’ performance during different discrimination and generalization experiments . In addition , it allowed us to study how different connections of the lobula neurons and Kenyon cells may also affect performance . Given our models employed no form of learning , it is all the more impressive that our simplified and experimentally disadvantaged simulated brains were able to generate largely similar results to actual bees . The Kenyon cell outputs of our models were achieved solely by the summation of either excitatory or inhibitory connections from the lobula orientation-sensitive neurons ( with predefined configurations , and fixed synaptic weights of either +1 or -1 respectively ) . These simulated Kenyon cell outputs allowed our simulated bees to discriminate and generalize the tested patterns with approximately 50% activation of their Kenyon cell populations ( due to the reciprocal lobula orientation-sensitive neurons to Kenyon cell connection types , see Methods ) ; we assumed , for comparison with our simple models , that some form of synaptic plasticity from the Kenyon cells to the mushroom body extrinsic neurons would allow the bees to associate the appropriate 50% active Kenyon cells to the rewarding training pattern , and from these adjusted synaptic weights make the behavioural decisions . However , neuronal recordings of the mushroom body lip , which receives olfactory input , shows just ~5% activation of the Kenyon cells mediated by a feedback inhibitory network in the mushroom body calyces [39] . It may be that when honeybees visit a correct pattern they can increase the firing rate or reduce the response latency of the Kenyon cells that fire , but potentially more importantly , may quiescent those Kenyon cells that incorrectly fired for the unrewarding , or punished , training pattern ( during differential training ) . In this case the 5% of the Kenyon cells that are active ( assuming the same value as for olfactory stimuli ) would potentially be optimal to associate the rewarding stimulus with sucrose reward . Additional research is required to see if this greater specificity would actually account for some of the honeybees’ higher performance over that of our current models . It should be noted that ~50% of the olfactory projection neurons to the mushroom bodies are highly active when a particular odour is presented [40] providing a population coding response to a given odour , this differs considerably to that of the optic lobe neurons that typically have more specific firing rate tuning curve responses to particular stimuli . Due to issues with harnessing bees during visual learning tasks we currently lack the ability to record Kenyon cell responses for anything but the simplest visual stimuli ( e . g . whole eye exposure to a single colour [41] ) . Unfortunately this means we do not yet have empirical evidence for the Kenyon cell activation level for visual stimuli . New research using walking bees in virtual reality rigs [42] may allow these activation levels , and Kenyon cell response changes , to be recorded during visual learning paradigms . These findings will undoubtedly provide vital information for the next generation of theoretical models , which could be used to understand the trial-by-trial learning process of bees . Despite the limitations mentioned above , our simulated bees still performed almost identically to the real bees when making simple generalizations and only dropped in performance when either the test patterns began to differ from the oriented edges presented in the rewarding patterns or the correct and incorrect test patterns became very similar ( Fig 3 ) . Here the difference in the honeybees’ exposure to the unrewarding as well as rewarding stimuli during training almost certainly contributed to the typical 5–10% performance advantage compared to our simulated bees , which only used the rewarding stimuli . Again , future behavioural and electrophysiological research may reveal how training paradigms affect the learning on the neuronal level , which would allow corresponding adjustments to the new theoretical models . During the offset pattern discrimination simulations ( Fig 2 ) we found that simply combining the neuronal firing rates of lobula orientation-sensitive neurons from each eye onto individual Kenyon cells would allow for pattern discrimination with an impressive location invariance of the perceived stimuli . By merging information from both eyes , a very coarse representation of the whole 270° bee eye horizontal field of view can be produced . Surprisingly , this non-retinotopic representation appears sufficient to discriminate quite complex visual patterns , removing the need for the bees to have to store an eidetic or ‘photographic’ view of the pattern . As a pattern is offset from the centre of the field of view , such that it is visible in one eye more than the other ( Fig 2 ) , then the firing rates of all eight neurons ( a type A and a type B lobula orientation-sensitive neuron in each of the four visual field regions—dorsal and ventral half of each eye ) will adjust according to the oriented edges each region now perceives . With the DISTINCT model , as the pattern is offset the changes in the total synaptic input per Kenyon cell ( compared to the zero offset pattern ) are quite pronounced—as these are directly influenced by the amount the lobula neurons response change due to the addition , or removal , of oriented edges in each separate region . Therefore , the DISTINCT model’s discrimination ability is impaired the further a pattern is offset . In contrast , with the MERGED model , although the lobula-orientation sensitive neurons’ firing rates are the same as the DISTINCT model , by combining the lobula neuron responses from the left and right eyes even as the pattern is offset the total summated Kenyon cell values remain similar to the summated values with no offset ( at least until the patterns begin to leave the field of view of both eyes ) . For example in the second generalization test the correct test stimuli had the orientation of bars in one quadrant of original rewarding pattern rotated through 90° ( Fig 3 ) ; here the DISTINCT model had a whole quadrant producing incorrect Kenyon cell responses , whereas in the MERGED model only a proportion of the whole dorsal or ventral field of view is altered and thus a smaller number of Kenyon cells ‘misfire’ . However , despite the typically good discrimination results over large offsets and the ability to discriminate when patterns are only partially visible , our results show that this mechanism may well come at the expense of discriminating certain types of stimuli . Complex spiral and octagonal patterns ( Fig 2 ) were not reliably discriminated by our simulated bee based on the MERGED model or by real honeybees [18] . Surprisingly , honeybees have been shown unable to discriminate a very simple pair of 90° cross patterns ( incorrect pattern rotated through 45° ) [11] ( Fig 4 ) , despite their apparent differences to a human observer . Simulations of these experiments once again showed the MERGED model’s simulated bee’s closer similarity to the honeybee behavioural results , with a sub 60% discrimination performance on these simple cross patterns , whereas the DISTINCT simulated bee achieved over 70% accuracy . Interestingly both of the simulated bees , and honeybees , were able to discriminate a pair of 22 . 5° rotated cross patterns easily ( incorrect pattern rotated through 90° ) ( Fig 4 ) . It may well be that in allowing the neuronal architecture of the honeybee brain to overcome location variance for common stimuli , it has compromised its ability to discriminate specific , arguably less important cue combinations . In a few specific instances our MERGED simulated bee failed to discriminate the tested pattern pairs , in contrast to the empirical results . The model’s inability to allow its simulated bee to discriminate the left / right reversal patterns in experiment four and six of the discrimination experiments ( Fig 2 ) and experiment four of the generalization experiments ( Fig 3 ) was no surprise as both the correct and incorrect test patterns presented the exact same orientations only in the reverse eyes , and hence produced the same summated input to the Kenyon cells , whereas the inability to discriminate the incorrect checkerboard pattern from the correct test patterns ( Fig 3 ) may be down to the lack of a predominant orientation in this stimulus causing lobula orientation-sensitive neuron outputs which were equally dissimilar from the rewarding patterns as the correct test patterns confusing the system . It is most likely that in these experiments and while observing other similar stimuli the honeybees use other visual features ( optic flow , symmetry , etc . ) to which our very simple models did not have access . In addition , the poor concordance of the MERGED model simulated bee results and the honeybees in the generalization experiments may also result from the experimental paradigm that allowed the bees to fixate on the pattern at close range and make their final decision from a fixed perspective . This would , for these experiments , be very similar to the better-performing DISTINCT model’s simulated bee with zero stimuli offsets . It is conceivable that honeybees have a combination of both DISTINCT and MERGED type lobula orientation-sensitive neuron to Kenyon cell configurations within their mushroom bodies . In this neuronally still simple scenario , attention-like processes could “selectively learn” the Kenyon cell responses that are good indicators of reward in a given experimental scenario . This might therefore account for some of the honeybees’ higher performance compared to that of our simulated bees based solely on the MERGED or DISTINCT models . Future work will investigate if there is an optimal distribution of distinct and merged lobula orientation-sensitive neuron connections to the Kenyon cells , or if synaptic plasticity is able to adjust the proportion of each connection type for a particular task . This modelling of bee visual processing and synaptic tuning may then be able to provide additional insights for machine vision applications where very lightweight computational solutions are required for object or landmark recognition , such as next generation self-drive vehicles and autonomous flight systems . Our research shows that very simple neuronal connections , which would be easily accommodated within the miniature brain of a bee , are able to facilitate seemingly complex visual cognitive tasks . In addition the merging of visual information from both eyes , as seen in the mushroom bodies of bees [16] , appears to be a very effective solution to partial occlusion and retinal location invariant pattern discrimination .
The simulated lobula ( 3rd optic ganglion ) large-field orientation-sensitive neurons used in our models were derived from the Yang & Maddess ( 1997 ) study on the honeybee ( Apis mellifera ) [13] . In these experiments , electrophysiological recordings where made from the lobula of tethered bees placed in front of CRT computer monitors; stimuli of oriented bars moving across one eye were presented at 30° angle intervals , in both the frontal and lateral eye regions . These neurons responded to the oriented bars moving anywhere across the whole width of the eye , but were maximally sensitive to orientations of 115° ( type A ) and 250° ( type B ) with angular half-widths of about 90° . We produced best-fit curves to both the reported type A and type B lobula orientation-sensitive neuron responses so that we could provide a theoretical neuronal response to a fixed 280-pixel edge at any orientation ( Fig 5 ) . Bees presented with two identically oriented bars simultaneously in both the frontal and lateral regions of the eye generated lobula orientation-sensitive neuron responses that were higher than for a single bar in either eye region but less than the summated responses [13] . A similar nonlinear response was seen in dragonflies ( Hemicordulia tau ) [14] where the response to an oriented moving bar would increase with the length of the presented bar . Assuming that these honeybee lobula neuronal responses are due to a nonlinear summation of smaller orientation detectors in the lower lobula or medulla , we used this more detailed response curve recorded in the dragonfly to generate a best-fit scale factor curve for when the length of a presented edge increases ( Fig 5 ) . This allowed us to scale the lobula orientation-sensitive neuron responses for any oriented edge based on its length compared to the fixed length used for our LOSN tuning curves . To account for multiple edges at different orientations in any one image , we again presume that the overall lobula orientation-sensitive neuron response is composed from smaller subunits in the medulla or early lobula and will vary with both the total length and abundance of all oriented edges within the receptive field that that neuron receives information from . We thus calculated the overall type A and type B responses for any given pattern using the edge length histogram datasets for all four quadrants of that pattern ( see below ) . For each quadrant and each lobula orientation-sensitive neuron type , we summated the proportion ( orientation edge length / total edge length ) of each edge orientation ( 0°-180° ) and multiplied it by the neural response for that orientation on our standard 280 pixel edge curve ( Fig 5 ) . This total value was then corrected by the scaling factor derived from the total edge length within that quadrant ( Fig 5 ) . This produced a type A and type B response ( Eq 1 ) for each quadrant of the visual field ( see Fig 6 ) and therefore eight lobula orientation-sensitive neuron responses in total for a given pattern . These image specific responses were saved with the pattern’s unique identification number ( UID ) and subsequently used as the sensory inputs to the Kenyon cells of our mushroom body models . Where LOSN: lobula orientation-sensitive neuron; x: LOSN type A or type B; q: visual field quadrant 1:4; H: matrix of edge lengths for each orientation ( 1° increments ) in each quadrant; C: response of LOSN type x to a 280 pixel edge at orientation; S: scale factor for given total edge length ( Fig 5 ) . Our first model , “DISTINCT” , uses excitatory and inhibitory connections from the lobula orientation-sensitive type A and type B neurons originating from each quadrant of the pattern , representing the equivalent dorsal and ventral visual fields of the bees left and right eyes , respectively ( see Fig 1 ) . This allowed us to evaluate discrimination and generalisation performance of visual patterns based on these lobula neurons alone . We used 86 different types of simple excitatory and inhibitory synaptic configurations of the lobula orientation-sensitive neurons to Kenyon cells to achieve the 25°–30° orientation acuity reported for honeybees during dual trial discrimination tasks [43] ( see Table 1 for Kenyon cell synapse configurations ) . The lobula neuron to Kenyon cell synaptic weights were fixed at +1 for the excitatory synapses , and at -1 for the EAI inhibitory synapses , such that the ± synaptic value of each Kenyon cell’s synapse would be the same as the single lobula orientation-sensitive neuron’s firing rate to which it connects ( with a small amount of noise applied , see below ) . This model could have just as easily been configured to receive , for example , just one type B input with a synaptic weight of +3 , which would have produced the exact same effect as three excitatory lobula orientation-sensitive neuron type B inputs ( Fig 1 ) . However , to reinforce the importance that there is no learning in our models , and to focus the investigation into the lobula neuronal responses , here we restrict the models to the most basic synaptic configuration , with all synaptic weights equal to ±1 . The model had 30 copies of each of these Kenyon cell configuration types per quadrant , resulting in a total of 10 , 320 Kenyon cells , which is still a small proportion of the 340 , 000 Kenyon cells in the honeybee mushroom bodies [44] . The theoretical Kenyon cell connections defined above ( Table 1 ) will each fire for a large number of perceived edge orientations and edge lengths . However , the combinatorial firing code of these 86 types allows small ranges of orientations to be uniquely identified by our models , and furthermore these edge orientations can be recognised invariant of the presented edge lengths since an almost identical combinatorial code of the fired Kenyon cells is produced if the same edge orientations are presented ( see below ) . Adding additional lobula orientation-sensitive neuron combinations would not increase the models ability to discriminate more specific angles , as the acuity is fundamentally constrained by the particular lobula neuron response curves , which often have the same firing rate for several adjacent orientations ( Fig 5 ) . It is most likely that within the honeybee mushroom bodies a large variety of random lobula neuron to Kenyon cell synaptic connections are initially established . Equally these synapses are almost certainly plastic , adapting the synaptic strengths , and even adding and removing lobula neuron synapses , during a bee’s foraging life [45] . In this way these Kenyon cells could become highly selective and fire only for particular rewarding visual inputs . In addition , the honeybee brain may be capable of adjusting the Kenyon cell synapse strengths to better account for noise in the lobula orientation-sensitive neuron responses and produce more effective combinatorial codes for identifying particular orientations than our models ( see Discussion ) . However , since this study is primarily concerned with the lobula orientation-sensitive neurons effectiveness as feature detectors and their affect on the honeybees’ ability to discriminate and generalize achromatic patterns , and not on learning or other ‘fine-tuning’ neuronal mechanisms , this additional model complexity of random connectivity and weight adaption was omitted . Each models’ Kenyon cell response , to a given pattern , was calculated by first summating the value of all its synapses ( number and type of synapses dependent on that Kenyon cells particular configuration type ( Table 1 ) ) . If this total summated synaptic input was greater than zero the output of the Kenyon cell was set to 1 ( fired ) . Otherwise the response was set to 0 ( completely inhibited ) . The individual Kenyon cell synaptic values were calculated by taking the firing rate of the connected lobula orientation-sensitive neuron , plus a small synaptic signal to noise distortion , and multiplying this by +1 for excitatory synapses and -1 for inhibitory ones . The noise was added to account for natural variation in both the lobula orientation-sensitive neurons’ responses when presented with the same pattern , and in pre- and post- synaptic neurotransmitter signals . Matlab’s ( Matworks ) AWGN ( add white Gaussian noise to signal ) function was used with a signal to noise ratio value of 30 . This setting produced approximately 2–5Hz variations on the 36Hz response of the type A lobula orientation-sensitive neuron at its maximal sensitivity and an edge length of 280 pixels . This would be similar to the response variation reported in the honeybee lobula neurons after the deduction of the neuronal background firing rates [13] . In this way the binary values of all 10 , 320 Kenyon cell responses were calculated; these values were stored in an array and saved cross-referenced to the pattern’s UID . Given the apparent non-retinotopic distribution of visual inputs from the corresponding left and right eye regions in the bee mushroom bodies [16] the second model “MERGED” was created to explore the effect of merging lobula orientation-sensitive neuron synaptic connections from both eyes onto the Kenyon cells . To keep our theoretical model simple and comparable to the DISTINCT model , we again relied on the 86 lobula neuron to Kenyon cell configuration types ( Table 1 ) . However , in this model , rather than the previous model’s segregation of Kenyon cells into different groups per quadrant , here just two distinct groups of Kenyon cells were formed; one group of Kenyon cells all received lobula orientation-sensitive neuron type A and type B inputs from the dorsal regions of both left and right eyes , and a similar group of Kenyon cells receiving the four lobula inputs from the ventral regions of the eyes ( see Fig 1 ) . This MERGED model again had 30 copies of each configuration type , which created in total 5 , 160 Kenyon cells . Each achromatic pattern used in this study was taken from the pdf document of the published behavioural papers . These images were scaled and centred to fit within a 150 x 150 pixel PNG image . Where pattern image resolution was insufficient , we recreated the patterns in Microsoft PowerPoint using the stimuli instructions provided in the papers’ method sections . For the offset discrimination experiments , the 150 x 150 pixel patterns were placed centrally within a larger white 300 x 150 pixel image and horizontally offset left and right between 0 and 200 pixels in 25 pixel increments to create a set of 17 test images per original pattern . For offsets greater than 75 pixels the original images were cropped accordingly ( see Fig 2 ) . All images were processed in Matlab ( Mathworks ) in the following way: Each experiment simulated in this study was composed of three patterns , the rewarding pattern ( CS+ ) used during the honeybee training and two novel test patterns used in the experimental evaluation trial . The test stimuli patterns that honeybees preferred during their trials were designated as correct test stimuli ( TSCOR ) and the incorrect test stimuli ( TSINC ) were accordingly the patterns the bees least preferred . To simulate the experiments from published behavioural work , we first pre-processed the lobula orientation-sensitive neuron responses for all the used patterns and compiled them in an experiment-specific unique Matlab ( Mathworks ) file ( hereafter referred to as “study file” ) . For each individual experiment within a study file , we defined the CS+ , TSCOR and TSINC pattern unique identifiers ( UIDs ) as well as recoding the behavioural results of the honeybees . For each model we loaded the study file , extracted the unique pattern image IDs for each experiment and the corresponding eight lobula neuron firing rate values and from these calculated the model’s Kenyon cell responses to all three patterns . This provided separate arrays of binary Kenyon cell responses for the three patterns ( rewarding stimulus , correct test stimulus and incorrect test stimulus ) , which we used to calculate the Euclidian distance from the rewarding stimulus array to the correct stimulus array , and rewarding stimulus array to the incorrect stimulus array ( see Fig 7 ) . The ratio of these two Euclidian distances produced a Kenyon cell similarity ratio for that experiment for a single simulation trial ( Eq 2 ) . Each experimental simulation was repeated one thousand times and the average , standard deviation , minimum and maximum Kenyon cell similarity ratio results of each experiment were recorded . Where KCSR: Kenyon cell similarity ratio; E ( x , y ) : Euclidian distance between x and y; CS+ , TSCOR , TSINC: array of Kenyon cell response values for the respective patterns . For the generalisation experiments , honeybees had been trained on multiple rewarding and unrewarding pattern pairs selected from relevant pools ( Fig 3 ) [19 , 20] . We followed the same procedure as above but created individual simulations for each possible pattern triplet combination . As the behavioural results also included the choice selections of different groups of bees trained on the reciprocal of the learned association ( i . e . rewarding patterns became unrewarding patterns , and vice versa ) , we used the published unrewarding training patterns as a new set of rewarding ( CS+ ) simulation patterns and paired them with the according correct and incorrect test patterns . Simulations were again performed one thousand times for all pattern triplet combinations . The Kenyon cell similarity ratio results for all combinations were then averaged to create an overall Kenyon cell similarity ratio value for that particular pattern test . Due to the difficulties attaining electrophysiological recordings from honeybees during visual learning tasks [41 , 48–50] we know almost nothing about how a bee’s final behavioural decision is underpinned by neuronal firing patterns in the visual system or mushroom bodies . However , we can assume that if the Kenyon cell responses to a presented test stimulus are very similar to those generated by a previously learnt rewarding training stimulus ( i . e . the same pattern is presented ) and the distractor pattern is very different to the learnt rewarding pattern , then the honeybee Kenyon cell similarity ratio would be almost 1 . 0 , and we would expect the bee to almost always visit the correct test pattern , with an experimental correct choice performance close to 100% . Similarly , if the correct and incorrect test patterns are different from each other and also different to the learnt rewarding pattern , but both produced Kenyon cell responses equally similar/dissimilar to that of the rewarding pattern ( i . e . Kenyon cell similarity ratio = 0 . 5 ) then we would expect the honeybee to visit each pattern equally likely , and therefore over multiple trials ( and multiple bees ) have an experimental ‘correct’ choice performance of approximately 50% . Furthermore , if the honeybees were trained on a particular rewarding pattern and then tested with a correct test pattern similar to this learnt stimulus and a very different incorrect test pattern , and then a second test conducted with the same correct pattern and a very similar incorrect pattern , we would again assume the honeybees correct choice accuracy for the first test would be far higher than the second test . Similarly , the Kenyon cell similarity ratio of the first experiment would undoubtedly be much higher than that of the Kenyon cell similarity ratio of the second experiment . Consequently , to allow us to compare our model simulation results directly against the empirical honeybee experimental results we make the following assertion: our models’ simulated bee performances for any given experiment are directly correlated to the average Kenyon cell similarity ratio of all simulation trials for that experiment . In this way if a model’s average Kenyon cell similarity ratio for a given experiment were 0 . 64 then its simulated bee’s overall experimental performance for selecting the correct test pattern would be 64% . It would have been possible to implement a probabilistic ‘Monte Carlo’ style binary response for the simulated bees to choose either the corrector incorrect test pattern per trial ( based on that simulation trial’s Kenyon cell similarity ratio result ) and subsequently calculate the proportion of correct choices ( as with honeybees ) . However , this would have added probabilistic variability , whereas the Kenyon cell similarity ratio values are variant on just the small amount of synaptic noise applied to the lobula orientation-sensitive neuron to Kenyon cell connections ( which is biologically relevant ) , therefore this additional probabilistic step was judged an unnecessary and potentially detrimental complication . The above assertion does have some limitations when assuming a direct comparable mechanism within the honeybee brain ( see Discussion ) , but nonetheless this provides an effective method for assessing how the lobula orientation-sensitive neuron responses , as well as their Kenyon cell connection configurations , affect the models’ performances over a wide range of pattern experiments . This mechanism also benefits from not needing to train and test an artificial neural network on each pattern experiment , and the inherent parameter tuning and subsequent performance evaluations that this approach would require . It would have been desirable to assess how our models correlated with the honeybees’ relative performances over all of the tested experiments . Each set of the original honeybee generalisation experiments [19 , 20] only provided a number of mean data points for comparison . In each study , the bees were tested on patterns that typically varied in one particular aspect ( e . g . number and orientation of bars in each pattern quadrant ) , but were similar otherwise . Moreover , the used publications addressed similar issues and used similar patterns . While this is a good approach when probing the limits of the learning abilities of bees , it also means that the data points are not independent due to pseudo-replication . A correlation coefficient involving data from multiple different experiments would , therefore , be misleading . Instead , we displayed our simulated bee experimental performance results ( equivalent to the Kenyon cell similarity ratio averages over all of that experiment’s simulations ) side-by-side with the empirical data . These were grouped into five batches of related generalization tasks , similar to that done in the original studies , so that the relative performance of the different simulated experiments could be assessed , and compared to that of the real honeybees’ relative performances on the same sets of pattern pairs . | We present two very simple neural network models based directly on the neural circuitry of honeybees . These models , using just four large-field visual input neurons from each eye that sparsely connect to a single layer of interneurons within the bee brain learning centres , are able to discriminate complex achromatic patterns without the need for an internal image representation . One model combining the visual input from both eyes showed an impressive invariance to the location of the test patterns on the retina and even succeeded with the partial occlusion of these cues , which would obviously be advantageous for free-flying bees . We show that during generalization experiments , where the models have to distinguish between two novel stimuli , one more similar to a training set of patterns , that both simple models have performances very similar to the empirical honeybee results . Our models only failed to generalize to the correct test pattern when the distractor pattern contained only a few small differences; we discuss how the protocols employed during training enable honeybees to still distinguish these stimuli . This research provides new insights into the surprisingly limited neurobiological complexity that is required for specific cognitive abilities , and how these mechanisms may be employed within the tiny brain of the bee . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"learning",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"honey",
"bees",
"nervous",
"system",
"social",
"sciences",
"electrophysiology",
"neuroscience",
"animals",
"learning",
"and",
"memory",
"simulation",
"and",
"modeling",
"cognitive",
"psychology",
"vi... | 2017 | Insect Bio-inspired Neural Network Provides New Evidence on How Simple Feature Detectors Can Enable Complex Visual Generalization and Stimulus Location Invariance in the Miniature Brain of Honeybees |
For the past five years , genome-wide association studies ( GWAS ) have identified hundreds of common variants associated with human diseases and traits , including high-density lipoprotein cholesterol ( HDL-C ) , low-density lipoprotein cholesterol ( LDL-C ) , and triglyceride ( TG ) levels . Approximately 95 loci associated with lipid levels have been identified primarily among populations of European ancestry . The Population Architecture using Genomics and Epidemiology ( PAGE ) study was established in 2008 to characterize GWAS–identified variants in diverse population-based studies . We genotyped 49 GWAS–identified SNPs associated with one or more lipid traits in at least two PAGE studies and across six racial/ethnic groups . We performed a meta-analysis testing for SNP associations with fasting HDL-C , LDL-C , and ln ( TG ) levels in self-identified European American ( ∼20 , 000 ) , African American ( ∼9 , 000 ) , American Indian ( ∼6 , 000 ) , Mexican American/Hispanic ( ∼2 , 500 ) , Japanese/East Asian ( ∼690 ) , and Pacific Islander/Native Hawaiian ( ∼175 ) adults , regardless of lipid-lowering medication use . We replicated 55 of 60 ( 92% ) SNP associations tested in European Americans at p<0 . 05 . Despite sufficient power , we were unable to replicate ABCA1 rs4149268 and rs1883025 , CETP rs1864163 , and TTC39B rs471364 previously associated with HDL-C and MAFB rs6102059 previously associated with LDL-C . Based on significance ( p<0 . 05 ) and consistent direction of effect , a majority of replicated genotype-phentoype associations for HDL-C , LDL-C , and ln ( TG ) in European Americans generalized to African Americans ( 48% , 61% , and 57% ) , American Indians ( 45% , 64% , and 77% ) , and Mexican Americans/Hispanics ( 57% , 56% , and 86% ) . Overall , 16 associations generalized across all three populations . For the associations that did not generalize , differences in effect sizes , allele frequencies , and linkage disequilibrium offer clues to the next generation of association studies for these traits .
Since its introduction in 2005 , the genome-wide association study ( GWAS ) design has become a powerful tool in human genetics to identify single nucleotide polymorphisms ( SNPs ) associated with common diseases or traits using an experimental design that does not require a priori biological knowledge . As of September 2010 , greater than 1 , 000 SNPs across the genome have been reported as genome-wide significant ( p≤5×10−8 ) for 165 traits [1] . An early analysis of the GWAS-reported SNPs demonstrated that most identified variants were intergenic or intronic [2] , suggesting either novel biology or that the functional variant has yet to be found . While GWAS have been successful in identifying novel associations , there are several limitations . First , the majority of GWAS have been conducted in populations of European-descent . There are several GWAS in populations of Asian-descent , and GWAS are just emerging for other populations such as African Americans [3]–[20] , Mexican Americans/Hispanics [9] , [20]–[26] , and American Indians [27] . It is possible that novel associations await discovery in these populations given the differing linkage disequilibrium ( LD ) patterns when compared with populations of European-descent [28] . Second , much work is needed to test SNPs discovered in case-control studies in more population-based , representative cohorts to determine if the associations generalize . Data on generalization will inform future fine-mapping [29] and discovery studies as well as provide clues to whether GWAS-identified SNPs are simply tagSNPs or are more likely to be true functional SNP ( s ) . A major goal of the Population Architecture using Genomics and Epidemiology ( PAGE ) study is to determine whether GWAS-identified variants generalize to diverse groups drawn from population-based studies [30] . Generalization is defined here as a significant association ( p<0 . 05 , uncorrected for multiple testing ) in a non-European population and a direction of genetic effect in the same direction as that of European Americans . In PAGE , variants identified in GWAS and well replicated in multiple studies are chosen for targeted genotyping in hundreds to thousands of European Americans ( ∼20 , 000 ) , African Americans ( ∼9 , 000 ) , American Indians ( ∼6 , 000 ) , Mexican Americans/Hispanics ( ∼2 , 500 ) , Japanese/East Asians ( ∼690 ) , and Native Hawaiians/Pacific Islanders ( ∼175 ) . All samples are linked to extensive demographic , health , and exposure data , making the PAGE study a rich resource for post-discovery generalization and characterization for common human diseases and traits . We present here PAGE study data on the replication and generalization for 49 SNPs associated with three common lipid traits: low-density lipoprotein cholesterol ( LDL-C ) , high-density lipoprotein cholesterol ( HDL-C ) , and triglycerides . Each of these three traits has numerous GWAS published in European ancestry individuals [30]–[43] but only a handful published in other populations ( such as Asians [44] and Micronesians [45] ) . Additional data are just now emerging from large sample sizes of diverse populations for generalization [32] , [46]–[51] and fine-mapping [52] of these lipid GWAS-identified SNPs . We demonstrate that the majority of the targeted GWAS-identified SNPs replicate in European Americans in PAGE and that many generalize to diverse populations . Both power and LD are explored as explanations of non-generalization , highlighting the complexities involved in properly interpreting results of even robust genetic associations such as these .
The PAGE study sites are diverse across multiple variables ( Table 1 and Table S1 ) . Together , the PAGE study consists of several populations: European Americans , African Americans , Mexican Americans/Hispanics , American Indians , Japanese/East Asians , and Native Hawaiians/Pacific Islanders . All PAGE study sites except WHI ascertained both men and women . Participant age varies widely across PAGE . For example , CHS ascertained on average older adults ( median age = 74 and 72 years for European and African Americans , respectively ) , CARDIA ascertained younger adults ( median age = 26 and 24 . 5 years for European and African Americans , respectively ) , and NHANES ascertained all ages of adults ( 18 years to 90 years; median age = 51 , 39 , and 40 years for European , African , and Mexican Americans , respectively ) . In addition to demographic differences , lifestyles and health differed across the PAGE study sites by population , including lipid lowering medication use and current smoking status . More Japanese participants ascertained by MEC reported lipid lowering medication use compared with other populations ascertained by other PAGE study sites: 38 . 3% versus <5–10% . American Indians from the Dakotas reported more smoking ( 42 . 2–47 . 8% ) than other American Indians ( 25–33% ) or other PAGE study site populations ( 6 . 3% to 35 . 3% ) . The differences in demographics , lifestyle , and health characteristics observed across the PAGE study sites and populations are reflected in the three traits studied here ( Table S1 ) . Given the diversity observed across the PAGE study sites , we performed all tests of association for HDL-C , LDL-C , and triglycerides unadjusted , minimally adjusted ( for age and sex ) , and adjusted for various demographic , lifestyle , and health variables . Coded allele frequencies are presented in Table 2 , Table 3 , Table 4 and in Figure S1 , by population . We calculated the Pearson correlation coefficient ( r ) and FST between European American coded allele frequencies and all other groups . The highest correlation was observed in the comparison with Mexican Americans/Hispanics ( 0 . 97 ) followed by American Indians ( 0 . 92 ) , Native Hawaiians/Pacific Islanders ( 0 . 90 ) , Japanese/East Asians ( 0 . 87 ) , and African Americans ( 0 . 84 ) . Compared with European Americans , the proportion of SNPs with FST values greater than 0 . 15 was smallest in Mexican Americans/Hispanics ( 0/49 SNPs ) and largest in African Americans ( 6/49 SNPs; 12% ) followed by Japanese/East Asians ( 5/46 SNPs , 11% ) . FST values were small for the remaining populations compared to European Americans , with 3% and 7% of SNPs with FST values greater than 0 . 15 for American Indians and Native Hawaiians/Pacific Islanders , respectively . A striking example of population differences in allele frequencies is FADS1 rs174547 . The T allele of FADS1 rs174547 is the major allele in three populations ( allele frequency = 0 . 66 , 0 . 91 , and 0 . 59 in European Americans , African Americans , and Japanese/East Asians , respectively ) , but is the minor allele in the other three populations ( allele frequency = 0 . 39 , 0 . 21 , and 0 . 42 in Mexican Americans/Hispanics , American Indians , and Native Hawaiians/Pacific Islanders , respectively ) . Compared to European Americans , FST for this SNP was largest in American Indians ( 0 . 34 ) followed by African Americans ( 0 . 15 ) . We also compared allele frequencies between the various PAGE study sites , within each racial/ethnic group . As demonstrated in Figure S2 , the allele frequencies of European Americans , African Americans , and Mexican Americans/Hispanics do not differ substantially across PAGE studies ( allele frequencies differ by less than ±0 . 10 ) . In contrast , over half of the SNPs genotyped in American Indians had allele frequency differences greater than ±0 . 10 , with three SNPs with allele frequencies that differed by more than ±0 . 25 . Comparisons are more difficult in Japanese/East Asians and Native Hawaiians/Pacific Islanders , as many SNPs were genotyped by only one PAGE study in these two racial/ethnic groups . We meta-analyzed tests of association for 27 , 19 , and 14 SNPs previously associated with HDL-C , LDL-C , and/or triglycerides , respectively , across European American populations collected by individual PAGE study sites ( Table S2 ) . For HDL-C , 23 of the 27 ( 85% ) SNPs tested were associated at p<0 . 05 assuming an additive genetic model and adjusting for age and sex ( Figure 1 and Table 2 ) . The four SNPs that did not replicate at this liberal significance threshold were rs471364 ( TTC39B ) , rs1883025 ( ABCA1 ) , rs4149268 ( ABCA1 ) , and rs1864163 ( CETP ) , all of which are intronic ( Table S2 ) . For LDL-C , only one ( intergenic MAFB rs6102059 ) of the 19 SNPs tested was not significantly associated at p<0 . 05 ( Figure 1 and Table 3 ) . Finally , for ln ( TG ) , all 14 SNPs tested were associated at p<0 . 05 ( Figure 1 and Table 4 ) . Of the associations that did not replicate in the European-descent populations from PAGE , four out of five had sufficient power ( >80% ) to detect the previously reported effect size: TTC39B rs471364 ( >99% power; HDL-C ) , CETP rs1864163 ( 80% power; HDL-C ) ; MAFB rs6102059 ( >90% power; LDL-C ) , and ABCA1 rs4149268 ( 99% power; HDL-C ) . ABCA1 rs1883025 , which did not replicate the expected association with HDL-C , did not have sufficient power to detect the reported effect size ( 68% power; n = 3 , 865 ) . We then compared the genetic effect sizes reported in the literature to the genetic effect sizes estimated from the meta-analysis of these population-based studies . We observed that the majority of the point estimates of effect size ( β ) were smaller than previously reported estimates . Using the HDL-C association results as an example , 15 out of the 23 ( 65% ) significant associations had effect estimates smaller than published effect estimates . We caution , however , that we did not formally test for significant differences between estimates and that these smaller effect estimates may or may not be significantly different than the published reports . However , it is interesting to note that 11 of our effect estimates differed from previous reports by more than 25% , including two HDL-C associations whose effect sizes differed by 50% or more from those in the literature ( ANGPTL4 rs2967605 and MLXIPL rs17145738; Table 2 and Table S2 ) . We meta-analyzed tests of association performed in African Americans for the same 27 , 19 , and 14 SNPs previously associated with HDL-C , LDL-C , and/or triglycerides in populations of European-descent . For all three traits studied , assuming an additive genetic model and adjusting for age and sex , approximately half of the tested GWAS-identified SNPs were associated at p<0 . 05: 12/27 ( 44% ) for HDL-C , 11/19 ( 58% ) for LDL-C , and 8/14 ( 57% ) for ln ( TG ) ( Figure 1 , Figure S3 , Table 2 , Table 3 , Table 4 , Table 5 ) . The majority of SNPs that failed to replicate in the meta-analysis for European Americans also failed to associate in the meta-analysis for African Americans . Interestingly , one SNP ( CETP rs1864163 ) was significantly associated with HDL-C in African Americans ( n = 451; CAF = 0 . 27; β = −2 . 79; p = 6 . 19×10−3 ) but not in European Americans ( n = 291; CAF = 0 . 23; β = −2 . 07; p = 0 . 13 ) . Other populations that were examined for select SNPs included American Indians , Mexican Americans/Hispanics , Japanese/East Asians , and Native Hawaiians/Pacific Islanders . Among American Indians , 9/21 ( 43% ) , 10/14 ( 71% ) , and 10/13 ( 77% ) of the SNPs tested for association with HDL-C , LDL-C , and ln ( TG ) , respectively , were associated at the liberal significance threshold of p<0 . 05 . For Mexican Americans/Hispanics , 14/27 ( 52% ) , 10/19 ( 53% ) , and 12/14 ( 86% ) SNPs were significantly associated at p<0 . 05 with HDL-C , LDL-C , and ln ( TG ) , respectively . Despite a small sample size , intronic CETP rs1864163 was significantly associated with HDL-C in Mexican Americans/Hispanics ( n = 265; CAF = 0 . 28; β = −2 . 98; p = 1 . 78×10−2 ) but not in European Americans ( n = 291; CAF = 0 . 27; β = −2 . 07; p = 0 . 13 ) , although the size and the direction of effect were similar . Venn diagrams representing the overlap of significant associations across the four major PAGE populations are presented in Figure S3 . The sample sizes for Japanese/East Asians and Native Hawaiians/Pacific Islanders are considerably smaller compared with the other populations examined . Despite the lower power to detect associations , significant associations were observed for both groups at a liberal significance threshold of p<0 . 05 . Among the 26 , 18 , and 13 SNPs tested for associations with HDL-C , LDL-C , and ln ( TG ) , respectively , there were nine ( 35% ) , three ( 17% ) , and three ( 23% ) SNPs significantly associated in the combined Japanese/East Asian group . For Native Hawaiians/Pacific Islanders , the group with the smallest sample size considered here , one SNP each was associated with HDL-C ( APOA1/C3/A4/A5 gene cluster rs28927680 ) and LDL-C ( APOB rs754523 ) out of the 24 and 18 SNPs tested for association , respectively . Three out of 12 SNPs tested for an association with ln ( TG ) were associated at p<0 . 05 ( PLTP rs7679 , MLXIPL rs17145738 , and APOA1/C3/A4/A5 gene cluster rs28927680 ) , with the latter at a significance of p<10−19 . For the 55 SNP-trait associations that replicated in European Americans , we determined which associations generalized across all four of our largest populations ( European Americans , African Americans , American Indians , and Mexican Americans/Hispanics ) . Generalization was based on two criteria: 1 ) level of significance ( i . e . p-value ) and 2 ) direction of effect ( i . e . positive or negative beta ) . SNPs that were significantly associated at p<0 . 05 and had the same direction of effect as European Americans in all populations studied were considered to have generalized . For HDL-C , five SNPs ( CETP rs3764261 , LPL rs6586891 , LIPC rs4775041 , LPL rs2197089 , and APOA1/C3/A4/A5 gene cluster rs3135506 ) met these criteria , and two SNPs ( LCAT rs2271293 and LPL rs328 ) were associated in three groups and trended towards significance in a fourth group ( p = 0 . 06 and p = 0 . 07 in Mexican Americans/Hispanics and American Indians , respectively; Table 2 ) . For LDL-C , six SNPs generalized across all four groups , if genotyped: APOB rs562338 , CELSR2/PSRC1/SORT1 rs599839 and rs646776 , PCSK9 rs11591147 , HMGCR rs12654264 , and LDLR rs2228671 ( Table 3 ) . Similarly for ln ( TG ) , six SNPs were significantly associated across the four largest populations: APOA1/C3/A4/A5 gene cluster rs964184 and rs3135506 , GCKR rs780094 , LPL rs328 , MLXIPL rs1714573 , and FADS1 rs174547 . In addition , for ln ( TG ) , two SNPs ( LPL rs2197089 and GCKR rs1260326 ) were associated in three groups and trended towards significance in a fourth group ( p = 0 . 07 in African Americans and p = 0 . 09 in American Indians , respectively ) . Among the 17 SNPs that generalized across the largest groups among the three lipid traits , only four ( 24% ) were either nonsense ( rs328 ) or missense SNPs ( rs3135506 , rs11591147 , and rs1260326; Table S2 ) . Based on our definition of generalization , several SNPs discovered and replicated in European-descent populations failed to generalize to other populations . There are several possible explanations for non-generalization , including power . To further investigate potential lack of power , we first performed post-hoc power calculations assuming an additive genetic model and liberal significance threshold ( 0 . 05 ) in each racial/ethnic group for each test of association . In these power calculations , we further assumed the observed genetic effect size ( beta ) from PAGE European Americans and the observed allele frequency , sample sizes , and trait mean/standard deviations from each non-European American population . By adding the power of all tested loci , we estimated the number of expected significant associations and compared this to the number of observed significant associations ( Table 5 ) . In general , the number of expected significant associations was greater than the number observed . African Americans consistently had fewer significant associations ( 11 , 11 , and 8 for HDL-C , LDL-C , and ln ( TG ) , respectively ) than expected ( 17 . 3 , 14 . 7 , and 11 . 9 for HDL-C , LDL-C , and ln ( TG ) , respectively ) based on power , regardless of the lipid trait being tested . More specifically , we were powered to detect in African Americans 17 of the 25 associations that replicated in European Americans but failed to generalize to African Americans . Compared to African Americans , differences between the observed and the expected number of associations for American Indians and Mexican Americans/Hispanics were less extreme . In fact , for ln ( TG ) , more significant associations were detected in these two populations than the PAGE study was powered to detect ( 8 . 4 and 10 . 4 expected; 10 and 12 observed for American Indians and Mexican Americans/Hispanics , respectively; Table 5 ) . We were powered to detect in American Indians nine of the 18 associations that replicated in European Americans but did not generalize to American Indians . Similarly , we were powered to detect in Mexican Americans/Hispanics eight of the 20 associations that replicated in European Americans but failed to generalize to Mexican Americans/Hispanics . To examine whether LD can account for the lack of generalization of the properly powered tests of association in African Americans , we examined LD patterns in HapMap Europeans ( CEU ) and West Africans ( YRI ) as well as those published in the literature for the genotyped SNPs and surrounding variation . For APOA1/C3/A4/A5 rs28927680 , previous studies in European-descent populations have noted that this SNP is in strong LD ( r2 = 0 . 98 ) with missense APOA5 rs3135506 [42] . APOA1/C3/A4/A5 rs964184 is also in moderate LD with missense rs3135506 ( r2 = 0 . 510 in CEU ) . However , neither rs28927680 nor rs964184 are in LD with missense rs3135506 ( r2 = 0 . 039 and r2 = 0 . 048 ) in YRI . Furthermore , APOA5 rs3135506 is associated with HDL-C in European Americans , African Americans , Mexican Americans/Hispanics , and American Indians ( Table 1 and Table 2 ) . Generalization of rs3135506 coupled with non-generalization and differences in YRI LD patterns for rs28927680 and rs964184 suggest that APOA5 rs3135506 is either the putative functional SNP for the association with HDL-C or in LD with the functional SNP . Although the exact mechanism is not yet known , molecular modeling [53] as well as in vitro [53] and in vivo [54] , [55] studies support the epidemiologic evidence that rs3135506 is functional . Other interpretations of LD patterns are more difficult . For example , CETP rs9989419 , which failed to generalize in African Americans for HDL-C despite sufficient power , is not in strong LD with obvious functional SNPs in CEU within 50 kb flanking the genotyped SNP . The strongest pair-wise LD ( r2 = 0 . 251 ) consists of intergenic and intronic SNPs , and these same SNPs have weak LD ( r2<0 . 03 ) or are not found in YRI . Similarly , LIPC rs261332 associated with HDL-C levels in European Americans but failed to generalize in African Americans . LIPC rs261332 is in strong LD ( r2>0 . 80 in CEU ) with SNPs in the 5′ flanking region of LIPC , but not in LD with these same SNPs in YRI ( r2<0 . 15 ) . Genetic variations in isolation are not the sole determinants of lipid trait distributions . Many environmental exposures and demographic variables are associated with lipid traits . To account for these variables , we meta-analyzed all tests of association for HDL-C , LDL-C , and ln ( TG ) adjusted for age , sex , body mass index , current smoking , type 2 diabetes , post-menopausal status , and current hormone use . Adjustment for these additional covariates did not appreciably alter the results compared with the models minimally adjusted for age and sex ( Figures S4 , S5 , S6 ) . Inclusion of previous myocardial infarction as a variable to the fully adjusted model also did not appreciably alter the results compared with the minimally adjusted models ( Figures S4 , S5 , S6 ) . All analyses presented thus far include fasting adult participants regardless of lipid lowering medication use . Many GWAS conducted for the lipid traits excluded participants on lipid lowering medication [40] , [42] , [43] given that these medications substantially lower LDL-C levels . We have included these participants for analysis as participants on lipid lowering medication could represent the upper extreme of the normal LDL-C distribution associated with a genetic profile found in a general population . Exclusion of these participants would preclude these meta-analyses from fully describing the extent and strength of associations relevant to these traits in a population-based setting . However , if genetic variation is associated with lipid concentrations and medication use lowers lipid concentrations , inclusion of participants on lipid lowering medications could bias associations towards the null . As a sensitivity analysis , WHI used detailed medication data available on a subset of participants , and performed the tests of association for HDL-C , LDL-C , and ln ( TG ) excluding and including participants on lipid lowering medication with the latter adjusted for medication usage using average effects estimated in Wu et al [56] for specific drug classes . Figure S7 suggests that both the point estimates and the confidence intervals of the genetic effects are similar for this female-only study whether participants are excluded or included and adjusted for medication use . We also performed a second sensitivity analysis: tests of association excluding participants on lipid lowering medication for all models . As detailed in Figures S8 , S9 , S10 , excluding participants on lipid lowering medication usage does not appreciably alter the results , with the possible exception of LDL-C associations in Japanese/East Asians . More specifically , two SNPs ( rs11206510 and rs1501908 ) became significantly associated with LDL-C after excluding participants on medications while two other SNPs ( rs562338 and rs6544713 ) were no longer significantly associated ( Figure S9 ) . The difference in significance for these four tests of association may be related to lipid lowering medication use; however , it is more likely due to statistical fluctuations from small samples sizes ( nInclude = 690; nExclude = 467 ) . Also of note , use of lipid-lowering medications was low ( <10% ) in the ARIC , CHS , NHANES , and WHI studies since the majority of study recruitment occurred before the introduction or widespread use of the recent generation of lipid-lowering medications . Medication use was higher in the MEC study ( 20–38% depending on the population ) , which contributed the majority of Japanese/East Asian samples .
Perhaps not unexpectedly , we were able to replicate most reported associations in European Americans . Regardless of significance , all but one of the tested SNPs had effect estimates in the same direction as the previously reported association from the literature . FADS1 rs174547 , which was significantly associated with decreased ln ( TG ) in this meta-analysis for European Americans , was associated with increased TG in European Americans from the Framingham Heart Study ( n = 7 , 423 ) [43] . HDL-C had proportionally ( 15% ) the greatest number of SNPs that failed to replicate in European Americans compared with LDL-C ( 5% ) and TG ( 0% ) despite the fact that we had sufficient power to detect the reported genetic effect size for many of these tests . TTC39B rs471364 was not associated with HDL-C levels despite a sample size of 18 , 089 and >99% power to detect the reported effect size . Neither ABCA1 rs4149268 nor rs1883025 was associated with HDL-C , although the latter test of association was underpowered ( 68%; n = 3 , 865 ) . Finally , as previously discussed , CETP rs1864163 was not associated with HDL-C in this European American dataset although we had 80% power to detect the reported genetic effect size . For LDL-C , only MAFB rs6102059 was not associated despite >90% power to detect the reported effect size . The reasons for non-replication in this European American dataset for properly powered tests of association are unclear . It is possible that we have overestimated our power to detect reported associations . The “winner's curse” and inflated genetic effect estimates from initial discovery are well known [57] , [58] . Indeed , for the five SNPs that did not replicate in this meta-analysis for European Americans , the association was described in only one GWAS each despite the fact that numerous GWAS [31] , [33]–[43] and a large meta-analysis [32] for these three traits have been conducted in populations of European-descent . The meta-analysis recently reported by Teslovich et al [32] did report significant associations between TTC39B rs581080 for HDL-C and MAFB rs2902940 for LDL-C . TTC39B rs581080 is in moderate linkage disequilibrium ( LD ) with rs471364 ( r2 = 0 . 49 in CEU HapMap ) , but MAFB rs2902940 is not in LD with rs6102059 ( r2 = 0 . 03 in HapMap CEU ) . A second possibility for our observed non-replication is heterogeneity among the PAGE studies . Because it is important to understand the degree to which associations are consistent across individual studies , we compared directions of effect ( betas ) across PAGE study sites for each test of association ( Figures S11 , S12 , S13 ) and performed tests of heterogeneity . Association results for TTC39B rs471364 , which meta-analysis result for HDL-C in European Americans was insignificant , had significant evidence for heterogeneity across studies ( pheterogeneity = 0 . 048; I2 = 58 . 25% ) . In four of the five PAGE study sites , the association between this SNP and HDL-C had consistent directions of effect; however , only one test of association was significant in European Americans ( p = 0 . 005 in EAGLE; Figure S11 ) . Only two other association results had evidence for heterogeneity among European Americans: FADS1 rs174547 for HDL-C ( pheterogeneity = 0 . 006; I2 = 75 . 73% ) and PCSK9 rs11206510 for LDL-C ( pheterogeneity = 0 . 048; I2 = 55 . 34% ) . However , for both of these loci , the tests of association were significant in European Americans and had similar directions of effect in all but one of the PAGE study sites ( Figures S11 and S12 ) . When taking into account power , significance , and direction of effect , most SNPs discovered in European Americans generalized to African Americans , Mexican Americans , and American Indians . Of note are the eleven tests of association significant in European Americans that did not generalize to African Americans despite having adequate power . Given that GWAS products are a mixture of tagSNPs and functional SNPs , it is likely that discovery in European Americans represents tagSNPs rather than the true functional SNP . Because linkage disequilibrium patterns differ across populations , tagSNPs genotyped directly in populations of non-European descent may not recapitulate the association observed in European-descent populations depending on the pattern of LD . The association of HDL-C and nonsynonymous rs3135506 versus tagSNPs rs28927680 in the APOA1/C3/A4/A5gene cluster in this analysis is an example of the effects of LD and the ability to generalize across populations . Evoking LD as an explanation for lack of generalization is appealing , but it does have limitations given that the functional SNP is not often obvious . All tests of association that did not generalize to African Americans had evidence of LD differences between CEU and YRI using the HapMap data . However , most of these SNPs are located in the intergenic and intronic regions . Further fine-mapping in both the discovery population as well as other diverse populations will be needed along with a better understanding of genetic variation and its relationship to biological function to identify the true functional SNPs for these traits . Among the five putative functional SNPs genotyped ( nonsynonymous rs11591147 , rs1260326 , rs3135506 , and rs1800961 and nonsense rs328 ) , all five replicated in populations of European-descent , and three of the five generalized to populations of non-European descent . One putative functional SNP that did not replicate across populations was HNF4A rs1800961 , likely due to low power because of the very low minor allele frequency in all subpopulations ( 0 . 0065 to 0 . 0398 ) . Both the direction and magnitude of effect , however , were consistent across groups . GCKR rs1260326 did not generalize to all populations of non-European descent but did generalize in three of the four populations tested and trended towards significance in American Indians ( p = 0 . 085; Table 4 ) . The major strengths and limitations of the PAGE study for lipids are sample size and diversity . The largest sample size is for samples of European-descent ( ∼20 , 000 ) , followed by African Americans and American Indians . The sample sizes for Mexican Americans , Japanese/East Asians , and Pacific Islanders/Native Hawaiians are smaller and consequently underpowered for tests of association as estimated from genetic effect sizes in the published European-descent discovery studies . Also , not all SNPs were genotyped in all PAGE studies , further affecting the power of the meta-analyses . An additional limitation is the lack of data related to lipid lowering medication . Ideally , all analyses would be adjusted for use of lipid lowering medication based on the type and dose of medication . In most PAGE studies , these data were not available and in many , use was low at baseline when blood samples were obtained . As we demonstrate in Supplementary material , inclusion of participants using lipid-lowering medication did not appreciably alter the results of the meta-analysis when compared with excluding these participants . While this finding may be useful for future studies , we caution that the majority of participants in this study were not on lipid lowering medications . In general , the cohorts and surveys included in PAGE are diverse with regard to demographics , genetic ancestry , lifestyle , health , and environmental exposure . Despite this diversity , very few tests of association from the meta-analysis exhibited evidence of heterogeneity . Overall , the majority of GWAS-identified SNPs for HDL-C , LDL-C , and TG replicated in European Americans and generalized to non-European-descent populations . These results suggest that the genotyped SNP either tags the functional SNP ( s ) common across these populations or that the genotyped SNP represents the risk SNP directly . SNPs that replicated in European Americans but did not generalize in the largest non-European-descent populations , despite adequate power , could represent priority associations that require fine-mapping and re-sequencing to identify the functional variant ( s ) .
All studies were approved by Institutional Review Boards at their respective sites ( details are given in Text S1 ) . PAGE study samples were drawn from four large population-based studies or consortia: EAGLE ( Epidemiologic Architecture for Genes Linked to Environment ) , based on three National Health and Nutrition Examination Surveys ( NHANES ) [59]–[61] , the Multiethnic Cohort ( MEC ) [62] , the Women's Health Initiative ( WHI ) [63] , [64] , and Causal Variants Across the Life Course ( CALiCo ) , a consortium of several cohort studies: Atherosclerosis Risk in Communities Study ( ARIC ) [65] , Coronary Artery Risk in Young Adults ( CARDIA ) [66] , Cardiovascular Health Study ( CHS ) [67] , Strong Heart Family Study ( SHFS ) [68] , and Strong Heart Cohort Study ( SHS ) [69] ( Table 1 ) . The PAGE study design is detailed in Matise et al [30] . Serum HDL-C , triglycerides , and total cholesterol were measured using standard enzymatic methods . LDL-C was calculated using the Friedewald equation [30] , [70] , with missing values assigned for samples with triglyceride levels greater than 400 mg/dl . For PAGE study sites with longitudinal data , the baseline measurement was used for analysis . A full description of each study , along with population-specific study characteristics , is presented in Text S1 and Table S1 . All SNPs considered for genotyping were previously associated with HDL-C , LDL-C , and/or triglycerides in published ( as of 2008 ) candidate gene and genome-wide association studies . A total of 52 SNPs were targeted for genotyping by two or more PAGE study sites . There is no overlap between samples used in this study and samples used in GWAS from which the SNPs were selected . The 52 targeted variants are located in or nearby 32 different genes/gene regions , with 12 of the gene/gene regions represented by two or more SNPs . Five SNPs are nonsynonymous , one SNP is a nonsense variant , and two SNPs are synonymous; the remainder are located in introns , flanking , or intergenic regions . The full list of targeted SNPs , their locations , and their previously associated lipid trait can be found in Table S2 . Cohorts and surveys were genotyped using either commercially available genotyping arrays ( Affymetrix 6 . 0 , Illumina 370CNV BeadChip ) , custom mid- and low-throughput assays ( TaqMan , Sequenom , Illumina GoldenGate or BeadXpress ) , or a combination thereof . Quality control was implemented at each study site independently . In addition to site-specific quality control , all PAGE study sites genotyped 360 DNA samples from the International HapMap Project and submitted these data to the PAGE Coordinating Center for concordance statistics [71] . Study specific genotyping details are described in Text S1 . Of the 52 targeted SNPs , three ( CETP rs1800775 , APOE rs429358 , and APOE rs7412 ) failed at all PAGE study sites that attempted genotyping; therefore , a total of 49 SNPs were tested in this analysis . All tests of association were performed by each PAGE study site using the same analysis protocol prior to meta-analysis . The study protocol excluded participants <18 years of age as well as non-fasting samples ( defined here as <8 hours ) . When triglyceride level was the dependent variable , participants with >1 , 000 mg/dl were excluded from analyses . Triglyceride ( TG ) levels were natural-log transformed ( ln ) prior to analysis . Linear regression was performed for fasting adults regardless of lipid lowering medication use with HDL-C , LDL-C , or ln ( TG ) as the dependent variable and a SNP as the independent variable , assuming an additive genetic model , stratified by race/ethnicity . The coded allele is reported in Table 2 , Table 3 , Table 4 . The beta estimate is per additional copy of the coded allele . For each SNP , four models were considered: 1 ) unadjusted , 2 ) adjusted for age ( continuous in years ) and sex , 3 ) adjusted for age , body mass index ( continuous in kg/m2 ) , current smoking ( yes/no; binary ) , type 2 diabetes ( yes/no; binary ) , post-menopausal status ( yes/no for females only; binary ) , and current hormone use ( yes/no for females only; binary ) , and 4 ) adjusted for age , body mass index , current smoking , type 2 diabetes , post-menopausal status , current hormone use , and previous myocardial infarction ( yes/no; binary ) . All PAGE study sites ( except for WHI , which is female only ) stratified models 3 and 4 by sex given the sex-specific variables ( post-menopausal status and hormone use ) prior to meta-analysis . Select PAGE study sites also included study site or site of ascertainment as a covariate in all models . Results from Model 2 ( adjusted for age and sex ) are reported in the main text while results from Models 1 , 3 , and 4 are presented in Figures S4 , S5 , S6 . Model 2 excluding participants on lipid-lowering medications are presented in Figures S8 , S9 , S10 . Meta-analyses , using a fixed-effects inverse-variance weighted approach and tests for effect size heterogeneity across studies , were performed using METAL [72] . P-values were not adjusted for multiple testing , and association results were plotted using Synthesis-View [73] , [74] , where indicated . Power calculations were performed using Quanto [75] , [76] assuming unrelated participants , an additive genetic model , the published effect size from European-descent populations listed in Table S1 , and the population-specific allele frequencies listed in Table 2 , Table 3 , Table 4 . Linkage disequilibrium was calculated using HapMap European ( CEU ) and West African ( YRI ) data accessed through the Genome Variation Server . FST was calculated using the Weir and Cockerham algorithm [77] . Aggregate data from the meta-analysis as well as individual tests of association from each PAGE study site will be made available via dbGaP [30] , [78] . NHGRI GWAS Catalog ( www . genome . gov/GWAStudies ) . Genome Variation Server ( pga . gs . washington . edu ) . Synthesis-View ( http://chgr . mc . vanderbilt . edu/ritchielab/method . php ? method=synthesisview ) . | Low-density lipoprotein cholesterol ( LDL-C ) , high-density lipoprotein cholesterol ( HDL-C ) , and triglyceride ( TG ) levels are well known independent risk factors for cardiovascular disease . Lipid-associated genetic variants are being discovered in genome-wide association studies ( GWAS ) in samples of European descent , but an insufficient amount of data exist in other populations . Therefore , there is a strong need to characterize the effect of these GWAS–identified variants in more diverse cohorts . In this study , we selected over forty genetic loci previously associated with lipid levels and tested for replication in a large European American cohort . We also investigated if the effect of these variants generalizes to non-European descent populations , including African Americans , American Indians , and Mexican Americans/Hispanics . A majority of these GWAS–identified associations replicated in our European American cohort . However , the ability of associations to generalize across other racial/ethnic populations varied greatly , indicating that some of these GWAS–identified variants may not be functional and are more likely to be in linkage disequilibrium with the functional variant ( s ) . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"genetic",
"polymorphism",
"genetic",
"association",
"studies",
"genetics",
"population",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"and",
"genomics"
] | 2011 | Genetic Determinants of Lipid Traits in Diverse Populations from the Population Architecture using Genomics and Epidemiology (PAGE) Study |
We describe the development and application of a suite of modular tools for high-resolution detection of proteins and intracellular protein complexes by electron microscopy ( EM ) . Conditionally stable GFP- and mCherry-binding nanobodies ( termed csGBP and csChBP , respectively ) are characterized using a cell-free expression and analysis system and subsequently fused to an ascorbate peroxidase ( APEX ) enzyme . Expression of these cassettes alongside fluorescently labelled proteins results in recruitment and stabilisation of APEX , whereas unbound APEX nanobodies are efficiently degraded by the proteasome . This greatly simplifies correlative analyses , enables detection of less-abundant proteins , and eliminates the need to balance expression levels between fluorescently labelled and APEX nanobody proteins . Furthermore , we demonstrate the application of this system to bimolecular complementation ( ‘EM split-fluorescent protein’ ) , for localisation of protein–protein interactions at the ultrastructural level .
Rapid and reliable protein localisation is critical for the functional characterisation of any protein of interest ( POI ) . Traditionally , this has been achieved through antibody-mediated methods or tagging with a fluorescent protein , such as GFP . The recent emergence of nanobodies ( small , single-domain antibodies amenable to cellular expression ) has allowed the development of new biotechnological tools based on the detection of epitopes in living cells [1 , 2] , although the availability of defined variable domains for antigen binding remains limiting . At the same time , the use of enzymatic tags such as the soybean ascorbate peroxidase ( APEX ) for ultrastructural detection of proteins provides an alternative to the use of traditional antibody labelling in electron microscopy ( EM ) [3] , with the advantage of protein localisation throughout the depth of whole cells or tissues making it compatible with the latest revolutionary 3D EM methods [4] . We have previously generated expression plasmids that encode a GFP-nanobody/binding peptide ( GBP ) for high-resolution detection of GFP-tagged proteins by electron microscopy . To achieve this , we genetically fused the GBP nanobody to the well-characterized soybean-derived enzyme APEX . When APEX–GBP is expressed in the presence of any GFP-tagged POI , its localisation can be determined by transmission EM following processing [5 , 6] . Here , we have developed and characterized a new suite of APEX/nanobody-mediated tools . As GFP and mCherry are the most broadly used fluorescent proteins in cell biology , we used cell expression to screen a library of putative mCherry-binding peptides ( ChBPs ) by single-molecule coincidence detection . We demonstrate the utility of a single mCherry nanobody for high-resolution , EM-based analysis of protein distribution and use this probe for correlative analyses . Furthermore , we generate conditionally stable ( cs ) nanobodies for both GFP and mCherry fused to APEX and show that degradation of unbound cs nanobodies by the proteasomal system reduces background APEX signals and results in an increased signal-to-noise ratio . Finally , we show that the new suite of APEX nanobody tools opens up entirely new avenues for EM localisation through the application of the csAPEX-nanobody system to bimolecular fluorescence complementation , allowing the detection and localisation of intracellular protein-protein interactions at the ultrastructural level .
To date , no modular systems exist to sensitively detect mCherry-tagged POIs to high-resolution for transmission electron microscopy . Therefore , we initially sought to generate a modular APEX-ChBP expression vector . We screened six sequences previously shown to have affinity for mCherry [7] by fluorescence cross-correlation spectroscopy in Leishmania tarentolae cell-free lysate [8] . Each peptide was first expressed fused to the open reading frame of GFP and assayed for self-association or cross-reactivity with GFP ( S1A–S1F Fig ) . ChBP1 and ChBP2 behaved as monomeric proteins ( S1A and S1B Fig ) , whereas ChBP3 , ChBP4 , ChBP6 , and ChBP8 demonstrated bursts of GFP signal above baseline monomeric protein behaviour ( S1C–S1F Fig ) , indicating a propensity for self-association . We next performed single-molecule coincidence detection after co-expression of mCherry-Caveolin1 ( Cav1 ) [9] . mCherry-Cav1 was selected as it generates stable , uniform , and membrane-associated oligomeric Cav1 , resulting in highly clustered mCherry tags within the confocal volume . Co-expression of GFP-tagged ChBP1 , ChBP3 , ChBP4 , and ChBP6 with mCherry-Cav1 did not result in significant coincidence between mCherry-Cav1 and GFP-tagged ChBP1 , suggesting that these peptides are inefficient at binding the mCherry tag in this context ( S1A and S1C–S1E Fig second and third panels ) . However , ChBP2-GFP and ChBP8-GFP demonstrated a considerable coincidence between the GFP-tagged ChBP and mCherry-Cav1 , with a coincidence ratio of Cherry to Cherry and GFP of approximately 0 . 5 , indicating a 1:1 binding ratio of GFP to Cherry ( S1B and S1F Fig second and third panel ) . We selected ChBP2 as the best-performing peptide in our analysis and incorporated this into our modular expression system ( Fig 1A; mammalian expression vector hitherto termed APEX-ChBP ) . To verify that this construct could be used for high-resolution EM , we co-transfected baby hamster kidney ( BHK ) cells with APEX-ChBP and three different subcellular markers: ( i ) mCherry to denote the cytoplasm , ( ii ) mCherry-Cavin1 to denote caveolae on the plasma membrane ( PM ) , and ( iii ) 2xFYVE-mCherry to denote early endosomes . Co-expression of the soluble APEX-ChBP and mCherry ( with subsequent DAB reaction in the presence of H2O2 and post-fixation with osmium tetroxide [OsO4] ) resulted in the accumulation of electron density in the cytoplasm of transfected cells ( Fig 1B ) . This observation closely mirrored the expression of GFP with APEX-GBP [6] . Cavin1 is a critical structural component of plasma membrane microdomains termed ‘caveolae’ and , when present at the PM , resides only within these domains [10] . When APEX-ChBP was co-transfected with mCherry-Cavin1 , the electron density generated by the APEX tag and the DAB reaction was restricted to the plasma membrane at structures with morphologies consistent with caveolae ( Fig 1C ) . Finally , we attempted to localize the phosphoinositide ( PI ) probe 2xFYVE-mCherry ( a marker of PI ( 3 ) P lipids ) , which are highly enriched within early endosomes [11] . Co-expression of 2xFYVE-mCherry and APEX-ChBP resulted in the specific accumulation of electron density surrounding structures consistent with early endosomal morphology ( Fig 1D ) . These data demonstrate that our APEX-ChBP vector can be used to localize mCherry-tagged proteins at ultrastructural resolution . As shown in Fig 1E–1H , use of the APEX-ChBP system is compatible with efficient correlative light and EM . Because the APEX2 probe is visible under both light and EM , this represents a simple alternative to more complex and currently widely used CLEM methods . The modular system for EM detection of fluorescently tagged POIs involves recruitment of APEX-tagged binding peptides to the fluorescent protein ( FP ) . Any unbound APEX nanobody will produce a diffuse cytosolic pool that will hinder detection of the POI and reduce the signal-to-noise ratio , particularly for low-abundance antigens . Recent work using the GBP nanobody has shown that manipulation of specific conserved residues produces a cs protein that is rapidly degraded by the proteasomal system in the unbound state [2] . We used this knowledge to generate csAPEX-GBP; ( schematically depicted in Fig 2A ) and introduced the analogous residue changes to APEX-ChBP ( generating csAPEX-ChBP ) . Expression of csAPEX-GBP in cells lacking GFP co-expression resulted in only negligible cytosolic APEX signal ( Fig 2B ) ; however , in a small number of cells , restricted electron density was observed in a punctate distribution ( Fig 2B inset ) . We hypothesise that this signal represents the residual expression of APEX-GBP in the process of proteasomal degradation . In contrast , co-expression of GFP produced a strong cytosolic signal ( Fig 2C , quantitated in S2A Fig ) and a complete loss of the punctate distribution observed in the csAPEX-GBP alone . The csAPEX-GBP protein showed efficient recruitment to different cellular compartments , including the plasma membrane , endosomes , and caveolae , showing the functionality of the csAPEX-GBP construct for detection of any GFP-tagged protein ( Fig 2D–2F ) . Consistent results were obtained with csAPEX-ChBP- and mCherry-tagged markers ( Fig 2G–2J ) . To confirm efficient degradation of our new , conditionally stable csAPEX-ChBP via the proteasomal pathway , we used the well-established proteasome inhibitor MG132 [12] . Cells expressing csAPEX-ChBP alone showed negligible reaction product following the DAB reaction , whereas cells expressing both csAPEX-ChBP and cytoplasmic mCherry showed intense staining throughout ( S2B Fig ) . However , following a 5-h supplementation with 10 μM MG132 , cells expressing csAPEX-ChBP alone retained DAB staining in the cytoplasm , indicating that , under normal conditions , csAPEX-ChBP is degraded by the proteasome . Bimolecular fluorescence complementation ( BiFC ) is a technique for testing pairwise protein-protein interactions in fixed or living cells by genetically tagging candidates with different halves of a “split” fluorescent protein [13] . If these candidates attain sufficient proximity , the full length fluorescent protein is reconstituted , can fold and emit photons under excitation by a suitable wavelength of light . We hypothesised that by using the conditionally stable APEX nanobody system , we should be able to extend the resolution of bifluorescence complementation to the ultrastructural level . Indeed , the nanobody binding site in GFP ( and its variants ) straddles the split site in commonly used BiFC pairs [13 , 14] . Furthermore , folding is absolutely required for the GFP–nanobody interaction , such that recognition of the unfolded halves of the split protein by GBP is a theoretical impossibility . Using this technique , we were able to directly visualize interactions between Cavin1 and Cavin3 by EM using split mVenus , a YFP derivative recognised by the GBP . We transfected BHK cells with vectors encoding Cavin1 fused to the N-terminal fragment of mVenus , ( Cavin1-mVenus1–155 ) , Cavin3 fused to the C-terminal fragment ( Cavin3-mVenus156–239 ) , and our csAPEX-GBP construct , schematically represented in Fig 2K . Using this technique , we were able to delineate surface caveolae and putative endocytic caveolar carriers associated with intracellular compartments ( Fig 2L , further examples in S2C and S2D Fig ) . The reciprocal experiment , in which the N- and C-terminal fragments of mVenus were exchanged , showed similar results ( Fig 2M ) . Unusually high-expressing cells were occasionally visible , showing aggregation of intracellular Cavin recognised by csAPEX-GBP ( Fig 2N ) . Transfection with just one-half of the split YFP most commonly showed no cytoplasmic staining ( Fig 2O ) . However , inclusions of increased density were sometimes noted in these controls ( Fig 2P , further example in S2E Fig ) and were absent from untransfected samples . This staining was clearly distinguishable from the specific signal shown in Fig 2L and 2M , although the importance of such controls is emphasized , particularly since different cell types may contain different numbers of proteasomes . These results clearly demonstrate that protein–protein interactions can be effectively visualized using bimolecular fluorescence complementation at the ultrastructural level using csAPEX-GBP . In summary , we have utilised cell-free expression and single-molecule analysis to screen a number putative ChBP for association with mCherry-tagged Caveolin-1 . The single nanobody we identify is a selective , high-affinity binder of the mCherry tag , lacks detectable self-aggregation or cross-reactivity with GFP , can be linked to APEX for high-resolution analysis of mCherry-tagged proteins in cell culture systems , and is compatible with correlative light and EM . We have also employed conditional stabilisation of both GFP and mCherry binding nanobodies fused to APEX2 which results in the generation of an APEX reaction product only when bound to their target fluorescent proteins . By degrading unbound APEX-BP protein , this modification facilitates an improved signal-to-noise ratio and circumvents any potential oversaturation of the APEX-BP vector . Finally , we have coupled the csAPEX-GBP system with bimolecular fluorescence complementation . This now allows direct visualisation of intracellular protein–protein interactions at the ultrastructural level , far beyond the resolution of light microscopy . This system is immediately applicable ( without any new cloning steps ) to any system in which the fluorescent split GFP system has been used . Unlike labelling on sections , APEX methods are compatible with 3D EM methods [4] such as focused ion beam-scanning EM , serial blockface-scanning EM , and electron tomography and can be used in whole animal systems [5] . As cellular function depends not on single proteins but on protein–protein interactions , these methods will be a vital complement to dynamic light microscopic methods .
Single-molecule spectroscopy was performed as previously described [9] . Briefly , samples ( 20 μl ) were loaded into a custom-made silicone 192-well plate adhered to glass coverslips ( ProSciTech Australia ) . Samples were analysed with two lasers ( 488 nm and 561 nm ) using a Zeiss LSM710 microscope with a Conforcor3 module for single-molecule counting and a single 488-nm laser for aggregation analyses . The fluorescence emission was filtered with 505–540-nm band pass filter ( GFP ) and 580-nm long-pass filter ( mCherry ) . Measurements were taken with photon counts in the approximate range of 750–2 , 000 which corresponds to a GFP concentration of around 1–2 . 5 μg/ml . Three replicates were carried out for each construct pair , and consistent results were obtained for each . BHK cells were passaged in Dulbecco’s Modified Eagle Medium ( Gibco ) supplemented with 10% Fetal Bovine Serum and L-Glutamine . Cells were seeded onto 35-mm culture dishes ( TPP ) , transfected with Lipofectamine 3000 as per the manufacturer’s instructions and processed for EM 24 h later . For bimolecular fluorescence complementation experiments , an 8-h incubation in 50 μM cyclohexamide prior to processing was used to reduce background staining . EM was performed exactly as described previously [5 , 6] . Briefly , cells were fixed with 2 . 5% glutaraldehyde in 0 . 1-M sodium cacodylate buffer for 1 h at room temperature . Cells were washed with cacodylate buffer to remove the fixative , then washed with DAB in cacodylate buffer for 1 min and subsequently treated with DAB in cacodylate buffer containing H2O2 for 30 min at room temperature . Cells were post-fixed with 1% OsO4 for 2 min to provide contrast . Cells were then washed in water and serially dehydrated in increasing percentages of ethanol before serial infiltration with LX112 resin in a BioWave microwave ( Pelco ) . Resin was polymerised to hardness at 60°C overnight . Ultrathin sections were cut on an ultramicrotome ( UC6: Leica ) and imaged at 80 kV on a JEOL1011 transmission electron microscope . Sections were not post-stained . Cells were grown on 35-mm gridded MatTek dishes ( with an in-plane alphanumeric code ) and co-transfected with nls-mCherry and APEX-ChBP . Live cell imaging was performed on an EVOS FL epifluorescent microscope ( ThermoFisher Scientific ) at 10x and 20x magnification . Cells were processed as described above with the following exceptions . Post-polymerisation , the flat-embedded cells were removed from the dish and the region of interest was trimmed using the now-imprinted grid coordinates on the block face . Ultrathin sections were cut , placed on a slot grid , and imaged on a Tecnai 12 transmission electron microscope fitted with a 4K x 4K LC1100 camera ( Direct Electron ) at 120 kV under the control of SerialEM . Low-magnification ( 4 , 400 XMag ) montages were acquired at a binning of 1 and stitched together using the Blendmont program in IMOD . Correlation of light and EM images was performed using Photoshop ( Adobe Inc . ) . Split mVenus constructs were made by first removing the Fos and Jun inserts from pcs_kmVenus1-155_FosLZ135-171 and pcs_kmVenus156-239_JunLZ253-289 using EcoRV/SpeI . Human Cavin1 and Cavin3 open reading frames were amplified by PCR using the primer tags forward 5′-AGCGGCGGCGGCTCTGATATC-3′ and reverse 5′-ACAAGAAAGCTGGGTACTAGT-3′ and subcloned using infusion ( BD ) . The series of ChBP-GFP expression vectors for L . tarentolae expression were constructed by PCR subcloning from the original templates [7] into the cell-free gateway cloning vector ‘N-term 8xHis eGFP pCellFree_G03’ [8] ( Genbank KJ541667 ) using the following primer tags: forward 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTC-3′ , reverse 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTT-3′ . Previously described vectors used for expression or subcloning were pmCherry-N1 ( Clontech PT3974-5 ) , pEGFP-N1 ( Clontech PT3027-5 ) , GFP-CaaX ( Kras ) [15] , GFP-2xFYVEhrs [16] , mCherry-2xFYVEhrs [17] , Cavin1-mCherry [10] , Cavin2-GFP and Cavin3-GFP [18] , pCSDEST2 [19] , pDEST-Tol2-pA2 , p5E-CMV/SP6 , pME-mCherry-CaaX ( Hras ) and p3E-pA [20] , APEX2-GBP , mKate2-P2A-APEX2-GBP , and pME-APEX2-NS [6] . All other constructs were made using the Multisite Gateway system ( Invitrogen ) . These new vectors have been deposited in the Addgene repository with the following identifiers: APEX2-csGBP ( 108874 ) , mKate2-P2A-APEX2-csGBP ( 108875 ) , APEX2-csChBP ( 108876 ) , EGFP-P2A-APEX2-csChBP ( 108877 ) , APEX2-ChBP ( 108878 ) , EGFP-P2A-APEX2-ChBP ( 108879 ) , H2B-mCherry ( 108880 ) , nls-mCherry ( 108881 ) , pME-nls ( 108882 ) , pME-H2B ( 108883 ) , p3E-mCherry ( 108884 ) , pME-mCherry-NS ( 108885 ) , mCherry-CaaX ( Hras ) ( 108886 ) , mVenusN-Cavin1 ( 108887 ) , mVenusC-Cavin1 ( 108888 ) , mVenusN-Cavin3 ( 108889 ) , mVenusC-Cavin3 ( 108890 ) , p3E-csGBP ( 108891 ) , p3E-ChBP ( 108892 ) , p3E-csChBP ( 108893 ) , p3E-APEX2 ( 108894 ) , pME-EGFP-P2A-APEX2-NS ( 108895 ) , and p3E-APEX2-P2A-EGFP ( 108896 ) . | The use of enzymatic tags such as the ascorbate peroxidase ( APEX ) for electron microscopic detection of proteins is changing electron microscopy ( EM ) in the same way that the use of GFP and related proteins caused a revolution in light microscopy . We previously developed expression plasmids encoding GFP-binding peptide ( or nanobody ) fused to APEX , which allows EM localisation of GFP-tagged proteins in vivo . Here , we have generated conditionally stable GFP- and mCherry-binding nanobodies fused to APEX . Using co-transfection of these APEX nanobodies with fluorescent-tagged constructs , we recruit APEX and detect the tagged proteins by electron microscopy . As unbound conditionally stable nanobodies are efficiently degraded by the proteasome , the signal to noise ratio is dramatically reduced . This enables detection of less abundant proteins and eliminates the need to balance expression levels between fluorescent-labelled and APEX nanobody constructs . Furthermore , and perhaps most exciting , is our application of this method to bimolecular fluorescence complementation—in which two tagged proteins interact—allowing the detection and localisation of protein-protein interactions in EM . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] | [
"methods",
"and",
"resources",
"protein",
"interactions",
"condensed",
"matter",
"physics",
"cytoplasmic",
"staining",
"membrane",
"proteins",
"microscopy",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",... | 2018 | Ultrastructural localisation of protein interactions using conditionally stable nanobodies |
In simple organisms like E . coli , the metabolic response to an external perturbation passes through a transient phase in which the activation of a number of latent pathways can guarantee survival at the expenses of growth . Growth is gradually recovered as the organism adapts to the new condition . This adaptation can be modeled as a process of repeated metabolic adjustments obtained through the resilencings of the non-essential metabolic reactions , using growth rate as selection probability for the phenotypes obtained . The resulting metabolic adaptation process tends naturally to steer the metabolic fluxes towards high growth phenotypes . Quite remarkably , when applied to the central carbon metabolism of E . coli , it follows that nearly all flux distributions converge to the flux vector representing optimal growth , i . e . , the solution of the biomass optimization problem turns out to be the dominant attractor of the metabolic adaptation process .
Constraint-based computational methods such as Flux Balance Analysis ( FBA ) are nowadays widely used when investigating metabolism of bacteria and other simple unicellular organisms [1 , 2] . Within the framework of FBA , a commonly accepted hypothesis is that biomass production has a special role: evolution has shaped cellular metabolism of these organisms so as to optimize growth , hence if growth is used as objective function of an optimization problem , the vector of fluxes found in correspondence of the optimum represents a plausible flux distribution for the organism . Although such a criterion is phenomenological , it is reasonable , and indeed the fluxes constructed by FBA methods describe well the empirical fluxes observed in many experimental situations , dealing with wild type organisms [3] , knockout mutants [4] , engineered strains , screenings of drugs [5] , nutrient shifts [6] or stress responses . For bacteria like E . coli , the short-term metabolic response to genetic and environmental perturbations is characterized by a growth arrest and by the activation of a number of latent pathways , a strategy which can favor the survival of the organism at the expenses of efficient biomass production [4 , 6 , 7] . This activation is however only transient , and most latent reactions become resilenced as the microorganism adapts to the new condition [7–9] . Although experimental data describing this adaptation process at metabolic , genomic , gene expression and proteomic level are starting to appear [6–12] , it is still unclear how this dynamical recovery is implemented by the organism . From the experimental data one can deduce for instance that the FBA criterion is inadequate to describe the transient response , but that it can still be used to characterize the end-point of the metabolic adaptation [4 , 6] . The two main criteria proposed in the literature to describe the metabolic response to a perturbation are MOMA ( Minimization Of Metabolic Adjustment [13] ) and ROOM ( Regulatory On/Off Minimization [14] ) . Both capture the idea that metabolism tends to minimize the adjustment with respect to the pre-perturbation fluxes at the expenses of growth , and for both criteria this results in a number of non-essential reactions being activated , which is coherent with the aforementioned experimental evidence [7 , 8] . However , these methods can provide only a static snapshot of the early adjustments that follow a perturbation . Attempts to model the dynamical changes happening during adaptation have been made for example using kinetic models [15] or combining pseudo steady-states of FBA with kinetic models as in dynamical FBA [16] , see [17] for an overview . Other types of proposals include the incorporation of extra time-dependent constraints in the model , representing for instance molecular crowding [18] or other growth-limiting factors [19] . An alternative to adding kinetic parameters or constraints is to combine multiple datasets , such as gene expression [10 , 20] and/or proteomic [12] , see [21 , 22] for reviews . The transcriptional or translational information obtained in this way can be used to tune the constraints of an FBA model , leading to improved matches with empirically observed fluxes [12 , 20] . None of these methods is however able to provide a systematic interpretation of how and why the organism accomplishes the adaptation , let alone to propose a mathematical principle combining adaptation and FBA . A possible way to obtain a dynamical description of metabolic adaptation is proposed in [23] . Starting from a non-adapted progenitor metabolism , a population of phenotypes is obtained through the resilencing of a single reaction ( i . e . , letting the corresponding flux become negligible ) . If the growth rate of the different phenotypes is taken as measure of fitness , then a selection biased towards the fittest phenotypes favors the recovery of growth , see Fig 1 . If the procedure is iterated , then a Markov chain is obtained . Since at each step of the chain the metabolism of the selected phenotype differs from its predecessor only for a single silenced reaction , it can be seen as a short-term adjustment and computed through a MOMA . For the central carbon metabolism , the resulting process of iterated metabolic adjustments leads rapidly to metabolic adaptation of the microorganism . It is shown in [24] that this model can be used to describe a series of experimental results dealing with adaptation to single carbon sources of various E . coli knockout strains [6] . In particular , it allows to achieve a good agreement with both the experimental growth rates and measured flux data reported in [6 , 8] . The aim of this paper is to take the approach one step further , by showing that for the core metabolism of E . coli , the Markov chains of recursive resilencings constructed in this way exhibit a single dominant end-point flux vector , and that this vector corresponds to one of the FBA optima , namely the parsimonious enzyme usage FBA optimum ( pFBA , minimizing the number of fluxes [12] ) . To do so , we compute a large number of trajectories for our Markov chains from random initial conditions , and show that they tend to become absorbed into the pFBA flux distribution or at least to become highly correlated with it . More specifically , the chain of pseudo steady-state fluxes computed through the adjustments that follow the resilencings steers the vast majority of all admissible flux vectors towards alignment with the pFBA vector , regardless of the norm ( and growth rate ) achieved by the flux vectors at the end-point of the process . In dynamical systems terms , we can say that the pFBA flux vector constitutes the dominant attractor of the fitness landscape associated to the process of metabolic adaptation . The fact that the single dominant peak of this landscape corresponds to the pFBA flux distribution sheds a novel perspective on FBA optimization , and may contribute to turning this phenomenological argument into a rigorous mathematical model .
In FBA [2] , the polytope of admissible steady state metabolic fluxes is represented by Γ = { v : S v = 0 , l⩽v⩽u } , where v is the vector of fluxes , of lower and upper bounds l = [ℓ1 … ℓn] and u = [u1 … un] , and S is the stoichiometric matrix . The FBA optimal flux vector is given by v FBA = arg max v ∈ Γ ξ T v , ( 1 ) where g = ξT v is the growth rate , i . e . , the linear combination of fluxes that constitutes the biomass reaction . When the optimum is degenerate , a secondary optimization criterion can be used to discriminate among equivalent optimal solutions . For example , overall enzyme investment is minimized by the pFBA solution vpFBA , which corresponds to minimization of the sum of the ( absolute values of the ) fluxes [12] . Following [23] , we assume that the adaptation dynamics form a stochastic process of recursive resilencings described by the Markov chain 𝓢k = {vk , Γk}k = 0 , 1 , 2 , … , where v0 is a randomly chosen initial condition in Γ0 = Γ . The stochastic process can be summarized as follows . At step k , assume the population of bacteria has an homogeneous metabolism , i . e . , all cells have the same nk active reactions with the same fluxes vk−1 ( this corresponds to a specific sampling of our Markov chains ) . From vk−1 , it is possible to obtain nk + 1 different phenotypes , corresponding to the resilencing of one of the enzymes ( nk possibilities ) or to the current phenotype remaining unchanged for another step . The nk possible silencings of a reaction yield the nk reduced polytopes Γk , i = Γk−1 ∩ {ℓi = ui = 0} , i = 1 , … , nk . The corresponding fluxes vk , i are computed via a MOMA projection on these reduced polytopes: v k , i = arg min v ∈ Γ k , i ‖ v - v k - 1 ‖ , i = 1 , … , n k , k = 1 , 2 , … where ‖ ⋅ ‖ is the Euclidean norm , see Fig 1 for a sketch . Each choice of vk , i leads to a possible growth rate: gk , i = ξT vk , i , i = 1 , … , nk . Viable phenotypes have gk , i > 0 while non-viable phenotypes ( e . g . when an essential reaction is suppressed ) have gk , i = 0 . In what follows these growth rates will be placed on the diagonal of a fitness matrix G k = [ g k , 0 g k , 1 ⋱ g k , n k ] where gk , 0 represents the current growth rate . To the nk + 1 possible choices gk , i , it is possible to associate selection probabilities through a basic replicator equation which uses the gk , i as fitness function . Denote pk , i , i = 0 , 1 , … , nk , the probabilities ( or frequencies ) associated to the gk , i . If Δt is the time duration of each step , then the replicator equation is p ˙ k = G k p k - ϕ ( p k ) p k τ ∈ [ 0 , Δ t ] , ( 2 ) where p k = [ p k , 0 p k , 1 ⋮ p k , n k ] and ϕ ( p k ) = ∑ i = 0 n k g k , i p k , i is the average fitness . The explicit solution of Eq ( 2 ) can be expressed as a Boltzmann distribution , see S1 Text for the details . In synthesis , in the two cases we can distinguish ( sketched in Fig . B of S1 Text ) one gets for the selection probabilities: uniform priors: at the begin of the time interval the selection probability is pk ( 0 ) = 1/ ( nk + 1 ) , where 1 = [ 1 … 1 ] T , i . e . , all choices are equiprobable . In this case the Boltzmann distribution for the selection probabilities at the end of the time interval is p k ( Δ t ) = 1 Z k ( Δ t ) e G k Δ t 1 where Zk ( Δt ) = ∑i=0nkegk , iΔt is a partition function . non-uniform priors: at the begin of the time interval the selection frequencies are not uniform but are themselves expressible as a Boltzmann distribution p k ( 0 ) = 1 Z k ( β k ) e G k β k 1 where βk has the interpretation of an inverse temperature . In this case , at the end of the time interval we obtain p k ( Δ t ) = 1 Z k ( β k + Δ t ) e G k ( β k + Δ t ) 1 . A through derivation of these selection probabilities is available in the S1 Text . In both cases described above pk ( Δt ) has the meaning of transition probability between the current state 𝓢k−1 = {vk−1 , Γk−1} and the possible states achievable at the k-th step 𝓢k , i = {vk , i , Γk , i} , i . e . , p k , i = ℙ ( X k = 𝒮 k , i ∣ X k − 1 = 𝒮 k − 1 ) , i = 0 , 1 , … , nk . Since the fluxes vk , i can take any value between lower and upper bound , the corresponding transition matrix is infinite dimensional . However , in order to understand the properties of the stochastic process we are considering , it is useful to look at its projection over the subspace of active reactions ( i . e . , over the binary equivalent of the polytope Γk ) . In terms of this projection , the possible states of the Markov chain are the 2r possible combinations of the r reactions of the network , see Fig 2 for a toy example with r = 4 . Denote 𝓩1 , … , 𝓩2r these discrete states and Pij = ℙ ( Xk = 𝓩i ∣ Xk−1 = 𝓩j ) the corresponding transition probabilities . As in our model the resilencings are irreversible , P is triangular , i . e . , it is completely reducible , see Fig 2D . Since in reality P is the result of a projection , it is P = P ( v ) , i . e . , the exact transition probabilities pk depend on the values of the fluxes and hence on v0 . However , even in the complete model the fully reducible structure is preserved . In particular it follows that a certain number of states 𝓩i must correspond to ergodic classes , i . e . , “absorbing” states in which the Markov chain stabilizes . Full reducibility implies that each ergodic class is composed of a single state . Periodic chains of states are impossible .
The metabolic adaptation process described in Methods and in Fig 1 is applied to the network that describes the central carbon metabolism of E . coli [25] . In order to do this , a large number of realizations of the Markov chain is produced . Even in a network of modest dimensions like that of E . coli central metabolism ( r = 95 reactions , see S1 Text for details ) , the number of possible discrete states 𝓩 of the Markov chain is enormous ( 295 ∼ 1028 ) , hence numerical computations are necessarily limited to a fraction of all possible trajectories . For this study , some 105 trajectories have been generated , starting from randomly chosen initial conditions in the polytope of admissible fluxes Γ and using various forms for the priors . Given the irreversibility of the resilencing , the number of steps required to reach an end-point state can be computed from the trajectories . A trajectory is considered absorbed into an end-point state ( i . e . , it has reached a local maximum of the fitness landscape ) when no further silencing happens for 5 consecutive steps . With this stopping condition , the expected time until absorption of our trajectories is 24 . 6 ± 4 . 1 steps . Except for pathological initial conditions ( those missing some essential reactions ) , all trajectories stabilize to flux vectors of positive growth . Nearly 20% of all trajectories reach the maximal growth computed by the FBA criterion , gFBA , and for nearly 50% of all trajectories the growth at the end-point is ≥ 0 . 85 gFBA , see Fig 3 . The remaining 50% of trajectories are more or less uniformly distributed in the interval 0 . 2 < g/gFBA < 0 . 85 . Much more remarkable is the correlation between the end-point flux distribution v and the flux distribution given by the pFBA criterion vpFBA: the mean of the correlation is 0 . 96 and the median is 0 . 98 , with 88% of all trajectories achieving a correlation of at least 0 . 9 , see Fig 3 . The meaning of this result is that nearly all initial conditions in Γ tend to become aligned with the flux distribution vpFBA , regardless of the biomass they can produce , see Fig . A of S1 Text for a sketch . The time evolution of the 3D histogram of Fig 3 during adaptation is shown in Fig . C of S1 Text . It can be seen that while randomly chosen initial conditions in Γ usually give a zero-growth , already with the first silencings growth starts to recover , and gradually improves in the first 10 steps of the Markov chain . During the transient , no significant intermediate peak is visible , meaning that many different routes are explored by the trajectories . After ∼ 10 steps , the high correlation / high growth peak starts to appear , and rapidly becomes dominant . Examples of the resulting trajectories are shown in Figs . D-F of S1 Text . For instance , the first row of Fig . D of S1 Text shows a set of trajectories originating from the same random initial condition , all converging towards vpFBA , although through slightly different paths . None of the trajectories of the second row of Fig . D of S1 Text instead achieves a growth rate higher than 0 . 75gFBA . However , all of the end-points flux vectors become aligned with vpFBA ( correlation higher than 0 . 97 ) . In this case the two values of g reached by the trajectories correspond to two different ergodic states , as can be seen by the grouping of the number of active reactions eventually reached . It should be observed how for this phenotype of non-optimal growth the number of reactions is much less than for vpFBA . This is indeed a constant pattern in our metabolic adaptation strategy . As can be seen in Fig . G of S1 Text , at the end-point the number R of active reactions of v and g are positively correlated: for strains that have sub-optimal growth more resilencings are possible i . e . , more directions with slow but positive Δg exist and are explored . In fact , Fig . G of S1 Text shows that indeed the length L of a trajectory is inversely correlated with the growth g of v . Other than the peak at high correlation / high growth , Fig 3 does not show any other sufficiently significant peak ( and nor does Fig . C of S1 Text ) . It is however worthwhile observing that a small fraction of trajectories is steered towards flux distributions of maximal growth different from vpFBA , i . e . , to alternative FBA optima . An example of such trajectory is shown in Fig . E of S1 Text ( bottom row ) : while most of the trajectories converge to vpFBA , a few do not ( one is shown in green ) , and stabilize in an alternative FBA flux vector of correlation 0 . 75 with vpFBA . Cases like this lead to a correlation corr ( vpFBA , vk ) which decreases when vk falls into the basin of attraction of a local maximum other than vpFBA . In the ensemble of the trajectories , however , these situations are unfrequent: if we look at the average of all trajectories , corr ( vpFBA , vk ) is always monotonically increasing , regardless of the final g achieved , see Fig 3 and Fig . H of S1 Text . Similarly , also g is monotonically growing on the vast majority of the trajectories ( Fig . I of S1 Text ) . Interestingly , if we start relaxing the assumption of irreversibility that characterizes a large fraction of the metabolic reactions ( 49 of the 95 reaction are irreversible in our network ) , then the convergence rate to vpFBA quickly decreases , in favor of other vFBA , see Figs . J-L of S1 Text . In particular , when all reactions are considered as reversible , then the correlation between vpFBA and v at absorption is completely lost , although optimal biomass is still achieved by most trajectories , see Fig . L of S1 Text . It follows directly from the linearity of the expression for the biomass that when a trajectory v becomes aligned with vpFBA , the growth rate it can produce depends on the norm of v . In fact , Fig 4 shows that there is a sharp direct proportionality between ‖v‖ at ergodicity and g: when the recursive process of silencings and adjustments leads to a v which is smaller in norm than vpFBA , also the corresponding g will be smaller than gFBA . In order to understand when an initial condition can lead to an end-point v of norm comparable to vpFBA , one can look at how many of the bounds that delimit the polytope of admissible fluxes at step k , Γk , become active during the adaptation ( i . e . , a flux for a reaction becomes equal to one of its lower or upper bounds ) . Fig . N of S1 Text shows that in the early part of the adaptation , trajectories that do not achieve high growth ( which , from Fig 4 , correspond to trajectories having ‖v‖ < ‖vpFBA‖ ) tend to saturate less than those achieving higher growth . Hence fluxes that tend to stay in the interior of the polytope rarely will reach a high ‖v‖ . From Fig . O of S1 Text it can be observed that the difference in active bounds concerns mostly certain specific pathways: in strains achieving high growth , uptake bounds on many exchange reactions tend to become saturated in the early transient , and so do upper bounds of pyruvate metabolism , signs of a more efficient use of the available resources . Coherently , Fig . P of S1 text says that high growth is achieved when TCA cycle and pentose phosphate pathway remain fully functional during adaptation . Notice that uptake bounds of gluconeogenic carbon sources such as acetate are almost never saturated in the high growth trajectories .
Mathematically , the dynamical model used in this paper to describe metabolic adaptation has many aspects in common with standard evolutionary models based on natural selection [26] . The only extra assumption we require is that the “selection” that leads to adaptation occurs only through the silencing of non-essential reactions . That a multitude of latent pathways becomes active after an environmental perturbation is a known fact experimentally [7 , 8] . That during adaptation these low-yield pathways tend to become resilenced is also a commonly accepted hypothesis [8 , 27] , supported for example by gene expression data . It has in fact been observed that e . g . after a change of carbon source a major rearrangement occurs at gene expression level , with more than 103 genes differentially expressed [8] . A similar pattern is observed also in response to a wide variety of stress factors [7] . After a strain has adapted to the new condition , however , most differentially expressed genes have returned to their baseline level , and so is probably the concentration of the corresponding enzymes . As shown in [7] , different stress responses elicit early metabolic responses that are less stereotypical than those observed at gene expression level . When growth is recovered , however , the metabolic profiles in the various cases show a high similarity . This picture is coherent with the presence of an attractor in flux space , which can compensate for possibly widely different flux distributions right after a perturbation . For metabolic responses such as the stress responses of [7] , it is unclear how to include the direct effect of the perturbation on the metabolic fluxes of an FBA model . To cope with this fact , in our Markov chains the initial condition for the flux vector , v0 , is chosen randomly in the polytope Γ , which implies that at the begin of the Markov chain most reactions are already active . When instead the effect of a specific perturbation can be explicitly included in the FBA model , then the Markov chains can be used to investigate also the early stages of the transient , with the activation of the latent pathways . This is the point of view taken in [24] , where the experimental setting of [8] is considered . It is shown in [24] that the shift from rich medium to single carbon sources for various E . coli mutants can be reproduced closely by the metabolic adaptation process described in the Methods section . Proceeding in this way corresponds to fixing specific initial conditions on the Markov chains , and following the specific family of trajectories that results from them ( activatory phase included ) . It becomes then interesting to see what happens when these “nominal” trajectories are compared to more general trajectories in which the initial fluxes v0 are randomly chosen in Γ . For glucose as single carbon source , the two types of trajectories are compared in Fig . M of S1 Text . As can be seen , for all 4 mutant strains there is a high correlation between the end-points achieved by the flux vectors , meaning that the specific pattern of transient activations of the latent pathways is not crucial to the achievement of the adapted flux distribution , as both types of trajectories converge towards vpFBA . Notice how the pgi mutant has a secondary peak at low correlation: this corresponds to a less frequent second phenotype of lower growth , described in [8] . Such a phenotype is sometimes achieved by both the nominal trajectories of [24] and the randomly initialized trajectories computed in this paper . A number of possible optimality criteria alternative to biomass optimization have been investigated in the literature [2 , 28–30] . Common choices are for example maximization of yield ( instead of biomass ) , maximization of ATP , minimization of overall intracellular flux ( i . e . , minimum enzyme investment ) , minimization of redox potential , etc . In [29] a thorough analysis of their coexistence/complementarity is carried out . By using reaction resilencing to progressively adjust the metabolism to the new environment , two of the most accepted among these criteria , biomass optimization and minimization of overall fluxes , are naturally combined . It is shown in [31] that in FBA irreversibility of a large fraction of metabolic reactions is a key factor in achieving optimal flux distributions that are sparse , as our pFBA is . Indeed also for our metabolic adaptation process irreversibility is key to convergence to vpFBA , as Figs . J-L of S1 Text clearly show . It is worth observing that when we start relaxing the assumption of irreversibility , what is lost is not the achievement of optimal growth , but only convergence to the sparsest degenerate solution of Eq ( 1 ) ( i . e . vpFBA ) . On the contrary , in the case of all reversible reactions the ratio g/gFBA achieved by the trajectories is even better than in Fig 3 , with a mean value for g of 0 . 91gFBA and a median value of 0 . 997gFBA , see Fig . L of S1 Text . Given that the irreversibility of the constraints follows from thermodynamic considerations [32] which are usually considered sufficiently reliable , our results provide novel evidence in favor of sparse optimal biomass solutions such as pFBA , and a novel point of view on the coexistence of optimality criteria such as biomass production and enzyme parsimony of the solution . They also confirm that the repeated resilencing process described in this paper is indeed an effective strategy for describing the recovery of growth that occurs in metabolic adaptation . It is worth remarking that the method used in this paper is fundamentally different from a dynamical FBA [16] . In the latter , in fact , growth is used as the objective function of an optimization problem , and the adjustments of the metabolic fluxes follow the gradient direction indicated by the solution of such a problem . In our case , instead , the growth rate is only used to shape the fitness landscape of a population of possible phenotypes ( corresponding to the possible silencings that can occur ) , but the metabolic adjustments are always computed through MOMA projections . In general , there is no a priori guarantee that a greedy fitness landscape constructed in this way i ) may be regular; ii ) may achieve maximal growth , and iii ) may lead to flux distributions that resemble those of the pFBA . In our trajectories , in fact , what we observe is that the fitness landscape is rugged , but the plethora of local maxima have all a very small basin of attraction , as opposed to the global maximum which attracts around 50% of all trajectories when we count based on growth . If instead we look at normalized flux distributions then the basin of attraction of v pFBA ∣ ∣ v pFBA ∣ ∣ grows to 90% of all v ∣ ∣ v ∣ ∣ . This tells us that for what concerns central metabolism , a procedure like the one described in this paper is substantially a monotonic process of alignement of vk on vpFBA . The robustness of the convergence is also reflected in the low sensitivity to the randomness of the Markov chains , see S1 Text and Fig . T of S1 Text for more details . In conclusion , one can say that simple flux reorganization rules based on local fitness are sufficient to drive the cell toward a more efficient use of the metabolic resources . It is quite remarkable that most of the trajectories end up in the pFBA optimum , without knowing it , and without ever using growth rate to update metabolic fluxes ( growth rate is used only for the selection probabilities in the resilencings; metabolic fluxes are always updated via MOMA ) . Clearly this fact provides a further evidence in favor of the FBA criterion , and one could even speculate that it provides a more fundamental principle , from which FBA follows as a corollary . | In modeling metabolic networks , concepts like biomass optimization are often used to determine flux distributions of simple organisms such as E . coli . Although they often give good results in practice , they normally rely on heuristic considerations like “evolution has tuned metabolic fluxes to optimize growth , hence optimizing growth gives reasonable fluxes” . The main result of this paper is to show that metabolic adaptation naturally leads to optimal growth , in the sense that the flux distribution associated to optimal growth is the dominant attractor of the fitness landscape of the metabolic adaptation process . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Metabolic Adaptation Processes That Converge to Optimal Biomass Flux Distributions |
The m-AAA protease preserves proteostasis of the inner mitochondrial membrane . It ensures a functional respiratory chain , by controlling the turnover of respiratory complex subunits and allowing mitochondrial translation , but other functions in mitochondria are conceivable . Mutations in genes encoding subunits of the m-AAA protease have been linked to various neurodegenerative diseases in humans , such as hereditary spastic paraplegia and spinocerebellar ataxia . While essential functions of the m-AAA protease for neuronal survival have been established , its role in adult glial cells remains enigmatic . Here , we show that deletion of the highly expressed subunit AFG3L2 in mature mouse oligodendrocytes provokes early-on mitochondrial fragmentation and swelling , as previously shown in neurons , but causes only late-onset motor defects and myelin abnormalities . In contrast , total ablation of the m-AAA protease , by deleting both Afg3l2 and its paralogue Afg3l1 , triggers progressive motor dysfunction and demyelination , owing to rapid oligodendrocyte cell death . Surprisingly , the mice showed premature hair greying , caused by progressive loss of melanoblasts that share a common developmental origin with Schwann cells and are targeted in our experiments . Thus , while both neurons and glial cells are dependant on the m-AAA protease for survival in vivo , complete ablation of the complex is necessary to trigger death of oligodendrocytes , hinting to cell-autonomous thresholds of vulnerability to m-AAA protease deficiency .
Oligodendrocytes are glial cells of the central nervous system ( CNS ) that produce myelin to enhance conduction velocity . Oligodendrocytes utilize high amount of energy to synthesize proteins and lipids to build up myelin [1] and depend on mitochondrial respiration heavily during differentiation and myelination , when they are dramatically susceptible to ischemia , energy deprivation , and oxidative stress [2 , 3] . It has been hypothesized that post-myelination oligodendrocytes can undergo a metabolic switch to glycolysis , and provide metabolic support to axons , by supplying lactate as an energy source [2 , 4] . In support of this hypothesis , oligodendrocyte-specific deletion of an essential assembly factor for complex IV , Cox10 , did not lead to axonal degeneration or demyelination [2] , strongly suggesting that these cells can survive a respiratory chain deficiency . Moreover , recent data suggest that oligodendrocyte mitochondria may be involved in specialized functions relevant for myelin maintenance , such as lipid synthesis , or fatty acid oxidation , rather than in ATP production [5] . The m-AAA protease is a large proteolytic complex in the inner mitochondrial membrane endowed with crucial and pleiotropic roles in mitochondria . It regulates the turnover of respiratory chain subunits [6–8] , controls ribosome assembly and thereby mitochondrial translation [9 , 10] , and affects mitochondrial dynamics [11] . In humans , the m-AAA protease is composed of two subunits , paraplegin and AFG3L2 , which form either homo-oligomeric ( AFG3L2 alone ) or hetero-oligomeric ( AFG3L2 and paraplegin ) hexameric functional complexes [12] . The mouse genome contains a third gene , encoding a functional m-AAA protease subunit , Afg3l1 , which can form either homo-oligomers or hetero-oligomers with AFG3L2 or paraplegin [12] . The discovery that both paraplegin and AFG3L2 are implicated in human neurodegenerative diseases has sparked increasing interest in the m-AAA protease . Recessive mutations in SPG7 , encoding paraplegin , lead to hereditary spastic paraplegia ( HSP ) [13] , a neurodegenerative disease affecting the long corticospinal motor axons , while dominant mutations in AFG3L2 cause spinocerebellar ataxia type 28 ( SCA28 ) [14] , associated with atrophy of the cerebellum . Moreover , a severe phenotype combining features of spastic paraplegia and ataxia associated with myoclonic epilepsy ( SPAX5 ) has been linked to a homozygous mutation in AFG3L2 [15] . A plethora of dysfunctional pathways have been unravelled in cells when the m-AAA protease is depleted , including reduced assembly of respiratory complexes [9 , 16 , 17] , COX deficiency , impaired mitochondrial translation , fragmentation of the mitochondrial network [9] , disturbance of mitochondrial anterograde transport [18] , and calcium dysregulation [19 , 20] . Neurons are extremely susceptible to decreased levels of the m-AAA protease , and cannot survive Afg3l2 deficiency [9] . The role of the m-AAA protease in glial cells is so far unknown . Here , we used an inducible Plp1-CreERT transgenic mouse line to delete Afg3l2 in a wild-type or Afg3l1-null background in adult oligodendrocytes . We found that AFG3L2 deficiency was tolerated by oligodendrocytes for a long time , but ultimately led to late-onset myelin abnormalities and axonal degeneration in the spinal cord . In contrast , deletion of both Afg3l2 and Afg3l1 , which completely abolishes the m-AAA protease , caused rapid cell death of targeted cells . Our study unravels a crucial role of the m-AAA protease in protection against cell death , independent from the metabolic profile of the cell , and demonstrates that different thresholds of m-AAA protease activity are required in neurons and glial cells .
AFG3L2 is highly expressed in the brain [12] , however its abundance in neuronal versus glial cells is unknown . We investigated the expression of subunits of the murine m-AAA protease in lysates from enriched neuronal cultures , purified astrocytes , and purified late oligodendrocyte progenitors by immunoblotting . Notably , we did not observe remarkable differences in the levels of AFG3L2 or SPG7 ( paraplegin ) in neurons versus glial cells ( Fig 1A ) . Minor differences in the expression levels correlated with the abundance of other mitochondrial markers such as SDHA ( Fig 1A ) . Our results are consistent with published quantitative proteomic data obtained from acutely isolated cell population in the brain [21] . These data raise the question of the functional role of the m-AAA protease in glial cells . To target oligodendrocytes , we utilized a well-established tamoxifen-inducible Plp1-CreERT transgenic line [22] . Sensitivity and specificity of the Cre expression was confirmed by crossing Plp1-CreERTwt/tg mice with a reporter transgenic line expressing a mitochondrially targeted YFP ( mt-YFP ) upon Cre recombination ( ROSA26+/SmY ) [23] . Tamoxifen was injected intraperitoneally for five consecutive days at P29 , a time point used in a previous study to induce Cox10 deletion using the same promoter [2] , and the corpus callosum analysed at P36 ( S1A Fig ) . mt-YFP expression was largely restricted to cells expressing both or either of the oligodendrocyte markers APC and Olig2 ( S1B and S1C Fig ) . We determined that approximately 50% of oligodendrocytes ( positives for either or both APC and Olig2 ) in the corpus callosum were targeted ( S1D Fig ) . To induce Afg3l2 deletion in oligodendrocytes , we crossed Plp1-CreERT mice with mice carrying a floxed allele of Afg3l2 [9] , and used the same protocol described above . We analysed Afg3l2fl/fl Plp1-Cretg/wt mice ( referred to as L2KO ) and compared them with Afg3l2fl/fl Plp1-Crewt/wt mice ( L2fl/fl ) , similarly treated with tamoxifen . L2KO mice did not show any apparent phenotype in the cage or weight loss compared to controls up to 90 weeks of age ( S2A and S2B Fig ) . However , at this old age , they displayed a mild but significant impairment in motor performance on an accelerating rotarod test ( Fig 1B ) . We then carefully analysed the brain and the spinal cord of L2KO mice to detect any sign of late-onset pathology . No obvious demyelination was detected up to 90 weeks of age , when we only observed the appearance of few bigger APC+ cells in the corpus callosum ( S2C Fig ) . Semithin sections of the spinal cord revealed no alterations at 56 weeks ( S2D Fig ) , but abnormal myelin profiles , characterized by myelin thickening and infoldings , and myelin whorls , indicative of axonal degeneration were visible at 90 weeks ( Fig 1C and 1D ) . Ultrastructural analysis of the spinal cord white matter disclosed a few axons characterized by thin myelin , and others showing adaxonal myelin detachment and vacuolization already at 56 weeks . These abnormalities were more prominent at 86 weeks , when degenerating axons surrounded by damaged myelin or containing accumulation of material were present ( Fig 1E ) . Oligodendrocytes contained enlarged mitochondria with disrupted cristae ( Fig 1E and S2E Fig ) , closely resembling those previously described in Afg3l2-deficient neurons [9 , 17] . Fragmentation of the mitochondrial network occurs at early time points after deletion of Afg3l2 in neurons [9] , and is caused by activation of the stress protease OMA1 , which in turn cleaves the dynamin-like GTPase OPA1 , leading to impaired mitochondrial fusion [11] . The functional role of this fragmentation in neurons is unclear , since they die shortly after showing this phenotype . To visualize mitochondrial morphology in targeted oligodendrocytes , we further crossed mice with the mt-YFP ( ROSA26+/SmY ) reporter line [23] . To exclude a toxic effect caused by mt-YFP and/or Cre expression , we used as controls mice haploinsufficient for Afg3l2 . At 8 weeks of age , oligodendrocytes in the corpus callosum of Afg3l2fl/+ Plp1-Crewt/tg ROSA26+/SmY mice had a tubular mitochondrial network , while in absence of Afg3l2 ( genotype: Afg3l2fl/fl Plp1-Cretg/wt ROSA26+/SmY ) mitochondria appeared swollen and fragmented ( Fig 1F and 1G ) , indicating that the residual m-AAA protease is not sufficient to prevent this stress response . The mt-YFP reporter also allowed us to trace the fate of targeted oligodendrocytes . In the corpus callosum , the total number of targeted mt-YFP+ cells and of APC+ oligodendrocytes was not significantly changed in absence of AFG3L2 compared to control mice at 56 weeks ( S2F Fig ) , in agreement with the lack of overt demyelination . However , while in control mice almost all mt-YFP+ cells were also APC+ , L2KO mice displayed a significantly increased number of mt-YFP+APC- cells ( S2G Fig ) . Thus , even though lack of Afg3l2 triggered early-onset pronounced mitochondrial morphology defects , oligodendrocytes survived for long time , and myelin alterations occurred only at very old age . The murine m-AAA protease subunit AFG3L1 is highly expressed in liver , kidney and heart , but is hardly detectable in the brain ( Fig 2A ) , consistent with previous data [24] . To rule out a major role of Afg3l1 in the mouse nervous system , we generated a full body knock-out of Afg3l1 , by deleting exons 2 and 3 ( S3A Fig ) . Analysis of Afg3l1 transcript levels from liver of 5-week-old mice showed the presence of residual mRNAs after splicing from exon 1 to either exon 4 or exon 5 ( S3B Fig ) . While exon 1–4 splicing gives rise to an out-of-frame transcript , splicing from exon 1 to 5 leads to an in-frame transcript potentially encoding a shorter protein that is devoid of large part of the mitochondrial targeting sequence . However , immunoblotting of liver mitochondria showed no mature AFG3L1 protein ( Fig 2B ) . Most importantly , blue-native PAGE demonstrated lack of assembled AFG3L1 in high-molecular weight m-AAA complexes ( Fig 2C ) , confirming that Afg3l1-/- mice are bona-fide knock-out . AFG3L2 and paraplegin abundance , as well as OPA1 processing , were not affected by lack of AFG3L1 ( Fig 2B ) . Afg3l1-deficient mice were born at the expected Mendelian ratio , showed a comparable growth curve to control littermates ( Fig 2D ) , were fertile , and did not show any evident phenotype up to 78 weeks of age . We carefully examined the brain and the spinal cord of Afg3l1+/- and Afg3l1-/- mice and detected neither obvious myelination defects nor axonal degeneration at least till 1 year of age ( Fig 2F and 2G ) . Ultrastructural analysis of mitochondria in the spinal cord did not revealed morphological abnormalities ( Fig 2G ) . Thus , in contrast to Afg3l2 , mouse Afg3l1 is dispensable in the central nervous system both in neurons and oligodendrocytes . Paraplegin cannot form homo-oligomeric functional complexes [12] , however in absence of AFG3L2 it may assemble together with AFG3L1 and constitute a functional m-AAA protease . Therefore , total ablation of m-AAA complexes can be achieved in oligodendrocytes by Plp1 promoter-driven recombination of Afg3l2 in a null Afg3l1 background . Double knock-out animals ( referred to as DKO; genotype Afg3l1-/- Afg3l2fl/fl Plp1-Cretg/wt ) were compared to control Afg3l1-/- littermates ( CTRL; genotype Afg3l1-/- Afg3l2fl/fl Plp1-Crewt/wt ) . Cre expression was induced by tamoxifen at 4 weeks , as previously described . Starting from 8 weeks of age , DKO mice failed to gain weight compared to CTRL mice ( Fig 3A ) . This difference persisted even after putting food pellets directly in the cage ( Fig 3B ) . Furthermore , DKO mice had decreased fat mass compared to CTRL mice ( Fig 3C ) . At about 13 weeks , DKO mice started to show signs of motor dysfunction . In a rotarod test DKO mice of 11–13 weeks of age spent less time on the rotating rod compared to CTRL mice ( Fig 3D ) . Moreover , the number of foot slips while walking on a 1 cm-wide beam was significantly increased at 13 weeks , and became dramatically higher at 28 weeks ( Fig 3E , S1 and S2 Movies ) . Surprisingly , DKO mice developed a progressive pattern of hair greying starting ventrally close to the forelimbs at 10 weeks , resulting in a grey belly at 17 weeks , and finally extending to the dorsal skin at 28 weeks of age ( Fig 3F ) . Thus , concomitant loss of Afg3l1 in oligodendrocytes strongly exacerbates the phenotypes observed in absence of Afg3l2 . The DKO mice therefore serve as valuable model to examine the significance of a complete m-AAA protease deficiency in myelinating cells . To shed light on the neurological phenotype of DKO mice , we examined brains and spinal cords from DKO and CTRL mice at different time points . At 4 weeks , before tamoxifen injection , the degree of myelination in CTRL and DKO mice was comparable both in the lumbar spinal cord and in the brain ( Fig 4A and S4A Fig ) . However , progressive demyelination was detected in the lumbar spinal cord of DKO , leading to the appearance of demyelinated and degenerating axons as well as dark cells at 28 weeks in the antero-lateral funiculus ( Fig 4A–4C ) . Ultrastructural analysis of the white matter of the spinal cord confirmed progressive demyelination with some axons showing adaxonal detachment of the myelin at 13 weeks , and pronounced signs of demyelination already at 18 weeks ( Fig 4D ) . Signs of secondary axonal degeneration , with accumulation of organelles and material in axons , were visible at 28 weeks ( Fig 4D ) . We identified several axons surrounded by thin myelin , and by oligodendrocytes with dark cytoplasm containing heterogeneous membranous material , probably of lysosomal origin ( Fig 4D ) . These cells have the characteristic of the dark oligodendrocytes , previously proposed to represent mature oligodendrocytes [25 , 26] . The g ratio , expressing the ratio between the diameter of the inner axon and the total fiber diameter , was significantly increased at this age ( Fig 4E ) . In agreement with these data , Gallyas’ silver staining of myelinated tracts in the brain showed prominent loss of white matter in the corpus callosum , the internal capsule , and the cerebellum at 28 weeks of age in the DKO mice ( S4B Fig ) . Progressive loss of myelin was confirmed by western blot analysis of myelin proteins in spinal cord and brain lysates at different time points ( S4C and S4D Fig ) . Concomitantly , we observed upregulation of GFAP , indicating reactive astrogliosis ( S4C , S4D and S5 Figs ) . At 28 weeks , activated microglia cells , which can be recognized by a change in morphology from small cells with slender processes to larger amoeboid-like cells with thick processes , were also detected in the corpus callosum , emphasizing the presence of a neuroinflammatory response ( S5 Fig ) . Since neuroinflammation is a very sensitive read-out of cell damage , we investigated whether cell demise underlies the phenotype . To this end , we crossed DKO mice with the mt-YFP reporter line to visualize targeted oligodendrocytes in vivo . Strikingly , mt-YFP+ cells were rapidly and progressively lost after 6 weeks , and only a few targeted cells remained at 28 weeks ( Fig 5A and 5B ) . Initially , loss of targeted oligodendrocytes was paralleled by a decrease in APC+ cells that was especially evident at 10 weeks both in the corpus callosum and in the spinal cord ( Fig 5A and 5C , S6A and S6B Fig ) . At 10 weeks , the percentage of mt-YFP+ cells in the corpus callosum that were also APC+ was significantly reduced in DKO mice in comparison with the control line carrying only the mt-YFP reporter , while the percentage of APC- Olig2- targeted cells was significantly increased ( Fig 5D ) . This result is reminiscent of what observed in L2KO mice ( S2G Fig ) . Surprisingly , the number of mature oligodendrocytes was recovered at 28 weeks in the DKO . This might be explained by the compensatory proliferation and differentiation of untargeted oligodendrocytes that still express Afg3l2 . Consistently , the size of APC+ cells in the DKO mice was increased at 28 weeks ( Fig 5A , S6C and S6D Fig ) , and the enlarged APC+ cells did not colocalize with mt-YFP+ cells ( Fig 5A ) . We found that the enlarged APC+ cells were in fact intensively stained for Olig2 ( S6D Fig ) , a transcription factor expressed at higher levels in migrating and remyelinating oligodendrocytes [27–29] . However , quantification of total Olig2+ cells ( both intensively and less intensively stained ) indicated no statistical difference in the corpus callosum of the CTRL and DKO mice ( S6E and S6F Fig ) . When we monitored mitochondrial morphology , taking advantage of the expression of the mt-YFP reporter in targeted oligodendrocytes , we found abnormal swollen mitochondria already at 6 weeks of age in the DKO mice ( Fig 6A ) . COX1 staining was preserved at this time , but was lost at 8 weeks in targeted oligodendrocytes of the DKO , indicating impairment of mitochondrial respiratory function ( Fig 6A ) . Moreover , cytochrome c was undetectable in several swollen mitochondria in DKO oligodendrocytes at 8 weeks ( Fig 6B ) . We found several oligodendrocytes showing features of dark cell death ( Fig 6C ) , a caspase-independent form of death , characterized by strong cytoplasmic condensation , chromatin clumping , ruffling of the cell membrane , but no blebbing of the nucleus or plasma membrane [30] . Consistently , in situ TUNEL assay showed only a few apoptotic cells in DKO mice at 7 weeks of age ( quantification in the corpus callosum: 4 . 13 ± 0 . 533 in the CTRL mice , 9 . 75 ± 0 . 85 in the DKO mice , n = 4 mice per genotype ) . In summary , these results suggest that the complete loss of m-AAA protease causes major mitochondrial dysfunction and death of mature oligodendrocytes followed by compensatory repopulation by untargeted oligodendrocytes . One surprising finding was that the DKO mice showed progressive hair greying ( Fig 3F ) . During embryonic development , the Plp1 promoter has been shown to target not only oligodendrocytes , but also Schwann cell ( SC ) precursors ( SCPs ) , bipotential progenitors of both SCs and melanoblasts [31 , 32] . The role of SCPs in the formation of new melanocytes in the adult , and during age-related hair greying remains unknown . Moreover , there is evidence for weak expression of the Plp1 promoter in embryonic melanocytes [33–35] , but its activity in adult melanoblasts and melanocytes is unclear . The observed greying of DKO mice thus raised the question whether melanoblasts were targeted and SCPs were affected . To this end , we performed fate-mapping experiments using the mt-YFP reporter line to establish which cells are targeted in our experiments . We administered tamoxifen at P29 for 5 days , and then collected the ventral or dorsal skin of wild-type mice at P36 . This time corresponds to the growth phase called anagen of the second hair cycle in the mouse . During hair follicle ( HF ) growth , unpigmented melanoblasts differentiate from melanocyte stem cells residing in or close to the hair follicle bulge area and migrate within the outer root sheath of the HF towards the hair matrix where they differentiate into fully mature pigmented melanocytes [36] . We identified mt-YFP+ signal in SCs in the subcutaneous nerve plexus or in the nerves surrounding the HFs , in the bulge area containing melanocyte stem cells , in melanoblasts located in the outer root sheath of the HFs , in pigmented bulbar melanocytes ( S7 Fig ) . Remarkably , targeting was more efficient in the ventral than in the dorsal skin , thus providing a potential explanation for the ventral to dorsal progression of hair greying ( S7 Fig ) . We then investigated the fate of targeted cells in the skin of DKO mice . At 10 weeks of age , we observed a strong reduction of mt-YFP+ melanoblasts in the outer root sheath in DKO mice ( Fig 7A ) . Moreover , non-myelinating SCs in the subcutaneous nerve plexus also showed a significant decrease of mt-YFP signal ( Fig 7A ) . At 28 weeks , although general skin structure was preserved in DKO mice , there was reduced fat deposited in the dermis ( Fig 7B and 7C ) , consistent with the observed general reduction of fat mass in these mice ( Fig 3C ) . Since HF cycling is largely non-synchronized in the mouse at 28 weeks , we shaved the back of the mice , selected pigmented areas of the skin ( containing HFs in anagen ) for biopsy , and stained sections with antibodies against c-KIT , a marker of melanocytes and melanoblasts . Strikingly , DKO mice showed a significant reduction of pigmented HFs ( 79 . 7% ± 2 . 9 in CTRL versus 28 . 0% ± 12 . 3 in DKO mice , n = 3 , Student’s t-test , p < 0 . 05 ) and c-KIT-positive melanoblasts and melanocytes ( Fig 7D ) . We conclude that hair greying is caused by progressive loss of melanocyte stem cells and melanoblasts that are targeted in our experiments . Given the fact that DKO mice showed a pathological phenotype in unmyelinated cutaneous nerves , we further examined if peripheral nerves were affected . The Plp1 promoter is known to be expressed at low level in adult SCs [22] . Consistently , we observed scattered mt-YFP signal within the sciatic nerve of wild-type mt-YFPtg/wt mice at 10 weeks of age , and noticed a reduction of this signal in DKO mice ( Fig 8A ) . Semithin sections of the sciatic nerve did not show a remarkable phenotype at 10 and 28 weeks ( Fig 8B ) , probably because of the very low number of targeted cells . We therefore performed ultrastructural analysis and found clear signs of pathology affecting preferentially small calibre unmyelinated fibers ( Fig 8C ) . These fibers are normally associated with non-myelinating SCs in the so-called Remak bundles that contain several axons wrapped by one individual SC . In a normal Remak bundle the cytoplasm of a SC separates individual axons . In the DKO several Remak bundles appeared affected with individual axons touching each other , and showing initial signs of axonal degeneration ( Fig 8C ) . Some alterations were also observed in a few large calibre myelinated axons . In most cases , these were characterized by enlargement of the inner adaxonal tongue , which contained large vacuoles or other material . Similar changes have been previously observed in Cnp knock-out mice [37 , 38] . These changes were more pronounced at 28 weeks , when also a few demyelinated axons were noted ( Fig 8C ) . Together , our data establish the vulnerability of both SCs and melanoblasts to loss of the m-AAA protease .
Although the m-AAA protease in the inner mitochondrial membrane is essential to preserve respiratory activity in neurons , nothing is known about cell autonomous requirements of this complex in glial cells in vivo . Here , we have generated mouse models expressing different levels of the m-AAA protease in adult myelinating cells . We found that these glial cells survive for long time with reduced levels of the m-AAA protease , but total absence of the m-AAA protease triggers rapid cell death . A main conclusion of our study is that the threshold of m-AAA protease activity allowing survival of neurons and myelinating cells is remarkably different . This is in line with the fact that mutations in AFG3L2 or SPG7 lead to distinct neurodegenerative diseases , characterized by a pure neuronal and axonal phenotype , respectively . What underlies the different cellular vulnerability ? We found no significant difference in the expression of AFG3L2 and paraplegin among astrocytes , oligodendrocytes , and neurons , excluding that different stoichiometry of the individual subunits of the m-AAA protease plays a crucial role . A possible explanation for our results is the different metabolic profile of adult oligodendrocytes and SCs that are able to survive using glycolysis alone when mitochondrial respiration is impaired . Consistently , oligodendrocyte-specific deletion of an essential assembly factor for complex IV , Cox10 , using the same promoter and Cre induction paradigm as in this study , did not lead to axonal degeneration , demyelination , or cell death up to 14 months of age [2] . Moreover , deletion of Tfam in SCs caused conspicuous respiratory deficiency , but did not affect their survival [39] . At odds with the hypothesis that oligodendrocytes compensate metabolically for respiratory deficiencies , is the fact that the complete ablation of the m-AAA protease is incompatible with cell survival . All targeted cells in the DKO mice , oligodendrocytes , SCs , and melanoblasts , showed mitochondria with dramatically abnormal morphology , and died shortly after removal of the complex . Non-myelinating SCs were affected earlier and more prominently than myelinating SCs , in agreement with previous findings of a peculiar susceptibility of these cells to mitochondrial dysfunction [39] . Oligodendrocytes showed features of dark cell death , very similar to those observed in neurons haploinsufficient for Afg3l2 [40] . It is conceivable that the respiratory function of mitochondria in post-myelinating oligodendrocytes and SCs is more important than previously thought . Moreover , deficiencies of the m-AAA protease likely have more severe effects on the oxidative capacity of the organelles than deletion of Cox10 or Tfam . Indeed , Cox10 deletion results in isolated complex IV defect , and that depletion of mtDNA upon loss of Tfam occurs after a rather long time [41] . In contrast , it was sufficient to ablate Afg3l2 alone in oligodendrocytes to trigger mitochondrial fragmentation and swelling , a stress response to defective turnover of de novo synthesized inner membrane proteins [8 , 42] , which was not observed in Cox10-deficient oligodendrocytes [2] . Consistently , deletion of Afg3l2 in adult neurons provoked neuronal loss much earlier than observed when a similar strategy was applied to delete Cox10 or Tfam [39 , 43] , and deletion of Afg3l2 in oligodendrocytes leads to a late-onset myelin phenotype . In the future , it will be important to develop more genetic models lacking specific mitochondrial proteins involved in respiratory function to fully understand the relevance of oxidative phosphorylation for energy metabolism of adult oligodendrocytes . An alternative , not mutually exclusive , explanation for the rapid cell demise induced by the lack of the m-AAA protease is the activation of a death pathway independent from energy deprivation . We recently found that loss of the m-AAA protease results in accumulation of constitutively active MCU-EMRE channels leading to mitochondrial Ca2+ overload , mitochondrial permeability transition pore opening and cell death [44] . Such a death pathway would be consistent with the swollen mitochondrial morphology , and the ultrastructural dark appearance of dying neurons and oligodendrocytes depleted of the m-AAA protease . Furthermore , the intrinsic vulnerability of neurons to Ca2+-dependent cell death may provide a rationale for their increased susceptibility to deficiency of the m-AAA protease compared to oligodendrocytes [45] . The fragmentation of the mitochondrial network triggered by Afg3l2 deletion in oligodendrocytes , as previously observed in neurons [9] , raises the question whether this stress-mediated response has different outcomes in oxidative versus glycolytic cells . Emerging data suggest a possible relationship between mitochondrial network morphology and the metabolic capacity of cells [46 , 47] . However , in oligodendrocytes , stress-induced mitochondrial fragmentation may be beneficial when transient , but become detrimental if persistent [42] , thus contributing to the late-onset myelin abnormalities in mice carrying Afg3l2 deletion in oligodendrocytes . In the central nervous system , our DKO model recapitulates features already observed in models of demyelination [48–50] . Loss of the myelin sheaths in DKO mice occurred a few weeks after oligodendrocyte cell death , consistent with previously described long-term stability of the myelin proteins and lipids [51–53] , and was associated with a regenerative response of oligodendrocytes leading to some degree of myelin repair , further highlighting the reparative potential of adult oligodendrocytes . However , we cannot completely exclude that part of the phenotype is caused by impaired myelin formation , since myelination is not totally complete at P28 when we injected tamoxifen . A consistent finding , both in L2KO and DKO mice , was an increased percentage of targeted cells that were negative for both APC and Olig2 . These cells may represent either oligodendrocytes that lose expression of these markers before dying , and/or oligodendrocyte precursors which fail to differentiate . In fact , experiments in wild-type mice using the mt-YFP reporter line showed that a small percentage of APC- Olig2- cells were targeted shortly after tamoxifen injection , but were hardly detectable at 10 weeks ( compare S1C Fig and Fig 5D ) . Tailored experiments need to be performed in the future to address whether the m-AAA protease or a tubular mitochondrial network are required during oligodendrocyte differentiation . Surprisingly , DKO mice lost weight and fat mass . Albeit we do not know at present the reason for this phenotype , we excluded that this is the result of neurological impairment , hampering to access food . It is conceivable that hypothalamic brain areas involved in feeding behavior may be affected by demyelination . Furthermore , we cannot rule out the possibility that the Plp1 promoter is expressed in other cell types than myelinating cells , contributing to this phenotype . An unexpected phenotype in DKO mice was premature and progressive hair greying . Melanocytes and SCs both arise from the neural crest . A previous study has identified both in chick and mouse two distinct migratory pathways producing melanocyte stem cells during development . At E10-E11 . 5 , melanoblasts delaminate from the murine neural tube and migrate dorsally between the dermomyotome and the epidermis to populate the skin . Later , at E12-E14 bipotential precursors of both melanoblasts and SC , the SCPs , leave the neural crest along a ventral migratory route along the nerves [31 , 32] . According to these studies , at around E13 , SCPs detaching from the nerve differentiate into melanoblasts and downregulate Plp1 , while those that remain attached to the nerve acquire SC properties [31] . This second migratory wave would contribute to melanoblasts in the limbs , the belly , and the dorsal skin , and has been identified performing fate-mapping experiments using a Plp1-CreERT2 transgenic line [31] . In our hands , at 4 weeks the Plp1 promoter is active not only in the nerve , but also in cells that are located in the bulge area and in the outer root sheath of the HF , and therefore have lost contact with the nerve , suggesting that they are melanoblasts . Importantly even a few melanocytes were targeted , as others have observed at developmental stages [34] . We therefore conclude that the Plp1-CreERT line is not suitable to determine whether SCPs contribute to melanocytes in the adult mouse . Aging-associated hair greying has been linked to increased differentiation and/or loss of melanocyte stem cells [36] . As in our model stem cells are lost , without an initial increase in differentiated pigmented cells , the latter is the most likely mechanism , coupled with the concomitant death of targeted melanocytes . Premature hair greying has been previously reported in the mutator mouse , a genetic model of accelerated aging caused by expression of a proof-reading-defective mitochondrial Polg DNA polymerase [54] . The mechanism of hair greying in this model has not been investigated in detail , and more studies are needed to understand the role played by mitochondrial dysfunction or intrinsic pathway of apoptosis in aging-related hair greying . Finally , we show here that constitutive deletion of Afg3l1 alone in the mouse does not lead to an evident neurological phenotype . This is consistent with the extremely low levels of expression in the brain . However , our study indicates that AFG3L1 can largely rescue deficiency of AFG3L2 in oligodendrocytes , by sustaining residual m-AAA activity likely in complex with the more abundant paraplegin subunit . Since Afg3l1 is a functional gene in the mouse but not in humans [55] , examining the phenotypic consequences of mutations in Afg3l2 or Spg7 in an Afg3l1 null background now offers a more suitable model mimicking the human situation . In summary , our data shed new light on functional requirements of the mitochondrial m-AAA protease in adult oligodendrocytes , help understanding cell-specificities in the context of the human pathologies , and provide insights in oligodendrocyte , SC , and melanoblast survival mechanisms .
All animal procedures were carried out in accordance with European ( EU directive 86/609/EEC ) , national ( TierSchG ) , and institutional guidelines and were approved by local authorities ( Landesamt für Natur , Umwelt , und Verbraucherschutz Nordrhein-Westfalen , Germany; approval numbers 87–51 . 04 . 2010 . A219 and 84–02 . 04 . 2015 . A402 ) . Plp1-CreERT mice [22] were purchased from Jackson Laboratory . Conditional Afg3l2fl/fl mice [9] and ROSA26+/SmY mice [23] were previously reported . Afg3l1-/- mice were commercially generated in C57BL/6N background at Taconic-Artemis . Plp1-CreERT mice were mated with ROSA26+/SmY mice to visualize mitochondria in oligodendrocytes . To obtain L2KO mice , Afg3l2fl/fl mice were crossed to Plp1-CreERT mice . As controls , we used Afg3l2fl/fl Cre-negative littermates . Mice of both genotypes were injected with tamoxifen as specified below . Afg3l1-/-Afg3l2fl/fl Plp1-Crewt/tg ( DKO ) mice were compared with Afg3l1-/-Afg3l2fl/flPlp1-Crewt/wt ( CTRL ) littermates injected with tamoxifen . To investigate mitochondrial morphology in L2KO mice or in DKO mice , when indicated , Afg3l2fl/flPlp1-Crewt/tg or Afg3l1-/-Afg3l2fl/flPlp1-Crewt/tg mice were crossed with ROSA26+/SmY mice . Tamoxifen ( T5648 , Sigma ) was dissolved in a corn oil/ethanol ( 9:1 ) mixture at a final concentration of 10 mg/ml . 1 mg tamoxifen was administrated by intraperitoneal injection once a day for 5 consecutive days to 4-week-old ( P28-30 ) mice . Unless specified , mice of either sex were used for experiments . The rotarod apparatus ( TSE systems ) was used to test motor ability and coordination . Mice were placed on a rotating rod ( accelerating model ) and the latency time to fall was recorded for each mouse up to a maximum of 300 seconds . Three tests were performed for three consecutive days . Mice were allowed to rest for 15 minutes after each test . In the beam walking test , mice were trained to walk on a 90 cm long and 3 cm wide beam , elevated by 30 cm on a metal support , for three times for three consecutive days . The actual test was performed by allowing the mice to walk on a 1 cm wide beam on the third day . The performance was filmed and the number of foot slips was quantified . The lean and fat mass of mice was measured with Bruker Minispec Live Mice Analyzer ( LF50H ) . Mice were deeply anesthetized with xylazine/ketamine ( 10 mg/100 mg per kg of body weight ) and perfused transcardially with PBS and 4% paraformaldehyde ( PFA ) . Brain , spinal cord , and the peripheral nerves were then dissected and postfixed in 4% PFA for histology and immunofluorescence or in 2% glutaraldehyde in 0 . 12 M phosphate buffer for electron microscopy . The skin was collected after shaving the mice and immersed in 4% PFA for 2–4 h at 4°C . For RNA extraction , western blot analyses and TUNEL assay , mice were sacrificed by cervical dislocation . Tissues were quickly collected and frozen in liquid nitrogen . RNA extraction was performed with TRIzol reagent ( Life Technologies ) according to the manufacturer specifications . cDNA was synthesized using SuperScript First-Strand Synthesis System ( Life Technologies ) . The sequence of primers used for RT-PCR are available upon request . Postfixed brain and spinal cord were embedded in 6% agar , and 30 μm sections were cut using a vibratome ( VS1000 , Leica ) . Gallyas’ staining was performed as previously described [56] and images were captured with slide scanner ( SCN400 , Leica ) . For immunofluorescence , free-floating sections were permeabilized and blocked in 0 . 4% Triton X-100 and 10% goat serum in TBS for one hour at RT . Primary antibodies were incubated overnight at 4°C , followed by incubation with secondary antibodies for 2 h at RT . Sections were mounted in FluorSave Reagent ( Calbiochem ) . Skin specimens were embedded in paraffin and sectioned at 5 μm thickness using a microtome ( RM2255 , Leica ) . Skin sections were stained with Haematoxylin solution ( MHS32 , Sigma ) and Eosin Y-solution 0 . 5% aqueous ( 1098441000 , Millipore ) . Antigen retrieval was conducted by boiling sections in 0 . 1 M citrate buffer ( pH 6 ) before immunofluorescence analysis . The following primary antibodies were used for immunofluorescence: APC ( 1:400 , OP80 , Calbiochem ) , COX1 ( 1:1000 , 459600 , Invitrogen ) , cytochrome c ( 1:1000 , 556432 , BD Pharmingen ) , Olig2 ( 1:500 , AB9610 , Millipore ) , GFAP ( 1:400 , 3670 , Cell Signaling ) , IBA1 ( 1:2000 , 019–19741 , Wako ) , GFP ( 1:1000 , ab6556 , Abcam ) , MBP ( 1:1000 , SMI94 , Covance ) , and c-KIT ( 1:1000 , 553352 , BD Pharmingen ) . All secondary antibodies were from Molecular Probes: anti-mouse Alexa Fluor 488 ( A-11029 ) , 546 ( A-21143 ) , anti-rabbit Alexa Fluor 488 ( A-11034 ) , 546 ( A-11035 ) , 594 ( A-21207 ) , and anti-rat Alexa fluor 488 ( A-11001 ) . All fluorescent images were acquired using an Axio-Imager M2 microscope equipped with Apotome 2 ( Zeiss ) or gSTED super-resolution and confocal microscope with HyD detector ( TCS SP 8 , Leica ) , as specified . When specified , Huygens Deconvolution software was employed . Quantification of APC+ , Olig2+ , and mt-YFP+ cells was performed manually on single plane images of brain coronal vibratome sections ( 30 μm ) . Three sections of each brain cut at comparable levels ( about -1 . 50 mm , -2 . 8 mm , -3 . 4 mm from the bregma ) were stained with the indicated antibodies . Two to five images of non-overlapping fields in the corpus callosum of one hemisphere were taken for each section and the number of positive cells/area was manually counted . 3 independent mice per genotype and time point were used for quantification . Quantification of APC+ cell size was performed on images acquired using the same exposure time using the measure function of the Axiovision software ( Zeiss ) . To visualize mitochondrial morphology in targeted cells , mice were crossed with ROSA26+/SmY mice . To visualize endogenous mt-YFP signal , sciatic nerves were dissected out , embedded in O . C . T . ( Tissue-Tek ) and were cut longitudinally at a thickness of 7 μm using a cryostat ( CM1850 , Leica ) . For cryosections , the skin was cryoprotected in 15% sucrose for 2 h and then in 30% sucrose overnight , embedded in O . C . T . ( Tissue-Tek ) , frozen on dry ice and sectioned with a cryostat ( CM1850 , Leica ) . 10 μm frozen sections were directly mounted for imaging . For all analyses performed in the brain , vibratome sections were incubated with an anti-GFP antibody to enhance the endogenous YFP signal . Fluorescent images were acquired using a gSTED super-resolution and confocal microscope with HyD detector ( TCS SP 8 , Leica ) , as specified . The circularity of mitochondria in targeted oligodendrocytes within the corpus callosum was measured using a macro of ImageJ . The circularity was calculated with the following formula: 4 *pi* ( area/perimeter^2 ) [57] . To detect TUNEL+ cells , the In Situ Apoptosis assay ( S7101 , Millipore ) was used on cryostat sections out following the manufacturer’s protocol . The number of TUNEL+ cells within the corpus callosum was quantified manually . 4 mice per genotype and 2–3 sections from each mouse were used for quantification . The corpus callosum and the lumbar spinal cord were post-fixed in 2% glutaraldehyde ( Sigma ) in 0 . 12 M phosphate buffer and were treated with 1% osmium tetroxide ( Sigma ) . After dehydration with ethanol and propylene oxide , tissues were embedded in Epon ( Fluka ) . Tissue in Epon-block was further trimmed and cut using an ultramicrotome ( EM UC7 , Leica ) . 1 μm semithin sections were prepared and stained with 1% toluidine blue for light microscopy . For electron microscopy , 70 nm ultrathin sections were cut and stained with uranyl acetate ( Plano GMBH ) and lead citrate ( Electron Microscopy Sciences ) . Images were taken by a transmission electron microscope ( CM10 , Phillips ) equipped with Orius SC200W camera . Quantification of dark cells and number of myelinated axons was performed on at least three semithin micrographs of the anterolateral funiculus of the lumbar spinal cord per mouse . Three independent mice per genotype were analyzed . The number of myelinated axons was quantified using ImageJ particle analyzer with the setting of size 50-infinity and circularity 0 . 3–1 . 0 . The g ratio was determined by measuring the ratio between the diameter of the axon and the diameter of the myelinated fiber on electron micrographs from 3 mice per genotype . Brains from P5 pups were removed and manually dissociated with the Neuronal Tissue Dissociation Kit ( 130-092-628 , Miltenyi Biotec ) . To purify oligodendrocytes , the cell suspension was further incubated with anti-O4 magnetic beads ( 130-096-670 , Miltenyi Biotec ) , washed and loaded onto 30 μm pre-separation filters fixed on top of MS MACS Columns ( 130-042-201 , Miltenyi Biotec ) , which were placed in a magnetic field of the MACS separator ( Miltenyi Biotec ) . The magnetic labeled O4+ cells were retained within the columns , while flow-through was collected . Finally , magnetically labeled cells were flushed out by firmly pushing the plunger into each column . Astrocytes were isolated from the cerebellum of P0-P3 newborn pups . The cerebellum was dissected in dissection solution ( 60% EBSS; 4% Glucose; 30 mM HEPES; 30% FCS III ) . After removing the meninges , the tissue was mechanically meshed , washed twice with EBSS solution ( 10% HEPES; 90% EBSS ) , and placed in glia medium ( 90% DMEM/F12 Hams media with L-glutamine; 9% FCS III; 1% Pen/Strep ) . The single-cell suspension was obtained by mechanical dissociation . The cells were then passed through a 100 μm Nylon cell strainer and the strainer was then washed with 5 ml of glia medium . The cells were centrifuged at 800g for 5 minutes at 4°C . The supernatant was removed and the pellet was resuspended in 10 ml glia medium and plated in 75 cm2 flasks previously coated with poly-L-lysine ( 0 . 1mg/ml ) . For enriched neuronal cultures , the cortex and the hippocampus were dissected from E16 . 5 mouse embryos . The meninges were removed prior mechanical dissociation of the tissue . Chemical dissociation was obtained with Trypsin solution ( 0 . 025% in HBSS ) for 15 minutes at 37°C . The tissue was then washed 3 times with HBSS for 5 minute at 37°C and triturated with Pasteur pipettes ( 0 . 5 mm opening size ) for 7 times . The cells were cultured in in Neurobasal plating Media . One week following the seeding , the cells were collected for protein extraction . Mitochondria from tissues were isolated in MOPS sucrose buffer ( 440 mM sucrose , 20 mM MOPS , 1 mM EDTA , 0 . 2 mM phenylmethylsulfonyl fluoride ) by differential centrifugation at 10 , 000 g . 100 μg of mitochondria were solubilized in 1 M ε-amino n-caproic acid , 50 mM Tris ( pH 7 . 0 ) and digitonin at a detergent to protein ratio of 4 g/g . BN-PAGE was performed as previous described [58] . Cells or tissues were lysed in RIPA buffer and immunoblot analysis was conducted as described [18] . The following primary antibodies were used: AFG3L1 ( 1:1000 ) [12] , AFG3L2 ( 1:1000 ) [12] , Paraplegin ( 1:500 ) [59] , SDHA ( 1:4000 , A11142 , Molecular Probes ) , β-III tubulin ( 1:1000 , T8660 , Sigma ) , MBP ( 1:2000 , SMI94 , Covance ) , CNP ( 1:500 , C5922 , Sigma ) , GAPDH ( 1:2000 , MAB374 , Chemicon ) , CNX ( 1:4000 , ADI-SPA-860 , Enzo ) , GFAP ( 1:2000 , Z0334 , Dako ) . All statistical analyses were performed using GraphPad Prism 6 software , presenting the data as mean ± standard deviation ( SD ) or as mean ± standard error of the mean ( SEM ) . If not stated otherwise , p value was determined by two-tailed unpaired Student’s t test . Statistical significance was defined as *p < 0 . 05 , **p < 0 . 01 and ***p < 0 . 001 . | Oligodendrocytes are cells of the central nervous system that produce the myelin sheath . Myelin production is extremely costly from the energetic point of view , and oligodendrocytes that are synthesizing myelin are particularly susceptible to mitochondrial dysfunction . However , the function of mitochondria in mature oligodendrocytes , after myelination is completed , has been poorly explored using genetic models . Here , we have generated and characterized mouse models expressing different levels of the m-AAA protease , a proteolytic complex preserving proteostasis of the inner mitochondrial membrane and respiratory activity . We show that oligodendrocytes are capable to cope with reduced levels of the complex , but undergo rapid death upon complete ablation of the m-AAA protease . Thus , the m-AAA protease is essential for cell survival , but oligodendrocytes are less vulnerable than neurons to a deficiency of the complex . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"nervous",
"system",
"enzymes",
"enzymology",
"brain",
"neuroscience",
"epithelial",
"cells",
"nerve",
"fibers",
"mitochondria",
"chromatophores",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"spinal",
"cord",
... | 2016 | The Mitochondrial m-AAA Protease Prevents Demyelination and Hair Greying |
There is renewed interest in effective measures to control Zika and dengue vectors . A synthesis of published literature with a focus on the quality of evidence is warranted to determine the effectiveness of vector control strategies . We conducted a meta-review assessing the effectiveness of any Aedes control measure . We searched Scopus and Medline for relevant reviews through to May 2016 . Titles , abstracts and full texts were assessed independently for inclusion by two authors . Data extraction was performed in duplicate and validity of the evidence was assessed using GRADE criteria . 13 systematic reviews that investigated the effect of control measures on entomological parameters or disease incidence were included . Biological controls seem to achieve better reduction of entomological indices than chemical controls , while educational campaigns can reduce breeding habitats . Integrated vector control strategies may not always increase effectiveness . The efficacy of any control programme is dependent on local settings , intervention type , resources and study duration , which may partly explain the varying degree of success between studies . Nevertheless , the quality of evidence was mostly low to very low due to poor reporting of study design , observational methodologies , heterogeneity , and indirect outcomes , thus hindering an evidence-based recommendation . The evidence for the effectiveness of Aedes control measures is mixed . Chemical control , which is commonly used , does not appear to be associated with sustainable reductions of mosquito populations over time . Indeed , by contributing to a false sense of security , chemical control may reduce the effectiveness of educational interventions aimed at encouraging local people to remove mosquito breeding sites . Better quality studies of the impact of vector control interventions on the incidence of human infections with Dengue or Zika are still needed .
The ongoing Zika virus outbreak in Central and South America which started in 2014 has attracted media attention and alarmed public health officials worldwide because of the high number of people affected , rapid transmission rate and association with immuno-neurological disorders ( eg . Guillain-Barré syndrome ) and newborn microcephaly [1–3] . It is feared that Zika virus will spread rapidly in the Americas as was the case for dengue and Chikungunya [2 , 4] . Dengue fever , Zika , Chikungunya and yellow fever viruses are all transmitted by Aedes aegypti mosquitoes and associated with significant disease burden globally . While yellow fever is the only disease that has an effective vaccine , its incidence is increasing and it was stated that yellow fever is making a comeback due to the increasing number of naïve population following the scaling back of mass vaccination and changing sociodemographic conditions [5 , 6] . Aedes is a genus of mosquitos which originated in Africa but are now found worldwide in tropical and subtropical zones . Establishment of Aedes mosquito , especially A . aegypti , has resulted in the epidemic spread of several arboviruses and linked to the current epidemic outbreak of Zika virus in South America [7] . The success of A . aegypti is linked to its opportunistic and high adaptability to the peridomestic environment exploiting any stagnant water as its breeding habitat [8] . Despite decades of Aedes mosquito control programmes , mosquito populations are widely established and abundant worldwide . Recognition of the link between Zika virus and newborn microcephaly in Brazil led to a concerted and renewed interest in Aedes control [7] . The World Health Organisation advice to control Aedes transmitted diseases is well implemented mosquito control measures that can effectively reduce disease transmission [8] . In order to assist the active implementation of Aedes control measures , we sought to provide a timely , up to date and evidence based synthesis of the literature . We carried out a meta-review or “systematic review of systematic reviews” [9 , 10] , to assess and synthesise evidence from systematic reviews and meta-analyses . Meta-reviews allow evidence to be summarised on topics for which multiple systematic reviews have already been published [9 , 11] . In addition , it may be possible to identify patterns of results not previously apparent , by taking into account a larger body of evidence than any individual systematic review captured . Meta-reviews provide a structured approach for exploring and explaining differences in systematic review conclusions , which may have resulted from variations in objectives , quality or other factors . This meta-review critically assessed systematic reviews that investigated the effectiveness of Aedes control interventions or protective measures against Aedes transmitted diseases .
In a previous meta-review investigating control strategies for a number of climate sensitive diseases , a broad search strategy retrieved five systematic reviews about dengue control [12] . For the current meta-review , the search was updated to retrieve recent systematic reviews on control of Aedes transmitted diseases ( published between January 2011 and May 2016 ) . Scopus and Medline Ovid databases were searched using the following search strategy: “ ( dengue OR chikungunya OR yellow fever OR Zika OR Aedes ) AND ( systematic review OR meta-analysis ) ” . This format was restricted to title , abstract and keyword fields . All systematic reviews reporting on the effectiveness of Aedes control measures were included . Reference lists from included reviews were screened for additional relevant reviews . Titles , abstracts and full texts were assessed independently for inclusion by two authors . Data extraction was performed independently in duplicate using a standardised form and differences were resolved by discussion . Data extracted included type of intervention , main outcome measure , number of included studies , type of control group ( pre-post , contemporary ) and pooled effect size ( when reported ) . The methodology and reporting were in accordance with the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” ( PRISMA ) [13] ( S1 Checklist ) . The general approach adopted in this meta-review was based on the 2nd edition of the World Health Organization’s Handbook for Guideline Development especially chapters 8 and 9 [14] . Vector control strategies were categorised as 1 ) Chemical controls ( including insecticide and larvicide applications ) , 2 ) Biological controls ( where a biological agent was used ) , 3 ) Educational campaigns ( focused on training and awareness of the general public with the aim of reduction/ elimination of breeding sites ) or 4 ) Integrated vector controls ( comprising two or more individual control strategies ) also known as Integrated Vector Management . The quality of the evidence was assessed using the GRADE score , recommended by the World Health Organisation [14] , ( http://www . gradeworkinggroup . org/ ) based on five criteria namely: risk of bias , imprecision , inconsistency , indirectness of evidence and publication bias [15] . Scores for each of these criteria were calculated and then combined for each intervention and by outcome measure . The overall score allowed to judge the quality of the evidence as very good , good , poor or very poor ( for details of scoring see S1 Table ) .
Control strategies were classified as chemical , biological , educational or integrated . For each included systematic review , control strategy , main outcome measure ( s ) , number of included studies and effectiveness were recorded for each intervention type . Effectiveness was usually reported as pooled effect size for entomological indices or clinical outcomes . When pooled effects were not reported , descriptive analyses described by the authors were extracted instead . Table 1 summarises the characteristics of the included reviews . Reviews reporting on effectiveness of chemical control were most common ( 8/13 ) , with 17 study arms ( per type of intervention and outcome measure ) . All eight reviews [16–23] reported on the effects of chemical control on entomological indices and 4/8 [16–18 , 22] on dengue incidence . Chemical control included insecticide spraying , insecticide treated curtains , nets and screens , and larvicide application ( particularly temephos ) . Biological control was assessed in six reviews ( 8 study arms ) and included copepods ( crustaceans in water storage that eat mosquito larvae ) , larvivorous fish , Bacillus thuringiensis israelensis ( Bti ) bacterium , predatory insects and turtles . Copepods ( n = 5 ) and Bti ( n = 4 ) were the most widely reviewed biological agents . A single biological strategy was assessed in three reviews ( 5 study arms ) and a combination of biological control strategies was assessed in three systematic reviews ( 3 study arms ) . For biological control , all six reviews [19 , 20 , 23–26] reported on entomological indices and two reviews [24 , 26] also reported on dengue cases . Four reviews ( 5 study arms ) reported on educational campaigns ( involving training , awareness raising and cleanliness incentives in households and/or for school children ) as the only disease control measure . Educational campaigns aimed to reduce breeding sites by removing or covering water containers and elimination of water collection micro-habitats in the peridomestic environment . All four reviews [16 , 17 , 20 , 27] reported on entomological indices and one [16] on dengue incidence . Integrated vector control strategies ( details in Table 1 ) were assessed in 9/13 systematic reviews ( 16 study arms ) . Entomological indices were reported on in all nine reviews [16 , 18–21 , 23 , 25 , 27 , 28] while only two reviews [25 , 27] reported on dengue cases . Educational campaigns and community action interventions focus on educating and encouraging community members to take steps to reduce disease risk through environmental modification in order to reduce or eliminate mosquito’s breeding sites . While educational campaigns are rarely used as the sole control measure , four reviews assessed the effect of this control strategy on dengue transmission . Bowman and colleagues included one RCT which assessed the effectiveness of community based environmental modification ( including clean up , education , mobilisation and use of water container covers ) on dengue incidence , finding a statistically significant reduction in dengue ( OR 0 . 22 , 95% CI 0 . 15 to 0 . 32 ) , but providing only low quality evidence due to unclear allocation concealment , lack of blinding and lack of reproducibility . Evidence on entomological indices came from two studies , which appeared to lead to reductions in Breteau , House and Container Indices , however , the evidence was of low quality [16] . Das and colleagues assessed the effectiveness of preventive community based education and cleanliness campaigns based on three pre- post studies [17] . 25% reduction in ovitrap index ( eggs found in traps per 100 houses ) ( RR 0 . 75 , 95%CI 0 . 62–0 . 91 ) was reported , however , the quality of the evidence was very low . Ballenger-Browning and colleagues assessed the effect of educational or behavioural interventions ( screening , cleaning or disposal of water containers ) based on five studies with contemporary control groups [20] . They reported 41 . 6% mean reduction of entomological indices ( range 4–87 . 6% ) , but this was very low quality evidence . Heintze and colleagues focussed on community-based control programmes ( educational meetings and materials ) based on five studies [27] . No pooled effect size was calculated as the authors found that most primary studies ( all showing reductions in entomological indices ) were of low quality , which was in accordance with the GRADE score . Integrated vector management refers to the simultaneous use of two or more control measures as detailed above . This type of control is favoured because it is thought to be more effective , which is reflected in the number of relevant systematic reviews ( 9/13 ) . George and colleagues reviewed the efficacy of temephos larvicide in water storage containers with other control measures ( chemical or biological vector control , education campaigns ) based on 16 studies [21] . Nine studies were pre-post design and seven were interventions with contemporary control groups ( including 3 RCTs ) , providing very low quality evidence . No pooled effect size was calculated . Although 11 / 16 studies showed that temephos application together with other chemical vector control methods reduced entomological indices , this benefit was either not sustained over time or failed to reduce the immature stages ( in 5 studies ) . The effectiveness of temephos depended on various factors including quality of delivery , water turnover rate , water type , organic debris , temperature and exposure to sunlight . In addition , long term success depended on political commitment and community participation . Limitations to temephos use and community effectiveness were identified as need for reapplication , cost , supplies , time consuming and laborious nature , high water turnover and temephos resistance as well as poor acceptability ( due to unpleasant odour and taste ) and limited local knowledge . Furthermore , it was reported that the use of temephos as part of an integrated strategy seemed to reduce implementation rate and effectiveness of source reduction and environmental management because of a false sense of security due to the belief that temephos application alone is sufficient to prevent dengue [21] . Han and colleagues assessed the effect of larvivorous fish in combination with other biological control measures and educational campaigns based on three studies ( 2 pre-post and 1 contemporary control group ) [25] . All studies reported reductions in entomological indices though no pooled effect was calculated . The quality of evidence was very low . The same review assessed the effect of larvivorous fish alone or as part of integrated control on dengue cases based on two studies , providing very low quality evidence [25] . The first study found no dengue cases in any village since the start of the intervention , and the other study reported a decline from 6 cases pre-intervention to zero cases post-intervention , but the authors stated that this could not be attributed solely to the intervention . Lima and colleagues investigated the effectiveness of integrated vector control combining biological , chemical and educational strategies based on 12 studies all with contemporary control groups [23] . The pooled significance statistic ( pw ) suggested statistical significance , but provided very low quality evidence . Al-Muhandis and Hunter focussed on the role of community based educational interventions either alone or in combination with chemical or biological control ( including indoor and outdoor insecticide spraying , larviciding , copepods , covering , removal and clean-up of water containers ) [28] . This review included 22 studies ( 6 pre-post and 16 contemporary control groups ) , and reported a pooled relative effectiveness of 0 . 25 ( 95%CI 0 . 17–0 . 37 ) for entomological indices , with very low quality evidence . The authors reported that 61% of the heterogeneity in outcome measures could be explained by the type of control group and time from intervention to assessment . Studies using pre-post design substantially overestimated intervention effectiveness compared to studies using contemporary controls . It was noted that the effectiveness of educational interventions was maintained for about 18 months , and the authors observed that adding chemical or biological control to educational campaigns did not add value or increase effectiveness [28] . This finding was also reported by Esu and colleagues [18] who stated that houses that received both educational and chemical control did not achieve significant reduction of entomological indices , while houses that received educational campaigns alone achieved significant reduction . The authors concluded that chemical spraying may create a false sense of security and thus reduce the beneficial effect of educational campaigns . Erlanger and colleagues assessed three types of integrated vector control strategies [19] . The first category focused on environmental management ( removal of unused and covering of water containers ) in combination with insecticide treated nets , curtains and screens and included 14 studies . The authors conducted pooled analyses , finding statistically significant reductions in three entomological indices , and providing very low quality evidence: Breteau Index ( pooled BI , 0 . 71 , 95% CI 0 . 55 to 0 . 90 ) based on 9 studies , Container Index ( pooled CI , 0 . 43 , 95% CI 0 . 31 to 0 . 59 ) and House Index ( pooled HI , 0 . 49 , 95% CI 0 . 30 to 0 . 79 ) ( both based on 10 studies each ) . The second integrated control category was environmental management in combination with outdoor and indoor spraying as well as bed nets and larviciding . The pooled effect sizes suggested improvements in Breteau and House Indices , but not Container Index , though the evidence was of very low quality: ( BI 0 . 33 , 95%CI 0 . 22 to 0 . 48 based on 11 studies , CI 0 . 17 , 95%CI 0 . 02 to 1 . 28 based on 9 studies and HI 0 . 12 95%CI 0 . 02 to 0 . 62 based on 8 studies ) . The third category was environmental management in combination with biological control , for which 5 primary studies were retrieved but no pooled effect was calculated as the studies reported on distinct entomological indices , providing very low quality evidence . Due to the consistent evidence of improvements in entomological indices , Erlanger and colleagues concluded that dengue vector control is effective in reducing vector populations [19] . However , their conclusion was not supported by the quality of evidence . The review did not report study methodology or assess study validity , study results were clearly heterogeneous , publication bias was unclear and no health outcomes were reported . The authors investigated intervention type as a source of heterogeneity and did not attempt to investigate whether excluding studies from pooling would bias their conclusions . Heintze and colleagues focussed on community-based educational control programmes in combination with chemical larvicide and larvivorous fish or copepods based on 11 studies ( 2 RCTs , 6 pre-post studies and 3 interrupted time series ) [27] . Each category was assessed separately and by outcome measure i . e . entomological indices and dengue incidence resulting in a very small number of primary studies per category . The authors reported that most studies were of low quality and concluded that the evidence of the effectiveness of community-based dengue control programmes is weak , which concurs with our GRADE score showing very low quality evidence for these interventions . In addition to the widely used control strategies discussed above , Bowman and colleagues reviewed the effect of insect repellents ( 1 study ) , mosquito coils ( 2 studies ) and mosquito traps ( 1 study ) on dengue incidence [16] . The use of insect repellents and mosquito traps were not associated with a protective effect , while mosquito coils were significantly associated with an increased risk of dengue incidence ( OR 1 . 44; 95% CI 1 . 09–1 . 91; p = 0 . 01 ) . The quality of the evidence was very low .
Most included systematic reviews focussed on reducing entomological indicators . Undeniably , vector presence is pivotal for disease transmission , yet , there is no clear evidence of quantifiable association between vector density and disease transmission in particular whether reducing vector abundance actually leads to less disease [29] . This shortcoming was noticed by only a few systematic reviews’ authors . For example Heintze and colleagues stated “our findings suggest that although community-based control strategies in addition to or together with biological and chemical vector control tools are able to reduce classical Aedes larval indices , it is unknown whether this reduces dengue transmission” [27] . Therefore , evidence about entomological indices only was downgraded in our quality assessment for intervention impacts on disease incidence . Indeed , out of eight reviews that assessed the effect of vector control on disease outcomes [16–18 , 22 , 24–27] , only two showed pooled statistically significant reduction in dengue incidence or positive serology [16 , 17] . Future research on the effect of vector control strategies should utilise RCT methodology , have longer durations and report disease-related outcomes . The strength of evidence for the effectiveness of any vector control intervention was uniformly low or very low . This means that while , in many cases , there was a suggestion of improvement , this was not scientifically rigorous , and we have little ability to compare effectiveness ( or cost effectiveness ) of different strategies . This was due to several reasons . Where dengue incidence was directly assessed , primary studies were generally observational , and intervention studies mostly assessed entomological outcomes ( both observational studies and indirect outcomes downgraded the strength of evidence ) . The risk of bias in included studies was generally scored as very high due to problems with allocation concealment and blinding in intervention studies and lack of quality assessment or problems with confounding and dissimilarity of comparator groups at baseline for observational studies . In order to truly understand the effectiveness of dengue ( and other vector-borne diseases ) control interventions , we need high quality randomised controlled trials with adequate blinding , allocation concealment and sample size reporting on disease outcomes for long enough follow-up period . Our review of the evidence was hampered by the quality of primary studies as well as some of the systematic reviews included . It was uncommon for systematic reviews to describe study methodology accurately or assess study validity appropriately or publication bias . These omissions have inevitably clouded our understanding of the levels of bias within the included primary studies [30] . It is possible that some evidence maybe of higher quality than assessed using the GRADE score , but in the absence of clear reporting , the quality of the evidence is downgraded . Where reviews did not assess the underlying validity of the included studies , particularly study methodology ( type of control group , randomisation , allocation concealment and blinding ) , the effectiveness of vector control strategies was more likely to be over-stated . Previous research established that using historical controls ( pre-post studies ) substantially over-estimated effectiveness compared to studies using contemporary control groups [28] . The use of historic controls is considered poor practice as most historical control groups are compromised [31 , 32] . Many studies in the review by Erlanger and colleagues [19] had pre-post design , which may explain their conclusions that “dengue vector control is effective in reducing vector populations” , even though our assessment suggests very low quality evidence . However , other reviews surveying some of the same evidence were more cautious , such as Heintze and colleagues [27] who concluded “Evidence that community-based dengue control programmes … can enhance the effectiveness of dengue control programmes is weak” , Ballenger-Browning and Elder said “Little evidence exists to support the efficacy of mosquito abatement programs owing to poor study designs and lack of congruent entomologic indices” [20] and Esu and colleagues stated “Based on a comprehensive search of available peer reviewed literature , the effectiveness of peridomestic space spraying in reducing dengue transmission has not been conclusively demonstrated” [18] . While systematic reviews represent high quality evidence , we acknowledge that they might exclude relevant studies due to strict inclusion criteria and are limited to dated evidence i . e . by the time of publication , the recent literature could comprise relevant studies ( potentially changing the body of evidence ) . Therefore , we attempted to provide a brief overview of latest relevant research that did not inform this meta-review , including novel vector control strategies that did not have ample body of evidence warranting consideration by systematic reviews’ authors . Further information is provided in S1 Text . It is worth bearing in mind that effectiveness of any disease control intervention is closely related to the specific settings of the study area . For example , Lazaro and colleagues found that copepods were effective in studies carried out in Vietnam , including long-term control of larval and adult A . aegypti and dengue incidence [24] . However , this success was not replicated in studies conducted elsewhere ( Costa Rica , Mexico , USA , Honduras , Laos ) . The authors attributed the success in Vietnam to community participation , environmental and/or biological factors . Tran and colleagues discussed social sustainability of copepods for dengue control in Vietnam , and reported that effectiveness varied between northern and central Vietnam ( high sustainability ) and south Vietnam ( low sustainability ) [33] . Limited knowledge and education , lack of government support , poor implementation and poor household monitoring were the main drivers of low sustainability and limited effectiveness [33] . Further investigations including qualitative research alongside RCTs may assist better understanding of crucial factors supporting or reducing the effectiveness of specific control interventions . The World Health Organisation ( WHO ) recommends “integrated approaches that tackle all life stages of the mosquito and fully engage communities” for the control of Zika and other Aedes transmitted diseases [8] . This is in accordance with Heintze and colleagues that “multifaceted interventions are more effective than single interventions because a larger variety of barriers for change can be addressed” [27] , which is also in line with social science theory [34] . However , two reviews found that adding additional chemical or biological interventions to educational campaigns did not increase efficacy [18 , 28] . This was attributed to a false sense of security following insecticide spraying [18] and the belief that temephos alone is sufficient to control dengue transmission [21] . Therefore , our review suggests that the WHO is correct to reiterate that the most effective intervention to control disease and protect populations is the elimination of mosquito breeding sites [8] , which would require sustained and ongoing education campaigns , resource allocation and good governance . This is particularly important considering the resilience of A . aegypti mosquitoes , with population numbers recovering and increasing shortly after vector control strategies have ceased [12] . While prevention of mosquito borne diseases has always focused on control of the mosquito vector , there is a debate about whether a rethink of control strategies is warranted . This is relevant considering the day biting pattern and low flight range ( <100 m ) of Aedes mosquitoes . These traits mean that vector control strategies should be focused not only on the peridomestic environment but also on day gathering places such as markets , schools , hospitals etc . and combined with better diagnosis and monitoring/ restriction of viremic persons’ movement , which has been found to be an important driver of dengue spatiotemporal clustering and disease spread [35] . In addition , relevant factors driving establishment of Aedes and spread of Aedes transmitted diseases need to be better understood and accounted for when designing control strategies such as international travel and trade , urbanisation , water storage practices , socioeconomic factors and global environmental change . All the primary studies included in the systematic reviews were undertaken in low and middle income countries ( LMICs ) . Caveats may need to apply if extrapolating public health research between LMICs and indeed to high income nations . The efficacy of any newly introduced vector control measure may depend on other control measures already in place [36] . Another knowledge gap identified here is the scarcity of data on cost effectiveness of vector control strategies in systematic reviews [27] . Bearing in mind that Health Economics is currently a major element in decision making processes , future studies should address this gap [37 , 38] . This is particularly important considering the significant burden of dengue and other vector-borne diseases ( including Zika and yellow fever ) and the international commitment to improve global health and eradicate poverty related diseases with finite financial means .
We identified thirteen systematic reviews assessing dengue or Aedes control strategies . Control strategies were categorised and the effect of interventions on entomological indices and disease incidence were recorded . Though some systematic reviews reported significant reduction of entomological indices , most reviews were considered to be of low to very low quality . This suggests that more high quality primary studies and well conducted systematic reviews that follow PRISMA reporting guidance and report on the quality of evidence [13] are still required for evidence based recommendations . The systematic reviews we assessed suggest that biological control achieves better and more sustainable reduction of entomological indices than chemical control . Educational campaigns and community engagement appear paramount in reducing breeding habitats in the peridomestic environment , although ongoing resources must be allocated to ensure educational interventions are maintained . Chemical control measures could be associated with a false sense of security leading to lesser community engagement with reduction/ elimination of breeding sites . Promising novel vector control strategies are being tested and would be a valuable addition to control mosquito borne diseases . | Various strategies for the control of mosquito-borne diseases exist and have been used for decades . The effectiveness of these control measures has been evaluated in several systematic reviews , however , their conclusions were contradicting . The current Zika outbreak in the Americas renewed the global health community’s interest in the control of Aedes transmitted diseases ( dengue , yellow fever and chikungunya ) . We sought to provide an up to date systematic review about the effectiveness of chemical , biological , educational and integrated vector control strategies . In addition , we looked at recent primary studies that were not included in any systematic review as well as novel tools for mosquito control . This meta-review provides a comprehensive list of systematic reviews on the effect of vector control interventions on entomological parameters ( most often indicators of vector density ) or disease incidence . Biological control was found to achieve higher reduction of mosquito populations than chemical control . Educational campaigns are essential to reduce breeding sites and interrupt disease transmission . Integrated vector control strategies may not always increase effectiveness . The quality of the evidence was low to very low for most interventions . The effectiveness of any control strategy is setting- dependent . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusions"
] | [
"larvicides",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"animals",
"copepods",
"developmental",
"biology",
"pest",
"control",
"crustaceans",
"infectious",
"disease",
"control",
"insect",
"vectors",
"zoology",
"research",
"and",
"analysis",
"methods",
"in... | 2016 | Public Health Interventions for Aedes Control in the Time of Zikavirus– A Meta-Review on Effectiveness of Vector Control Strategies |
Specific intestinal microbiota has been shown to induce Foxp3+ regulatory T cell development . However , it remains unclear how development of another regulatory T cell subset , Tr1 cells , is regulated in the intestine . Here , we analyzed the role of two probiotic strains of intestinal bacteria , Lactobacillus casei and Bifidobacterium breve in T cell development in the intestine . B . breve , but not L . casei , induced development of IL-10-producing Tr1 cells that express cMaf , IL-21 , and Ahr in the large intestine . Intestinal CD103+ dendritic cells ( DCs ) mediated B . breve-induced development of IL-10-producing T cells . CD103+ DCs from Il10−/− , Tlr2−/− , and Myd88−/− mice showed defective B . breve-induced Tr1 cell development . B . breve-treated CD103+ DCs failed to induce IL-10 production from co-cultured Il27ra−/− T cells . B . breve treatment of Tlr2−/− mice did not increase IL-10-producing T cells in the colonic lamina propria . Thus , B . breve activates intestinal CD103+ DCs to produce IL-10 and IL-27 via the TLR2/MyD88 pathway thereby inducing IL-10-producing Tr1 cells in the large intestine . Oral B . breve administration ameliorated colitis in immunocompromised mice given naïve CD4+ T cells from wild-type mice , but not Il10−/− mice . These findings demonstrate that B . breve prevents intestinal inflammation through the induction of intestinal IL-10-producing Tr1 cells .
Recent advances in metagenomic analysis of intestinal bacteria have revealed that inflammatory bowel diseases ( IBD ) is associated with dysbiosis in the intestinal microflora [1] , [2] , [3] . In support of these human studies , analysis of mice lacking NLRP6 has revealed that altered composition of intestinal symbiotic bacteria contributes to the pathogenesis of colitis [4] . Probiotics , live microorganisms which confer a health benefit on the host when administered in appropriate amounts , have been used for the treatment of IBD [5]–[8] . Probiotics have been shown to modulate the intestinal symbiotic bacteria leading to the maintenance of intestinal homeostasis [9] , [10] . Modulation of microbiota by probiotics has been shown to be elicited by antagonizing pathogenic bacteria through the reduction of luminal pH , inhibition of bacterial adherence , or production of anti-microbial molecules [8] . Probiotics have also been shown to enhance barrier functions of intestinal epithelial cells [11] . Thus , several mechanisms for the cross-talk between probiotics and the host have been postulated . Recent accumulating evidence has indicated that intestinal commensal microbiota has a great influence on the host intestinal immune system [12]–[14] . Commensal microbiota has been shown to induce IgA-mediated responses and development of Th1/Th17 effector T cells as well as regulatory T ( Treg ) cells [15]–[17] . More recently , a specific microbiota that induces development of Th17 cells or Treg cells has been demonstrated . Segmented filamentous bacteria ( SFB ) , which have been previously shown to induce IgA-producing cells in the small intestine , were shown to induce Th17 cell development in the small intestine of mice [18] , [19] . A human symbiotic bacterium , Bacteroides fragilis has been shown to mediate Toll-like receptor 2 ( TLR2 ) -dependent development of Foxp3+ Treg cells in the large intestine [20]–[22] . Clostridium species mediate TLR-independent induction of Foxp3+ Treg cells in the large intestine [23] . Thus , several selective intestinal bacteria promote development of intestinal T cells via distinct mechanisms . Most recently , microbiota-dependent induction of Foxp3+ Treg cells has been shown to be required for the establishment of intestinal CD4+ T cell homeostasis [24] . Additionally , commensal microbiota has been shown to educate Foxp3+ Treg cells to acquire the antigen-specific repertoires of their T cell receptors [25] . Probiotics have also been shown to directly modulate the host immune system , especially the induction of Foxp3+ Treg or TGF-β-bearing Treg cell development [26]–[29] . Thus , several mechanisms for intestinal bacteria-dependent development of Foxp3+ Treg cells have been postulated . Intestinal homeostasis is maintained by regulatory T cell populations consisting of two major CD4+ T cell subsets; Foxp3+ Treg cells and IL-10-producing type 1 regulatory T ( Tr1 ) cells [30] . Tr1 cells modulate immune responses via mechanisms distinct from those used by Foxp3+ Treg cells [31] . Indeed , Tr1 cells do not express the master Treg transcription factor Foxp3 , and are induced by distinct cytokines such as IL-10 and IL-27 [32] , [33] . Tr1 cells are abundant in the intestinal lamina propria [34] , yet it remains unclear how Tr1 cells develop in the intestine . In this study , we analyzed the effect of two probiotic strains , Bifidobacterium breve and Lactobacillus casei , on intestinal T cells responses . Oral administration of B . breve , but not L . casei , resulted in increased IL-10 production from colonic CD4+ T cells , without enhancing Foxp3 expression . B . breve-induced IL-10-producing CD4+ T cells possessed properties of Tr1 cells , as evidenced by expression of cMaf , Il21 , and Ahr . B . breve-dependent Tr1 cell induction was mediated by intestinal CD103+ dendritic cells via TLR2/MyD88-dependent production of IL-10 and IL-27 . B . breve administration ameliorated intestinal inflammation in immunocompromised mice transferred with naïve CD4+ T cells in an IL-10-dependent manner . These findings establish the mechanisms for Tr1 cell induction by the probiotic B . breve , which modulates the host immune responses .
Lactobacillus casei strain Shirota and Bifidobacterium breve Yakult strain have been proven to be beneficial for the treatment of several diseases such as diabetes mellitus , arthritis and inflammatory bowel diseases [35]–[40] . In order to analyze the effect of these probiotic strains on the intestinal homeostasis , we orally treated C57BL/6 mice with L . casei and B . breve ( 109 bacteria each ) for 3 months . We first analyzed fecal microbiota using both quantitative PCR and reverse transcription-quantitative PCR methods targeting rDNA and rRNA , respectively [41] . Administration of L . casei and B . breve did not induce a significant change in the number and composition of microbiota ( Text S1 , Table S1 ) . Because several microbiota have been shown to induce differentiation of intestinal CD4+ T cells [17] , we analyzed production of IL-10 , IL-17 , and IFN-γfrom CD4+ T cells in the small intestine and large intestine of mice orally treated with L . casei and B . breve . The number of IL-10- , IL-17- , and IFN-γ-producing T cells in both the small intestine and the large intestine was not altered in mice administered with L . casei ( Figure 1A , B ) . In B . breve-treated animals , the number of IL-17- and IFN-γ-producing T cells in the small intestine and the large intestine was not significantly changed . However , the number of IL-10-producing T cells was increased about two-fold in the large intestine , but not altered in the small intestine , spleen , and mesenteric lymph nodes ( MLN ) ( Figure 1C , D and Figure S1 ) . Thus , administration of B . breve in C57BL/6 mice selectively increased the number of IL-10-producing CD4+ T cells in the large intestine without modulating intestinal microbiota . We next analyzed the effect of B . breve on the BALB/c mouse strain . BALB/c mice were orally treated with B . breve ( 109 bacteria ) for the indicated time before expression of IL-10 in CD4+ T cells of the large intestinal lamina propria was analyzed . The number of colonic IL-10-producing T cells increased after 2 weeks of treatment , and by 3 weeks the number of IL-10-producing cells had doubled ( Figure 2A , C ) . Because IL-10 has been shown to be produced from Foxp3+ and Foxp3− populations of intestinal T cells , we analyzed expression of Foxp3 in colonic T cells in B . breve-treated BALB/c mice . The number of Foxp3+ CD4+ T cells in the large intestine was not altered in B . breve-treated mice ( Figure 2B , D ) . Therefore , we orally administered B . breve into Foxp3-GFP mice , and analyzed IL-10 expression in the colonic CD4+ T cells 3 weeks after beginning treatment . The number of IL-10-producing cells was increased in the Foxp3− population , but not in the Foxp3+ population of B . breve-treated mice ( Figure 2 E , F ) . Thus , B . breve administration selectively increased IL-10-producing Foxp3− CD4+ T cells in the large intestine . We next analyzed how B . breve induces IL-10-producing T cells . Because intestinal dendritic cells ( DCs ) modulate T cell differentiation into effector or regulatory T cells , CD11c+ cells were isolated from the colonic lamina propria , stimulated with B . breve or L . casei , and then co-cultured with splenic naïve CD4+ T cells . After 4 days of the co-culture , T cells were harvested and stimulated with coated anti-CD3 mAb and soluble anti-CD28 mAb . CD4+ T cells co-cultured with B . breve-treated , but not L . casei-treated , intestinal DCs produced high amounts of IL-10 ( Figure 3A ) . We analyzed the effect of other Bifidobacterium spp . Intestinal DCs treated with B . adolescentis or B . bifidum did not induce IL-10 production from co-cultured CD4+ T cells , although B . longum-treated DCs moderately induced IL-10-producing T cells ( Figure 3B ) . Thus , B . breve strongly induced IL-10-producing T cells via activation of intestinal DCs . In contrast to high induction of IL-10 , B . breve-treated intestinal CD11c+ cells did not induce Foxp3 expression in co-cultured CD4+ T cells ( Figure 3C ) . IL-10-producing Foxp3− T cells have been characterized as type 1 regulatory T ( Tr1 ) cells expressing c-Maf , aryl hydrocarbon receptor ( Ahr ) and IL-21 [42]–[44] . Therefore , we analyzed expression of cMaf , Ahr and Il21 . Expression of cMaf , Ahr and Il21 was increased in CD4+ T cells co-cultured with B . breve-treated , but not L . casei-treated , intestinal CD11c+ cells ( Figure 3D ) . These findings indicate that B . breve-treated intestinal DCs promote the induction of IL-10-producing Tr1 cells . Intestinal DCs consists of two major subsets; CD103+ CX3CR1− CD11b− DCs ( CD103+ DCs ) and CX3CR1+ CD11b+ DCs ( CX3CR1+ DCs ) [45] , [46] . Therefore , we analyzed which subset mediates B . breve-dependent Tr1 cell development . CD103+ DCs and CX3CR1+ DCs were isolated from the colonic lamina propria , treated with B . breve , and then co-cultured with naïve CD4+ T cells . CD4+ T cells co-cultured with B . breve-treated CD103+ DCs , but not CX3CR1+ DCs , produced high amounts of IL-10 ( Figure 4A , B ) . B . breve caused a dose-dependent increase in IL-10 production from T cells co-cultured with CD103+ DCs ( Figure S2 ) . CD103+ DCs have been shown to induce Foxp3+ Treg cells [47] , [48] . Indeed , CD103+ DCs induced low levels of Foxp3 expression on co-cultured CD4+ T cells even in the absence of TGF-β or retinoic acid ( Figure 4B ) . However , B . breve-treated CD103+ DCs did not induce Foxp3 expression , but induced enhanced IL-10 production in co-cultured T cells . Next , we analyzed whether intestinal CD103+ DCs in B . breve-treated mice instruct Tr1 cell development . Since CD103+ DCs have been postulated to sample intestinal antigens in the lamina propria and move to MLN where they induce Foxp3+ T cells [49] , we analyzed CD103+ DCs in MLN and colonic lamina propria . C57BL/6 mice were fed with B . breve for 3 weeks , before CD103+ DCs were isolated from MLN and colonic lamia propria , and co-cultured with naive CD4+ T cells . CD4+ T cells co-cultured with CD103+ DCs from the colonic lamina propria of B . breve-fed mice showed higher IL-10 production , with lower levels observed in CD4+ T cells co-cultured with MLN CD103+ DCs ( Figure 4C ) . Thus , intestinal CD103+ DCs possess an enhanced capacity to induce Tr1 cells by B . breve treatment in mice . These findings indicate that intestinal CD103+ DCs are responsible for B . breve-dependent Tr1 cell development . IL-10 was originally shown to induce Tr1 cells [50] . Subsequently , IL-27 was identified as a growth and differentiation factor for Tr1 cells [51]–[53] . Therefore , we analyzed expression of these key cytokines in B . breve-treated CD103+ DCs . B . breve treatment increased expression of Il27p28 , Ebi3 ( both of which encode subunits of IL-27 ) , and Il10 in CD103+ DCs ( Figure 5A ) . Furthermore , neutralizing mAb to IL-10 or IL-27 severely or moderately blocked B . breve-mediated development of Tr1 cells , respectively , and combination of both mAbs almost completely blocked Tr1 cell development . In contrast , neither a retinoic acid receptor antagonist LE540 nor anti-TGF-β neutralizing Ab inhibited B . breve-mediated Tr1 cell induction ( Figure 5B and Figure S3 ) . These findings indicate that IL-10 and IL-27 , which are produced from B . breve-treated CD103+ DCs , mediate Tr1 cell development . In order to corroborate these findings , we analyzed Il10−/− and Il27ra−/− mice . We first treated CD103+ DCs from the colonic lamina propria of wild-type or Il10−/− mice with B . breve , before co-culturing them with wild-type naïve CD4+ T cells . CD4+ T cells co-cultured with B . breve-treated Il10−/− DCs produced severely decreased levels of IL-10 ( Figure 5C ) . Then , CD4+ T cells were isolated from the spleen of Il27ra−/− mice and co-cultured with B . breve-treated wild-type CD103+ DCs . IL-10 production from Il27ra−/− T cells was severely decreased ( Figure 5D ) . Taken together , these findings demonstrate that IL-10 and IL-27 , which are produced by B . breve-treated CD103+ DCs , cooperatively mediateTr1 cell induction . We next analyzed which signaling pathway is responsible for B . breve-dependent production of IL-10 and IL-27 from CD103+ DCs . Several pattern recognition receptors mediate activation of innate immunity through the recognition of microbe-associated molecular patterns [54] . Therefore , we analyzed the involvement of Toll-like receptor ( TLR ) signaling using Myd88−/− mice . In intestinal CD103+ DCs from Myd88−/− mice , B . breve-induced expression of Il27p28 , Ebi3 , and Il10 was severely reduced ( Figure 6A ) . Furthermore , wild-type CD4+ T cells , which were co-cultured with B . breve-treated Myd88−/− CD103+ DCs , failed to produce IL-10 ( Figure 6B ) . These findings indicate that the TLR signaling pathway in CD103+ DCs is critically involved in B . breve-mediated Tr1 cell development . We further analyzed which TLR mediates B . breve-mediated responses . B . breve-induced expression of Il27p28 , Ebi3 , and Il10 was severely reduced in intestinal CD103+ DCs of Tlr2−/− mice ( Figure 6C ) . In addition , B . breve-treated Tlr2−/− CD103+ DCs did not promote the development of IL-10-producing T cells ( Figure 6D ) . B . breve-treated Tlr4−/− and Tlr9−/− CD103+ DCs induced IL-10-producing cells normally ( Figure S4 ) . CD103+ DCs treated with the TLR2 ligand , but not TLR4 or TLR5 ligand , induced Tr1 cells , albeit reduced when compared with B . breve ( Figure S5 ) . The critical involvement of the TLR2-mediated pathway in B . breve-dependent Tr1 induction was further confirmed in Tlr2−/− mice orally administered with B . breve for 4 weeks ( Figure 6E , F ) . In Tlr2−/− mice , B . breve treatment did not increase the number of IL-10-producing CD4+ T cells in the colonic lamina propria . Taken together , these findings demonstrate that the TLR2/MyD88-dependent pathway in CD103+ DCs mediates B . breve-mediated Tr1 cell induction . Probiotic strains of bacteria have been shown to be used for the treatment of several diseases including IBD [5]–[8] . Therefore , we analyzed the effect of oral B . breve treatment in intestinal inflammation caused by transfer of naïve CD4+ T cells into immune-compromised severe combined immunodeficiency ( SCID ) mice . Daily treatment with B . breve markedly improved the severity of intestinal inflammation ( Figure 7 A , D , E ) . In B . breve-treated SCID mice , IL-10 concentration in the colonic tissues was increased , whereas IFN-γ concentration was decreased ( Figure 7C ) . We then analyzed whether IL-10 was responsible for the prevention of intestinal inflammation . SCID mice were transferred with naïve CD4+ T cells from Il10−/− mice and orally treated with B . breve . No effect on the amelioration of intestinal inflammation in SCID mice given Il10−/− CD4+ T cells was observed ( Figure 7B , D , E ) . These findings demonstrate that T cell-derived IL-10 suppresses T cell-dependent intestinal inflammation in B . breve-treated SCID mice .
In the present study , we show that probiotic B . breve promotes development of IL-10-producing Tr1 cells in the colon without altering the composition of intestinal commensal flora . Intestinal CD103+ DCs mediate B . breve-induced development of Tr1 cells via the TLR2/MyD88-dependent induction of IL-27 and IL-10 . Recent accumulating evidence has indicated that specific microbiota influence the development of intestinal T cells . Segmented filamentous bacteria have been shown to induce Th17 cells in the small intestine [18] , [19] . Polysaccharide A ( PSA ) of B . fragilis has been shown to promote Foxp3+ Treg cell development via TLR2 expressed on T cells in the large intestine [21] , while Clostridium species have been shown to induce Foxp3+ Treg cells in the colon through TGF-β induction of epithelial cells [23] . Several probiotic strains of commensal bacteria have also been shown to induce Foxp3+ Treg cells or TGF-β expressing Treg cells [27]–[29] , [55] . Several studies have also indicated that selective probiotics induce IL-10 production in the intestine or the development of IL-10-producing T cells in vitro [26] , [29] , [56] . However , the precise mechanism by which probiotics induce IL-10-producing T cells in the intestinal lamina propria remained unknown . This study clearly demonstrates that a probiotic strain of bacteria , B . breve , promotes development of Foxp3− Tr1-type of T cells . Several recent studies have demonstrated that colonization of specific microbiota in germ-free mice induced development of Treg cells and Th17 cells [18] , [19] , [21] , [23] . However , oral administration of probiotic B . breve did not induce colonic Tr1 cells in germ-free mice . This might be due to that fact that B . breve has a low ability to colonize in the intestine by itself . As was the case in other studies [18] , [19] , [21] , [23] , germ-free mice received single administration of B . breve . However , due to the low ability to colonize in the intestine , B . breve might not be able to induce Tr1 cell development by single administration . Alternatively , this probiotic strain might require assistance by other commensal bacteria to be uptaken or recognized by intestinal DCs . A low ability for colonization in the intestine of B . breve might correlate with the fact that oral administration of this bacterium did not induce apparent change in the composition of commensal microbiota . Tr1 cells were identified as the second subset of CD4+ regulatory T cells [50] . Both Foxp3+ Treg cells and Tr1 cells are critically involved in the maintenance of intestinal homeostasis [30] . In vitro studies demonstrated that IL-10 and IL-27 are critical for the induction of Tr1 cells [51]–[53] . The present study shows that intestinal Tr1 cells are induced by both IL-10 and IL-27 , which is produced by intestinal CD103+ DCs that are exposed to B . breve . However , Tr1 cells are present in the intestinal lamina propria of mice that are not fed with B . breve [34] . In this regard , given that there are many types of Bifidobacterium species in the intestine ( Table S1 ) , these indigenous Bifidobacterium might contribute to development of intestinal Tr1 cells . Indeed , our data suggest that B . longum , one of indigenous commensal bacteria , moderately induced Tr1 cells . The B . breve-induced increase in Tr1 cells was observed in the large intestine , but not in the small intestine . This might be due to the characteristics of B . breve , which preferentially colonize in the large intestine rather than the small intestine [57] . B . breve-induced Tr1 cell development depends on the TLR2/MyD88 pathway . The TLR pathways play mandatory roles in the elimination of pathogenic microorganisms [54] . Previous studies indicated that mice deficient in MyD88 , TLR2 , or TLR4 were highly sensitive to intestinal inflammation induced by dextran sodium sulfate treatment [58] , [59] . However , the mechanism for the TLR-dependent maintenance of gut homeostasis remains unclear . This study demonstrates that the TLR2 pathway in DCs is beneficial for the suppression of intestinal inflammation via induction of IL-10-producing Tr1 cells . It is interesting to note that Tr1 cells are present in Tlr2−/− and Myd88−/− mice , indicating that Tr1 cell development in the intestine in steady states is induced independently of TLR signaling . The TLR-independent induction of intestinal Tr1 cells might be induced by other , so-far unknown , bacteria . Our in vitro experiments clearly indicate that intestinal CD103+ CX3CR1− CD11b− DCs respond to B . breve and promote Tr1 cell development . Intestinal CD103+ DCs residing in the colonic lamina propria and MLN showed enhanced capacity to induce IL-10-producing Tr1 cells after B . breve treatment . CD103+ DCs from MLN were less effective in Tr1 cell induction compared to the lamina proprial CD103+ DCs . Thus , it is possible that CD103+ DCs in MLN and the colonic lamina propria have differential characteristics in Tr1 cell induction . In addition , it remains unclear how CD103+ DCs sense B . breve in the intestinal mucosa . CX3CR1-expressing intestinal DCs have been shown to extend their dendrites into the intestinal lumen to sample luminal contents [60] . However , CD103+ DCs do not express the CX3CR1 that is required for dendrite extension . Several metabolites produced by commensal microbiota have been shown to influence host cell gene expression [61] . However , culture supernatants of B . breve did not induce IL-10 production from T cells co-cultured with CD103+ DCs , indicating that B . breve directly acts on intestinal DCs ( Figure S6 ) . Elucidating how CD103+ DCs recognize B . breve in the intestinal lamina propria would be a future interesting issue . IL-10-producing Tr1 cells can be induced by UV-irradiated B . breve , or even sonicated B . breve ( Figure S7 ) . These findings indicate that components of B . breve directly act on intestinal DCs , possibly by interacting with TLR2 , and promote Tr1 cell development . TLR2 has been shown to recognize a unique polysaccharide structure ( PSA ) of B . fragilis to induce Foxp3+ Treg cells [21] . The probiotic strain of B . breve used in this study also possesses a unique structure of polysaccharide in their cell walls [62] . Therefore , it would be interesting in the future to analyze whether the polysaccharide of B . breve is recognized by TLR2 to induce Tr1 cells . Identification of such B . breve components that activate the TLR2 pathway will lead to development of a new effective agent for the treatment of IBD . In contrast to the development of Tr1 cells promoted by B . breve , L . casei did not have any effect on the differentiation of intestinal T cells , although it is well known as a beneficial probiotic strain possessing several health-promoting effects [63] . In this regard , several mechanisms of action of probiotics , other than the influence on the host T cell development , have been postulated [7] , [8] . These include enhancement of barrier functions of epithelial cells , modification of commensal flora , and effects on dendritic cells and monocytes/macrophages . Several mechanisms of Lactobaciluus species-mediated actions have been reported [55] , [56] , [64] . Our results indicate that each probiotic strain has their specific modes of action on the host . VSL#3 containing several probiotics ( three bifidobacteria , five lactobacilli and Streptococcus salivarius subsp . thermophilus ) have been reported to have potent effects on host health and diseases [26] . This might be due to the synergistic effect of these different probiotic strains that have distinct mechanisms of actions . In the present study , we show that a probiotic bacterium , B . breve , induces intestinal Tr1 cells and thereby improves intestinal inflammation . Analysis of the effect of this probiotic-dependent Tr1 cell development on other disease models will expand the application of B . breve as a therapeutic agent .
All animal experiments were carried out in strict accordance with the Guidelines for Animal Experimentation of the Japanese Association for Laboratory Animal Science . The protocol was approved by the committee for Animal Experiments of Osaka University ( Permit Number: 21-058-0 ) . Lactobacilus casei strain Shirota ( L . casei ) and Bifidobacterium breve Yakult strain ( B . breve ) were as described [35] , [40] . B . bifidum ( Yakult strain YIT10347 ) , B . adolescntis ( ATCC15703 ) , and B . longum ( ATCC15707 ) were used for experiments . For oral treatment of mice , freeze-dried preparations of L . casei and B . breve were dissolved with distilled water , and 1×109 bacteria were administered . A sachet of B . breve contained 4×109 freeze-dried living bacteria , cornstarch , and hydroxypropyl cellulose as vehicle . Placebo sachet of B . breve contained only cornstarch and hydroxypropyl cellulose . A sachet of L . casei consisted of 5×109 freeze-dried living bacteria with lactose , cornstarch , powdered skim milk , crystallized cellulose and hydroxypropyl cellulose . The placebo sachet of L . casei consisted of only common excipients . For in vitro stimulation , B . breve was inoculated in GAM broth ( Nissui Pharmaceutical ) supplemented with 1% ( w/v ) glucose , and cultured for 24 h at 37°C under anaerobic conditions , and then centrifuged and the pellets were suspended with culture media . The number of B . breve was measured by culturing on MRS agar plate . Neutralizing anti-mouse IL-10 was purchased from BD biosciences , anti-mouse IL-27p28 , and anti-TGF-β ( 1D11 ) blocking antibodies were purchased from R&D systems . Anti-mouse CD3 ( 145-2C11 ) and CD28 ( 37 . 51 ) were obtained from BioLegend . LE540 was purchased from WAKO Chemicals ( Tokyo , Japan ) . BALB/c and C57BL/6J mice were purchased from CLEA Japan or Japan SLC . CB17-SCID mice were obtained from CLEA Japan . Il10−/− , Foxp3eGFP were purchased from Jackson laboratories , and Myd88−/− , Tlr2−/− , Tlr4−/− and Tlr9−/− mice were generated previously [65] . Il27ra−/− mice were kindly provided by Amgen [66] . These mice were backcrossed eight or more generations onto BALB/c or C57BL/6J . C57BL/6J mice were orally administered with L . casei or B . breve ( 109 bacteria each ) as well as placebo daily with gastric tubes for 3 months . Alternatively , probiotics were orally introduced into BALB/c , C57BL/6J , or Foxp3eGFP mice for 1–4 weeks . Methods for the analysis of fecal bacteria are described in Text S1 . For flow cytometry , the following antibodies were used: PerCP/Cy5 . 5-conjugated anti-CD4 ( GK1 . 5 ) , Alexa Fluor 647-conjugated CD11c ( N418 ) , CD62L ( MEL-14 ) , streptavidin-conjugated PE/Cy7 from BioLegend , FITC-conjugated anti-CD11b ( M1/70 ) , CD25 ( 7D4 ) , and PE-conjugated CD103 , anti-mouse CD16/32 ( Fcγ III/II receptor ) from BD PharMingen , PE-conjugated CD44 ( IM7 ) , Alexa Fluor 647-conjugated Foxp3 ( FJK-16s ) , and biotin-conjugated CX3CR1 from eBiosciences . Flow cytometric analysis was performed using a FACS Canto II flow cytometer ( BD Biosciences ) with FlowJo software ( Tree Star ) . Lamina propria DCs and lymphocytes were isolated as previously described [67] with simple modifications . Briefly , colons and small intestines were opened longitudinally and vigorously rinsed in PBS . Intestines were shaken in HBSS containing 5 mM EDTA and 5% fetal bovine serum ( FBS ) for 20 min at 37°C . After removal of epithelial layers and fat tissues , the intestines were cut into small pieces and incubated with RPMI 1640 containing 5% FBS , 1 mg/ml of collagenase D ( Roche Diagnostics ) , 1 mg/ml of dispase ( Invitrogen ) and 40 µg/ml of DNase I ( Roche Diagnostics ) for 1 h at 37°C in a shaking water bath . The digested tissues were washed with HBSS containing 5 mM EDTA . Cell suspensions were filtered through a 40 µm cell strainer into chilled PBS and centrifuged . Cell suspensions from enzyme digestion were then applied to a Percoll ( GE Healthcare ) gradient ( for DCs: 30% percoll on top , 75% percoll on the bottom , and for lymphocytes: 40% percoll on top , 80% percoll on the bottom ) by centrifugation at 780 g for 20 min at 25°C . The cells at interface were taken and washed twice with FACS buffer . For purifying lamina propria DC subsets , single cell suspensions were treated with anti-mouse Fcγ receptor antibody for 5 min at 4°C . Cells were then stained with CD11c-APC , CD11b-FITC , CD103-PE and CX3CR1-PE-Cy7 and subsequently sorted using a FACSAria ( BD Biosciences ) to a purity >98% . The cells were used immediately for each of experiment . To prepare single-cell suspensions from spleens , they were ground between glass slides and passed through a 40 µm cell strainer . Splenocytes were treated with RBC lysis buffer ( 0 . 15 M NH4Cl , 1 mM KHCO3 , 0 . 1 mM EDTA ) for 5 min and washed twice with PBS . For FACS sorting , cells were stained with PerCP/Cy5 . 5-conjugated anti-CD4 ( Biolegend ) , APC-conjugated anti-CD62L , FITC-conjugated anti-CD25 and PE-conjugated anti-CD44 ( BD Biosciences ) . Naïve CD4+ T cells were sorted using a FACSAria for CD4+CD62LhighCD25−CD44low . The purity of the sorted cells was routinely >98% . The intracellular expression of IFN-γ , IL-17 , and IL-10 in CD4+ T cells was analyzed using the Cytofix/Cytoperm Kit Plus ( with Golgistop; BD Biosciences ) according to the manufacturer's instructions . In brief , lymphocytes obtained from the intestinal lamina propria were incubated with 50 ng/ml of phorbol myristate acetate ( PMA; Sigma ) and 5 µM of calcium ionophore A23187 ( Sigma ) and Golgistop in complete RPMI1640 at 37°C for 4 h . Surface staining was performed with PerCP/Cy5 . 5-conjufated anti-CD4 for 20 min at 4°C . After Fix/Perm treatment for 20 min , intracellular cytokine staining was performed with PE-conjugated anti-IL-10 , FITC-conjugated anti-IFN-γ , and APC-conjugated anti-IL-17 for 20 min . Data were acquired using a FACS Canto II and analyzed using FlowJo software . Alternatively , for intracellular staining for Foxp3 and IL-10 , cells were stained using the Foxp3 Staining Buffer set ( eBiosciences ) . Colonic DC subsets ( 5×104 ) were incubated with the same number or the indicated number of L . casei or B . breve in 100 µl of complete RPMI1640 media for 24 h in a round-bottom 96 well plate . DCs were then washed with PBS and naïve CD4+ T cells ( 5×104 ) were added into the culture with 2 µg/ml soluble anti-CD3 mAb . After 4 days , T cells were collected , washed and counted . The same numbers of T cells were re-stimulated with plate-bound anti-CD3 mAb ( 2 µg/ml ) and soluble anti-CD28 mAb ( 2 µg/ml ) for 24 h . Re-stimulated T cell cytokine production in the supernatants was analyzed by ELISA ( R&D systems ) . Alternatively , T cells were re-stimulated with 50 ng/ml of PMA and 5 µM of calcium ionophore A23187 for 6 h before intracellular cytokine staining was performed as described above . Golgistop was added for the last 2 h . Total RNA was isolated with the RNeasy Mini Kit ( Qiagen ) , and 1–2 µg of total RNA was reverse transcribed using M-MLV reverse transcriptase ( Promega ) and random primers ( Toyobo ) after treatment with RQ1 DNase I ( Promega ) . Complementary DNAs were analyzed by qPCR using the GoTaq qPCR Master Mix ( Promega ) on an ABI 7300 system ( Applied Biosystems ) . All values were normalized to the expression of Gapdh encoding glyceraldhyde-3-phosphate dehydrogenase , and the fold difference in expression relative to that for Gapdh is shown . Amplification conditions were: 50°C ( 2 min ) , 95°C ( 10 min ) , and 40 cycles of 95°C ( 15 s ) and 60°C ( 60 s ) . The following primer sets were used: cMaf , 5′-AATCCTGGCCTGTTTCACAT-3′ and 5′-TGACGCCAACATAGGAGGTG-3′; Il21 , 5′-GCCAGATCGCCTCCTGATTA-3′ and 5′-CATGCTCACAGTGCCCCTTT-3′; Il27p28 , 5′-TTCCCAATGTTTCCCTGACTTT-3′ and 5′-AAGTGTGGTAGCGAGGAAGCA-3′; Ebi3 , 5′-TGAAACAGCTCTCGTGGCTCTA-3′ and 5′-GCCACGGGATACCGAGAA-3′; Il10 , 5′-TTTCAAACAAAGGACCAG-3′ and 5′-GGATCATTTCCGATAAGG-3′; and Gapdh , 5′-TGTGTCCGTCGTGGATCTGA-3′ and 5′-CCTGCTTCACCACCTTCTTGA-3′ Naive CD4+CD62LhighCD25−CD44low splenic T cells from BALB/c mice or Il10−/− mice ( BALB/c background ) were purified and intraperitoneally transferred into SCID mice ( 3×105 cells per mouse ) . B . breve ( 109 bacteria ) were fed by oral gavage from 3 days before the transfer to the end of the experiments . Weight changes were monitored every day . The mice were sacrificed , and the colons were examined histochemically after haematoxylin and eosin staining . Alternatively , the colons were cut into small pieces after wash and cultured for 24 h . Then , culture supernatants were collected and the level of IL-10 , IL-17A and IFN-γ was measured by ELISA ( R&D systems ) . Paraffin-embedded colon samples were sectioned and stained with hematoxylin and eosin . Severity of colitis was evaluated by the standard scoring system as previously described [68] . Five regions of the colon ( cecum; ascending , transverse , and descending of colon; and rectum ) were graded semiquantitatively from 0 ( no change ) to 5 ( most severe change ) . The grading represents an increasing incidence and degree of inflammation , goblet cell loss , ulceration and fibrosis in the lamina propria . The scoring was performed in a blinded manner . Images of hematoxylin and eosin staining and May-Grunwald-Giemsa staining were taken using Biozero ( Keyence ) . Statistical analysis was performed using PRISM 4 software . Unpaired student's t-test and Mann-Whitney U test were used to determine the significance of experiments . P values of less than 0 . 05 were considered statistically significant . | Unlike induction of Foxp3+ regulatory T cell development , it remains unclear how intestinal environmental factors regulate development of another regulatory T cell subset , Tr1 cells that produce IL-10 . In this study , we reveal that a probiotic strain , Bifidobacterium breve induces IL-10-producing Tr1 cells that express c-Maf , IL-21 , and Ahr via activation of intestinal CD103+ DCs in the large intestine . Using several gene-targeted mice , we show that B . breve-induced development of IL-10-producing Tr1 cells is dependent on DC secretion of IL-10 and 27 via a TLR2/MyD88 pathway . We finally show that B . breve ameliorated T cell-dependent colitis in immunocompromised mice via T cell production of IL-10 . These findings demonstrate that B . breve maintains intestinal homeostasis through the induction of intestinal IL-10-producing Tr1 cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"gastroenterology",
"and",
"hepatology"
] | 2012 | Probiotic Bifidobacterium breve Induces IL-10-Producing Tr1 Cells in the Colon |
Many types of epigenetic profiling have been used to classify stem cells , stages of cellular differentiation , and cancer subtypes . Existing methods focus on local chromatin features such as DNA methylation and histone modifications that require extensive analysis for genome-wide coverage . Replication timing has emerged as a highly stable cell type-specific epigenetic feature that is regulated at the megabase-level and is easily and comprehensively analyzed genome-wide . Here , we describe a cell classification method using 67 individual replication profiles from 34 mouse and human cell lines and stem cell-derived tissues , including new data for mesendoderm , definitive endoderm , mesoderm and smooth muscle . Using a Monte-Carlo approach for selecting features of replication profiles conserved in each cell type , we identify “replication timing fingerprints” unique to each cell type and apply a k nearest neighbor approach to predict known and unknown cell types . Our method correctly classifies 67/67 independent replication-timing profiles , including those derived from closely related intermediate stages . We also apply this method to derive fingerprints for pluripotency in human and mouse cells . Interestingly , the mouse pluripotency fingerprint overlaps almost completely with previously identified genomic segments that switch from early to late replication as pluripotency is lost . Thereafter , replication timing and transcription within these regions become difficult to reprogram back to pluripotency , suggesting these regions highlight an epigenetic barrier to reprogramming . In addition , the major histone cluster Hist1 consistently becomes later replicating in committed cell types , and several histone H1 genes in this cluster are downregulated during differentiation , suggesting a possible instrument for the chromatin compaction observed during differentiation . Finally , we demonstrate that unknown samples can be classified independently using site-specific PCR against fingerprint regions . In sum , replication fingerprints provide a comprehensive means for cell characterization and are a promising tool for identifying regions with cell type-specific organization .
In mammals , replication of the genome occurs in large , coordinately firing regions called replication domains [1]–[7] . These domains are typically one to several megabases , roughly align to genomic features such as isochores , and are closely tied to subnuclear position , with transitions to the nuclear interior often coupled to earlier replication , and transitions to the periphery to later replication [4] , [5] , [8] , [9] . Given their connections to subnuclear position and remarkably strong correlation to chromatin interaction maps [3] , replication profiles provide a window into large-scale genome organization changes important for establishing cellular identity . The organization of replication domains is cell-type specific , and a larger number of smaller replication domains is a property of embryonic stem cells ( ESCs ) [3]–[5] . Importantly , in both humans and mice , induced pluripotent stem cells ( iPSCs ) reprogrammed from fibroblasts display a timing profile almost indistinguishable from ESCs , suggesting that replication profiles may also be used to measure cellular potency [3] , [5] . While a wide-range of cell classification methods are actively used , the most common practice for verifying identity is to monitor a handful of molecular markers , some of which are shared with other cell types . Genome-wide classification of features such as DNA methylation [10]–[12] , transcription [13] , [14] and histone modifications [15] , [16] have in principle more potential to accurately distinguish specific cell types . However , these features of chromatin are highly dynamic at any given genomic site [17] , and most measurements require high-resolution arrays and costly antibodies . Moreover , recent reports highlight the unstable nature of transcription and related epigenetic marks in multiple embryonic stem cell lines [18] , [19] . By contrast , since replication is regulated at the level of large domains , replication profiles are considerably less complex to generate and interpret than other molecular profiles . Timing changes occurring during differentiation are on the order of several hundred kilobases and are highly reproducible between various stem cell lines [3]–[5] . They are also robust to changes in individual chromatin modifications , retaining their normal developmental pattern in G9a ( −/− ) cells despite strong upregulation of G9a target genes and near-complete loss of H3K9me2 [8] . Here , we describe a method for classifying cell types—replication fingerprinting—based on genome-wide replication timing patterns in mouse and human ESCs and other cell types . We applied the method to 67 ( 36 mouse and 31 human ) whole-genome replication timing datasets to demonstrate the feasibility of classifying cell types using a minimal set of cell type-specific regions . After identification , these regions were used to classify two independent samples using site-specific PCR . We also demonstrate that loss of pluripotency is accompanied by consistent changes in replication timing , implicating the replication program as an important factor in maintaining pluripotency and revealing a novel fingerprint for pluripotent stem cells .
In addition to our previously reported replication profiles , BG02 hESCs were differentiated to mesendoderm and definitive endoderm as previously described [20] , as well as ISL+ mesoderm and smooth muscle cultured in defined medium ( Methods ) , and profiled for replication . Replication profiles were generated as described previously [3]–[5] , [21] . In brief , nascent DNA fractions were collected in early and late S-phase , differentially labeled , and co-hybridized to a whole-genome CGH microarray . The ratio of early and late fraction abundance for each probe—“replication timing ratio”—represents its relative time of replication . Values from individual probes are then smoothed using LOESS ( a locally weighted smoothing function ) , and plotted on log scale ( Figure 1 ) . Replication profiles generated in this way are freely available to view or download at www . ReplicationDomain . org [22] , and those analyzed in this report are summarized in Table S1 . Figure 1 illustrates the basic concept of replication fingerprinting . Two exemplary profiles each for D3 embryonic stem cells ( ESCs; blue ) and D3 ESC-derived neural precursor cells ( NPCs; green ) are overlaid . Given that most of the genome is conserved in replication timing between any two cell types ( e . g . 80% conserved between ESCs and NPCs [4] ) , the first challenge is to choose genomic regions that are differentially replicated within a set of cell types . We define a “replication fingerprint” of a cell type as a set of genomic regions useful for classification , along with their associated replication timing values . For a simplified example , we show exemplary fingerprint regions for a segment of chromosome 7 ( Figure 1A , gray bars ) . Note that the four regions change dramatically upon differentiation to neural precursors ( e . g . , ESC2 vs . NPC1; Figure 1A , B ) , but have replication timing values that are well conserved between replicate experiments ( e . g . , ESC1 vs . ESC2 ) . We and others have observed similarly widespread changes in replication profiles between any two different cell types profiled to date [1] , [3]–[5] , [7] . As classification methods require a measure of distance between samples , we defined the distance between replication profiles as the sum of absolute differences in replication timing in fingerprinting regions ( Figure 1B ) . To select an optimal set of fingerprinting regions we maximize a “distance ratio , ” representing the ratio of the average distance between unlike cell types to the average distance between equivalent cell types ( Figure 1C ) . This ratio is maximized by selecting regions that are consistently different in replication timing between different cell types , but consistently similar between equivalent types . Importantly , the assignment of unlike vs . equivalent cell types is user-defined and flexible , allowing selection of features that best distinguish any group of cells from any other , such as ESCs from NPCs , normal from disease-related cells , or pluripotent from committed cells . While Figure 1 shows a simplified example of four regions distinguishing ESCs from NPCs , real-world classification requires the ability to make distinctions genome-wide between many cell types , making manual selection of regions impractical . Therefore , to make the method generally applicable , we developed an automated algorithm based on Monte Carlo sampling [23] to select regions that best distinguish between all available cell types in genome-wide replication datasets . Alternative approaches evaluated for feature selection and classification included Bayesian networks , nearest neighbor methods , decision trees and SVMs , which were comparably successful only for smaller collections of cell types . We chose to explicitly maximize distances between cell types in the method described here in anticipation of translating cell classification to more convenient empirical assays with a limited number of features , because larger timing differences are easier to verify empirically and are more robust to experimental and biological variation . In practice , replication fingerprinting is a feature selection problem . Although most genome-wide approaches are both simple and comprehensive , we found that genome-wide correlations and distances , while a good first approximation of the relatedness between cell types , are not ideal for classification as the small amount of noise in regions with conserved replication timing is compounded over this relatively large fraction of the genome ( Figure S1 ) . We therefore wish to exclude domains that are noisy ( having high technical or biological variability ) , irrelevant ( conserved in all cell types ) , or redundant ( containing overlapping information ) . To achieve this , we first remove regions with conserved replication timing between cell types , resulting in a set of informative regions that can be further optimized by a Monte Carlo selection algorithm . Figure S2 depicts the Monte Carlo algorithm . To reduce noise from individual probe measurements , replication profiles are first averaged into nonoverlapping windows of approximately 200 kb . This window size represents a balance between sizes of the regions that change replication timing during development ( 400–800 kb ) , and the number of probes needed for timing changes to be deemed statistically significant ( 35–180 probes are contained in each window depending on the probe density of the array platform; see Methods , Table S2 ) . An initial set of regions with the highest replication timing changes in the set of replication profiles are chosen to exclude regions with conserved replication timing , and half of these starting regions are randomly selected to calculate initial distances between cell types . At each iteration of the algorithm , a region can be added to the set of fingerprint regions , removed from the set , or swapped with an unused region . Using a Metropolis-Hastings criterion [23] , [24] , moves that improve the overall distance ratio are accepted with higher probability than those that do not; after 20 , 000 or more such moves , a final set of fingerprinting regions is selected . As depicted in Figure 2 , the fingerprinting algorithm selects domains with large and reproducible replication timing differences between cell types , discarding those with minimal or variable changes . Before selecting optimal regions ( Figure 2A , C ) , the average distance between “like” and “unlike” cell types are similar , translating into classification errors for randomly selected domains ( Figure 2C ) as well as the whole genome ( Figure S1 , red shading ) . After selection , the separation in distances between like and unlike types becomes very distinct ( Figure 2B , D ) , even for closely related cell types ( Figure 3 ) . These regions similarly highlight distinctions between cell types both in correlations ( Figure S3 , S4 , S5 , S6 , S7 , S8 ) , and distance matrices between cell types ( Figure S9 , S10 , S11 , S12 ) . Since Monte Carlo selection is stochastic , different sets of fingerprinting regions can be selected in different runs . To evaluate the stability of regions included in replication fingerprints , we applied the algorithm 100 times for each type of human and mouse fingerprint constructed ( Figure S13 ) . Results demonstrate that fingerprinting regions are well-conserved among multiple rounds of selection , with the top 10–14 regions selected in 100/100 trials in each case . For all subsequent classification , we used regions included in at least 75/100 fingerprinting runs . As the distances between profiles derive from either the same or different cell types ( Figure 2C ) , their distributions can be used to create a general classifier ( Figure 2C , D , Figure 3A ) , with an error rate proportional to the overlap in distances shared by “like” and “unlike” cell type comparisons ( Figure 2C , D , blue shading ) . This allows us to state a level of confidence for a given prediction , as well as estimate the similarity of a cell type to others . To refine this classification , we applied the k-nearest-neighbor rule [25] ( kNN; k = 3 ) to assign cell types according to the three most similar profiles in the training set . Distances below the threshold – θ = 2 . 4 in Figure 2D – are hypothesized to derive from similar cell types , and are used with kNN to classify profiles according to the closest profiles in the training set . Distances above the threshold are presumed to derive from different cell types , preventing kNN from classifying highly divergent RT profiles as the cell type of the most similar known profile . To test the ability of our method to select suitable regions for classification , we applied it to predict the known identity of 9 mouse and 7 human cell types with 36 and 31 total experimental replicates , respectively . Datasets used for prediction are summarized in Table S1 , with most described in detail in previous publications [3]–[5] . Rough classification of each experiment into like and unlike cell types by a distance ratio cutoff was accurate in 951/961 ( 99 . 0% ) human and 1250/1296 ( 96 . 5% ) mouse comparisons respectively ( Figure 3A , B ) . Refining this classifier by using kNN to assign cell types according to the three most similar profiles in the training set resulted in correct predictions for 36/36 mouse and 31/31 human replication timing profiles ( Figure 3C , D ) . Strikingly , even closely related cell types could be reliably distinguished using this method , such as mouse ESCs and early primitive ectoderm-like stem cells ( EPL/EBM3 ) , and two day intermediates of human ESC differentiation into endomesoderm ( DE2; day 2 ) and definitive endoderm ( DE4; day 4 ) . Thus , replication profiles appear capable of distinguishing among a wide array of cell types in early mouse and human development . The use of all experimental data in a selection algorithm often results in overfitting the model to a limited set of observations . For this reason , machine-learning algorithms are commonly trained and tested on different subsets of data ( termed cross-validation ) . To determine whether overfitting is occurring in our selection method and assess the degree to which fingerprinting domains are generally cell type-specific , we performed leave-one-out cross-validation ( LOOCV ) with each of the available experiments by constructing fingerprints using all but one experimental replicate , and testing classification on the remaining replicate . In all cases ( 31/31 human , 36/36 mouse ) , correct predictions in the excluded profile confirmed that fingerprinting regions remained consistent with cell type , and that most cell-line-specific differences were discarded ( Figure 3C , LOOCV column ) . This was also true for a cell line with only one replicate ( mouse 46C neural precursor cells ) , implying that most of the regions of differential replication timing useful for classification are shared between cell lines . To simulate the classification of a cell type not yet encountered in the training set , we tested predictions after selecting fingerprinting regions with all replicates of a given cell type excluded ( Figure 3C , LCTO column ) . This confirmed that most cell types not yet encountered were correctly classified as “Unseen” ( 7/7 cell types in human , 7/9 in mouse ) . However , two cases in which profiles were ambiguous were between neural precursors ( NPCs ) and mouse epiblast-like stem cells ( EpiSCs , EBM6 ) , suggesting that closely related cell types are more accurately distinguished when examples of each type are included in the training set . One of the most striking features of replication timing is its widespread consolidation into larger replication domains during neural differentiation , concomitant with global compaction of chromatin [3] , [4] . This consolidation , along with recovery of ESC replication timing by induced pluripotent stem cells ( iPSCs ) , suggested that replication patterns in specific regions of the genome are associated with the pluripotent state . Further , if certain timing changes are a stable property of cellular commitment , they may provide a unique opportunity to evaluate differentiation capacity using replication-timing patterns . To explore this , we analyzed the differences in replication profiles between collections of pluripotent/reversible ( ESCs , iPSCs , EBM3 ) and committed cell types in 13 human and 21 mouse cell lines ( Figure 4A ) . In each case , we created a stringent consensus fingerprint for classification consisting of regions found in >75/100 runs ( 18 regions each in mouse and human ) , and examined genes in the top 200 fingerprint regions ( ∼2% of the genome ) to characterize a more inclusive sample . Genes and regions found to consistently switch to earlier or later replication as pluripotency is lost are provided in Tables S3 , S4 , S5 , S6 . Strikingly , several regions displayed conserved , significant differences in timing between all pluripotent and committed cell types ( Figures 4A , S10 , S12 ) . As with general fingerprints , classification into pluripotent or committed types could be performed unambiguously ( 36/36 cases in mouse , 31/31 in human ) , even with regions selected with the test profile excluded ( LOOCV column ) . Several of the genes consistently switching to later replication in mouse and human pluripotency fingerprints have known roles in maintaining pluripotency ( for instance , Dppa2 and Dppa4 in both species , and DKK1 in human; Tables S4 and S6 ) . In addition , two classes of genes stood out from this analysis that showed significant switches to later replication in both species: a large cluster of protocadherins ( PCDs ) , and the majority of the Hist1 cluster of core histone genes ( Table S7 ) . The former are developmentally regulated genes with broad involvement in neural development and cell-cell signaling [26] , [27] , and switch to later replication in all committed mouse and human cell types . The latter Hist1 cluster was later replicating in 8/8 committed cell types in mouse and 5/6 in human ( not lymphoblasts ) , and includes several core histone genes that were downregulated up to 2 . 5-fold in NPCs . These results are intriguing in light of previous reports of histone downregulation during development [28] , as well as a hyperdynamic chromatin phenotype in ESCs that involves higher exchange rates of histone H1 [29] and is required for efficient somatic cell nuclear reprogramming in Xenopus oocytes [30] . Importantly , all of the histone H1 genes are found in this cluster , suggesting that regulation of global H1 abundance may provide a mechanism for the overall chromatin compaction and consolidation of replication timing observed during neural differentiation [3]–[5] . To characterize the genes included in the mouse pluripotency fingerprint , we compared them to a previous class of genes that showed lineage-independent switches to later replication in mouse ESC differentiation , and failed to revert to ESC-like expression in three separately derived samples of partial iPSCs ( clusters 15 and 16 in Figure 7 of Hiratani et al . , 2010 ) . Remarkably , 200 out of 217 genes in the top 100 mouse pluripotency regions belonged to this class , despite very different methods for deriving them ( Figure 5A ) . All of the fingerprint genes switched to later replication , and at the transition between early and late epiblast stages where cell fates become restricted [5] . Most genes also had reduced expression in late epiblast and neural progenitor stages ( average 1 . 66-fold reduction in transcription from ESC/EBM3 to EBM6/NPCs ) . Thus , some of these genes may make prime candidates for improving the efficiency of iPSC production , or for reverting human ESCs to a more naive , mouse ESC-like state . However , the overlap between human and mouse pluripotency fingerprint genes , while significant , was much lower ( Figure 5A ) , and this was true even when comparing human ESCs to developmentally analogous mouse EpiSCs [3] , [31] . Therefore , many pluripotency-associated genes and loci may be species-specific , consistent with recent studies that underscore considerable differences between mouse and human pluripotency networks [32] , [33] . This low alignment is also accounted for by a general drop in overall alignment in regions with the greatest developmental switches in replication timing ( Figure 5B ) , which are those preferentially selected by the fingerprinting algorithm . Of the genes conserved in the fingerprints of both species ( indicated by boldface type in Tables S4 and S6 ) , most belong to the aforementioned large class of protocadherins . However , Dppa2 and Dppa4 are also conserved , as well as genes with no known roles in maintaining pluripotency ( Cast , Riok2 , Lix1 ) that reside within the same replication units as pluripotency fingerprint genes in both species . Other core pluripotency genes remain relatively early replicating in both species ( Pou5f1[Oct4] , Sox2 , Nanog ) , and are likely regulated by other mechanisms . For instance , Sox2 belongs to a class of genes with strong promoters ( HCP , or high CPG content promoters ) generally unaffected by local replication timing [4] , [34] . One potential application of replication fingerprints is in the development of PCR kits for epigenetic classification , particularly for cell types or disease samples with no known aberrations in transcription or sequence . To confirm that fingerprint regions can be translated into a classification scheme using site-specific PCR , we classified two unknown samples representing cell types that were analyzed previously , but that were derived from different cell lines than the original set used for training . The experiment was performed in a blind manner in which the experimenter had no prior knowledge of the regions or cell types being tested . Primers were assembled against sequences within 10–20 kb from the center of each fingerprint region , and the replication times of each region were quantified as the “relative early S phase abundance” ( relative abundance of a sequence in nascent strands from early S phase ) , as previously described [35] ( Figure 6A ) . PCR-based timing values were rescaled for consistency with the original scale of the array datasets used in training , and distances were calculated between the unknown samples and other human profiles in fingerprint regions ( Figure 6B ) . Using the same methods as in prior classifications , these distances correctly identified the two unknown samples as lymphoblasts and hESCs , respectively; the three known datasets with the smallest distances were each of the correct cell type .
Our method for cell typing through replication fingerprinting addresses a well-recognized need for comprehensive methods to assess cellular identity and differentiation potential in stem cell biology . Unlike other molecular markers , replication is regulated at the level of large , multi-megabase domains , making comprehensive , genome-wide profiles relatively simple to generate and interpret [36] . In particular , the robust stability of replication timing profiles in stem cells [8] , and wide divergence between cell types make them a promising candidate for classification . While the functional role for the replication program is not yet understood , its conservation between human and mouse cell culture models of development support its functional significance . We and others have shown a substantial correlation ( R2 = 0 . 42–0 . 53 ) in replication patterns between mouse and human cell types , with timing patterns of embryonic stem cells , neural precursor cells , and lymphoblastoid cells most closely aligned to their cognate in the other species [1] , [3] . The important role for replication is further corroborated by its remarkably strong link to genome organization [3] , and its ability to confirm the mouse epiblast identity of human ESCs genome-wide and with an epigenetic property [3] , [31] . By comparison , methods for cell typing using DNA methylation , gene expression , histone modifications or protein markers are well suited to some applications [10]–[16] , but may not be informative for certain fractions of the genome , or may rely on genome features that cannot distinguish between similar cell states . We therefore envision replication fingerprinting as a complement to existing cell typing strategies that may be used for samples unsuitable for traditional methods , or for additional confidence in assessing cell identity in cases where this is critical , such as regenerative medicine . One caveat to consider in these applications is that replication profiles , similar to other genome-wide methods , are an ensemble aggregate from many cells , making measurement of homogeneity difficult . In addition , as with other supervised classification approaches , the method is informative only for cell types ( classes ) available during training . However , our fingerprinting method is in principle applicable to any data type , and may be modified to select discriminating features in other epigenetic profiles . A major advantage of our fingerprinting method is in selection of a minimal set of regions that allow for classification with a straightforward PCR-based timing assay and a reasonably small set of primers , particularly if only cell-type specific regions are examined . Our results suggest that a standard set of 20 fingerprint loci can be effective for classification , but the number of regions queried can be adjusted based on the confidence level required . The sole requirement for replication profiling is the collection of a sufficient number of proliferating cells for sorting on a flow cytometer . Consistently , just as replication fingerprints can be generated for particular cell types or general categories of cells , features of replication profiles allow for the creation of disease-specific fingerprints , which may be valuable for prognosis . In addition to cell typing applications , replication profiling is informative for basic biological questions . Here , we have identified regions that may undergo important organizational changes upon differentiation , which include a class of gene that fail to reverse expression in partial iPSCs , and the majority of mouse and human histone H1 genes . Human lymphoblasts retained early replication in H1 genes , which may be explained by their high rate of proliferation . Since highly developmentally plastic regions ( including pluripotency fingerprint regions ) are poorly conserved ( Figure 5B ) the evolutionary conservation of cell-type specific timing patterns must be driven by the moderately changing majority of the genome . The recent derivation of mouse ESC-like human stem cells with various methods raises an intriguing question [37]: will naïve hESCs align better to mESCs than to mEpiSCs for replication timing as they have for transcription ? Although pluripotency is currently assessed by marker gene expression or laborious complementation experiments , replication timing assays in regions uniquely early or late replicating in pluripotent cells provide a tractable method to predict the pluripotency of various cell types , as well as insights into conserved genome organizational changes during differentiation .
Mouse replication timing datasets are described in Hiratani et al . , 2010 . Briefly , mouse embryonic stem cells ( ESCs ) from D3 , TT2 , and 46C cell lines were subjected to either 6-day ( 46C ) or 9-day ( D3 , TT2 ) neural differentiation protocols to generate neural progenitor cells ( NPCs ) [4] , [5] . For D3 , intermediates were also profiled after 3 ( EBM3 ) and 6 ( EBM6 ) days of differentiation . Muscle stem cells ( myoblast ) and induced pluripotent stem cells ( iPSCs ) reprogrammed from fibroblasts were collected as described for human and mouse [38]–[40] . For human timing datasets , neural precursors were differentiated from BG01 ESCs as described in Schulz et al . , 2004 [3] , [41] . Lymphoblast cell lines GM06990 and C0202 were cultured as previously described [2] , [42] . Differentiation of BG02 hESCs to mesendoderm ( DE2 ) and definitive endoderm ( DE4 ) was performed by switching from defined media ( McLean et al . [20] ) to DMEM/F12+100 ng/ml Activin A 20 ng/ml Fgf2 for two and four days , respectively , with 25 ng/ml Wnt3a added on the first day . Mesoderm and smooth muscle cells were derived by adding BMP4 to DE2 cells at 100 ng/ml . Using custom R/Bioconductor scripts [43] , [44] , microarray data from Hiratani et al . 2008 , Hiratani et al . 2010 , and Ryba et al . , 2010 were normalized to equivalent scales , and averaged in nonoverlapping windows of approximately 200 kb . Additional profiles for human ESCs , definitive endoderm , mesendoderm , mesoderm , and smooth muscle were derived , normalized and scaled equivalently , as described [45] . Profiles shown in Figure 1 and Figure 4 were smoothed using LOESS with a span of 300 kb . Selection of fingerprint regions was performed as described using custom R/Bioconductor scripts . Regions of non-conserved RT ( 2000/10994 mouse , 2000/12625 human ) were first selected based on standard deviation , then optimized using a Monte Carlo algorithm ( Figure S2 ) . Using the Metropolis-Hastings criterion for Monte Carlo with simulated annealing [23] , [24] , moves are accepted when exp ( ( dRbest−dR ) /T ) >i , where dR is the distance ratio of the proposed move , dRbest is the current best distance ratio , T is a temperature parameter that decreases geometrically during the simulation , and i is a random number from 0 to 1 . Cell type classification was performed using absolute distances between experiments measured from replication timing in fingerprint regions , using the k-nearest neighbor rule with k = 3; i . e . , each profile was categorized according to the three nearest profiles . Crossvalidation was performed to select an appropriate value for k , with k = 3 chosen as the smallest value that yielded 100% classification accuracy after leave-one-out crossvalidation ( LOOCV ) to allow classification of cell types with fewer replicates . For LOOCV results , each experiment was sequentially left out during Monte Carlo selection , and the resulting regions were used to predict the identity of the excluded experiment . To test prediction on cell types not yet encountered , all profiles for a given cell type were left out during region selection ( LCTO ) , and cell type was predicted using the resulting regions . All data analysis was performed using custom R scripts and Bioconductor packages [43] , [44] . For each fingerprint region depicted in Table 1 , 10–20 kb from the center of the region was sent to NCBI Primer-Blast ( http://www . ncbi . nlm . nih . gov/tools/primer-blast/ ) to design several PCR primer sets with product sizes of 150–350 bp , using standard parameters . Forward and reverse primer pairs displaying the greatest specificity were chosen . Primer sets were verified for specificity and product size using the In-Silico PCR tool at the UCSC genome browser ( http://genome . ucsc . edu/cgi-bin/hgPcr ) . PCR reactions were set up using 1 . 25 ng genomic DNA and 1 uM each of forward and reverse primers in 12 . 5 uL scaled according to the instructions of Crimson Taq DNA Polymerase ( NEB ) . Thirty six cycles of PCR ( empirically determined to be unsaturated for amplification ) were performed according to manufacturer's conditions with annealing temperature of 62°C . One-third of the reaction was analyzed on a 1 . 5% agarose gel containing ethidium bromide . The gel was scanned by Typhoon Trio ( GE Healthcare ) and band intensity was quantified by Image Quant TL ( GE Healthcare ) . After the background was subtracted , signal intensity from the early S fraction was divided by the sum of those from early S and late S fractions from each sample , as described [35] . PCR timing values were converted to array RT scale ( equivalent root-mean-square ) using the scale function in R , and distances were calculated against other cell types as previously performed . GSE18019 , GSE20027 | While continued advances in stem cell and cancer biology have uncovered a growing list of clinical applications for stem cell technology , errors in indentifying cell lines have undermined a number of recent studies , highlighting a growing need for improvements in cell typing methods for both basic biological and clinical applications of stem cells . Induced pluripotent stem cells ( iPSCs ) —adult cells reprogrammed to a pluripotent state—show great promise for patient-specific stem cell treatments , but more efficient derivation of iPSCs depends on a more comprehensive understanding of pluripotency . Here , we describe a method to identify sets of regions that replicate at unique times in any given cell type ( replication timing fingerprints ) using pluripotent stem cells as an example , and show that genes in the pluripotency fingerprint belong to a class previously shown to be resistant to reprogramming in iPSCs , identifying potential new target genes for more efficient iPSC production . We propose that the order in which DNA is replicated ( replication timing ) provides a novel means for classifying cell types , and can reveal cell type specific features of genome organization . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"differentiation",
"algorithms",
"developmental",
"biology",
"stem",
"cells",
"induced",
"pluripotent",
"stem",
"cells",
"dna",
"replication",
"dna",
"cell",
"potency",
"embryonic",
"stem",
"cells",
"biology",
"molecular",
"biology",
"microarrays",
"computer",
... | 2011 | Replication Timing: A Fingerprint for Cell Identity and Pluripotency |
The use of mutagenic drugs to drive HIV-1 past its error threshold presents a novel intervention strategy , as suggested by the quasispecies theory , that may be less susceptible to failure via viral mutation-induced emergence of drug resistance than current strategies . The error threshold of HIV-1 , , however , is not known . Application of the quasispecies theory to determine poses significant challenges: Whereas the quasispecies theory considers the asexual reproduction of an infinitely large population of haploid individuals , HIV-1 is diploid , undergoes recombination , and is estimated to have a small effective population size in vivo . We performed population genetics-based stochastic simulations of the within-host evolution of HIV-1 and estimated the structure of the HIV-1 quasispecies and . We found that with small mutation rates , the quasispecies was dominated by genomes with few mutations . Upon increasing the mutation rate , a sharp error catastrophe occurred where the quasispecies became delocalized in sequence space . Using parameter values that quantitatively captured data of viral diversification in HIV-1 patients , we estimated to be substitutions/site/replication , ∼2–6 fold higher than the natural mutation rate of HIV-1 , suggesting that HIV-1 survives close to its error threshold and may be readily susceptible to mutagenic drugs . The latter estimate was weakly dependent on the within-host effective population size of HIV-1 . With large population sizes and in the absence of recombination , our simulations converged to the quasispecies theory , bridging the gap between quasispecies theory and population genetics-based approaches to describing HIV-1 evolution . Further , increased with the recombination rate , rendering HIV-1 less susceptible to error catastrophe , thus elucidating an added benefit of recombination to HIV-1 . Our estimate of may serve as a quantitative guideline for the use of mutagenic drugs against HIV-1 .
The high mutation rate of HIV-1 coupled with its massive turnover rate in vivo results in the continuous generation of mutant viral genomes that are resistant to administered drugs and can evade host immune responses [1] , [2] . The design of drugs and vaccines that exhibit lasting activity against HIV-1 has remained a challenge [3]–[6] . A promising strategy to overcome this challenge has emerged from insights into viral evolution gained from the molecular quasispecies theory [7] , [8] . The theory predicts that a collection of closely related but distinct genomes , called the quasispecies , exists in an infected individual when the viral mutation rate is small . When the mutation rate is increased beyond a critical value , called the error threshold , the quasispecies delocalizes in sequence space , inducing a severe loss of genetic information–a phenomenon termed error catastrophe–and compromising the viability of the viral population . It is widely believed therefore that viral mutation rates may have been evolutionarily optimized to lie close to but below their error thresholds so that viral diversity , and hence adaptability , is maximized while genomic identity is maintained [9]–[11] . Consequently , a small increase in the viral mutation rate may trigger an error catastrophe . In accordance , 4-fold increase in the mutation rate induced a dramatic 70% loss of poliovirus infectivity in vitro [9] . Chemical mutagens have been employed successfully to enhance the mutation rates of a host of other viruses [10]–[13] including HIV-1 [14]–[17] . An HIV-1 mutagen is currently under clinical trials [18] . Identification of the host restriction factor APOBEC3G ( A3G ) has suggested that mutagenesis might also be a natural antiviral defence mechanism ( reviewed in [19] , [20] ) . A3G ( and , to a smaller extent , APOBEC3F ) induces G to A hypermutations in HIV-1 , which when unchecked can severely compromise the viability of HIV-1 . Interestingly , HIV-1 appears to have evolved a strategy to resist A3G . The HIV-1 protein Vif targets A3G for proteasomal degradation and suppresses its mutagenic activity . Vif thus presents a novel drug target . Inhibiting Vif may enable A3G to exert mutagenic activity adequate to compromise HIV-1 . Indeed , significant efforts are underway to develop potent HIV-1 Vif-inhibitors [21] . The use of mutagenesis as an antiviral strategy requires caution because increasing the mutation rate to values below the error threshold could prove counterproductive . The quasispecies theory predicts that a suboptimal increase in the mutation rate would result in an increase in viral diversity that may not be accompanied by a substantial loss of genetic information , which in turn may facilitate the emergence of mutant genomes resistant to drugs and/or host-immune responses [22] . The mutagenic activity of drugs and of host-factors like A3G is dose-dependent [9] , [17] , [23] . It is important , therefore , to identify the minimum exposure to mutagenic drugs that would ensure that the error threshold of HIV-1 is crossed . The error threshold of HIV-1 is not known . Translation of the predictions of the quasispecies theory to HIV-1 has remained a challenge: The theory considers the asexual reproduction of a haploid organism with an infinitely large population size , whereas HIV-1 is diploid , undergoes recombination [24]–[27] , and is estimated to have a small effective population size in vivo , ∼102–105 cells [28]–[37] . Several studies have advanced the quasispecies theory to account for the diploid nature of HIV-1 and recombination [38]–[50] . The small effective population size of HIV-1 in vivo , however , renders the deterministic formalism of the quasispecies theory inadequate . Population genetics-based stochastic simulations have been resorted to as an alternative [37] , [40] , [41] , [47] , [49] , [51]–[53] . Such simulations often make significant departures from the quasispecies theory that may render an error catastrophe untenable . For instance , a sharp error catastrophe may not occur with certain fitness landscapes [54]–[58] . Further , in the large population size limit , the simulations may not converge to the predictions of the quasispecies theory [59] . Indeed , whether population genetics- or quasispecies theory-based approaches are more appropriate for describing viral evolution has been the subject of an ongoing debate [56] . We have recently developed stochastic simulations of HIV-1 evolution in vivo that incorporate key aspects of the HIV-1 lifecycle and the underlying evolutionary forces , namely , mutation , multiple infections of cells , recombination , fitness selection using a landscape representative of HIV-1 , and random genetic drift [37] , [47] , [49] . The simulations quantitatively described data of the evolution of viral diversity and divergence in HIV-1 infected individuals over several years following seroconversion , indicating that the simulations faithfully mimicked HIV-1 evolution in vivo [37] . Here , we applied the simulations to determine the structure of the HIV-1 quasispecies and estimate its error threshold . In the limit of large population sizes and in the absence of recombination , our simulations converged to the quasispecies theory , thus bridging the gap between population genetics- and quasispecies theory-based approaches to describing viral evolution and suggesting the existence of an error threshold for HIV-1 . We estimated the error threshold of HIV-1 to be ∼2–6-fold higher than its natural mutation rate . HIV-1 thus appears to survive close to its error threshold and may be readily susceptible to mutagenic drugs .
We performed simulations as follows . Uninfected cells were synchronously infected by a pool of identical virions , each cell potentially infected by multiple virions . Viral genomic RNA in cells were then reverse transcribed to proviral DNA . Reverse transcription involved mutation and recombination . The proviral DNA were transcribed to viral genomic RNA , which were assorted into pairs and released as progeny virions . Virions from the pool of progeny virions were selected according to their relative fitness to infect a new generation of uninfected cells , and the cycle was repeated . Following several thousand generations and several such realizations , the expected structure of the viral quasispecies at a given mutation rate was determined . Simulations at different mutation rates allowed identification of the error threshold . Details of the simulation procedure and parameter values employed are presented in Methods . We present first the evolution of the frequencies of genomes in different Hamming classes in one realization of our simulations ( Fig . 1 ) . Hamming class contains genomes carrying mutations with respect to the fittest , or master , sequence; thus , , where is the genome length . Without loss of generality , we let the fittest sequence be the founder sequence ( Fig . S1 ) . Thus , initially , the distribution of genome frequencies was localized at Hamming class zero . As time ( or the number of generations ) progressed , mutant genomes arose and higher Hamming classes were populated ( Fig . 1A ) . The average number of mutations contained in the proviral pool gradually increased and the peak of the frequency distribution shifted to higher Hamming classes . After a certain number of generations , here ∼500 , the distribution became steady; no net shift occurred from generation 500 to 10000 . Correspondingly , the Shannon entropy , , rose from zero at the start and attained the steady value , , of by generation 500 ( Fig . 1B ) . We averaged the above frequencies over the last 1500 generations and over several realizations of our simulations to obtain the expected frequency distribution at steady state . The latter distribution yielded the structure of the viral quasispecies ( Fig . 1A ) . Upon increasing the mutation rate , , the quasispecies shifted to higher Hamming classes indicating the increasing accumulation of mutations ( Fig . 2A ) . The peak Hamming class ( i . e . , the Hamming class with the maximum frequency ) shifted gradually from to as increased from to substitutions/site/replication . ( Note that nucleotides in Fig . 2A . ) At this point , a small increase in to substitutions/site/replication produced a remarkable jump in the peak Hamming class to . Subsequent increases in again caused only gradual shifts in the peak Hamming class . This jump was more dramatic with larger genome lengths . With nucleotides , the peak Hamming class jumped from to when increased from to substitutions/site/replication ( Fig . 2B ) . Correspondingly , jumped from 0 . 24 to as increased from to substitutions/site/replication ( Fig . 2C ) . implied that all possible genomes occurred with equal frequencies . The number of distinct genomes in Hamming class is . Thus , if all genomes occurred with equal likelihood , the Hamming class frequencies would follow . Indeed , we found that the quasispecies structure obtained by our simulations was identical to the latter distribution of Hamming class frequencies ( Fig . 2B inset ) , confirming that all genomes occurred with equal likelihood when . Thus , the jump in indicated the transition to error catastrophe . The transition from low to occurred over a narrow range of values of . For within this range , the quasispecies structure was bimodal because error catastrophe occurred in some realizations and not in others depending on the stochastic variations encountered . For illustration , we present several independent realizations of our simulations at three values of , namely , , and substitutions/site/replication ( Fig . 3 ) , where the first is well below the transition from low to , the second is in the transition region , and the third is well above the transition in Fig . 2B . With substitutions/site/replication , in each realization rose from zero and reached in generations ( Fig . 3A ) . There was little variation between the realizations . With substitutions/site/replication , rose from zero and reached in generations , again with little variation between the different realizations ( Fig . 3C ) . With substitutions/site/replication , however , we found substantial variation between realizations ( Fig . 3B ) . rose from zero and reached a plateau value of in generations . In some realizations , remained at this value till the end , i . e . , 10000 generations . In other realizations , at some intermediate time , which differed from realization to realization , rose sharply from and reached 1 . remained at 1 subsequently . Averaging the Hamming class frequencies thus yielded the bimodal structure of the quasispecies observed for substitutions/site/replication ( Fig . 2B ) , where realizations with yielded the peak at Hamming class and realizations with yielded the peak at Hamming class . Our aim was to identify the smallest value of at which error catastrophe was ensured . We found that when stochastic variations became insignificant and error catastrophe occurred nearly invariably . We therefore identified the smallest for which as the error threshold , . Thus , and substitutions/site/replication for and nucleotides , respectively , in Fig . 2C . To test whether our simulations converged to the quasispecies theory , we performed simulations with parameter values that mimic the assumptions employed in the quasispecies theory . We let infection/cell and crossovers/site/replication to represent the asexual reproduction of effectively haploid individuals . We chose a large population size , cells , and a small genome length , nucleotides , to approximate the infinite population size limit ( ) . We employed the single peak fitness landscape , typically employed in calculations of the quasispecies theory , which we implemented by letting viral production be virions/cell for cells infected with the master sequence and virion/cell for all other cells and then selecting virions with equal probability from the viral pool . We also solved the equations of the quasispecies theory using the latter fitness landscape ( Methods ) . Remarkably , our simulations were in excellent agreement with the quasispecies theory for a wide range of mutation rates ( Fig . 7A ) . To test the robustness of this agreement , we performed simulations with two other fitness landscapes , an exponential landscape , , where the relative fitness declined nearly linearly ( at rate per mutation ) with the number of mutations from the master sequence , , and the experimental landscape above rescaled to the smaller genome length . In both these cases , we let virions/cell in our simulations and selected virions in proportion to their relative fitness . Again , our simulations were in excellent agreement with solutions of the quasispecies theory using the latter fitness landscapes ( Figs . 7B and C ) . Thus , with large population sizes , our simulations were in quantitative agreement with the quasispecies theory . With smaller population sizes , our simulations predicted trends that were consistent with previous finite population models of genomic evolution . Further , with parameter values representative of HIV-1 infection in vivo , we showed previously that our simulations quantitatively described patient data of the evolution of viral diversity and divergence over extended durations ( ∼10–12 years ) [37] , giving us confidence in our simulations . We employed our simulations to estimate the error threshold of HIV-1 . We performed simulations with parameter values that mimic patient data of viral genomic diversification quantitatively ( Methods ) . We previously analyzed data of viral diversity and divergence from 9 patients [68] and found that with infections/cell , following observations of Jung et al . [24] , the best-fit values of varied from 400–10000 cells across the patients with a mean of cells [37] . Accordingly , we performed simulations here with , , and cells . We found a sharp error catastrophe with , and substitutions/site/replication , respectively ( Fig . 8A ) . A smaller frequency of multiple infections of cells , mimicking the observations of Josefsson et al . [69] , was also able to capture the same patient data with higher best-fit values of [37] . Then , except for one patient ( Patient 11 ) , for whom was 105 cells , the best-fit values of were in the range of 1500–10000 cells . Recognizing that the dependence of on was weak for large , we performed simulations with , , cells ( where 5000 cells was the mean for the remaining 8 patients ) using drawn from a distribution mimicking the observations of Josefsson et al . We found again that a sharp error catastrophe occurred with , and substitutions/site/replication for the three cases ( Fig . 8B ) , close to the estimates above . The modest increase of with again displayed the dependence ( , Fig . 8B inset ) and yielded substitutions/site/replication for cells and substitutions/site/replication for . Taken together , our simulations predict that HIV-1 undergoes a sharp error catastrophe and estimate the error threshold to be in the range substitutions/site/replication .
The success of mutagenic drugs against HIV-1 hinges on reliable estimates of the error threshold of HIV-1 , which are currently lacking . The assumptions employed in the quasispecies theory render it inadequate for describing HIV-1 evolution . Here , we have employed population genetics-based simulations of HIV-1 evolution to examine the susceptibility of HIV-1 to mutation-driven error catastrophe . With these simulations , we found that HIV-1 experienced a sharp error catastrophe at a mutation rate of substitutions/site/replication . Our simulations incorporated key evolutionary forces underlying the within-host genomic diversification of HIV-1 and were shown previously to be in agreement with longitudinal patient data of viral diversity and divergence [37] , giving us confidence in our estimate of the error threshold . That the estimated error threshold is ∼2–6 fold higher than the natural mutation rate of HIV-1 in vivo , substitutions/site/replication [63] , [70] , suggests that HIV-1 exists close to its error threshold . The mutation rate of HIV-1 thus appears to be evolutionarily optimized to maximize diversity while retaining genomic identity . A relatively small ( 2–6 fold ) increase in the mutation rate may thus drive HIV-1 past its error threshold , presenting a quantitative guideline for mutagenic drugs . The quasispecies theory has presented remarkable insights into viral evolution and suggested new strategies of intervention [9] , [71]–[74] . Yet , its ability to describe viral evolution comprehensively is limited , as recognized by Eigen himself [75] , by its assumptions of , for instance , an infinitely large population size , asexual reproduction of haploid organisms , and an isolated peak fitness landscape where all mutants are equally less fit than the master sequence . The last 40 years have seen significant efforts to relax these assumptions and tailor the quasispecies theory to specific organisms , especially HIV: Several , more complex and more realistic fitness landscapes have been employed [37] , [49] , [54] , [55] , [58] , [60] , [76]–[78] . Simultaneously , population genetics-based approaches , which naturally consider stochastic effects associated with finite populations , have been developed [59] , [60] , [79]–[85] . The latter descriptions , however , while painting a more realistic picture of the organisms considered , often make marked deviations from the key predictions of the quasispecies theory . In particular , finite population models may not converge to the quasispecies theory in the infinite population limit [59] , or with more complex fitness landscapes , a sharp error catastrophe may cease to occur [54]–[58] . Consequently , questions arise of the relative merits and appropriateness of using the quasispecies theory or population genetics-based approaches to describe viral evolution ( reviewed in [56] ) . Here , we showed that our simulations converge to the quasispecies theory in the large population size limit , indicating that quasispecies theory is not at odds with population genetics-based descriptions at least of HIV-1 . In a related study , convergence of similar population genetics-based descriptions to the quasispecies theory has been established formally [86] . Importantly , with a fitness landscape representative of HIV-1 [64] , and with other parameters that mimic patient data , our simulations predict that a sharp error threshold exists for HIV-1 . In our simulations , the error threshold scaled linearly with , where C is the population size of cells , in agreement with previous studies [59] , [60] . We note that some studies using alternative simulation strategies found a linear scaling with [80] . The origin of this discrepancy in the dependence of the error threshold on C remains to be established . Nonetheless , the weak dependence of the error threshold on C implies that our estimate of the error threshold remains robust to any increase in the effective population size in vivo either due to inter-patient variations or due to uncertainties in the estimates of model parameters . We showed previously that estimates of the effective population size of HIV-1 in vivo were sensitive to the frequency of multiple infections of cells , M , and the recombination rate [37] . Few estimates of M in vivo are available . While one study of infected splenocytes in two patients found that most cells were multiply infected with a mean of 3–4 proviruses per cell [24] , recent evidence from peripheral blood mononuclear cells of several acute and chronically infected individuals suggests that multiple infections of cells may be rare [69] , and hence the influence of recombination weak [51] , [61] . Using parameters corresponding to either observation , we found that our simulations captured patient data of viral diversification with appropriate values of C [37] . Using both combinations of M and C that matched patient data , we estimated the error threshold of HIV-1 here and found that the estimates were close , suggesting that uncertainties in the frequency of multiple infections did not significantly affect our estimate of the error threshold . The role of recombination in HIV-1 evolution has remained difficult to interpret [2] , [39] , [87] . Just as recombination can bring favorable mutations together , it can also drive favorable combinations of mutations apart , raising questions more generally about the evolutionary origins of the ubiquitously present recombination and sexual reproduction , often referred to as the paradox of sex [39] , [88] , [89] . The benefit of sex has recently been suggested to arise from the subtle interactions of random genetic drift , selection , and recombination in finite populations [90] . When the population size is small , negative linkage disequilibrium ( ) is generated by the Hill-Robertson effect [91] . Recombination lowers the absolute value ( magnitude ) of , which when enhances diversity and favors selection [89] , [92]–[94] . Indeed , our simulations showed that as the recombination rate increased , the quasispecies shifted to lower peak Hamming classes and spread wider , implying greater average fitness and diversity . In agreement , we showed previously that the mean fitness and diversity of the viral population increased with recombination when the population size was small [47] . An added advantage of recombination that we found here was that the error threshold also increased with recombination , rendering the quasispecies more resistant to mutation-driven loss of genetic information . In an earlier study , recombination was found in contrast to decrease the error threshold [38] . The latter study , however , considered an infinitely large population size with a single peak landscape , which is expected to generate . Accordingly , the lowering of by recombination decreases diversity and is therefore expected to lower the error threshold . generated by the Hill-Robertson effect underlies the enhancement of the error threshold due to recombination in our simulations . Given that host factors such as A3G combat HIV-1 by increasing the viral mutation rate [19] , [20] , recombination , in synergy with Vif-induced degradation of A3G , may serve to stall the onslaught of A3G and establish lasting infection . The population sizes we employed were obtained by fits of our simulations to patient data [37] . The census population size of HIV-1 is ∼107–108 infected cells [95] . Yet , the effective population sizes obtained by several independent studies are small and lie in the range of ∼102–105 cells ( reviewed in [35] ) . The effective population size is defined as the size of the population in an idealized model of evolution that has the same population genetic properties as that of the natural population [96] . The reasons underlying the differences between the census and effective population sizes of HIV-1 remain to be established; bottlenecks introduced by the immune system and other selection pressures [36] , asynchronous infections of cells [97] , pseudohitchhiking [98] , and metapopulation structure [99] may all contribute to the small effective population sizes estimated , but their roles in HIV-1 evolution are yet to be fully elucidated . We employed a fitness landscape that is a measure of the relative replicative ability of various HIV-1 mutants determined using in vitro assays [64] . The landscape suggests that the predominant fitness effects depend on the number and not on the specific combinations of mutations , allowing us to group genomes into Hamming classes [100] . Simpler fitness landscapes , such as multiplicative landscapes , were not compatible with patient data [37] . More comprehensive fitness interactions are beginning to be unraveled [66] , [101] . The resulting fitness values [66] have been shown to be correlated with the viral load in vivo [102] . Under certain limiting conditions , we found that the latter interactions yielded a fitness landscape consistent with the landscape we employed above ( Methods and Fig . S7 ) . Further , our estimates of the error threshold were robust to minor variations in the fitness landscape . For instance , allowing lethal mutations using a truncated landscape , where genomes with fitness below a certain threshold were assumed replication incompetent , did not substantially alter the error threshold ( Fig . S7 ) . We recognize that lethal mutations can occur more frequently; for instance , 40% of random mutations in an RNA viral genome were found to be lethal [103] . Such a scenario is estimated to increase the error threshold for an infinitely large population size and a single peak fitness landscape by a factor of ∼5/3 [104] , [105] . Understanding the influence of major variations in the fitness landscape is computationally prohibitive and awaits future studies . Finally , we recognize that we have assumed uniform recombination rates and either uniform or nucleotide-specific mutation rates across the HIV-1 genome , whereas mutation [63] and recombination hot-spots [106] , [107] are known to exist within HIV-1 . Estimation of the error threshold of HIV-1 from experimental studies of viral mutagenesis-induced loss of viral infectivity has not been possible because of several confounding effects . For instance , 2–3 fold increase in the mutation rate obliterated HIV-1 infection in vitro [14] , [15] , in agreement with our present findings . The agreement , however , is not conclusive because establishing that the loss of infectivity in vitro is due to an error catastrophe is not straight-forward . The loss of infectivity may be due to an error catastrophe , as demonstrated with poliovirus [9] , but may also arise from other effects: At mutation rates above the natural mutation rate but below the error threshold , production of defective genomes may drain resources within cells , compromising the production of viable genomes and causing extinction of the viral population [13] . Thus , whether viral extinction necessarily implies crossing the error threshold remains unclear . Conversely , crossing the error threshold may not imply viral extinction; the latter may require crossing an alternative ‘extinction’ threshold , where each viral particle produces less than one progeny that infects a cell , akin to the epidemiological threshold for extinction of disease [108] . ( Note that in our present simulations , infection was sustained by keeping the pool of infected cells constant . ) Viral extinction may also be determined by the influence of mutations on protein stability and its impact on viability [109] , [110] . Establishing which of these phenomena underlies the observed loss of viral infectivity in vitro remains a challenge . Finally , we recognize that the dynamics of the transition to error catastrophe , which remains poorly characterized , is also of importance to mutagenic strategies targeting HIV-1 . For instance , 9–24 serial passages were required for loss of viral infectivity in vitro [14] . In a recent clinical trial with an HIV-1 mutagen , no viral load decline was observed in patients following 124 days of treatment although the mutational patterns were altered [18] . This absence of apparent antiviral activity was attributed to the lack of knowledge of both the level and the duration of exposure of the drug necessary to compromise the viability of HIV-1 [18] , reiterating the importance of reliable estimates of the error threshold and of the timescales of the transition . Our estimate of the error threshold together with the dose-response data of the drug may help determine the level of drug necessary to induce an error catastrophe in HIV-1 . Further , although we focused here on identifying the structure of the HIV-1 quasispecies and estimating its error threshold , our simulations present a framework for determining the time required to ensure completion of the transition to error catastrophe , thus elucidating guidelines for the duration of treatment with mutagenic drugs .
We employed parameter values representative of HIV-1 infection in vivo [37] . Variations are mentioned below and in the text and figures . We let nucleotides and ρ = 8 . 3×10−4 crossovers/site/replication [46] . We fixed to 3 infections/cell following Jung et al . [24] , or let follow a distribution–similar to that observed by Josefsson et al . [69]–where 77% of the cells were singly , 19% doubly , and 4% triply infected [37] . With each , we chose an appropriate that matched patient data [37] . Following recent estimates of the basic reproductive ratio of HIV-1 in vivo [111] , we let P = 10 infectious progeny virions/cell . A majority of HIV-1 mutations are transitions [70]; as a simplification , we therefore ignored transversions , insertions and deletions . We spanned a wide range of mutation rates in order to identify the error threshold . We let selection follow the fitness landscape derived in [37] to capture corresponding experimental data from [64] . Accordingly , the relative fitness of genome is represented by , where is the minimum fitness of sequences , is the Hamming distance at which , and is analogous to the Hill coefficient [37] . The fitness of a virion is determined by the average Hamming distance of its two genomes from the master sequence . We let simulations proceed to 10000 generations ( ∼30 years ) . We examined the influence of variations in some of these parameters as mentioned below . According to the quasispecies theory [7] , [8] , the structure of the quasispecies is obtained as the dominant eigenvector of the value matrix , . We constructed the mutation matrix by recognizing that its element , , is the probability that genome mutates to genome , with the Hamming distance between genomes and . The selection matrix is a diagonal matrix with elements , the relative fitness of the respective genomes . We employed three different fitness landscapes: the experimental landscape above , the isolated peak landscape , and the exponential landscape ( see above ) . We computed the dominant eigenvector of and normalized it so that . The Hamming class frequencies were then . We performed computations using a program written in MATLAB . | Currently available antiretroviral drugs curtail HIV infection but fail to eradicate the virus . A strategy of intervention radically different from that employed by current drugs has been proposed by the molecular quasispecies theory . The theory predicts that increasing the viral mutation rate beyond a critical value , called the error threshold , would cause a severe loss of genetic information , potentially leading to viral clearance . Several chemical mutagens are now being developed that can increase the mutation rate of HIV-1 . Their success depends on reliable estimates of the error threshold of HIV-1 , which are currently lacking . The quasispecies theory cannot be applied directly to HIV-1: the theory considers an infinitely large population of asexually reproducing haploid individuals , whereas HIV-1 is diploid , undergoes recombination , and is estimated to have a small effective population size in vivo . We employed detailed stochastic simulations that overcome the limitations of the quasispecies theory and accurately mimic HIV-1 evolution in vivo . With these simulations , we estimated the error threshold of HIV-1 to be ∼2–6-fold higher than its natural mutation rate , suggesting that HIV-1 survives close to its error threshold and may be readily susceptible to mutagenic drugs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"theoretical",
"biology",
"evolutionary",
"biology",
"hiv",
"population",
"genetics",
"biology",
"computational",
"biology",
"viral",
"diseases",
"infectious",
"disease",
"modeling",
"evolutionary",
"theory"
] | 2012 | Stochastic Simulations Suggest that HIV-1 Survives Close to Its Error Threshold |
Representational models specify how activity patterns in populations of neurons ( or , more generally , in multivariate brain-activity measurements ) relate to sensory stimuli , motor responses , or cognitive processes . In an experimental context , representational models can be defined as hypotheses about the distribution of activity profiles across experimental conditions . Currently , three different methods are being used to test such hypotheses: encoding analysis , pattern component modeling ( PCM ) , and representational similarity analysis ( RSA ) . Here we develop a common mathematical framework for understanding the relationship of these three methods , which share one core commonality: all three evaluate the second moment of the distribution of activity profiles , which determines the representational geometry , and thus how well any feature can be decoded from population activity . Using simulated data for three different experimental designs , we compare the power of the methods to adjudicate between competing representational models . PCM implements a likelihood-ratio test and therefore provides the most powerful test if its assumptions hold . However , the other two approaches—when conducted appropriately—can perform similarly . In encoding analysis , the linear model needs to be appropriately regularized , which effectively imposes a prior on the activity profiles . With such a prior , an encoding model specifies a well-defined distribution of activity profiles . In RSA , the unequal variances and statistical dependencies of the dissimilarity estimates need to be taken into account to reach near-optimal power in inference . The three methods render different aspects of the information explicit ( e . g . single-response tuning in encoding analysis and population-response representational dissimilarity in RSA ) and have specific advantages in terms of computational demands , ease of use , and extensibility . The three methods are properly construed as complementary components of a single data-analytical toolkit for understanding neural representations on the basis of multivariate brain-activity data .
The measurement of brain activity is rapidly advancing in terms of spatial and temporal resolution , and in terms of the number of responses that can be measured simultaneously [1] . Modern electrode arrays and calcium imaging enable the recording of hundreds of neurons in parallel . Electrophysiological signals that reflect summaries of the population activity can be recorded using both invasive ( e . g . the local field potential , LFP ) and non-invasive techniques ( e . g . scalp electrophysiological measurements ) at increasingly high spatial resolution . Modern functional magnetic resonance imaging ( fMRI ) enables us to measure hemodynamic activity in hundreds of thousands of voxels across the entire human brain at sub-millimeter resolution . In order to translate advances in brain-activity measurement into advances in computational theory [2] , researchers increasingly seek to test representational models that capture both what information is represented in a population of neurons , and how it is represented . Knowing the content and format of representations provides strong constraints for computational models of brain information processing . We refer to hypotheses about the content and format of brain representations as representational models , and address here the important methodological question of how to best test such models . Referring to an activity pattern as a “representation” constitutes a functional interpretation [3] , which requires not only that the represented variable ( such as a perceptual property , some cognitive content , or an action parameter ) is encoded in the pattern of activity in a format that can be read out by downstream neurons , but also that the information is actually used by other brain regions and , thus , serves a functional purpose [4] . The representational interpretation therefore ultimately needs to be supported by evidence for a cause-and-effect relationship between the activity and downstream neural and behavioral responses . Testing causal effects of activity patterns is beyond the scope of the observational methods considered in this paper . However , we note that a good brain-computational model must , as a necessary condition , be able to explain the format in which information it is encoded in the task-relevant brain regions . For a population code to constitute an explicit representation , another area must be able to read out the represented variable directly using a neurobiologically plausible readout mechanism , such as linear or radial-basis-function decoding [2 , 5 , 6] . Note that this definition of explicit does not restrict us to highly localized codes , such as the “grandmother neuron” [7] , but encompasses widely distributed codes . An example of an implicit representation is the representation of object category in the retina . The retina clearly contains information about object category , and an aspect of its function is to convey this information . However , it does not explicitly represent object category . Multiple stages of nonlinear tranformation along the ventral visual stream are required to render the category of an object explicit . Inferior temporal cortex contains a representation of object category [8 , 9] , along with representations of much additional information [10] . Many researchers have used linear decoding methods to reveal explicit information in neural representations [11–13] . Representational models , as considered here , go one step further: they fully characterize the representational geometry , defining all represented features in a region , how strongly each of them is represented ( signal to noise ratio ) , and how the activity patterns associated with different features relate to each other . Representational models therefore fully specify the representational content of an area . To define representational models formally , we need to consider two complementary perspectives on activity data , as illustrated in Fig 1 . The activity of many neurons , or more generally measurement channels ( neurons , electrodes , or fMRI voxels ) , can be measured across a range of experimental conditions ( stimuli , movements , or tasks ) . Thus , each channel will have an activity profile , which can be plotted as a point in the space spanned by the experimental conditions ( Fig 1B ) . A representational model specifies a probability distribution of activity profiles in the space spanned by the experimental conditions . It treats the true activity profiles as a random variable and predicts , for each possible activity profile , the probability of observing a measurement channel exhibiting that profile . It does not predict the activity profile for each individual channel actually measured . The motivation for this approach derives from the idea that the computational function of a region does not depend on specific neurons having specific response properties , but on the fact that certain features can be read out from the population by downstream neurons . The probability distribution over activity profiles determines which features can be read out from the code and the signal-to-noise ratio of the readout . By basing further analyses on the probability distribution of the activity profiles , we disregard ( 1 ) which neuron fulfills which function , and ( 2 ) where neurons are located within a cortical area . Furthermore , by focusing on the second moment of the distribution we ignore ( 3 ) the degree to which the information about a given represented feature is concentrated in a few neurons ( as in single-cell selectivity for a represented feature ) or spread out over the population . Ignoring these aspects may be viewed as an advantage or a disadvantage , depending on the level of description that a researcher is interested in . We argue that treating activity profiles as random vectors is a simplification that is useful for drawing computational insights from population activity measurements . In this paper , we show that the multivariate second moment of the activity profiles fully defines the representational geometry and with it all the information that can linearly or nonlinearly decoded . In particular , under the assumption of Gaussian noise the second moment determines the signal-to-noise ratio with which any feature can be decoded . We discuss three established methods for adjudicating between representational models: encoding analysis , pattern-component modeling ( PCM ) and representational similarity analysis ( RSA , see Table 1 ) . We show that these three techniques all exclusively rely on information contained in the second moment . This core commonality enables us to consider these methods in the same formal framework . In encoding analysis [14 , 15] , representational models are defined in terms of the underlying features ( Fig 2A ) . Each activity profile can be characterized by a linear combination of such features . Examples include Gabor filters [16] ( for a low-level visual representation ) , abstract semantic dimensions [17] ( for a cognitive representation ) , and force , direction or hand position [18–20] ( for a movement representation ) . The importance of each feature in each channel is measured by a feature weight . Feature weights are considered first-level parameters in our framework , as they describe the individual activity profiles , as opposed to second-level parameters that describe the distribution of the activity profiles ( Table 1 ) . The large number of parameters ( number of features in the model times number of channels in the measurements ) engenders a danger of overfitting . Encoding models are therefore commonly evaluated using cross-validation: The feature weights are estimated on a training set , and the model is evaluated in terms of its performance at predicting left-out data [14] . The test data may consist in a sample of experimental conditions not used in training , so as to test the model’s generalization performance [15 , 16] . While many studies use simple linear regression to estimate the weights [15 , 21] , it is increasingly common to use a regularization penalty ( for example the L2 norm of the vector of weights ) [16 , 17] . We will show that regularization is not merely a technical trick used in fitting a given model . Instead , the regularization ( and its implicit distributional assumptions ) are an essential part of the representational hypothesis that is tested . Without it , encoding models do not specify a probability distribution with a finite second moment and thus do not define the linear decodability of different features . Pattern component modeling [22] is based on an explicit generative model of the process that produced the data and can be considered a Bayesian approach . The true activity profiles are assumed to have a multivariate Gaussian distribution in the space spanned by the experimental conditions ( Fig 2B ) . This formulation enables us to evaluate the marginal likelihood of the observed activity profiles under the probability distribution specified by the model . Thus , we do not fit any first-level parameters ( feature weights ) and hence reduce the risk of overfitting . This enables us to compare models with different numbers of features without having to correct for model complexity . If the assumptions of the generative model hold , PCM implements the likelihood-ratio test between models [23] , which by the Neyman-Pearson lemma [24] , is the most powerful test of its size . In theory , therefore , PCM should yield more accurate inferences than any of its competitors , that is it should be able to more sensitively adjudicate among competing models . Finally , representational similarity analysis ( RSA [9 , 25 , 26] ) approaches the problem from a complementary perspective . Rather than considering the activity profiles of the measurement channels as points in the space spanned by the conditions ( Fig 1B and 1D ) , it considers the activity patterns associated with the experimental conditions as points in the space spanned by the measurement channels ( Fig 1C and 1E ) . RSA then uses the representational distances ( Fig 2C ) between the conditions as a summary statistic . We will see that these distances again exclusively depend on the second moment of the distribution of activity profiles . Having obtained a matrix of dissimilarities between activity patterns ( the representational dissimilarity matrix , RDM ) , RSA then tests models by comparing the observed distances to the distances predicted by each representational model . This can be done by calculating rank-based correlations [27] or Pearson correlations [28] . Here we show that for near-optimal inferences it is important to take the co-dependence structure of the distance estimates into account , for example by using a multivariate normal approximation to the joint distribution of the cross-validated Mahalanobis distances [29 , 30] . In the remainder of the paper , we first introduce the second moment of the activity profiles and explain why it is the sufficient statistic of the representational geometry and thus of linear and nonlinear decodability . We then define the three methods in detail , and show how they related to the second moment . Finally , using simulated data and models taken from our fMRI work , we assess the statistical efficiency , i . e . how well these methods adjudicate between two or more competing representational models given limited data . We also compare the methods in terms of their computational efficiency .
All symbols used in the following derivations are summarized in Table 2 . First , we define U to be the matrix of noiseless activity profiles with K ( number of experimental conditions ) rows and P ( number of measurement channels ) columns . Each row of this matrix is an activity pattern , the response of the whole population to a single condition . Each column of this matrix is an activity profile ( Fig 1A ) . Because we are interested in the distribution of activity profiles , but not in the activity profiles per se , we consider the columns of U to be a random variable . This is an essential step underlying our common framework , which is justified by the fact that , for the purpose of reading out information , the different measurement channels are exchangeable ( see introduction ) . We assume that the activity profiles are repeatedly measured , with the data consisting of M independent partitions , each containing at least one activity measurement for each condition and measurement channel . In the context of fMRI , a partition will consist of a separate phase of data acquisition , e . g . a scanner run . The activity estimates U ^ ( m ) of partition m are the true patterns U plus noise E ( m ) . The noise captures both neural trial-by-trial variability of the activity pattern in a single condition , as well as measurement noise . U ^ ( m ) = U + E ( m ) ( 1 ) For the purposes of this paper , we assume that the noise is Gaussian , and independent and identically distributed ( i . i . d . ) across conditions and partitions ( homoscedasticity ) . Possible dependence within each partition , however , can be easily accounted for [29 , 31] . The discussion below further assumes that the noise is also i . i . d . across different measurement channels ( isotropicity ) . However , noise in fMRI , MEG , and even invasive electrophysiology exhibits strong correlations between neighboring locations in the brain . To account for these dependencies , we employ multivariate noise normalization ( i . e . spatial prewhitening ) , which has been shown to increase the reliability of inference [32] . Across all measurement channels , we estimate the P × P variance-covariance matrix across trials , ΣP and then regularize the estimate by shrinking it towards a diagonal matrix [33] . In the context of fMRI , we can use the residual time series from the fitting of the time-series model to estimate noise covariance [32 , 34] . We then post-multiply our activity estimates by Σ ^ P - 1 / 2 , rendering the model errors in the channels approximately uncorrelated . If multivariate noise normalization is not performed or is incomplete , inference will be suboptimal in all three methods ( for details see [29] ) . In this section , we show that , under the assumption of Gaussian noise , the second moment of the activity profiles fully characterizes the decodability of any feature that describes the experimental conditions . The fact that the second moment determines what can be decoded provides a motivation , from the perspective of brain computation , for using the second moment matrix as a summary statistic . The nth moment of a scalar random variable u is E ( un ) , where E ( ) denotes the expected value . Here we use a multivariate extension of the concept , with the second moment of the random vector u defined as the matrix E ( uuT ) , the expected outer product of the activity profiles , where the expectation is across the measurement channels . The second-moment matrix of the activity profiles is given by G ≡ ∑ j = 1 P u . , j u . , j T / P = U U T / P . ( 2 ) Thus , each cell of this matrix contains the scaled inner product of two activity patterns . Before calculating G , some investigators subtract the mean activity across measurement channels for each condition from the data . In this case , Eq 2 becomes the variance-covariance matrix of the activity profiles –the second moment around the mean activity profile . Here we do not remove the mean , but use the second moment around zero . From the perspective of a neuron that reads out the activity pattern of an area , any difference between activity patterns across conditions can be used to decode information . Some features ( for example , stimulus intensity ) may be encoded in the mean activity over all measurement channels . Other properties ( for example , stimulus identity ) may be encoded in relative activity differences , with some measurement channels responding more to one condition , and others to a different condition . The second moment around zero captures both of these potentially meaningful differences . Any feature of the conditions that we might want to decode can be defined by a K × 1 vector f with one entry per condition , which describes how the feature varies across conditions . To obtain a linear read-out estimate f i ^ for the feature fi for a given condition i , we weight each channel’s observed activity using the P × 1 read-out vector v: f ^ i = u ^ i , . v . ( 3 ) We would like the estimate f ^ to have very different values for two trials that differ on the feature value , while showing small differences for trials that have the same feature value . fT U is the pattern that encodes the feature . We are looking for the readout vector v that maximizes the ratio S between the sum-of-squares of the readout of the feature and the sum-of-square of the readout of the noise: S = v T U T ff T Uv v T E T ff T Ev ( 4 ) The solution to this equation is commonly known as Fisher’s linear discriminant [35] , which , under the assumption of homoscedastic Gaussian noise , is the best achievable linear decoder . If the noise is isotropic ( or the data is adequately pre-whitened ) , then ET ffT E = Ib , where b is a constant . The denominator then depends only on the norm of the read-out vector v , not on its direction , and can be ignored when v is constrained to have a norm of 1 . The best readout vector v is then given by the first eigenvector of the matrix UT ffT U , and the quality of the best readout is determined by the corresponding eigenvalue . Non-zero eigenvalues ( eig ) of a square matrix are invariant to cyclic permutations of the product order: e i g U T ff T U = e i g f T UU T f = P e i g f T Gf ( 5 ) Therefore , the quality of the best linear decoder for any feature ( as defined by f ) is fully characterized by fT Gf . By the same logic , the second moment determines decodeability in general . Assume you want to use a non-linear decoder to estimate feature f . For this to be possible , one has to ensure that one can decode the difference between any two conditions that have different feature values . Once that distinction is made , one can use a arbitrary non-linear function ( i . e . table lookup ) to read out the feature for each condition . Because the second moment defines the decodeability of any pair of two experimental conditions , it also defines whether a feature can be read out non-linearly . Importantly , the second moment is only a sufficient statistic for decodeability under the assumption that the readout neurons can integrate information from the entire population that constitutes the code , i . e . , it as capable of any arbitrary linear transform of the input data . If the readout neuron is only partially connected , it becomes important to what extent particular information is concentrated in restricted sets of neurons . This information is captured in the higher moments of the activity profile distribution , a point to which we will return in the Discussion . For a fully connected readout neuron that can weight activities in any arbitrary way , the second moment is a sufficient statistic of the decodable information . The methods in this paper were first developed in the context of fMRI data analysis , and our examples will come from this domain . A simple way to apply the analyses to fMRI data is to use as activity estimates ( U ^ ( m ) ) the regression coefficients , or “beta”-weights , from a first-level time series analysis [36 , 37] . The time-series model accounts for the hemodynamic lag and the temporal autocorrelation of the noise . The activity estimates usually express the difference in activity during a condition relative to rest . Activity estimates commonly co-vary together across fMRI imaging runs , because all activity estimates within a partition are measured relative to the same resting baseline . This positive correlation can be reduced by subtracting , within each partition , the mean activity pattern ( across conditions ) from each activity pattern . This makes the mean of each measurement channel ( across condition ) zero and thus centers the ensemble of points in activity-pattern space that is centered on the origin . Rather than using the concatenated activity estimates from different partitions , encoding analysis and PCM can also be applied directly to time series data . As a universal notation that encompasses both situations , we can use a standard linear mixed model [38]: Y = ZU + XB + ϵ , ( 6 ) where Y is an N × P matrix of all activity measurements , Z the N × K design matrix , which relates the activity measurements to the K experimental conditions , and X is a second design matrix for nuisance variables . U is the K × P matrix of activity patterns ( the random effects ) , B are the regression coefficients for these nuisance variables ( the fixed effects ) , and E is the matrix of measurement errors . If the data Y are the concatenated activity estimates , the nuisance variables typically only model the mean pattern for each run . If Y consists of time-series data , the nuisance variables typically capture additional effects such as time-series drifts and residual head-motion-related artifacts . All three methods can also be applied to recordings of single cell activity or neurophysiological potentials [9 , 25] . The activity estimates can then be firing rates estimated over a temporal window for each trial , or the power in different frequency bands over time . Because the trial-by-trial variability of firing rates will usually increase with the mean firing rate , it is advisable to use the square root of firing rates to make the data conform better to the assumption that the variance of the noise is independent of the signal [39] . Here we focus on models that treat the activity patterns U as static snapshots . To exploit the temporal detail provided by electrophysiological recordings , the analyses can be either performed using a sliding window over the time course of the trial [40–42] , or by “stacking” the time series and conditions , resulting in a activity matrix with TK rows [43] . An encoding model characterizes the structure of the representation in terms of a set of features [14–17] . We will show in the following that encoding models are representational models as definied by the second moment of the activity profiles . For this to be the case , however , the use of regularized regression is a crticial factor . We will therefore first present the encoding apporach in general , and then show why regularisation is important to test for distributions with a definied second moment . In general . the value of each feature for each experimental condition is coded in the model matrix M ( K conditions by Q features ) . The feature weight matrix W ( Q features by P channels ) then determines how the different model features contribute to the activity profiles of different measurement channels to produce the predicted activity patterns U: U = MW . ( 7 ) Geometrically , we can think of the features as the basis vectors of the subspace , in which the activity profiles reside ( Fig 2A ) . An alternative to cross-validation is to evaluate the likelihood of the measured activity profiles under the representational model . This approach is taken in pattern-component modeling [22] . We start with a generative model of the activity profiles ( Eq 6 ) . We consider the activity profiles ( columns of U ) to come from a multivariate Gaussian distribution with zero mean and second-moment matrix G . To account for other nuisance effects ( mean activity for each partition , low-frequency drift , etc ) , the model also has some fixed-effects regressors ( B ) . We are not interested in fitting U per se , but simply want to evaluate the likelihood of the data under different models , marginalized over all possible values of U . The marginal distribution for each channel ( columns of matrix Y ) takes the form of a multivariate normal: y . , j ∼ N X b . , j , V θ V θ = ZG s Z T + I σ ϵ 2 θ = s , σ ϵ 2 . ( 16 ) The predicted covariance matrix of the activity measurements for each person is the function of the model ( as encoded in the second-moment matrix G ) and two second-level parameters ( θ ) : one that determines the strength of the signal ( s ) and one that determines the variance of the noise ( σ ϵ 2 ) . In determining the likelihood , we remove the fixed effects using the residual forming matrix R = I − X ( X T V − 1 X ) − 1 X T V − 1 ( 17 ) We need to then account for the removal of these fixed effects by evaluating the restricted likelihood l ( Y | G , θ ) [47]: l Y | G , θ = - N P 2 log 2 π - P 2 log V - 1 2 trace Y T R T V - 1 RY - P 2 log X T V - 1 X . ( 18 ) To evaluate the fit of a model , the scaling and noise parameters need to be determined . For fMRI data , these two parameters can vary widely between different brain regions and individuals , and are not meaningful in themselves . We therefore replace θ with point estimates that maximize Eq 18—i . e . , the approach uses Empirical Bayes , or Type-II maximum likelihood for model comparison [45] . Because every model has the same two free second-level parameters , even models that are based on different numbers of features can be compared directly . An efficient implementation of this algorithm can be found in the open-source Matlab package for PCM [48] .
When evaluating encoding models without using regularization , one compares the subspaces spanned by the respective model features . To make different models distinguishable , one typically needs to reduce the dimensionality of the model matrix M , for example by using only the eigenvectors with the n highest eigenvalues of the predicted second-moment matrix . The decision to use a given number regressors is somewhat arbitrary: For example , Leo et al . [21] used 5 “synergies” ( i . e . principal components of the kinematic data of 20 movements ) , as these explained 90% of the variance of the behavioral data . Here we explore systematically how the number of principal components influences model selection . For each experiment , we simulated data sets with a fixed signal-to-noise ratio ( Exp . 1 and Exp . 3: s = 0 . 3 , Exp 2: s = 0 . 1; σ ε 2 = 1 ) , and compared model selection accuracies using a number of principal components ranging between one and the maximum number . We used both cross-validated performance measures , R c v 2 ( Eq 10 ) and r ( the correlation between predicted and observed values; Eq 11 ) to perform model selection . Fig 5A–5C shows the percentage of correct model selections for Experiments 1-3 . Results for encoding analysis without regularization are shown in blue . The dimensionality that differentiated best between competing models was 2 , 3 , and 5 features , respectively . As more features were included , the number of correct model selections declined . When the number of features was the same as the number of conditions minus 1 ( due to the mean subtraction ) , i . e . the models became saturated , model selection accuracy fell to chance . This is expected , as two saturated models span exactly the same subspace and hence make identical predictions ( Fig 3D ) . Using correlations as selection criterion led to more accurate decisions than using R c v 2 . Correlations ( Fig 5D–5F , blue lines ) were generally positive and peaked at a number of features that was slightly higher than the optimal dimensionality for model selection . R c v 2 values for encoding without a prior were all negative ( and are therefore not visible ) , because the approach does not account for the noise in the data and hence leads to predictions that are too extreme—i . e . the approach over-predicts the scale of the data . Correlations are insensitive to this problem as they allow for arbitrary scaling between predicted and observed values . From a Bayesian perspective , employing regularization ( Eq 13 ) is equivalent to adding a prior to the feature weights . Note that this changes the representational hypotheses tested . For example , the models for Experiment 3 , based on the neural network representations , now predicted not only that some weighted combination of the neural network features can account for the data , but more specifically that the distribution of activity profiles should match the distribution of activity profiles of the original neural network simulation . In the model matrix , we scaled each principal component of G with the square root of the eigenvalue ( Eq 15 ) , such that we could employ ridge regression to obtain the best linear unbiased predictor for the held-out data patterns . For encoding models with a prior , model selection performance increased with increasing number of features ( red lines , Fig 5A–5C ) . Thus , dimensionality reduction of the model is not necessary here . Furthermore , model selection was always more powerful with than without a prior when correlation was used for model selection . This reflects the fact that the prior provides additional information about the models to be compared . It enables us to compare well-defined distributions of activity profiles instead of just subspaces . For Experiments 2 and 3 , the R c v 2 criterion performed substantially worse than the correlation between predicted and observed activity patterns . The difference between the two criteria arises from the fact that correlations allow for an arbitrary scaling between predicted and observed activity patterns , whereas R c v 2 penalizes deviation in scale . The scaling of the prediction in turn strongly depends on the choice of the scalar regularization coefficient . This fact is illustrated in Fig 6 , where we simulated data from Exp . 2 with a fixed noise and signal strength , and varied the regularization coefficient systematically . While R c v 2 is highly sensitive to the choice of the regularization coefficient , the correlation criterion is not . Because the regularization coefficient is determined separately for each cross-validation fold and model , deviations from the optimal ridge will decrease model selection accuracy for the cross-validated R c v 2 criterion , but not for the correlation criterion . In sum , using regularization improves model selection performance , even if the encoding model has fewer features than conditions or measurements . Rather than just comparing subspaces , the implicit prior on the weights means that a more specific hypothesis is being tested . From this perspective , it is unsurprising that we can adjudicate between these hypotheses with greater accuracy . Furthermore , the use of correlation instead of the predictive R c v 2 makes model selection more robust against variations in the regularization coefficient . When evaluating models with RSA , there is no need to restrict the model to a specific number of features—the second-moment matrix from all features can determine the predicted distances . As an empirical dissimilarity measure , we used the crossnobis estimator [32] and compared the predicted to the measured RDM . To select the winning model , we used rank-based correlation of dissimilarities [27] , Pearson correlation , correlation with a fixed intercept ( Eq 24 ) , and the likelihood of the observed distances under the normal approximation ( Eq 26 ) using the full variance-covariance matrix of the estimated dissimilarities . For Experiment 1 ( Fig 7 ) , rank-based correlation performed substantially worse than the other criteria . The lower performance of rank correlation may have been exacerbated here by the fact that the two models predict relatively similar dissimilarity ranks . However , we expect lower performance for rank correlation in general , because this approach does not use all the information in the measured RDMs . It forgoes the assumption of a linear relationship between predicted and measured dissimilarities and therefore does not exploit the information in the continuous magnitudes of the dissimilarities . Likelihood-based RSA yielded the best decisions; slightly better than Pearson correlation and fixed-intercept correlation . The advantage of the likelihood-based approach was clearer for Exp . 2 and 3 . Here , it led to about 10 percentage points greater accuracy of the decisions than the next-best RSA approach . This advantage is likely due to the fact that Pearson correlations and especially fixed-intercept correlations ( Eq 24 ) are sensitive to the observed value for the largest predicted dissimilarities , as these data points have a large leverage on the estimated regression line . Indeed , some of the models for Exp . 2 and 3 contain a few especially large dissimilarities , which will influence the model fit strongly . The likelihood-based approach incorporates the knowledge that large dissimilarities are measured with substantially larger variability [29] , and hence discounts their influence . Notably , rank-based correlation performed relatively well on these models as compared to Pearson correlation , likely because rank correlation is robust to outliers and less dominated by the large predicted distances . In sum , these simulations show that it is advantageous to take the covariance structure of the measured dissimilarities into account whenever the additional assumptions this requires are justified . In the same simulations , we also applied the direct likelihood-ratio test , as implemented by PCM . As all the assumptions of the generative model behind PCM are met in the simulation , we would expect , by the Neyman-Pearson lemma [24] , that this method should provide us with highest achievable model selection accuracy . Model selection performance ( dotted line in Fig 7 ) was indeed systematically higher than for the best RSA-based method . For direct comparison of the so far best methods—PCM , likelihood-based RSA , and encoding analysis with regularization ( using correlations as a model selection criterion ) —we simulated the three Experiments at a single signal strength ( Fig 8 ) . In this simulation , PCM resulted in 1 . 48 , 3 . 01 and 2 . 86 percentage points ( for Exp . 1-3 , respectively ) better model selection accuracy than likelihood-based RSA , and 1 . 98 , 1 . 17 and 0 . 85 percentage points higher model selection accuracies than an encoding analysis using correlations . PCM never performed worse than another method and performed significantly better than the other two approaches in 4 of 6 total comparisons across the three experiments ( Fig 8 ) . There were no significant performance differences between RSA and encoding analysis . Overall , however , all three methods were very close in performance . A practical concern is the speed at which the model comparison can be performed . This is usually not important when evaluating the model fit on a small number of participants or ROIs . However , if a larger number of models is evaluated continuously over the cortical surface using a searchlight approach [52 , 53] , or in data sets with large numbers of participants , computational cost becomes a practical issue . While we cannot treat this issue exhaustively , we provide here a brief overview over the computation time required for the three methods for our specific examples and implementation . In general , the computation time will of course depend on the number of conditions , the number of channels , the exact variant of each technique . Our goal here is simply to give the reader a starting point for making a choice for a particular application , trading off computational and statistical efficiency . Both RSA and PCM operate on the inner product matrix of the activity estimates , thus the computational costs for these approaches is virtually independent of the number of voxels . PCM works on the MK × MK inner product matrix of the activity estimates , whereas RSA operates on a K × K matrix of distances between conditions . For a small number of conditions , this explains the favorable computational cost of RSA . However , when using likelihood-based RSA , the covariance matrix of the distances needs to be calculated and inverted . The size of this matrix is ( K ( K-1 ) /2 ) 2 and it therefore grows with the 4th power of the number of conditions K . For Exp . 3 ( Fig 8F ) with K = 96 , this is computationally costly , whereas PCM still only needs to invert matrices of size ( MK ) 2 . Using RDM-correlation-based model selection ( whether rank , Pearson or fixed-intercept ) , RSA is much more computationally efficient ( not shown ) . For encoding models , conducting the actual ridge regression for each cross-validation fold ( dark blue area ) is extremely fast and efficient . The main cost of the technique lies in the determination of the optimal ridge coefficient ( light blue area ) . In our simulations , we use restricted maximum likelihood estimation ( Eq 18 ) to do so—therefore this cost is always M times higher than for PCM alone . Depending on the implementation , generalized cross-validation [46] may offer a considerable speed-up . If very high speeds are required , one could use a constant ridge coefficient and accept the possible loss in model selection accuracy . In sum , while PCM is computationally feasible across the three experiments , encoding models were less efficient in the present implementation and likelihood-based RSA was less efficient than PCM for the condition-rich scenario of Experiment 3 . Alternative variants of encoding models ( with fixed ridge coefficient ) and RSA ( with correlation-based model selection ) are less statistically efficient , but beat PCM in terms of computational efficiency .
There is a fundamental difference between encoding models with and without weight priors . Without a prior on the feature weights , encoding models test how well the subspace spanned by the model features captures the observed activity profiles . For models to be discriminable , the dimensionality ( i . e . the number of features ) of each model must be substantially lower than the number of experimental conditions . As the number of model dimensions increases , the subspaces of competing models increasingly overlap . Once the number of features matches the number of experimental conditions , their subspaces comprise the entire space of activity profiles , each perfectly fits the training data , and their predictions for unseen data become identical . A subspace specifies what activity profiles are possible and what activity profiles are impossible ( though they might still arise as estimates because of the noise ) . A subspace might be conceptualized as an infinite flat distribution over the subspace dimensions , with 0 probability outside the subspace . However , a uniform distribution on an infinite interval has an infinite second moment and hence does not specify the neural representation uniquely . L2-norm regularization ( i . e . ridge regression ) is equivalent to imposing a Gaussian prior on the regression weights . With such a prior , the representational model specifies a probability distribution with a finite second moment . When changing the form of regularization , one also changes the implicit prior , and hence the representational model that is being tested . Thus , regularization is not simply a trick for stabilizing the fit . Instead , the weight prior forms an integral part of the model , which determines the strength with which each feature is encoded according to the model . Choosing a specific form of regularisation therefore constitutes a decision about the neuroscientific hypothesis to be tested rather than a methodological consideration . Encoding models do not support inferences about the particular feature set generating a representation , because infinitely many feature sets can span the same space . Even when using a prior , the feature set that characterizes a given representational model is not unique . Features should not in general be constrained to be orthogonal in the space of experimental conditions , because the structure of the model is not usually meant to depend on the experimental conditions chosen . Whether the features chosen are orthogonal or not , there is an infinite number of basis sets of features that express the same representational model ( inducing the same second moment of activity profiles , Eq 3 ) . For example , two equally long correlated feature vectors can equally well describe a distribution with elliptical isoprobability-density contours ( Fig 3A ) as two orthogonal features , with one vector longer than the other . Thus , when one representational model is shown to be superior to others , it does not imply anything special about the feature set chosen to express that model . These complications need to be kept in mind in the interpretation of the results of encoding model analyses . It is very tempting to attribute meaning to the particular features , especially when they are mapped onto the cortical surface [17 , 21] . When interpreting these maps , one needs to remember that a feature set only describes a distribution of activity profiles , and that very different maps can emerge when the same distribution is described by a rotated set . In PCM and RSA , the equivalence of different feature sets is made explicit , as they lead to the same second-moment and representational dissimilarity matrices . When using RSA to test representational models , the crossnobis estimator provides a highly reliable measure of dissimilarity with the added advantage of having an interpretable zero-point [32] . Rank-based , Pearson , and fixed-intercept correlation provide fast and straightforward ways of measuring the correspondence between predicted and observed distances , so as to select the representational model most consistent with the data . However , using simple correlations ignores the dependence of the distance estimates , as well as their unequal variances . In other words , the sampling distribution of the estimated RDM in the space spanned by the dissimilarities ( one dimension per pair of conditions ) is not isotropic . This problem is addressed in likelihood-based RSA , which uses a multivariate-normal approximation to the sampling distribution of the crossnobis RDM estimate [29] . The approximation provides an analytical expression for the statistical dependency of distance estimates , as well as their signal-dependent variances . In the simulations , likelihood-based RSA was shown to be more powerful than correlation-based RSA . Its model-selection accuracy was only slightly below the theoretical upper bound , as established by PCM . Likelihood-based RSA might therefore become the approach of choice when comparing representational models using crossnobis estimates . There are situations , however , in which the models are not specific enough to support ratio-scale predictions of representational dissimilarities . Moreover , for measurement modalities like fMRI , it might be undesirable to assume a linear relationship between predicted and measured representational dissimilarities . Rank-correlation-based RSA [25 , 27] provides a robust method that is not dependent on the assumption of a linear reflection of the underlying neural dissimilarities in the data RDM . It is also more computationally efficient in the context of condition-rich designs . Likelihood-based RSA becomes computationally expensive as the number of conditions increases . A practical compromise might be to only use the diagonal of the variance-covariance matrix , which would dramatically reduce computational complexity at the expense of neglecting dependencies among dissimilarity estimates . For all simulations , model selection using PCM [22] was better than competing methods . This is not surprising , as the data were simulated exactly according to the generative model underlying this approach ( Gaussian distribution of noise and signal , independence across voxels ) . In this case , PCM implements the likelihood-ratio test , which by the Neyman-Pearson lemma [24] is the most powerful test . Beyond confirming what we know from theory , the simulations were important because they revealed how close the other two techniques come to the theoretical upper bound established by PCM . Results showed that encoding models with a prior and likelihood-based RSA perform near-optimally . In practice , we therefore expect these three techniques to provide similar answers . When its assumptions hold , PCM has clear advantages for model comparison , providing optimal power at reasonable computational cost . However , the other two techniques have other advantages that make them attractive for specific applications . RSA using RDM correlation for model selection gives up statistical efficiency for computational efficiency , and beats PCM at the latter . When rank correlation is used to compare RDMs , the inference does not rely on a linear relationship between the true dissimilarities and the estimated dissimilarities , an assumption that might be violated in many contexts . RSA also provides readily interpretable intermediate statistics ( cross-validated distances ) , which are closely related to linear decoders for all pairs of stimuli . These statistics can be used to test whether two conditions have different activity patterns [27 , 29] , or whether the dissimilarity is larger for one pair than for another pair of conditions . Multidimensional scaling of the stimuli on the basis of their representational dissimilarities also provides an intuitive visualization of the representational structure [25] , which can be very helpful in the generation of novel representational hypotheses . In contrast , PCM sometimes demands complicated approaches to answer simple questions: For example , to test the hypothesis that a difference between two conditions is encoded , one would need to fit one model that allows for separate patterns and one model that does not—and then compare the marginal likelihood of these models . Furthermore , PCM requires the noise to be explicitly modeled , whereas RSA removes the bias arising from noise through cross-validated distances . Encoding analysis explicitly estimates the first-level parameters that describe the response for each individual voxel . This enables the mapping of the estimated features onto the cortical surface to study their spatial distribution [17 , 21] . In sum , the three methods are deeply related in that they test hypotheses about the second moment of the activity profiles . However , each method constitutes a unique perspective on the data and supports different kinds of exploratory analyses . We view the methods as complementary tools that are part of a single coherent toolkit for analyzing representations . An important issue , which we have not touched upon so far , is whether to perform model comparison on single or multiple voxels . While RSA and PCM are usually applied to groups of voxels ( such as for ROIs or searchlights ) , encoding models are often compared on the single-voxel level . This tendency , however , is not strictly inherent in methodological constraints: The searchlight approach for RSA and PCM can be reduced to single voxels , and encoding models can be combined with multi-voxel searchlights . Analyses with coarser granularity give up some spatial precision of the map in exchange for greater statistical power . Searchlight mapping boosts power ( 1 ) by locally combining the evidence , ( 2 ) by enabling the use of a multivariate noise normalization , and ( 3 ) by reducing the effective number of multiple comparisons [54] . There is no reason to assume that a single-voxel searchlight is always the optimal choice when balancing spatial precision and power . Based on our previous results [32] , we expect that ignoring voxel dependencies will entail a loss of sensitivity when making inferences on representational models for regions of interest comprising multiple responses . Whenever a model is fitted using experimental data , its parameters will necessarily be overfitted to the data to some extent . Assessing the performance of a fitted model therefore requires independent test data . An important question is whether the test data should consist in independent measurements for the same experimental conditions or in measurements for a fresh sample of experimental conditions ( e . g . a different sample of visual images ) . The simple answer is that it depends on the inference we would like to make . If our hypothesis is restricted to the present set of conditions ( e . g . five finger movements ) , we need only account for overfitting to the noise in the data and require different measurements for the same conditions . If our hypothesis is about a population of conditions ( e . g . all natural images ) , we need to account for overfitting to the condition sample and require measurements for an independent random sample of conditions from the population of conditions covered by our hypothesis . However , overfitting only needs to be accounted for when the model being tested had parameters fitted in the first place . Encoding models always require independent test sets to account for the over-fitting of the first-level parameters of the representational model ( feature weights ) . RSA and PCM , by contrast , rely explicitly on summary statistics of the responses . Therefore , only second-level parameters related to the strength of the signal and noise need to be fitted ( see Table 1 ) . Because the representational models considered here had the same number of such second- level parameters , they could be compared directly . Decoding is widely used in multivariate analysis of brain imaging data [11–13] . Can it serve us also as a tool for comparing representational models ? While one can use standard decoding approaches to determine whether specific features are represented in an area or not , it does not lend itself to the comparison of full representational models ( as defined here ) . Representational models determine ( via the second moment matrix ) the decodability of any linear feature , not just a restricted set of features . This is most obvious in RSA , where the RDM assembles all pairwise condition discriminabilities . It is of course possible to use decoding in the context of the methods considered here . For example , some studies have used encoding models to decode stimuli [15 , 16 , 21] . Decoding accuracy on held-out data can then serve , instead of correlation or R c v 2 , as a performance evaluation of an encoding model . While this approach is motivated by the intuitive demonstration of mind reading , it does not provide a particularly natural or powerful approach to adjudicating between representational models . Alternatively , we could use classification accuracy as a measure of dissimilarity between two conditions in the context of RSA [55] . However , classification essentially converts a continuous measure of dissimilarity into a binary label of correct / incorrect . It is therefore expected to be less informative than the underlying continuous measure , and we have shown previously that this entails a loss of sensitivity in practice [32] . In sum , decoding is not particularly useful for the evaluation of representational models [14 , 23] and should therefore be limited to situations , in which the quality of the decoding itself is the measure of interest . All models considered here were “fixed” , i . e . , they did not include free parameters that would change the predicted second-moment matrix . In many applications , however , the relative importance of different features ( for example encoding strength for orientation and color ) are unknown . In this case , the predicted second moment can be expressed as the weighted sum of different pattern components , i . e . G = ∑i ωi Ci [22 , 56–58] , with the weights being free second-level parameters . In other situations , G is a nonlinear function of free model parameters: For example , G depends non-linearly on the spatial tuning width in population receptive field modeling [59] . Both RSA and PCM already provide a mechanism to estimate such parameters , as both approaches need to estimate the signal strength parameters s by maximizing the respective likelihood function ( Eqs 17 and 28 ) —and the analytical derivatives of the likelihood ( Eqs 17 and 28 ) with respect to the parameters are easily obtained . In the context of encoding approaches using ridge regression , free model parameters that change the model structure would result in independent scaling of different features , rotations , or extensions of the model matrix M . At the time of writing there are no published examples of such parameter optimization in the context of cross-validated encoding models that we know of . The inclusion of free parameters into the model also enables the specification of measurement models . Representational models ideally test hypotheses about the distribution of activation profiles of the core computational elements—i . e . neurons . When using indirect measures of brain activity such as fMRI or MEG , the distribution of activity profiles across measurement channels is also influenced by the measurement process , which samples and mixes neuronal activity signals in the measurement channels [30 , 60–63] . As the underlying brain computational models become more specific and detailed , the corresponding measurement models will also have to be improved . We focused on approaches that characterize the distribution of activity profiles by its second moment . If the true distribution of the activity profiles is a multivariate Gaussian , then the second moment fully defines the distribution of activity profiles . However , a representational hypothesis may not only predict that the response for condition A is uncorrelated to the response for condition B , but , for example , that channels either respond to A or B , but not to both A and B . Such tuning is for example prevalent in primary visual cortex , where neurons ( and voxels ) respond to a stimulus in a one specific part of the visual field , but less often two or more disparate locations [59] . This would correspond to a non-Gaussian prior on the feature weights . In a recent publication , Norman-Haignere and colleagues [64] suggested a likelihood-based method , in which the Gaussian prior on the feature weights W is replaced with a Gamma distribution , essentially providing a non-Gaussian extension of PCM . It will be interesting to determine to what degree such non-Gaussian distributions are present in fMRI or single-cell data , and what computational function these may serve . It is important to stress that the approaches considered here are still appropriate when the distribution of activity profiles is truly non-Gaussian . Even in the non-Gaussian case , the second moment determines the representational geometry and thus the decodability of all possible features . It therefore remains essential for characterizing the representation . Taking into account higher moments of the activity profile distribution would enable us to distinguish between representations that afford the same decoding of features ( assuming that readout neurons have access to the entire code ) , but achieve this by distinct population codes . If advances in brain-activity measurements are to yield theoretical insights into brain computation , they need to be complemented by analytical methods to test computational models of information processing [65] . The main purpose of this paper was to provide a clear definition of one important class of models—representational models—and to compare three important approaches of testing these . We have shown that PCM , RSA and encoding analysis are all closely related , testing hypotheses about the distribution of activity profiles . Moreover , all three approaches , in their dominant implementations , are sensitive only to distinctions between representations that are reflected in the second moment of the activity profiles . Thus , these three methods are properly understood as components of a single analytical framework . Each of the three methods has particular advantages and disadvantages and preferred areas of application . We hope that the general framework presented here will enable researchers to combine these approaches to make progress revealing the computational mechanisms of biological brains . | Modern neuroscience can measure activity of many neurons or the local blood oxygenation of many brain locations simultaneously . As the number of simultaneous measurements grows , we can better investigate how the brain represents and transforms information , to enable perception , cognition , and behavior . Recent studies go beyond showing that a brain region is involved in some function . They use representational models that specify how different perceptions , cognitions , and actions are encoded in brain-activity patterns . In this paper , we provide a general mathematical framework for such representational models , which clarifies the relationships between three different methods that are currently used in the neuroscience community . All three methods evaluate the same core feature of the data , but each has distinct advantages and disadvantages . Pattern component modelling ( PCM ) implements the most powerful test between models , and is analytically tractable and expandable . Representational similarity analysis ( RSA ) provides a highly useful summary statistic ( the dissimilarity ) and enables model comparison with weaker distributional assumptions . Finally , encoding models characterize individual responses and enable the study of their layout across cortex . We argue that these methods should be considered components of a larger toolkit for testing hypotheses about the way the brain represents information . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"statistical",
"noise",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"engineering",
"and",
"technology",
"signal",
"processing",
"neuroscience",
"simulation",
"and",
"modeling",
"magnetic",
"resonanc... | 2017 | Representational models: A common framework for understanding encoding, pattern-component, and representational-similarity analysis |
The histone H2A variant H2A . Z is essential for embryonic development and for proper control of developmental gene expression programs in embryonic stem cells ( ESCs ) . Divergent regions of amino acid sequence of H2A . Z likely determine its functional specialization compared to core histone H2A . For example , H2A . Z contains three divergent residues in the essential C-terminal acidic patch that reside on the surface of the histone octamer as an uninterrupted acidic patch domain; however , we know little about how these residues contribute to chromatin structure and function . Here , we show that the divergent amino acids Gly92 , Asp97 , and Ser98 in the H2A . Z C-terminal acidic patch ( H2A . ZAP3 ) are critical for lineage commitment during ESC differentiation . H2A . Z is enriched at most H3K4me3 promoters in ESCs including poised , bivalent promoters that harbor both activating and repressive marks , H3K4me3 and H3K27me3 respectively . We found that while H2A . ZAP3 interacted with its deposition complex and displayed a highly similar distribution pattern compared to wild-type H2A . Z , its enrichment levels were reduced at target promoters . Further analysis revealed that H2A . ZAP3 was less tightly associated with chromatin , suggesting that the mutant is more dynamic . Notably , bivalent genes in H2A . ZAP3 ESCs displayed significant changes in expression compared to active genes . Moreover , bivalent genes in H2A . ZAP3 ESCs gained H3 . 3 , a variant associated with higher nucleosome turnover , compared to wild-type H2A . Z . We next performed single cell imaging to measure H2A . Z dynamics . We found that H2A . ZAP3 displayed higher mobility in chromatin compared to wild-type H2A . Z by fluorescent recovery after photobleaching ( FRAP ) . Moreover , ESCs treated with the transcriptional inhibitor flavopiridol resulted in a decrease in the H2A . ZAP3 mobile fraction and an increase in its occupancy at target genes indicating that the mutant can be properly incorporated into chromatin . Collectively , our work suggests that the divergent residues in the H2A . Z acidic patch comprise a unique domain that couples control of chromatin dynamics to the regulation of developmental gene expression patterns during lineage commitment .
Precise control of gene expression is critical for lineage commitment and proper development in all multicellular organisms . Regulation of chromatin structure has emerged as an important mechanism for modulating gene expression patterns in response to developmental cues . While post-translational histone modifications can influence chromatin structure and transcriptional activity , less is known about the role of histone variants . Histone variants are incorporated in a replication-independent manner and appear to mark structurally and functionally distinct chromatin domains [1]–[3] . The histone H2A variant H2A . Z is highly conserved among eukaryotes and is of particular interest because it plays an essential but poorly understood role in metazoan development including mammals [4]–[6] . H2A . Z has been implicated in a range of DNA-mediated processes such as gene expression , DNA repair , and genomic stability [7]–[9] . Notably , H2A . Z is required for proper execution of developmental gene expression programs during embryonic stem cell ( ESC ) differentiation [10] , suggesting that H2A . Z has specialized functions to regulate lineage commitment . A role for H2A . Z in gene regulation is supported by genome-wide localization studies showing that this variant flanks the nucleosome-free region at transcription start sites in a wide range of cell types [11] , [12] . In particular , H2A . Z is incorporated at the majority of H3K4me3 modified promoter nucleosomes including bivalent promoters in ESCs that harbor both H3K4me3 and H3K27me3 , marks of Trithorax and Polycomb , respectively [10] , [11] . Bivalent promoters in ESCs are associated with lineage specific genes that are poised , but remain competent for activation [13] , [14] . These studies suggest that H2A . Z contributes to formation of distinct chromatin states and that its incorporation at bivalent promoters may be necessary to allow for induction of lineage programs in response to developmental cues . Consistent with this idea , H2A . Z levels decreased at promoters upon gene activation [11] , [15]–[17] . H2A . Z also occupied regulatory elements such as enhancers and boundary elements [11] , [18] , [19] , indicating that H2A . Z is incorporated at regions that are subject to considerable chromatin regulation . H2A . Z shares extensive homology with the major type histone H2A throughout the histone fold domain . However , divergent regions in the amino- and carboxy-terminal domains as well as the L1 loop region within the histone fold suggest that the two histones have different structural and functional properties . In vitro biophysical studies showed that H2A . Z incorporation stabilizes the dimer-tetramer interface and strongly favors formation of 30 nm fibers over formation of higher order chromatin folding that require fiber-fiber interactions when compared to canonical H2A , consistent with the idea that unique features of the variant contribute to specialized chromatin domains [20]–[22] . Moreover , H2A . Z and the histone H3 variant H3 . 3 can occupy the same nucleosome resulting in a double variant nucleosome that is enriched at active promoters as well as at highly regulated chromatin regions [23] , [24] . These hybrid nucleosomes are characterized as unstable and highly salt labile , consistent with its presence in dynamic chromatin domains [25] , [26] . Thus , dissecting the features of H2A . Z that distinguish it from core H2A is key to understanding its functional specialization and may provide new insights into the essential role of this variant during development . The H2A . Z carboxy-terminal acidic patch comprises substitutions of H2A residues Asn89 , Asn94 , Lys95 for Gly92 , Asp97 , and Ser98 . While the overall structure of the H2A . Z nucleosome appears similar to H2A-containing nucleosomes , these three divergent residues form an uninterrupted acidic patch that extends across the surface of the H2A . Z octamer resulting in a solvent-exposed cavity in the center of the nucleosome [27] . In Drosophila , domain-swap experiments demonstrated that the H2A . Z carboxy-terminus including the acidic patch is essential for development [28] . Later work in Xenopus involving site-specific mutagenesis of the divergent H2A . Z acidic patch residues resulted in embryos that exhibited significant developmental defects [6] . These studies suggest that the unique H2A . Z acidic patch plays an important role in establishing a novel chromatin state that is essential for embryonic development . In yeast , replacement of two residues in the H2A . Z acidic patch resulted in low nucleosome occupancy at the PHO5 promoter , suggesting that the acidic patch is necessary for interaction with its deposition complex and for proper incorporation into chromatin [29] . However , the amino acids mutated in this study are conserved between H2A . Z and canonical H2A , indicating that this phenotype is not likely specific to H2A . Z . Other loss-of-function studies in yeast have shown that the H2A . Z carboxy-terminal docking domain is critical for H2A . Z function; however , these studies examined truncations that retained an intact acidic patch domain [30] , [31] . Thus , we currently lack a detailed mechanistic understanding of how this domain regulates chromatin conformation during development and whether it plays a similar role in mammals . We investigated the role of the divergent H2A . Z acidic patch during ESC differentiation . ESCs are an ideal model for investigating how H2A . Z influences mammalian development because these cells maintain the potential to differentiate into all somatic cell types [32] , [33] . We generated a mutant form of H2A . Z ( denoted H2A . ZAP3 ) , where the three divergent acidic patch residues in H2A . Z are replaced with the corresponding H2A amino acids . H2A . ZAP3 ESCs maintained the ability to self-renew , but these cells failed to differentiate properly . We found that while H2A . ZAP3 interacted with its deposition complex and displayed a highly similar distribution pattern compared to expression of the wild-type H2A . Z transgene ( denoted H2A . ZWT ) , its enrichment levels were reduced at target promoters and were particularly diminished at the +1 nucleosome . Further analyses revealed that H2A . ZAP3 was less tightly associated with chromatin compared to H2A . ZWT suggesting that the mutant is more dynamic . Notably , bivalent genes that are poised for activation in ESCs displayed significant changes in expression compared to active genes suggesting that the poised state is more sensitive to H2A . Z regulation . Moreover , this group of genes showed reduced levels of the repressive chromatin mark H3K27me3 at H2A . ZAP3 target gene promoters compared to H2A . ZWT and displayed higher H3 . 3 enrichment , a variant associated with high chromatin flux . Consistent with this observation , we further showed that H2A . ZAP3 displayed higher mobility in chromatin compared to H2A . ZWT by fluorescence recovery after photobleaching ( FRAP ) . Remarkably , ESCs treated with the transcriptional inhibitor flavopiridol partially restored the H2A . ZAP3 mobile fraction to wild-type levels and resulted in an increase in H2A . ZAP3 occupancy at target genes . Collectively , our results demonstrate that the divergent H2A . Z acidic patch mediates chromatin dynamics and indicate that control of H2A . Z dynamics is critical for the regulation of gene expression patterns during lineage commitment .
The H2A acidic patch domain resides on the nucleosome surface , and in the case of H2A . Z contains three divergent residues that comprise an extended acidic patch [27] ( Figure 1A ) . While disruption of the H2A . Z acidic patch results in early developmental defects in Drosophila and Xenopus , how it functions to regulate chromatin structure and whether this domain plays a similar role in mammals is unknown [6] , [28] . In mouse , H2A . Z is encoded by two isoforms that differ by only 3 amino acid residues , denoted H2AFZ ( H2A . Z ) and H2AFV [34] , [35] ( Figure S1A ) . H2A . Z knockout mice die around the time of implantation [5] , suggesting that H2AFV is unable to compensate for loss of H2A . Z and that the two isoforms are functionally distinct . We analyzed the relative abundance of H2A . Z and H2AFV in ESCs by mass spectrometry . While we detected both isoforms in ESCs , H2A . Z is ∼20-fold more abundant than H2AFV ( Figure S1B ) . Thus , given the essential role of H2A . Z and its abundance in ESCs compared to H2AFV , we focused on dissecting the function of the divergent acidic patch in H2A . Z . Since H2A . Z and H2AFV share nearly identical amino acid sequences , available antibodies cannot distinguish between the two isoforms . To circumvent this limitation , we generated ESC lines that harbor a stably integrated Tet-inducible H2A . Z transgene fused to YFP ( Figure 1B ) . Upon induction by doxycycline , we sorted for YFP expression and collected cells that displayed transgene expression comparable to endogenous H2A . Z levels for further analysis . To specifically test transgene function , we integrated short hairpins directed against the H2A . Z 3′ UTR into the inducible ESC lines ( Figure 1B ) . This system allowed for targeted depletion of endogenous H2A . Z ( denoted H2A . ZKD for knock-down ) ( Figure 1C , Figure S1D ) without affecting transgene levels or H2AFV expression ( Figure 1D , Figure S1C ) . Consistent with our previous work , H2A . ZKD ESCs displayed typical colony morphology , cell cycle kinetics , and normal expression of the pluripotency marker Pou5f1 ( Oct4 ) suggesting that H2A . Z depletion does not affect self-renewal ( Figure 1E , Figure S1E , F ) . To test differentiation potential , wild-type H2A . Z-YFP ( denoted H2A . ZWT ) and H2A . ZKD ESCs were induced to differentiate by allowing these cells to aggregate into embryoid bodies ( EBs ) in the absence of the pluripotency growth factor LIF . EBs are similar to egg cylinder stage embryos , albeit disorganized , and are capable of differentiation into tissues representing the three germ layers . H2A . ZKD ESCs failed to differentiate properly , and lacked the distinct differentiated structures observed in Day 10 EBs compared to H2A . ZWT ( Figure 1F ) . Additionally , EBs from H2A . ZKD ESCs failed to activate developmental genes to levels observed in H2A . ZWT cells ( Figure 1I ) , consistent with the idea that H2A . Z regulates lineage programs [11] . Importantly , expression of H2A . ZWT rescued the H2A . ZKD phenotype as measured by the restoration of ESC gene expression patterns as well as their capacity for multi-lineage differentiation ( Figure 1F , I ) , whereas H2A-YFP did not compensate for loss of H2A . Z ( data not shown ) . These data indicate that expression of the H2A . ZWT transgene recapitulates normal H2A . Z function . Thus , we used this system to dissect the role of the H2A . Z acidic patch during lineage commitment . We replaced the divergent H2A . Z acidic patch residues by site-directed mutagenesis of Gly92 , Asp97 , and Ser98 to the equivalent residues in H2A- Asn89 , Asn94 , and Lys95 ( Figure 1A ) . Similar to H2A . Z depletion , expression of the acidic patch mutant ( denoted H2A . ZAP3 ) in H2A . ZKD ESCs did not affect self-renewal , colony morphology , or levels of the pluripotency marker OCT4 ( Figure 1G ) . We observed , however , that H2A . ZAP3 EBs were smaller , morphologically distinct , and failed to differentiate properly compared to H2A . ZWT EBs as demonstrated by the lack of differentiated cell types at Day 10 and the inability to activate developmental gene expression programs during lineage commitment ( Figure 1H , I ) . The smaller size of H2A . ZAP3 EBs was not a result of altered cell cycle kinetics , proliferation , or differences in levels of apoptotic cells relative to H2A . ZWT ( Figure S1E–I ) . Moreover , the number of cells recovered from Day 10 H2A . ZAP3 EBs was comparable to that recovered from H2A . ZWT , and H2A . ZKD EBs ( Figure S1J ) . Rather , we observed a larger number of small EBs in H2A . ZAP3 cultures compared to H2A . ZWT . Taken together , these data suggest that the divergent residues in the H2A . Z acidic patch are necessary for proper ESC differentiation . H2A . Z containing nucleosomes occupy majority of promoters from yeast to human as determined by genome-wide localization studies and its incorporation is critical for proper gene regulation [10] , [11] , [36] . Thus , we analyzed the localization pattern of the acidic patch mutant relative to H2A . Z . Given that current H2A . Z antibodies cannot distinguish between H2A . Z and H2AFV isoforms , we performed ChIP-Seq in ESCs using GFP antibodies that recognize the H2A . Z-YFP tag . This analysis revealed significant H2A . Z-specific enrichment at ∼11 , 000 promoters in ESCs ( Table S1 ) . We found by ChIP-Seq that H2A . Z enrichment overlapped with the majority of H3K4me3 promoters , a histone modification that is normally associated with transcriptional competence ( Figure 2A ) [11] , [37] . Our H2A . Z data are similar to other genome wide reports using pan H2A . Z antibodies in ESCs [11] , [36] , indicating that the YFP tag does not affect H2A . Z incorporation . In addition to its enrichment at active promoters , H2A . Z was enriched at bivalent promoters consistent with our prior ChIP-chip analysis [10] . Bivalent promoters are characterized by the enrichment of H3K4me3 and H3K27me3 marks , and these genes display low expression levels and reduced enrichment of RNA Polymerase II ( RNAP2 ) ( Figure 2A , B top ) [13] , [14] . In contrast , H2A . Z enrichment at active promoters ( H3K4me3-only ) was coincident with a strong RNAP2 peak at the TSS and higher expression of associated genes ( Figure 2A , B bottom ) . Our analysis also revealed a bimodal H2A . Z distribution pattern around TSSs with a marked enrichment at the +1 nucleosome ( Figure 2C ) , consistent with reports showing that H2A . Z flanks the nucleosome-depleted region [11] . Notably , we observed a broader distribution pattern of H2A . Z and H3K4me3 at bivalent genes compared to their enrichment at active genes suggesting that the chromatin structure differs at these two classes of promoters . In addition to promoters , we found that H2A . Z was enriched at a subset of distal enhancers identified in ESCs [38] ( Table S1 ) , similar to recent data using a pan-H2AZ antibody [36] . The enrichment at promoters as well as distal regulatory elements suggests that H2A . Z is incorporated at regions subject to considerable chromatin regulation . We next analyzed H2A . ZAP3 localization across the ESC genome by ChIP-Seq . While H2A . ZAP3 enrichment was globally decreased , it occupied a highly similar set of promoter regions as well as distal enhancers ( Figure 2C , Table S1 ) . For example , H2A . ZAP3 displayed a similar overall spatial pattern of enrichment , albeit reduced compared to H2A . ZWT as shown across the large bivalent region encompassing the HOXA locus in ESCs ( Figure 2D ) . Importantly , H2A . ZAP3 was expressed at similar levels as H2A . ZWT ( Figure 1D , Figure S2A , B ) , indicating that reduced H2A . ZAP3 enrichment was not due to its lower abundance in ESCs . Interestingly , we observed a dramatic reduction in H2A . ZAP3 enrichment downstream of the TSS in ESCs , which is thought to mark the +1 nucleosome ( Figure 2C ) . A number of studies have suggested that the +1 nucleosome possesses significant regulatory potential and that remodeling of this promoter nucleosome may be important for controlling gene expression by recruiting RNAP2 or by facilitating transcriptional elongation [39] , [40] . Collectively , these data suggest that altered levels of H2A . ZAP3 as well as a particular reduction in the +1 nucleosome at promoters have consequences on the regulation of gene expression states . One possibility for the observed lower levels of H2A . ZAP3 is that replacement of the divergent amino acids with those of H2A lead to its incorporation via a similar pathway as core histones . We expected that if H2A . ZAP3 is more broadly distributed along chromosomes , then its overall chromatin-associated fraction would be similar or higher relative to H2A . ZWT . To this end , we performed chromatin fractionation and probed for H2A . ZWT and H2A . ZAP3 using GFP antibodies in the respective transgenic ESC line ( Figure S2B ) . Semi-quantitative immunoblots showed that the fraction of H2A . ZWT associated with chromatin is approximately 1 . 85 fold higher than H2A . ZAP3 suggesting that depletion of the mutant at TSSs does not lead to its random accumulation in chromatin ( Figure 2E ) . Given recent evidence demonstrating that H2A . Z is redistributed during the cell cycle from promoters to heterochromatin regions in mouse trophoblast stem ( TS ) cells [41] , next we examined the levels of H2A . ZAP3 at these regions . We first analyzed metaphase chromosomes ( a time point when H2A . Z is enriched at heterochromatin in TS cells ) and found that unlike the broad distribution of H2A across the chromosome H2A . ZAP3 ESCs showed a non-uniform distribution pattern and depletion at centromeric regions similar to H2A . ZWT ( Figure S2C ) . Consistent with this observation , we found both H2A . ZAP3 and H2A . ZWT showed overall low enrichment at repetitive elements associated with heterochromatin including major satellite repeats in cycling cells ( Figure S2D ) . Additionally , we found no change in the levels of H3K9me3 , a modification highly enriched in heterochromatin , in H2A . ZWT and H2A . ZAP3 ESCs as well as in Day 5 RA differentiated cells by ChIP-qPCR ( Figure S2E ) . The lower levels of H2A . Z at heterochromatin regions in ESCs compared to TS cells may reflect differences in cell cycle dynamics between the two cell types , as ESCs have a notably short G1 . Together , these data suggest that mutations in the H2A . Z acidic patch do not result in the inappropriate incorporation of H2A . ZAP3 or disruption of heterochromatin . The observed lower levels of H2A . ZAP3 at promoters suggested that the mutant is either more dynamically associated with chromatin or that it is not properly incorporated by its deposition complex . In yeast , deletion of either the carboxy-terminal docking domain or mutation of conserved residues within the acidic patch resulted in low H2A . Z occupancy at target genes [29]–[31] . This decrease in H2A . Z occupancy was attributed to the inability of the carboxy-terminal mutant to interact with its deposition complex SWR1 . Site-specific incorporation of H2A . Z in mammalian cells is accomplished by the ATPase complex SRCAP ( Snf2-related CREBBP activator protein ) , and in yeast H2A . Z removal appears to require the INO80 complex [42] , [43] . Thus , we tested the ability of H2A . ZAP3 to interact with components of both of these complexes . We performed co-immunoprecipitation followed by immunoblot and found that H2A . ZAP3 interacted with SRCAP ( catalytic component of SRCAP ) and RUVBL1 ( component of Tip60 , SRCAP , and INO80 complexes ) similar to H2A . ZWT ( Figure S2F , Figure 2F ) . While it is possible that H2A . ZAP3 renders the deposition complex catalytically inactive , this scenario is unlikely because endogenous H2A . Z enrichment at promoters was unchanged over multiple passages in cells that also expressed the H2A . ZAP3 transgene ( Figure S2G ) . While H2A . Z specific chaperones such as CHZ1 have been identified in yeast [44] , the CHZ1 homolog or chaperones that play similar roles in mammalian cells have not been studied in detail . Nevertheless , NAP1 is also a critical chaperone for histone incorporation in mammals including H2A . Z so we sought to test whether alterations in the acidic patch affected its interaction with this histone chaperone . Thus , we quantified the interaction between NAP1L1 and H2A . ZAP3 , H2A . ZWT , as well as core H2A ( Figure S2H , I ) . We found by co-immunoprecipitation using GFP antibodies that the amount of NAP1L1 bound to H2A . ZAP3 was modestly higher ( <2-fold ) compared to H2A . ZWT ( Figure S2H , I ) . Surprisingly , we observed that lower levels of NAP1L1 co-immunoprecipitated with H2A compared to H2A . ZWT ( Figure S2H , I ) . This result may be due to the much higher levels of stably associated H2A in chromosomes , of which only a fraction of H2A would interact with NAP1L1 unlike the continuous dynamic replacement of H2A . Z . Thus , these data are consistent with the idea that H2A . ZAP3 is more dynamic , resulting in a higher fraction of the mutant available to interact with histone chaperones . We next tested the idea that H2A . ZAP3 is less tightly associated with chromatin compared to H2A . ZWT . To this end , we performed salt titrations on nuclei isolated from H2A . ZWT and H2A . ZAP3 ESCs . This analysis showed that while H2A . ZWT was stably associated with chromatin up to 500 mM NaCl and largely depleted at 1M NaCl ( dimer loss typically occurs between 600–800 nM NaCl ) , a fraction of H2A . ZAP3 dissociates at the lower salt concentrations ( 75 mM–200 mM NaCl ) ( Figure 2G ) , suggesting that H2A . ZAP3 is less tightly associated with chromatin . The finding that H2A . ZAP3 is less tightly associated with chromatin by salt titration is consistent with the higher proportion of H2A . ZAP3 associated with NAP1L1 in ESCs . While we cannot rule out that H2A . ZAP3 incorporation is less efficient or that a small fraction is randomly distributed similar to incorporation of H2A , our data are consistent with the model that the unique H2A . Z extended acidic patch is critical for regulating its dynamic association with chromatin . These data also suggest that control of H2A . Z dynamics is important for regulation of gene expression programs during lineage commitment . H2A . Z knock-out mice die around the time of gastrulation when complex gene expression patterns are established during embryogenesis [5] . Moreover , ESCs depleted of H2A . Z fail to execute developmental gene expression programs when signaled to do so [10] . Furthermore , in vitro biophysical studies showed that an intact H2A . Z acidic patch is necessary for the ability of a nucleosomal template to fold into a 30-nm chromatin fiber and for the efficient repression of transcription [45] , [46] . Thus , we analyzed global gene expression patterns in H2A . ZWT and H2A . ZAP3 cells in ESCs ( Day 0 or D0 ) and at Day 3 ( D3 ) of ESC differentiation ( Table S2 ) . We observed that a subset of genes showed higher expression levels in H2A . ZAP3 at D0 and that many of these genes remained expressed at higher levels in Day 3 H2A . ZAP3 EBs relative to H2A . ZWT ( Figure 3A ) . These de-repressed genes have functions in vasculature development , pattern formation , and embryonic morphogenesis ( e . g . Tbx20 , Hoxb4 , Foxc1 ) ( Figure 3A ) . Genes that displayed higher expression levels in H2A . ZAP3 specifically at D3 , function in steroid biosynthesis , signaling , and growth ( e . g . Cyp51 , Mvd , Wnt5a ) . In contrast , genes with lower expression levels at D3 in H2A . ZAP3 have roles in differentiation and transcription regulation ( e . g . Notch4 , Spag1 , Neurod1 , Wnt5a ) . These results indicate that H2A . ZAP3 incorporation leads to significant changes in the expression of genes with important developmental functions . While H2A . Z is enriched at the majority of H3K4me3 marked promoters , we previously observed that bivalent genes exhibited significant changes upon H2A . Z depletion [10] . In ESCs , H2A . Z incorporation at bivalent promoters is required for precise regulation of developmental programs during the initial stages of lineage commitment . Notably , genes that showed changes in H2A . ZAP3 ESCs comprised a large cohort of developmental regulators ( Figure 3A ) . Thus , we further tested the connection between H2A . Z and the regulation of bivalent genes . We compared the differentially regulated genes in H2A . ZAP3 cells ( D0 and D3 ) with genes containing either bivalent ( H3K4me3 and H3K27me3 ) or active ( H3K4me3 only ) histone marks . We found a significant overlap between genes that are differentially regulated and genes with bivalent marks ( P<10−100 , Figure 3B ) , whereas no significant overlap was observed with H3K4me3 only ( active ) genes ( Figure S3A , top ) . Reciprocally , we grouped target genes as active or bivalent according to histone modification patterns and compared the expression levels of these two groups . Similarly , bivalent genes ( H3K4me3 , H3K27me3 ) in H2A . ZAP3 cells showed significant deviations in expression relative to H2A . ZWT , whereas active genes did not exhibit significant differences ( Figure S3A , bottom , Table S2 ) . We focused further attention on the class of bivalent genes and compared their expression in H2A . ZAP3 relative to H2A . ZWT during ESC differentiation using the Nanostring mCounter assay as an independent measure of gene expression ( Figure 3C , Table S4 ) . The subset of genes in the Nanostring probe set included developmental regulators , lineage specific genes as well as pluripotency factors and housekeeping genes , comprising a subset of known H2A . Z target genes and negative controls ( Table S4 ) . Consistent with our RNA-Seq data , genes involved in differentiation and pattern specification were expressed at higher levels in H2A . ZAP3 ESCs ( Figure 3C , D ) . Interestingly , bivalent genes are expressed at higher levels in H2A . ZAP3 relative to H2A . ZKD ESCs , suggesting that incorporation of the mutant results in a distinct chromatin state compared to loss of H2A . Z . Consistent with this idea , H2A . ZKD ESCs failed to activate lineage markers in D3 EBs [10] , whereas many of these genes were expressed at higher levels in H2A . ZAP3 EBs ( Figure 3C , D ) . The distinct transcriptional output of H2A . ZAP3 relative to H2A . ZKD may be due to the replacement of H2A . Z with H2A rather than its loss [36] , whereas incorporation of H2A . ZAP3 may result in a constitutively dynamic nucleosome . Given that H2A . Z is enriched at active genes and poised bivalent genes in ESCs ( Figure 2A ) , next we analyzed H3K4me3 and H3K27me3 patterns by ChIP-qPCR in H2A . ZWT and H2A . ZAP3 ESCs ( Figure 3E , F ) . While H3K4me3 enrichment patterns did not vary significantly between H2A . ZWT and H2A . ZAP3 at either class of H2A . Z target genes ( Figure 3E ) , we found that H3K27me3 enrichment was reduced at bivalent genes in the acidic patch mutant ESCs . This observation is consistent with the de-repression of bivalent genes in H2A . ZAP3 ESCs . Together , these results suggest that the divergent H2A . Z residues play key roles in the formation of specialized chromatin domains that are necessary for maintenance of the poised state and for responding to developmental cues . H2A . Z exists in hybrid nucleosomes along with the histone H3 variant H3 . 3 whose incorporation marks active promoters , enhancers , and insulator elements [23]–[25] . These double variant nucleosomes are characterized as highly unstable and salt labile , consistent with its enrichment in regions of highly dynamic chromatin . Given that H2A . ZAP3 appears to be loosely associated with chromatin relative to H2A . ZWT by salt titration , next we asked whether H3 . 3 was enriched at bivalent promoters in H2A . ZAP3 ESCs compared to H2A . ZWT . Since H3 . 3 differs from H3 by only three amino acids , antibodies against H3 . 3 cannot be used to effectively distinguish the variant from canonical H3 . To circumvent this issue , we transfected an H3 . 3 C-terminal HA-Flag construct into H2A . ZWT and H2A . ZAP3 transgenic ESCs depleted of endogenous H2A . Z ( Figure S3B ) . Similar lines were created with H3 . 1 C-terminal HA-Flag constructs ( Figure S3B ) . ChIP-qPCR analyses using Flag and HA antibodies revealed that the ratio of H3 . 3 enrichment relative to H3 . 1 was higher at target promoters in H2A . ZAP3 ESCs relative to H2A . ZWT ( Figure 3G ) . Given that H3 . 3 is associated with hyperdynamic chromatin , including active genes in ESCs [24]–[26] , [47] , the increase in H3 . 3 enrichment at TSSs is consistent with the dynamic nature of H2A . ZAP3 nucleosomes and de-repression of target genes in these cells . Taken together , our results demonstrate that the regulation of bivalent genes is highly sensitive to H2A . Z incorporation compared to active genes . In particular , the finding that a subset of genes remained highly expressed in H2A . ZAP3 cells during lineage commitment suggests that the acidic patch may be necessary to maintain the poised , silent chromatin state at bivalent genes . Nucleosome dynamics have important functional consequences on gene regulation [47]–[49] . Our data suggest that disruption of the divergent residues in the acidic patch results in a more dynamic association of H2A . ZAP3 in chromatin . Single cell analysis by fluorescence recovery after photobleaching ( FRAP ) has been used previously to probe the mobility and dynamics of chromatin-associated proteins in mammalian nuclei [48] , [50]–[52] . FRAP studies showed that the mobility of core histones H2B-GFP and H3-YFP are significantly higher in ESCs compared to differentiated cell types [47] . These analyses led to the idea that ESC chromatin is in a hyperdynamic and transcriptionally-permissive state , whereas heterochromatin formation during differentiation leads to a decrease in core histone dynamics [47] . Furthermore , in vitro biophysical studies showed that H2A . Z-containing nucleosomal arrays impeded higher order chromatin folding compared to canonical nucleosomes , suggesting that H2A . Z incorporation may contribute to a unique chromatin environment [53] . Thus , we analyzed the recovery kinetics of H2A . ZWT-YFP relative to H2A-YFP in ESCs . As expected , H2A-YFP displayed recovery kinetics in ESCs similar to previous reports for H2B-GFP ( ∼20% mobile fraction ) ( Figure 4A , B , Table S5 ) [47] . Interestingly , the dynamics of H2A . ZWT was reduced compared to H2A in ESCs . For example , the mobile fraction of H2A was ∼20% compared to ∼13% for H2A . ZWT ( P<0 . 02 ) ( Figure 4B ) . Importantly , unbleached photo-imaging controls showed that imaging conditions did not incur inadvertent bleaching and loss of histone fluorescence signal during the experiment ( Figure S4A , B ) . The higher mobility of H2A in ESCs is consistent with a global , transcriptionally permissive chromatin environment in ESC , while the slower recovery of H2A . ZWT is suggestive of a more specialized and distinct H2A . Z chromatin state . Our data are also consistent with prior work indicating that H2A . Z incorporation leads to a more stable nucleosome [21] , [54] and that H2A . Z nucleosomes promote formation of 30 nm fibers [53] . Next we probed H2A . ZWT dynamics upon lineage commitment . ESCs were differentiated by addition of retinoic acid ( RA ) for five days and fluorescence recovery was measured on the final day . Consistent with an increase in chromatin condensation during lineage commitment due to heterochromatin formation , we found that H2A was less dynamic upon RA differentiation , as indicated by the reduction in the mobile fraction ( ∼20% in ESCs versus ∼12 . 5% in RA differentiated cells , P<0 . 02 ) ( Figure 4C , F ) . In contrast , the recovery rate of H2A . ZWT was similar in ESCs and differentiated cells ( Figure 4D , F , Table S5 ) . The inability of H2A . Z to form more highly condensed chromatin structures typical of heterochromatin regions may explain why H2A . ZWT dynamics does not display a further reduction upon ESC differentiation . These data also suggest that H2A . Z dynamics is regulated via a different pathway compared to H2A . Given that H2A . ZAP3 enrichment was lower in ESC chromatin by ChIP-Seq , we investigated its mobility in ESCs . We found that H2A . ZAP3 recovers significantly faster than H2A . ZWT and displays recovery kinetics similar to H2A ( Figure 4A , B ) . For example , the H2A . ZAP3 mobile fraction was higher in ESCs compared to H2A . ZWT ( 20% versus 13%; P<0 . 006 ) and similar to H2A , suggesting that the divergent acidic patch residues play an important role in regulating H2A . Z dynamics . These observations are consistent with in vitro solution studies demonstrating that H2A . Z-containing nucleosome arrays harboring mutations in the acidic patch exhibited chromatin folding kinetics similar to H2A [55] . In contrast , the H2A . ZAP3 mobile fraction remained significantly higher upon differentiation compared with H2A , suggesting that H2A . ZAP3 nucleosomes are structurally distinct ( P<0 . 044 ) ( Figure 4E , F , Table S5 ) . Taken together , these observations indicate that the reduced occupancy of H2A . ZAP3 in ESCs is likely due to its increased dynamics and suggest that the divergent H2A . Z acidic patch is important for formation of specialized chromatin states . The difference in H2A . ZAP3 mobility upon RA differentiation relative to H2A may result from failure of these cells to properly differentiate . To determine if the mobile fraction was reflective of impaired differentiation capacity or purely a result attributable to H2A . ZAP3 dynamics , we introduced an H2A-mCherry transgene into H2A . ZWT and H2A . ZAP3 ESCs . To determine the mobile fraction of H2A-mCherry in both H2A . ZWT and H2A . ZAP3 , we performed FRAP on Day 0 and Day 5 RA-differentiated ESCs . We found that the H2A dynamics was similar in H2A . ZWT and H2A . ZAP3 ESCs ( Figure 4G , Figure S4C , D , Table S5 ) ( 12 . 7% and 11 . 4% respectively ) , suggesting that expression of the mutant variant does not significantly alter global H2A dynamics in the undifferentiated state . The relative differences in the percent mobile fraction for H2A-mCherry and H2A-YFP ( 22% versus 12 . 7% ) is likely due to differences in their respective fluorophore properties [56] . Upon RA differentiation , we found H2A dynamics decreased in H2A . ZWT ( 4 . 3% in RA cells ) , consistent with formation of condensed heterochromatin during lineage commitment ( Figure 4G , Figure S4C , Table S5 ) . Similarly H2A dynamics was also reduced in H2A . ZAP3 retinoic acid induced cells ( 6 . 7% in RA cells ) , indicating that these cells form chromatin structures more similar to differentiated cell types ( Figure 4G , Figure S4D , Table S5 ) . Collectively , these data suggest that the acidic patch region is necessary for proper regulation of H2A . Z dynamics and that its incorporation directly controls gene expression during lineage commitment . Active transcription is accompanied by rapid exchange of histone H2A/H2B dimers to accommodate the transiting RNA polymerase [48] , [49] , [51] , [52] . Given that H2A . ZAP3 incorporation leads to altered chromatin dynamics and changes in gene expression , we hypothesized that the increase in H2A . ZAP3 mobility was linked to transcription . To test this idea , H2A . ZWT , H2A . ZAP3 , and H2A ESCs were treated with flavopiridol , a reversible inhibitor of CDK9 that rapidly decreases RNAP2-dependent transcription . Consistent with previous observations , treatment of H2A . ZWT ESCs with 1 µM flavopiridol for 2 hrs resulted in a 30–44% decrease in RNAP2-dependent transcripts [57] , [58] ( Figure 5A ) . Notably , transcript levels were restored to control levels 2 hrs after the wash step . Prior FRAP studies in differentiated cells ( HeLa ) demonstrated a 3% reduction in the H2B mobile fraction upon treatment with the transcription inhibitor 5 , 6-Dichlorobenzimidazole 1-β-D-ribofuranoside ( DRB ) [52] . While this overall fraction is low , it is expected that only a small subset of total H2B is associated with transcription given its broad distribution across the genome . Similarly , we found that H2A treated with flavopiridol exhibited an approximately 6% decrease in the mobile fraction in ESCs by FRAP relative to the DMSO control ( P<0 . 02 ) ( Figure 5B , Figure S5A ) . Importantly , H2A dynamics were restored to normal levels 2 hrs after removal of flavopiridol , suggesting that at least 6% of H2A in ESCs is linked to transcription ( Figure 5B , Table S5 ) . In contrast , we did not detect a measurable change in H2A . ZWT dynamics upon flavopiridol treatment as compared with untreated ESCs , suggesting H2A . ZWT dynamics is not specifically dependent on active transcription ( Figure 5C , Figure S5B , Table S5 ) . This result is consistent with the idea that H2A . Z incorporation results in an inherent steady-state dynamic at promoters [21] , [54] . It is also possible that ESCs treated with flavopiridol show a moderate reduction in the H2A . ZWT mobile fraction that cannot be effectively resolved by FRAP [59] . In contrast , H2A . ZAP3 dynamics was decreased in ESCs upon flavopiridol treatment relative to H2A . ZWT ( Figure 5D and Figure S5C ) . For example , we detected a ∼6% decrease in the H2A . ZAP3 mobile fraction upon flavopiridol treatment ( P<0 . 01 ) similar to H2A , which was restored to near normal levels 2 hrs after flavopiridol removal . These observations were independently confirmed using DRB , an irreversible inhibitor of CDK9 and RNAP2-dependent transcription ( Figure S5D–G , Table S5 ) . Thus , our analysis suggests that the altered dynamics observed in H2A . ZAP3 ESCs are , in part , linked to transcription . These data are also consistent with disruption of a repressive chromatin state and higher expression of bivalent genes in H2A . ZAP3 cells . To determine if the transcription-dependent decrease in H2A . ZAP3 dynamics is coincident with increase in H2A . ZAP3 occupancy at target genes , we performed ChIP on H2A . ZAP3 ESCs treated with DMSO , flavopiridol ( Flavo ) and 2 hours post flavopiridol removal ( Wash ) . Consistent with our hypothesis , we found a significant increase in H2A . ZAP3 enrichment at target gene promoters with flavopiridol and partial reversal of this trend upon removal of flavopiridol ( Figure 5E ) , suggesting that the decrease in H2A . ZAP3 dynamics upon flavopiridol treatment is coincident with higher H2A . Z chromatin occupancy . These data also support our findings that H2A . ZAP3 is properly incorporated at promoters . Similar ChIP analyses in H2A . ZWT ESCs revealed a more modest increase in H2A . ZWT occupancy at some but not all target genes upon flavopiridol treatment relative to H2A . ZAP3 ( Figure 5F ) . This is also consistent with the FRAP data which indicates minimal change in H2A . ZWT mobile fraction upon flavopiridol treatment . These data are also consistent with in vitro biophysical studies showing that an intact H2A . Z acidic patch is necessary for the ability of a nucleosomal template to fold into a 30-nm chromatin fiber and for the efficient repression of transcription [45] , [46] . Taken together , our work demonstrates that the divergent acidic patch is an important structural feature that mediates H2A . Z dynamics and maintenance of chromatin states necessary for regulation of inducible gene expression programs during lineage commitment .
We analyzed the role of the three divergent H2A . Z acidic patch residues during ESC differentiation . While disruption of the acidic patch domain did not affect the global distribution of H2A . Z , its levels at target gene promoters were reduced , suggesting that the divergent acidic patch domain is an important determinant of H2A . Z incorporation or dynamics . We found that H2A . ZAP3 is less stably associated with chromatin relative to H2A . ZWT by salt extraction studies , suggesting that an intact acidic patch is necessary for stabilizing H2A . Z nucleosome structure . The dynamic association of H2A . ZAP3 with chromatin is also consistent with its lower enrichment at genomic sites . Prior in vitro studies show that H2A . Z forms stable 30 nm fibers at the expense of self-association [53] , indicating that the unique acidic patch may be important for mediating distinct higher order chromatin structures . H2A . Z is found at most H3K4me3 promoters of both active and poised genes and often flanks the nucleosome free region ( NFR ) around TSSs . We observed a particular reduction in H2A . ZAP3 levels in regions corresponding to the nucleosome downstream of the TSS , which likely represents the +1 nucleosome . The +1 nucleosome possesses significant regulatory potential and it is an important mediator of how a gene responds to environmental and developmental cues [39] . For example , a physical interaction between +1 nucleosome and stalled RNAP2 was demonstrated in Drosophila S2 cells and studies in C . elegans and Drosophila suggested that H2A . Z incorporation at the +1 nucleosome is important for maintaining the paused state of promoters [61]–[63] . Consistent with these findings , the accessibility of the TSS is determined in part by the displacement of the +1 nucleosome , which is important for recruitment of RNAP2 . Additionally , the +1 nucleosome appears to be important for regulating transcriptional elongation through interactions with Mediator components [40] . Thus , H2A . Z incorporation at the +1 nucleosome may be critical for determining the specific transcriptional response to developmental cues . A similar phenomenon regarding the role of the H2A . Z +1 nucleosome has been observed in plants in response to changes in ambient temperature [15] . Notably , specific transcriptional responses cannot be achieved by replacement with H2A reinforcing the idea that H2A . Z incorporation leads to specialized chromatin states . Similarly , while H2A can substitute for H2A . Z upon loss of the variant in ESCs , target genes fail to activate properly during ESC differentiation [10] , [36] . In contrast , incorporation of H2A . ZAP3 at TSSs , which is associated with a decrease in the +1 nucleosome , would allow for recruitment of RNAP2 and/or transcriptional elongation consistent with the persistent higher expression of bivalent genes during ESC differentiation ( see below ) . Together , these data point to a critical role for regulating H2A . Z dynamics , particularly at the +1 nucleosome . The decrease in H2A . ZAP3 enrichment led to changes in gene expression particularly at bivalent genes during ESC differentiation . H2A . Z is largely associated with active genes whose promoters are marked by H3K4me3 and it is enriched at a smaller subset of bivalent genes ( comprising H3K4me3 and H3K27me3 modifications ) in ESCs that are silent yet competent for activation . Why is the regulation of bivalent genes more sensitive to mutations in the acidic patch relative to active genes ? We propose that the acidic patch is necessary for mediating H2A . Z dynamics or nucleosome stability upon incorporation at both poised and active target genes . Mutations in the acidic patch increase H2A . Z mobility and possibly alter the stability of the +1 nucleosome , which leads to expression of bivalent genes . Conversely , the increase in H2A . Z dynamics may have little impact on the expression of actively transcribed genes since elongation is occurring at these genes and perhaps other modes of regulation prevail to maintain genes in an active state [48] . This idea is also consistent with the observation that H2A . Z levels decrease upon gene activation [8] . In addition to H2A . Z , Polycomb Repressive Complexes ( e . g . PRC1 and PRC2 ) are enriched at bivalent genes [10] , [11] . PRCs are important for establishment and maintenance of repressed chromatin states and play key roles in dynamic regulation of gene expression during lineage commitment [64] . We suggest that disruption of the H2A . Z acidic patch alters the chromatin state and loss of PcG-mediated repression at bivalent genes . Consistent with this idea , we found that H3K27me3 levels were reduced at bivalent genes in H2A . ZAP3 ESCs , whereas H3K4me3 levels remained largely unchanged . These data further support our prior findings that functional coordination between PRCs and H2A . Z is important for regulating developmental gene expression patterns during ESC differentiation [10] , [65] . Thus , the unique H2A . Z acidic patch may generally restrict the mobility of nucleosomes in ESCs . Changes in local chromatin compaction due to incorporation of the mutant may lead to changes in gene expression and subsequent loss of PRCs leading to a shift in the balance of chromatin associated factors that promote activation . In support of this model , we found that bivalent promoters gained H3 . 3 in H2A . ZAP3 ESCs , a histone variant typically incorporated at highly dynamic chromatin regions . Together , our data suggest that the divergent H2A . Z acidic patch is necessary for the formation of specialized chromatin states that allow for maintenance of the poised state and for proper induction of gene expression programs in response to developmental cues ( Figure 6 ) . Given that H2A . Z displayed reduced chromatin association in ESCs compared to H2A . ZWT , we used single cell analysis to further analyze its dynamic association with chromatin . Interestingly , FRAP measurements revealed that H2A . Z was less dynamic than H2A as indicated by its reduced mobile fraction . These data are consistent with the idea that H2A . Z incorporation leads to the formation of a more stable nucleosome and the H2A . Z containing nucleosome arrays promote formation of 30 nm fibers [53] . While FRAP measurements represent overall steady-state chromatin dynamics , little is known regarding the on- and off-rates of H2A . Z . We suggest that this equilibrium is altered upon disruption of the divergent acidic patch residues resulting in a higher off rate and increased transcriptional output at genes that required H2A . Z for maintenance of the poised state ( Figure 6 ) . Moreover , our data suggest that the divergent acidic patch is a critical determinant of H2A . Z dynamics and that this region may be important for rapid activation of developmental programs . Remarkably , we show that while H2A . ZAP3 appears more dynamic , treatment with the transcriptional inhibitor flavopiridol led to a decrease in its mobile fraction as well as increased occupancy at target promoters . Thus , H2A . ZAP3 dynamics may also be regulated by the local chromatin environment as well as by interactions with other factors . Consistent with this idea , studies in Arabidopsis showed that H2A . Z is an important component of the thermosensory response . For example , H2A . Z was less sensitive to nuclease digestion at silent , inducible genes at normal temperature [15] . Upon activation of these genes by increasing temperature , H2A . Z-containing nucleosomes become more accessible to nuclease digestion suggesting that the variant is dynamically remodeled in response to inductive cues . Thus , it remains possible that H2A . Z modifications or interaction with other factors may mediate the function of the divergent acidic patch . H2A . Z localization and dynamics is regulated from yeast to human in large part by two ATP-dependent remodelers , SRCAP and INO80 , [66]–[69] . Thus , future studies will be important to determine how the divergent acidic patch cooperates with chromatin remodelers . For example , INO80 appears to be an important regulator of H2A . Z removal by catalyzing the exchange of H2A . Z for H2A [69] . Thus , one possibility is that H2A . ZAP3 is a better substrate for INO80 leading to a higher off rate and increased dynamics . Interestingly , Swc2 in yeast , a conserved subunit of SRCAP , binds directly to H2A . Z [70] and recent work indicates that this interaction may function as a “lock” that prevents rapid exchange of H2A . Z in chromatin [71] . Notably , domain swap experiments in yeast , where the H2A . Z C-terminal domain including the extended acidic patch is replaced with the analogous region of H2A blocks interaction with Swc2 and many other components of the SWR1 remodeling complex [70] . In contrast , deletion of the H2A . Z C-terminal region distal to the extended acidic patch maintained interactions with SWR1 [30] . Consistent with these observations , we found that H2A . ZAP3 interacted with SRCAP; however , it is possible that mutations in the extended acidic patch interfere with a specific interaction with YL1 , the mammalian Swc2 homologue . Thus , an alternative scenario is that H2A . ZAP3 incorporation leads to SRCAP-mediated exchange of H2A . Z for H2A , which would likely influence dynamics . The acidic patch has been proposed to mediate interactions with histone tails from neighboring nucleosomes , and this domain may also form a novel interaction surface for recruitment of other chromatin factors and downstream effectors [27] , [53] , [55] , [60] . Thus , the unique H2A . Z acidic patch may interact with distinct subsets of factors that determine its specialized functions compared to H2A . In support of this idea , a recent proteomics analysis identified a number of chromatin-associated proteins enriched with H2A . Z [72] . These interactions may be important for regulating transcriptional output at target genes , ultimately allowing the cell to tune signals into specific responses that lead to changes in cell fate . In addition to its divergent structural features , H2A . Z is also subject to numerous posttranslational modifications ( PTMs ) . For example , acetylation of the H2A . Z amino-terminus is a hallmark of active genes and has been implicated in gene regulation and chromosome stability [73]–[76] , whereas the carboxy-terminal domain is modified by ubiquitylation and sumolyation , and these marks have been implicated in heterochromatin formation and DNA repair , respectively [77] , [78] . As such , PTMs may regulate H2A . Z dynamics by promoting or inhibiting interactions with regulatory factors or by affecting its interactions with neighboring nucleosomes . Overall , our work provides critical insights into the role of the divergent H2A . Z acidic patch residues as a structural determinant that links chromatin dynamics , gene regulation , and ultimately cell fate . Moreover , these data provide a mechanistic explanation for the essential role of the divergent acidic patch during metazoan development .
V6 . 5 ( 129SvJae and C57BL/6 ) ESCs were plated with irradiated murine embryonic fibroblasts ( MEFs ) and grown under typical ESC conditions on gelatinized tissue culture plates . Briefly , cells were grown in Knockout DMEM ( Invitrogen ) supplemented with 10% fetal bovine serum ( Hyclone ) , leukemia inhibitory factor [16] , non-essential amino acids ( Invitrogen ) , L-glutamine ( Invitrogen ) , and penicillin/streptomycin ( Invitrogen ) as previously described [79] . ESCs used for ChIP-Seq , RNA-Seq and qRT-PCR experiments were plated without MEFs for the final passage . H2A . ZWT-YFP , Mutant H2A . ZAP3-YFP and H2A-YFP constructs were generated from the pAd-cDNA ( Addgene ) , which contains a dox-inducible CMV-promoter . The H2A . ZAP3 mutant was generated by replacing H2A . Z residues G92 , D97 , and S98 with the corresponding H2A amino acids ( N , N and K respectively ) using the instructions provided in the Quikchange Site Directed Mutagenesis Kit ( Stratagene ) . We then replaced GFP in the vector with YFP ( YFP exhibits lower background for imaging ) to generate an in frame fusion C-terminal to H2A . Z . The resulting lentiviral constructs were transfected into 293 cells using the protocol outlined by the RNAi consortium ( BROAD Institute , http://www . broadinstitute . org/rnai/public/ ) . The viral supernatant generated 48 hrs after transfection was used to infect KH2 ESCs [80] to generate wild-type and mutant H2A . Z transgenic ESC lines . The YFP transgenic ESCs were induced with 1 µg/ml of doxycycline and FACS sorted for YFP positive cells . Lentiviral constructs expressing short hairpins specifically directed at the 3′ UTR of endogenous H2A . Z were introduced into the wild-type and mutant H2A . Z transgenic KH2 ESC lines . Sequences of the different H2A . Z 3′UTR-directed hairpin oligos are as follows: sh#1 5′- AACAGCTGTCCAGTGTTGGTG-3′; sh#2 5′- AATTAGCCTTCCAACCAACCA-3′ . Hairpin oligos were annealed and cloned into pLKO . 1 vector ( Sigma ) as detailed by the RNAi consortium , BROAD ( http://www . broadinstitute . org/rnai/trc/lib ) . Since KH2 cells are puromycin resistant [80] , blasticidin was used as a selection marker for the generation of endogenous H2A . Z-depleted transgenic KH2 ESCs . The puromycin marker in the pLKO . 1 vector was removed by digestion with BamHI and KpnI and replaced with blasticidin . The blasticidin cDNA was PCR amplified from pLenti6 . 2/V5-DEST Gateway Vector ( Invitrogen ) . V6 . 5 ( 129SvJae and C57BL/6 ) and the YFP transgenic ESCs were cultured as previously described [79] . The endogenous H2A . Z-depleted transgenic KH2 ESCs were cultured in the presence of blasticidin ( 5 µg/ml ) on blasticidin-resistant feeder cells [81] . Retinoic acid-induced differentiation was performed by plating mESCs onto gelatinized tissue culture plates without MEFs and grown in the same mESC media as above but without LIF and supplemented with 1 µM all-trans retinoic acid ( Sigma , R2625 ) . Cells were collected after five days . mESCs were induced to form embryoid bodies [68] by plating on non-adherent plates from a starting density of 1–2 million ESCs per 10 ml of mESC media lacking LIF . RNA was extracted using TRIzol ( Invitrogen , 15596-018 ) or Izol ( 5PRIME , 2302700 ) . Purified RNA was reverse transcribed using SuperScript III ( Invitrogen , 18080-044 ) or M-MLV reverse transcriptase ( Invitrogen , 28025-013 ) and random hexamers according to manufacturer protocols . Quantitative PCR reactions were performed with SYBR Green Master Mix ( Roche ) . Primer sequences are listed in Table S4 . Relative mRNA levels were determined in triplicate for each transcript using the manufacturer's software ( Advanced Relative Quantification with Roche Lightcycler 480 Software Version 1 . 5 ) using control genes ( Hprt , Gapdh , or Tubb5 levels ) for normalization . RNA was isolated using the protocol mentioned above . The purified RNA was the subjected to oligo ( dT ) selection , fragmentation and first and double strand synthesis with the Illumina Tru-Seq kit ( RS-930-20 01 ) according to the manufacturer's instructions . DNA fragments above 30 bp was purified using SPRI-TE beads ( Beckmann Coulter , Agencourt , A63880 ) according to manufacturer's instructions . The purified DNA was end-repaired and single A bases for adaptor ligations . The adaptor-ligated DNA was then subjected to double SPRI-TE purification to select for ∼200 bp fragments . These fragments were enriched and barcoded by PCR for multiplexing . A final SPRI-TE purification was performed to clean up the barcoded RNA-Seq libraries . RNA-Seq data were aligned against the mouse reference genome ( mm9 ) using Tophat 1 . 4 . 1 and Bowtie 0 . 12 . 7 in single-end mode , tolerating up to 2 mismatches and matches up to 20 locations transcriptome-wide . Gene expression was quantified with Cufflinks 1 . 3 . 0 , and pairwise differential expression was analyzed using Cuffdiff 1 . 3 . 0 . H3K4me3 and H3K27me3 enrichment peaks were obtained from our previous work [38] . Histone mark enrichment peaks whose boundaries fell within a region +/−2 kilobases ( kbs ) of a transcription start sites ( TSS , based on RefSeq annotation of the mm9 mouse genome assembly ) were identified and the number of H3K4me3 and H3K27me3 TSSs were recorded . RNA-Seq-based isoform expression data from Cufflinks ( in RPKMs ) were summed over each TSS and used as a metric for the transcriptional output originating from each TSS ( Table S2 ) . Median summed isoform expression levels were computed across all isoforms with H3K4me3- or H3K4me3/H3K27me3-marked TSSs in the different ESCs ( H2A . ZWT , H2A . ZKD , and H2A . ZAP3 ) and EB differentiation time points and compared by t-test . Similarly , the overlap between these subsets and the genes differentially regulated upon H2A . Z modulation was tested using the hypergeometric test . Finally , cumulative distribution of isoform expression differences upon H2A . Z modulation were computed by stratifying log2-tranformed isoform fold changes based on the presence of H3K4me3 and H3K27me3-bound regions within +/−2 kbs of the TSS . The significance of the differences between the cumulative distributions plots was assessed using the Kolmogorov-Smirnov test . All the sequencing results in the article have been deposited in GEO under the accession number GSE40065 . ChIP was performed as described [38] with the following modifications . Diagenode Bioruptor ( UCD-200 ) was used to sonicate with 30 cycles of 30 sec on , 30 sec off . The samples were sonicated in 15 ml polystyrene tubes at 4°C while samples were immersed in ice cold water . Antibodies used for ChIP include: GFP ( Abcam , ab290 ) , H2A . Z ( Abcam , ab4174 ) , H3K27me3 ( Cell Signaling Technology , #9733 ) , H3K4me3 ( Millipore , #07-473 ) , Suz12 ( Bethyl , A302-407A ) and RNAP2 Ser5P ( Abcam , ab5131 ) . ChIP enriched DNA was quantified by quantitative PCR and the data analyses performed as described in [10] . qPCR reactions using SYBR Green ( Roche ) and gene-specific primers ( Table S4 ) were performed on ChIP and whole cell extract ( WCE ) DNA . Reactions were performed in triplicate on the Roche LightCycler 480 . % Input was calculated with the following formula: % Input = 2 ( Cp ( WCE ) -Cp ( IP ) ) × ( % WCE ) . Approximately 200 ng of DNA was submitted to SPRI-works Fragment Library System I ( Beckman Coulter ) for each library prepared and sequenced on Illumina GAII . ChIP-seq reads were aligned to the mm9 genome assembly using Bowtie 0 . 12 . 3 , allowing for two mismatches . To determine regions of the genome enriched for H2A . Z , mapped reads were extended to 200 bp ( average fragment length ) and allocated in 25-bp bins . A Poissonian model was used to determine statistically enriched bins with a P-value threshold set at 1E-9 as described previously [82] . Additionally , we required that genomic bins were at least 5 fold over input to be considered enriched peaks . All the sequencing results in the article have been deposited in GEO under the accession number GSE40065 . ESCs were plated in the absence of MEFS on 0 . 2% gelatin in 2 mm Lab-Tek Chambered #1 borosilicate cover glass chambers containing a 1 mm glass slide 24 hours prior to imaging . Phenol-red free DMEM ( Invitrogen 21063-045 ) was used to make ESC media for imaging purposes to minimize medium auto fluorescence . For differentiation experiments , ESCs were treated with 1 µM retinoic acid and plated on 0 . 2% gelatin in Lab-Tek Chambered cover glass 24 hours prior to imaging . YFP transgenic cells were additionally induced with 1 µg/ml of doxycycline 24 hours prior to imaging . Confocal fluorescence imaging was performed on a LSM 510 microscope ( Carl Zeiss , Jena , Germany ) with a 514 nm wavelength laser for YFP excitation , a 520–550 nm bandpass emission filter and a 100× 1 . 4 N . A . oil immersion objective . FRAP experiments were performed by photobleaching for a short 5 . 4 sec exposure to 100% laser intensity . To minimize error in the quantification of fluorescence recovery induced by the movement of unbleached chromatin into the bleached region due to the dynamic nature of ESC chromatin and changes in nuclear morphology , a 20 pixel sub-region within the 75 pixel bleached region was used was used to quantify fluorescence recovery using custom-written MATLAB routines ( Mathworks , Natick , MA ) . In all cases the bleached region was sufficiently smaller than the size of the nucleus to minimize the effect of the original loss of fluorescent protein due to bleaching on maximal recovery . In any case , such loss would result in a uniform underestimation of mobile fractions in all tested instances , and not affect relative differences . Images were acquired using 30 sec acquisition intervals for approximately 14 mins with scan speed and imaging intervals optimized to minimize photobleaching during the recovery process . A minimum of 14 cells were analyzed for each FRAP experiment , with background correction performed by normalizing intensities to the maximum initial mean intensity in the bleach spot prior to photobleaching to generate individual FRAP curves . Mean data shown in results constitute the average of individual recovery curves . Mobile fractions ( MF ) were calculated from individual curves using MF = ( Idip−Isat ) / ( 1−Idip ) , where Idip is the value of the mean intensity immediately after the bleaching pulse , and Isat is the mean intensity at the end of the monitored recovery period . Student's t-test was used to calculate the statistical significance of the differences between estimated mobile fractions . For the flavopiridol experiments , cells were treated with 1 µM flavopiridol 2 hours before conducting photobleaching experiments . The post-wash FRAP experiments were performed on ESCs 2 hrs after washing away the inhibitor . For DRB experiments , ESCs were treated with 15 µM DRB for 2 hours and then imaged for FRAP as described above . All mobile fractions and the respective standard errors are listed in Table S5 . Histone extracts were prepared by harvesting the cells first with ice-cold PBS supplemented with 5 mM sodium butyrate ( 106–107 cells ) . The cells were then washed in cold filter sterilized Lysis buffer ( 0 . 25M Sucrose , 3 mM CaCl2 , 1 mM Tris pH 8 . 0 , 0 . 5% NP-40 ) . The nuclei were spun down at 3900 rpm for 5 min at 4°C . The supernatant was removed leaving the pellet of 50 µl of nuclei . This pellet was washed again with Wash solution ( 300 mM NaCl , 5 mM MgCl2 , 5 mM DTT , 0 . 5% NP40 ) and spun under the same conditions as mentioned above . The supernatant was removed and the nuclei were resuspended in 50 µl of Extraction solution ( 0 . 5M HCl , 10% glycerol , 0 . 1M 2-mercaptoethylamine-HCl ) . The nuclei were then left in ice for 30 min after which they were spun at 13 , 000 rpm for 5 min at 4°C . The supernatant was transferred to siliconized tubes and 500 µl of acetone was added . The resulting solution was incubated at −20°C overnight . The resulting histone precipitate was spun at 13 , 000 rpm for 5 min for further analyses . Rabbit anti-H2A . Z antibody ( Abcam , ab4174 ) and Rabbit anti-GFP antibody ( Abcam , ab290 ) was used for western blot at concentrations recommended by the manufacturer . Histone extracts , lysates and co-immunoprecipitated samples were resolved on SDS-PAGE gels and transferred on PVDF membrane ( Bio-rad , 162-0177 ) using the Mini-Trans-Blot ( Bio-rad , 170-3930 ) . The transferred blot is then blocked with 5% milk in TBST ( 0 . 1% Tween-20 in 1× Tris-buffered saline-pH 7 . 4 ) for 1 hour at room temperature and incubated with the indicated antibodies prepared in TBST with 5% milk . The blots were then washed with TBST three times and incubated with species-specific HRP-conjugated secondary antibodies ( Calbiochem ) . After incubation with secondary antibodies , the immunoblots was washed again with TBST and visualized using HRP substrate from Biorad ( 170-5070 ) . Cells were fixed with 4% paraformaldehyde for 20 minutes at room temperature , washed 3× with PBS , permeabilized in PBS with 0 . 2% TritonX , 0 . 1% Tween-20 for 30 minutes at room temperature , washed twice with PBS and blocked with PBS containing 0 . 1% Tween-20 , 2% CCS ( Cosmic Calf Serum , Invitrogen ) for 1 hour at room temperature . The cells were then stained with anti-Oct4 antibody ( mouse monoclonal , Santa Cruz , sc-5279 , 1∶100 ) for 1 hour at room temperature , then washed twice more with block and stained with anti-mouse secondary ( Alexafluor 594 ) . After washing twice more , the cells were imaged using Nikon Eclipse Ti-S . For histology , the EBs were fixed in 10% formalin for 20–30 mins and washed with PBS . Fixed EBs were then dehydrated by treating with a gradient of 70% , 80% , 90% and 100% ethanol for 20 mins each . The dehydrated EBs were finally cleared in xylene and embedded in paraffin overnight at 60°C . Sections of EBs at 0 . 4 microns were generated and placed on slides for staining . The sections are deparaffinized with xylene , rehydrated and stained with Harris Hematoxylin ( Surgipath , 01560 ) and Eosin ( Polyscientific , s176 ) . ESCs were cultured in 13C6/15N2-lysine ( +8 ) 13C6/15N4-arginine ( +10 ) ( “SILAC heavy” ) or naturally occurring lysine and arginine ( “SILAC light” ) medium according to [83] . Histones were purified from ES and differentiated cells essentially as described [84] except that a Zorbax C8 HPLC column was employed ( Agilent ) . Each 1 min fraction collected from the HPLC separation of the histones was subjected to SDS-PAGE . Subsequent LC-MS/MS experiments were performed on an LTQ Velos-Orbitrap mass spectrometer ( ThermoFisher Scientific ) fed by an Agilent 1200 nano-HPLC system ( Agilent ) following procedures analogous to those described elsewhere [85] . First , the Coomassie-stained visible bands on PAGE separations of the HPLC fractions were interrogated by tryptic and chymotryptic digestion . Peptides unique to H2A . Z ( and not derived from other H2A variants ) were detected in bands of ∼14 kDa and ∼20 kDa . The ∼14 kDa variant of H2A . Z co-HPLC-separated with Histone H4 ( ∼12 kDa ) while the ∼20 kDa variant co-HPLC-separated with Histone H2B ( ∼14 kDa ) . Next , these bands ( from a parallel preparation ) were subject to in-gel propionylation using propionic anhydride according to [86] . To study the C-terminal ubiquitination of H2A . Z , chymotryptic peptides were analyzed . To study the N-terminal acetylation of H2A . Z , tryptic peptides were analyzed . M/z values corresponding to the various acetylated and ubiquitin-residual peptides ( The proteases will cleave ubiquitin as well as H2A . Z , leaving a branched peptide residual ) were calculated . Separate acquisition methods were designed for the study of acetylation or ubiquitination . Selective-ion monitoring ( SIM ) windows were designed around these m/z as appropriate and data-independent MS/MS scans were acquired at these m/z ratios as dictated by each experiment . The sample was introduced to the mass spectrometer via liquid chromatography with conditions identical to those described [87] . Chromosome spread analysis was performed as described in [88] . The cells were treated with a final concentration of 1 µg/ml of Nocodazole for 8 hours at 37°C . The treated cells were then washed with PBS . The mitosis-arrested cells were then collected and spread on slides using a Cytospin 4 ( Thermo , Waltham , MA ) [89] . The spread cells were then fixed with paraformaldehyde and permeabilized with detergent . The fixed slides were then stained for CENP-A at a dilution of 1∶800 ( Cell Signaling Technology , #2048 ) and visualized with Alexa 594 secondary antibodies ( 1∶500 ) . The DNA was visualized using DAPI and the GFP signal was from the YFP-fused transgene expression . | Elucidating how regulation of chromatin structure modulates gene expression patterns is fundamental for understanding mammalian development . Replacement of core histones with histone variants has recently emerged as a key mechanism for regulating chromatin states . The histone H2A variant H2A . Z is of particular interest because it is essential for embryonic development and for proper execution of developmental gene expression programs during cellular specification . ESCs provide a good model for investigating the function of H2A . Z during lineage commitment because these cells can generate an unlimited number of equivalent descendants while maintaining the capacity to differentiate into any cell type in the organism . Divergent regions in H2A . Z are likely key for functional specialization , but we know little about how these differences contribute to chromatin regulation . Here , we show that the unique H2A . Z acidic patch domain is necessary for regulation of lineage commitment during ESC differentiation by linking transcription to chromatin dynamics . Our work provides a critical foundation for elucidating how H2A . Z incorporation is key to cell fate determination . These findings are particularly important given that H2A . Z has been implicated in many diseased conditions , including cancer . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"cell",
"biology",
"biochemistry",
"developmental",
"biology",
"biology"
] | 2013 | H2A.Z Acidic Patch Couples Chromatin Dynamics to Regulation of Gene Expression Programs during ESC Differentiation |
Cellular senescence involves epigenetic alteration , e . g . loss of H3K27me3 in Ink4a-Arf locus . Using mouse embryonic fibroblast ( MEF ) , we here analyzed transcription and epigenetic alteration during Ras-induced senescence on genome-wide scale by chromatin immunoprecipitation ( ChIP ) -sequencing and microarray . Bmp2 was the most activated secreted factor with H3K4me3 gain and H3K27me3 loss , whereas H3K4me3 loss and de novo formation of H3K27me3 occurred inversely in repression of nine genes , including two BMP-SMAD inhibitors Smad6 and Noggin . DNA methylation alteration unlikely occurred . Ras-activated cells senesced with nuclear accumulation of phosphorylated SMAD1/5/8 . Senescence was bypassed in Ras-activated cells when Bmp2/Smad1 signal was blocked by Bmp2 knockdown , Smad6 induction , or Noggin induction . Senescence was induced when recombinant BMP2 protein was added to Bmp2-knocked-down Ras-activated cells . Downstream Bmp2-Smad1 target genes were then analyzed genome-wide by ChIP-sequencing using anti-Smad1 antibody in MEF that was exposed to BMP2 . Smad1 target sites were enriched nearby transcription start sites of genes , which significantly correlated to upregulation by BMP2 stimulation . While Smad6 was one of Smad1 target genes to be upregulated by BMP2 exposure , Smad6 repression in Ras-activated cells with increased enrichment of Ezh2 and gain of H3K27me3 suggested epigenetic disruption of negative feedback by Polycomb . Among Smad1 target genes that were upregulated in Ras-activated cells without increased repressive mark , Parvb was found to contribute to growth inhibition as Parvb knockdown lead to escape from senescence . It was revealed through genome-wide analyses in this study that Bmp2-Smad1 signal and its regulation by harmonized epigenomic alteration play an important role in Ras-induced senescence .
Cellular senescence was first described as the limited replicative capacity of primary cells in culture [1] . Activated oncogenes can induce premature form of cellular senescence , and cells fall into irreversible arrest to block cellular proliferation [2] , [3] . In addition to cell death programs such as apoptosis and autophagy , oncogene-induced senescence is recognized as a potent barrier against oncogenic transformation , suppressing unscheduled proliferation of early neoplastic cells [4]–[7] . Replicative senescence and oncogene-induced senescence are known to comprise activation of tumor suppressor pathways including p16Ink4a-Rb and p19Arf ( p14ARF in human ) -p53 signaling cascades . Genetic and epigenetic inactivation of these genes in cancer supported their crucial roles in senescence as barriers to tumorigenesis [8] , [9] . Although the roles of RB and p53 signaling pathways in senescence are undisputed , it has become clear that other factors are also involved . Expression of secreted factors , or “senescence-messaging secretome” , has been proposed as an example of such mechanisms [10] , [11] . The induction of senescence required several secreted factors including members of Wnt , insulin , transforming growth factor-β , plasmin and interleukin signaling cascades [11] . Epigenetic mechanism is also suggested to play important roles in senescence . When human fibroblasts senesced , heterochromatic regions condensed to form senescence-associated heterochromatic foci , where regions with histone H3K9 trimethylation ( H3K9me3 ) gathered [12] , and were recently shown to restrain DNA damage response [13] . Expression of Jhdm1b , a demethylase specific for H3K36me2 , caused cell immortalization or leukemic transformation depending on its demethylase activity on p15Ink4b , and its knock down resulted in cellular senescence [14] , [15] . INK4A and ARF region in young cells was repressed by H3K27me3 imposed by the Polycomb Group proteins , and the repressive mark was lost during oncogene-induced senescence , resulting in expression of p16 and p19; the loss of repressive mark was also detected when mouse embryonic fibroblast ( MEF ) underwent stress-induced senescence around seven passages [16]–[19] . Jmjd3 , a histone demethylase for H3K27 , was found to be essential in senescence , and its knock down lead to escape from senescence sustaining repression of p16 by H3K27me3 [20] , [21] . In the previous studies , we comprehensively analyzed aberrant promoter DNA methylation in colorectal cancer and reported three distinct DNA methylation epigenotypes [22] , [23] . Distinct methylation epigenotypes significantly correlated to different oncogene mutation statuses , suggesting that epigenotypes of cancer might perhaps be requisite phenotype of aberrant methylation to escape from oncogene-induced senescence by inactivation of critical factors of senescence [23] , [24] . To gain insight in phenotype of critical gene inactivation in oncogene-mutation ( + ) cancer , we aim to clarify critical genes/signals/phenomena in oncogene-induced senescence in normal cells in this study . Here we perform genome-wide analyses of epigenetic and gene expression changes in Ras-indeced senescence using mouse embryonic fibroblasts ( Figure S1 ) . We show that Bmp2/Smad1 signal is critical in Ras-induced senescence , and is regulated by coordinated epigenomic alteration . We further examine downstream target genes of this critical signal on genome-wide scale , and show that the epigenomic regulation of the signal involves disruption of negative feedback loop , and that activated downstream targets actually include a gene to contribute to growth arrest .
To induce cellular senescence , mouse embryonic fibroblasts after two passages ( MEFp2 ) was infected with retrovirus of oncogenic Ras ( RasV12 ) with N-terminal FLAG tag and cultured through day 10 ( Figure S2A ) . RasV12-infected cells ( RasV12 cells ) showed significant increase in number of SA-βgal ( + ) cells , compared to MEFp2 , MEF passed three more times without infection ( MEFp5 ) , mock-infected cells ( Mock cells ) , and wild type Ras ( RasG12 ) -infected cells ( Figure 1A and Figure S2B ) . Global gene expression analysis was performed using expression array . In RasV12 cells on day 10 , 822 genes were upreglated and 735 genes downregulated , by >5-fold compared to MEFp2 ( Tables S1 , S2 ) . Gene annotation enrichment analysis suggested that genes related to secreted protein ( P = 1 . 8×10−19 ) , extracellular region ( P = 1 . 2×10−21 ) , and differentiation/development ( P = 3 . 8×10−10 ) , e . g . Bmp2 and Igfbp3 , were upregulated , supporting the importance of secreted factor expression in senescence . Genes related to cell cycle ( P = 7 . 2×10−22 ) such as Cdc6 and Mcm5 were enriched in downregulated genes , indicating growth arrest . Also genes related to secreted protein ( P = 7 . 9×10−18 ) and extracellular region ( P = 9 . 2×10−14 ) such as Bmp4 and Tgfb2 were enriched in downregulated genes , suggesting that dynamic control of secretome by activation and repression of secreted factors occurred during senescence . To analyze epigenomic gene regulation during Ras-induced senescence , we selected H3K4me3 as an active mark and H3K27me3 as a repressive mark , and mapped them by Chromatin immunoprecipitation ( ChIP ) -sequencing . As reported , H3K27me3 mark at p16Ink4a-p19Arf locus in MEFp2 was markedly lost in RasV12 cells ( Figure 1A ) . ChIP-sequencing of H3K4me3 showed concurrent gain of the active mark around p16 transcription start site ( TSS ) , which reflected increase of p16 expression in RasV12 ( Figure 1B ) . By quantitative ChIP-PCR , significant gain of H3K4me3 and loss of H3K27me3 were validated in RasV12 cells , compared to MEFp2 ( Figure 1C ) . Gain of H3K4me3 and loss of H3K27me3 were also detected at intermediate level in Mock and RasG12 cells ( Figure 1C ) . Expression of p16 was also partially increased in Mock and RasG12 cells , at the similar level to MEFp5 ( Figure 1B ) . These indicated that p16 expression could be induced partially by gain of H3K4me3 and H3K27me3 during passages , which was in agreement with the previous report of gradual H3K27me3 loss in stress-induced senescence during 5–7 passages [16] , [18] , but more marked alteration occurred at this locus in Ras-induced senescence . Enrichment of Ezh2 , a member of the Polycomb Group proteins , was also analyzed by ChIP-PCR , and it was significantly decreased around p16 TSS in RasV12 cells compared to MEFp2 ( Figure 1D ) . When analyzing distribution of 36-bp reads mapped around TSS of 20 , 232 genes , the mapped reads were enriched within ±2 kb of TSS , mainly ±1 kb of TSS ( Figure S3A ) , for both H3K4me3 and H3K27me3 . We counted mapped reads within a window of genomic region , so that the number of mapped reads per million reads within a window is regarded as epigenetic status of the center position of the window . Within ±2 kb from TSS of each gene , the maximum number of mapped reads per million reads in a window size of 300 bp ( H3K4me3 ) or 500 bp ( H3K27me3 ) was regarded as the epigenetic status of each gene . A wider window was necessary for H3K27me3 because distribution of H3K27me3 was rather wide than H3K4me3 ( Figure 1A and Figure S3A ) . The number of genes with repressive H3K27me3 mark was generally decreased in RasV12 cells ( Figure S3B ) , in agreement of the previous reports [20] , [21] that expression of Jmjd3 was increased during senescence , whereas expression of Ezh2 was decreased ( Figure S4 ) . It was expected that genes activated by losing H3K27me3 might exist other than p16 and p19 , because of the decrease of genes with H3K27me3 mark in RasV12 cells . Among 20 , 232 genes with epigenomic alteration analyzed , 16 , 793 genes were also analyzed for expression on array ( Figure S5 ) . For epigenetic status of H3K4me3 , 9 , 164 genes in MEFp2 and 8 , 841 genes in RasV12 showed >4 reads per million reads around TSS , and regarded as H3K4me3 ( + ) . Similarly , 7 , 140 and 7 , 354 genes respectively with <3 reads per million reads were regarded as H3K4me3 ( - ) . Markedly higher expression levels of H3K4me3 ( + ) genes than H3K4me3 ( - ) genes were confirmed by comparing the mean of expression levels ( Figure S5A ) . For H3K27me3 , 2 , 612 and 2 , 370 genes with >1 . 5 reads per million reads around TSS were regarded as H3K27me3 ( + ) , and 13 , 205 and 12 , 841 genes with <1 were as H3K27me3 ( - ) in this study . H3K27me3 ( + ) genes were markedly repressed than H3K27me3 ( - ) genes ( Figure S5B ) . Among 284 genes losing H3K27me3 in RasV12 cells , 30 genes losing H3K27me3 and gaining H3K4me3 simultaneously , like p16 , showed significant enrichment in upregulated genes among the 284 genes ( P = 0 . 000007 , Kolmogorov-Smirnov test , Figure 2A ) . Among the 30 genes ( listed in Table S3 ) , Bmp2 , a secreted factor for BMP/SMAD pathway , was found to be the most upregulated secreted factor and activated more than p16 ( Figure 2A ) . Interestingly , 110 genes modified bivalently in MEFp2 showed loss of H3K27me3 and sustained H3K4me3 mark in RasV12 cells , but did not show significant enrichment in upregulated genes ( P = 0 . 9 , Figure 2A ) . Not only genes with H3K27me3 loss , but also there were as many as 239 genes showing H3K27me3 gain in RasV12 cells . Nine genes gaining H3K27me3 and losing H3K4me3 simultaneously showed significant enrichment in downregulated genes ( P = 0 . 0004 , Figure 2B . Genes are listed in Table S4 ) . Very interestingly , two of the nine genes were Smad6 and Nog , inhibitors for BMP-SMAD pathway [25] . The majority , 189 of the 239 genes , had neither modification in MEFp2 with very low expression levels . These genes acquired de novo H3K27me3 mark in RasV12 cells , but did not show any more downregulation ( P = 1 , Figure 2B ) . Around TSS of Bmp2 , a secreted factor for BMP-SMAD pathway , loss of H3K27me3and gain of H3K4me3 were validated by quantitative ChIP-PCR ( Figure 3A , 3B ) . ChIP-PCR also showed that Ezh2 enrichment was significantly decreased around Bmp2 in RasV12 cells ( Figure S6 ) . ChIP-PCR showed that H3K4me3 and H3K27me3 levels in MEFp2 were sustained in Mock and RasG12 , but specifically altered in RasV12 cells ( Figure 3B ) . Quantitative RT-PCR showed very low level of Bmp2 expression in MEFp2 , Mock cells and RasG12 cells , but marked increase to 91 . 6-fold in RasV12 cells ( Figure 3C ) . Bmp2 activation thus occurred specifically in Ras-induced senescence , different from p16 that partially showed increased expression and histone methylation alteration during passages ( Figure 1C ) . Retrovirus to express shRNA against Bmp2 ( shBmp2 ) was infected together with RasV12 infection , to knock down Bmp2 to 0 . 04–0 . 08 fold on days 3 , 7 , and 10 ( Figure 3D ) . Bmp2-knocked-down RasV12 cells escaped from senescence with decreased number of SA-βgal ( + ) cells compared to RasV12 cells . While Smad1/5/8 is known to serve principally as substrates for BMP receptors [26] , western blotting analysis revealed phosphorylation of Smad1/5/8 in RasV12 cells ( Figure 3E ) . Decrease of Smad1/5/8 phosphorylation level was also shown in Bmp2-knocked-down RasV12 cells ( Figure 3E ) , and continual cell growth faster than Mock cells ( Figure 3F ) . To confirm that this escape from senescence was specifically due to Bmp2 knockdown , Bmp2-knocked-down RasV12 cells were cultured with recombinant BMP2 protein ( rBMP2 , R&D systems #355-BM ) at 0 , 20 and 200 ng/mL in culture medium with 10% serum . The cells showed increased number of SA-βgal ( + ) cells in dose-dependent manner , even to the level of RasV12 cells when rBMP2 was at 200 ng/mL ( Figure 3G , 3H ) . The level of Smad1/5/8 phosphorylation was increased when rBMP was added ( Figure 3I ) , and growth curve showed growth arrest similar to senescent RasV12 cells ( Figure 3F ) . These results indicated that Bmp2 upregulation plays an important role in Ras-induced senescence . There was no increase of SA-βgal ( + ) cells when Mock cells or MEF cells without infection were exposed to rBMP2 at 200 ng/mL , indicating that increase of BMP2 alone is not enough to induce cellular senescence ( Figure 3G ) . As for Smad6 , a specific inhibitor for BMP-SMAD pathway , gain of H3K27me3 and loss of H3K4me3 in RasV12 cells were found and validated by quantitative ChIP-PCR ( Figure 4A , 4B ) . H3K27me3 and H3K4me3 levels in MEFp2 were sustained in Mock and RasG12 cells , and altered specifically in RasV12 cells . This indicated that these alterations of histone methylation were not detected in stress-induced senescence during passages , but specifically occurred in Ras-induced senescence , like Bmp2 . Markedly decreased expression of Smad6 to 0 . 05-fold specifically in RasV12 cells was also validated by quantitative RT-PCR , while there was no repression of Smad6 during passages ( Figure 4C ) . Ezh2 enrichment was also analyzed by ChIP-PCR ( Figure 4D ) . This histone methyltransferase for H3K27 was significantly increased around TSS of Smad6 in RasV12 cells . It was indicated that Ezh2 was recruited to this de novo H3K27 trimethylation site , and that repression mechanism by de novo H3K27me3 was still active although Ezh2 expression level itself was downregulated during senescence , and Jmjd3 expression level was upregulated ( Figure S4A ) . Smad6 with N-terminal Myc tag was introduced to MEF by retroviral infection together with RasV12 virus , and their simultaneous expression was confirmed by cellular immunofluorescence ( Figure 4E and Figure S7 ) . Western blotting analysis and cellular immunofluorescence showed decrease of Smad1/5/8 phosphorylation in Smad6-introduced RasV12 cells compared to RasV12 cells ( Figure 4F , 4G ) . Smad6-introduced RasV12 cells showed decreased number of SA-βgal ( + ) cells compared to RasV12 cells ( Figure 4H and Figure S8 ) and showed continual cell growth faster than Mock cells or Smad6-introduced Mock cells ( Figure 4I ) . These data indicated that Smad6 repression was important in Ras-induced senescence . Nog , another inhibitor for BMP-SMAD pathway , was repressed to 0 . 06-fold in RasV12 cells also by losing H3K4me3 and gaining H3K27me3 ( Figure 5A , 5B ) . Introduction of Nog cDNA by retrovirus infection together with RasV12 resulted in its overexpression and escape from senescence ( Figure 5B–5D ) . To clarify whether Nog at the physiological expression level could inhibit cellular senescence , Nog-transgenic ( Nog-Tg ) mice under Krt19 promoter [27] was used next , since the transgene was expected not to be modified with de novo H3K27me3 . Krt19 was expressed in MEFp2 at much higher level compared to brain and testis , confirming that Krt19 promoter is active in MEF ( Figure S9 ) . Nog-Tg female mouse was crossed with C57B6 , to establish and pool Tg ( - ) and Tg ( + ) MEFs from embryos of the same mother . Tg ( + ) MEF showed Nog expression at similar level to wild type MEFp2 and Tg ( - ) MEF ( Figure 5E ) . While Tg ( - ) MEF showed Nog repression by RasV12 infection similarly to wild type MEF , Tg ( + ) MEF did not show Nog repression by RasV12 infection and showed continual growth faster than Tg ( - ) MEF ( Figure 5E–5G ) . These indicated that Nog repression was also important in Ras-induced senescence . It was reported that oncogenic Ras induces DNA methylation-mediated epigenetic inactivation in NIH3T3 cells [28] , and that EZH2 directly controls DNA methylation [29] , [30] . We therefore performed bisulfite sequencing to analyze DNA methylation statuses of 5′ regions of Smad6 and Bmp2 where increase or decrease of Ezh2 was confirmed ( Figure 4D , Figure S6 ) . There was no methylation alteration of these regions in RasV12 cells compared to MEFp2 ( Figure 6A ) . Also , Dnmt1 expression level was not altered during Ras-induced senescence ( Figure 6B ) . To gain insight whether oncogenic Ras induces DNA methylation-mediated inactivation in MEF on genome-wide scale , we performed methylated DNA immunoprecipitation ( MeDIP ) -seq in MEFp2 and RasV12 cells ( Figure 6C , 6D ) . Although MeDIP is reported to be not accurate to detect DNA methylation in low-CpG regions , it is powerful screening method to detect candidate methylation regions in high-CpG regions , e . g . promoter CpG islands [22] , [23] , [31] . Increase of methylation was detected only in three candidate genes , and the increase was considered as a noise in genome-wide analysis because the increase was not validated by bisulfite sequencing ( Figure 6D , 6E ) . Bisulfite sequencing was performed for five more genes which showed slight increase of methylation in MeDIP-seq , but there was no methylation alteration in RasV12 cells compared to MEFp2 ( Figure 6F ) . Human fibroblast IMR90 was infected with RasV12 retrovirus ( RasV12-IMR90 cells ) . It was confirmed by SA-βgal staining on day 7 that cells fell into premature senescence ( Figure S10A ) . Real-time RT-PCR showed that BMP2 expression was markedly increased to 145-fold in RasV12-IMR90 cells , while SMAD6 and NOG expressions were decreased to 0 . 32-fold and 0 . 15-fold , respectively ( Figure S10B ) . Nog was introduced in IMR90 by retroviral infection with RasV12 , and Nog-induced RasV12-IMR90 cells showed continual cellular growth ( Figure S10C ) , suggesting that BMP2-SMAD1 is also an effector program in human fibroblasts . Since Bmp2 upregulation , Smad6 repression , and Nog repression were shown to contribute to Ras-induced senescence , downstream target genes of Bmp2-Smad1 signal are further analyzed on genome-wide scale . Smad1 binding sites in MEF were analyzed by exposing MEF to rBMP and ChIP-sequencing using anti-Smad1 antibody ( Figure 7A and Figure S11 ) . Smad1 mostly bound to gene regions; 1 , 103 ( 75% ) out of 1 , 479 Smad1 binding sites were located within 10 kb from 20 , 232 RefSeq genes , and 818 sites ( 55% ) were within 5 kb from their TSS . Using GADEM ( http://www . niehs . nih . gov/research/resources/software/gadem/ ) [32] , GGGGCGGGGC was extracted as highly enriched motif within Smad1 binding region in both whole genomic and TSS regions ( Figure 6B , Figure S12 ) . Using DME ( http://rulai . cshl . edu/dme/ ) [33] , it was confirmed that very similar motifs e . g . GGGCGGGGC ( Figure 7B ) or GGGGCGGGGM ( Figure S13 ) were enriched . This was in good agreement with the canonical SMAD1-bound GC-rich elements [26] , [34] , [35] and the previous report that the sequence GGCGGGGC was enriched within Smad1/5 binding regions in ES cells and pulled down SMAD proteins [36] . Genes with Smad1 binding site at TSS regions were significantly enriched in active genes in MEF , especially in genes upregulated by rBMP exposure ( Figure 7C ) , suggesting that Smad1 binding correlates to gene upregulation . Smad1 target genes upregulated most by rBMP exposure included Smad6 , which was upregulated by 4 . 5-fold in MEF ( Figure 7D , 7E ) . These indicated that Bmp2/Smad1 signal in MEF could be controlled by negative feedback through Smad1 regulation on Smad6 . However , Smad6 was repressed in RasV12 cells by H3K27me3 , so when Bmp2-knocked-down RasV12 cells was exposed to rBMP2 , Smad6 level was still suppressed lower than the level in MEFp2 ( Figure 7E ) . Smad1 target genes repressed in RasV12 cells were not limited to Smad6 . H3K27me3 gain during Ras-induced senescence was detected in 50 Smad1 target genes , which were enriched in genes repressed in RasV12 cells , e . g . Atoh8 . Atoh8 was highly upregulated in BMP2 exposure , but repressed in RasV12 cells with decrease of H3K4me3 from 8 . 7 to 1 . 8 and increase of H3K27me3 mark from 1 . 0 to 1 . 8 ( Figure 7C , Figure 8A and 8B . Gene list is available in Table S5 ) . It was reported that Atoh8 was , like Id1 , suggested to be a direct target of BMP-SMAD signal [37] . On the contrary , Smad1 target genes without increased repressive mark were shown to keep upregulation . Among 838 Smad1 target genes , 581 with no increase of H3K27me3 , or 156 showing decrease of H3K27me3 , were significantly enriched in genes upregulated in RasV12 cells ( P = 0 . 01 and P = 0 . 004 , respectively , Figure 8A ) . If Bmp2/Smad1 signal is critical in senescence , the most upregulated target genes are expected to include genes with growth suppressor function . To choose such candidate genes , the most upregulated target genes were screened using promoter methylation data of our previous methylated DNA-immunoprecipitation ( MeDIP ) -chip analyses of human cancer cells [22] , [23] ( Table S6 ) , since such genes may possibly be frequently inactivated in human cancer . The most upregulated targets then included Parvb , which showed promoter methylation in human cancer cell lines HCT116 and DLD1 ( Table S6 ) . When MEF senesced , Parvb showed increase of H3K4me3 from 8 . 6 to 16 . 8 , and decrease of H3K27me3 from 1 . 0 to 0 . 6 ( Figure 8B ) . Real-time RT-PCR validated increase of Parvb expression in RasV12 cells , and also when exposed to rBMP2 ( Figure 8C ) . When Parvb was knocked down to 0 . 05-fold by shRNA , SA-βgal ( + ) cells were partially decreased and cells showed continual growth ( Figure 8D and Figure S14 ) . Western blot analysis showed decrease of Akt phosphorylation in exposure to a growth factor or serum when Parvb with C-terminal V5 tag was introduced in MEF ( Figure 8E ) .
In this study , we examined H3K4me3 and H3K27me3 marks for genome-wide analysis of epigenomic changes , revealing that activation of Bmp2-Smad1 signal is important in Ras-induced senescence and it is regulated by dynamic epigenomic alteration in coordinated manner . Different from p16 , H3K4me3 and H3K27me3 marks on Bmp2 was not altered during passage in cell culture , but specifically altered in RasV12 cells to induce its marked upregulation , leading to Smad1/5/8 phosphorylation and cellular senescence . Decrease of Ezh2 and increase of Jmjd3 were detected in RasV12 cells at similar levels to MEFp5 , Mock cells and RasG12 cells . This may contribute to partial increase of p16 expression in MEFp5 , Mock cells and RasG12 cells , and partial decrease of H3M27me3 mark on p16 in stress-induced senescence during passages as reported [16] , [18] . However , the alterations on p16 were more markedly detected in RasV12 cells , and the alterations on Bmp2 and Smad6 were specifically detected in Ras-induced senescence and did not occur during passages . It is noteworthy that de novo formation of H3K27me3 occurs on Smad6 in RasV12 cells in spite of general decrease of Ezh2 and increase of Jmjd3 . The mechanism how these epigenetic regulations are programmed is largely unknown , but one possible answer might be non-coding RNA [38] , [39] . PRC2 was reported to be recruited in trans to its target gene by virtue of its association with HOTAIR , a 2 . 2 kb non-coding RNA in the HOXC locus [40] . Oncogenic Ras inhibited expression of ANRIL ( antisense non-coding RNA in the INK4 locus ) ; ANRIL showed binding to CBX7 within PRC1 and SUZ12 in PRC2 , and was important in repressing the protein-coding genes of INK4b/ARF/INK4a locus in cis to regulate senescence [41] , [42] . Ezh2 recruitment was increased in Smad6 , and decreased in Bmp2 and p16 ( Figure 1D , Figure 4D , Figure S6 ) . It would be interesting to analyze whether any non-coding RNAs recruit PRC to Smad6 and Bmp2 in cis or trans , and their expression alterations contribute to epigenetic alterations of these genes during Ras-induced senescence . Gene repressions by other epigenetic mechanism than Polycomb , such as H3K9 methylation , would be interesting to be analyzed next . Human fibroblasts in senescence are reported to suppress DNA damage response by forming heterochromatic foci , where regions with methylated H3K9 gathered [12] . Amplification of SETDB1 , a methyltransferase for H3K9 , was recently reported to play an accelerating role in melanoma onset [13] , while knockout of Suv39h1 , another histone methyltransferase for H3K9 , caused escape from senescence of lymphocytes [43] , suggesting necessity of adequate control of H3K9 methylation . Genome-wide analyses of methylated H3K9 and other epigenomic marks as well would be helpful to obtain the whole picture of epigenomic alteration and its importance in senescence . As for DNA methylation , it was reported that oncogenic Ras induces DNA methylation-mediated epigenetic inactivation in NIH3T3 cells and that 28 responsible genes including DNMT1 are required for the methylation [28] . DNA methylation statuses at 5′ regions of Smad6 and Bmp2 were not altered , however , indicating that expression changes of these genes during senescence were not due to DNA methylation . Dnmt1 level was not altered in RasV12 cells , either . Increase of methylation was detected only in three candidate genes by MeDIP-seq analysis , and the increase was considered as a noise in genome-wide analysis because the increase was not validated by bisulfite sequencing . Five more genes were chosen for bisulfite sequencing , because Sfrp1 was reported to be methylated by oncogenic Ras in NIH3T3[28] , and four other genes were chosen randomly from genes with slight increase of methylation in MeDIP-seq . There was no methylation alteration in RasV12 cells compared to MEFp2 , either . Although MeDIP is not accurate to detect DNA methylation in low-CpG regions [22] , [31] , it was suggested that DNA methylation unlikely occurs in Ras-induced senescence , at least high-CpG regions e . g . promoter CpG islands . The discrepancy between the previous report of NIH3T3 and our MEF result may be because MEF falls into cellular arrest by oncogenic stress and there might be no time enough to induce DNA methylation alteration . In NIH3T3 , cells transform by oncogenic Ras , and may have time enough to acquire DNA methylation during continuing proliferation . Or , two independent cells , NIH3T3 ( ATCC #CRL-1658 ) and K-ras-transformed NIH3T3 ( ATCC #CRL-6361 ) , were compared in the previous NIH3T3 study [28] , so the result might be different if one NIH3T3 clone is analyzed at time courses before and after induction of activated Ras . As for BMP-SMAD signals , utilization of four BMP type 1 receptors depends on BMP ligands; BMP2 and BMP4 utilize BMPR1A and BMPR1B , BMP6 and BMP7 bind principally to ACVR1 , and BMP9 is a ligand for ACVRL1 and ACVR1 [26] . We reported that Smad6 inhibited BMPR1A/BMPR1B preferentially to ACVR1/ACVRL1 , and inhibited BMP2-induced Smad1/5 phosphorylation more prominently than BMP6-induced Smad1/5 phosphorylation [44] . This is in agreement with the current results that Smad6 could cause decreased phosphorylation of Smad1/5 and escape from senescence , and that coordination of Bmp2 upregulation and Smad6 repression was critical in Ras-induced senescence . Our genome-wide analysis showed that Smad6 was a Smad1 target gene that could be highly upregulated by exposure to BMP2 , but strongly repressed in RasV12 cells with de novo H3K27me3 mark . Previous reports showed that BMP-activated Smad1/5 activates Smad6 expression through interaction with the Smad6 promoter [45] , [46] . These suggested that Smad6 repression with de novo H3K27 methylation blocks negative feedback loop to sustain the effect of upregulated Bmp2 , i . e . activation of Bmp2-Smad1 signal in Ras-induced senescence . In other words , dynamic H3K27me3 alteration is suggested to repress selectively the genes which could negatively control senescent signal , and to activate selectively genes which could positively affect senescent signal ( Figure 9 ) . In fact , another BMP-SMAD inhibitor , Nog , was also repressed by increased H3K27me3 mark . While ChIP-seq analysis did not show Smad1 binding site around Nog TSS , Nog was also highly upregulated by rBMP2 exposure ( Figure 6C ) and repressed by increased H3K27me3 mark in RasV12 cells ( Figure 5A , 5B ) . This might suggest that Nog repression could also be a disruption of negative feedback loop , though Nog is not a direct downstream target of Bmp2-Smad1 . Parvb , which possessed Smad1 binding site around its TSS , was upregulated in exposure to BMP2 or in RasV12 cells , and its knock down lead to escape from senescence . While PARVA was reported to bind to integrin-linked kinase ( ILK ) and play a critical role in cell survival by promoting membrane recruitment of Akt and its activation by phosphorylation , PARVB was reported to compete PARVA in binding to ILK and reverse its oncogenic effect by repressing ILK kinase activity [47] , [48] . As PARVB introduction was reported to suppress cellular growth of breast cancer cells with decreased Akt phosphorylation [49] , [50] , Parvb introduction in MEF also decreased phosphorylation of Akt in exposure to a growth factor or serum ( Figure 7E ) . It was suggested that Parvb might be one of Bmp2-Smad1 target genes playing a positive role in growth inhibition , at least partly , and selectively and effectively activated through simultaneous inactivation of negative regulators . We chose Parvb on the assumption that candidate genes downstream of BMP-SMAD might be inactivated by DNA methylation in full-blown cancers , but other downstream genes that were not methylation target in analyzed cancer cell lines might also play a positive role in senescence . Aberrations in BMP-SMAD signal have been frequently reported in human cancer . Juvenile polyposis syndrome , an inherited syndrome with high risk of colorectal cancer , is caused by germline mutation of BMPR1A or SMAD4 [51] , and importance of BMP signal is supported by its mouse model with transgenic Nog expression or with Bmpr1a inactivation [52] , [53] . BMP2 expression was lost in microadenoma of familial adenomatous polyposis , while BMP2 was expressed in mature colonic epithelial cells , promoting apoptosis and differentiation and inhibiting proliferation [54] . Inactivation of BMPR1A , BMPR2 , and SMAD4 was frequently observed in sporadic colorectal cancer , correlating to loss of Smad1/5/8 phosphorylation [55] . Colon epithelial polyps were developed even by alteration of BMP pathway in the stromal microenvironment , using mice with conditional inactivation of Bmpr2 in the stroma [56] . About prognosis , Smad6 expression was reported to be elevated in 40% of non-small cell lung cancer , and correlated to poorer outcome [57] . BMP2 upregulation was reported in senescence of other cell types , such as vascular smooth muscle cells [58] . Considering frequent RAS gene mutation in cancer , e . g . colon ( ∼40% ) and non-small cell lung cancers ( ∼30% ) [59] , further experiments are to be performed to clarify which cell types Bmp2-Smad1 signal is critical in oncogene-induced senescence , and whether Bmp2-Smad1 signal and its target genes are disrupted in cancer with association to oncogene mutation .
MEF was established from 13 . 5 embryonic day embryos of C57/B6 as reported [60] . After cells were passed twice ( MEFp2 ) , cells were infected with retroviruses for 48 hours . Then cells were exposed to 4 µg/mL puromycin for selection during days 0–3 , and were passed on days 3 , 7 , and 10 . Human fibroblast IMR90 ( JCRB9054 ) was purchased from Health Science Research Resources Bank ( Osaka , Japan ) , and 2 µg/mL puromycin were used for selection after retrovirus infection . Total RNA was collected using TRIzol ( Invitrogen , Carlsbad , CA ) . This study was certified by Animal Ethics Committee in Tokyo University . Nog-Tg mice using keratin 19 gene promoter and mouse Nog cDNA were previously established [27] , and were crossed with wild type C57/B6 mice five times to obtain C57/B6 background . Nog-Tg female mouse was crossed with C57B6 , and Tg ( - ) and Tg ( + ) MEFs were established from 13 . 5 embryonic day embryos of the same mother . Each embryo was minced separately , and Tg ( - ) and Tg ( + ) MEFs were pooled after genotyping each MEF , and used for experiments . Retroviral vectors for Ras was constructed by cloning cDNAs for wild type HRAS ( RasG12 ) and mutated HRAS ( RasV12 ) by reverse-transcription PCR products from HMEC and SK-BR3 cell RNA , respectively , with N-terminal FLAG tag into pMX vector that contains puromycin resistance gene ( a kind gift from T . Kitamura ) . Mock pMX vector ( Mock ) , and vectors containing RasG12 and oncogenic RasV12 were transfected into plat-E packaging cells ( a kind gift from T . Kitamura ) using FuGENE 6 Transfection Reagent ( Roche , Germany ) to prepare retroviruses . Smad6 cDNA with N-terminal 6x Myc tag , Nog cDNA with C-terminal V5 tag , and Parvb cDNA with C-terminal V5 tag were also cloned into pMX vector . To knock down Bmp2 or Parvb , double strand oligonucleotide DNA to express small hairpin RNA against Bmp2 ( shBmp2 ) or Parvb ( shParvb ) , respectively , was cloned into RNAi-Ready pSIREN-RetroQ Vector ( Clontech , CA ) . Viral packaging for Smad6 , Nog , shBmp2 and shParvb retrovirus vectors was also done using plat-E cells . Retroviruses of RasV12 and Nog for human fibroblast were prepared using Retrovirus Packaging Kit Ampho ( #6161 , TaKaRa Bio Inc , Shiga , Japan ) . For genome-wide transcription analysis , GeneChip Mouse Genome 430 2 . 0 Array ( Affymetrix ) was used as described [61] . The GeneChip data were analyzed using the Affymetrix GeneChip Operating Software v1 . 3 by MAS5 algorithms , to obtain signal value ( GeneChip score ) for each probe . For global normalization , the average signal in an array was made equal to 100 . Gene annotation enrichment analysis was done at DAVID Bioinformatics Resources ( http://david . abcc . ncifcrf . gov/ ) . Array data is available at GEO datasets ( #GSE18125 ) . MEFp2 and infected cells at day 10 were cross-linked with 1% formaldehyde for 10 min and were prepared for ChIP . ChIP using anti-H3K4me3 ( ab8580 , abcam , rabbit polyclonal ) , H3K27me3 ( 07–142 , Upstate , rabbit polyclonal ) , or Ezh2 ( #39103 , Active Motif , rabbit polyclonal ) antibody was performed as described previously [62] . For ChIP using anti-Smad1 antibody ( BioMatrix , mouse monoclonal ) , MEFp2 cells were starved for 16 hours and exposed to rBMP2 protein ( #355-BM , R&D systems ) at 25 ng/mL in serum-free medium for 1 . 5 hours . Cells were cross-linked with 1 mM Disuccinimidyl Glutarate ( Thermo Scientific , Rockford , IL ) for 20 min and 1% formalin for 10 min , and ChIP was performed similarly . For MeDIP , genomic DNA of MEFp2 and RasV12 cells was fragmented by sonication , and immunoprecipitated by anti 5-methylcytocine monoclonal antibody ( kindly supplied by Dr . K . Watanabe , Toray Research Center , Inc . ) , as we previously reported[22] , [23] , [63] . MeDIPed sample and Input sample underwent MeDIP-PCR to check enrichment of methylated regions in MeDIPed sample . Sample preparation for ChIP- and MeDIP-sequencing was performed according to the manufacturer's instructions ( Illumina ) , and sequencing was performed using Solexa Genome Analyzer II [61] . 36-bp single end reads were mapped to the NCBI Build #36 ( UCSC mm8 ) reference mouse genome , using the Illumina pipeline software v1 . 4 . The numbers of uniquely mapped reads for MEFp2 were 10 , 845 , 082 ( H3K4me3 ) , 11 , 519 , 151 ( H3K27me3 ) , 9 , 663 , 324 ( DNA methylation ) and 5 , 688 , 804 ( Input ) , those for RasV12 cells were 13 , 246 , 871 ( H3K4me3 ) , 9 , 894 , 241 ( H3K27me3 ) , 11 , 319 , 506 ( DNA methylation ) and 6 , 126 , 206 ( Input ) , and that for Smad1 ChIP-sequencing was 9 , 417 , 307 . Window sizes of 300 bp for H3K4me3 , 500 bp for H3K27me3 , 500 bp for DNA methylation and 300 bp for Smad1 , were used to calculate the number of mapped reads per million reads at the center of the window . Sequencing data is also available ( #GSE18125 ) . Aliquots of protein were subjected to SDS/PAGE and were transferred to nitrocellulose , and the resulting immunoblots were visualized using Amersham ECL Plus ( GE Healthcare ) and LAS-3000 ( Fujifilm , Japan ) . Phosphorylated Smad1/5/8 was detected using antibody against phospho-Smad1/5/8 ( Cell Signaling ) as primary antibody , and green-fluorescent Alexa Fluor 488 dye-labeled anti-rabbit antibody ( Invitrogen ) as secondary antibody . RasV12 with N-terminal FLAG tag and Smad6 with N-terminal Myc tag were detected using antibody against FLAG ( F7425 , Sigma , rabbit polyclonal ) and Myc ( 9E10 , Santa Cruz , mouse monoclonal ) as primary antibody , respectively , and Alexa Fluor 594 anti-rabbit antibody and Alexa Fluor 488 anti-mouse antibody ( Invitrogen ) as secondary antibody . Photographs were taken with Biozero BZ-8100 ( KEYENCE , Osaka , Japan ) . MEFp2 and infected MEFs on day 10 , and infected IMR90 on day 7 underwent SA-βgal staining as previously described [64] . Infected MEFs were counted on days 3 , 7 , 10 , 14 , 17 , 21 using Countess automated cell counter ( Invitrogen ) and seeded at density of 1×105 cells/6-cm dish for every passage . Infected IMR90 were counted on days 4 , 8 , 12 and 16 similarly , and seeded at density of 2 . 5×105 cells/6-cm dish . Mean number of three dishes was calculated and used to draw growth curve . Real-time PCR was performed using iCycler Thermal Cycler ( Bio-Rad Laboratories ) as previously described [65] . The experiment was triplicated and mean and standard error were calculated and shown . Primer information is in Tables S7 and S8 . DNA methylation status was analyzed by bisulfite sequencing as previously described [65] . Briefly , 500 ng of genomic DNA of MEFp2 and RasV12 cells underwent bisulfite treatment , and were finally suspended in 20 µL of distilled water . For bisulfite sequencing , 1 µl was used as a template for PCR with primers common for methylated and unmethylated DNA sequences . The primers and PCR conditions are available at Table S9 . PCR products were cloned into pGEM-T Easy vector ( Promega ) , and 9–10 clones each were cycle-sequenced using T7 and Sp6 primers . | To avoid becoming cancer cells , cells have a barrier system to block cellular proliferation by falling into irreversible growth arrest , so-called cellular senescence . For future strategy of cancer treatment , it is important to understand how cancer occurs , and investigation of underlying mechanism in senescence can lead to clarification of carcinogenesis mechanism . Epigenetic mechanism including DNA methylation and histone modification may be important to regulate gene expressions properly in senescence . Here , taking advantage of recent technical and methodological advance of genome-wide analyses , we examine epigenome and gene expression alteration in senescence induced by Ras oncogene . We identify that Bmp2-Smad1 signal is critical . We further examine downstream target genes of this critical signal on a genome-wide scale . We show dynamic and coordinated H3K27me3 alteration , e . g . activation of Bmp2 by loss of H3K27me3 , repression of the signal inhibitors and the negative feedback loop by gain of H3K27me3 , and selective activation of downstream target genes that may contribute to growth arrest . Our findings are helpful in understanding the importance of epigenetic regulation and a critical signal in the physiological barrier system against oncogenic transformation and the importance of disruption of BMP-SMAD signal in cancer , and they may provide an idea how cancer with Ras mutation occurs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome",
"sequencing",
"genome",
"analysis",
"tools",
"cancer",
"genetics",
"genome",
"scans",
"genetics",
"epigenetics",
"biology",
"genomics",
"genetics",
"and",
"genomics"
] | 2011 | Activation of Bmp2-Smad1 Signal and Its Regulation by Coordinated Alteration of H3K27 Trimethylation in Ras-Induced Senescence |
Geobacter species are of great interest for environmental and biotechnology applications as they can carry out direct electron transfer to insoluble metals or other microorganisms and have the ability to assimilate inorganic carbon . Here , we report on the capability and key enabling metabolic machinery of Geobacter metallireducens GS-15 to carry out CO2 fixation and direct electron transfer to iron . An updated metabolic reconstruction was generated , growth screens on targeted conditions of interest were performed , and constraint-based analysis was utilized to characterize and evaluate critical pathways and reactions in G . metallireducens . The novel capability of G . metallireducens to grow autotrophically with formate and Fe ( III ) was predicted and subsequently validated in vivo . Additionally , the energetic cost of transferring electrons to an external electron acceptor was determined through analysis of growth experiments carried out using three different electron acceptors ( Fe ( III ) , nitrate , and fumarate ) by systematically isolating and examining different parts of the electron transport chain . The updated reconstruction will serve as a knowledgebase for understanding and engineering Geobacter and similar species .
Microorganisms play a major role in the global carbon cycle . Insights into the various mechanisms and energetic constrains which govern their behavior will advance our understanding of carbon fluxes and might ultimately allow for a rational perturbation of the carbon cycle . Key features of the carbon cycle are the conversion of organic and inorganic carbon and the energy flow through the system . Qualitative and quantitative knowledge of carbon assimilation and electron flow to and from key microorganisms is critical when evaluating certain aspects of the carbon cycle . Members of the genus Geobacter are ubiquitous in the soil environment , and have been described to utilize various organic substrates while transferring electrons to insoluble metals externally [1] . Furthermore , certain Geobacter species , such as G . metallireducens have been reported to transfer electrons directly to poised electrodes [2] and even to other microbes [3] , a process coined direct interspecies electron transfer or DIET . Quantitative assessment of carbon and energy flow in G . metallireducens by computational modeling approaches therefore provides valuable insight into the role of this bacterium in the carbon cycle . Geobacter metallireducens , the first Geobacter species that was isolated [4] , serves as a pure culture model for the study of many of the important reactions that Geobacter species catalyze in the biogeochemistry of anaerobic soils and sediments , groundwater bioremediation , and several bioenergy applications [5] . For example , Geobacter species play a major role in the biogeochemical cycling due to their ability to couple the oxidation of organic compounds to the reduction of Fe ( III ) and Mn ( IV ) oxides [5] . G . metallireducens was the first microorganism shown to be capable of the anaerobic degradation of aromatic hydrocarbons [6] and Fe ( III ) is an important electron acceptor for the removal of aromatic hydrocarbons in many contaminated subsurface environments [5] . Direct electron transfer from electrodes to microorganisms to drive anaerobic respiration has potential applications in bioenergy and bioremediation [7] . Constraint-based reconstruction and analysis ( COBRA ) is a powerful method for characterizing the content of an organism , or systems of organisms , and understanding the limits of its collective functionality [8] . Metabolic network reconstruction ( the most widely utilized form of COBRA ) enables the enumeration of the genome-wide machinery ( i . e . , enzymes , uptake systems , etc . ) in an organized fashion for use in modeling [9] . With a reconstructed network for an organism , predictions can be made about its functionality when combined with physiological data in a modeling framework . Further , a validated and accurate network can be utilized for prospective design and engineering of cellular networks [10] . There is a history of modeling Geobacter sp . using COBRA [11] . One of the first studies utilizing COBRA and Geobacter was for G . sulfurreducens [12] . The key findings of this study were an initial reconstruction and examination of the extracellular electron transport , the examination of the efficiency of internal biomass biosynthetic pathways , and predictions of gene deletion phenotypes . A subsequent study branched off to build a reconstruction of G . metallireducens based on the original content of the G . sulfurreducens reconstruction [13] . The initial G . metallireducens reconstruction was used to examine the efficiency of pathway usage in the network along with yield on a variety of substrates . In this work , an updated reconstruction was built and analyzed to better understand key capabilities of G . metallireducens . The updated reconstruction effort was fueled by the appearance of a more complete genome annotation [14] and new data available for the electron transport chain and key metabolic content [15] , [16] .
An updated reconstruction of G . metallireducens GS-15 , iAF987 , was generated by reconciling an existing genome-scale reconstruction [13] and an updated genome annotation , performing a bottom-up reconstruction of additional metabolic pathways . This new reconstruction was functionally tested for performance under known growth conditions ( Figure 1A ) . The final reconstruction contained 987 genes , 1284 reactions , and 1109 metabolites . In the first phase , the existing reconstruction was compared to the updated genome annotation [14] to identify a list of agreements , discrepancies , and scope for expansion . A distinct periplasm compartment was determined to be important as G . metallireducens has the unique ability to transfer electrons extracellularly [5] . Thus , characterizing the electron transfer pathways from the cytosol through the periplasm to the extracellular space was crucial for understanding this unique capability . Furthermore , the addition of the periplasm compartment allows for a more accurate representation of metabolism , such as p-cresol and 4-hydroxybenzyl alcohol degradation , which partially occurs in the periplasm [17] . A wild-type and a reduced ‘core’ biomass objective function [18] were formulated to validate whether the reconstruction could generate the appropriate biomass components necessary to replicate , and for use in simulation to predict the growth rates of the organism on the different substrates . Gaps were filled in the network using data characterizing growth of G . metallireducens GS-15 on 19 different carbon sources/electron donors with Fe ( III ) as the electron acceptor , and the SMILEY algorithm [19] ( see Text S1 ) . The iAF987 reconstruction was compared to the previous version [13] and an automatically generated reconstruction from the ModelSEED framework [20] ( Figure 1B ) was used to identify and evaluate newly reconstructed and unique content . The ModelSEED reconstruction was found to have 114 unique genes that were not present in the iAF987 reconstruction . Of these , 86 genes were involved in macromolecular synthesis , DNA replication , and protein modifications that are beyond the scope of a metabolic network , and 8 of them did not have a specific reaction association in the ModelSEED ( i . e . , generic terms such as aminopeptidase , amidohydrolase ) . Of the remaining 20 genes , only two ( Gmet_0988 and Gmet_2683 ) were added to the reconstruction as isozymes for existing reactions; the other 18 assignments conflicted with our functional annotation of the genome and thus were not included . The iAF987 reconstruction contains 227 genes not in either reconstruction , thus representing a significant advancement of coverage . These newly included genes encode several unique pathways , encoding 325 unique reactions , which have not previously appeared in a collection of 14 representative reconstructions ( Table 1 ) from the UCSD database from which iAF987 was constructed and is internally consistent ( see Text S1 for a detailed comparison of iAF987 to previous work ) . Transcriptomic data profiling a growth shift from acetate to the aromatic compound benzoate was integrated with the metabolic model to validate its content . Specifically , the computational analysis was performed using the MADE algorithm [21] which uses the statistical significance of changes in gene expression to create a functional metabolic model that most accurately recapitulates the expression dynamics . Of the 987 genes in the metabolic model , transcriptomic data indicated that the expressions of 857 genes do not change significantly and 130 were differentially expressed ( >2-fold and p-value<0 . 05 ) . Specifically , 77 genes were up-regulated and 53 were down-regulated during this shift . The MADE algorithm predicted that during this metabolic shift , the expression of 885 genes in iAF987 do not change significantly and 102 genes are differentially expressed . Of the 102 differentially expressed genes , the MADE algorithm predicted the up-regulation of 70 genes and down-regulation of 32 genes . Specifically , the model predicted upregulation of 70 genes , while data indicated 77 genes to be upregulated during this shift . Similarly while the model predicted downregulation of 32 genes , the data actually indicated that 53 genes were downregulated during this shift . The model-based prediction of change in expression disagreed with the in vitro transcriptomic data for only 28 of the 987 genes leading to 97% overall agreement ( for a more detailed breakdown , see Text S1 and Table S1 ) . Among the genes differentially expressed during the shift , the genes encoding for benzoyl-CoA reductase were up-regulated over 100-fold during benzoate growth . It was determined that this key enzyme that links the degradation of aromatic substrates to central metabolism is not ATP driven as previously thought [13] , but is likely membrane bound and proton translocating [16] . Thus , a proton translocating reaction was added to the reconstruction for this step in metabolism . A translocation stoichiometry of 3 protons per electron was determined to be the likely extent of coupling through a thermodynamic analysis ( see Text S1 ) . Similar transcriptomic analyses for growth shifts on two other aromatic electron donors ( i . e . , toluene and phenol ) yielded 86% and 84% agreement , respectively ( Table S1 ) . These findings will likely broaden our knowledge of how G . metallireducens can be utilized for bioremediation . The genome of G . metallireducens GS-15 encodes two out of the six known carbon fixation pathways [22] . The pathways which were reconstructed in iAF987 are the reductive citric acid cycle ( rTCA ) and the dicarboxylate–hydroxybutyrate cycle [22] ( Figure 2A ) . Key enzymes for the rTCA include the 2-oxoglutarate synthase ( abbreviated OOR2r in the reconstruction ) and ATP-citrate lyase ( ACITL ) , both of which enable the citric acid cycle to run in reverse . For the dicarboxylate–hydroxybutyrate cycle , the key enzyme is 4-hydroxybutyryl-CoA dehydratase ( 4HBCOAH ) . The rTCA and the dicarboxylate–hydroxybutyrate cycles share four reactions . The reconstruction of these carbon fixation pathways led to the prediction of a new growth condition for G . metallireducens ( Table 2 ) . With the expanded content including the carbon fixation pathways , it was computationally predicted that G . metallireducens can grow with formate as the electron donor and Fe ( III ) as the electron acceptor when a computational screen of all possible media combinations was performed with the model ( Table 2 ) . Investigating the resulting flux distributions revealed that the CO2 derived from formate oxidation is reduced via the rTCA to form acetyl-CoA which is subsequently assimilated into biomass . The electrons derived from formate oxidation are split between running the rTCA and for Fe ( III ) reduction . The energy gained by formation of a proton gradient during Fe ( III ) reduction was instrumental for providing the required ATP for carbon fixation . This prediction of growth on formate and Fe ( III ) was experimentally validated , Figure 2B . The requirement of CO2 fixation for growth of G . metallireducens solely on formate and CO2 as carbon sources is further highlighted by a study which examined G . sulfurreducens reducing Fe ( III ) with formate as the electron donor . While G . sulfurreducens was able to reduce Fe ( III ) with formate as electron donor , it required the addition of 0 . 1 mM acetate to assimilate cell carbon ( i . e . , grow ) [23] . This was attributed to the lack of an rTCA in G . sulfurreducens , specifically due to the absence of the ATP-dependent citrate lyase . Overall , this example represents a unique power of the model to rapidly generate hypotheses in silico that can then be verified experimentally . This particular result proves to be of great interest for examining carbon fixation . To reconstruct the electron transport system of G . metallireducens , key steps involving electron transfer to the terminal electron acceptor were subject to a thermodynamic analysis ( Text S1 ) . Specifically , the feasibility of proton translocation and the theoretical maximum proton translocation stoichiometry was determined [24] , [25] . Subsequently , the included content was analyzed with physiological data from growth screens to predict stoichiometries for the key reactions of the electron transport system ( ETS ) . This was performed in a model-driven iterative process with an approach that delineates the energetics of extracellular electron transfer by examining three distinct modules ( Figure 3A–B ) . These modules were characterized by representative electron acceptors; fumarate , nitrate , and Fe ( III ) , respectively . Reduction of fumarate in G . metallireducens requires the strain to harbor the dcuB gene [26]; nitrate is reduced via the ETS and the nitrate reductase [14] , [27]; extracellular Fe ( III ) requires several extracellular c-type cytochromes and pili [28] , which have been shown in the closely related G . sulfurreducens to have metal-like conductivity [29] . Acetate consumption has previously been described as a preferred optimal growth condition for G . metallireducens [27] , therefore it was used as the primary electron donor when analyzing the ETS in this study . The pathways and key ETS reactions for this conversion are shown in Figure 3C and evidence for the inclusion of the reactions in iAF987 is given in the Text S1 . The first step in the modeling process was to examine the fumarate reductase reaction and maintenance energies necessary for predicting phenotypes using constraint-based analysis ( Step 1 , Figure 3C ) . To examine this content , data was generated using the G . metallireducens dcuB strain and utilized in simulations . An electrogenic bifunctional fumarate reductase/succinate dehydrogenase ( FRD2rpp ) was included based on recent biochemical evidence in a similar species [15] . This inclusion was validated with the growth data of G . metallireducens on acetate and fumarate using the dcuB strain . The earlier version of the G . metallireducens model [13] and the current model ( iAF987 ) without an electrogenic fumarate reductase were unable to produce a feasible solution when constrained with experimentally measured acetate and fumarate uptake rates ( see Table 3 ) . However , the electrogenic fumarate reductase enabled the model to reproduce the experimental observations and was assigned to translocate two protons per two electrons transferred to fumarate . It was determined that two protons were necessary to drive the endergonic oxidation of succinate with menaquinol as part of the TCA cycle during Fe ( III ) respiration and nitrate reduction ( see Text S1 ) . Maintenance energies were also evaluated in this step by choosing a growth condition that eliminated any use of the external electron transferring reactions ( so these reactions could be examined in isolation later ) . G . metallireducens dcuB strain was grown in batch using acetate/fumarate medium and also in a chemostat at a set growth rate ( Table 3 ) . The growth rate determined in this study was similar to that previously reported ( 0 . 114 hr−1 vs 0 . 105 hr−1 , respectively ) [26] . From these two conditions , it was possible to estimate the maintenance costs ( GAM and NGAM ) for the model using an established procedure [9] , [18] . The calculated costs were 79 . 20 mmol ATP gDW−1 for the GAM and 0 . 81 mmol ATP gDW−1 hr−1 for the NGAM . These values are similar to those found previously for G . sulfurreducens [12] , thus it provided confidence in the use of the GAM and NGAM throughout the study . The additional content included in the ETS was further analyzed using experimental data . The second step in the modeling process to examine extracellular transfer was to examine the energetics of the menaquinone cytochrome oxidoreductase reaction ( CYTMQOR3 ) in the ETS ( Step 2 , Figure 3B ) . This was performed by switching the electron acceptor from fumarate to nitrate , while keeping acetate as the electron donor . Assuming the nitrate reductase translocates two protons per two electrons ( a value that has been verified [30] ) , the translocation stoichiometry of the menaquinone cytochrome oxidoreductase can be isolated . Analyzing the energetics of the CYTMQOR3 reaction , it was determined that a likely number of three protons are translocated per two electrons transferred to the cytochrome pool ( see Text S1 ) . Using this stoichiometry and simulating growth with acetate and limiting nitrate ( see Table 3 ) , the predicted optimal growth rate was calculated to be 0 . 054 hr−1 as compared to the experimentally determined value of 0 . 050 hr−1 ( in this experiment , the growth rate is equal to the set dilution rate in a steady state chemostat ) . Thus , it was concluded that the proton translocating assignment of the CYTMQOR3 reaction was consistent with the experimental results of the growth screen . The final step taken to predict the energetic cost of transferring electrons to an external substrate was to examine growth of G . metallireducens on acetate and Fe ( III ) after reconciling the other components of the ETS consistent with observed phenotypic data . Figure 3C , step 3 shows the reconstructed path from the inner membrane cytochromes to the outer membrane cytochromes and eventually to reduce Fe ( III ) . An energetic cost was estimated in terms of an ATP cost proportional to the flux of electrons to Fe ( III ) . Growth screens of wild type G . metallireducens ( acetate/Fe ( III ) ) were performed in triplicate in both batch and chemostat cultures . When culturing with Fe ( III ) as an electron acceptor , a measurement of biomass is challenging given that the optical density is used to calculate the amount of reduced iron ( see Methods ) . Therefore , the ratio of acceptor produced to donor consumed was used to compare to simulations , as it is more precise than a biomass-normalized uptake and production rate that was calculated by measuring protein content in the chemostat . Furthermore , the ratio of acceptor to donor calculated for both the batch and chemostat conditions was very similar , thus providing a consistent experimental comparison ( see Table 3 ) . A Phenotypic Phase Plane ( PhPP ) analysis [31] was used to compare model-predicted performance to the experimentally measured acceptor to donor ratio . It was calculated that a cost of one proton translocated across the inner membrane per one electron transferred ultimately to Fe ( III ) best matched the line of optimality in the PhPP analysis ( see Figure 3 C , Figure S1 ) . Further , this cost is very close to 0 . 3 ATP per electron transferred ultimately to Fe ( III ) , as the ATP synthase in the cell converts protons to ATP at a ratio of 3 . 33 protons per ATP ( see Text S1 ) . Further analysis of this modeling approach with phenotypic data on different electron donors ( butanol , ethanol , and pyruvate ) yielded the same cost of external electron transfer ( see Table S3 ) . Thus , it was hypothesized that this is the approximate cost for external electron transfer to iron and the reactions and costs were built into the iAF987 reconstruction as such . This cost can now be further validated for different external electron transfer processes that G . metallireducens is known to carry out . The work presented here demonstrates how constraint-based modeling and reconstruction can be applied to generate hypotheses that can be tested experimentally . Specifically , modeling revealed a non-obvious culturing condition where carbon fixation could be directly examined . Further , the cost of external electron transfer could be quantified using an iterative and systematic approach . Carbon sequestration is of great biotechnological interest [32] . By understanding the mode of growth for CO2 fixation , computational predictions can be used to guide genetic modifications which enhance the rate of CO2 fixation . Specifically , this could be in the form of reaction knockouts or identification of genes which could be targeted for overexpression which are predicted to enhance CO2 fixation . The reconstructed model advances our knowledge for this unique species and provides a platform for further analysis and hypothesis formulation for environmental and biotechnology applications .
Curation of the genome annotation of G . metallireducens was continued after the initial publication [14] with additional insights from curation of the genome annotations of Geobacter bemidjiensis [33] , Pelobacter carbinolicus [34] , and other species ( M . Aklujkar , unpublished ) , with extensive reference to the MetaCyc database [35] . The updated annotation was submitted to NCBI with reference numbers CP000148 and CP000149 . The reconstruction was generated in a four-step process . First , the updated genome annotation for G . metallireducens was entered into the UCSD SimPheny ( Genomatica , San Diego , CA ) database [14] . Next , the existing G . metallireducens reconstruction [13] was mapped to the updated genome using the existing gene-protein-reaction associations [8] and all of the pathways excluding membrane lipid biosynthesis , lipopolysaccharide biosynthesis , murein biosynthesis and degradation , and transport were entered into the SimPheny framework if an exact match for the reaction was present in the UCSD SimPheny database . If an exact match for the reaction did not exist in the UCSD SimPheny database on the level of metabolites participating in the reaction , they were manually evaluated for inclusion ( see below ) . Next , a comparison of the metabolic content included in the updated genome annotation ( Dataset S1 ) that was not in SimPheny was performed . Manual evaluation of new content or disagreements from the annotation and existing genome-scale reconstruction consisted of gathering genetic , biochemical , sequence , and physiological data and reconciling this information to determine the likelihood of each reaction being present in the organism . This manual curation process has been described and reviewed several times [8] , [9] . In the manual review process , the KEGG database ( www . genome . jp/kegg/ ) , the ModelSEED database [20] , and primary literature ( see Dataset S2 ) were used extensively in the manual curation process . Confidence scores were given for each reaction along with noteworthy evidence used to justify inclusion of a given reaction . The BOFs were formulated using a previous template [36] and are included in Dataset S3 . The biomass content previously determined for the close species Geobacter sulfurreducens was used to determine the breakdown of macromolecules [12] except that for total carbohydrate as the distribution in the murein , lipopolysaccharide , and cytosolic fractions was not indicated . Further , the genome annotation [14] was used for the breakdown of chromosome bases , a study on lipid and lipopolysaccharide chain length was used for the breakdown of acyl chain length [37] , and the remaining content was approximated using the full profile presented for the gram-negative bacterium E . coli [36] . It should be noted that prediction of growth rate and unmeasured uptake rates are relatively insensitive realistic variations in biomass macromolecular weight fractions [36] . The BOF components included in the core BOF were extrapolated from the core BOF formulated for E . coli as Geobacter sp . have the same gram-negative cell morphology . The COBRA Toolbox 2 . 0 [38] and the SimPheny framework ( Genomatica , Inc . , San Diego , CA ) were used for simulations . Constraints used to simulate growth in ferric citrate medium are presented in Text S1 . For the evaluation of the maintenance energies , the best-fit values of acetate and fumarate were set to the exact values . Growth screens were performed in triplicate using cultures in 125 mL serum bottles under anoxic conditions using Fe ( III ) or fumarate ( with the dcuB mutant studies ) as an electron acceptor . The composition of the fumarate medium and ferric citrate medium were the same as previously described [39] and [40] , respectively . The concentrations of the electron donors utilized in the experiments were 15 mM acetate , 20 mM ethanol , and 10 mM butanol . For the growth screens , the cells were passed in the media in which they were being tested for at least three passages before the growth screen was performed . For the dcuB mutant strain , the correlation between OD and biomass that was used was 0 . 4561 gDW L-1 OD600-1 . This value was determined by growing the dcuB cells in freshwater medium , taking OD measurements at various time points , and weighing the dried biomass at various points of the growth curve . The cells were dried overnight in an oven and weighed on filter paper , with a correction for the amount of weight lost by a filter paper that did not contain cells . Analytes were quantified by HPLC using an Aminex 87-H ion exchange column at 65° C . The mobile phase was 5 mM H2SO4 at an isocratic flow of 0 . 5 mL/minute . Sample injection volume was 10 µL . Products were identified by retention time using ultraviolet detection at 210 nm and refractive index detection at 30° C internal temperature and 45° C external temperature and quantified by relating peak area to those of standards . Fe ( II ) was measured using the ferrozine assay as described [40] . Growth rates were calculated by determining the exponential growth phase region from a series of samples taken for each growth screen ( typically , 7–10 samples were taken over the entire screen ) . For this corresponding exponential growth phase region , the ratio of an analyte to the gDW of the sample was determined using a linear fit obtained through the least-squares method ( ‘regress’ function in MATLAB ) . This value ( mmol gDW-1 ) was then multiplied by the growth rate to get a corresponding uptake or production rate . The ratios were subsequently multiplied by the growth rate to get the uptake or production rates . Averages and standard deviations were reported and calculated from three biological replicates for each experiment . For the G . metallireducens GS-15 dcuB stain , the fumarate and succinate rates were independently calculated and then averaged for each replicate as there is a 1∶1 ratio for the dcuB exchange and experimental accuracy for each analyte differed slightly . The fumarate uptake rate was calculated using the net fumarate and malate concentrations measured [41] . Malate was observed when growing with ethanol and butanol as electron donors with the dcuB strain . For the acceptor rates in these two conditions , the succinate production rate was used as the rate of the donor to minimize compounding measurement error ( although the effective fumarate uptake rate was very similar , <27% difference in any individual replicate ) . Chemostat cultures of G . metallireducens GS-15 wild-type and dcuB cultures were grown at 30°C in anaerobic continuous culture vessels as previously described [42] . The GS-15 wild-type strain was grown under donor limiting conditions with 5 mM acetate and 55 mM ferric citrate , the dcuB strain was grown under acetate limiting conditions with 5 mM Acetate and 27 . 5 mM fumarate . A protein conversion value of protein content ( ug/mL ) = 0 . 4344*dry-cell-weight ( ug/mL ) was used to calculate rates . To determine if G . metallireducens was capable of growth with formate , cells were adapted to 20 mM formate in a ferric-citrate medium [40] and not provided with any other electron donor . After the third transfer with formate as the sole electron donor , samples were withdrawn anaerobically and substrate consumption was monitored . Formate was detected via high performance liquid chromatography . Reduction of Fe ( III ) to Fe ( II ) was monitored with the ferrozine assay . Experiments were run with triplicate cultures . The number of cells was monitored over time until the cultures reached a plateau of Fe ( II ) production . To fix cells , 900 µL culture was withdrawn anaerobically and mixed with 100 µL glutaraldehyde ( 25% ) . Fixed cells were stored at −20°C until processed . All samples were defrosted , placed on filters , filtered and washed with sterile water prior to staining with acridin orange for 3 minutes . Filters were washed and dried . Cell counts were carried out for three biological samples by manually counting at least 5 microscopic fields ( Area 5826 µm2 per field ) per sample , with the help of a cell counting application in ImageJ ( http://rsbweb . nih . gov/ij/ ) . Control cultures consisting of ( i ) cells in ferric-citrate medium with no formate , and ( ii ) ferric-citrate medium with formate and no cells ( i . e . , cell-free ) were also performed . For the microarray experiments performed , G . metallireducens was grown anaerobically with Fe ( III ) citrate ( 55 mM ) as the electron acceptor and one of the following electron donors: acetate ( 10 mM ) , benzoate ( 1 mM ) , phenol ( 0 . 5 mM ) , or toluene ( 0 . 5 mM ) . Cells were grown in 1 L bottles and harvested during mid-exponential phase by centrifugation . The cell pellet was immediately frozen in liquid nitrogen and stored at −80 °C . RNA was isolated from triplicate cultures grown on each electron donor with a modification of the previously described method [43] . Briefly , cell pellets were resuspended in HG extraction buffer [44] pre-heated at 65 °C . The suspension was incubated for 10 minutes at 65 °C to lyse the cells . Nucleic acids were isolated with a phenol-chloroform extraction followed by ethanol precipitation . The pellet was washed twice with 70% ethanol , dried , and resuspended in sterile diethylpyrocarbonate-treated water . RNA was then purified with the RNA Clean-Up kit ( Qiagen , Valencia , CA , USA ) and treated with DNA-free DNASE ( Ambion , Woodward , TX , USA ) . The RNA samples were tested for genomic DNA contamination by PCR amplification of the 16S rRNA gene . cDNA was generated with the TransPlex Whole Transcriptome Amplification Kit ( Sigma ) . Whole-genome microarray hybridizations were carried out by Roche NimbleGen , Inc . ( Madison , WI , USA ) . Triplicate biological and technical replicates were conducted for all microarray analyses . Cy3-labeled cDNA was hybridized to oligonucleotide microarrays based on the G . metallireducens genome and resident plasmid sequences ( accession number NC007515 and NC007517 at GenBank ) . All the microarray data has been deposited with NCBI GEO under accession number GSE33794 . For each metabolic shift ( benzoate vs acetate , toluene vs acetate , phenol ) , the fold-changes in expression level and p-value ( t-test ) were computed using ArrayStar 4 . 02 ( DNASTAR , Madison , WI , USA ) . This was used in conjunction with the iAF987 metabolic model to predict the metabolic adjustment using the MADE algorithm [21] . Additionally , the constraints for the iAF987 metabolic model were set to simulate growth on the respective substrates . MADE analysis was implemented using the TIGER toolbox [45] . The output of this analysis consisted of the genes predicted by MADE algorithm to significantly change in expression during the concerned metabolic shift . | The ability of microorganisms to exchange electrons directly with their environment has large implications for our knowledge of industrial and environmental processes . For decades , it has been known that microbes can use electrodes as electron acceptors in microbial fuel cell settings . Geobacter metallireducens has been one of the model organisms for characterizing microbe-electrode interactions as well as environmental processes such as bioremediation . Here , we significantly expand the knowledge of metabolism and energetics of this model organism by employing constraint-based metabolic modeling . Through this analysis , we build the metabolic pathways necessary for carbon fixation , a desirable property for industrial chemical production . We further discover a novel growth condition which enables the characterization of autotrophic ( i . e . , carbon-fixing ) metabolism in Geobacter . Importantly , our systems-level modeling approach helped elucidate the key metabolic pathways and the energetic cost associated with extracellular electron transfer . This model can be applied to characterize and engineer the metabolism and electron transfer capabilities of Geobacter for biotechnological applications . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"biotechnology",
"applied",
"microbiology",
"bioengineering",
"systems",
"biology",
"biological",
"systems",
"engineering",
"microbial",
"metabolism",
"microbial",
"physiology",
"industrial",
"microbiology",
"engineering",
"and",
"technology",
"biology",
"and",
"life",
"sci... | 2014 | Constraint-Based Modeling of Carbon Fixation and the Energetics of Electron Transfer in Geobacter metallireducens |
Dysregulation of AMPK signaling has been implicated in many human diseases , which emphasizes the importance of characterizing AMPK regulators . The tumor suppressor FLCN , responsible for the Birt-Hogg Dubé renal neoplasia syndrome ( BHD ) , is an AMPK-binding partner but the genetic and functional links between FLCN and AMPK have not been established . Strikingly , the majority of naturally occurring FLCN mutations predisposing to BHD are predicted to produce truncated proteins unable to bind AMPK , pointing to the critical role of this interaction in the tumor suppression mechanism . Here , we demonstrate that FLCN is an evolutionarily conserved negative regulator of AMPK . Using Caenorhabditis elegans and mammalian cells , we show that loss of FLCN results in constitutive activation of AMPK which induces autophagy , inhibits apoptosis , improves cellular bioenergetics , and confers resistance to energy-depleting stresses including oxidative stress , heat , anoxia , and serum deprivation . We further show that AMPK activation conferred by FLCN loss is independent of the cellular energy state suggesting that FLCN controls the AMPK energy sensing ability . Together , our data suggest that FLCN is an evolutionarily conserved regulator of AMPK signaling that may act as a tumor suppressor by negatively regulating AMPK function .
Birt-Hogg-Dubé syndrome ( BHD ) is an autosomal dominant neoplasia disorder that was originally described by Hornstein and Knickenberg in 1975 and by Birt , Hogg , and Dubé in 1977 as a disorder associated with colon polyps and fibrofolliculomas of the skin [1] , [2] . Toro et al . recognized in 1999 that BHD patients were also predisposed to develop kidney cancer mostly of the onococytic , chromophobe , or mixed subtype [3] . However , later studies showed a predisposition for all subtypes of kidney cancer including clear cell and papillary subtypes [4] . In addition , BHD confers an increased risk of pulmonary cysts , spontaneous pneumothorax , and cysts of the kidney , pancreas , and liver [3] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] . The gene responsible for BHD , FLCN , was mapped to chromosome 17p11 . 2 by linkage analysis [15] , [16] and identified in 2002 by positional cloning [14] . FLCN encodes a novel cytoplasmic 64kDa protein FLCN , which is expressed in most epithelial tissues [17] . BHD patients carry a loss of function germline mutation in one FLCN allele and acquire a second hit somatic mutation or loss of heterozygosity ( LOH ) in the remaining wild-type copy in their renal tumors [18] , [19] . In addition , strains of rats , mice , and dogs with a germline mutation in the Flcn gene developed spontaneous kidney tumors with a loss of function in the second allele pointing to a tumor suppressor function of FLCN [20] , [21] , [22] , [23] . However , homozygous deletion of Flcn resulted in embryonic lethality in these species [24] , [25] . Finally , ablation or restoration of FLCN in human cancer cells revealed tumor suppressor function in xenograft and soft agar assays [24] , [26] . Though the FLCN protein presents no significant homology to any known protein , it is highly conserved from unicellular organisms ( yeast ) through mammalian species ( rodents , dog , humans ) . Moreover , two 130 kDa folliculin-interacting proteins , FNIP1 and FNIP2 have been identified [27] , [28] , [29] and implicated in some of the FLCN phenotypes in B-cell and stem cell differentiation , and the regulation of apoptosis upon DNA damage [30] , [31] , [32] , [33] . Several studies identified both FLCN and FNIP1/2 as AMPK ( 5′AMP-activated protein kinase ) binding proteins [28] , [34] , [35] , [36] . However , no clear role for FLCN/FNIP1/2 in AMPK function has been described , since both inhibition and stimulation of AMPK have been reported upon loss of function of these genes [32] , [37] . Strikingly , the majority of naturally occurring FLCN mutations predisposing to BHD were predicted to generate truncated proteins unable to bind AMPK pointing to an essential role of this interaction in the tumor suppressor function . Since we and others have observed that FLCN regulates cellular metabolism [37] , [38] , [39] , we hypothesized that FLCN may regulate cellular energy metabolism through its interaction with AMPK . AMPK is an evolutionarily conserved master regulator of energy metabolism [40] , [41] , [42] . When energy levels drop , AMP or ADP bind to the γ regulatory subunit of AMPK and induce an allosteric conformational change [43] , [44] . This change leads to the activation of AMPK through phosphorylation of a critical threonine residue ( Thr172 ) in the catalytic subunit and inhibition of its dephosphorylation . When animals and cells encounter stressful environmental conditions leading to lower energy levels , activated AMPK phosphorylates downstream metabolic targets to generate ATP and maintain bioenergetics [40] , [41] , [42] . For instance , AMPK activates autophagy , a lysosome-dependent degradation process that recycles cytosolic components to generate new cellular components and produce energy [45] . Recently , AMPK was shown to activate autophagy via binding and phosphorylation of the autophagy initiation kinase ULK1 , Beclin 1 , and Vps34 [46] , [47] , [48] . Since studies in mammalian cells have led to unclear roles for FLCN in AMPK function , we decided to study the genetic relationship between FLCN and AMPK in the model organism C . elegans . FLCN and AMPK are conserved in C . elegans and loss-of-function mutants are viable . AMPK activation promotes lifespan extension in C . elegans [49] , [50] , [51] , [52] and increases resistance to oxidative and other stresses [51] , [52] , [53] , [54] , [55] , [56] , [57] . Here we show that FLCN controls a distinct evolutionarily conserved energy stress pathway by acting as a negative regulator of AMPK function . Loss of FLCN function led to constitutive AMPK activation , which increased autophagy , resulting in inhibition of apoptosis , higher bioenergetics , thereby enhancing survival to several metabolic stresses . Specifically , we find that the chronic activation of autophagy upon loss of FLCN modifies cellular metabolism , providing an energetic advantage that is sufficient to survive metabolic stresses such as oxidative stress , heat , and anoxia . We confirmed these C . elegans results in FLCN- deficient mouse embryo fibroblasts ( MEFs ) and human cancer cells demonstrating strong conservation of this pathway throughout evolution . Our results demonstrate that FLCN inhibits AMPK and autophagy functions , which may lead to inhibition of tumorigenesis .
The FLCN gene product is highly conserved from C . elegans to humans with 28% identity and 50% similarity ( Figure 1A ) . To determine the function of FLCN and whether it genetically interacts with AMPK in C . elegans , we used a strain carrying the flcn-1 ( ok975 ) mutation . The ok975 mutation is an 817 bp insertion-deletion , predicted to truncate the protein at residue 141 resulting in a null or loss-of-function allele ( Figure 1B ) . In accordance , the C . elegans FLCN-1 polyclonal antibody that we developed recognized a gene product at the predicted size in N2 wild-type but not in flcn-1 ( ok975 ) animals ( Figure 1C ) . Importantly , we did not detect obvious developmental or morphological defects in flcn-1 ( ok975 ) animals compared to wild-type . The C . elegans AMPK ortholog ( aak-2; α2 catalytic subunit ) modulates longevity and tolerance to stresses including oxidative stress , heat , anoxia , and dietary restriction [49] , [50] , [51] , [52] , [54] , [56] . Since we did not observe a difference in lifespan between wild-type and flcn-1 ( ok975 ) animals ( Figure 1D and Table S1 ) , we investigated the function of FLCN-1 in stress response by treatment of animals with mild and acute oxidative stress . The flcn-1 ( ok975 ) mutation conferred an increased resistance to 4 mM and 100 mM paraquat ( PQ ) , a superoxide inducer [58] , which could be rescued in two different transgenic lines expressing FLCN-1 ( Figures 1E and 1F and Tables S2 and S3 ) . In addition , treatment with flcn-1 RNAi increased the resistance of wild-type animals to low and high concentrations of PQ but did not further increase the resistance of the flcn-1 ( ok975 ) mutant animals , supporting that the ok975 mutation is a loss-of-function allele ( Figures 1G and 1H ) . A similar resistance phenotype was observed upon H2O2 treatment and was rescued with the two transgenic lines expressing flcn-1 ( Figures 1I and S1A ) . To exclude the possibility that the changes in stress resistance are not due to effects on lifespan of the animals , assays performed on 4 mM PQ were accompanied with lifespan controls ( Table S2 ) . In conclusion , these results demonstrate that loss of FLCN increases resistance to oxidative stress in C . elegans . Since FLCN binds to AMPK in mammalian cells , we aimed to determine whether flcn-1 and aak-2 interact genetically in C . elegans . Similarly to published results [50] , [51] , [52] , [54] , aak-2 ( ok524 ) mutant animals were more sensitive to PQ stress compared to wild-type ( Figures 2A and 2B and Tables S2 and S3 ) . Strikingly , flcn-1 ( ok975 ) ; aak-2 ( ok524 ) double mutant animals displayed reduced survival upon treatment with 4 mM PQ ( Figure 2A and Table S2 ) and 100 mM PQ ( Figure 2B and Table S3 ) , similarly to aak-2 ( ok524 ) single mutants , indicating that aak-2 ( ok524 ) is required for the flcn-1 ( ok975 ) phenotype . The C . elegans AMPKα1 homolog ( AAK-1 ) was previously shown to be dispensable for oxidative stress resistance [54] . Accordingly , the aak-1 ( tm1944 ) mutation did not abolish the increased survival of flcn-1 ( ok975 ) mutants to PQ ( Figure 2C and Table S3 ) . To further test whether the increased survival of flcn-1 ( ok975 ) mutants was also dependent on PAR-4 , the C . elegans ortholog of LKB1 and major upstream kinase of AMPK [40] , [54] , we measured survival to PQ upon par-4 loss . Interestingly , par-4 ( it57 ) , a strong loss of function allele , only partially suppressed the flcn-1 ( ok975 ) survival phenotype , leading to a significant increase in the resistance of flcn-1 ( ok975 ) ; par-4 ( it57 ) animals to PQ compared to par-4 ( it57 ) animals alone , suggesting that additional inputs might activate AAK-2 to mediate survival ( Figure 2D and Table S3 ) . Based on these results , we anticipated that loss of flcn-1 might lead to a constitutive activation of AAK-2 since the observed flcn-1 loss of function phenotype is similar to the reported AAK-2 overexpression in terms of oxidative stress resistance [49] , [52] . Although we did not observe an increased abundance of phospho-AAK-2 at residue Thr234 ( corresponding to Thr172 in human AMPKα ) in flcn-1 ( ok975 ) animals compared to wild-type , our data demonstrate a significant increase in phospho-AAK-2 levels in flcn-1 ( ok975 ) ; par-4 ( it57 ) double mutants compared to par-4 ( it57 ) animals ( Figure 2E and 2F ) . This is consistent with the observation that par-4 is not fully required for the stress resistance phenotype ( Figure 2D and Table S3 ) . To exclude the possibility that the increased phosphorylation of AAK-2 is due to a flcn-1-dependent increase in total AAK levels , we measured the mRNA expression of aak-1 and aak-2 in wild-type and flcn-1 ( ok975 ) animals and did not observe a significant difference ( Figure S1B ) . Interestingly , the residues flanking the mammalian AMPKα ( Thr172 ) are conserved in AAK-2 ( Thr234 ) , but are different in AAK-1 and therefore , AAK-1 is unlikely to be detected by the antibody used ( Figures 2E , lane 5 ) . The increased phosphorylation of AAK-2 only in par-4 ( it57 ) mutants can be explained by the fact that PAR-4 is the major kinase that phosphorylates AAK-2 in certain tissues of the animal or cellular sub-compartments , which would mask the FLCN-dependent phosphorylation signal on AAK-2 induced by other upstream kinases [54] . Taken together , these results imply that flcn-1 negatively regulates aak-2 in C . elegans , and that loss of flcn-1 confers an aak-2-dependent resistance to oxidative stress . The insulin/IGF-1-like ( DAF-2 ) /FOXO3a ( DAF-16 ) and target of rapamycin ( TOR ) signaling pathways are known to control lifespan and stress response in C . elegans and other organisms and have been linked to AMPK signaling [50] , [52] , [59] , [60] , [61] . While daf-2 ( e1370 ) mutants exhibited an increased survival to PQ compared to wild-type animals , the flcn-1 mutation further increased the resistance of daf-2 ( e1370 ) animals ( Figure S2A and Table S3 ) . Consistently , daf-16 ( mu86 ) slightly reduced but did not suppress the resistance of the flcn-1 ( ok975 ) animals to PQ ( Figure S2B and Table S3 ) . Moreover , we found that the PQ resistance of the flcn-1 ( ok975 ) animals treated with TOR ( let-363 ) RNAi was significantly higher than wild-type animals fed with the same RNAi ( Table S3 ) . Transcriptional upregulation of ROS detoxification enzymes prior to stress could explain the increased survival of flcn-1 ( ok975 ) [62] . We did not observe a significant increase in the gene expression of superoxide dismutases ( sod-1 , sod-2 , sod-3 , sod-4 , and sod-5 ) or catalases ( ctl-1 , ctl-2 , and ctl-3 ) , in flcn-1 ( ok975 ) mutants when compared to wild-type ( Figure S2C ) . Furthermore , we quantified the oxidative damage to protein and DNA . Levels of protein carbonylation and DNA damage were equal in both wild-type and flcn-1 ( ok975 ) animals under basal conditions and were similarly induced after PQ treatment ( Figures S2D and S2E ) . Taken together , these findings suggest that the increased survival of flcn-1 ( ok975 ) mutant to PQ may not be dependent on classical oxidative stress resistance mechanisms . A likely mechanism of survival upon loss of flcn-1 on PQ might involve inhibition of apoptosis . Autophagy has been shown to mediate resistance to oxidative stress across evolution without a clear mechanistic explanation [63] . To investigate whether the increased oxidative stress resistance of flcn-1 ( ok975 ) mutants was due to autophagy , we measured autophagy levels using several methods . Using electron microscopy , we noticed the frequent appearance of autophagic vacuoles ( Figure 3B ) in flcn-1 ( ok975 ) mutants at the basal level compared to wild-type animals which increased under PQ treatment ( Figures 3A and 3C ) . To confirm this result , we used a reporter strain that carries the integrated transgene expressing GFP::LGG-1 ( LC3 ortholog ) . LC3 localizes to pre-autophagosomal and autophagosomal membranes , and GFP-positive puncta are thought to represent autophagosomal structures in this strain [64] , [65] , [66] . To exclude effects on the transgene expression , we determined the mRNA levels of LGG-1 in wild-type and flcn-1 ( ok975 ) animals and in the GFP::LGG-1 and flcn-1; GFP:: LGG-1 transgenic lines . In both cases , the transcript levels of LGG-1 were not significantly different between wild-type and flcn-1 ( ok975 ) animals demonstrating equal expression ( Figures S3A and S3B ) . Importantly , we observed a significant increase in the number of GFP-LGG-1 positive puncta in flcn-1 ( ok975 ) mutants compared to wild-type animals under basal conditions ( Figure 3D ) . Consistently , treatment of GFP::LGG-1 animals with flcn-1 RNAi increased the number of GFP-LGG-1 puncta ( Figure 3D ) . Previous studies in yeast , C . elegans , and mammalian cells have demonstrated that LGG-1-II ( or LC3-II ) is degraded inside the autolysosomes , and that the GFP fragment is resistant to degradation and accumulates when autophagy is induced [67] , [68] , [69] , [70] . Therefore , we performed western blot analysis on wild-type and flcn-1 protein extracts to assess the level and cleavage of GFP-LGG-1 . Importantly , western blot analysis showed that both cleaved LGG-1-II and released GFP were increased in flcn-1 ( ok975 ) mutants , indicating higher autophagic activity ( Figure 3E ) . AMPK has recently been shown to directly induce autophagy in mammals via phosphorylation of autophagy proteins including ULK-1 , VPS-34 and BEC-1 [46] , [47] , [48] . Moreover , loss of aak-2 reduced autophagy in daf-2 mutant animals , while aak-2 overexpression induced autophagy [46] . Based on these results , we questioned whether the increased autophagy in flcn-1 ( ok975 ) animals depends on aak-2 . Importantly , RNAi treatment against aak-2 significantly reduced the number of puncta in flcn-1 ( ok975 ) mutants , demonstrating an aak-2-dependent mechanism ( Figure 3D ) . Inhibition of autophagy genes in C . elegans reduced survival to certain stresses such as anoxia , starvation and pathogens [66] , [71] , [72] . However , the requirement of autophagy genes in resistance to oxidative stress was not previously reported . We aimed to determine whether the increased survival of flcn-1 ( ok975 ) animals to oxidative stress was dependent on autophagy . Strikingly , inhibition of the essential authophagy genes atg-7 and bec-1 using RNAi markedly abolished the resistance of flcn-1 ( ok975 ) to PQ ( Figures 3F–3I and Tables S2 and S3 ) . Taken together , these results demonstrate that loss of flcn-1 induces autophagy , which is required for flcn-1-mediated stress resistance . Autophagy is a process that generates catabolic substrates for mitochondrial ATP production and allows cellular macromolecules to be recycled . Since we did not observe a difference in oxidative damage between wild-type and flcn-1 ( ok975 ) mutant , and since PQ is known to severely decrease ATP levels by inhibiting oxidative phosphorylation [73] , [74] , [75] , we wondered if flcn-1 is mediating an increased resistance to energy stress by employing autophagy as a source of energy . To test this hypothesis , we measured ATP levels prior and after 13 hours of 10 mM PQ treatment . Strikingly , we found that flcn-1 mutant animals have higher levels of ATP before PQ treatment compared to wild-type ( Figure 4A ) . As expected , PQ treatment decreased the ATP levels in both wild-type and flcn-1 , yet ATP levels in flcn-1 ( ok975 ) nematodes remained higher than wild-type . Importantly , flcn-1 ( ok975 ) mutants treated with PQ exhibited equal amounts of ATP when compared to the non-treated wild-type animals ( Figure 4A ) . To test if the increased energy in flcn-1 ( ok975 ) mutants is dependent on autophagy , we treated the wild-type and flcn-1 nematodes with atg-7 RNAi and measured ATP levels . Strikingly , downregulation of autophagy completely suppressed the increased ATP levels in flcn-1 ( ok975 ) mutants in presence or absence of PQ ( Figure 4A ) . To further confirm that loss of flcn-1 confers resistance to low energy levels , we measured the resistance of wild-type and flcn-1 ( ok975 ) nematodes to heat stress and anoxia , both of which are known to result in a strong depletion of energy [76] . Accordingly , when exposed to 35°C , the mean survival of flcn-1 ( ok975 ) animals was significantly higher compared to wild-type ( Figure 4B and Table S4 ) . In addition , the recovery rates after a 26 hours anoxic injury were faster in flcn-1 ( ok975 ) compared to wild-type ( Figure 4C and Table S5 ) . In conclusion , our data describe a novel mechanism for AAK-2-dependent resistance to oxidative stress , which depends on maintenance of energy homeostasis via autophagy . The interplay between autophagy and apoptosis determines the decision between life and death which is very important for the genetic integrity of the cell [77] . The activation of autophagy has been shown to protect against cell death in C . elegans and mammals [72] , [77] , [78] , [79] . To see whether flcn-1 controls apoptosis in animals , we determined the number of apoptotic cell corpses in the gonad arms of wild-type and flcn-1 animals upon PQ treatment . As expected , we found that PQ treatment significantly increased the number of apoptotic corpses in wild-type animals ( Figure S4A ) . However , the increase in flcn-1 was much lower suggesting that loss of flcn-1 protects against cell death . To determine whether the decreased cell death in flcn-1 nematodes depends on the activation of autophagy , we pretreated the wild-type and flcn-1 ( ok975 ) nematodes with atg-7 RNAi and then measured the number of apoptotic corpses upon PQ treatment . Importantly , the inhibition of the autophagy gene atg-7 increased the number of apoptotic corpses , up to the same level , in both wild-type and flcn-1 suggesting that the FLCN-1-dependent activation of autophagy protects against cell death ( Figure S4A ) . Importantly , the apoptotic pathway is conserved from animals to mammals . When cells are destined to die , the BH3 only protein EGL-1 binds and inhibits the BCL-2 homolog CED-9 , which activates the caspase CED-3 and leads to death [80] . Therefore , we treated wild-type and flcn-1 ( ok975 ) animals with egl-1 and ced-3 RNAi and assessed their survival to 100 mM PQ . Importantly , downregulation of egl-1 or ced-3 using RNAi increased the resistance of wild-type animals which was not observed in flcn-1 ( ok975 animals suggesting that the inhibition of the apoptotic pathway leads to the increased resistance ( Figures S4B–S4D and Table S3 ) . Consistently , treatment of wild-type and flcn-1 ( ok975 ) animals with ced-9 RNAi abolished the survival of flcn-1 animals ( Figure S4 and Table S3 ) . Importantly treatment of aak-2 ( ok524 ) animals with ced-3 RNAi did not increase their resistance to PQ suggesting that this phenotype depends on AAK-2 ( Table S3 ) . The increased stress response by inhibition of apoptosis has been recently reported [81] . Although it is not clear whether the apoptosis inhibition in the gonad arms delays organismal death in C . elegans or whether apoptotic genes acquire non-apoptotic functions , our data suggest an AMPK-dependent involvement of the apoptotic pathway in the increased survival of flcn-1 animals to PQ stress ( Table S3 ) . To test whether the FLCN functions that we identified in C . elegans are evolutionarily conserved , we used wild-type ( Flcn+/+ ) and knockout ( Flcn−/− ) MEFs . First , we examined the cellular resistance to serum starvation ( -FBS ) , which reduces energy levels and induces oxidative stress in a physiological manner ( Figure 5A ) [58] . Flcn−/− MEFs were unaffected by serum starvation , as demonstrated by a significant maintenance of cell survival after 4 days of serum starvation , which eliminated almost 80% of wild-type cells . Rescue of wild-type FLCN expression ( resc . ) reverted this protective phenotype ( Figure 5A ) . Consistent with these data , Flcn−/− MEFs were more resistant to 2 mM H2O2 treatment compared to wild-type ( Figure S5A ) . Moreover , in accordance with the C . elegans results , phospho-AMPK levels were increased upon loss of Flcn in MEFs , which could be rescued by expression of wild-type Flcn , suggesting that FLCN acts a negative upstream regulator of AMPK ( Figures 5B and S2B ) . Additionally , down regulation of Flcn by shRNA in MEFs lacking AMPKα ( Ampk−/− ) did not increase resistance to serum starvation suggesting that the increased resistance to starvation-induced stress depends on AMPK ( Figures 5C and S5B ) . Next , we asked whether the AMPK activation in Flcn−/− MEFs could be further activated . Importantly , phosphorylation levels of AMPK and its target ACC were maximal in Flcn−/− cells and did not further increase upon serum starvation ( Figure S5C ) . Similarly , treatment with AICAR ( 5-amino-1-β-D-ribofuranosyl-imidazole-4-carboxamide ) , an AMP analogue , increased AMPK activation in wild-type as marked by elevated pACC levels but not in Flcn−/−MEFs ( Figure S5C ) . These results demonstrate that loss of FLCN leads to maximal AMPK activation , which appears uncoupled from its energy sensing function . Moreover , we wondered whether loss of FLCN also increases autophagy in MEFs similarly to the results obtained in C . elegans . Importantly , Flcn−/− MEFs displayed an increased number of autophagosomes at the basal level and under serum starvation conditions when compared to wild-type cells as determined by the GFP-LC3 reporter ( Figures 5D and 5E ) . To validate these results , we used a GFP-mCherry/LC3 reporter [82] . Upon physiological pH in newly formed autophagosomes or when autophagy is impaired , both GFP and mCherry colocalize in puncta whereas upon lysosomal fusion and acidification , the GFP signal is lost and only mCherry is detected . As expected , the number of mCherry puncta was increased in Flcn−/− MEFs , pointing to a normal lysosomal acidification and completion of autophagy ( Figure S6A ) . In addition , chloroquine ( CQ ) treatment , which inhibits the acidification of autolysosomes , further increased the number of autophagosomes in Flcn−/− MEFs suggesting that the autophagy process is not impaired ( Figures 5E and S6A ) . We also measured the rate of conversion of LC3I to LC3II by western blot analysis . The ratio of LC3II to LC3I was increased in Flcn−/− MEFs at the basal level and was reverted by FLCN rescue ( Figure S6B ) . In agreement with the heightened autophagy , we observed an increase in the activating AMPK-dependent phosphorylation site at the autophagy initiating kinase ULK1 ( Figure S5C ) . To determine whether the increased resistance to serum starvation was due to increased autophagy as we observed in C . elegans , we inhibited autophagy using CQ or Bec-1 shRNA . Inhibition of autophagy strongly suppressed the survival advantage upon serum starvation in Flcn−/− MEFs ( Figures 5F and S6C ) . Finally , we measured apoptosis in response to serum starvation and inhibition of autophagy in MEFs ( Figure S7D ) . Apoptosis was strongly increased in wild-type MEFs and suppressed in Flcn−/− MEFs upon serum starvation . Furthermore , inhibition of autophagy using CQ abolished the suppression of apoptosis in Flcn−/− MEFs . In conclusion , these results correspond well with the data we obtained in C . elegans . Next , we determined whether the chronic activation of autophagy is leading to an energy surplus , which is required for the stress resistance phenotype conferred by loss of FLCN . Similarly to what we found in C . elegans , Flcn−/− MEFs displayed increased ATP levels under basal conditions ( Figure 5G ) . Serum starvation decreased ATP levels in wild-type MEFs , while Flcn−/− MEFs maintained ATP levels at wild-type levels ( Figure 5G ) . Importantly , inhibition of autophagy with CQ abolished the increase in ATP in Flcn−/− MEFs ( Figure 5G ) . The ATP , ADP , and AMP levels as well as phospho-creatine , a short term energy reserve , were all increased at the basal level in Flcn−/− cells as compared to wild-type , suggesting a general increase in cellular metabolism and energy storage ( Figures S7A and S7C ) . Importantly , serum starvation decreased energy levels in Flcn−/− cells down to wild-type basal levels ( Figures S7A and S7C ) . Normally , AMPK is activated upon energy deprivation ( increased ADP/ATP and AMP/ATP ratios ) , which is inconsistent with the increased basal ATP levels conferred by loss of FLCN [40] , [42] . To understand this discrepancy , we determined the adenylate energy charge of wild-type and Flcn−/− MEFs . The adenylate energy charge , which is expressed by the ratio [ATP] + 0 . 5[ADP]/[ATP]+[ADP]+[AMP] , was proposed as a convenient indicator of the cellular energy status [83] . Under normal conditions , ATP levels were increased in Flcn−/− MEFs but the total energy charge was comparable to wild-type ( Figure S7B ) . However , serum starvation significantly reduced the energy charge in wild-type MEFs but not in Flcn−/− MEFs , suggesting that Flcn−/− cells derive their energy from an intracellular source of ATP [83] ( Figure S7B ) . To determine whether loss of FLCN in human cancer cells also conferred an advantage in energy homeostasis , we used the follicular thyroid carcinoma cells FTC-133 lacking FLCN expression , which we rescued for FLCN using stable transfection . First , we confirmed the findings that we obtained with wild-type and Flcn−/− MEFs . As expected , loss of FLCN conferred an increased phosphorylation of AMPK as well as a higher LC3 cleavage demonstrating that AMPK activation and autophagy in FLCN-deficient FTC cells is elevated compared to rescued cells ( Figure S8A ) . Next , we aimed to determine whether the increased autophagy promoted by loss of FLCN heightens ATP levels at normal conditions in FTC cells . Similarly to what we observed in MEFs , loss of FLCN requires autophagy to increase ATP levels ( Figure S8B ) . To determine whether autophagy contributes to anchorage-independent growth conferred by loss of FLCN , we performed soft agar assays in the presence or absence of autophagy inhibition using atg7 shRNAs . As expected , loss FLCN significantly increased the number of colonies growing in soft agar in an autophagy-dependent manner ( Figures S8C and S8D ) . Taken together , these data demonstrate that loss of FLCN leads to an autophagy-dependent increase in ATP levels enabling FLCN-deficient animals/cells to resist metabolic stresses , which could constitute a tumor suppression mechanism .
Maintenance of cellular bioenergetics and management of oxidative stress are essential for life . Here we highlight the discovery of an evolutionarily conserved signal transduction pathway mediated by the tumor suppressor FLCN and AMPK that is essential for resistance to metabolic stress . Loss of FLCN in C . elegans and mammalian cells leads to constitutive activation of AAK/AMPK , which in turn increases autophagy . Chronic activation of autophagy leads to increased ATP production and confers resistance to energy depleting stresses by inhibition of apoptosis . Together , our data identify FLCN as a key regulator of stress resistance and metabolism through negative regulation of AMPK . Several questions arise from these results . First , how is AAK-2 being activated in flcn-1 ( ok975 ) animals upon loss of PAR-4 ? Several upstream AMPK kinases other than LKB1 have been identified in mammalian cells , have been shown to affect AMPK activity [40] . Although these kinases have not been linked to AAK-2 in C . elegans , our data showed a significant basal phosphorylation of AAK-2 in par-4 ( it57 ) mutant animals . This is in agreement with a recently published study showing that starvation and mitochondrial poisons increased phospho-AAK-2 levels in par-4 ( it57 ) mutant animals , and that the starvation-induced aak-2 phenotypes were partially dependent on PAR-4 [84] . The fact that the AMPK signaling pathway is evolutionarily conserved suggests that AMPK upstream kinases other than PAR-4 are likely to exist in C . elegans . For instance , Pak1/Camkk-beta was first identified in yeast as Snf-1/AMPK-activating kinase and was proven later to act upstream of AMPK in mammalian systems [85] , [86] , [87] . Very recently , CAMKII overexpression was shown to increase lifespan in C . elegans , although the link to AAK-2 was not investigated [88] . Together , our data demonstrate that FLCN-1-dependent regulation of AAK-2 mediates an important novel pathway for stress resistance . Interestingly , this pathway is distinct from previously described AAK-2-mediated oxidative stress resistance mechanisms that involve ROS detoxification [51] , [52] , [89] . Another unexpected finding is that loss of flcn-1 did not modulate C . elegans longevity under normal growth conditions . The observed increased AAK-2 activation upon loss of flcn-1 is masked by PAR-4-dependent phosphorylation of AAK-2 . General overexpression of AMPK extends lifespan and increase stress resistance [46] , [49] , [50] , [51] , [52] , [53] , [56] , [90] . Our data suggest that the signaling cascade downstream of FLCN-1/AAK-2 is different from the AAK-2 responses that modulate longevity . It is possible that the PAR-4-dependent activation of AAK-2 extends lifespan and increases stress resistance , while the AAK-2 activation by other upstream kinases only increases resistance to stress . Another possibility would be that a tissue-specific or sub-cellular AAK-2 activation might lead to different outcomes . Importantly , our data indicate that loss of FLCN-1 extends lifespan only upon treatment with high concentrations ( 100 uM ) of the DNA synthesis inhibitor 5-fluoro-2′-deoxyuridine ( FUDR ) ( Figure S9F and Table S1 ) , a phenotype that has been recently reported by Gharbi et al . [91] . This drug is frequently used in C . elegans aging studies to prevent eggs from hatching and has been recently reported to “artifactually” affect lifespan in mitochondrial C . elegans mutants and modulate metabolism in the daf-2 mutant strain [92] , [93] , [94] . It is not clear why flcn-1 ( ok975 ) animals exhibit an extension of lifespan only upon treatment with FUDR and what would be the potential mechanism of FUDR action . FUDR may act as metabolic stressor especially that high FUDR concentrations above 100 uM seem to be required to observe the effect on lifespan in flcn-1 ( ok975 ) animals ( Figure S9 ) . Interestingly , lower concentrations of FUDR ( 5–10 uM ) that are also sufficient to inhibit germ line proliferation had no effect on lifespan . Here we show that the enhanced resistance to oxidative stress in the absence of FLCN-1 does not result from a decrease in oxidative damage or an increased transcriptional upregulation of ROS-detoxifying enzymes [51] , [52] , [89] . Instead , we show that loss of FLCN-1 activates AAK-2 thereby inducing autophagy . Accordingly , downregulation of unc-51 , the ortholog of the autophagy kinase ULK1 , was shown to suppress the increased number of positive GFP::LGG-1 foci upon overexpression of the kinase domain of AAK-2 in C . elegans [46] . More evidence was gathered in mammalian systems to support the direct activation of autophagy by AMPK [46] , [47] , [48] . In addition , we show that autophagy is required for the increased ATP in flcn-1 ( ok975 ) animals and Flcn−/− MEFs suggesting that the chronic activation of autophagy in the absence of FLCN recycles building blocks to produce ATP promoting stress resistance . When energy levels drop in the cell , AMP or ADP bind to the γ regulatory subunit of AMPK and induce an allosteric conformational change [43] , [44] , which leads to AMPK activation through phosphorylation of Thr172 in the catalytic subunit via inhibition of dephosphorylation activities . It is striking that loss of FLCN induces AMPK and autophagy in flcn-1 ( ok975 ) mutant animals and Flcn−/− MEFs , which exhibit high energy levels . These observations suggest that FLCN might be involved in the control of the energy sensing ability of AMPK . The increased activation of AMPK despite high energy levels has been recently reported upon inhibition of two other inhibitors of AMPK activity [95] , [96] . The roles of AMPK and autophagy in cancer are puzzling [97] , [98] , [99] , [100] . Both AMPK activation and autophagy have been shown to acquire anti- and pro-tumor functions [97] , [98] , [99] , [100] . Our results imply that the AMPK-dependent activation of autophagy might be essential for FLCN-deficient tumor cells to acquire an energetic advantage and drive tumorigenesis . In analogy with our results , autophagy was recently shown to be required for tumor growth in many cancer models [101] , [102] , [103] . A similar role for VHL , another renal tumor suppressor , in the regulation of autophagic events in renal cell carcinomas has been recently described [104] . In fact , the inhibition of autophagy by MiR-204 suppressed the tumor growth in VHL-deficient cells [104] . Moreover , the LC3B/ATG5-dependent autophagy was shown to be required for the development of VHL-deficient renal cell carcinomas in nude mice [104] . We suggest that the AMPK-dependent autophagy activation upon loss of FLCN promotes the survival of transformed cells , which normally undergo severe metabolic stresses as caused by hypoxia and lack of blood vessels [105] . In agreement with our results , four groups have recently reported that AMPK activation drives tumorigenesis via metabolic stress adaptation of different tumors [99] , [106] , [107] , [108] . In fact , AMPK is the best-characterized target of the tumor suppressor LKB1 and it was shown recently that loss of LKB-1/AMPK had a positive effect on tumor initiation but a negative effect on tumor progression/dissemination [98] , [109] . The fact that FLCN negatively regulates AMPK strongly implicates that it exerts physiological functions other than being a tumor suppressor . In the past ten years , FLCN was reported to be involved in the regulation of apoptosis , rRNA synthesis , TGF-β signaling , B-cell and stem cell differentiation , ciliogenesis , mitochondrial biogenesis , TOR signaling , epithelial polarization , and cytokinesis without the elucidation of the molecular mechanism [24] , [25] , [26] , [31] , [33] , [35] , [39] , [110] , [111] , [112] , [113] . While this report was under review , two reports have shown that loss of folliculin leads to mTOR inhibition and that it is involved in nutrient sensing via acting as a GTPase activating enzyme for the RAG GTPases [114] , [115] . It is possible that FLCN acts in two complexes . It binds to RAGs under starvation conditions leading to mTOR inhibition , while in normal conditions FLCN would bind AMPK , inhibiting its activity . However , it is not clear how the reported inhibition of mTOR activity upon loss of FLCN could lead to tumorigenesis , since mTOR was shown to be hyper-activated in tumors of BHD patients and mice devoid of FLCN [25] . In conclusion , we used the model organism C . elegans to decipher a genetic pathway , which is regulated by FLCN . We show that FLCN negatively regulates the activity of AMPK , which leads to increased autophagy , energy and survival to metabolic stress . Moreover , we confirmed conservation of this pathway in mammalian cells and suggest that chronic AMPK activation upon loss of FLCN potentiates tumorigenesis via increased autophagy leading to metabolic stress resistance and inhibition of apoptosis , which are two hallmarks of cancer cells [116] .
The FLCN-1 nematode polyclonal antibody was generated in rabbits with a purified GST-FLCN-1 recombinant protein by the McGill animal resource center services . Commercial antibodies and reagents used in this study are listed in Text S1 . C . elegans strains were obtained from the Caenorhabditis Genetics Center ( see Table S6 ) . The flcn-1 ( ok975 ) strain was outcrossed eight times to wild-type . Nematodes were maintained and synchronized using standard culture methods [117] . The RNAi feeding experiments were performed as described in [118] , and bacteria transformed with empty vector were used as control . For all RNAi experiments , phenotypes were scored with the F1 generation except for aak-2 knockdown ( F2 ) . Lifespan assays were performed according to standard protocols [119] . Expression constructs were generated using the pPD95 . 77 vector . pRF4 rol-6 ( su1006 ) was used as a co-injector marker . Transgenic lines were generated by microinjection into the gonad of adult hermaphrodite using standard techniques . 2 . 8 kb endogenous promoter of flcn-1 was generated by PCR from wild-type genomic DNA ( Forward primer 5′AAAACTGCAGCGTCTTCTCGTTTCACAGTAGTCA-3′ and reverse primer: 5′GCTCTAGATTGAATTCTGTAAAAACATGAATTTGA-3′ ) and cloned into pPD 95 . 77 at PstI and XbaI sites . flcn-1 cDNA was obtained from an RT–PCR reaction performed on wild-type animals RNA extracts using the following: forward primer 5′GCTCTAGAATGCAAGCAGTAATAGCACTTTGT-3′ and Reverse primer 5′CGGGATCCACGAGCAGTAGAGGTTTGAGACTG-3′ . flcn-1 cDNA was subsequently cloned into pPD 95 . 77 ( GFP expression plasmid-with flcn-1 endogenous promoter region ) at XbaI and BamHI sites . Primary MEFs were isolated from C57BL/6 E12 . 5 Flcn floxed mice . Ampk+/+ and Ampk−/− MEFs were described in [93] . Cell lines were maintained in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) , 100 U/ml penicillin and 100 µg/ml streptomycin ( Invitrogen ) . For details on shRNA procedures , transfections , FLCN deletions and immortalization see supporting information . Resistance to acute oxidative stress ( 100 mM PQ and H2O2 ) was determined as described in [52] , [54] . Chronic oxidative stress was assessed on thirty post-fertile animals using 4 mM PQ and survival was measured daily . For heat stress , one-day adult animals were transferred to NGM plates and exposed to 35°C . Survival was measured at indicated time points until all animals died . Concerning anoxia stress , one-day old adult animals were transferred to NGM plates and left in a Bio-Bag Environmental Chamber Type A ( Becton Dickinson Microbiology Systems ) for 26 hours at 20°C . Recovery rates were scored at indicated time points . For MEFs , cells were seeded ( 2×104 cells ) in 12-well plates and FBS-free media was added 24 hours after plating . Cell numbers were counted daily and survival rates were determined as the percent cell number compared to day 0 . For C . elegans , autophagy levels were assessed in hypodermal seam cells of L3 animals using the GFP::LGG-1 transgenic reporter strain DA2123 ( See Table S6 ) . For MEFs , wild-type and Flcn−/− cells were infected with the pMigR-1-LC3-GFP or the mcherry/GFP-LC3 constructs [82] , seeded on coverslip ( 50000 cells in 6 well-plate ) , serum starved for 12 hours and fixed with 4% paraformaldehyde . Autophagic-GFP positive puncta were quantified in at least 200 cells . Pictures from nematodes and MEFs were acquired with a Zeiss fluorescence confocal microscope . For C . elegans , synchronized young adults were collected and washed in M9 buffer . Pellets were treated with three freeze/thaw cycles and boiled for 15 min . ATP content in C . elegans was measured using an ATP determination kit ( Invitrogen ) and in MEFs using CellTiter-Glo Luminescent Cell Viability Assay ( Promega ) . For C . elegans , levels were normalized to protein levels and in MEFs normalized to cell number . Genomic DNAs from worm pellets were purified using Phenol/Chloroform extraction and treated with RNase A for 1 hour at 37°C . OxiSelect oxidative DNA damage ELISA assay was performed with 8 µg of DNA following manufacturer's instructions ( Cell Biolabs ) . Protein oxidative damage was assessed using Oxyblot Protein Oxidation Detection Kit ( Millipore ) . Synchronized young adult nematodes were harvested and total RNA was extracted with Trizol , purified using the RNeasy kit ( Qiagen ) . Quantitative real-time PCR ( qRT-PCR ) was performed using Express SYBR Green qPCR supermix ( Invitrogen ) and LightCycler480 system ( Roche ) . Catalase and SOD transcripts were normalized to housekeeping genes cdc42 , pmp-3 , and Y45F10D . 4 using Genorm [120] . AAK transcripts were normalized against cdc-42 . For primer sequences see Table S7 . Cells and synchronized young adult nematodes were washed with ice-cold PBS and M9 respectively and lysed in the AMPK lysis buffer [121] supplemented with the complete protease and phosphatase inhibitors ( Roche ) , 1 mM DTT , and benzamidine 1 µg/ml . Worm pellets were sonicated and cleared by centrifugation . Percent pAAK-2/pAMPK levels were quantified using ImageJ software and normalized for the AMPK levels . Synchronized young adult nematodes were treated with 50 mM PQ/M9 or M9 alone for 2 hours , washed and plated on NGM plates allowing 30 minutes recovery . TEM immersion fixation and embedding was performed according to [122] . See Text S1 . FTC cells were trypsinized , counted and resuspended in complete DMEM/F12 media . Two layered soft agar assay were undertaken in six well plates . The bottom layer contains 0 . 6% agar in complete DMEM/F12 media . The second layer encompasses 0 . 3% agar mixed with 0 . 5 million cells . Plates were cultured at 37°C in 5% CO2 . For worms , apoptotic germ cell corpses were visualized using Acridine Orange ( AO ) as described in [123] . Worms were incubated for 2 hours in M9 with or without 50 mM PQ in OP50 , supplemented with 2 µl/ml AO ( stock of 10 mg/ml ) . Worms were then washed and transferred into light-protected recovery NGM plates for 45 min before visualization . In MEFs , apoptosis levels were determined using the Annexin V: PE Apoptosis Detection Kit I ( BD Pharmingen ) according to manufacturer's protocol . Fluorescence intensity corresponding to apoptosis levels was detected using FACSCalibur flow cytometer ( excitation 488; emission 575/26; BD Biosciences ) . Targeted metabolite analysis was performed on an Agilent 6430 triple quadrupole mass spectrometer equipped with a 1290 Infinity UPLC system ( Agilent Technologies ) . Metabolites were separated using a 4 . 0 µm , 2 . 1×100 mm Cogent Diamond Hydride column ( MicroSolv Technology Corporation ) . Quantification was accomplished using MassHunter Quantitative Analysis software ( Agilent ) . See Text S1 for details . Data are expressed as means ±SEM . Statistical analyses for all data were performed by unpaired student's t-test , ANOVA , using Excel ( Microsoft , Albuquerque , NM , USA ) , SPSS ( IBM , Armonk , NY , USA ) and prism software ( GraphPad ) . For lifespan curve comparisons we used the Log-rank Mantel Cox test using GraphPad from Prism Statistical significance is indicated in figures ( * P<0 . 05 , **P<0 . 01 , ***P<0 . 001 ) or included in the supplementary tables . | The FLCN gene is responsible for the hereditary human tumor disease called Birt-Hogg-Dube syndrome ( BHD ) . Patients that inherit an inactivating mutation in the FLCN gene develop lung collapse as well as tumors in the kidney , colon , and skin . It is not clear yet what the exact function of this protein is in the cell or an organism . In this study , we used a simple model organism ( the round worm C . elegans ) to study the function of FLCN . We found that it is involved in the regulation of energy metabolism in the cell . FLCN normally binds and blocks the action of another protein ( AMPK ) , which is involved in the maintenance of energy levels . When energy levels fall , AMPK is activated and drives a recycling pathway called autophagy , where cellular components are recycled producing energy . In the absence of FLCN in worms and mammalian cells , like in tumors of BHD patients , AMPK and autophagy are chronically activated leading to an increased energy level , which makes the cells/organism very resistant to many stresses that would normally kill them , which in the end could lead to progression of tumorigenesis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"genetics",
"of",
"disease",
"biochemistry",
"signal",
"transduction",
"cellular",
"stress",
"responses",
"animal",
"genetics",
"cancer",
"genetics",
"model",
"organisms",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"bioene... | 2014 | Folliculin Regulates Ampk-Dependent Autophagy and Metabolic Stress Survival |
Buruli ulcer ( BU ) , caused by Mycobacterium ulcerans , is a neglected tropical disease frequently leading to permanent disabilities . The ulcers are treated with rifampicin and streptomycin , wound care and , if necessary surgical intervention . Professionals have exclusively shaped the research agenda concerning management and control , while patients’ perspective on priorities and preferences have not explicitly been explored or addressed . To get insight into patient perception of the management and control of Buruli ulcer a mixed methods research design was applied with a questionnaire and focus group discussions among former BU patients . Data collection was obtained in collaboration with a local team of native speakers in Ghana . A questionnaire was completed by 60 former patients and four focus group discussions were conducted with eight participants per group . Former patients positively evaluated both the effectiveness of the treatment and the financial contribution received for the travel costs to the hospitals . Pain experienced during treatment procedures , in particular wound care and the streptomycin injections , and the side-effects of the treatment were negatively evaluated . Former patients considered the development of preventive measures and knowledge on the transmission as priorities . Additionally , former patients asked for improved accessibility of health services , counselling and economic support . These findings can be used to improve clinical management and to guide the international research agenda .
Buruli ulcer ( BU ) is a devastating skin and soft tissue infection prevalent in several tropical and subtropical endemic areas worldwide , with the highest prevalence in West Africa [1 , 2] . BU is caused by infection with Mycobacterium ulcerans ( M . ulcerans ) . It is the third most common mycobacterial disease in immunocompetent persons [3 , 4] . In early disease stages BU presents as a painless nodule , plaque or oedema . Without treatment these lesions go on to ulcerate presenting with characteristic undermined edges . The lesions are divided in three categories; Category I: a single lesion <5cm in diameter . Category II: a single lesion 5-15cm in diameter . Category III: single lesion >15cm in diameter , multiple lesions , lesions at critical sites such as eye , breasts , genitalia and osteomyelitis [5] . Drug treatment for BU consists of a combination of antibiotics given for an eight weeks period . The current World Health Organization ( WHO ) -recommended regimen is rifampicin ( 10 mg/kg once daily , oral tablet ) combined with streptomycin ( 15mg/kg once daily , intramuscular injection ) . Wound care is performed three times a week to daily depending on the severity of the wound . Almost 30% of patients indicate pain during the wound care procedure [6] . Physiotherapy is required to promote healing of the wounds and to prevent disabilities [7] . Over 90% of the early detected , limited cases ( <10cm diameter ) can be treated with antibiotics alone and have a good quality of life in long-term follow-up [8 , 9] . Health seeking behaviour among BU patients is complex [10–12]; If BU patients delay in seeking medical care , they run an increased chance to develop complications [13–15] . Patients who present with larger lesions are more at risk of developing contractures , functional limitations , and social participation restrictions [15 , 16] . Surgical treatment , including skin grafting may be necessary in patients with complicated lesions or without satisfying response to antimicrobial treatment . The pathological mechanisms , tissue necrosis , immune suppression as well as the painlessness of the ulcers , are mediated by the toxin mycolactone produced by M . ulcerans [2 , 17] . Although M . ulcerans is believed to be acquired from environmental sources in endemic areas , the exact mode of transmission is currently incompletely understood . Risk factors identified that increase susceptibility for BU include being aged below 15 years or over 49 years , poor hygiene of existing wounds , living close to stagnant water , insect bites , the water sources used and the activities near them [18 , 19] . The current research priorities as stated by WHO are the mode of transmission , development of methods for early diagnosis , drug treatment and new treatment modalities , vaccines , cultural and social-economic studies and lastly , incidence , prevalence and mapping of BU [20] . Patient’s perspectives on a disease may however lead to different priorities compared to researcher’s priorities [21–23] . This study questions former BU patients about their experiences and the congruent priorities for research and interventions to improve management of BU .
The study was conducted in two endemic rural areas in Ghana , near Agogo Presbyterian Hospital and Nkawie/Toase Governmental hospital in Agogo and Nkawie respectively . Both hospitals are important BU treatment centers in their areas . In 2015 , the hospital in Agogo treated 17% of the new BU patients registered nationally . A mixed method design was used consisting of questionnaires with open-ended and multiple choice questions and focus group discussions . Between June and November 2012 the questionnaires were administered . Thereafter between August and November 2014 focus group discussions were organized in order to further discuss the topics identified , combined with patients’ experiences , perceptions and suggestions for improvements and future research . Four focus group discussions were held; the number of focus group discussions chosen was based on reported evidence that 70–80% of relevant topics were identified after four discussions [24] . Former BU patients residing in the catchment area of Agogo and Nkawie were eligible for participation . Participants were at least 16 years old , healed from a small BU lesion ( Category I or II ) within a year after start of treatment . Community volunteers were instrumental in localising them on their last known address in hospital administration . The questionnaires were administered to 71 participants . These participants in the BURULICO trial between 2006 and 2009 could be retrieved for follow-up visits on long-term consequences as described in Klis et al [25] . The focus group discussions were conducted with eight randomly selected participants per group , selected by purposive sampling based on hospital records . The aim was to obtain a heterogeneous group concerning age , profession and education to encourage the discussion . One female group and one male group discussion were conducted in each area , so in total 16 females and 16 males participated in the discussion . Male and female groups were chosen due to existing gender inequalities in traditional Ghanaian communities [26 , 27] . Therefore it was expected that participants would be more open and feel more at ease to express their opinions and experiences in homogenous groups [28 , 29] . Due to a high rate of illiteracy among the participants , all questionnaires were administered orally in their native language . Three trained local hospital staff members administered the questionnaires in a quiet private place . The focus group discussions were performed by a local team organised according to the principles specified on organising focus group discussions in low-income countries [24] . Preliminary testing , training of the moderator and note taker , and a pilot study with hospital staff were completed before data collection . The participants were given a reimbursement for their transportation and participation when the groups were finished . The recordings of the focus group discussions were then translated and transcribed by the note taker , who supplemented the recordings with his notes made during the meetings . The translation and transcription were checked by a second member of the research team . Qualitative analysis was performed on the final English transcripts using open coding and axial coding . During the open coding , the English transcripts were read carefully and codes were assigned , which was done individually by two people ( AV and RW ) . The initial codebook was developed based on the open coding of two interviews . Thereafter , the two separate codebooks were discussed and combined in regular meetings . In case of disagreement other researchers ( YS and JDZ ) were consulted in order to reach consensus . Throughout data collection the codebook was adapted based on new codes that emerged from the data . After the open coding , the various codes were subjected to a process of axial coding in order to develop categories and identify connections . The study protocol was approved by the Committee on Human Research , Publication , and Ethics of the Kwame Nkrumah University of Science and Technology and the Komfo Anokye Teaching Hospital , Kumasi ( reference number CHRPE/RC/158/14 ) . Informed consent or assent for the questionnaires and all focus group discussions was explained and obtained and the forms were signed by signature or thumbprint by participants and by parents or legal representatives in participants <18 years .
All participants approached for the questionnaires were willing to participate . Of the 71 administered questionnaires , 11 were incomplete and 60 were included in analysis . The majority of these participants were females ( N = 42; 70% ) . The age of the participants ranged between 16 and 77 years , with a median of 23 years ( IQR ( 25–75% ) 18–31 ) . All these participants had lesions < 10 cm in diameter . One of the 33 persons approached for the focus group discussion refused due to work and time-related issues . The age distribution was from 18 till 73 . Only 3 former patients in this group had had category 3 lesions . The results of the questionnaire on the multiple choice questions exploring which aspects of suffering from BU was important , with a Likert scale reflecting the degree of their dissatisfaction , are shown in Fig 1 . 49 former patients ( 82% ) reported not to have disliked the treatment aspects they were interviewed about in the open questions , but in the multiple choice questions , only 4 ( 6 . 7% ) indicated no dislike for any of the aspects mentioned . More than 50% of the participants disliked the pain they experienced during treatment . Other striking results include that 60% of participants disliked the uncertainty about how they contracted the disease . For 22% of these participants , this was their only dislike . Furthermore , during the open questions , 27 participants ( 45% ) didn’t mention anything . 15 participants ( 25% ) found it bothersome that they ran into trouble with work or school . Being bothered by side-effects , pain , stigma and having to walk a long distance to receive treatment were also mentioned . Moreover , when asked what they disliked , 11 participants ( 18% ) mentioned injections , pills , dressings and punch biopsies . When asked what they liked , a third of the participants reported to appreciate the small incentive they received . The effectiveness of treatment was mentioned as well; 13 participants ( 22% ) enjoyed the way they were welcomed and treated in the hospital . Additionally , singular aspects of treatment were mentioned as well as their effectiveness . We did not explore why these things were considered positive .
The questionnaires helped to identify topics for the focus groups discussion but were administrated two years earlier than the focus groups discussions . During this period patients’ experiences and opinions potentially could have changed . As BU treatment and management did not change in these two years , we do not expect any impact on our results . Furthermore , only former patients with small lesions ( Category I or II ) were included in the study . Former patients or patients with larger lesions might have indicated other priorities . Not all participants had BU in the same period , it might introduce recall bias . However , this recall problem could not be detected during the study and it was not mentioned by participants . In the overall results there was a lack of argumentation and support of statements made by our study participants . This may be because participants are not aware of other possibilities to do certain treatment aspects differently . Additionally , Ghana is known as a hierarchical culture that is quite collectivist; ‘The doctor knows best’ [34] . This may have led participants to not critically appraise their doctors and the treatment they received . Another reason might be that the participants , though assured they should say anything they had on their mind , nonetheless did not feel free to express all of their thoughts . The participants of the focus group discussions easily reached agreement with limited discussions as a result , despite all focus groups consisting of a heterogeneous group and the facilitator stimulating discussion . Possible explanations can be that the former patients had similar experiences concerning treatment or that a negative group dynamic occurred . However , during the discussion all were encouraged equally to express their opinion . During the discussion equal opportunities for expression were monitored and all participants were involved in order to minimize effects of group dynamics . We therefore conclude that limitations due to negative group dynamics were minimal . Members of the local BU team were working on the study; this might have affected the participants’ sense of freedom of expression , though all members not vital for the discussion were asked to leave the area where the groups met . To promote participants to express their opinions a former BU patient was designated as moderator . By inviting a former BU patient , and not a staff member , we aimed that participants would feel more comfortable during the discussion , knowing that the moderator could identify with them , and encourage the discussion effectively .
The preferences and priorities of BU patients appeared generally in agreement with the international research agenda . Additional preventive measures , accessibility and pain were mentioned as important subjects . More counselling would benefit some patients , since the disease may have a huge impact on life , economically as well as socially . Community based rehabilitation was suggested , thus enlarging the public support for the disease as well as benefitting the patients directly . In addition , for clinical purposes , more explanation on the course of treatment and expected side effects appears necessary , since a lack of knowledge on these aspects was indicated . Doctors are limited in their time to explain issues extensively , but giving more elaborate information on the management of the disease and possible side-effects appears to be useful to enhance compliance to treatment . This information could aid in creating a patient based BU research agenda in which patient empowerment in neglected tropical diseases may improve healthcare . | Buruli ulcer ( BU ) is a skin and soft tissue infection possibly leading to deformities and long term consequences with severe impact on patients’ lives . It is one of the neglected tropical diseases . The current research agenda is created solely by health professionals , whereas patients might indicate different priorities . Therefore we conducted qualitative research on former patients’ experience with the BU management and on the priorities for the research agenda as envisaged by patients . We found that the preferences of former patients appeared generally in agreement with the international research agenda . Additionally , they indicated that additional preventive measures , better accessibility to care and pain management were important subjects for them . More counselling and social support in the form of community-based rehabilitation were expected to contribute in dealing with the impact of the disease and to enlarge a public platform for support . For clinical management we found that more information on the disease appears necessary , since former patients indicate they did not know much about the course of treatment and expected side-effects . Involving patients can improve BU management and its research agenda . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"health",
"services",
"research",
"disabilities",
"tropical",
"diseases",
"health",
"care",
"bacterial",
"diseases",
"research",
"design",
"signs",
"and",
"symptoms",
"pharmaceutics",... | 2016 | Former Buruli Ulcer Patients’ Experiences and Wishes May Serve as a Guide to Further Improve Buruli Ulcer Management |
Tsetse are vectors of pathogenic trypanosomes , agents of human and animal trypanosomiasis in Africa . Components of tsetse saliva ( sialome ) are introduced into the mammalian host bite site during the blood feeding process and are important for tsetse’s ability to feed efficiently , but can also influence disease transmission and serve as biomarkers for host exposure . We compared the sialome components from four tsetse species in two subgenera: subgenus Morsitans: Glossina morsitans morsitans ( Gmm ) and Glossina pallidipes ( Gpd ) , and subgenus Palpalis: Glossina palpalis gambiensis ( Gpg ) and Glossina fuscipes fuscipes ( Gff ) , and evaluated their immunogenicity and serological cross reactivity by an immunoblot approach utilizing antibodies from experimental mice challenged with uninfected flies . The protein and immune profiles of sialome components varied with fly species in the same subgenus displaying greater similarity and cross reactivity . Sera obtained from cattle from disease endemic areas of Africa displayed an immunogenicity profile reflective of tsetse species distribution . We analyzed the sialome fractions of Gmm by LC-MS/MS , and identified TAg5 , Tsal1/Tsal2 , and Sgp3 as major immunogenic proteins , and the 5'-nucleotidase family as well as four members of the Adenosine Deaminase Growth Factor ( ADGF ) family as the major non-immunogenic proteins . Within the ADGF family , we identified four closely related proteins ( TSGF-1 , TSGF-2 , ADGF-3 and ADGF-4 ) , all of which are expressed in tsetse salivary glands . We describe the tsetse species-specific expression profiles and genomic localization of these proteins . Using a passive-immunity approach , we evaluated the effects of rec-TSGF ( TSGF-1 and TSGF-2 ) polyclonal antibodies on tsetse fitness parameters . Limited exposure of tsetse to mice with circulating anti-TSGF antibodies resulted in a slight detriment to their blood feeding ability as reflected by compromised digestion , lower weight gain and less total lipid reserves although these results were not statistically significant . Long-term exposure studies of tsetse flies to antibodies corresponding to the ADGF family of proteins are warranted to evaluate the role of this conserved family in fly biology .
Tsetse flies are vectors of pathogenic trypanosomes , which cause Human African Trypanosomiasis ( HAT ) , also known as Sleeping Sickness . In west and central Africa , the parasite Trypanosoma brucei gambiense causes a chronic but nearly always fatal disease , while in east of the Rift valley , Trypanosoma brucei rhodesiense causes an acute disease that is rapidly fatal if untreated [1] . Devastating epidemics in the 20th century resulted in tens of thousands of deaths in sub-Saharan Africa [2] . WHO has recently reported that epidemics that devastated Africa since 1980s have come under control , with case numbers declining below 10 , 000 for the first time in 2009 [3] . Many HAT endemic countries , including Central African Republic , Chad , Congo , Côte d’Ivoire , Uganda and Sudan , with disease occurring in remote areas , have limited access to surveillance , treatment and control measures [4] . In countries such as Guinea , the first country affected by HAT epidemics in West Africa , surveillance activities were eliminated especially in the context of the EBOLA epidemic . In addition to HAT , nagana or Animal African Trypanosomosis ( AAT ) , caused by Trypanosoma brucei brucei and the related parasites , Trypanosoma congolense and Trypanosoma vivax , limits effective cattle rearing across ten million square kilometers of Africa [5] and has wide implications for land use , agricultural practices and nutrition [6] . Natural transmission of the parasite to the mammalian host requires the insect tsetse host ( genus Glossina ) . Based on geographic distribution , behavioral , molecular and morphological characteristics , the genus Glossina is split into three species complexes: subgenera Fusca , Morsitans and Palpalis [7] . The Palpalis group consists of the major HAT transmitting species associated with forest galleries and thickets along riverine ecosystems , including Glossina palpalis palpalis ( Gpp ) and Glossina palpalis gambiensis ( Gpg ) in west Africa; and Glossina fuscipes spp . in Democratic Republic of Congo , northern Angola , southern Congo , western Tanzania and Kenya , Uganda , Rwanda , Burundi , and southern Sudan . The Morsitans group consists of vectors of HAT and AAT in east and central Africa , including two closely related species , Glossina morsitans morsitans ( Gmm ) and Glossina pallidipes ( Gpd ) , both associated with savannah type ecosystems [8–10] . The role of saliva has been widely documented for successful blood feeding of insects . Analysis of the secreted salivary gland ( SG ) proteins present in the saliva ( termed sialome ) of different blood sucking insects have identified functionally conserved molecules that disarm host hemostasis and inflammatory/immune processes . Based on transcriptomic and proteomic approaches the major tsetse SG proteins are known to include the anticoagulant thrombin inhibitor ( tsetse thrombin inhibitor , TTI ) , two putative adenosine deaminases ( tsetse salivary growth factors 1 and 2; TSGF-1 and TSGF-2 ) , salivary apyrase ( 5’ nucleotidase-related SG protein 3; Sgp3 ) , antigen5-related allergen ( tsetse Antigen 5; TAg5 ) and two putative endonucleases ( tsetse SG proteins 1 and 2; Tsal1 and Tsal2 ) [11–18] . Because the often fast evolving insect saliva proteins can be species specific with unique immunogenic properties , the potential use of saliva antigens as biomarkers for host exposure to different tsetse species has been recently investigated [19–22] . In addition to enabling successful blood feeding , sialome proteins also influence pathogen transmission processes at the bite site . For infections caused by sand fly transmitted Leishmania spp . , fly saliva has been shown to increase lesion size and parasite burden , and enhance the infection rate [23–26] . In the case of Rhodnius prolixus , which transmit Trypanosoma cruzi , parasite infection is enhanced by immunosuppressant mechanisms of the reduviid bug saliva [27–29] . Tsetse saliva also facilitates T . brucei infection in mice , possibly resulting from reduced host inflammatory responses [30] . Given the critical role sialome proteins can play in the infection outcome , vaccinating the mammalian host against saliva proteins has been suggested as a means to reduce pathogen transmission , or host feeding ability [31 , 32] . Here we compared the major sialome proteins from four tsetse species that belong to two different subgenera: subgenus Palpalis ( Gpg and Gff ) , and subgenus Morsitans ( Gmm and Gpd ) . Gpg and Gff are among the most important human disease transmitting tsetse species , while Gmm and Gpd prefer non-human hosts . We characterized the immunogenic components of the sialome , and determined the serological cross-reactivity that major saliva proteins exhibit between the different species complexes . We focused on the abundant protein family TSGF and characterized the genomic aspects of this family in different tsetse species . Finally , we evaluated the potential use of TSGF proteins as mammalian vaccine antigens to reduce tsetse fitness through passive-immunity approach in experimental mice .
Mouse experiments were carried out in strict accordance with the recommendations of the Care and Use of Laboratory Animals of the National Institutes of Health . All of the animals were handled according to Yale University Institutional Animal Care and Use Committee ( IACUC ) approved Protocols 2011–07266 and 2014–07266 ( 2011–07266 renewed on June 27 , 2014 ) . The cattle sera from Uganda were collected by the National Livestock Resources Institute ( NaLIRRI ) Veterinary team . Prior to the collections , the protocols were developed by NALIRRI Institutional Animal Care and Use Committee and were submitted to and approved by the Uganda National Council for Science and Technology ( UNCST ) as specified in Reference Number HS 1061 , Dec . 1 , 2011 . The veterinary team obtained the required permission for obtaining the cattle sera from individual owners . Six week old male C57BL/6 mice were used for all experiments . In each case mice were individually housed . Gmm ( Westwood ) are maintained in the insectary at Yale University . Puparia from Gff , Gpg and Gpd were imported from the Institute of Zoology laboratory at Slovak Academy of Science according to USDA Research Permit 30355 to S . A . All flies were maintained at 25°C with 50–55% relative humidity , and received defibrinated bovine blood every 48 h using an artificial membrane system [33] . SG dissections and saliva collection were performed as previously described [34] with some modifications . Three days after receiving their last blood meal , flies were immobilized and SGs were microscopically dissected and pooled in ice cold sterile PBS ( phosphate-buffered saline , 137 mM NaCl , 2 . 7 mM KCl , 2 . 4 mM KH2PO4 , 10 mM Na2HPO4 , pH 7 . 4 ) . After 1 hour of incubation on ice , all samples were spun down at 2300 g for 10 min , and the supernatant was collected and identified as sialome . Six week old C57BL/6 mice were sedated and exposed to 7–10 tsetse bites 3 times/week for 3 weeks . Two weeks after the final tsetse exposure , blood was collected via cardiac puncture and allowed to clot at room temperature for 20 min , after which samples were centrifuged for 15 min at 600 g at room temperature . The serum fraction was removed and stored at -20°C for long-term storage or 4°C while in use . The sera from five mice exposed to the same tsetse species were combined for immunoblot analysis . The recombinant ( rec ) TSGF-2 protein expression and anti-sera has been described [13] . To generate additional recProteins , the coding sequences of tsal1 , and tsgf1 were amplified without the signal peptide region from Gmm SG cDNA . The primers used in the amplification process were for tsal1 ( Forward: 5’- CTATGAGCTCTCGTTAAAAATACCAGAGAG and Reverse: 5’- CTCAGCGGCCGCATTAAATTTTAACAAATTATTA ) ; and for tsgf1 ( Forward: 5’- GTACGGATCCGAAGTGAACAAAGCTTATC and Reverse: 5’- GTACCTCGAGTTTCTCCTTCTTTCAAG ) . PCR amplification products were cloned into the pET-28a vector ( Novagen ) , and transfected into Escherichia coli BL21 strain for expression . recTsal1 ( 43 kD ) , and recTSGF-1 ( 54 kD ) proteins were purified using the His bind purification kit ( Novagen , Cat # 70239–3 ) . Purified recTsal1 and recTSGF-1 were analyzed by SDS-PAGE and the protein bands were sliced and 500 μg protein was used with adjuvant to generate polyclonal sera in rabbits commercially ( Cocalico Biologicals , Inc ) . Same amount of total sialome proteins obtained from dissected SG ( or extracts from the same number of dissected salivary glands ) were analyzed by 12% SDS-PAGE under reducing conditions and either stained by coomasie blue , or transferred to nitrocellulose membranes ( BioRad , Cat # 162–0112 ) according to standard protocols [35] . Protein concentration was detected by Nanodrop 2000 Spectrophotometer ( Thermo Scientific , Cat# ND2000PR14 ) . For immunoblot analysis , the concentrations of the primary antibodies used were: 1:200 for anti-Gmm and anti-Gff saliva , 1:10 , 000 for anti-recTsal1; 1:20 , 000 for anti-recTSGF-1 and 1:5 , 000 for anti-recTSGF-2 . The secondary antibody goat anti-mouse IgG ( H+L ) -HRP conjugate ( BioRad , Cat # 170–6516 ) and goat anti-rabbit IgG ( H+L ) -HRP conjugate ( BioRad , Cat # 170–6515 ) were diluted 1:20 , 000 before use and SuperSignal west pico chemiluminescent substrate ( Thermo Scientific , Cat# 34080 ) was added for detection and visualized using Molecular Imager ChemidocTM XRS+ ( BioRad , Cat #170–8265 ) following the manufacturer’s instructions . In addition , saliva collected from different tsetse species were separated on 10% native PAGE analysis at 4°C with Tris buffer ( 25 mM Tris , 200mM glycine ) [17] using molecular weights ranging from 14 , 000–500 , 000 for non-denaturing PAGE analysis ( Sigma_Aldrich , Cat# MWND500 ) for size estimation . For immunoblot , the proteins were transferred to nitrocellulose membrane using transfer buffer without SDS and the membranes were blotted as described above . Sera were obtained from cattle maintained in the Kibuku and Manafwa districts in south-eastern Uganda . The location and age of the cattle from which sera were collected are listed in S1 Table . Given that individual animal responses and/or exposure to tsetse bites could vary widely in the natural state , we chose to combine sera from 5–6 cattle in the same age group for immunoblot analysis . Sera obtained from cattle used to maintain the Gff colony at The National Livestock Resources Research Institute ( NaLIRRI ) , Tororo were used as positive control for exposure . Negative control serum used was commercially obtained FBS ( Sigma , Cat # 12105 ) . For immunoblot analysis primary antibodies were diluted 1:2 , 000 and the secondary antibody goat anti-bovine IgG ( Thermo Scientific , Cat # PA1-28700 ) was diluted 1:20 , 000 before use . Gmm sialome components were separated on 12% SDS PAGE , stained by commassie blue and the visible abundant protein bands were sliced and separated into two fractions . The first fraction contained the protein bands identified as immunogenic based on Immunoblot analysis with anti-saliva antibodies generated in mice . The second fraction contained protein bands identified as non-immunogenic in the Immunoblot analysis . The protein components of both fractions were subjected to LC-MS/MS analysis at W . M . Keck Facility at Yale University . Briefly , peptides were separated on a Waters nanoACQUITY ( 75 μm x 250 mm eluted at 300nl/min ) with MS analysis on a LTQ Orbitrap mass spectrometer . Mascot distiller and the Mascot search algorithm were used for searching in the NCBI database . Confidence level was set to 95% within the MASCOT search engine for protein hits based on randomness . Genome data from Gmm , Gpd , Glossina austeni ( Gau ) , Gff and Glossina brevipalpis ( Gbr ) were obtained from Vectorbase ( https://www . vectorbase . org/ ) . For SG transcriptome , we used both EST and Illumina data [36 , 37] . Transcriptome denovo assembly and mapping were analyzed using CLC Genomics Workbench ( CLC bio , Cambridge , MA ) . Blast , genome annotation and sequence alignment were performed by CLC Main Workbench ( CLC bio , Cambridge , MA ) . The published Gmm TSGF sequences and Drosophila melanogaster ( Dm ) sequences were used to identify homologs in other Glossina species by Blast . The ADA motif associated with each homolog was verified by BlastP analysis . Signal peptides were predicted by SignalP ( http://www . cbs . dtu . dk/services/SignalP/ ) . Phylogenic trees were generated using CLC Main Workbench . Jukes Cantor method was used to measure the protein distance and Neighbor Joining method was used to generate the tree . Bootstrap analysis was performed by 1000 replicates . Four mice were injected intraperitoneally with a total of 400 μg purified IgG ( ImmunoPure ( A/G ) IgG Purification Kit , Pierce , Cat# 44902 ) , corresponding to 200 μg of anti-Gmm recTSGF-1 and recTSGF-2 IgG , respectively . Control mice ( n = 4 ) were similarly injected with 400 μg of purified pre-immune rabbit IgG . Rabbit IgG antibody titers were assessed from each mouse by ELISA 24 h , and 12 days after IgG transfer . Two 96 well plates were coated ( 18 h , 4°C ) with 10 μg/ml of purified recTSGF-1 or TSGF-2 in 50μl of coating buffer ( 0 . 05 M Na2CO3 , 0 . 05 M NaHCO3 , pH 9 . 6 ) , respectively . Plates were washed five times with 200 μl washing buffer ( PBS , pH 7 . 4 , containing 0 . 1% ( v/v ) Tween 20 ) and were then blocked with blocking buffer ( 5% milk in PBST ) for 2 h at room temperature . Plates were washed again and serum samples from mice were 1:1250 diluted and 50μl were added in duplicate wells for each dilution . Rabbit anti- recTSGF-1 or anti-recTSGF-2 antibodies were included as positive control , and normal mice sera were used as negative control . Samples and controls were incubated for 1 . 5 h at room temperature . For secondary antibody HRP-conjugated goat anti-rabbit IgG ( 1:5 , 000 ) in blocking buffer was added and plates were incubated for 1 h at room temperature . Following 5 washes , 50 μl chromogenic substrate ( TMB ) were added to each well , plates were incubated for 8 min , and the reaction was terminated with 50 μl H2SO4 stop solution and absorbance at 450 nm was measured with reduction at 630 nm using ELISA plate reader . Newly emerged virgin male and female flies ( n = 64 male and 64 female ) were separated into 16 individual cages with 8 female or male flies per cage . One cage of male flies and one cage of female flies were each randomly assigned to either a TSGF passively immunized mouse or a control mouse , such that the same 16 flies ( 8 male and 8 female ) were fed on only their assigned mouse for the duration of the experiment . Each cage of flies was weighed 24 h before they received their first blood meal and then again immediately after they fed on mice . Flies were exposed to the same mouse at 2 , 5 , 8 and 11 days post antibody transfer , and allowed to feed for 15 min following the IACUC approved protocols . On each blood meal , one cage of male and one cage of female flies were allowed to feed on one mouse . For detecting engorgement variation , the combined weight of flies ( n = 8 ) in each cage ( n = 16 ) was measured before and after each blood meal on mice to an accuracy of 0 . 1 mg . Total weight change was also measured by comparing the weight of flies 24 h before the first blood meal and 72 h after the 4th blood meal . The fly survival data was recorded every three days over the experimental period . At the end of the experiment , total lipid levels were determined using a vanillin assay as previously described [38] . Briefly , flies were collected 72 h after their 4th blood meal on mice , and allowed to dry at 0% RH ( relative humidity ) at 60°C . Individual flies were homogenized in 0 . 5 ml of chloroform:methanol ( 2:1 ) and 0 . 1ml of the supernatant was moved into a 5 ml glass tube and the solvent was fully evaporated at 90°C . 0 . 4 ml of concentrated sulfuric acid was added into the dried lipid and heated at 90°C for 30 min . 4 ml vanillin reagent was added to the acid/lipid mixture . Samples were measured spectrophotometrically at 525 nm , and total lipid content was calculated against a lipid standard ( canola oil ) [38 , 39] . Statistical analyses were performed using the Mann Whitney test in the GraphPad Prism 6 software package .
We analyzed the protein profiles of the sialomes as well as SG extracts from Gmm , Gpd , Gff and Gpg by SDS-PAGE analysis ( Fig 1 and S1 Fig ) . Based on staining intensity , sialomes contain several abundant protein fractions , and based on banding profiles , the sialomes of the Morsitans group species ( Gmm and Gpd ) are more similar to one another than to the species in the Palpalis group ( Gpg and Gff ) . Comparison of the sialome components of the Morsitans group species shows three protein fractions of similar sizes ( Fig 1 , Lane 1 labeled 5–7 in Gmm , and Lane 2 labeled 3–5 in Gpd , respectively ) . In addition , two high molecular weight proteins of about 150 kD in size in Gmm ( Lane 1 , bands 1 and 2 ) can be reproducibly detected , while Gpd has a single protein band around this size ( Lane 2 , band 1 ) and Gmm has a protein ( band 3 ) that runs slightly higher in size than in Gpd ( band 2 ) . Gmm sialome also has an approximately 60 kD protein fraction ( labelled 4 ) , which is reproducibly absent from Gpd . The overall profiles of the Gff and Gpg sialome proteins are also similar to one other , but differ from the Morsitans group both in size and relative abundance ( Lanes 3 and 4 ) . There are several protein bands in Gff and Gpg sialomes of around 200 kD in size ( Lanes 3 and 4 , band 1 , respectively ) in addition to five major highly reproducible protein fractions . The most abundant protein fraction in the Morsitans group is about 40 kD in size ( Lanes 1 and 2 , bands 6 and 4 , respectively ) , while the most abundant sialome protein of Palpalis species is about 60 kD in size ( band 3 in Lanes 3 and 4 ) . Our results suggest the sialome protein profile is comparable between the species of the two groups , Morsitans and Palpalis . To determine the antigenic potential of the sialome proteins , we performed immunoblotting with anti-Gmm and anti-Gff saliva antibodies generated in mice , respectively ( Fig 2 ) . Gmm anti-saliva antibodies consistently recognized four major protein fractions in Gmm ( Fig 2A , Lane 1 ) : two proteins of about 150 kD ( bands 1 and 2 ) , 40 kD ( band 6 ) , and 25 kD ( band 7 ) in size . In contrast , the same Gmm anti-saliva antibodies recognized only one protein of about 150 kD in Gpd sialome ( Fig 2A , Lane 2 , band 1 ) . Anti-Gmm saliva antibodies did not show serological cross reactivity with either Gff or Gpg sialomes ( Lanes 3 and 4 , respectively ) . We next used anti-Gff saliva antibodies to analyze the Gmm and Gpd sialome preparations using the same immunoblotting approach . We detected reliably only two large proteins ( >250 KB in size ) from Gmm and Gpd sialomes ( Fig 2B , Lanes 1 and 2 , respectively ) , although very weak hybridizing bands of around 55 KD and two bands of around 150K were also detected . By contrast , anti-Gff saliva antibodies detected multiple proteins from the Gff sialome ( Fig 2B , Lane 3 ) ; two high molecular weight protein fractions ( band 1 and 2 ) , and three proteins ranging in size from 40–55 kD ( bands 3–5 in Lane 3 ) , and generated a similar profile with Gpg sialome , except that band 4 was not detectable and the intensity of band 5 was less pronounced ( Lane 4 ) . The immunoblots were repeated using different sialome preparations and with anti-saliva antibodies generated in different mice . These results collectively showed the Gmm , Gpd and Gpg species-specific profiles to be highly reproducible ( S2 Fig ) . Native PAGE analysis under non-denaturing conditions and subsequent immunoblot analysis were also performed ( S3 Fig ) . Similar to the results obtained under denaturing conditions , anti-Gmm saliva antibodies detected one major protein band in Gmm and Gpd sialomes analyzed under non-denaturing conditions , while no cross-reactivity was noted with Gpg sialome ( S3B Fig ) . The anti-Gff saliva antibodies detected one strong and several less pronounced bands in the Gpg , and only a single weak band in the Gpd sialome , but no cross-reactivity was noted with Gmm sialome ( S3C Fig ) . To identify the immunogenic and non-immunogenic components of the sialome , we combined the Gmm sialome proteins into two fractions as “immunogenic” and “non-immunogenic” based on our immunoblot analysis ( Fig 3 , bands 1 , 4 and 5 versus bands 2 and 3 , respectively ) . We subjected the two fractions to LC-MS/MS analysis and used the Gmm transcriptome database to predict the putative peptides present in each fraction [36] . This analysis identified the immunogenic proteins as TAg5 , Tsal1/Tsal2 , and Sgp3 , while the non-immunogenic fraction included the 5'-nucleotidase family and four members of the Adenosine Deaminase Growth Factor ( ADGF ) family ( Fig 3 , S1 Table ) . We had previously characterized two members of the ADGF family , TSGF-1 and TSGF-2 [13] , as abundant proteins in the salivary glands . The other two members of the ADGF family , Salivary Secreted Adenosine ( Genbank Number ADD20094 ) and Adenosine Deaminase-related-Growth Factor C ( Genbank Number ADD20092 ) , were previously annotated in the Gmm database as ADGF-3 and ADGF-4 , respectively [40] . We generated polyclonal rabbit antibodies against the recombinant ( rec ) Gmm Tsal1 ( 40 kD ) , TSGF-1 ( 55 kD ) and TSGF-2 ( 60 kD ) and used these antibodies for immunoblot analyses of the same amount of sialome extracts obtained from different tsetse species ( Fig 4 ) . Unlike anti-saliva antibodies generated through the natural fly bite , antibodies against Gmm recTsal1 and Gmm recTSGF-1 recognized the corresponding proteins from the sialomes of both Morsitans and Palpalis group flies ( Fig 4A and 4B ) . In contrast , recTSGF-2 antibodies detected the corresponding 60KD protein from Gmm and Gpg , but not from Gpd ( denoted by * in Fig 4C ) . Thus , although the Gmm TSGF-1 and TSGF-2 proteins are non-immunogenic when introduced to mice through the natural feeding route , the corresponding rec-proteins in association with adjuvant appear to be immunogenic in rabbits . The recTSGF-2 antibodies recognized the same protein in Gpg sialome ( Fig 4D lane 3 ) that showed the strongest signal with anti-Gff saliva antibodies ( Fig 4D lane 4 ) suggesting that the most immunogenic fraction in the Palpalis group sialome corresponds to TSGF-2 . To understand anti-saliva specific antibody responses in animals living under natural fly challenge in endemic areas , we obtained sera from young ( less than 8 month old ) and old cattle ( aged 8–15 years ) from the Kibuku and Manafwa areas of Uganda ( S2 Table ) . In Kibuku , Gff is the predominant tsetse species , while in Manafwa Gpd is more abundant ( Personal communication with Dr . Loyce Okedi , NaLIRRI ) . We also obtained positive control sera from cattle that were used to maintain the Gff colony in the insectary in Uganda , and commercial FBS was used as negative control . Immunoblotting showed that older cattle in both Kibuku and Manafwa regions contained antibodies that recognized only a few species-specific sialome proteins ( Fig 5 ) . The signal we detected from young cattle was weaker than that obtained from older cattle while no signal was detected with the negative control FBS ( S4 Fig ) . Sera obtained from animals in either region identified two proteins of 25 KD and 40 KD in Gmm and Gpd sialomes , which based on size correspond to Antigen 5 and Tsal , respectively . The blotting signal of the 25 KD protein from Gmm or Gpd saliva was much weaker compared to the signal detected with the 40 KD band . Sera obtained from animals in Kibuku district reacted more strongly with the Gpg sialome , possibly reflecting the greater abundance of the Palpalis group flies in Kibuku . Sera obtained from cattle in Kibuku as well as Gff exposed cattle detected up to seven protein fractions in the Gpg sialome ( Fig 5A and 5C , respectively ) , the strongest signal corresponding to TSGF-2 based on size ( 60 KD ) . In addition to the seven distinct bands , the blots showed a smear around 30KD ( Fig 5A and S3A Fig ) , which was less pronounced with positive control sera ( Fig 5C ) . Such a smear was also noted in a previous study where sera obtained from humans in Gff epidemic area were used to detect responses to Gpg and Gff sialomes [41] . The smear may result from degradation products or a saliva component unique to Palpalis group flies , such as the bacterial symbionts associated with SG tissue as also suggested by the previous study [41] . Thus , cattle living under constant tsetse challenge exhibited only a low response to a small fraction of the sialome components similar to the laboratory experimental system in mice . Furthermore , field studies further support that the anti-sialome responses in naturally exposed animals reflect a species complex-specific profile , and titers are related to the intensity and length of fly challenge as was also noted before [19] . We focused on the ADGF protein family because based on our immunoblotting data , this family of proteins do not appear to elicit a strong immune response despite being highly abundant in the Gmm sialome ( Fig 3 ) . Using TSGF-1 and TSGF-2 from Gmm and ADGF A-E from Dm , we searched the Gmm transcriptome and genome databases for related proteins , and identified 7 putative members for the ADGF family . We refer to the members of the ADGF family in tsetse as TSGF-1 and TSGF-2 and ADGF 3–7 [40 , 42] . The genes encoding TSGF-1 and -2 , and ADGF 3–5 are located on the same genomic contig spanning over a 32 kb region , while ADGF 6–7 are organized on the other scaffold ( Fig 6A ) . With the exception of ADGF-5 , which is abound 6000 bps , all members of the ADGF family are around 1500–3000 bps in size . Based on their genomic structure , TSGF 1–2 and ADGF 3–5 have 6 or 7 exons , while ADGF-6 and -7 have 3 and 2 exons , respectively ( Fig 6A ) . The putative ADGF proteins are 485–544 aa in size , with the exception of ADGF-5 which encodes a shorter putative protein of 202 aa due to the presence of a premature stop codon . With the exception of ADGF-7 , all putative ADGF proteins contain a secretory signal peptide ( S3 Table ) . Nucleotide and deduced protein sequences were compared between the Gmm ADGF family members and Drosophila ADGFs ( Table 1 ) . Four members of the ADGF family , TSGF 1–2 , and ADGF 3–4 , showed the highest similarity , which ranged from 50–60% at the nucleotide level , and 40–50% at the protein level . We searched genome and transcriptome data available for multiple tsetse species in the Morsitans ( Gmm , Gpd and G . austeni ( Gau ) ) , Palpalis ( Gpg and Gff ) , and Fusca ( G . brevipalpis ( Gbr ) ) groups to compare the ADGF family in Glossina spp . ( Listed in S4 Table ) . Phylogenic analysis of ADGF family proteins from different Glossina species together with members of the Dm ADGF family are shown in Fig 6B . Overall , the relationships among the ADGF family members from different tsetse species reflected the species phylogeny determined by ITS-2 analysis [43] . The ADGF genes from members of the same subgenus were more closely related with each other than those from another subgenus . The ADGFs from Gbr were consistently the most divergent among the five tsetse species analyzed , confirming the earlier separation of the Fusca subgenus in Glossina evolution . Phylogenetic analysis confirmed the close relatedness of tsetse TSGF-1 , TSGF-2 , ADGF-3 and ADGF-4 with Dm ADGF-C . The analysis also showed that tsetse ADGF-5 is related to Dm ADGF-D , tsetse ADGF-6 to Dm ADGF A , while tsetse ADGF-7 did not have a Dm homolog . Based on the sequence and phylogenetic relatedness along with co-localization on the genome , genes encoding TSGF 1–2 and ADGF 3–4 may represent a recent expansion in the genus Glossina . Interestingly , neither Gpd nor Gau genomes , encode a TSGF-2 homolog based on genomic and transcriptomic analysis . Analysis of the transcriptome data from Gmm , Gpd , Gpg and Gff indicated expression of only TSGF 1–2 and ADGF 3–4 in the salivary gland tissue . Among these four ADGF subtypes , TSGF-1 is the highest expressed gene in Morsitans group species ( Gmm and Gpd ) , corresponding to 57% of total ADGF expression in Gmm , and to 95% in Gpd , respectively . In Gmm salivary glands , TSGF-2 and ADGF 3–4 also have significant expression ( 7 . 0–13 . 9 RPKM ) , while in Gpd , which lacks TSGF-2 genomic locus , there was little to no expression of ADGF 3–4 ( Fig 6C ) . In the Palpalis group ( Gpg and Gff ) salivary glands , the major expressed subtype was TSGF-2 , corresponding to 83% in Gff and 74% in Gpg of total ADGF expression , respectively ( Fig 6C ) . Despite sharing a common origin , these four genes display isotype specific expression profiles in the different tsetse species . Since our results indicated that Gmm TSGF proteins are largely invisible to the mice immune system when introduced upon tsetse bites , we reasoned that TSGF family may have some important function ( s ) during the blood feeding process . To test our hypothesis , we passively immunized mice with both the rec-TSGF-1 and rec-TSGF-2 antibodies , and included a control group that received pre-immune sera . We performed ELISA to ensure that antibody titers remained high in mice during the experimental period . We also confirmed that sera from mice that received the pre-immune sera did not cross react with rec-TSGF antigens ( S5 Fig ) . We allowed teneral virgin male and female Gmm flies to receive four blood meals on the passively immunized mice and at the conclusion of the experiment evaluated flies for feeding efficiency and fitness effects by measuring engorgement , mortality and total lipid levels ( Fig 7 ) . Gmm fed on immunized mice showed slightly lower , but not statistically significant engorgement , when measured after the second and fourth blood meals in comparison to the control groups , 12 . 6% versus 17 . 0% , respectively ( Fig 7A ) . The mortality rates between the experimental and control groups were similar during the experimental duration , 62 . 5% in flies that received the anti-TSGF antibodies versus 60 . 9% in the control group , respectively ( Fig 7B ) . Over the experimental period , flies maintained on control mice had slightly higher total weight change ( 9 . 9 mg ) than those maintained on passively immunized mice ( 8 . 1 mg ) , although this difference was not statistically significant ( P>0 . 1 ) ( Fig 7C ) . The total lipid levels in the control group ( 0 . 20 mg ) also showed a slightly higher but not significant ( P>0 . 1 ) change from experimental flies ( 0 . 15 mg ) ( Fig 7D ) . Thus , exposure of Gmm for two weeks to mice that had anti-TSGF-1 and -2 antibodies resulted in a slight detriment to their blood feeding ability as reflected by compromised digestion , lower weight gain and less nutritional resource availability although these results were not statistically significant .
Blood feeding insects transmit disease agents to humans and animals worldwide . For successful blood feeding , insect saliva contains a variety of molecules ( termed sialome ) with functions important for evading the hematological and immune system of the vertebrate host at the bite site [29] . For long-term success , it is important for the sialome components not to trigger strong host immune responses , which can otherwise interfere with the insects’ feeding ability . Our results from four different tsetse species that belong to two different species complexes of Glossina confirm that the major abundant sialome proteins do not induce high immunogenic responses in both laboratory mice and in cattle living under tsetse challenge in endemic areas . Our immunoblotting results show that the few sialome proteins with immunogenic potential when introduced through the natural feeding route show limited cross-reactivity between different fly species , particularly among those that belong to different tsetse subgenera . We evaluated the potential function ( s ) of one of the most abundant sialome proteins , TSGF-1 and TSGF-2 , which belong to ADGF family of proteins with adenosine deaminase activity . Our data suggest that limited exposure of flies to mice that have passively received anti-recTSGF 1–2 antibodies results in a negative trend on tsetse fitness parameters . Both the abundance and the specific members of the ADGF family of proteins expressed in the sialome of different tsetse species may influence tsetse’s long-term fitness , host preference and disease transmission characteristics . Studies on the proteins present in the sialomes of different blood feeding insects provide fundamental information for saliva research with practical applications for disease control [44 , 45] . Our analysis of the major sialome proteins from four tsetse species indicates that these proteins are fast evolving and that flies belonging to different species complexes display similar protein profiles and species-specific immunogenicity in experimental mice . In Gmm , Sgp3 , Tsal and Antigen 5 have higher immunogenicity than the other abundant proteins , such as 5’ nucleotidase and TSGF family ( Figs 2 and 3 ) . However in Gff , proteins that belong to the ADGF family display the highest immunogenicity potential . Immunogenicity of sialome proteins , and cross reactivity of saliva mediated immunological responses among different vector sub-species have been analyzed in several other blood-feeding insects [29 , 46–48] . In the mosquitoes Aedes communis , Aedes aegypti and Anopheles stephensi , most of the saliva antigens appear to be species-specific and only weak cross-reactivity is observed with heterologous immune sera [47] . The Leishmania spp . sand fly host species Phlebotomus papatasi , Phlebotomus sergenti , and Lutzomyia longipalpis also have unique saliva protein profiles , and sera from mice exposed to these three species specifically do not exhibit extensive cross reactivity [48] . Studies in different tsetse species similarly reported varying serological responses to salivary proteins [19 , 21 , 22 , 49] . Caljon et al . evaluated host antibody responses to the highly immunogenic family of the endonuclease-like Tsal proteins using mice previously exposed to multiple tsetse species and rec Tsal1 proteins . Based on the cross reactivity they observed , they suggest that detection of anti-rTsal1 IgGs could be a promising serological indicator of tsetse fly presence [49] . The anti-saliva antibodies we generated in this study represent a low fly challenge , and our immunoblotting analysis did not detect obvious cross reactivity for Tsal1 proteins between the species we analyzed in the different tsetse subgenera . This difference may reflect the varying sensitivity associated with the different methods we used for detecting the immunological responses , but also may be related to the varying intensity of fly challenge the experimental mice were subjected to . In the study by Somda et al . , the anti-saliva responses in experimental cattle were shown to depend on the number and frequency of fly challenge . They reported that G . m . submorsitans ( Gms ) saliva showed a broad cross reactivity with sera of cattle exposed to different tsetse fly species . But sera from cattle exposed to G . m . submorsitans ( Gms ) exhibited weaker cross reactivity to Gpg saliva than sera from animals exposed to Gpg bites [19] . The antisera we developed in the laboratory could differ from the sera generated in animals in Africa that are naturally bitten by tsetse flies . The composition of the saliva we used from the cultured tsetse fly lines might vary especially since the colony flies we tested are maintained on artificial feeding systems that do not rely on the anticoagulation functions of saliva . In fact , previous research on sand flies have found that laboratory fly saliva can induce better protection against Leishmania infection than saliva from wild-caught or recently colonized sand flies [29 , 50] . Moreover , the duration and frequency of the fly bites the animals in Africa are naturally exposed to also differ from the laboratory experimental conditions . Furthermore , the sera obtained from native animals would contain responses to other blood sucking insect bites the animals are exposed to . To investigate the immunological responses of animals living under natural tsetse challenge , we compared the cross-reactivity of sera collected from cattle in two districts of Uganda where tsetse species distribution varies . We used the closely related Gpg for our analysis with endemic cattle sera as we did not have access to Gff fly saliva during the course of our later studies due to colony collapse . We also used Gmm and Gpd for our analysis with natural sera . Despite these potential variations , generally , our analysis with endemic cattle sera confirmed our findings with experimental mice in that sialomes from different fly species generate varying immune signatures . The cattle sera from Kibuku district , where the Palpalis group species Gff is the dominant tsetse species , had stronger interaction with the closely related Gpg sialome than with either Gpd or Gmm . In contrast , cattle sera from Manafwa area , where Gpd is present , showed stronger interaction with Gmm and Gpd sialomes . This would be expected as Morsitans group flies have strong preference for cattle , thus it is likely that immune responses of the cattle would reflect this bias . Among the sialome proteins , Gpg TSGF-2 showed the strongest signal with sera from Kibuku area , while the same sera could not detect TSGF-2 from Gmm and Gpd by immunoblot analysis . Beyond the presence of the Palpalis group flies in Kibuku area , the strong response we detected with TSGF-2 may also reflect the varying abundance of this protein in different species sialomes . In concordance , the transcriptomics analysis indicates that TSGF-2 is expressed at low levels in Gmm , and is missing from Gpd genome while TSGF-2 is expressed at high levels in Gpg ( Fig 6C ) . Analysis of sera from cattle aged 8–15 years showed a stronger immunological response to specific sialome antigens than cattle less than 8 months old , confirming that anti-saliva antibody titers increase over time related to frequency of fly challenge . Nevertheless , the magnitude of these responses were quite restricted even in older cattle that likely received many tsetse bites , suggesting that cattle repeatedly exposed to bites may eventually gain tolerance to the bites of those species . In previous works , human and cattle in tsetse epidemic and free areas were also compared to illustrate that anti-saliva antibody titers vary by fly bites in rainy and dry seasons where tsetse densities fluctuate [19 , 51] . Immunogenic components of saliva have been exploited as biomarkers for exposure to different arthropod bites , including ticks , sandflies and mosquitoes [45 , 46 , 52 , 53] . The potential use of tsetse saliva as a biomarker of exposure has also been investigated using human and cattle sera [19–21] . The saliva antigens in Gpg were analyzed to detect human exposure to tsetse flies in West Africa [21] . Sera from humans in Uganda scored positive for saliva-specific IgGs by ELISA detection , and against recombinant Gmm Tsal proteins by immunoblotting [51] . Tsal protein has been reported in previous works to be a good antigen to detect human and animal exposure to tsetse bites [49 , 51] . In our study , analysis of the cattle sera from both tsetse epidemic areas commonly recognized Tsal proteins in Gmm , Gpd and Gpg , which also confirms previous findings . While Tsal protein responses might be good for evaluating the risk for general tsetse exposure , TSGF-2 would be particularly useful for detecting exposure to Palpalis group fly challenge based on our immunoblot analysis . In addition a TSGF-1 specific peptide corresponding to aa 18–43 has been proposed as a good biomarker of tsetse exposure in Gpg epidemic area [22] . The ADGF family in tsetse has seven members , but only TSGF 1–2 and ADGF 3–4 are preferentially expressed in salivary gland . These four related genes are co-localized in the Gmm genome , indicating a recent gene duplication event . All four genes are related to Dm ADGF-C , which is actually not a highly expressed subtype in Drosophila spp . [54] . In Drosophila spp . brain and salivary glands , the isotype ADGF-D is expressed , which is more closely related to tsetse ADGF-5 [54] . It is possible that the expansion of TSGF 1–2 and ADGF 3–4 may have evolved with the blood feeding diet of tsetse , suggesting that ADGFs may play important function ( s ) for tsetse’s blood feeding process . The transcriptome data indicated that the four members of the ADGF family genes are expressed at varying levels in the salivary glands of different tsetse species . In the Morsitans group , Gmm and Gpd , TSGF-1 is the highest expressed member of the family , while in the Palpalis group , Gff and Gpg , TSGF-2 , is the highest expressed member . The differences in the abundance of varying ADGF isotypes in the sialomes of different species may contribute to the varying saliva immunogenicity we detected but also may influence their ability to feed on specific hosts or reflect adaptation to their preferential hosts . While the Palpalis group flies feed preferentially on humans and are efficient vectors of human-infective trypanosomes , the Morsitans group tsetse prefer feeding on ungulates and are more efficient vectors of the animal disease causing trypanosomes [13 , 55] . Analysis of transcriptomes from normal and parasitized salivary glands of Gmm indicate that the expression of all four ADGF genes are significantly down regulated in infected flies [37] . It remains to be seen whether the variation in the abundance of the different ADGF isotypes we noted in the different species complexes may influence the varying vector competence noted in the different tsetse host species [56] . Adenosine deaminase ( ADA ) deficiency is lethal in Drosophila [42] . Studies have shown that in Drosophila both ADGF-C ( the most closely related protein family member in Dm to the four tsetse ADGF family proteins ) and ADGF-D are mitogenic in vitro , stimulating cell proliferation by depleting extracellular adenosine [54 , 57] . Our previous analysis had shown that Gmm saliva has highest ADA activity , with G . p . palpalis ( Gpp ) having much less , and Gau no ADA activity [13] . Gpp is closely related to Gpg and Gff in the Palpalis subgenus , while Gau is closely related to Gpd , and both Gau and Gpd have lost the TSGF-2 locus . It is possible that differences in ADA activity levels in the different tsetse species saliva may influence blood feeding processes and vector competence traits . Adenosine deaminase was found in several other blood feeding insects including sand fly and mosquitoes [13 , 58–61] . In mosquitoes , ADA activities were detected in the saliva of Culex pipiens quinquefasciatus ( vector of avian malaria and West Nile virus ) and A . aegypti ( vector of Dengue and Yellow Fever viruses ) , but not in the Anopheline mosquito Anopheles gambiae ( vector of human malaria ) [58] . In the sand fly , ADA activity has been detected in L . longipalpis and Phlebotomus duboscqi saliva , but not in P . papatasi , Phlebotomus argentipes , Phlebotomus perniciosus or Phlebotomus ariasi [61] . To understand the functional role of ADGF protein family in tsetse , we maintained flies on mice that were passively transferred anti-recTSGF-1 and recTSGF-2 IgGs . Although overall the blood intake , weight and total lipid levels in the anti-TSGF blood meal receiving group were lower than the control after four blood meals , these differences were not statistically significant between the two groups . In Caljon et al’s study , anti-saliva immunized mice also showed no negative effect on tsetse fly blood feeding efficiency and survival [51] . It is possible that the four blood meals the flies received during our study were not sufficient to cause a detrimental effect on major host fitness parameters . Furthermore the TSGF-1 and TSGF-2 polyclonal sera we used may not block ADGF 3 and -4 specified activities , which may compensate due to potential functional redundancies . Finally , it is possible that passive transfer of rabbit IgGs to mice may mount an anti-rabbit immunoglobulin response , resulting in a serum sickness reaction later in the process . Thus , formation of mouse anti-rabbit immunoglobulin immune complexes late in the experiment may have reduced the available amount of anti-tsetse protein antibodies . Thus , future studies would focus on experiments where anti-recTSGF mice IgGs may be tested using the same bioassay . Previous studies have reported that adenosine in saliva can help initiate perception of pain in vertebrates , and also have vasodilatory , anti-platelet aggregation and lymphocyte-immunosuppressive activities [62] while inosine can potently inhibit production of inflammatory cytokines [60] . Adenosine deaminase activity in insect saliva may play a role in regulating the concentration of saliva adenosine and inosine levels . It is possible that because the mice used in our study were anaesthetized while exposed to fly bites , the mice may not feel the itch or pain caused by saliva adenosine . So the effect of blocking ADA activity by antibodies may not cause host behaviors that might influence the amount of blood the flies are able to take up at each feeding . The expression levels of the different TSGF proteins in saliva of different tsetse species vary . Furthermore , TSGF proteins expressed in the different tsetse species vary in their immunogenicity as also indicated by studies reported by Dama E et al . , [21] . It remains to be seen whether these differences influence host feeding preferences of the different tsetse species . In summary , both the composition and the immunogenicity potential of the sialome proteins vary in the different tsetse species groups . In Gmm , the major immunogenic proteins are TAg5 , Tsal1/Tsal2 , and Sgp3 , while in Gpg the immunogenic proteins include the Adenosine Deaminase Growth Factor ( ADGF ) family . We show that of the seven members of this family , only 4 are expressed at varying levels in the salivary gland of different tsetse species . The different ADGF subtypes expressed in the different tsetse species may contribute to the varying levels of immunogenicity this family displays . The relationship of ADGF family proteins with host vector competence traits , as well as their potential role as bio-marker of exposure to the different tsetse species complexes merits further investigations . | Insect saliva contains many proteins that are injected into the mammalian host during the blood feeding process . Saliva proteins enhance the blood feeding ability of insects , but they can also induce mammalian immune responses that inhibit successful feeding , or modulate the bite site to benefit pathogen transmission . Here we studied saliva from four different tsetse species that belong to two distant species groups . We show that the saliva protein profiles of different species groups vary . Experimental mice subjected to fly bites display varying immunological responses against the abundant saliva proteins and the antigenicity of the shared saliva proteins in different tsetse species differs . We show that one member of the ADGF family with adenosine deaminase motifs , TSGF-2 , is non-immunogenic in Glossina morsitans in mice , while the same protein from Glossina fuscipes is highly immunogenic . Such species-specific immune responses could be exploited as biomarkers of host exposures in the field . We also show that short-term exposure of G . morsitans to mice passively immunized by anti-TSGF antibodies leads to slight but not statistically significant negative fitness effects . Thus , future investigations with non-antigenic saliva proteins are warranted as they can lead to novel mammalian vaccine targets to reduce tsetse populations in the field . | [
"Abstract",
"Introduction",
"Material",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Immunogenicity and Serological Cross-Reactivity of Saliva Proteins among Different Tsetse Species |
The modeling of large-scale communicable epidemics has greatly benefited in the last years from the increasing availability of highly detailed data . Particullarly , in order to achieve quantitative descriptions of the evolution of epidemics , contact networks and mixing patterns are key . These heterogeneous patterns depend on several factors such as location , socioeconomic conditions , time , and age . This last factor has been shown to encapsulate a large fraction of the observed inter-individual variation in contact patterns , an observation validated by different measurements of age-dependent contact matrices . Recently , several works have studied how to project those matrices to areas where empirical data are not available . However , the dependence of contact matrices on demographic structures and their time evolution has been largely neglected . In this work , we tackle the problem of how to transform an empirical contact matrix that has been obtained for a given demographic structure into a different contact matrix that is compatible with a different demography . The methodology discussed here allows to extrapolate a contact structure measured in a particular area to any other whose demographic structure is known , as well as to obtain the time evolution of contact matrices as a function of the demographic dynamics of the populations they refer to . To quantify the effect of considering time-dynamics of contact patterns on disease modeling , we implemented a Susceptible-Exposed-Infected-Recovered ( SEIR ) model on 16 different countries and regions and evaluated the impact of neglecting the temporal evolution of mixing patterns . Our results show that simulated disease incidence rates , both at the aggregated and age-specific levels , are significantly dependent on contact structures variation driven by demographic evolution . The present work opens the path to eliminate technical biases from model-based impact evaluations of future epidemic threats and warns against the use of contact matrices to model diseases without correcting for demographic evolution or geographic variations .
During recent years , models on disease transmission have improved in complexity and depth , integrating high-resolution data on demography , mobility and social behavior [1 , 2] . Specifically , the topology of social contacts plays a major role in state-of-the-art modeling [3–8] . The complete knowledge of the network of contacts through which an epidemic spreads is usually unreachable or impossible to implement , and for modeling purposes it is useful to remain at the coarse level of age-groups . Under this view , the population under study is divided into different groups according to its age distribution and different contact rates are assumed among these groups . Age-dependent contact patterns give powerful insights on the transmission of diseases where epidemiological risk is age-dependent , either as a result of behavioral or physiological factors . Relevant examples are influenza-like diseases [6–10] , pertussis [11] , tuberculosis [12 , 13] , and varicella [14] . Furthermore , they are instrumental for modeling and implementing more efficient interventions [15 , 16] . Given the utmost importance of contact heterogeneities , the study of age-dependent social mixing has become a priority in the field . In 2008 , Mossong et al . [17] published a seminal work with the measurements of age-dependent contact rates in eight European countries ( Belgium , Finland , Germany , Great Britain , Italy , Luxembourg , Netherlands and Poland ) via contact diaries . Due to the high cost of gathering empirical data on social contacts , Fumanelli et al . [18] proposed an alternative path consisting on building synthetic contact patterns via the modelling of virtual populations . Nevertheless , other authors have followed the original route opened by Mossong et al . , measuring empirically the age-dependent social contacts of other countries such as China [19] , France [20] , Japan [21] , Kenya [22] , Russia [23] , Uganda [24] or Zimbabwe [25] , as well as the Special Administrative Region of Hong Kong [26] , thus expanding significantly the available data on social mixing in the last few years . In these studies , participants are asked how many contacts they have during a day and with whom . This allows to obtain the ( average ) number of contacts that an individual of a particular age i has with individuals of age-group j . The resulting matrix is not symmetric due to the different number of individuals in each age-group . However , it is precisely the demographic structure that imposes constraints in the entries of this matrix , as reciprocity of contacts should be fulfilled at any time ( i . e . , the total number of contacts reported by age-group i with age-group j should be ideally equal in the opposite direction ) . Therefore , an empirical contact matrix , that has been measured on a specific population , should not be used directly , without further considerations , in another population with a different demographic structure . This issue has important consequences in the field of disease modeling . As contact matrices play a key role in disease forecast , it is essential to assure that the matrices implemented are adapted to the demographic structure of the population considered in order to avoid biased estimations . For some short-cycle diseases like influenza , the time scale of the epidemic is much shorter than the typical times needed for a demographic structure to evolve . That means that , typically , the demographic structure can be safely considered constant [10] , and the eventual evolution of the contact matrix can be neglected throughout the simulation of an outbreak . For these diseases , the problems might arise when modelers use contact matrices that are not up to date -for instance , one might wonder whether the patterns reported in [17] in 2008 can be used nowadays , a decade later , during which all the European countries analyzed in that study aged significantly . The same issue appears when a contact matrix measured in a given location ( e . g . , a specific country ) is directly used to simulate disease spreading in another region or country with a different population structure . The previous considerations are even more troublesome for the case of persistent diseases that need long-term simulations , for which the hypothesis of constant demographic structures does not hold anymore [12] . In those cases , contact matrices should continuously evolve during the simulation to reflect the effect that an evolving demography should exert on contact structures . Furthermore , it remains unknown to what extent the variations between contact matrices coming from different geographic settings are due to differences in the demographic structures , divergent cultural traits and/or methodological differences between studies . For instance , elderly people exhibit higher contact rates with children in African countries than in Europe [25] . This could be explained by the different demographic structures: one might expect to observe higher contact rates toward the younger age strata in Africa than in Europe because their populations have a higher density of young individuals . However , it is not clear yet whether the demographic structure is the only driver of geographical heterogeneity between empirical contact matrices . The problems that arise when exporting contact patterns across settings have been noticed in previous studies , specially in what concerns matrix reciprocity . Recently , in [27] , Prem et al . proposed a method to export European contact patterns to different settings around the world in a way that preserves reciprocity . Similarly , in other epidemiological studies , when implementing heterogenous contact patterns , modelers apply different corrections to solve the problem of non-reciprocity [7 , 8 , 11 , 28 , 29] . However , a general discussion on the side implications of these corrections and their range of applicability is still missing . The main focus of this work is to study how age contact matrices , originally obtained for a specific setting ( country and year ) , can be adapted to different demographic structures -i . e . , to another ( location and/or time ) setting . To this end , we first study the magnitude of the reciprocity error incurred when the adaptation of empirical social contacts to different age structures is ignored , thus justifying the need of studying possible projections that solve this problem . Next , we analyze different methods to perform these adaptations , highlighting the differences induced in the contact patterns by the use of these methods . We also compare empirical contact matrices of 16 countries and regions in different areas worldwide filtering the influence of the demographic structure . This allows us to isolate the differences between contact patterns that are caused by any other factors , such as socio-cultural traits or methodological aspects , from those caused by demographic variability across settings . Finally , we implement a Susceptible-Exposed-Infected-Recovered ( SEIR ) dynamics to study the differences in prospected incidences that arise when applying the methods analyzed to project social contact matrices .
For this work we have gathered 16 different contact matrices coming from several geographic settings: 8 from the POLYMOD project [17] ( Belgium , Finland , Germany , Great Britain , Italy , Luxembourg , Netherlands and Poland ) , China [19] , France [20] , Hong-Kong [26] , Japan [21] , Kenya [22] , Russia [23] , Uganda [24] and Zimbabwe [25] . There are some methodological differences between these studies , thus some pre-processing to homogenize the matrices is required . Specifically , we need to transform them to the same definition of contact matrix and adapt them to the same age-groups . Once this is done , we perform a reciprocity correction ( valid for the demographic structure corresponding to the country and year where the survey took place ) , and we normalize the matrices so that the mean connectivity is equal to one . Details can be found in the Supplementary Information . Data regarding the time evolution of demographic structures , either observed in the past or projected until 2050 , have been retrieved from the UN population division database [30] . The basic problem explored in this work is: how can we transform the ( empirical ) contact matrix Mi , j , that has been measured for a specific demographic structure Ni , into a different contact matrix M i , j ′ that is compatible with a different demographic structure N i ′ ? This could mean to adapt data obtained in one specific country to another different region that has a different demography . But the problem can appear even if we remain in the same geographical setting , as a contact matrix measured at a specific time τ , could not be valid for an arbitrary time t if the demographic structure of that population has changed . In the following sections , we formulate the problem of non-reciprocity and we present and discuss different methods of using contact matrices in an arbitrary demographic structure . Summing up , in this work we discuss up to four different methods in order to adapt contact patterns estimated in a given setting to a different one for which there is no available data . In Table 1 we provide a summary of the main properties of each method . The first of them , called M0 , consists of applying the original contact structures available on the setting to study with no correction . This , as previously discussed , leads to contact structures that violate the requirement of total contacts reciprocity . A second approach , called M1 , consists of a direct correction of the reciprocity bias , which suffers however from another conceptual issue , namely , it does not preserve intrinsic connectivity . This means that , under M1 , the number of contacts that an individual in age-group i has per unit time with individuals in another age-group j , will not be proportional to the density of available contactees in j when adapting the matrix across settings . Considering these conceptual limitations , these two elementary approaches should be avoided whenever demographic data is available , in favour of alternative methods such as M2 or M3 . As for M2 and M3 , the main difference between them involves the presence or absence of a global factor multiplying the entire contact matrix when comparing their outcomes on the same setting . While both methods similarly respect reciprocity and intrinsic connectivity requirements , overall connectivity is not preserved under M2 , but it is under M3 . Concerning their application to disease transmission modelling , the relevance of this difference depends on the modelling context . On the one hand , we have situations where an incipient epidemic phenomenon starts in a setting that is different -either in time or space- from the one where contact data is available , and its basic infectiousness has to be calibrated from its early stages using a transmission model . This usually happens with emergent diseases , yet uncharacterised , as well as with pathogens whose transmission dynamics is highly variable due to high mutation rates ( typically virus ) . In these contexts , modelers are forced to re-calibrate global infectiousness , among other key epidemiological parameters , for every outbreak . Also , if the typical duration of the outbreak is smaller than the time-scale during which demographic dynamics occurs ( e . g . from weeks to months ) , then contact structures can be safely considered invariant during the simulation of the event . In these contexts , using M2 or M3 leads to largely similar outbreak descriptions . The reason is that the independent calibration of the infectiousness at the beginning of the outbreak absorbs the changes in global connectivity that are the only difference between the contact matrices produced by M2 or M3 . This means that , under this scenario , the main difference between the methods will translate into the inference of arbitrarily different infectiousness parameters after model calibration to describe the same epidemic event . A paradigmatic example of this kind of situation is the modeling of seasonal influenza , that typically involves calibration of each year strains’ infectiousness at the early onset of the season outbreak . In other contexts , whenever real-time model calibration is not an option , or the epidemic simulations need to extend over time periods that are not short enough to exclude demographic dynamics ( e . g . from years to decades ) , the lack of control that M2 provides regarding overall connectivity makes more advisable the usage of M3 . One cardinal example for this kind of situation is the simulation of a persistent disease like tuberculosis , whose description requires models to run over decades [12] . However , the description of short-cycle diseases might require the usage of M3 instead of M2 too whenever calibration is not an option and the infectiousness of the pathogen is to be accepted from an a-priori source . Summing up , using each of the different methods here described can result into significantly different projected contact patterns and modelers should be aware of the implications that this has on disease modelling . To illustrate such implications , in the next section we explore the quantitative implications of using each of the methods discussed here , by comparing the contact-structures themselves and simulating epidemic phenomena where contacts are described according to each of them .
In order to study the error incurred when using M0 , we transform the contact matrices obtained from empirical studies in different geographic settings to new matrices that correspond to the same location but at different years ( that could be past records or future projections ) . As the population changes over time , the new matrices incorporate the population demographies of the same setting across time . We define the reciprocity error as the coefficient of variation of the number of contacts measured in both directions , which gives us a matrix that we will call non-reciprocity matrix ( NRi , j ) . It is an antisymmetric matrix , in which a positive value of the entry ( i , j ) means that there are more contacts from i to j than in the opposite direction , and viceversa . A value of 0 would mean that the contacts between i and j are well balanced . More details can be found in the Supplementary Information . In Fig 1 we represent the demographic structures of Poland ( panel A ) and Zimbabwe ( panel B ) for different years alongside the corresponding non-reciprocity matrices . In the case of European countries ( Poland in panel A as an example ) , demographic structures have suffered from an ageing process during the last decades , which is predicted to continue in the future . This ageing tends to provoke negative values under the diagonal for the matrices NRi , j when we assumed past demographic structures , while the opposite will occur in the future . The behaviour for African countries ( Zimbabwe in panel B ) is slightly different , as their demographies have been more stable for the last decades and only now they are beginning to age faster . In brief , when we use directly a contact pattern in a demographic structure that is younger than when it was measured , it will lead to an overestimation of the contact rate of ( and the force of infection corresponding to ) the youngest age-groups . The opposite will occur when we use contact patterns in an older population . In Fig 1C we represent the evolution of the proportion of non-reciprocal contacts for all 16 geographic settings studied here ( see Supplementary Information ) . This magnitude is equal to zero in the year when the contact matrix was measured , as we have applied a correction for the empirical matrices to fulfill reciprocity at the reference case . However , it dramatically increases as we move far from the year of the survey . In the examples shown here , only two years before/after the survey time , the fraction of non-reciprocal contacts already reaches 5% . Note that methods M1 , M2 and M3 are well balanced by construction , thus NRi , j = 0 for every ( i , j ) when using any of them . We next study the evolution of the ratio between the age-dependent contact rates and an homogeneous mixing scenario . This ratio gives us the matrix Γi , j , defined as the intrinsic connectivity in Eq 4 . The entries of Γi , j are bigger than 1 when the interactions between age-groups i and j surpasses what it is expected from the case of homogeneous mixing , and smaller than 1 in the opposite case . See the Supplementary Information for more details . In Fig 1D and 1E we show 4 snapshots of the ratio of the intrinsic connectivity and the original survey ( Γ i , j ′ / Γ i , j ) obtained using M1 for Poland and Zimbabwe respectively . Each panel corresponds to an adaptation of the contact matrix to the population demography of the countries 10 and 20 years before and after the survey ( i . e . , the 4 matrices correspond to t = τ − 20y , t = τ − 10y , t = τ + 10y and t = τ + 20y ) . We can see that , even if M1 corrects the appearance of non-reciprocity , this method changes the tendency of some age-groups to mix with respect to others . Specifically , we can see that M1 will over-represent contacts between young individuals ( and under-represent contacts between old individuals ) as the population gets older . Furthermore , the previous results are quantitatively important . For instance , if we were to use the contact matrices that we have from Poland ( measured in 2005 ) today ( 2018 ) , we would have that the ratio Γ i , j ′ / Γ i , j surpasses 1 . 5 for some specific age-group pairs , while it goes down to almost 0 . 5 in others , or , in other words , the usage of M1 , which does not take into account the changes in the fractions of individuals in each age-strata that occurred between 2005 and 2018 , causes a bias of more than 50% in the contact densities projected between certain age groups . Consequently we say that M1 does not preserve intrinsic connectivity . The density correction ( M2 ) avoids this problem , as it explicitly considers a fixed intrinsic connectivity matrix ( Γi , j as defined in the Methods section ) that is modified according to the density of each age-group ( see Eq 3 ) . In Fig 2A and 2B we represent the contact patterns obtained with M2 and M3 for Poland and Zimbabwe , respectively , in different years . We see how , specially in the case for Zimbabwe , as the population gets older , the values of the matrix below the diagonal ( contacts toward young individuals ) fade in favor of contacts toward older individuals as those age-groups gain more representation . As for the mean connectivity ( Fig 2C ) , considering the evolution of contact patterns in M2 or considering them constant ( M0 ) leads to the same qualitatively behaviour , although variances are higher with M2 . These trends are decreasing in Europe and increasing in Africa . M0 and M1 have the same mean connectivity , as M1 consists basically of a rewiring of those connections that exist in M0 in order to correct for reciprocity . M3 is a normalization of M2 so the connectivity is constant in this case . The intrinsic connectivity matrices Γi , j that we obtain for every country allow us to compare the contact patterns of different settings once the influence of demography has been accounted for , and removed . In Fig 3A we represent these matrices for the 16 geographic settings analyzed in this work . Just by visual inspection we can identify some distinctive features: European matrices are more assortative and present higher interaction intensities among young individuals than African ones . To formalize this observation , in Fig 3B , we place the different matrices in a two dimensional plot defined by the proportion of overall connectivity produced by young individuals and the assortativity coefficient ( see Supplementary Information for the definition of these quantities ) . African and European countries cluster around different values of these two magnitudes: specifically , in African countries we found less assortativity and the contacts are less dominated by young individuals than in the European countries . As for the Asia region we see that Japan and China have significantly higher assortativity and fraction of contacts among young individuals than either African or European countries . In turn , Hong Kong , with its particular geographic idiosyncrasy- a special administrative region , predominantly urban , with one of the highest population densities in the world- , presents an intrinsic connectivity matrix that is more similar to one from a European country than from China or Japan . Up to now , we have shown that there are several ways to deal with demographic change and evolving populations regarding the structure of the contact patterns for a given population . We next address how these different methods impact disease modeling . To this end , we implement a Short cycle SEIR model ( details can be found in the Supplementary Information ) to study a situation where a short-cycle , influenza-like pathogen appears in a given location in subsequent times . We consider two different modelling scenarios . In scenario 1 the pathogen infectiousness is independently calculated in each outbreak to ensure that all outbreaks have the same reproductive numbers independently of the eventual changes in contact matrices . By doing this , we aim at simulating a situation where a pathogen appears recurrently on a population , and its modelling relies on independent calibration of each outbreak . Then , in scenario 2 , we model a situation when independent outbreak recalibration is not possible ( or pertinent ) , and the infectiousness is assumed to be known ( and constant ) in all outbreaks . Under these hypothetical scenarios , we would like to know how different would be the predicted size of the epidemic as a result of considering different contact matrices coming from the different projection methods proposed in this work . In particular , scenario 1 is instrumental to distinguish the outcomes from models M0 and M1 from either M2 or M3 . However , in this case the infectiousness is recalibrated in each event to ensure that all outbreaks have the same reproductive numbers . As a consequence , since the contact matrices derived from M2 and M3 only differ by a global scaling factor , the recalibration procedure absorbs the differences between M2 and M3 , making them indistinguishable . In turn , scenario 2 simulates a situation where the election between M2 or M3 becomes of central relevance , since the basic reproductive number of outbreaks will now depend on the contacts produced by each method . These two scenarios are designed to recapitulate the two paradigmatic modeling situations discussed in the Methods overview section: the case where a short outbreak of a relatively unknown pathogen has to be modelled upon infectiousness calibration ( scenario 1: emergent pathogens , influenza , etc . ) versus the case where calibration is not an option , or model simulations extend in time ( scenario 2: persistent diseases and/or a-priori known pathogens ) . The results of this exercise are presented in Fig 4 ( scenario 1 ) and Fig 5 ( scenario 2 ) . Regarding scenario 1 , in Fig 4 , panel A we can see that , while methods M0 and M1 predict lower age-aggregated incidences in European countries in 2050 with respect to 2000 , M2 reduces these differences and the incidences are comparable for both years or even positive ( M3 is not included here , for it would produce exactly the same results of M2 ) . A different situation is observed in Africa , where M0 and M1 predict an increase in incidence in the future while using M2 would lead to a decrease , though differences remain small ( less than 5% of variation ) . In panel Fig 4B we represent , for two examples of Europe and Africa ( Poland in purple and Zimbabwe in orange ) , the temporal evolution of the incidence observed with the different methods . Furthermore , we represent the age-specific incidence for both countries in three different years: 2010 , 2030 and 2050 ( Panel Fig 4C ) . The age-distribution of the incidence evidences the differences in connectivity patterns between Poland and Zimbabwe . While the incidence in elderly people drops in Poland ( as the contact rates for those age-groups also drop ) , it remains high in Zimbabwe for the same age-groups . The different methods of implementing contact rates also affect the age-specific incidence . In panel Fig 4D we represent the relative variation in age-specific incidence obtained with methods M0 and M1 with respect to M2 for Poland and Zimbabwe . In Poland we see that M0 and M1 tend to underestimate the incidence specially among the elder age-groups . In Zimbabwe M0 tends to overestimate the incidence among young individuals , while with M1 we encounter both effects: and overestimation among the youngest and a underrepresentation among the eldest . The reshaping of the age-specific incidence between models is coherent with the changes in topology already studied . For the case of M0 , i . e . , maintaining the contact patterns constant in time , we have that in the future , as the demographic structure shifts to older populations , contacts toward children will be overrepresented and contacts toward adults will be underrepresented . At first order we can obviate the contacts that are far from the diagonal , and establish that M0 mainly underrepresents contacts between adults and overrepresents contacts between young individuals ( in the context of aging populations ) . Thus , we will obtain an underrepresentation of the incidence in adults , and the opposite in children . However , as the eldest age-groups increase their population in Europe , they dominate the dynamics and cause and underestimation of the global incidence that eventually affects all age-groups . In African countries , where the contact patterns are less assortative than European countries , this effect is smaller . Besides , as African populations are still young even in 2050 , the overestimation of young contacts dominates the dynamics , and the differences in incidence are mainly positive . The situation is similar for M1 . As represented in Fig 1D and 1E , for M1 we also have an underrepresentation of contacts between adults and an overestimation between young individuals , yielding to similar results to M0 . In scenario 1 , where the infectiousness β is recalibrated in each outbreak , the mean connectivity does not play a role in the size of the outbreak . Thus M2 and M3 lead to the same outbreaks’ description , with the exception of the inferred values of β needed to produce them , which would contribute , nonetheless , to different evaluations of the epidemiological risk . This dynamical equivalence emanates only from the assumption that reproductive numbers can be measured at the early stages of any of the epidemics being simulated in each year , which is a conservative -often optimistic- assumption . However , in the alternative scenario where no initial calibration is possible or prescribed , and constant infectiousness values are accepted through all possible times , the equivalence between M2 and M3 is broken ( scenario 2 , shown in Fig 5 ) . As discussed in the methods overview section , this is conceptually similar to the task of producing long term forecasts of persistent diseases [12] , based on epidemiological parameters calibrated on an initial time-window . As we show in Fig 5 , when we do not recalibrate the infectiousness , M2 and M3 show a very different behaviour . While M3 leads to an outbreak size that is essentially invariant in time -due to stochasticity- , the outcome predicted from M2 is highly variable . Specifically , we see how European countries produce outbreak sizes that decrease in time while the opposite occurs for African countries , which matches the evolution of the mean connectivity as shown in Fig 2C . Regarding the age distribution of the incidence under M3 ( Fig 5C ) , we see a similar pattern to the one seen in scenario 1 . The comparison of the age distributions from methods M2 and M3 ( Fig 5D ) shows that the differences between both methods , already discussed at the aggregated level , also occur in the same direction within all age groups . All together , these results illustrate how a poor adaptation of the contact patterns observed in the past in a given country to a later time point can translate into epidemiological forecasts that are highly biased . On the one hand , we have seen how the limitations of M0 and M1 at describing reciprocity and intrinsic connectivity patterns translate into inconsistent results . On the other hand , regarding the comparison between the two methods based on the density correction for available contactees -M2 and M3- , we have seen how the introduction of a normalization term in M3 aimed at preserving the overall connectivity is specially relevant in the cases where epidemiological parameters cannot be calibrated at the early stages of the epidemic phenomena to be modelled .
Summarizing , empirical contact patterns belong to a specific time and place . If we want to integrate the heterogeneity of social mixing into more realistic models , we need to address how ( and in what cases ) to export contact patterns from empirical studies to the populations we want to study . In this work , we have studied and quantified the significant bias incurred when a specific contact pattern is blindly extrapolated to the future ( or the past ) , even if we remained inside the same country where those contacts were measured . In fact , only a couple of years after the measurement of these contact patterns , the changes in the age structure of the population make them vary significantly . Thus , for any meaningful epidemic forecast based on a model containing age-mixing contact matrices , we would need to adapt them taking into account the evolution of the demographic structures . Moreover , as we have shown , even for cases that do not expand into long periods of time and a constant demography could be assumed , it is necessary to make an initial adaptation of whatever empirical contact structure we want to implement , into the specific demographic structure of our system . We have also seen how these relevant differences in the topology of contacts yield to significant consequences for the spreading of a disease . Applying different methods to deal with contact patterns leads to important differences not only in the global incidence for a SEIR model , but also on age-specific incidences . Having such an important impact for the spreading of a disease , the insights provided by this work should be taken into consideration by modelers and also by public health decision-makers . In a similar way , we have explored the differences between the contact patterns of different geographic settings . Thus , we have found the existence of some specific characteristics beyond the underlying demographic pyramids , which warns against exporting contact patterns across different geographic areas ( i . e . continents ) . Since there are different intrinsic connectivity patterns ( i . e . , once demography effects have been subtracted ) across countries , it is also likely that there exists a time-evolution of the intrinsic connectivity inside the same setting . Although it is impossible to predict how society will change in the future , we should always take this into account as a limitation in any forecast for which the heterogeneity in social mixing is a key element . Finally , we note that there are some limitations that could affect quantitatively the results shown in this work . First of all , we have derived the contact patterns of the different studies according to the demographic structures of the specific country for the year the survey took place . Thus , we are implicitly assuming that the setting where the different surveys were performed are comparable with the national data in terms of their demographic pyramids . In other words , we assume that the surveys are representative of the population at large . This is likely true for most of the geographic settings analyzed , but there are certain cases in which this might not be the case , either because of small study size or putatively biased recruitment of participants . Besides , as we have already discussed in the Methods section , the different granularity ( i . e . , definition of the age-groups ) used throughout the bibliography studied also imposes some limitations when comparing the data . It is also worth pointing out that , although in this work we have focused on age-structured systems ( which has had its relevance in recent history of epidemiology ) , the problem studied here can be extrapolated to other models that might categorize their individuals based on other different traits that determine their social behavior . The results reported here and their implications open several paths for future research . One is related to the social mixing patterns themselves . In order to predict the large-scale spreading of a disease , multiple scales need to be integrated and coupled together . This implies that when integrating different spatial scales , we need to deal with different contact matrices and local demographies . For instance , in developed countries , it is known that the structure of the population is not the same in the most central or most populated cities as compared to smaller ones or the countryside . Thus , nation-wide demographies and surveys to infer contact matrices might need to be disaggregated . What is the right spatial scale to measure both quantities is an interesting and unsolved question . In this sense , here we have limited our simulated disease scenario to the case of isolated populations , but it remains to be seen what are the effects over a meta-population framework , in which we have mobility between sub-populations of potentially very different demographic structures . We plan to explore these issues in the future . | Large scale epidemic outbreaks represent an ever increasing threat to humankind . In order to anticipate eventual pandemics , mathematical modeling should not only have the capacity to model in real time an ongoing disease , but also to predict the evolution of potential outbreaks in different locations and times . To this end , computational frameworks need to incorporate , among other ingredients , realistic contact patterns into the models . This not only implies anticipating the demographic structure of the populations under study , but also understanding how demographic evolution reshapes social mixing patterns along time . Here we present a mathematical framework to solve this problem and test our modeling approach on 16 different empirical contact matrices . We also evaluate the impact of an eventual future outbreak by simulating a SEIR scenario in the countries and regions analyzed . Our results show that using outdated or imported contact matrices that do not take into account demographic structure or its evolution can lead to largely misleading conclusions . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"people",
"and",
"places",
"poland",
"infectious",
"diseases",
"age",
"groups",
"medicine",
"and",
"health",
"sciences",
"zimbabwe",
"european",
"union",
"pathology",
"and",
"laboratory",
"medicine",
"infectious",
"disease",
"epidemiology",
"epidemiology",
"pathogens",
... | 2018 | Projecting social contact matrices to different demographic structures |
The lysin motif ( LysM ) containing proteins can bind chitin and are ubiquitous in various organisms including fungi . In plant pathogenic fungi , a few LysM proteins have been characterized as effectors to suppress chitin-induced immunity in plant hosts and therefore contribute to fungal virulence . The effector mechanism is still questioned in fungus-animal interactions . In this study , we found that LysM proteins are also present in animal pathogenic fungi and have evolved divergently . The genome of the insect pathogen Beauveria bassiana encodes 12 LysM proteins , and the genes were differentially transcribed by the fungus when grown in different conditions . Deletion of six genes that were expressed by the fungus growing in insects revealed that two , Blys2 and Blys5 , were required for full fungal virulence . Both proteins could bind chitin and Blys5 ( containing two LysM domains ) could additionally bind chitosan and cellulose . Truncation analysis of Blys2 ( containing five LysM domains ) indicated that the combination of LysM domains could determine protein-binding affinity and specificity for different carbohydrates . Relative to the wild-type strain , loss of Blys2 or Blys5 could impair fungal propagation in insect hemocoels and lead to the upregulation of antifungal gene in insects . Interestingly , the virulence defects of ΔBlys2 and ΔBlys5 could be fully restored by complementation with the Slp1 effector from the rice blast fungus Magnaporthe oryzae . In contrast to Slp1 and Blys2 , Blys5 could potentially protect fungal hyphae against chitinase hydrolysis . The results of this study not only advance the understanding of LysM protein evolution but also establish the effector mechanism of fungus-animal interactions .
Insect pathogenic fungi such as the ascomycete species Beauveria bassiana and Metarhizium robertsii have been developed as promising biocontrol agents [1 , 2] . Fungal species such as B . bassiana , M . robertsii and M . acridum have also been investigated as genetically tractable systems to unravel the mechanisms of fungus-insect interactions [3] . Similar to plant pathogenic fungi , various strategies ranging from cell wall remodeling to the secretion of immune suppressors have been employed by insect pathogens to evade insect immune responses [3 , 4] . For example , the coat protein Mcl1 can be highly expressed by M . robertsii during fungal growth in the insect hemocoel ( body cavity ) to camouflage fungal cells from hemocyte recognition and encapsulation [5] . The blastospores of B . bassiana isolated from insect hemolymph have also been found with shielded carbohydrate epitopes to counteract insect immune defenses [6 , 7] . It has also been reported that small molecules such as the cyclopeptide destruxins produced by Metarhizium species and the red pigment oosporein produced by B . bassiana can facilitate fungal infection of insect hosts by inhibiting host immunities [8 , 9] . Relative to the well-understood mechanisms of fungus-plant interactions [10] , the effector-mediated perturbation of host immunities has not been fully elucidated in animal pathogens including both insect and mammalian pathogenic fungi . The chitin in cell walls is a well-characterized pathogen-associated molecular pattern ( PAMP ) in plant pathogenic fungi that can trigger host immune responses [11] . The lysin motif ( LysM ) -containing receptor kinases have been investigated in plants for mediating recognition of PAMP chitin during microbe-plant interactions [12 , 13] . The diverse LysM proteins are also widely distributed in the fungal kingdom and have been characterized as chitin-binding effectors in plant pathogens to deregulate host immunity [14] . Of these , Ecp6 is the first characterized LysM effector that contributes to the virulence of the tomato leaf mold Cladosporum fulvum [15 , 16] . It was later found that LysM proteins can also function as virulence factors in other plant pathogens , such as Slp1 in the rice blast fungus Magnaporthe oryzae [17] , Mg1LysM and Mg3LysM in the wheat pathogen Mycosphaerella graminicola [18] , ChElp1 and ChElp2 in the plant anthracnose fungus Collectotrichum higginsianum [19] , and Vd2LysM in the soil borne wilt disease Verticillium dahliae [20] . The virulence effect of LysM proteins in animal pathogenic fungi is still unclear . Functional studies of Ecp6 and Slp1 revealed that these LysM effectors could sequester the chitin oligosaccharides released from the cell walls of invading fungal hyphae to prevent the activation of plant immunity and/or outcompete the host immune receptor for chitin binding [16 , 17 , 21] . The chitin in fungal cell walls can be targeted and degraded by plant chitinases [12] . LysM effectors , such as the Mg1LysM and Mg3LysM of M . graminicola , can protect fungal hyphae against the hydrolytic activity of plant-derived chitinases [22] . In contrast , both Slp1 and Ecp6 cannot protect fungal cells from chitinase hydrolysis [16 , 17] . LysM proteins are also widely distributed in animal fungal pathogens [23 , 24] , and chitin can trigger immune responses in both mammals and insects [25] . These data would suggest the existence of a LysM effector machinery in fungus-animal interactions , which , however , has been questioned due to the consideration of the non-intimate relationships between fungi and animals compared to that which exist for plasmalemma-enveloped fungus and plant cells [11 , 26] . Molecular evidence is still required to substantiate these arguments . Our previous genomic analysis of insect pathogenic fungi identified an array of plant pathogen-like effectors , including LysM proteins [24 , 27] . In this study , we characterized the function of LysM proteins encoded in the genome of the insect pathogenic fungus B . bassiana and found that divergent proteins can bind chitin polymers and fungal cell walls to deregulate insect immune defenses . Of particular interest , we found that the Slp1 effector from the rice blast fungus could restore the virulence defect of the gene deletion mutants against insect hosts .
We first performed bioinformatic analyses of fungal LysM proteins by including those from insect and mammalian pathogenic fungi . Consistent with previous observations [14 , 28] , we found that the number of LysM proteins varied highly among fungal species ( S1 and S2 Tables ) . For example , 12 LysM proteins are present in the genome of B . bassiana ( termed Blys1-Blys12 , Fig 1 ) whereas 13 in Metarhizium robertsii , eight are present in M . oryzae and 18 in C . higginsianum . Fungal LysM proteins are usually cysteine-rich and vary in length [14] . By plotting protein length versus the cysteine ratio for each protein , we found that the 282 examined fungal proteins could be divided into two groups: one group containing proteins of 90–850 aa and 2–7% cysteine residues ( g1 group ) , and the other group containing proteins > 850 aa and 2 . 5–4% cysteine ( g2 group ) ( S1 Fig ) . In the g2 group , the proteins from the plant pathogens are highly underrepresented ( 5/96 ) compared to those from the insect ( 41/133 ) and mammalian ( 10/53 ) pathogens . In addition , we found that most LysM proteins contain a signal peptide ( 78% of 133 examined proteins from insect pathogens ) ( S1 Table ) , which is considerably higher than the genome-wide average ( ca . 15% ) for secretable proteins [29] . As evident in 12 LysM proteins from B . bassiana ( Fig 1 ) , varied numbers ( 1–7 ) of LysM domains ( termed as LysMs for abbreviation ) are present in various fungal proteins , which may or may not contain the additional ChtBD1 type of chitin-binding domain , the Glyco_18 type of chitinase domain and/or the Hce2 effector domain ( Pfam: PF14856 ) ( S2 Fig ) . Phylogenetic analysis indicated that the examined proteins could be grouped into different lineages partially through association with protein structures . For example , most of the large LysM proteins that contain the ChtBD1 , Glyco_18 and/or Hce2 domains are clustered together ( Cluster I ) , as are the intracellular proteins with a single LysM domain ( Cluster II , including Blys4 ) ( S2 Fig ) . For Blys1-12 of B . bassiana , Blys10 and Blys11 , and Blys9 and Blys12 are close to each other , whereas the rest of the Blys proteins fall into different lineages . Individual LysM sequences were also retrieved from each protein for phylogenetic analysis , and the results indicated that more than five clustering patterns could be obtained ( S3 Fig ) . Even most of LysMs were clustered together independent of their sequential positions within the parental proteins ( C4 cluster ) , the LysM1 ( C2 cluster ) and LysM2 ( C3 cluster ) from those proteins containing additional ChtBD1 and Glyco_18 domains could be grouped into respective lineages ( S3 Fig ) , i . e . , the relationships with protein structures . The LysMs of the effectors Ecp6 , Slp1 and Mg3LysM are more similar to the counterparts from bacteria with zero or one cysteine residue [14] . Consistently , the LysM domains from these proteins were clustered together into a basal lineage ( S3 Fig ) . In addition , further analysis of LysM sequence consensus indicated that the LysMs from plant pathogens could be divided into two types , i . e . , the pattern similar to those from bacteria ( S4A and S4B Fig ) , and the pattern containing four cysteine residues ( S4C Fig ) . The LysMs from insect pathogens contain four cysteine residues ( S4D Fig ) , i . e . , the typical fungal-specific LysMs that can putatively form two disulfide bridges within each domain [14] . The divergently evolved LysM proteins suggest functional diversities of these proteins in fungal biology . As indicated above , 12 LysM proteins are encoded by B . bassiana , and these proteins vary in length and contain a different number of LysM and/or other domains ( Fig 1 ) . Except for Blys4 , all other proteins each contain a signal peptide . We performed gene expression analysis by growing the fungus in various conditions , including the in vivo infection stages within the insect hemocoels . The results indicated that these genes were differentially expressed by the fungus ( Fig 2A ) . For example , Blys7 and Blys8 were highly transcribed in the conidia . Relative to growth on solid medium , fewer genes were expressed by the fungus growing in an artificial liquid medium . In particular , six genes ( i . e . , Blys2 , Blys4 , Blys5 , Blys6 , Blys7 and Blys8 ) were differentially transcribed by the fungus during in vivo infection of insect hosts . The closely related Blys10 and Blys11 , and Blys9 and Blys12 remained largely silent in B . bassiana under the examined conditions . The LysM effectors expressed by plant pathogens during colonization of hosts are required for fungal virulence [11] . To examine the virulence contribution of LysM proteins in B . bassiana , we performed homologous recombination-mediated deletions of the six genes upregulated by the fungus growing in insects . Different null mutants were obtained and verified by reverse-transcription PCR ( RT-PCR ) ( S5A Fig ) . Deletion of these genes had no obvious negative effect on fungal growth on potato dextrose agar ( PDA ) and PDA amended with Calcofluor White . However , ΔBlys4 and ΔBlys5 became relatively tolerant against H2O2-induced oxidative stress when compared to the wild type ( WT ) ( S5B Fig ) . Both the WT and mutant cultures could not grow at 37°C . We conducted both injection and topical infection bioassays using the last instar larvae of the wax moth Galleria mellonella to compare the virulence difference between the WT and null mutants of B . bassiana . The estimation and statistical comparison of the median lethal time ( LT50 ) values indicated that the deletions of Blys2 and Blys5 , but not the other genes , significantly ( P < 0 . 01 ) impaired fungal virulence in both types of bioassays ( Table 1; S6 Fig ) . For example , during the injection assays , the LT50 values of ΔBlys2 ( 3 . 133 ± 0 . 095 d ) and ΔBlys5 ( 2 . 944 ± 0 . 108 d ) was significantly ( P < 0 . 01 ) extended compared to that of the WT ( 2 . 600 ± 0 . 097 d ) . Similar to the virulence contributions of the LysM effectors in plant pathogens [11 , 28] , both Blys2 and Blys5 are therefore required for the full virulence of B . bassiana to infect insect hosts . Unfortunately , the trials to delete both the Blys2 and Blys5 genes were not successful for reasons that remain unclear . Additional mutants were generated to overexpress Blys2 in the WT strain , whereby Blys2 expression was made under the control of the constitutive promoter of either the GpdA or laccase gene of B . bassiana [30] . We also tried to rescue the Blys2 and Blys5 deletion mutants using the M . oryzae Slp1 gene under the control of the GpdA promoter for transformation . Both Blys2 and Slp1 could be successfully transcribed by the obtained mutants ( Fig 2B ) . The injection and topical infection bioassays indicated that the LT50 values of Blys2 overexpression mutants had no obvious difference ( P > 0 . 1 ) from those of the WT . Interestingly , complementation of ΔBlys2 with Slp1 fully restored the virulence defect of ΔBlys2 ( Table 1; S7A and S7B Fig ) . Likewise , the statistical difference between WT and ΔBlys5::Slp1 was non-significant ( χ2 = 2 . 081; P = 0 . 149 ) ( S7C Fig ) . Nevertheless , the full sequence and two LysM domains of Slp1 exhibited no obvious conservation and phylogenetic relatedness with those of Blys2 and Blys5 ( S8 Fig ) . It has been found that the LysM domains determine the protein-binding affinity to chitin or specificity to other type of carbohydrates [14] . Blys2 contains five LysMs whereas Blys5 has two ( Fig 1 ) . To determine their binding abilities to different carbohydrates , the cDNAs of both genes were cloned into the vector for expression in Escherichia coli . In addition , different truncated forms of Blys2 with a reduced number of LysM domains were also expressed . The purified proteins were all soluble ( Fig 3A ) . The saturated affinity binding assays were performed using the chitin polymers extracted from the conidial cell wall of B . bassiana , chitin beads , chitosan and cellulose . The results indicated that , similar to Ecp6 [16] and Slp1 [17] , both Blys2 and Blys5 could bind fungal cell wall chitin and chitin beads . However , in contrast to Blys2 , Blys5 could also target chitosan and cellulose ( Fig 3B ) . The truncated forms of Blys2 , i . e . , Blys2D1-2 ( Blys2 only contains first two LysM domains ) , Blys2D1-4 , Blys2D2-5 , Blys2D3-5 and Blys2D4-5 , could also bind fungal chitin and chitin beads . Unlike Blys2D1-2 and Blys2D1-4 , the other forms also slightly bound chitosan and cellulose . In addition , relative to the full protein and other truncated forms , the binding ability of the forms Blys2D3-5 and Blys2D4-5 was reduced , as these truncated proteins were also detected in the supernatant samples ( Fig 3B ) . Different LysM proteins play a distinct role in protecting fungal cells from chitinase hydrolysis [17 , 22] . To determine the protection potential of Blys2 and Blys5 against chitinase , the WT spores were germinated in a liquid medium for 16 hrs . The germlings were incubated with each protein and then treated with chitinase cocktails to compare the difference in formation of protoplasts . The results indicated that a similar ratio ( P = 0 . 1618 ) of protoplasts was released from the Blys2-treated sample whereas a significantly fewer number of protoplasts was formed from the Blys5-incubated germlings ( P < 0 . 0001 ) when compared to the mock control ( the WT germlings without protein incubation ) . Consistent with the previous report that Slp1 cannot protect fungal hyphae from chitinase degradation [16] , non-significant difference ( P = 0 . 0919 ) was observed between the mock and WT::Slp1 samples ( Fig 3C ) . Thus , Blys5 but not Blys2 can protect fungal hyphae against the hydrolytic activity of chitinase . A recent report indicated that Blys2 was identified as a cell wall protein from both the blastospores and hyphal bodies of B . bassiana [7] . To determine the secretion and localization feature of Blys2 , a green fluorescent protein ( GFP ) was fused in frame at the C-terminus of Blys2 . The cleaved form of Blys2 without the signal peptide ( Blys2-SP ) was also generated , and both cassettes were placed under the control of the GpdA gene promoter for transformations of the WT strain of B . bassiana . Both types of proteins could be successfully expressed and detected in the mycelial-protein samples of B . bassiana by Western blot analysis using an antibody against the GFP protein . However , in contrast to the WT Blys2 , Blys2-SP could not be detected in the culture medium . The result confirmed that Blys2 is an extracellular protein ( Fig 2C ) . To examine the localization of Blys2 , fungal cells were stained with the fluorochrome Calcofluor White for chitin labeling . We found that the Blys2-GFP signal could be detected in the chitin-stained cell walls of various cell types ( labelled as WT::gp-Blys2-GFP ) , including conidial spores , hyphae , blastospores and hyphal bodies ( Fig 4A ) . This result was consistent with the observation that Blys2 can bind fungal cell wall chitin ( Fig 3B ) . By contrast , smeared GFP signals were observed in mutant cells ( labelled as WT::gp-Blys2-SP-GFP ) expressing Blys2-SP-GFP ( Fig 4B ) . This finding was similar to those obtained from the cells ( labelled as WT::gp-GFP ) that were only transformed with a GFP gene ( Fig 4C ) . In addition , consistent with the chitin-binding nature of Slp1 [16] , we found that GFP-fused Slp1 could also be detected in the mutant ( labelled as WT::gp-Slp1-GFP ) cell walls of B . bassiana ( Fig 4D ) . In comparison to other types of cells , chitin-staining , Blys2-GFP and Slp1-GFP signals were more weakly detected in hyphal bodies that were harvested from insect hemocoels . This result could be due to the occurrence of cell wall re-modification with reduced contents of chitin and β-glucans during fungal growth in insect body cavities [31] . Having established that Blys2 and Blys5 are required for full fungal virulence and can bind fungal cell wall chitin , we performed further experiments to investigate the mechanism of protein virulence contribution . Thus , the spores of the WT , ΔBlys2 , ΔBlys5 and Slp1-rescued mutants were injected into the last instar of wax moth larvae , and the insects were bled at various times to examine fungal developments . We found that the typical cellular immune responses , i . e . , hemocyte encapsulation and melanization [5] , similarly occurred in insects against both WT and mutant spores up to 24 hrs post injection ( Fig 5 ) . However , in contrast to ΔBlys2 , the WT , Blys2 overexpression ( i . e . , the mutants WT::gp-Blys2 and WT::lp-Blys2 ) and Slp1-rescued mutants ( ΔBlys2::Slp1 ) began to produce hyphal bodies 36 hrs post treatments . Subsequently , significantly fewer ( P < 0 . 0001 ) free-living cells were produced by ΔBlys2 when compared to the WT 48 hrs post injection . We also found that , relative to the WT , the propagation of ΔBlys5 was considerably ( P < 0 . 0001 ) impaired in insect hemocoels as well . However , no obvious differences ( P > 0 . 05 ) were observed between the WT and other mutants ( Fig 6A ) . In addition , a quantitative real-time PCR ( qRT-PCR ) analysis indicated that the expression of the antifungal gallerimycin gene could be more highly ( P < 0 . 05 ) induced in insects infected by ΔBlys2 than by the WT ( Fig 6B ) . In comparison to the WT , deletion of Blys5 could also lead to a higher level ( P < 0 . 05 ) of induction of the antifungal gene expression in insects ( Fig 6C ) . However , no significant difference was observed between the WT and other mutants including the Slp1-rescued mutants . Thus , deletion of either Blys2 or Blys5 could impair fungal propagations and ability in suppressing immune responses in insects , and the mutant defects could be complemented with the Magnaporthe Slp1 gene .
To perturb chitin detection by host cells , plant pathogens have evolved various strategies to deregulate host immune responses [10 , 11] . It has been posited that animal pathogens may not require any effector due to the non-intimate relationships between pathogens and animal cells [11 , 26] . In this study , we report that the genomes of animal pathogenic fungi encode different numbers of LysM domain-containing proteins . Our functional studies revealed that two of 12 LysM proteins , i . e . , Blys2 and Blys5 , are required for the full virulence of the insect pathogen B . bassiana . These proteins can be secreted and function as bona fide effectors by targeting fungal cell wall chitin to deregulate insect host immunities . Of particular interest , the gene-rescue experiment with the Slp1 effector from the rice blast fungus M . oryzae could restore the virulence defect of ΔBlys2 and ΔBlys5 against insect hosts . The results of this study confirm that animal pathogens can employ a similar strategy to that used by plant pathogenic fungi for the effector-mediated evasion of host immune defenses to facilitate fungal infection . LysM proteins are widely distributed in different organisms , from bacteria to fungi to plants [14] . Protein length diversity , domain number and sequence variations of LysM proteins are observed in different fungi , even between closely related fungal species . For example , 12 LysM proteins are present in the genome of B . bassiana but 14 in B . brongniartii . The inter- and intra-specific functional diversities of LysM proteins are still unknown . We found that the 12 Blys genes were differentially regulated by B . bassiana , and only the Blys2 and Blys5 genes that were activated in insect hemocoels were virulence factors . A recent proteomic investigation of the cell wall proteins of B . bassiana indicated that Blys8 could be identified from the hyphal bodies isolated from insect hemocoels [7] . We found that deletion of Blys8 did not impair fungal virulence . Considering that B . bassiana is also a plant endophyte [32] , other Blys proteins may be involved in fungal interactions with plants that remain to be determined . Variations in LysM protein numbers were also observed among different strains of the plant pathogen V . dahliae . Moreover , a lineage-specific LysM effector but not the core LysM proteins ( those present in all strains ) was found to contribute to the virulence of the strain to alternative plant hosts [20] . This finding suggests that the LysM effector may play a role in influencing fungal host ranges . For insect pathogenic Metarhizium species , 13 LysM proteins are present in the genome of the generalist pathogen M . robertsii , whereas only four are present in the specialist species M . album and M . acridum [27] . This variation further suggests that the LysM protein may be connected with fungal lifestyles , including host specificity . In addition to suppress chitin-triggered immunity , the LysM proteins ChElp1 and ChElp2 of C . higginsianum are also required for the appressorium-mediated penetration of plant cells [19] . Interestingly , the LysM protein Tal6 from the mycoparasite Trichoderma atroviride could specifically inhibit the germination of the spores of Trichoderma species but not other fungi [33] . Thus , the exact function ( s ) of LysM protein remains to be determined in a species-specific manner . Varied numbers of LysM domains ( from 1–7 ) are present in different proteins . Except for the lineage-specific clustering pattern of the LysM1 and LysM2 domains retrieved from those proteins containing additional ChtBD1 and Glyco_18 domains , most LysM domains are clustered independent of their positions within the parental proteins ( S3 Fig ) . Functional diversities of these variations are unclear . Structural analysis of the effector Ecp6 ( containing three LysM domains ) indicated that LysM1 and LysM3 can tightly form an intrachain dimer to mediate chitin binding with ultra-high affinity whereas the remaining LysM2 binds chitin with low affinity [21] . The effectors Slp1 , ChElp1 , ChElp2 and Vd2LysM each have two LysM domains , whereas Mg1LysM only contains a single LysM [17 , 19 , 20] . The formation of LysM dimers between the compartmented domains is therefore not applicable to these proteins that all can bind chitin but not chitosan , xylan and cellulose . We performed truncation studies of Blys2 and found that the five truncated forms retained the ability to bind chitin . However , the Blys2D2-5 form could additionally bind chitosan and cellulose with low affinity , whereas the losses of LysMs 1–2 ( i . e . , Blys2D3-5 ) or LysMs 1–3 ( i . e . , Blys2D4-5 ) reduced the mutant proteins’ chitin-binding affinity ( Fig 3B ) . These results would suggest that the LysM1 of Blys2 might affect the protein binding specificity whereas the LysM2 and LysM3 domains might determine the protein’s chitin-binding affinity . Unlike Blys2 , Blys5 with two LysM domains could additionally bind chitosan and cellulose . It has been reported that the Tal6 of Trichoderma , with seven LysM domains , could bind chitosan but not the chitin and cell walls of various fungal species . However , the truncated form of Tal6 that contains the last four LysM domains could bind colloidal chitin but not chitin powder and chitin flakes [33] . Thus , both the sequence and combination of LysM domains jointly determine the carbohydrate-binding specificity and/or affinity of LysM proteins . To suppress the chitin-induced immune responses in plants , the LysM effectors upregulated by fungal pathogens can function as a competitive inhibitor of plant chitin receptors , a scavenger of chitin oligomers and/or a protective coat to shield fungal cells from the hydrolytic activity of plant chitinases [11 , 12] . In contrast to the identification of LysM kinase receptors in plants [34] , chitin receptor has not been identified in insects . Nevertheless , chitin oligomers could activate the expression of antimicrobial peptides in insects [35] . It has also been reported that the alternative chitinases encoded in insects might play a role in immune defense against chitin-containing pathogens [36] . In this respect , it is not surprising that entomopathogenic fungi have evolved similar strategies to suppress chitin-induced immunity in insects . Similar to the finding that extracellular ChElp2 is localized in fungal cell walls of C . higginsianum at the biotrophic interface [19] , secreted Blys2 can target the cell walls of B . bassiana . In addition , we found that the loss of Blys2 could lead to the delay of fungal cell escape from hemocyte encapsulation and the upregulation of antifungal peptide gene in insects . Deletion of Blys5 could also result in a higher level of activation of antifungal gene expression in insects when compared to the infection by the WT strain of B . bassiana . In contrast to Blys2 , Blys5 can additionally protect fungal hyphae against chitinase degradation . Taken together , these results suggest therefore that the Blys2 effector could coat and protect the cell walls of insect pathogens from host cell recognition while Blys5 could additionally shield fungal cells from the hydrolysis of insect chitinases , the non-redundant functions of these two proteins in evasion of insect immune defenses . The study of Slp1 in M . oryzae revealed that the effector is a competitive inhibitor of the rice receptor CEPiB to suppress chitin-induced immune responses in rice cells [17] . Since the virulence defects of ΔBlys2 and ΔBlys5 could be heterologously restored by Slp1 , it cannot be ruled out that both Blys2 and Blys5 may also be able to outcompete the chitin receptor of insects , if any , to deregulate host immunities . Cell wall re-modifications occur during fungal invasion of the insect body cavity [3] . Consistent with a previous observation [31] , we found that the chitin content of cell walls was reduced when the fungus was growing in insect hemocoels ( Fig 4 ) . This result could help explain why the overexpression of Blys2 and Slp1 did not lead to an obvious increase of fungal virulence that might be due to the saturation of the chitin substrate . As indicated above , Blys5 can additionally bind chitosan , a biopolymer that is rich in insects and has an immediate antifungal activity [37] . Thus , besides its shield effect against chitinases , Blys5 may additionally contribute to the detoxification of insect chitosan to facilitate fungal growth in insects . In conclusion , we report the presence of diverse LysM proteins in animal pathogenic fungi and reveal that , similar to the LysM effectors of plant pathogenic fungi , the extracellular LysM proteins are virulence factors of the insect pathogen B . bassiana . Alternative proteins can bind chitin , coat fungal cell walls , deregulate insect immunities and/or protect fungal cells from host chitinase damage to facilitate fungal infection . Intriguingly , we found that the highly divergent plant pathogen effector Slp1 could restore the virulence defect of the Blys2 and Blys5 deletion mutants . The results of this study not only expand our knowledge of LysM protein evolution and functional diversification/similarity but also establish that the employment of effectors to evade host immunities similarly occurs during fungal interactions with animal hosts .
The WT and mutants of the B . bassiana strain ARSEF 2860 were routinely cultured on PDA ( BD Difco ) at 25°C for two weeks for conidial spore isolation . The rice blast fungus M . oryzae strain 70–15 was used to amplify the Slp1 gene . For liquid incubation , fungi were grown in Sabouraud dextrose broth ( SDB , BD Difco ) at 25°C in a rotatory shaker . Conidium suspensions were prepared in 0 . 05% ( v/v ) Tween-20 and filtered through four layers of sterile lens-cleaning tissues to remove hyphal fragments for different experiments . The WT and mutant strains were also grown on PDA amended with H2O2 ( 3 mM ) and Calcofluor White ( 200 μg/ml ) at 25°C for different times [38] . The cultures were additionally incubated at 37°C to compare the stress responses between the WT and mutant strains . The proteins containing LysM domain ( s ) were retrieved from the genomes of 13 insect pathogenic fungi ( S1 Table ) , 13 plant pathogenic fungi , and nine mammalian pathogenic fungi ( S2 Table ) using the program HMMER 3 . 1 ( http://hmmer . org/ ) . The obtained sequences were manually curated for further analysis of signal peptide , cysteine-residue ratio , transmembrane feature and length of the LysM domain proteins . A phylogenetic tree was constructed for the LysM domain-containing proteins retrieved from the genomes of the 13 insect pathogenic fungi mentioned above and the three plant pathogenic fungi that have been investigated and genome sequenced: M . oryzae , F . oxysporum and Zymoseptoria tritici . Thus , the full sequences of each protein were aligned using the program MUSCLE 3 . 8 . 31 [39] , and a maximum likelihood tree was constructed using RAxML ( ver . 3 . 1 ) using a WAG model and a bootstrap test of 1 , 000 replicates . The conserved sites at LysM domains extracted from the Pfam PF01476 of proteins from bacteria ( 8 , 738 proteins ) , proteins from insect pathogens ( 133 ) and plant pathogens ( 96 ) were searched and plotted using the program GLAM2 [40] . The LysM domain sequences extracted from the selected proteins ( S1 and S2 Tables ) were also used for neighbor-joining of phylogenetic analysis using MEGA ( ver . 7 . 0 ) [41] with a bootstrap test of 1 , 000 replicates . Twelve LysM domain-containing protein genes are encoded by B . bassiana [30] . To examine the expression of these genes , total RNA was extracted from the mycelia or blastospores harvested from SDB and the conidial spores or hyphae from the PDA plates . To determine gene expression during fungal in vivo infection , the last instar larvae of wax moth ( G . mellonella ) were individually injected with 10 μl of spore suspension ( 107 spores/ml ) for 60 hrs . Insect hemolymph was collected on ice , and fungal hyphal bodies were harvested by gradient centrifugation using Centricoll ( Sigma-Aldrich ) [30] . Each RNA sample was converted to cDNA using an AffinityScript multiple-temperature cDNA synthesis kit ( Toyobo ) . qRT-PCR analysis was performed using a SYBR Premix Ex Taq kit ( Takara ) containing the primer pairs for different genes ( S3 Table ) on an ABI Prism 7000 system ( Applied Biosystems ) . The β-tubulin gene ( BBA_07018 ) of B . bassiana was amplified as an internal control . To determine insect antifungal gallerimycin gene expression , the last instar wax moth larvae were individually injected with the spore suspensions of WT and mutants for 36 hrs . Insect fat bodies were then dissected on ice and collected for RNA extraction to quantify the expression of the antifungal gene [9] . Based on the RT-PCR analysis , the genes Blys2 , Blys4 , Blys5 , Blys6 , Blys7 and Blys8 were found to be transcribed by the fungus during growth in insect hemocoels . Targeted gene deletion of these sixe genes was individually performed by homologous recombination via Agrobacterium-mediated fungal transformation as previously described [42] . Briefly , the 5′- and 3′-flanking sequences were amplified using genomic DNA as a template with the appropriate primer pairs ( S3 Table ) . The products were purified , digested with restriction enzymes and then inserted into the corresponding sites of the binary vector pDHt-bar ( conferring resistance against ammonium glufosinate ) to generate different plasmids ( S4 Table ) for transformation of the WT strain . In addition , the LysM effector Slp1 gene of M . oryzae [17] was amplified with the primers Slp1F/Slp1R and placed under the control of a constitutive GpdA gene ( BBA_05480 ) promoter , and the cassette was integrated into the binary vector pDHt-ben ( conferring resistance against benomyl ) to transform the deletion mutants ΔBlys2 and ΔBlys5 for heterologous complementation . The Slp1-gene cassette was also cloned into the plasmid pDHt-bar to transform the WT strain for constitutive expression . A laccase gene ( BBA_08183 ) promoter was also used to control Blys2 to transform the WT because this laccase gene was highly transcribed by the fungus during growth in insect hemocoels [30] . The obtained mutants ( S4 Table ) were verified by PCR and RT-PCR analyses using the corresponding primer pairs ( S3 Table ) . To determine protein localization , Blys2 and Slp1 were individually fused in frame at the C-termini with a GFP protein , and the cassettes were placed under the control of the GpdA gene promoter . The binary vectors were used to transform the WT strain of B . bassiana . The truncated Blys2 without signal peptide region ( Blys2-SP ) was also fused with the GFP protein . The WT and obtained mutants were grown under various nutrient conditions , and the fungal cells were harvested for microscopic observations . For chitin staining , fungal cells were incubated with 0 . 01% Calcofluor White ( Sigma-Aldrich ) buffered in 10% KOH for 1 min and rinsed twice with phosphate buffered saline ( PBS , pH 8 . 0 ) before examination using a confocal microscopy ( TCS SP8 , Leica ) . To determine the secretion of Blys2 , the obtained mutants engineered with GFP-fused Blys2 with and without SP regions were grown in SDB medium in a rotatory shaker at 220 rpm for three days . The cultures were filtered through filter paper , and the filtrates were further treated with a syringe filter unit ( GP/SLGP033RS , Millipore ) to remove fungal cells . Extracellular proteins were concentrated by precipitation with ammonium sulfate . The samples were centrifuged at 15 , 000× g for five minutes , and the proteins were reconstituted in Tris-HCl buffer ( pH 8 . 0 ) and kept at 4°C . Mycelial samples were washed twice with sterile water and homogenized for total protein extraction . Extracellular and intracellular protein samples were separated using a 12% SDS-PAGE gel , and the Western blotting analysis was performed using the anti-GFP antibody ( Abcam , China ) . To determine the binding feature of Blys2 and Blys5 to cell wall chitins , conidia of B . bassiana were harvested from the two-week old of PDA plates , and suspended in 1 ml of 0 . 05% ( v/v ) Tween-20 . The suspension was filtered through two layers of sterile filter paper to remove fungal hyphae . The conidia were collected by centrifugation , washed twice with distilled water and then resuspended in 5% ( w/v ) KOH for boiling for 30 min . After cooling down , cell wall chitin sample was collected by centrifugation at 15 , 000× g for 3 min , and the pellets were then washed with distilled water for three times , and resuspended in the solution of 40% H2O2 and glacial acetic acid ( 1:1 ) for boiling in water for 45 min [43] . Chitin was collected by low speed of centrifugation . The obtained pellets were washed for three times and resuspended in PBS ( pH 8 . 0 ) for experiments . To determine the effect of LysM domain number on chitin binding , we performed Blys2 protein truncations and polysaccharide binding assays . Thus , the primer pairs BF1/BR2 , BF1/BR4 , BF2/BR5 , BF3/BR5 and BF4/BR5 ( S3 Table ) were used to amplify the regions containing the LysM domains 1–2 , 1–4 , 2–5 , 3–5 and 4–5 of Blys2 , respectively . The cDNA of Blys5 was amplified with the primers B5F and B5R . The purified products were integrated into the expression vector pET28b containing the His ×6 tag at the C-terminus using the Gateway clone system ( Invitrogen ) . The plasmids were individually transformed into the BL21 ( DE3 ) strain of E . coli for expression . For binding assays , the purified proteins ( at a final concentration of 20 μg/ml each ) were individually incubated with 3 μg of fungal chitin isolated above , chitin beads ( Bioleaf , China ) , chitosan ( Sigma-Aldrich ) and cellulose ( Sigma-Aldrich ) in a total volume of 800 μl of water for 30 min with gentle rotating at room temperature . The samples were centrifuged at 15 , 000× g for 5 min [15] . The supernatants were collected , and the pellets were washed with PBS for three times . A SDS-PAGE analysis was conducted to detect the proteins in the supernatants and those bound to polysaccharides . To determine the potential effect of Blys2 and Blys5 on protecting fungal cell walls against chitinase , the WT and WT::Slp1 spores were germinated in SDB ( at a final concentration of ca . 1 × 105 spores /ml ) for 16 hrs . The germlings were harvested by centrifugation and washed twice with the 0 . 1 M potassium phosphate buffer containing 0 . 7 M KCl and 5 mM MgSO4 . The WT germlings were pre-incubated with either Blys2 or Blys5 ( 10 μg ) for 2 hrs , and additional WT ( mock control ) and the WT::Slp1 germlings were treated with the buffer for 2 hrs . The samples were washed twice with the buffer and then treated with the buffered chitinase ( Sigma-Aldrich; 0 . 2% , w/v ) mixtures containing β-glucuronidase ( Sigma-Aldrich; 0 . 4% , w/v ) , cellulase ( Sigma-Aldrich; 0 . 4% , w/v ) and lysozyme ( Yeasen; 0 . 4% , w/v ) [5] for 2 hrs at 30°C under gentle shaking . The rate of protoplast formation was estimated for each sample by examining 50 microscopic fields . Insect bioassays were conducted against the newly emerged last instar larvae of wax moth G . mellonella . Conidial suspensions were prepared for both topical infection ( 1 × 107 conidia/ml ) and injection ( 1 × 106 conidia/ml ) assays . Each treatment had three replicates with 15 insects each , and the experiments were repeated twice . For injection assays , spore suspensions ( 10 μl each ) were injected into the base of the second proleg of the insects . Additional insects were injected and bled at various times to examine insect cellular immune response and fungal developments within the insect hemocoels . Insect mortality was recorded every 12 h after the treatments , and the LT50 values were calculated for each strain by Kaplan-Meier analysis . The differences were estimated between the WT and each mutant by the Log-rank test with the program SPSS ( ver . 19 ) [44] . | Insect pathogenic fungi are of importance for both applied and basic research . Relative to the advances in understanding fungus-plant interactions , the mechanisms of the molecular pathogenesis of entomopathogenic fungi are rather limitedly understood . In particular , the machinery of effector-mediated inhibition of host immunity has not been well established in fungus-insect interactions . LysM effectors have been characterized as virulence factors in plant pathogens to suppress chitin-triggered immunity in plants . We found that the divergent LysM proteins are also present in animal pathogens . By using the insect pathogen Beauveria bassiana as a model , we revealed that two of 12 encoded LysM protein genes Blys2 and Blys5 that were transcribed by the fungus growing in insects are required for full fungal virulence against insect hosts . Interestingly , the virulence defects of ΔBys2 and ΔBys5 could be fully restored by complementation with the divergent Slp1 effector from the plant pathogen Magnaporthe oryzae . Both Blys2 and Blys5 can deregulate insect immune responses , and the latter can additionally protect fungal cells from chitinase hydrolysis . The findings of this study establish the contribution of LysM effectors to fungal virulence against insect hosts . | [
"Abstract",
"Introduction",
"Results",
"Discussions",
"Materials",
"and",
"methods"
] | [
"chitin",
"invertebrates",
"fungal",
"genetics",
"animals",
"fungi",
"plant",
"science",
"materials",
"science",
"plant",
"pathology",
"macromolecules",
"materials",
"by",
"structure",
"polymers",
"polymer",
"chemistry",
"mycology",
"proteins",
"chemistry",
"insects",
"... | 2017 | Divergent LysM effectors contribute to the virulence of Beauveria bassiana by evasion of insect immune defenses |
Enzymes of the M32 family are Zn-dependent metallocarboxypeptidases ( MCPs ) widely distributed among prokaryotic organisms and just a few eukaryotes including Trypanosoma brucei and Trypanosoma cruzi , the causative agents of sleeping sickness and Chagas disease , respectively . These enzymes are absent in humans and several functions have been proposed for trypanosomatid M32 MCPs . However , no synthetic inhibitors have been reported so far for these enzymes . Here , we present the identification of a set of inhibitors for TcMCP-1 and TbMCP-1 ( two trypanosomatid M32 enzymes sharing 71% protein sequence identity ) from the GlaxoSmithKline HAT and CHAGAS chemical boxes; two collections grouping 404 compounds with high antiparasitic potency , drug-likeness , structural diversity and scientific novelty . For this purpose , we adapted continuous fluorescent enzymatic assays to a medium-throughput format and carried out the screening of both collections , followed by the construction of dose-response curves for the most promising hits . As a result , 30 micromolar-range inhibitors were discovered for one or both enzymes . The best hit , TCMDC-143620 , showed sub-micromolar affinity for TcMCP-1 , inhibited TbMCP-1 in the low micromolar range and was inactive against angiotensin I-converting enzyme ( ACE ) , a potential mammalian off-target structurally related to M32 MCPs . This is the first inhibitor reported for this family of MCPs and considering its potency and specificity , TCMDC-143620 seems to be a promissory starting point to develop more specific and potent chemical tools targeting M32 MCPs from trypanosomatid parasites .
Members of the Trypanosomatidae family comprise parasitic organisms that cause highly disabling and often fatal diseases in humans and animals . The species that are responsible for human infections are Trypanosoma brucei , which cause Human African trypanosomiasis ( HAT ) , Trypanosoma cruzi , the etiological agent of Chagas disease ( American trypanosomiasis ) , and Leishmania spp . , which cause different forms of leishmaniasis . Together , these vector-borne diseases constitute a substantial public health problem for which there is not a satisfactory treatment [1] . Major side-effects , and in some cases low effectiveness , are common problems associated with existing therapy . This situation makes imperative the development of new chemotherapeutic options . In this context , new drugs based on unique aspects of parasite biology and biochemistry are of great interest , particularly in the case of emerging resistance to traditional treatments [2–4] . In this scenario , proteases have become popular targets as these enzymes play key functions in parasite biology; namely nutrition , cell cycle progression , invasion and pathogenesis , among others . The M32 family of metallocarboxypeptidases ( MCPs ) contains a group of hydrolases , which although being broadly distributed among prokaryotic organisms , are only present in a few eukaryotes including some green algae and trypanosomatids [5] . This unique phylogenetic distribution , in particular the absence of M32 enzymes in metazoans , has been considered an attractive trait due to the high specificity/selectivity potential of this family for drug target development . Within the Trypanosomatidae family several conserved M32 MCPs have been characterized [5–10] . Nonetheless , the cellular or biological functions of these proteins are currently unknown , as well as their essentiality status . In T . brucei , the genome-wide study by Alsford et al . ( 2011 ) reported no significant lost-of-fitness after induction of T . brucei MCP-1 ( TbMCP-1 ) RNAi in bloodstream and procyclic stages , as well as in the differentiation from procyclic to bloodstream forms [11] . More recently , however , it has been shown that TbMCP-1 null mutant strains display extended doubling times in culture , suggesting that this enzyme might contribute to the adaptive fitness of the bloodstream form [12] . On the basis of their biochemical properties and stage-specific expression , the L . major M32 carboxypeptidase has been implicated in the catabolism of peptides and proteins to single amino acids required for protein synthesis [7] . The restricted substrate preference of T . cruzi MCP-1 ( TcMCP-1 ) , plus its strong structural similarity to angiotensin I-converting enzyme ( ACE ) , neurolysin and thimet oligopeptidase [8] , have also pointed out a possible regulatory role of this family in the metabolism of small peptides . In fact , it has been shown that TcMCP-1 can produce des-Arg9-bradykinin [6] , a peptide that promotes the process of cell invasion through B1 receptors by the T . cruzi trypomastigotes [13] . In this sense , two reports have suggested that M32 peptidases are secreted by trypanosomatids [14 , 15] , a fact that is in agreement with this hypothesis . In the current scenario , the availability of selective small-molecule modulators of M32 MCPs activity would be of great value to ask mechanistic and phenotypic questions in both biochemical and cell-based studies . However , no inhibitors have been reported to date for these enzymes or other members of this family . Recently , a diverse collection of ~ 1 . 8 million compounds from the proprietary library of GlaxoSmithKline ( GSK ) has been run through whole-cell phenotypic screens against L . donovani , T . cruzi and T . brucei . As a result , three anti-kinetoplastid chemical boxes of ~200 compounds each were assembled and open sourced [16] . The guiding design criteria for these molecule sets were chosen to include structures from different chemical families that are likely to be active against a wide variety of targets . By taking advantage of this diversity , we identified the first inhibitors of the M32 family of MCPs within the GSK HAT and CHAGAS chemical boxes . As model enzymes of the M32 family we employed TcMCP-1 and TbMCP-1 , which have similar basic amino acid preference at the P1´ position and share 71% of protein sequence identity [5 , 6] .
To evaluate compounds in the HAT and CHAGAS chemical boxes , we devised a continuous assay for each MCP , based on FRET ( fluorescence resonance energy transfer ) peptides . We carried out the optimization process in 384 well plates , the same format used for the screening of the compound collections . For the selection of the most suitable substrate for the HTS assay , we initially assayed six FRET peptides against both enzymes . These were recently designed considering subsite preferences ( P1´-P4 ) of TcMCP-1 and TbMCP-1 [12] . However , because no peptide was completely satisfactory for both enzymes , we selected independent substrates , Abz-LKFK ( Dnp ) -OH and Abz-RFFK ( Dnp ) -OH , for TcMCP-1 and TbMCP-1 assays , respectively . After substrate selection , a convenient enzyme concentration in the assay was determined through the activity of 2-fold dilutions of TcMCP-1 and TbMCP-1 at a fixed substrate concentration ( Fig 1A and 1B ) . Moreover , the Selwyn test [17] revealed no enzyme inactivation under the conditions tested ( Fig 1C and 1D ) . Thus , for a wide range of enzyme concentrations ( for both MCPs ) , the V0 vs . [E]0 curves showed a linear behavior ( Fig 1E and 1F ) . In particular , for [TcMCP-1]0 < 0 , 34 nM and [TbMCP-1]0 < 1 , 53 nM , the rate of the substrate hydrolysis remained constant for at least 40 minutes , a suitable time to perform the screening ( Fig 1A and 1B ) . The best balance between TcMCP-1 activity on Abz-LKFK ( Dnp ) -OH substrate ( estimated as dF/dt ) and the time over which the reaction displayed linear kinetics was achieved at [TcMCP-1]0 = 0 , 17 nM . Under these conditions , the enzyme showed the typical hyperbolic behavior predicted by the Michaelis-Menten equation ( Hill coefficient = 1 , 06 ) and an estimated KM value of 2 , 23 ± 0 , 28 μM ( Fig A in S1 Text ) . Similarly , when the TbMCP-1 concentration was fixed at 1 , 25 nM we obtained a KM value on Abz-RFFK ( Dnp ) -OH substrate of 0 , 37 ± 0 , 06 μM ( Hill coefficient = 1 , 03 ) ( Fig A in S1 Text ) . To afford the best opportunity to find compounds with different inhibition modalities , we decided to employ balanced assay conditions ( i . e . KM/[S] = 1 ) [18] . Using these conditions , preliminary characterization experiments of both optimized assays showed good general performance , with a dynamic range ( μC+—μC- ) higher than 15 RFU/sec , a μC+/μC- ratio ≥ 50 , good reproducibility ( VC < 5% ) and a Z´ factor value in the range 0 , 6–0 , 8 . Using the same lot of substrate and enzyme , the 404 compounds present in the HAT and CHAGAS chemical boxes were screened at a single fixed dose ( 25 μM ) . Each plate included 24 positive and negative controls , plus 16 wells containing 31 , 25 mM EDTA ( inhibition control ) alternately located in columns 11 , 12 , 23 and 24 . In general , for each MCP , both plates presented highly similar Z´ scores although best values were obtained for the TbMCP-1 assay presumably due to the lower background signal of the Abz-RFFK ( Dnp ) -OH substrate . To avoid the interference of highly fluorescent compounds , an auto-fluorescence cut-off value equal to 2x105 RFU was used to accept or discard a molecule from the subsequent analysis . Using this limit , ~19% of the compounds were eliminated for TcMCP-1 and TbMCP-1 assays . Statistics are summarized in Table 1 . As shown in Table 2 , if we consider a cut-off value ≤ 3 standard deviations from the control mean ( μc+ - 3σc+ ) , 70 and 132 inhibitory molecules were retrieved for TcMCP-1 and TbMCP-1 , respectively . To reduce the number of resultant hits , we explored other two thresholds focusing only in outliers: i ) those compounds showing slopes >3σ standard deviations above the average of all slopes in the plate ( control independent ) and ii ) those compounds showing an inhibition percentage >3σ standard deviations above the average for the plate ( control dependent ) . Interestingly , both criteria retrieved exactly the same list of compounds for TcMCP-1 ( n = 5 ) while for TbMCP-1 the intersection between this two groups was lower ( 2 out of 4 compounds ) . In the secondary screening we decided to include all compounds that showed ≥ 40% of inhibition ( TcMCP-1: 23 compounds; TbMCP-1: 27 compounds ) . To estimate IC50 for the resulting hits , two-fold serial dilutions , ranging from 7 , 5 pM to 62 , 5 μM , were analyzed against both recombinant MCPs using identical assay conditions as in the primary screening . Prior to the analysis of the complete dataset , we examined whether there was a correlation between the inhibition percentages in the primary ( compound concentration 25 μM ) and secondary screening , using only the data corresponding to a compound concentration of 31 , 5 μM . This was important to assess consistency of data , as both screening rounds were performed without technical replicates due to limitation of compound stocks . For TcMCP-1 , 9 compounds presented similar behavior in both screenings ( correlation coefficient r2 = 0 , 9868; slope = 1 , 146 ) ( Fig 2A ) whereas 7 molecules failed to reach ≥ 40% of inhibition threshold ( n = 6 ) or displayed no inhibition ( n = 1 ) ( correlation coefficient r2 = -0 , 518; slope = 0 , 2595 ) . Additionally , 7 compounds performed better in the secondary screening ( correlation coefficient r2 = 0 , 5156; slope = 1 , 2749 ) . For the T . brucei enzyme , consistent results in both assays were achieved only by 8 compounds ( correlation coefficient r2 = 0 , 9349; slope = 1 , 080 ) ( Fig 2B ) . About 45% of the samples did not repeat the ≥ 40% of inhibition criterium ( n = 10 ) or did not inhibit ( n = 2 ) TbMCP-1 ( correlation coefficient r2 = 0 , 1163; slope = 0 , 3173 ) . Finally , another 7 molecules performed better in the secondary screening than in the first round . Despite the observed round to round discrepancies ( Table A in S1 Text ) , we decided to continue curve analysis for all the compounds , with the exception of the three that showed no inhibition at 31 , 5 μM during secondary screening . For TcMCP-1 , five compounds ( TCMDC-143620 , TCMDC-143422 , TCMDC-143456 , TCMDC-143209 and TCMDC-143385 ) showed an IC50 value ≤ 10 μM ( Fig 3A and Table 3 ) . In good agreement , the four more potent molecules ( TCMDC-143620 , TCMDC-143422 , TCMDC-143456 and TCMDC-143209 ) also inhibited the T . brucei enzyme ( Table 3 ) . Compounds TCMDC-143385 and TCMDC-143172 ( which display an IC50 ~10 μM for TcMCP-1 ) did not reach the 40% inhibition threshold in the TbMCP-1 primary screening and were left out from the secondary analysis . Other potent molecules , namely TCMDC-143409 and TCMDC-143323 were specific inhibitors of T . brucei enzyme or produced little inhibition on TcMCP-1 ( < 30% ) ( Fig 3B and Table 3 ) . The structure of the top-five inhibitors for each enzyme is shown in Fig 3C . To first assess the possibility that these lead compounds have shared structural features that help explain their bioactivity profile , we performed three different clustering strategies: one using Tanimoto similarity ( Fig B in S1 Text ) , one based on shared substructures ( overlap of Maximum Common Subgraphs , MCS ) ( Fig C in S1 Text ) , and the third one based on shared physicochemical properties ( Fig 4 ) . Whereas the Tanimoto clustering was expected to be inconclusive based on the premises used to assemble the chemical boxes ( one or two putative chemotypes per box [16] ) ; the clustering based on physicochemical properties also showed no significant correlation between these properties and the observed IC50s . Similarly , MCS clustering provided no insights into candidate substructures guiding the activity or specificity of the compounds against each enzyme . In all three strategies , the clusters not only group up dissimilar potencies , but also mix compounds with different enzyme specificity . To determine the number and type of Zinc-binding groups ( ZBGs ) among the compound leads , an MCS analysis was performed using an ad hoc curated [20 , 21] database of ZBGs . From a total of 48 groups available in the database , only six of them were found among 24 of the 30 lead compounds: pyridine ( 14 compounds ) , sulfonamide ( 7 compounds ) , imidazole ( 4 compounds ) , pyrazole ( 3 compounds ) , diol ( 1 compound ) and hydrazide ( 1 compound ) . The majority of compounds ( 24 out of 30 ) presented at least one ZBG in the structure . More specifically , 15 with a single group and 9 with two groups were found . All compounds and their corresponding ZBGs have been summarized in Fig D in S1 Text . Considering the abundance of ZBGs and heteroatom-containing moieties in the hits , we evaluated the possibility of a nonspecific mechanism of inhibition ( involving metal chelation ) for the top-five inhibitors identified in the screening for each enzyme . Because M32 MCPs show a strong topological similarity with ACE [22] , we chose this enzyme to estimate the IC50 value for each molecule . As done for the MCPs essays , ACE activity was analyzed employing a FRET substrate , Abz-FRK ( Dnp ) P-OH , at a concentration equal to the apparent KM of the enzyme ~3 μM [23] . Experiment set up is summarized in Figs E and F in S1 Text . For comparative purposes , captopril , a potent competitive ACE inhibitor , was included in the analysis ( IC50 ~1 nM ) ( Fig 5A ) . Under these conditions , no inhibition could be detected for any of the compounds evaluated , thus suggesting that these molecules are not promiscuous metallocarboxypeptidase inhibitors ( Fig 5B , 5C and 5D ) but are instead specific inhibitors of M32 MCPs .
M32 MCPs have an unusual phylogenetic distribution ( with trypanosomatids being among the few eukaryotic genomes encoding these enzymes ) . Hence M32 MCPs from parasites arose naturally as interesting candidates for drug target development . Furthermore , the current lack of knowledge about the cellular and/or physiological role ( s ) of these enzymes makes the identification of potent inhibitors a task of great significance , as these compounds may be used as molecular probes to potentially identify natural substrates , to recognize the specific pathways in which they are involved or , hopefully , to perform their chemical validation as drug targets . In this work , we describe the first drug-like inhibitors of TcMCP-1 and TbMCP-1 , two closely related MCPs from the human pathogens T . cruzi and T . brucei , respectively . Our starting point were the GSK HAT and CHAGAS boxes , two small collections containing non-redundant , chemically diverse and highly bioactive compounds [16] , which could facilitate future optimization efforts . Although we initially aimed for a common assay for both MCPs , we soon realized that the use of different FRET substrates for each enzyme resulted in better general performance of the individual assays ( considering signal robustness , temporal duration of linear kinetics , dynamic range , μC+/μC- ratio and Z´ factor ) . Surprisingly , the substrates that resulted most suitable for the developed HTS assays were not , in any case , those that showed the best values of kcat , KM and kcat/KM in their previous kinetic characterization [12] . Although different assays were used to screen these collections , we were able to find specific inhibitors for both enzymes , and perhaps more important , mutual inhibitors; suggesting the consistency of inter-assay results . Of note , specific inhibitors for each enzyme were distributed evenly among HAT and CHAGAS boxes with no apparent bias . This fact confirms the importance of not circumscribing the search to just the pathogen-specific box , but instead to widen the search to all the boxes available , as previously observed for T . cruzi cysteine peptidase cruzipain [24] . Due to the limited amount of compound stocks , we decided to implement the screening of chemical boxes in singlet , with primary evaluation of all compounds at a fixed dose and further dose-response analysis of unconfirmed hits in a secondary screening . As expected , given the error-prone nature of the single-well ( single dose , single replicate ) measurements used in primary screening , significant discrepancies in inhibition were observed for some compounds in comparison to secondary dose-response evaluation . These discrepancies are common and may be due to a variety of factors [25] . Besides intrinsic compound-specific and experimental data variability [26] , these factors may include solubility issues ( given that in primary and secondary screenings both the final concentration and serial-dilution protocol were different ) , differential stability of compounds in stock ( 10 mM ) and working ( 2 mM ) solutions [27] , unintended absorption of the compounds to different containing materials during storage , moderate dose-dependent quenching effects of compounds on fluorescence readouts , among others [28] . In addition , although we included 0 , 01% Triton X-100 in assay buffer , compound-specific aggregate formation was not tested and thus , cannot be dismissed . As mentioned , we identified in this work eight molecules able to inhibit both MCPs . These mutual inhibitors came from both boxes in similar numbers , as previously noted for enzyme-specific compounds . Interestingly , in all cases they were more potent inhibitors of TcMCP-1 , for reasons that are as yet unclear . Importantly , four of these compounds proved to be inactive on ACE , a Zinc-dipeptidyl carboxypeptidase involved in various physiological and physiopathological conditions in mammals [29] which shows significant structural similarity to M32 enzymes [22 , 30] . This fact strongly suggests that despite the structural resemblance and the small number of compounds tested here , the identification of inhibitors with high selectivity for trypanosomatid M32 MCPs over ACE can be achieved , a point in favor to the specific druggability of these enzymes . The identified inhibitors display high structural diversity , with many showing only marginal similarity to the other hits , hence representing different structural clusters and presumably , different inhibitory scaffolds . In this regard , the presence of “unpaired” hits is not surprising , considering that no more than two members of the same structural cluster were included per box during collection assembly [16] and that “twin” compounds might well not pass the activity or auto-fluorescence filters included in this work . Among the identified inhibitors , only TCMDC-143265 and TCMDC-143551 share similar core structures , thus probably populating the same cluster and sharing a common active scaffold . A significant part of both molecules is identical and adopts the same spatial conformation ( Fig G in S1 Text ) , with the largest differences located around the benzamide ring . Besides the obvious differences in the length and position of sulfonamide substituents , the chlorine substitution in position 2 imposes a ~90° rotation of the benzamide ring in TCMDC-143265 compared to TCMDC-143551 , where all ring systems are almost coplanar . Interestingly , these structural differences seem to dictate the selectivity toward TcMCP-1 , as TCMDC-143551 inhibits both enzymes whereas TCMDC-143265 is specific for TbMCP-1 . Even for this pair of compounds , there is no evident substructure responsible for M32 MCPs bioactivity; though this is probably a biased observation due to the lack of well-defined structural features for M32 MCPs inhibitors . Although the crystallographic structure of TcMCP-1 has been determined [8] and subsite specificity have been explored for both enzymes using FRET substrate libraries [12] and mutagenesis [6 , 8] , little is yet known about how substrates are accommodated into the catalytic groove , which residues are key determinants of subsite specificity and the significance of the hinge-type movement between L and R domains in the stabilization of enzyme-substrate or enzyme-inhibitor complexes . With all these gaps to fill , it seems risky to speculate about the modes of interaction of these new inhibitors with TcMCP-1 and TbMCP-1 . However , a presumptive explanation can be put forward . As in the case of many other metallopeptidase inhibitors , it is likely that inhibition of trypanosomatid M32 MCPs occurs throughout the perturbation of the coordination sphere of the catalytic metal ion ( presumably Zn2+ in the case of TcMCP-1 and TbMCP-1 , by extension from other M32 enzymes [31] ) . Typically , synthetic metallopeptidase inhibitors achieve preliminary affinity and target selectivity through the formation of stabilizing interactions with specific residues within the active site; while a ZBG is responsible for metal chelation , enhancing binding affinity , modulating selectivity and disrupting catalytic activity [32] . For the majority of the inhibitors presented here , it was possible to identify typical ZBG or at least , heteroatom-containing groups able to establish a coordinative bond with a Zn2+ ion ( Fig D in S1 Text ) . For those compounds , an inhibition mechanism like the one described above is possible . For other molecules not having a Zn-coordinating group , the most plausible explanation is that inhibition occurs as a result of the prevention of substrate binding by the partial occupancy or the deformation of the catalytic cleft by the inhibitor molecule , as previously observed for Non-Zinc-Binding inhibitors of other metallopeptidases [33] . The vast majority of the hits identified here inhibit one or both MCPs in the micromolar range , with only a few of them showing potencies <10 μM . Outstandingly , TCMDC-143620 inhibits TcMCP-1 in the sub-micromolar range ( it also inhibits TbMCP-1 , but with potency ~7-fold lower ) . This is the most potent inhibitor described so far for an enzyme of the M32 family and seems a promising candidate for further structure-based optimization . The unusually high flexibility of the M32 MCPs around the active site [31 , 34] prevented us to use a docking approach to get insights of the binding mode of this compound within TcMCP-1 and TbMCP-1 catalytic clefts . However , the TCMDC-143620 molecule seems able to form a variety of stabilizing interactions . These may include hydrophobic and electrostatic interactions , hydrogen bonding and the coordination to the metal ion through the pyridine ring . In addition , the presence of a central sulfonamide group and a distal nitrile group add further interaction possibilities to this molecule . For example , the sulfonamide group has been extensively incorporated into metallopeptidase inhibitors due to its ability to improve the enzyme-inhibitor binding by different mechanisms . These mechanisms include: i ) direct formation of hydrogen bonds to the enzyme backbone , ii ) properly redirection of bulky groups into enzyme pockets by inducing a twist in the structure of the inhibitor molecule and iii ) even cooperate with other chelating groups in the coordination of the catalytic metal ion [35] . Similarly , the nitrile group in TCMDC-143620 can establish polar interactions , hydrogen bonds or react with serine or cysteine side chains to form covalent adducts which would greatly stabilize inhibitor binding [36] . Interestingly , the nitrile group is also able to form coordinative bonds with a variety of metal ions including Co2+ , Mn2+ , Fe3+ , Cu2+ and Zn2+ [37] . Thus , a possible role of this group in the direct coordination of the catalytic metal ion cannot be discarded at present . The determination of the crystallographic structure of TcMCP-1 or TbMCP-1 in complex with TCMDC-143620 would provide a definitive answer to these questions as well as important clues to undertake the future lead-optimization of this hit . A preliminary analysis of the bioactivity profile of TCMDC-143620 ( https://pubchem . ncbi . nlm . nih . gov/compound/91800813 ) indicates that it shows potent activity against T . cruzi in culture and only moderate but measurable activity on T . brucei and L . donovani . Also , this compound exhibits moderate cytotoxicity on mammalian cell NIH 3T3 ( IC50 = 13 μM ) but resulted inactive on HepG2 ( IC50 > 100 μM ) . Considering target-specific assays; this compound has a single bioactivity report . TCMDC-143620 was found to be a potent inhibitor ( IC50 = 79 nM ) of T . cruzi sterol 14-α demethylase ( CYP51 ) enzyme , which is involved in the ergosterol biosynthesis pathway and was considered until recent years as a promissory therapeutic target for Chagas disease [38 , 39] . The inhibition of this target is probably the cause of its reported anti-T . cruzi activity . This might also explain , at least partially , the moderate cytotoxic and anti-T . brucei and L . donovani activities reported for this compound , considering the global similarities of enzymes within CYP51 family [40 , 41] . Although involved in other studies as part of the GSK CHAGAS Box [42] , no further information is currently available from the evaluation of TCMDC-143620 against other molecular targets , except for our previous cruzipain study [24] where it was found to be inactive ( ~7 , 5% of cruzipain inhibition at 25 μM ) . A complete profile of the off-target activity of TCMDC-143620 would be critical for future optimization efforts in order to achieve a suitable M32 MCPs probe from this compound . In summary , 30 micromolar-range inhibitors , presenting both high structural diversity and novelty , have been discovered for TcMCP-1 and/or TbMCP-1 by using continuous , fluorescent-based and HTS-capable enzymatic assays . The best hit shows sub-micromolar affinity for TcMCP-1 , inhibits TbMCP-1 in the low micromolar range and , like other potent hits , is inactive on ACE . Considering its potency and specificity , this molecule seems to be a promissory starting point to develop more specific and potent tools to expand our understanding of the biochemistry and biological role ( s ) of M32 MCPs from trypanosomatid parasites and , hopefully , to assess in a near future their value as drug targets .
Triton X-100 , MOPS ( 3- ( N-morpholino ) propanesulfonic acid ) , DMSO , EDTA and captopril were purchased from Sigma-Aldrich . Substrates Abz-RFFK ( Dnp ) -OH and Abz-LKFK ( Dnp ) -OH were from GenScript ( Piscataway , NJ , USA ) . Black solid bottom polystyrene Corning NBS 384-well plates were from Sigma-Aldrich ( CLS3654-100EA ) . TcMCP-1 ( MEROPS ID: M32 . 003 ) and TbMCP-1 were expressed as GST fusion proteins in E . coli BL21 ( DE3 ) Codon Plus and purified as previously described [6 , 8] . The HAT and CHAGAS chemical boxes [16] were provided by GlaxoSmithKline . The collection comprised 404 compounds , prepared as 10 mM stock solutions in DMSO ( 10 μL each ) and dispensed in 96 well plates . For primary screening , a working solution ( final concentration of 2 mM ) for each compound was prepared by 1/5 dilution in DMSO while 1 μL of the 10 mM stock solution was used for secondary screening of selected compounds , as previously described [24] . The final concentration of compounds tested in primary screening was 25 μM , while the compound concentrations assayed in secondary screening ranged from 7 , 5 pM to 62 , 5 μM . TbMCP-1 and TcMCP-1 activities were assayed fluorometrically with Abz-RFFK ( Dnp ) -OH and Abz-LKFK ( Dnp ) -OH substrates , respectively , in 100 mM MOPS pH 7 , 2 containing 0 , 01% Triton X-100 . Assays were performed in solid black 384-well plates ( final reaction volume ~80 μL ) and the hydrolysis of the K ( Dnp ) -OH group was monitored continuously at 30 °C with a Beckman Coulter DTX 880 Multimode Reader ( Radnor , Pennsylvania , USA ) using standard 320 nm excitation and 420 nm emission filter set . For each MCP , final substrate concentration was set to a value KM /[S] ~ 1 . Optimal enzyme concentration was selected from 2-fold serial dilutions to match three criteria: ( i ) being linearly proportional to V0 , ( ii ) display robust signal evolution at substrate concentration chosen and ( iii ) display linear kinetics for enough time to perform several reading cycles ( at least 8 cycles , minimum time between cycles: 264 sec ) through the 384-wells . In all cases , EDTA ( final concentration 31 , 25 mM ) was used as positive inhibition control . To perform the primary screening , 1 μL of each compound ( 2 mM in DMSO , final concentration in the assay: 25 μM ) , EDTA ( 500 mM , final concentration in the assay: 31 , 25 mM ) were dispensed into 384-well Corning black solid-bottom assay plates . Then , 40 μL of 100 mM MOPS , 0 , 01% Triton X-100 pH 7 , 2 containing TbMCP-1 ( 2 , 50 nM ) or TcMCP-1 ( 0 , 34 nM ) were added to each well , plates were homogenized ( 30 seg , orbital , medium intensity ) and each well subjected to a single autofluorescence read ( ex/em = 320/420 nm ) . Plates were incubated in darkness for 15 min at 30 °C and then 40 μL of Abz-RFFK ( Dnp ) -OH ( 4 μM ) or Abz-LKFK ( Dnp ) -OH ( 0 , 8 μM ) in assay buffer were added to each well to start the reaction . After homogenization ( 30 seg , orbital , medium intensity ) , the fluorescence of the Abz group ( ortho-aminobenzoic acid ) ( ex/em = 320/420 nm ) was acquired kinetically for each well ( 8 read cycles , one cycle every 300 seconds ) . Considering our previous experiences , the auto-fluorescent cut-off was arbitrarily set at 2x105 RFU to discard highly interfering compounds . All compounds were assayed in singlet ( without replicates ) due to the limited availability of stocks . Raw screening measurements were used to determine the slope ( dF/dt ) of progression curves by linear regression for control and non-interfering compound wells . In the case of control-dependent hit selection criteria , percent inhibition percentage ( %Inh ) was calculated for each compound according to the following equation: Inh=100∙[1− ( dFdtWELL−μC− ) ( μC+−μC− ) ] ( 1 ) where dF/dtWELL represents the slope of each compound well and μC+ and μC− the average of MCP ( no-inhibition ) and substrate ( no-enzyme ) controls , respectively . Compounds selected from primary screening were re-tested in a dose-response manner ( final concentration ranging from 7 , 5 to 62 , 5 μM ) using identical assay conditions . To avoid any positional and/or association bias , we randomly defined the row position for each compound . One μL of compounds stock ( 10 mM in DMSO ) and EDTA ( 31 , 25 mM ) were added to the first well of column 1 , followed by addition of 40 μL of 100 mM MOPS , 0 , 01% Triton X-100 pH 7 , 2 buffer . After addition of 20 μL of the same buffer to subsequent wells of the plate , 22 serial 2-fold dilutions were made horizontally . The last two positions of every row were used , alternatively , for C+ and C− controls to reduce any positional and/or association bias . Then , 20 μL of activity buffer containing TbMCP-1 or TcMCP-1 were added to each well , except for those corresponding to C−; completed with 20 μL of activity buffer . After homogenization , 15 minutes of incubation at 30°C and autofluorescence measurement , the substrate ( in activity buffer ) was added to the previous mix . Data collection and processing were performed exactly as described above . Percentage of M32 MCPs residual activity was calculated for each condition according to the following equation: %Res . ActMCP=100∙[ ( dFdtWELL−μC− ) ( μC+−μC− ) ] ( 2 ) where dF/dtWELL represents the slope of each compound well and μC+ and μC− the average of MCP ( no-inhibition ) and substrate ( no-enzyme ) controls , respectively . The IC50 and Hill slope parameters for each compound were estimated by fitting the four-parameter Hill equation to experimental data from dose-response curves using the GraphPad Prism program ( version 5 . 03 ) . Purified rabbit lung ACE ( EC 3 . 4 . 15 . 1 ) was purchased from Sigma-Aldrich . Enzyme activity was assayed fluorimetrically with Abz-FRK ( Dnp ) P-OH ( ex/em = 320/420 nm ) as substrate in buffer 0 , 1 M Tris-HCl , 50 mM NaCl , 10 mM ZnCl2 , pH 7 . 0 containing 0 , 01% Triton X-100 as indicated in [23] . Selected compounds were tested in a dose-response manner ( final concentration ranging from 7 , 5 pM to 62 , 5 μM ) using identical assay conditions employed with both MCPs . Captopril ( 15 pM—125 μM ) was used as inhibition control . Three separate compound clustering routines were used . One of them derived from calculated or predicted molecular features , and the other two directly inferred from different distance metrics between compounds: one using Tanimoto similarity and another one using the overlap score calculated in a MCS ( Maximum Common Subgraph ) pipeline . The Tanimoto distance compound clustering was performed to rapidly find compound pairs , if available , within the leads . OpenBabel 2 . 4 . 1 [43] was used to export molecule MDLs from SMILES format , available from GSK chembox summary . For Tanimoto clustering , the indexes were calculated using ChemFP 1 . 3 [44] with ob2fps bindings and simsearch -NxN as parameter . ChemFP results were parsed and analyzed using an ad hoc perl script , setting the distance ( D ) between compounds as D = 1—Tindex . The distance matrix was built using melt and acast from R Data table package [45] . To assess the MCS clustering , all compounds were imported into a R script using Chemminer [46] and further analyzed using fmcsR [47] for batch MCS calculations . For the molecular feature clustering , a perl script was built to run XlogP3 v3 . 2 . 2 [48] through all lead compounds . Features used to build distance matrix , along with their corresponding values , can be found in Table 4 . All clustering plots were achieved using the R base hierarchical clustering tool , hclust . To find ZBGs among lead compounds , a curated database of such chemotypes was first created ( Table B in S1 Text ) . Structures were drawn using Marvin Sketcher ( Chemaxon ) and exported to SMILES format . This database was then imported to R and processed similarly to the MCS clustering , though instead of calculating overlapping scores between compounds , the overlapping score was determined for each compound against all ZBGs in the database . Only those compound-ZBG pairs where overlap was complete ( score = 1 and , hence , ZBG completely contained in the lead compound ) were counted as a match . | In recent years , the pharmaceutical company GlaxoSmithKline announced the disclosure of small collections of antiparasitic compounds to facilitate research and drug development for three of the main Tropical Neglected Diseases- i . e . Human African Trypanosomiasis , Leishmaniasis and Chagas Disease . These collections include new chemical entities with potential novel mechanisms of action that are likely to be active against a wide variety of targets . Taking advantage of these open access molecules , we successfully set up medium-throughput screening assays to find the first inhibitors of two metallocarboxypeptidases of the M32 family , a group of proteolytic enzymes proposed to play several roles in the biology of trypanosomatids including peptide catabolism , maintenance of parasite adaptive fitness and hydrolysis of bioactive peptides from the human host . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"chemical",
"bonding",
"medicine",
"and",
"health",
"sciences",
"enzymes",
"tropical",
"diseases",
"enzymology",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"fluorophotometry",
"trypanosoma",
"brucei",
"protozoans",
"neglected",
"tropical",
"diseases",
"enzyme",
"... | 2019 | Potent and selective inhibitors for M32 metallocarboxypeptidases identified from high-throughput screening of anti-kinetoplastid chemical boxes |
Delineating the strategies by which cells contend with combinatorial changing environments is crucial for understanding cellular regulatory organization . When presented with two carbon sources , microorganisms first consume the carbon substrate that supports the highest growth rate ( e . g . , glucose ) and then switch to the secondary carbon source ( e . g . , galactose ) , a paradigm known as the Monod model . Sequential sugar utilization has been attributed to transcriptional repression of the secondary metabolic pathway , followed by activation of this pathway upon depletion of the preferred carbon source . In this work , we demonstrate that although Saccharomyces cerevisiae cells consume glucose before galactose , the galactose regulatory pathway is activated in a fraction of the cell population hours before glucose is fully consumed . This early activation reduces the time required for the population to transition between the two metabolic programs and provides a fitness advantage that might be crucial in competitive environments .
Microbial cells are continuously bombarded by diverse and changing combinatorial environmental stimuli . To survive and reproduce , a cell must accurately detect , assess , and selectively respond to these signals . Specifically , in competitive and unpredictable environments , cells need to constantly integrate information about the nature and quantities of nutritional substrates to scavenge maximum nutritional value [1] . Organisms that can balance the anticipation of future environmental shifts without sacrificing the rate of reproduction by excess metabolic burden exhibit a fitness advantage . However , metabolic strategies for achieving this balance have not been thoroughly explored . Studies of the response of microbial cells to the availability of multiple sugars has a long history , starting with the seminal work of Dienert in yeast [2 , 3] and Monod in bacteria [4 , 5] . When presented with both glucose and galactose , microbial cells consume these carbon substrates in a sequential manner rather than simultaneously metabolizing both , resulting in two separate growth phases [5] . In the first phase , cells preferentially metabolize the sugar on which they can grow the fastest ( glucose , in this case ) . Upon glucose depletion , cells transition to metabolizing the less preferred sugar after a “lag phase” in growth corresponding to the time needed to produce the necessary enzymes . Therefore , this response , classically known as “catabolite repression” , posits that the synthesis of the enzymes needed to metabolize the less preferred sugar is inhibited across the whole population . This inhibition is relieved by depletion of the preferred sugar , which triggers the diauxic shift . Crucially , in this model , the sequential consumption of the two sugars is generally attributed to the sequential expression of the enzymes needed for their metabolism [6] . Previous work has discussed regulatory strategies of microbial populations that constitute variations of the classical Monod model of uniform catabolite repression , and demonstrated that these strategies can facilitate adaptation to environmental change . Specifically , evolutionary tuning of the duration of the lag phase has been shown to be a crucial variable for fitness of microbial populations in fluctuating environments [7 , 8] . For example , heterogeneity in the expression of the Lac operon in Escherichia coli has recently been shown to modify the growth rates of single cells during the transition from glucose to lactose metabolism [9] . Furthermore , evolution of E . coli in environments with mixtures of carbon sources has been shown to trigger genetic mutations that produce phenotypic population diversification as a consequence of key trade-offs in carbohydrate metabolism [10] . Similarly , yeast cells that were evolved in cycling maltose and glucose environments displayed reduced catabolite repression and enhanced cell-to-cell variability in the gene expression of the maltose pathway ( MAL ) in mixed environments with maltose and glucose [8] . Interestingly , the evolved strains that displayed salient heterogeneity in MAL gene expression across the population also had a lower growth rate on glucose and a reduced lag phase upon shift to maltose media . Galactose is a less preferred sugar than glucose for the yeast Saccharomyces cerevisiae , and its metabolism requires activation of the galactose gene regulatory network ( GAL ) . The GAL pathway is a well-studied circuit that includes a set of regulatory genes ( Gal1p , Gal2p , Gal3p , Gal80p , and Gal4p ) for sensing and signaling and enzymatic genes ( Gal1p , Gal7p , and Gal10p ) that transform galactose into glucose-6-phosphate as an entry point for glycolysis ( S1 Fig . ) . Previous studies have shown that the GAL pathway can exhibit two stable gene expression states ( ON and OFF ) in a uniform environment in response to intermediate concentrations of galactose [11–14] , and this bistability was demonstrated to be dependent on the Gal1p and Gal3p positive feedback loops [12 , 14] . However , the gene expression dynamics of the GAL pathway in response to combinatorial glucose and galactose inputs and the relationship between the gene expression response and the diauxic shift have not yet been carefully characterized . In this work , we demonstrate that while S . cerevisiae cells indeed undergo sequential sugar consumption in the presence of combinations of glucose and galactose , the synthesis of the enzymes needed for the metabolism of galactose is not necessarily sequential . Specifically , we find that for a large combinatorial space of glucose-galactose inputs , a subpopulation of cells arises where the galactose transcriptional program is induced hours before the depletion of glucose . We demonstrate that this heterogeneous strategy is beneficial for rapid growth during the metabolic transition from glucose to galactose . These data suggest that the response of microorganisms to combinatorial environments may frequently involve diversification of phenotypes across a population . Furthermore , this strategy integrates direct environmental sensing with an anticipation of future environmental shifts . As such , it constitutes an elaboration on bet-hedging mechanisms , which often rely on stochastic fluctuations to produce subpopulations with different phenotypes without a dominant input from the environment [15] .
We studied the time-resolved response of a population of yeast cells to combinatorial inputs of glucose and galactose using an automated flow cytometry setup that measures gene expression approximately every 20 min for 14 h ( Fig . 1A ) [16] . This technology enabled us to measure the galactose ( GAL ) pathway activity dynamics in single cells using the epimerase GAL10 promoter ( pGAL10 ) driving Venus ( Yellow Fluorescent Protein , YFP ) and to dissect the quantitative growth patterns of the microbial culture . The culture was maintained in exponential phase prior to transfer to a microtiter plate . The microtiter plate was inoculated at low cell density and diluted every 20 min for 3–6 h with fresh media while shaking at 30°C ( see Materials and Methods ) . This period of growth and dilution before induction with the mixed sugar stimulus ensured that the growth history of the culture did not significantly influence the response of the cell population ( S2 Fig . ) . For sufficiently low glucose concentrations , pGAL10 induced as a single monomodal distribution . By contrast , pGAL10 did not activate over the course of the experiment for glucose concentrations significantly higher than those of galactose . These behaviors recapitulate previously observed phenotypes [17 , 18] . However , for a large spectrum of combinatorial glucose-galactose inputs aggregating around the regime of equal concentration of these two sugars , we observed the emergence of a bimodal gene expression response in which only a fraction of the population induced pGAL10 ( Fig . 1A , B ) . A single time point measurement has previously observed this bimodality in the GAL system for a mixture of glucose and galactose [11] . However , our dynamic measurements reveal that this bimodality is transient since the cohort of OFF cells uniformly switched ON following a delay . The promoters of the galactokinase GAL1 ( pGAL1 ) , permease transporter GAL2 ( pGAL2 ) and transferase GAL7 ( pGAL7 ) exhibited similar gene expression patterns , indicating that transient bimodality is a general feature of the GAL pathway in response to a mixture of glucose and galactose ( S3 Fig . ) . Our flow cytometry data did not show a significant population of cells with intermediate fluorescence levels for bimodal populations and for a given dual-sugar input , the fraction of GAL ON cells did not change significantly over time in the bimodality region ( highlighted box in S4A , B Fig . ) . Furthermore , previous studies have demonstrated that the GAL system can only exhibit stochastic transitions between states in the absence of the GAL80 negative feedback loop but not in wild type [19] . To test for stochastic switching during the bimodal regime , we sorted cells induced with 0 . 25% glucose and 1% galactose into GAL ON and GAL OFF populations using pGAL10 expression following 4 h of induction ( S5A Fig . ) . We then transferred the sorted subpopulations into one of three environments: ( 1 ) 0 . 25% glucose , ( 2 ) 1% galactose , or ( 3 ) a mixture of 0 . 25% glucose and 1% galactose ( S5B , C Fig . ) . The ON cells gradually turned off over a period of 7 . 3 h in 0 . 25% glucose , and no detectable population of cells switched to the OFF state in either the galactose or the glucose and galactose mixture environment ( S5C Fig . ) . These data strongly suggest that there is no stochastic switching from the GAL ON to the GAL OFF states . The GAL OFF cells remained in an OFF state in the environment with 0 . 25% galactose and activated the GAL pathway over a period of approximately 3 . 5 h in 1% galactose . However , we observed that a fraction of OFF cells induced pGAL10 when transferred into the mixed sugar condition . These cells are likely deriving from an intermediate population prior to the sorting process , since time-lapse microscopy measurements of the gene expression of single cells grown in microfluidic devices ( CellASIC , see Materials and Methods ) did not display stochastic switching between the ON and OFF states for a period of 6 h when induced with 0 . 25% glucose and 1% galactose and then switched to an environment containing 0 . 1% glucose and 1% galactose ( n = 8 OFF or ON colonies shown in S6A , B Fig . ) . However , further experiments are needed to categorically rule out the possibility of stochastic switching from the GAL OFF to the GAL ON state before glucose depletion . Using a Gaussian mixture model ( GMM ) to deconvolve the two populations ( see Materials and Methods ) , we quantified three measures of the response: the time to early activation for conditions with a detectable early activated population ( δa ) , the delay between early and late activation for conditions with transient bimodality ( δg , highlighted panels ) , and the fraction of ON cells quantified at the midpoint between the half-max of the early and delayed activation responses ( FON-mid ) ( Fig . 1B , see Methods ) . δa was modestly increased by glucose and reduced by galactose ( Fig . 1C ) . By contrast , δg , showed a substantial linear increase as a function of initial glucose ( highlighted panels in Fig . 1A and Fig . 1D ) . However , δg was not significantly modified by the initial galactose concentration ( S7A Fig . ) . FON-mid significantly increased with the initial galactose level and was reduced by the initial glucose concentration for any given concentration of galactose ( Fig . 1E ) . δg and FON-mid were modified in a set of mutants including regulators of the GAL pathway and glucose repression , suggesting that these phenotypes are modulated by a complex molecular program involving many factors ( S1 Text and S8A–D Fig . ) . To probe the dependence of this phenomenon on growth history , we compared the response of populations from stationary and exponential phase to a range of glucose and galactose inputs . Both conditions displayed a transient bimodal response , but stationary phase cells displayed a higher fraction of ON cells than cells deriving from exponential growth phase for any given sugar combination ( S9 Fig . ) . Bimodality was also present in populations with different carbon source histories including YP alone , YP + 3% glycerol and YP + 3% ethanol . However , FON was larger for cells grown in glycerol and ethanol than cells grown in YP alone ( S10 Fig . ) . These data indicate that transient diversification of the population into two expression states in response to dual-sugar inputs is a robust phenotype in our strain that persists for a range of growth histories . The quantitative properties of the transient bimodal response are , however , fine-tuned by population history . The existence of a subpopulation of cells in which the galactose transcriptional pathway was active in the transient bimodality regime suggested that the population might be consuming galactose concurrently with glucose . To test this hypothesis , we measured glucose , galactose , and the fraction of ON cells ( FON ) as a function of time in response to 0 . 1% glucose and 0 . 1% galactose , a condition in which the population exhibits a bimodal response ( see Materials and Methods ) . Our data recapitulated the known sequential order of sugar utilization , with glucose being depleted before appreciable galactose consumption ( Fig . 2A ) . Interestingly , FON scored by pGAL10 , pGAL1 , pGAL2 or pGAL7 increased immediately following the dual-sugar stimulus and transiently plateaued before the cells consumed the available glucose ( Figs . 2A , B and S3 and S4 ) . The initial concentration of glucose determined the duration of this plateau ( S4C Fig . ) . FON underwent a second increase to approximately 100% precisely at the time of total glucose depletion . Therefore , the timing of the delayed activation of the repressed subpopulation , and consequently the magnitude of δg , were determined by the time of glucose depletion . In agreement with this hypothesis , δg was inversely related to the initial population size N0 , which modifies the rate of sugar consumption ( S7B , C Fig . ) . In addition , a population that received a first step of glucose and galactose , followed by an additional step input of glucose after 5 h , had a significantly larger δg than a population that received only the initial dual-sugar input , further corroborating the fact that δg is tuned by the concentration of glucose ( S7D , E Fig . ) . In contrast to δg , FON-mid was approximately equal in conditions that received one or two steps of glucose , suggesting that the second glucose input did not induce substantial switching between OFF and ON states ( S7F Fig . ) . Our results indicate that galactose was not significantly consumed for hours despite the presence of a substantial subpopulation of GAL ON cells , making glucose the major contributor to biomass production up to the point of glucose depletion . Furthermore , during the bimodal epoch following an input of 0 . 1% glucose and 0 . 1% galactose , only approximately 2 . 7% of galactose was consumed when 78% of glucose was depleted ( Fig . 2D ) . The expected amount of galactose consumption by the population of GAL ON cells is significantly higher if these cells were fully consuming galactose based on a culture that received only 0 . 1% galactose ( S11A–C Fig . ) . This lack of significant galactose consumption could be due to failure of galactose to permeate into the cell or a different mechanism by which inhibition of galactose metabolism does not require transcriptional repression of the GAL genes . Gal1p and Gal3p are the only known sensors of galactose and these proteins function intracellularly , suggesting that a sufficient amount of galactose was entering the cells to induce pathway activation [20] . Furthermore , we observed a significant induction of pGAL2 , accumulation of the fluorescently tagged Gal2 permease in the activated subpopulation , localization of Gal2p-Venus to the membrane , and strong correlation between a Gal2 fluorescent protein fusion and pGAL10 in the presence of mixtures of glucose and galactose ( S3 and S12 Figs . ) . We also found that level of Gal2p was not limiting for the activation of the GAL pathway using a TET inducible promoter to vary the concentration of Gal2p in a strain deleted for the endogenous GAL2 gene . In this strain , we assessed the fraction of GAL ON cells ( as quantified using a pGAL10-Venus reporter ) in response to simultaneous addition of 0 . 5% galactose and a range of glucose levels ( S12D Fig . ) . We did not observe a significant dependence of the fraction of ON cells on aTc concentration , and hence on Gal2p levels . These data are consistent with the previous observation that the inhibition of galactose consumption in response to a glucose pulse occurs on a timescale faster than can be explained by changes in transcriptional regulation or protein degradation [21] . To understand how the structure of the GAL regulatory network could generate the observed transient bimodality in response to dual-sugar inputs , we constructed a simplified mathematical model of this circuit based on canonical knowledge about the galactose system ( equations are described in the S1 Text ) . In our model , the galactose input activates the signal transducer Gal1p ( G1 ) forming G1* , which inhibits the repressor Gal80p ( G80 ) from sequestering the transcriptional activator Gal4p ( G4 ) , thus leading to GAL gene activation ( Fig . 3A ) . The inhibition of G80 liberates G4 to induce expression of G1 and G80 , establishing a positive and negative feedback loop . Since glucose has been shown to reduce the activity of the GAL system , we coupled this model to an input of glucose [22] . We modeled GAL repression by glucose assuming that a repressor R ( such as Mig1 ) , can be activated by the glucose signal forming R* , which can then repress the promoters of GAL1 and GAL4 . R links the glucose input to GAL repression , an additional feature to several previous models of the GAL system [12–14] that is well supported by literature [22–25] and follows a previous study that similarly simplified the complexity of the glucose repression network into a single module [26] . The GAL network has been shown to exhibit memory of galactose and glucose exposure , suggesting bistability as the source of bimodality in this system [14 , 27] . A salient qualitative feature of our mathematical model is that it can undergo a bifurcation from monostability to bistability as a function of its two inputs: glucose and galactose ( Fig . 3B ) . In response to low-glucose and high-galactose inputs , the model exhibits one steady-state , corresponding to the experimental ON state ( high total G1 levels ) . For high glucose and low galactose inputs , the only steady-state corresponds to the OFF state ( low total G1 levels ) . Similar concentrations of the two inputs produce two stable steady-states that correspond to the bimodality observed in the experiments . We also considered the possibility that glucose influences the GAL pathway through modulation of the growth rate , leading to different dilution rates of the proteins in the GAL network . However , such a model lacking an explicit repressor could not robustly recapitulate our data ( S13 Fig . , and equations are described in the S1 Text ) . The glucose repressor model predicts the emergence and disappearance of bistability in the GAL system as a function of time for a given dual sugar input . By assuming that the system traverses a series of quasi–steady-states as a function of decaying sugar concentration ( highlighted panels in Fig . 1 ) , a given model trajectory crosses through a region of bistability , which is then transformed to monostability as glucose drops below a critical threshold ( bifurcation point ) due to cellular consumption ( representative trajectory in Fig . 3B ) . This transition from bistable to monostable behavior at the glucose bifurcation point corresponds to the synchronized delayed activation of the repressed cohort of cells . The observed monomodal activation for sufficiently low glucose concentrations ( left of highlighted panels in Fig . 1B ) and significantly delayed activation ( right of highlighted panels in Fig . 1B ) are also explained by the model ( S14A , B Fig . ) . Furthermore , the model indicates that if the glucose concentration were maintained above its value at the bifurcation point , for example by replenishment of glucose ( Fig . 3C ) , then the window of time where bistability exists in the system would be extended . This is precisely the case since cultures that received an initial pulse of glucose and galactose followed by a second pulse of glucose exhibited bimodality for a longer period of time compared to a culture that received only the initial sugar mixture ( S7D , E Fig . ) . In addition to explaining a potential origin of transient bimodality , the model made predictions about different features of the system . First , the model had predictions about the role of feedback loops . Removing the GAL80 feedback loop in the model augmented the range of glucose and galactose concentrations that produced bistability . This prediction was qualitatively consistent with our data showing that the range of glucose and galactose inputs that produced experimental bimodality was expanded in a strain lacking the Gal80p feedback loop ( S15 Fig . ) . We also predicted using the model that irrespective of the identity of the negative regulator R , a reduction in repression strength by decreasing the binding affinity of R to the promoters of G1 and G4 would shift and contract the window of bistability ( S14C , D Fig . ) . Consistent with this expectation , a mutant of the glucose-dependent negative regulator Mig1 ( representing a component of R in the model ) which has a reduced affinity for the Cyc8-Tup1 complex and therefore a diminished capability for glucose repression exhibited a smaller δg than wild type ( S8B-1 Fig . ) . Halving the dosage of the Cyc8-Tup1 repression complex similarly exhibited a smaller δg than wild type ( S8B-2 Fig . ) [28] . The bifurcation hypothesis provided by the model also implied that the amount of time required for glucose to decrease to a threshold concentration corresponding to the bifurcation point in the system ( δb ) should decrease if galactose is added at different times following the glucose input ( rather than concomitantly with glucose ) ( S3C Fig . ) . This increasing delay in the galactose input signifies a decreasing glucose concentration in the culture due to cellular consumption at the time of galactose addition , and hence a reduced window of time for bistability . The delayed activation response corresponds to the loss of bistability as glucose crosses a threshold bifurcation point . Hence δg and δb reflect similar properties of the system . To experimentally test this prediction , we applied a step input of 0 . 1% galactose at different times to a set of cultures that had all received 0 . 1% glucose from time zero . Galactose was added at 0 , 3 . 1 , 4 . 2 , 5 . 3 , and 6 . 3 h following the glucose stimulus to different cultures ( arrows in Fig . 3D ) . Matching the trend of decreasing δb in the model ( Fig . 3C ) , bimodality emerged at the time of the galactose input and δg contracted and eventually disappeared with the increased delay in this input ( right panel in Fig . 3D and S16 Fig . ) . Finally , the model indicated that the response time of the system to transition from the OFF to the ON state decreases as glucose decays . The system’s response time is dictated by both the domain of attraction and the magnitude of the dominant eigenvalue of the ON steady-state ( S1 Text ) , which both increase as glucose decreases ( Fig . 3E ) . Therefore , the response time of the fraction of ON cells should decrease in the delayed galactose experiment . Corroborating this insight , our experimental data demonstrated a decrease in the response time of FON with the increase in the delay of the galactose input ( Fig . 3F , G ) . Therefore , our model provides a framework that explains the transition of the GAL system between different phenotypic modes and demonstrates the modulation of the quantitative properties of this network by its environmental inputs . We next probed the physiological impact of the observed anticipatory induction of the GAL regulatory program in advance of galactose consumption by analyzing the relationship between the timing of GAL pathway activation and the population’s growth rate and metabolism . To do so , we quantified the concentrations of glucose , galactose , and growth rates for the different cultures that were subjected to delayed galactose inputs in the experiment described above . Glucose decayed at a similar rate irrespective of when galactose was added ( Fig . 4A ) . By contrast , galactose consumption was delayed for cultures that received galactose at 3 . 1 , 4 . 2 , and 5 . 3 h compared to the culture that received glucose and galactose simultaneously at time zero despite the fact that not all glucose was depleted at the time of galactose addition ( Fig . 4B , C ) . Furthermore , the delay in galactose consumption was increased commensurately with the delay in galactose administration . Notably , during the metabolic shift between carbon sources , the population that received galactose simultaneously with glucose exhibited a transient growth rate advantage , reaching approximately 15% compared to the population that received this sugar after a 6 . 3 h delay ( Fig . 4D and S17 Fig . ) . Since the growth rate is proportional to the current size of the population in exponential phase , the significance of this fitness difference increases with each cell generation . Overall , the delay in galactose input caused a monotonic increase in the transient growth defect , which was manifested as an increase in the “lag” time between the two phases of growth documented by Monod ( S17A Fig . ) [4] . Growth rate measurements indicate that the presence of galactose did not benefit the population of cells until total glucose depletion ( S17 and S18 Figs . ) . Taken together , these data indicate that the induction of the GAL pathway many cell generations before these genes are required provides a transient fitness advantage during the shift between carbon sources . This beneficial pre-emptive induction of the GAL pathway genes only occurs in a subset of the population . To investigate the tradeoffs that might motivate this bimodal induction , versus a uniform strategy in which all the cells in the population pre-emptively but coherently induce the GAL pathway , we sought to control GAL gene expression independently of galactose . To do so , we used an estradiol inducible Gal4 chimera in a strain lacking endogenous Gal4p [29 , 30] . In this strain , we could activate GAL gene expression on demand at specific times before glucose depletion in cultures subjected to a mixture of 0 . 1% glucose and 0 . 1% galactose at time zero ( S19A Fig . ) . Since the synthetic inducible system is not connected to the feedback structure of the natural circuit , GAL gene expression was monomodal ( graded ) as opposed to bimodal . In this strain , early activation of the GAL pathway generated a lower consumption rate of glucose compared to late activation , therefore revealing that constitutive GAL gene expression can inhibit glucose consumption ( S19B Fig . ) . The expression level of pGAL10 induced by this synthetic system was very similar to the expression level of pGAL10 in the wild type ( S19A Fig . ) . Therefore , this effect is not likely to be a consequence of over- or under-expression of the GAL genes . Constitutive induction of the GAL pathway through over-expression of Gal3p also reduced the glucose consumption rate ( S20 Fig . and S1 Text ) . Together , these data highlight an important tradeoff that the system has to balance: induction of the GAL genes before they are required results in faster galactose consumption upon glucose depletion , facilitating the transition between carbon substrates . At the same time , wholesale induction of these genes across the entire population comes at the cost of a reduced rate of glucose consumption . In agreement with this observation , the GAL repressed subpopulation in the wild type had approximately 15% faster growth rate on average than that of the activated subpopulation in the transient bimodal region ( S21 Fig . and S1 Text ) . However , in this bimodal regime , the glucose consumption rate of the whole population was not saliently reduced by GAL gene expression in a subpopulation of cells ( Fig . 4A ) . To further highlight the tradeoffs of the bimodal strategy adopted by the population in combinatorial carbon environments , we tracked the growth and gene expression of 13 single ON or OFF colonies over time induced with 0 . 25% glucose and 1% galactose using time-lapse microscopy in microfluidic chambers ( CellASIC , see Materials and Methods ) . In a first experiment , cells were grown for 3 h in 0 . 25% glucose and 1% galactose and then switched to 1% galactose ( S22 Fig . ) . The GAL ON cells grew faster following the abrupt switch to galactose media than the OFF cells , corroborating our flow cytometry results that demonstrated that pre-induction of the GAL genes hours prior to the metabolic transition provides a fitness advantage . During the initial 3 h in the combinatorial environment prior to the switch to galactose media , we could not detect a significant difference in the growth rates of the OFF and ON cells , likely because this period of time , which spans approximately two cell doublings , cannot reveal such difference . Instead , we hypothesized that a longer experiment and graded reduction in the glucose concentration is necessary to detect a difference in the subpopulation growth rates . To test this notion , we exposed a culture to a mixture of 0 . 25% glucose and 1% galactose for 3 h followed by a mixture of 0 . 1% glucose and 1% galactose for an additional 3 h ( S6 Fig . ) . In the lower glucose environment , the OFF cells grew significantly faster than the ON cells demonstrating the fitness cost imposed on cells by the induction of their GAL program in an environment where these genes are not required ( S6C Fig . ) .
In this work , we demonstrate that a combinatorial input of glucose and galactose triggers diverse regulatory states across a population of cells . This transient bimodality establishes the co-existence of two subpopulations of cells—one that prepares hours in advance for a future shift in carbon metabolism and a second that defers pathway activation over many cell generations until these genes are required upon glucose depletion . The fraction of cells that occupies each state is tuned by the dual-sugar mixture , standing in contrast to canonical models in which the output of a pathway is proportionally matched to the level of its inputs in all cells of the population [31 , 32] . Anticipatory responses to environmental change have been documented in a number of cellular systems [33 , 34] . For example , correlations between heat shock and low oxygen in the human gut are thought to cause E . coli to trigger both responses in the presence of any single one of these environmental cues [33] . Here , we show anticipatory induction of galactose utilization genes many hours before glucose depletion . A new study appearing in the same issue documents that early induction of the GAL pathway in the presence of both glucose and galactose is a general response of many S . cerevisiae isolates [35] . Consistent with our data showing that early synthetic induction of the GAL genes can shorten the lag phase upon glucose depletion , the Wang et al . study further confirms that in these strains , the length of the diauxic lag is strongly determined by how early the galactose pathway is induced . Similar conclusions were reached in a study of the maltose pathway in yeast [8] , indicating that preemptive induction of metabolic programs might be a general strategy that microorganisms use to enhance their preparedness for depletion of a preferred nutrient in mixed-nutrient environments . A second feature in our data is that for many combinations of mixed glucose and galactose concentrations , only a fraction of the population undergoes early induction of the GAL program . This fraction is further dependent on the concentration of the sugars . The GAL ON subpopulation grows slower than the GAL OFF subpopulation in the presence of glucose , suggesting that while early induction of the GAL program offers a delayed benefit by shortening the lag phase upon glucose depletion , it also incurs an immediate cost when glucose is present . Our data indicate that this cost is partially the result of a slowdown in glucose consumption . Similarly , the natural isolate with the earliest GAL induction ( and shorter lag phase ) in the Wang et al . study was found to exhaust glucose slower than the strain with delayed GAL induction ( and longer lag phase ) [35] . These observations point to a tradeoff in which populations balance loss of immediate fitness with future benefit . There are two broad strategies by which organisms can strike this balance to adapt to the environmental statistics of their particular niche . First , they can tune the parameters of the early activation of the GAL program in the whole population , delaying or accelerating its time and/or its induction kinetics . Alternatively , as we show in this paper , they can implement a stochastic diversification strategy , akin to bet-hedging [15 , 36 , 37] , whereby only a fraction of the population undergoes early GAL gene activation , therefore bearing the immediate cost in the presence of glucose and reaping its benefits upon glucose depletion . However , this fraction ( and hence the cost and benefit balance ) is dependent on the state of the environment ( the concentration of glucose and galactose ) . As a result , the strategy can be viewed as an elaboration of bet-hedging resembling the proposed mechanism of “stochastic sensing” [1 , 38 , 39] . Intriguingly , the different natural isolates probed in the Wang et al . study exhibit an array of population behaviors , ranging from strains exhibiting early unimodal induction of the GAL program to others showing transient bimodality in gene expression across the population [35] . Furthermore , evolved yeast cells after cycling between maltose and glucose inputs exhibited an array of behaviors including “specialist” mutants that have higher fitness under the glucose-only condition at the expense of a longer lag phase , and “generalist” mutants with a shorter lag phase and increased fitness during transition between carbon sources , but at the expense of decreased fitness during growth on glucose [8] . Interestingly , some of the evolved strains displayed a bimodal expression of the MalS gene across the population , although this phenotype was not dynamically characterized . Therefore , we postulate that while early GAL gene induction is likely to be a general anticipatory strategy of glucose depletion , the timing of induction and degree of heterogeneity in gene expression across the population varies among different strains , possibly as a result of distinct selective pressures and environmental statistics [40] . We observed that GAL ON cells did not appreciably metabolize galactose until glucose depletion . Although our data do not pinpoint the exact mechanism by which galactose metabolism might be inhibited in GAL ON cells , several possibilities exist . Previous studies examining this issue to produce biotechnologically efficient yeast strains documented that glucose and galactose are consumed simultaneously in cells lacking the glucose kinase Hxk2p [41] . Furthermore , in S . cerevisiae , glucose was shown to block the maltose pathway ( MAL ) by a novel mechanism at the signaling level , which is also linked to Hxk2p [42] . Finally , all but one of the evolved “generalist” strains with the shorter lag phases in the New et al . study harbored mutations in HXK2 , which led to reduced catabolite repression of the MAL genes [8] . It is therefore possible that inhibition by glucose proceeding though HXK2 can have both transcriptional ( catabolite repression ) and post-translational inhibitory components , and is general to both the GAL and MAL pathways . If this were the mechanism at play , then GAL gene expression and galactose metabolism would be at least partially decoupled in the cell . However , other mechanisms might also explain our observations . For example , our data in Fig . 2 show that the bimodal regime following a dual-sugar input is delayed in pGAL1 and pGAL10 versus pGAL2 and pGAL7 . Therefore , it is possible that temporal ordering in the expression of the GAL genes provides a timer-like mechanism for galactose consumption . An intriguing outcome of this model is that different yeast strains or species can modulate this timing based on their niche requirement , hence possibly consuming galactose while glucose is still available . However , further studies are needed to unravel the intricate molecular mechanisms that shape the coupling of gene expression and metabolic decisions of cells and the fitness trade-offs involved in specialization to a single environment versus rapid adaptation to a shift in environmental conditions . It will be fascinating to systematically dissect the combinatorial space of different regulatory and metabolic strategies in other S . cerevisae strains and different microorganisms and link their adoption to specific challenges for survival and reproduction in complex competitive environments .
All yeast strains were derived from W303 . All plasmids used in this study were derived from a set of yeast single integration vectors containing selectable markers and targeting sequences for the LEU2 , HIS3 , TRP1 , and URA3 loci . These vectors were linearized by digestion with PmeI and transformed using standard yeast transformation techniques . The sequences for the GAL1 , GAL2 , GAL7 , GAL10 , MIG1 , and ADH1 promoters were 646 , 743 , 729 , 657 , and 658 bp upstream from the start codons for these genes , respectively . The genotypes for these strains are listed in S3 Table . Several strains used in this study have been described previously [14] . Cells were grown at 30°C for all experiments . For all experiments except microscopy , cells were grown overnight in yeast peptone media ( YP ) containing 10 g/L yeast extract and 20 g/L peptone for approximately 12 h . This culture was diluted to an optical density ( OD600 ) of approximately 0 . 05–0 . 1 and grown for an additional 5–6 h in a 10 ml volume ( final OD600 ranged between 0 . 3–0 . 7 ) . This single well-mixed culture was then diluted appropriately with fresh YP media into a 96-well microtiter plate to a final OD600 of less than 0 . 02 ( 500 μL volume ) . Cells were grown in the 96-well plate and diluted every 20 min by a liquid-handling robot with YP media for 2–4 h prior to induction with glucose and galactose for all experiments except the direct comparison of stationary and exponential phase cultures in S9 Fig . [16] . This growth and dilution period further ensured that the cell population was in exponential phase at the time of the sugar stimulus . For the experiments shown in S9 Fig . , stationary phase cultures were grown for 26 h without dilution until saturation due to nutrient depletion in the media . Exponential phase cells were grown and diluted as explained above . An outgrowth period was not performed for this experiment . Further details about the flow cytometry automation can be found in [16] . Single-cell fluorescence was measured on a LSRII analyzer ( BD Biosciences ) . A blue ( 488 nm ) laser was used to excite YFP and emission was detected using a 530/30 nm filter . For each measurement , 1 , 000–20 , 000 cells were collected . As described previously [16] , a 500 µl culture volume was used in 96-well plate format for the automated flow cytometry measurements . For step-response experiments , a 30 µl sample was removed from the culture for measurement on the cytometer at each time point and 30 µl of fresh YP media containing the appropriate 1X concentration of glucose and galactose was used to maintain a constant culture volume . The sugar pulse experiments shown in S2B , S9 and S10 Figs . were performed by the addition of a range of glucose and galactose concentrations at the beginning of the experiment , and 30 µl of fresh YP media lacking the two sugars was added at each time point . A 60 ml culture volume was used for the flask experiments in which the sugar concentrations were quantified . Less than 5% of the total volume was removed over the course of the experiment to quantify the single cell fluorescence , sugar concentrations , and absorbance at 600 nm ( OD600 ) . OD600 was measured on a Nanodrop 2000c spectrophotometer ( Thermo Scientific ) . The ratio of YFP fluorescence to side scatter was used to quantify the total fluorescence per cell . Flow cytometry distributions were analyzed using a Gaussian mixture model algorithm ( GMM , MATLAB ) and each distribution was classified as either unimodal or bimodal , as described in [10] . The delay time δa was computed as the time required to reach the half-max of the mean of the activated subpopulation . δg was defined as the difference between the time required to reach the half-max of the mean of the activated and repressed subpopulations . The fraction of ON cells ( FON ) was computed as the fraction of the cell population higher than a fluorescence threshold ( 10−0 . 2 a . u . ) that corresponds to approximately the lowest density of single-cell fluorescence between the OFF and ON expression states . FON-mid was quantified at the midpoint between the half-max of the activated ( δa ) and repressed subpopulations . The response time was defined as the time to reach the half-max of FON ( FON = 0 . 5 ) . At each time point , individual cells were assigned to the OFF and ON states using the FON threshold on gene expression described above . The subpopulation growth rates were computed as the slope of a line fit to the log2 of the number of cells that accumulated in the OFF and ON states over time . Glucose and galactose were measured using the Amplex Red glucose oxidase and galactose oxidase kits ( Molecular Probes , Life Technologies ) . A Safire II plate reader ( Tecan ) was used to quantify the fluorescence . A standard of known concentration for each sugar was used to determine the quantitative relationship between the fluorescence and sugar concentration . The maximum concentration of sugar was in the linear range as shown in S11 Fig . Time-lapse microscopy was performed using the Y04C plates from CellASIC ONIX microfluidic platform ( EMD Millipore ) . For microscopy experiments , cells were grown overnight in minimal media containing all amino acids , yeast nitrogen base and 2% filter-sterilized raffinose . In the morning , cells were diluted to an OD600 of 0 . 1 in minimal media containing 1% galactose , 0 . 25% glucose and grown for 4 h prior to inoculation of the microfluidic chamber . For the diauxic shift experiment ( S22 Fig . ) , the media also contained 2% raffinose . The mid-log phase cells ( final OD600 was approximately 0 . 3–0 . 5 ) were sonicated briefly , centrifuged for 2 min at 2 , 000 rpm and then resuspended in a smaller volume of minimal media containing 1% galactose and 0 . 5% glucose such that the OD600 was equal to 0 . 8–1 . 5 . 1µl of the concentrated cells was loaded into the Y04C microfluidic chamber and 200 µl of minimal media containing glucose and/or galactose was loaded into the flow wells . The chambers were loaded with 8 psi and media continuously flowed at 5 psi over the course of the experiment . Cells were imaged every 15 min at 30°C on a Nikon Ti-E equipped with a Perfect Focus System and a Coolsnap HQ2 CCD camera ( Photometrics ) located at the Nikon Imaging Center ( UCSF ) . Imaging was performed using a Plan Apo 20x/0 . 25 and Plan Apo 40x/0 . 25 objectives . Wild-type cells were grown in YP media overnight . In the morning , the saturated culture was diluted to an OD600 of 0 . 1 and induced with 1% galactose and 0 . 25% glucose for 4 h at 30°C in YP media . After resuspending the cells in 1X PBS , they were sorted based on pGAL10 fluorescence using a FACS Aria II ( BD Biosciences ) . 3e6 ON cells and 500 , 000 OFF cells were collected . Following the sort , the ON and OFF cells were centrifuged for 10 min at 2000 rpm . These cells were used to inoculate three 10 ml cultures containing 0 . 25% glucose , 1% galactose and 1% galactose and 0 . 25% glucose in YP media and were grown at 30°C . Samples were taken approximately every 40 min ( S5B , C Fig . ) for 7 . 3 h and the fluorescence was measured on an LSRII analyzer ( BD Biosciences ) . For microscopy data , cell budding events over time were analyzed manually . Single-cell fluorescence was quantified in ImageJ using the ROI toolbox to track single cells over multiple frames . Single-cell fluorescence was defined as: integrated density − ( area of selected cell × mean fluorescence of background ) . The population growth rates were defined as the finite difference of the natural logarithm of the OD600 measurement divided by the change in time in hours . We used custom code for mathematical modeling written in MATLAB ( Mathworks ) and Mathematica ( Wolfram Research ) . Details about the model construction are provided in the S1 Text . The parameter values are listed in S1 Table and S2 Table . The domain of attraction of the ON steady state was defined as the fraction of initial conditions that were assimilated by the ON equilibrium point and was determined by randomly sampling 5 , 000 initial conditions using the Latin Hypercube Method [14] . A minimum and maximum bound on the concentration of each species was used based on the parameters of the model . The dominant eigenvalue was defined as the eigenvalue of smallest absolute value of the linearization at the ON equilibrium point . | To survive in resource-limited and dynamic environments , microbial populations implement a diverse repertoire of regulatory strategies . These strategies often rely on anticipating impending environmental shifts , enabling the population to be prepared for a future change in conditions . It has long been known that cells optimize nutritional value from mixtures of carbon sources , for example glucose and galactose , by sequential activation of regulatory programs that allow for metabolizing the preferred carbon source first before metabolizing the secondary carbon source . Using automated flow-cytometry , we mapped the dynamical behavior of populations simultaneously presented with a large panel of different glucose and galactose concentrations . We show that , counter to expectations , in populations presented with glucose and galactose simultaneously , the galactose regulatory pathway is activated in a fraction of the cell population hours before glucose is fully consumed . We demonstrate that the size of this fraction of cells is tuned by the concentration of the two sugars . This population diversification may constitute a tradeoff between the benefit of rapid galactose consumption once glucose is depleted and the cost of expressing the galactose pathway . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Population Diversification in a Yeast Metabolic Program Promotes Anticipation of Environmental Shifts |
Advancing age is the dominant risk factor for most of the major killer diseases in developed countries . Hence , ameliorating the effects of ageing may prevent multiple diseases simultaneously . Drugs licensed for human use against specific diseases have proved to be effective in extending lifespan and healthspan in animal models , suggesting that there is scope for drug repurposing in humans . New bioinformatic methods to identify and prioritise potential anti-ageing compounds for humans are therefore of interest . In this study , we first used drug-protein interaction information , to rank 1 , 147 drugs by their likelihood of targeting ageing-related gene products in humans . Among 19 statistically significant drugs , 6 have already been shown to have pro-longevity properties in animal models ( p < 0 . 001 ) . Using the targets of each drug , we established their association with ageing at multiple levels of biological action including pathways , functions and protein interactions . Finally , combining all the data , we calculated a ranked list of drugs that identified tanespimycin , an inhibitor of HSP-90 , as the top-ranked novel anti-ageing candidate . We experimentally validated the pro-longevity effect of tanespimycin through its HSP-90 target in Caenorhabditis elegans .
Increasing life expectancy in developed countries is revealing advancing age as the primary risk factor for numerous diseases [1] . Thus , identifying interventions that can ameliorate the effects of ageing , and consequently delay , prevent or lessen the severity of age-related conditions , are needed . Extensive research in laboratory animals has demonstrated that the ageing process is malleable and that dietary , genetic and pharmacological interventions can improve health during ageing , extend lifespan and combat pathologies [2] . Furthermore , humans who live to advanced ages show lower late-life morbidity ( disease burden ) than those who die earlier , indicating that compression of morbidity is achievable [3] . Although pharmacological interventions may prove to ameliorate the effects of ageing in humans , development of new drugs for this purpose would present significant difficulties , because of the need to treat healthy individuals in clinical trials over long periods for multiple outcomes . For this reason , it is more feasible to repurpose drugs already approved for specific diseases , or that passed their safety tests but failed against their original indication , than to target ageing itself with new drugs [4 , 5] . With this goal in mind , researchers have begun to conduct human clinical trials to assess the anti-ageing properties of drugs approved to treat human medical conditions , and that extend lifespan and healthspan in animal models . Some examples include the anti-diabetic drugs metformin ( National Clinical Trial ( NTC ) number: NCT02432287 ) [6] and acarbose ( NCT02953093 ) , the immunosuppressant sirolimus ( NCT02874924 ) and related compounds [7 , 8] , and the nutraceutical resveratrol ( NCT01842399 ) . Two natural metabolites , the NAD precursors nicotinamide riboside ( NCT02950441 ) and nicotinamide mononucleotide [9] are also being investigated . The development of computational methods to complement and accelerate this approach , by prioritising approved drugs that could ameliorate human ageing , is needed . Several bioinformatic methods have been developed to identify potential geroprotective drugs . For instance , caloric restriction ( CR ) mimetics have been identified , by comparing genes differentially expressed in rat cells exposed to sera from CR rats and rhesus monkeys with gene expression changes caused by drugs in cancer cell lines [10] . Structural and sequence information on ageing-related proteins have been combined with experimental binding affinity and bioavailability data to rank chemicals by their likelihood of modulating ageing in the worm Caenorhabditis elegans and the fruit fly Drosophila melanogaster [11] . Drug-protein interaction information has also been used to predict novel pro-longevity drugs for C . elegans , by implementing a label propagation algorithm based on a set of effective and ineffective lifespan-extending compounds and a list of ageing-related genes [12] . A similar approach used a random forest algorithm and chemical descriptors of ageing-related compounds from the DrugAge database [13] together with gene ontology ( GO ) terms related to the drug targets [14] . Enrichment of drug targets has been assessed for a set of human orthologues of genes modulating longevity in animal models to identify new anti-ageing candidates [15] . Despite the increasing interest in drug-repurposing for human ageing , research has tended to focus on predicting life-extending drugs for animal models . However , the translation from non-mammalian species to humans is still a challenge , and certain aspects of ageing may be human-specific . Only a few studies have focused on data from humans . For instance , Aliper et al . ( 2016 ) [16] applied the GeroScope algorithm [17] to identify drugs mimicking the signalome of young human subjects based on differential expression of genes in signalling pathways involved in the ageing process . Another study by Dönertas et al . ( 2018 ) [18] correlated a set of genes up- and down-regulated with age in the human brain with drug-mediated gene expression changes in cell lines from the Connectivity Map [19] . In the present study , we rank-ordered drugs according to their probability of affecting ageing , by measuring whether they targeted more genes related to human ageing than expected by chance , by calculating the statistical significance of the overlap between the targets of each drug and a list of human ageing-related genes using a Fisher’s exact test [20] . Additionally , to enhance the power of the approach , we mapped the drugs’ gene targets and ageing-related genes to pathways ( KEGG , Reactome ) , gene ontology terms ( biological processes , cellular components , molecular functions ) and protein-protein interactions , and repeated the analysis . We found that , independently of the data source used , the analysis resulted in a list of drugs significantly enriched for compounds previously shown to extend lifespan in laboratory animals . We integrated the results of 7 ranked lists of drugs , calculated using the different data sources , into a single list , and we experimentally validated the top compound , tanespimycin , an HSP-90 inhibitor , as a novel pro-longevity drug .
The drug-ageing association was inferred by comparing drug-gene interactions with gene-ageing associations . Fig 1 presents an overview of the procedure to prioritise the compounds . A dataset containing the interactions between drugs and proteins was built based on data from the STITCH database [21] . Only drugs targeting human proteins and successfully mapped to the DrugBank database [22] using the UniChem resource [23] were kept ( Fig 1A ) . The dataset was composed of 18 , 393 interactions between 2 , 495 drugs and 2 , 991 proteins . More than half of the drugs ( 51 . 1% ) in the dataset are approved for human use , 18 . 6% are in some phase of the approval process and 28 . 4% have been shown to bind to disease targets in experiments . We obtained a set of ageing-related genes from the Aging Clusters resource [24] . A total of 1 , 216 ageing-related genes discovered in at least 2 among 4 categories of studies were selected . These 4 categories are human genes: i ) changing expression with age or CR in different tissues ii ) whose DNA methylation levels changes with age iii ) associated with age-related diseases and iv ) in manually curated databases of genes linked with longevity in genetic studies [25] , associated with cellular senescence [26] or showing ageing-related effects in animal models in addition to evidence for a causative role in human ageing [27] . We determined if there was evidence supporting an association between drugs and ageing-related genes by calculating the statistical significance of the overlap between the gene targets of each drug and the ageing-related genes ( Fig 1B ) . From the 1 , 147 drugs analysed , 19 were statistically enriched for ageing-related targets after multiple testing correction ( Tables 1 and S1 ) . To assess the capability of the method to prioritise pro-longevity compounds , we compared the list of top-ranked compounds with the DrugAge database [13] . Six out of the 19 drugs have already been reported to significantly extend the lifespan of at least one model organism ( S1 Text ) , while only 1 was expected by chance ( p < 0 . 001 ) . Additionally , using literature mining , we identified studies showing the association with ageing of cAMP analogues [28] , selenium [29 , 30] and tanespimycin [31 , 32] . In contrast , we also found evidence for the DNA-mediated , pro-ageing ( anti-longevity ) effects of doxorubicin [33] , cisplatin [34] and hydrogen peroxide [35] . We performed an interaction-based similarity analysis and found that the genotoxic compounds clustered separately from the other drugs , suggesting that they have a similar mechanism of action ( S1 Text ) . Similarities were also identified regarding the mechanisms of action of sorafenib and regorafenib , bexarotene and GW-501516 , and lastly sirolimus and ECGC , in agreement with a previous study [36] . Although drugs interact directly with proteins , proteins do not act alone and interact with other proteins within pathways to perform different functions . Anti-ageing effects are likely to be mediated through altered pathway activity and function , and we therefore investigated if we could enhance the prediction of pro-longevity drugs using other biological annotations as comparators . Therefore , we calculated the pathways and gene functions enriched in ageing-related genes , together with the proteins that interact with them . A total of 82 KEGG and 54 Reactome pathways were enriched in this set of genes , as well as 1 , 177 biological processes , 69 cellular components and 103 molecular functions . In addition , we calculated that 676 proteins interacted with the set of ageing-related genes . These terms , mapped at different biological levels , were defined as the set of ageing-related terms ( Fig 1C–left ) . Equivalently , drugs were then associated with these terms through association with their targets using the list of genes defining each term according to the DAVID knowledgebase [37] and the biological database network [38] . This mapping procedure resulted in a set of terms from each data source related to each drug ( drug-related terms ) ( Fig 1C–right ) . Analogously to the gene-based association analysis , we calculated for each biological level if the overlap between ageing-related terms and drug-related terms was statistically significant using a Fisher’s exact test . This procedure generated 6 lists of ranked compounds in addition to the gene-based analysis ( S1 Table ) . Notably , when we evaluated the correlation between the ranking of compounds in the different lists ( Fig 2A ) , we observed a moderate correlation ( Kendall’s coefficient of concordance W = 0 . 58 , p-value = 1 . 02E-266 ) . The highest correlations were observed between the results from biological processes and cellular components ( Kendall’s tau = 0 . 51 , p-value < 2 . 2E-16 ) , while the lowest was observed between cellular components and genes ( Kendall’s tau = 0 . 16 , p-value = 3 . 289E-11 ) . Because in any enrichment analysis there is a potential for research bias , we performed random permutations to simulate the enrichment of each drug for a different set of terms at each level . None of the top-ranked drugs in each list ranked higher than in the analysis in more than 1 . 7% of the simulations ( Table A in S1 Text ) . We also quantified the capability of the strategy to prioritise pro-longevity compounds by calculating for each list the fraction of known pro-longevity compounds ( ranked by p-value ) among the fraction of drugs considered in each analysis ( Fig 2B ) . The enrichment for pro-longevity compounds was quantified by calculating the area under the curve ( AUC ) generated by plotting these two variables . The maximum AUC was obtained when biological processes ( AUC = 0 . 69 ) was used as the comparator ( Fig 2 in S1 Text ) . The use of genes showed the lowest enrichment when non-statistically significant drugs are considered ( AUC = 0 . 59 ) , which suggests that the use of higher biological levels to calculate the inference improves the prediction capabilities , and that the use of genes leads to a loss of power to rank drugs targeting a low proportion of ageing-related genes , which is observed in Fig 2B as a loss of enrichment after 25% of the drugs were ranked . We evaluated if the AUCs were statistically significant by calculating the AUC from the simulations generated to quantify the research bias . The p-value for each curve was calculated by determining the number of simulated results with an AUC equal or higher than the analysis . All lists showed a higher enrichment than expected by chance ( AUC > 0 . 5 and p-value < 0 . 05 , Table B in S1 Text ) . When we only considered the first 20 top-ranked drugs , we observed that using biological processes or cellular components to perform the comparison showed the highest proportion of pro-longevity drugs ( 45% ) , while only 2 pro-longevity drugs ( 10% ) were found among the top 20 drugs when KEGG pathways were used . Since our method is based solely on interactions , it should be able to prioritise both pro- and anti-longevity drugs ( as we observed in Table 1 ) . For this reason , and although the number of drugs reported to decrease lifespan in animal models is smaller than the set of pro-longevity drugs , we decided to repeat the enrichment analysis using anti-longevity drugs ( Fig 2C ) . As expected , ( because of the dataset size ) the enrichment for anti-longevity drugs was lower than for pro-longevity drugs ( Fig 2 in S1 Text ) . The highest AUC was observed when cellular components was used ( AUC = 0 . 63 ) , while using genes showed the lowest enrichment for anti-longevity drugs ( AUC = 0 . 54 ) . Because there are various cutoff values that can be selected to define the dataset of drug-protein/protein-protein interactions and enriched GO terms/pathways , we also repeated the analysis using different confidence scores ( for STITCH and STRING ) and p-value cutoff ( for Gene Ontology and Pathways ) , to explore its influence on the performance of the method . In the case of the enrichment for pro-longevity drugs , we do not observe a major change in the AUC when higher or lower confidence scores were used ( Fig 2 in S1 Text ) , while the selection of a lower p-value cutoff leads to the same ( GO:CC and KEGG ) or lower ( GO:MF and Reactome ) enrichment and the use of higher p-value cutoff to a decrease ( GO:MF and Reactome ) or increase ( GO:CC and KEGG ) in the AUC . A similar lack of trend was observed in the enrichment for anti-longevity drugs . In general , the cutoffs that we used initially ( p-value of 0 . 05 and confidence score of 700 ) maximised the enrichment for pro-longevity drugs when genes , Reactome pathways and molecular functions were used . Considering the lack of overlap between the ranked lists using the different data sources , we decided to integrate the results into a single list accounting for the complexity of multitiered effect of drugs by calculating their ranking average in the different analyses . The combination generated a list equally enriched as the maximum AUC obtained by the previous analysis ( AUC = 0 . 69 ) . Among the top 10 drugs with the best average ranking ( Tables 2 and S1 ) , we found 3 drugs that have extended lifespan in animal models ( trichostatin [39] , geldanamycin [10] and celecoxib[40] ) . Half of these 10 drugs are classified as kinase inhibitors , while 8 are indicated as anti-cancer drugs and 7 are approved for human use . Leading the joint ranking was tanespimycin , also known as 17-AAG , a well-characterized HSP-90 inhibitor that has been shown to activate the transcription factor HSF-1 and induce a heat shock response [32] . As a proof-of-principle , we decided to investigate whether tanespimycin could activate HSF-1 and extend lifespan in the nematode worm C . elegans . To test the efficacy of tanespimycin dosing in C . elegans , we grew worms expressing mCherry under the control of an HSF-1 responsive promoter [41] on solid media plates containing various doses of tanespimycin . Worms were exposed to tanespimycin continuously from the first larval stage ( L1 ) of development , or exclusively from the first day of adulthood . Worms grown continuously on tanespimycin plates exhibited a dose-dependent activation of the HSF-1 transcriptional reporter , starting at 25 μM and peaking at 100 μM ( Fig 3A and 3B ) . Similarly , exposure to tanespimycin plates exclusively in adulthood resulted in significant activation of the HSF-1 reporter at 50 and 100 μM concentrations . No markers of toxicity were observed in any treatment groups , except for the 100 μM larval group , which were developmentally delayed by 24 hours and had a significantly reduced brood size ( S1 Fig ) , consistent with chronic HSP-90 inhibition [42] . Together , these data demonstrate that tanespimycin activates HSF-1 in C . elegans and that treatment exclusively in adulthood is not associated with overt toxicity . We next sought to determine whether tanespimycin treatment could extend lifespan in C . elegans . To circumvent potential longevity effects arising from delayed development and reproduction , we exposed worms to 100 μM tanespimycin plates from the first day of adulthood . Tanespimycin treatment significantly extended median and maximal lifespan compared to vehicle-treated controls ( Fig 3C ) . To determine whether the effects of tanespimycin on lifespan require hsp-90 , we also exposed worms to tanespimycin treatment in the presence of hsp-90 ( RNAi ) . Consistent with previous reports , [43] hsp-90 ( RNAi ) treatment significantly reduced hsp-90 mRNA levels compared to empty vector treated controls ( Fig 3D ) , and significantly shortened C . elegans lifespan ( Fig 3C ) [43] . Furthermore , upon depletion of HSP-90 , tanespimycin treatment no longer increased lifespan compared to vehicle controls ( Fig 3C ) . These data suggest that tanespimycin treatment extends lifespan in an hsp-90 dependent manner , but that severe depletion of HSP-90 is toxic to animals , despite the activation of protective stress responses .
This study was designed to infer and rank drugs matched to ageing at multiple levels of biological activity using a simple statistical test . In an initial gene-centric analysis , 19 drugs were identified as candidates expected to modulate ageing in humans . A major finding was that 6 of the statistically significant drugs , resveratrol , genistein , simvastatin , epigallocatechin gallate , celecoxib and sirolimus , have already shown lifespan-extending properties in experimental studies in model organisms . This statistically significant enrichment suggests that , despite its simplicity , the method is able to prioritise pro-longevity compounds . We then expanded the analysis to higher levels of biological complexity , and again found a statistically significant enrichment for pro-longevity drugs in all cases . The results of the analysis at different levels showed a moderate correlation . Compounds ranked high on average included trichostatin , geldanamycin and celecoxib , 3 drugs with pro-longevity effects in animal models [10 , 39 , 40] . The compound ranked highest on average was tanespimycin , an HSP-90 inhibitor , shown to acts as a senolytic agent by killing human senescent cells without affecting the viability of healthy cells [31] and to ameliorate disease phenotypes in Drosophila models of Huntington’s disease and spinocerebellar ataxia [32] . We found that tanespimycin treatment extended median ( 23% ) and maximum ( 16% ) lifespan in C . elegans , through its target HSP-90 , possibly through the induction of cytoprotective pathways . Tanespimycin must act through more than one mechanism as a geroprotector , because cellular senescence has not been reported to occur in C . elegans . Evidence from the literature supports the senolytic action of other drugs that we identified as potentially geroprotective . Dasatinib , a kinase inhibitor ranked 7th on average , has been reported to induce apoptosis in senescent preadipocytes [44] . Combination of dasatinib and quercetin , which also inhibits HSP-90 , induced apoptosis in senescent murine mesenchymal stem cells and mouse embryonic fibroblasts in vitro , improved cardiovascular function in aged mice , and decreased bone loss and age-related symptoms in progeroid mice [44] . Three of the top 10 compounds from the combined ranked list have previously been proposed as anti-ageing candidates for humans using bioinformatic analysis . Specifically , tanespimycin , geldanamycin and trichostatin were among the 24 drugs predicted by Dönertas et al . ( 2018 ) [18] and Calvert et al . ( 2016 ) [10] . In contrast , we did not observe any overlap with the top results from Fernandes et al . ( 2016 ) [15] possibly due to the use of a different drug-protein interaction database ( DGIdb [45] ) or source of ageing data . Even though our method focused on predicting drugs affecting human ageing and not ageing of animal models , we noticed that three of the drugs in our final list ( Table 2 ) also overlap with the results of Ziehm et al 2017 [11] ( i . e . sorafenib , imatinib , dasatinib ) and one with the results of a previous study conducted by Snell et al 2018 [46] ( i . e . erlotinib ) , meaning that 70% of our top 10 drugs candidate drugs have been previously predicted to influence longevity by other drug-repurposing methods for ageing . Similar enrichment-based methods that combine multiple levels of biological information have been used for drug-repurposing for Rheumatoid arthritis , Parkinson’s disease and Alzheimer’s disease [47 , 48] , but not , to our knowledge , to identify anti-ageing drugs . Using annotated databases , our method evaluated the enrichment for pro- and anti-longevity effects of all compounds analysed , rather than only those with significant scores , and we observe that in all cases pro- and anti-longevity compounds are ranked higher than expected by chance . Unfortunately , the research bias towards publishing the results of experiment that lead to an extension of the animal model lifespan instead of drugs causing negative or no effect on longevity , makes it difficult to assess the specificity of the method ( false discovery rate ) or to implement machine learning models using a balanced positive and negative set . Although tanespimycin acts as a senolytic [31] , and has been predicted to be geroprotective by two previous studies [10 , 18] , we have demonstrated its effect on longevity experimentally . A limitation of this study is that it is based on previous knowledge about drug-protein interactions , which for non-commonly studied drugs is incomplete . This may explain why we observed many anti-cancer and well-known drugs in our results . While we assessed this bias using permutations and we found no significant effect on our results , further research is needed to increase the drug-protein interactome data using , for example , high-throughput technologies like those currently available for kinases [49] . While we combined the results from the different data sources using a strategy based on ranks , we hypothesise that integration of these results using other methods may lead to a list with a higher enrichment for pro-longevity drugs . Additionally , further experimental testing is required on the lists produced in this study , particularly those generated by using gene ontology terms , which presented the highest enrichment for pro-longevity drugs . An inherent limitation of inferred associations is that they do not provide information about the directionality of the effect , which in this case means that it is unknown if the drugs will ameliorate ageing or the opposite . While we indirectly assessed this using an interaction-based similarity analysis between the drugs , resulting in clusters or pairs of drugs with similar mechanism of action , experiments should be conducted to determine the effects of each drug on ageing . Finally , a practical limitation is that we validated the results of this study using experiments in animal models although we used human data to perform the analysis . Although testing the effects of drugs on human ageing is challenging , progress is starting to be made . A clinical trial conducted by Mannick et al . ( 2018 ) [8] showed that pharmacological inhibition of the mammalian target of rapamycin in humans by dactolisib plus everolimus enhances the response of elderly people to immunisation against influenza and reduces the rate of subsequent infections . Moreover , a recent short-term clinical trial of sirolimus established its safety in healthy individuals [50] . Similarly , supplementation of nicotinamide ribose , identified as a possible CR mimetic , stimulated NAD+ metabolism in healthy individuals aged 55 to 79 years [51] . Some mechanisms of ageing may be confined to humans and their near relatives , and ideally , the bioinformatic findings should be evaluated in humans , initially through genetic epidemiology and ultimately through clinical trials .
Independently of the biological level , the drug-ageing associated was inferred by calculating the statistical significance of the drug-related terms and ageing-related terms using a Fisher’s exact test . Drugs were associated with ageing at the following biological levels: gene , pathways ( KEGG , Reactome ) , functions ( GO:BP , GO:CC , GO:MF ) and indirect protein interactions . The universe was defined as all the terms on each level associated with at least one drug . Thus , drugs with a lower p-value modulate a higher proportion of ageing-related terms than that expected by chance . To control for the false discovery rate , we used the Benjamini and Yekutieli adjustment [57] . A p-value lower than 0 . 05 after multiple testing correction was considered significant . Some drugs have been more studied than others , which could bias the results towards drugs with a higher proportion of discovered targets . To evaluate the impact of this research bias , we randomly selected the same number of terms that were used as ageing-related terms 1 , 000 times , and we repeated the statistical analysis . Then we counted the times the statistically significant drugs appeared on the same or lower ranking . We expected that drugs associated with many terms would rank higher independently of the random set generated . Each of the drug lists generated were ranked by the p-values obtained from the statistical analysis . Then , we transformed the ranking of the drug into a value ranging from 0 to 1 . A set of 142 pro-longevity drugs and 30 anti-longevity drugs present in the DrugAge and DrugBank databases were used to determine the occurrence and ranking of pro- and anti-longevity compounds in the lists , respectively . The ranking was then scaled into a value between 0 to 1 . The AUC between the variables describing the pro-longevity drugs and drugs analysed was calculated using the function AUC from the DescTools package ( https://cran . r-project . org/package=DescTools ) . To measure its statistical significance , we calculated the AUC of the lists previously generated to measure the research bias , and we counted the number of simulations with an equal or higher AUC . | Human life expectancy is continuing to increase worldwide , as a result of successive improvements in living conditions and medical care . Although this trend is to be celebrated , advancing age is the major risk factor for multiple impairments and chronic diseases . As a result , the later years of life are often spent in poor health and lowered quality of life . However , these effects of ageing are not inevitable , because very long-lived people often suffer rather little ill-health at the end of their lives . Furthermore , laboratory experiments have shown that animals fed with specific drugs can live longer and with fewer age-related diseases than their untreated companions . We therefore need to identify drugs with anti-ageing properties for humans . We have used publically available data and a computer-based approach to search for drugs that affect components and processes known to be important in human ageing . This approach worked , because it was able to re-discover several drugs known to increase lifespan in animal models , plus some new ones , including one that we tested experimentally and validated in this study . These drugs are now a high priority for animal testing and for exploring effects on human ageing . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"caenorhabditis",
"animals",
"drug",
"screening",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"pharmacology",
"research",
"and... | 2019 | Using the drug-protein interactome to identify anti-ageing compounds for humans |
Dengue viruses ( DENV ) cause debilitating and potentially life-threatening acute disease throughout the tropical world . While drug development efforts are underway , there are concerns that resistant strains will emerge rapidly . Indeed , antiviral drugs that target even conserved regions in other RNA viruses lose efficacy over time as the virus mutates . Here , we sought to determine if there are regions in the DENV genome that are not only evolutionarily conserved but genetically constrained in their ability to mutate and could hence serve as better antiviral targets . High-throughput sequencing of DENV-1 genome directly from twelve , paired dengue patients’ sera and then passaging these sera into the two primary mosquito vectors showed consistent and distinct sequence changes during infection . In particular , two residues in the NS5 protein coding sequence appear to be specifically acquired during infection in Ae . aegypti but not Ae . albopictus . Importantly , we identified a region within the NS3 protein coding sequence that is refractory to mutation during human and mosquito infection . Collectively , these findings provide fresh insights into antiviral targets and could serve as an approach to defining evolutionarily constrained regions for therapeutic targeting in other RNA viruses .
Dengue , caused by one of four dengue viruses ( DENV ) , is the most important arboviral disease in the world . Recent estimates indicate DENV infects 400 million people annually and over half of the world’s population lives in regions endemic for this debilitating and potentially life-threatening disease [1] . DENVs are single-stranded , positive sense RNA viruses within the Flaviviridae family of viruses and are transmitted from human-to-human in most parts of the world by Aedes aegypti and Aedes albopictus [2] . The DENV genome is approximately 11kb long and encodes a single viral polyprotein that is then post-translationally cleaved into three structural proteins—the capsid ( C ) , pre-membrane ( prM ) and envelope ( E ) —and seven non-structural ( NS ) proteins , NS1 , NS2a , NS2b , NS3 , NS4a , NS4b and NS5 . The known and suspected functions of these proteins have been reviewed elsewhere [3 , 4] . The viral coding region is flanked by a short 5’ untranslated region ( UTR ) and a longer 3’ UTR , both of which have been shown to associate with host factors and form secondary and tertiary structures that are required for viability of the virus [3] . Dengue prevention relies solely on vector control , which in most places has not resulted in sustainable reduction in disease incidence . While vaccine development has made important strides recently , the efficacy against all four DENV serotypes is variable and protection against infection is incomplete [5 , 6] . An antiviral drug that specifically combats DENV remains a much-needed tool against this global scourge . Antiviral drug development has mostly focused on compounds targeting conserved regions of the viral genome . Despite such an approach , drug resistance has developed rapidly , particularly for RNA viruses . RNA viruses are indeed notorious for their ability to adapt quickly to selective pressure from the host immune system and/or antivirals [7 , 8] . This adaptability can largely be attributed to their existence as a population and the error-prone characteristics of their RNA-dependent RNA polymerase ( RdRp ) [9–11] . These features combine to make RNA viruses able to quickly adapt to selective pressure from the host or antiviral treatment by exploring available sequence space [12–22] . Combination therapy is thus required to prevent rapid emergence of drug resistant strains and this strategy have been successful for human immunodeficiency virus ( HIV ) and hepatitis C virus ( HCV ) [23 , 24] . However , such a therapeutic approach may not be suitable for viruses such as dengue or chikungunya . The cost of treatment would increase with each additional drug and the tropical world , where these viruses are prevalent and cause significant economic burden , may not be able to afford the treatment needed . Identification of regions within the DENV genome that are not only evolutionarily conserved but also genetically constrained could thus pinpoint potent and resilient targets for monotherapy that minimizes risk of resistance emergence . To this end , we analyzed intra-host genetic diversity of DENV1 at day 1–3 and again at 4–7 following onset of fever in 12 dengue patients . The sera from these patients were then intra-thoracically inoculated into both Ae . aegypti and Ae . albopictus and analyzed after 10 days of infection ( Fig 1a ) . This method of viral delivery to the mosquito bypasses the bottlenecking event the virus encounters in the midgut barrier [25] and was necessary due to the limited amount of patient sera available . This method does , however , allow us to explore the full mutational space available to the virus when not confronted by this bottleneck thereby allowing a more complete picture of which areas in the genome tolerate a degree of variability without sampling hundreds of natural infections . Conversely , we were also able to identify those regions where variability was significantly reduced . These areas of reduced variation , hereby referred to as constrained , likely represent residues lethal to the virus if mutated . Using the resolution enabled by next generation sequencing ( NGS ) technologies [12] , we show that there is an abundant accumulation of intra-host viral population diversity in both humans and mosquitoes . Unexpectedly , we observed specific variations in the DENV genome in Ae . aegypti not present in Ae . albopictus , suggesting that amid the stochastic variations , there are distinct changes critical for DENV to thrive in each mosquito host . Importantly , we also show that there are regions of constraint within the viral genome that are refractory to variation in both human and mosquito .
The intra-host genetic diversity of DENV1 was analyzed in 12 individuals that were enrolled in the early dengue infection and outcome ( EDEN ) study ( S1 File ) [26 , 27] . Consensus sequences of DENV1 isolated from these 12 individuals and grown in C6/36 cells have been reported previously [28] . DENV1 genomic material from paired serum samples was also taken at fever day 1–3 ( early ) and 4–7 ( late ) from each patient . The patients were a mixture of primary and secondary infection and the final diagnosis for each of them are shown in S1 File [28] . We tested for differences between primary and secondary infections using the non-parametric Wilcoxon-Mann-Whitney test but none were statistically significant . DENV1 genomic material from these samples was PCR amplified and deep sequenced . These same serum samples were also inoculated intrathoracically into 4-day old female Ae aegypti and Ae albopictus . After 10 days of incubation , the ten mosquitoes for each serum sample were pooled to minimize sampling bias from individual mosquitoes . DENV1 was then PCR amplified from the total RNA and deep sequenced on an Illumina or Solid sequencing platform ( Fig 1a and S2 File ) . In order to check whether different sequencing technologies ( Solid and Illumina ) had an effect on the type of SNPs detected , we performed Fisher's exact test on the number of Transition and Transversion SNPs of serum samples sequenced by Solid and Illumina . The results of this analysis suggest that there are no significant differences in any gene between Solid and Illumina sequencing ( S3 File ) . Overall , our deep sequencing data shows positional variance throughout the DENV1 genome ( S4 , S5 and S6 Files ) . The 17 positions where consensus discordance was observed in at least three of the twelve viruses are shown in Fig 1b and the complete list of all consensus changes observed in our data set are described in S7 File . These consensus changes fall within the coding sequence and the 3’UTR . Two of these consensus changes , one in prM and the other in NS5 also resulted in changes to the protein coding sequence . Besides the limited number of consensus changes , there are a large number of positions throughout the viral genome that display a degree of intra-host viral diversity . We refer to these types of positions as having ‘variance’ . To distinguish variants from the average sequencing error rate , we used the program Lofreq , which identifies single nucleotide polymorphisms by incorporating base-call quality scores as error probabilities into its model and assigns a p-value to each variant [29] . These analyses identified seven positions within the DENV genome that possessed this level of reproducible plasticity: two in the E gene , one in the NS1 gene , one in the NS3 gene , one in the 2k peptide at the C terminus of the NS4a gene and two in the NS5 gene ( S4 File ) . Since our samples were extracted from the same patients at two time points during their infection and then directly inoculated into the two mosquito vector species , we were able to track these changes in the genetic diversity across the viral genome over time . More specifically , we compared the proportion of base calls at each position in the DENV genome in early and late serum samples as well as between human and Ae . aegypti or Ae . albopictus in both early and late stages of acute dengue . Our results indicate that during the course of the human infection , changes in the intra-host genetic diversity were more prevalent in the NS1 , NS2A and E genes ( NS2A vs NS2b Bonferroni corrected p-value [Bcp] = 0 . 008; NS2A vs NS3 Bcp<0 . 001; NS2A vs NS4B Bcp = 0 . 02; NS2A vs NS5 Bcp<0 . 001; E vs NS3 Bcp = 0 . 006; also NS1 vs NS3 Bcp = 0 . 002 ) . The average number of changes occurring over the course of four days of human infection is 86 or ~0 . 0020 changes/position/day of human infection . In Ae . albopictus , changes were observed in E , NS1 , NS4A ( 2k peptide ) and NS5 genes . In Ae . aegypti , changes were observed in prM , E , NS1 , NS3 , NS4A ( 2k peptide ) and NS5 genes ( Fig 2 and S5 File ) . Two of the most commonly observed changes at 2719 ( NS1 ) and 6782 ( 2k peptide ) were observed in both species of mosquito and suggests that selection pressure on these residues is likely to be a common mechanism shared between the species . Interestingly , there were changes that were unique to Ae . aegypti infection , notably at 9986 and 9998 in the NS5 gene . These changes suggest that differential selection pressures may be applied on selected nucleotide residues in the DENV genome by Ae . aegypti but not by Ae . albopictus . For each detectable change in the DENV genome over the course of the human infection , we defined whether the proportion of base calls at each position moved towards or away from the consensus base after 10 days of incubation in the vector . Our results indicate that the variance acquired between early and late serum samples undergo a reversion back towards the sequence in the early serum sample after 10 days of incubation in either vector ( Fig 3 ) . We then asked whether this reversion was happening in specific regions or was a more general mechanism . Our data suggest that reversion is largely a general phenomenon that occurs across the majority of the viral genome ( S8 File ) . The majority of the reversion events are small oscillations in the overall composition at each position; however , larger consensus-level changes were also observed ( S9 File ) . Selection pressure and genetic drift were measured by calculating the dN/dS ratio . Overall it can be seen from the mean dN/dS ratio for each group that there is likely a purifying selection pressure against non-synonymous mutations ( S10 File ) . The Mann-Whitney test was used to compare single protein coding sequences against the rest of the polyprotein and results suggest that there are no significant differences within the human samples ( Early and Late ) , whereas significant differences were identified within the Ae . aegypti and Ae . albopictus samples ( S11 File ) . The transition/transversion analysis and the Shannon diversity index and Shannon equitability measurements suggest that there is a decrease in mutation frequency from early to late samples in Ae . aegypti and human whereas an increase in the mutation rate was observed in the Ae . albopictus samples ( S12 and S13 Files respectively ) . We also questioned whether the two time points for each species were likely to come from the same population and results suggest that the dN/dS ratio is significantly different between the Ae . aegypti early and late samples ( S14 File ) . From the Ts/Tv ratio comparison , the ratio is significantly different ( <0 . 01 ) between the Early and Late samples in Ae . aegypti , Ae . albopictus and human . This trend is consistent with the results from the Shannon diversity index and Shannon equitability measurements ( S13 File ) . The heterogeneity observed at positions 9986 and 9998 ( NS5 ) in the DENV genome fall within the RdRp domain of NS5 and correspond to amino acids 541 ( Thr → Ala ) and 545 ( Leu → Leu ) at junction of the “palm” and the α14 alpha helix “finger” of the RdRp domain of the protein , respectively ( Fig 4 ) [30] . That these observations were unique to Ae . aegypti suggests that these are not random events but are responses to species-specific selection pressure . To test this possibility experimentally , we constructed an infectious clone of DENV1 isolated from the same outbreak in Singapore in 2005 [26] but from a patient not among the 12 studied here . This infectious clone was constructed with the exact nucleotide sequence of the virus ( GenBank: EU081230 ) that was isolated in the C6/36 Ae . albopictus derived cell line [26] . In vitro transcribed RNA was electroporated into BHK cells and harvested supernatant was inoculated intrathoracically into both Ae . aegypti and Ae . albopictus . The initial starting material and time points of 5 , 10 and 21 days post intrathoracic inoculation were then sequenced and the data analyzed in the same manner as described above . Although the number of replicates in this experiment is limited , our results nevertheless indicate that the changes observed in NS1 and the 2k peptide ( Fig 2 ) are recapitulated in both Ae . aegypti and Ae . albopictus ( Fig 4 ) . The two Ae . aegypti specific residues in the NS5 gene , 9986 and 9998 ( amino acid positions 541 and 545 ) , arose 21 days after the infectious clone-derived virus was incubated in Ae . aegypti but not in Ae . albopictus ( Fig 4 ) . Collectively , these findings demonstrate that species-specific selective pressures act to select for variance in specific positions on the DENV genome . The “palm” domain of the DENV RdRp is the most structurally conserved domain among all known polymerases [30] . Although there is no specific catalytic activity associated with residue at position 541 , the Thr→Ala substitution may alter the angle of the “finger” relative to the “palm” and by doing so; alter the enzymatic properties of the RdRp . The nucleotide change at position 9998 does not translate to an amino acid shift at position 545 and its functional significance in regards to RdRp activity is not clear . However , examination of the predicted RNA secondary structure in this region suggests that nucleotides 9986 and 9998 interact with each other in a previously uncharacterized stem-loop structure ( Fig 4C and 4D ) [31] . The observed A→G and/or C→U changes at bases 9986 and 9998 respectively are predicted to strengthen this interaction ( see reduced dG in MFE structure ) . Other structures and sequences within the virus have been shown to be essential for the virus in a species-specific manner and this may be another mechanism the Ae . aegypti vector uses to control DENV replication [32] . To identify regions of constraint within the DENV1 genome , we aggregated all predicted single nucleotide variants ( SNVs ) to detect regions with ( i ) a local enrichment in intra-host SNV calls ( mutational hotspots ) and ( ii ) a significant depletion in variants ( mutational cold-spots ) ( Fig 5 ) . This type of analysis complements classical approaches of finding evolutionarily conserved regions through multiple sequence alignments and can reveal functionally important , though otherwise not easily detectable regions [29] . No hotspots predicted in more than one sample of either the mosquito or the human isolates could be detected . However , at least four samples of the infectious clone experiment were observed to have a hotspot in the envelope protein ( bases 1789–1814 ) . As the virus used in the infectious clone experiment was in vitro transcribed and electroporated into BHK cells for packaging we suggest that this particular hotspot can be attributed to the markedly different selection pressures in cell culture conditions than encountered by the virus in vivo [33] . Overall the mosquito samples had 12 coldspots covering 1064 total positions whereas the human samples show only 2 coldspots covering 220 total positions , which is consistent with previous reports [29] . The absence of coldspots in NS1 and NS2A has been observed before [29] , but its significance is unclear . Mosquito , human and infectious clone samples largely display an absence of shared mutational coldspots ( i . e . regions that show intra-host constraint ) with the notable exception of coldspots within the multifunctional NS3 gene ( Fig 5 ) . NS3 is comprised of a protease domain and an ATP-driven helicase with two subdomains . The large coldspot discovered in the mosquito samples covers all three of these domains . Structurally , this region clusters around the ATP-binding domain and around the interaction site with NS2B ( Fig 5b ) [34] . The coldspot in NS5 is in the fingers domain of the RdRp and forms part of the zinc-binding pocket ( Fig 5a and 5c ) [30] . Although the function of this zinc-binding pocket is unknown , it is a feature shared with the West Nile virus RdRp and is likely to be functionally important [34] . In order to test the hypothesis that coldspot regions would make good antiviral targets , siRNA’s were designed to target the hotspot in the E gene ( starting at position 1848 ) , the coldspot in NS3 ( starting at position 4794 ) and a region near the Ae . aegypti specific mutations in NS5 ( starting at position 10014 ) . A non-targeting ( NT control ) was used as a control for the experiment and DENV1 genome copies relative to GAPDH were measured by RTPCR at 24 and 48 hours post infection . All DENV1 siRNA’s were significantly different than the NT control ( p<0 . 001 ) at both 24 and 48 hours post infection ( Fig 5d ) . The siRNA’s targeting the E gene and NS3 were indistinguishable from each other in their effect at both 24 and 48 hours post infection . Interestingly , although the siRNA targeting NS5 that contains the Ae . aegypti specific mutations was not statistically significant from the E gene and NS3 siRNA’s at 24 hours post infection , it was statistically less effective by the 48 hour timepoint ( p = 0 . 0008 and p = 0 . 0003 respectively ) ( Fig 5d ) .
In this study , we have used high-throughput parallel sequencing to analyze intra-host genetic diversity in the DENV genome directly from serum samples obtained from dengue patients and intra-thoracically infected mosquitoes . Our data show that numerous positions within the DENV genome exhibit a high degree of plasticity over the course of infection within the human and mosquito hosts and that several of these changes are in functionally significant domains of the viral coding sequence . Maintenance of genome plasticity within viral populations is a poorly understood process however; it is possible that it may be critical for the overall fecundity of the virus [9] . Interestingly , members of the mosquito-borne clade of flavivirus have been observed to be more genetically stable over time than other RNA viruses [10 , 33 , 35–39] . Previous studies have found that repeated passage in a single host is likely to result in a consensus genome that is highly divergent from its original source [40 , 41] . In studies measuring viral variance by clonal analysis of viral amplicons , significant intra-host variation of the virus has been observed in laboratory passaged DENV and other flaviviruses such as West Nile virus [33 , 36 , 37 , 42 , 43] . Serial in vitro or in vivo passages in mosquitoes show significant amounts of consensus changes throughout the genome [15 , 33 , 37 , 44] . The long-term stability of these viruses may therefore be due to differential selection pressures exerted by the human and mosquito hosts that result in a net conservation of the viral genome [33 , 37] . Our results suggest that there is an abundant accumulation of intra-host viral population diversity in humans and mosquitoes . Consistent with the prevailing theory however , our study indicates that the changes that accrued during infection in the human host predominantly revert back to the ‘original state’ as the virus transits through the mosquito . Intriguingly , this reversion is occurring despite bypassing the bottlenecking event of midgut barrier escape . This suggests that even when these humanized variants are given the opportunity to replicate in the mosquito body , they are outcompeted by the original population observed early in the human infection . Given the broad distribution across the viral genome and their stochastic appearance in our data set , these larger changes likely represent changes to locations tolerated in the human host but not in the mosquito vectors . The proportion of these variants that are then able to disseminate through the salivary gland infection/escape barrier and thus infect the next human host is also of interest and will be the subject of future study [45] . Taken together , these data provide evidence that cycling between the two hosts restricts the overall diversity in DENV genome , making it phylogenetically more stable than other RNA viruses that propagate within a single species . Our study also cautions on performing phylogenetic analyses on consensus sequences alone; especially those derived from serially passaged virus . We have also identified species-specific variations that have not been previously reported . Some of these variations are recurrent among the samples we have tested suggesting that there are positions with a high degree of plasticity in the DENV genome . The locations of these positions appear to depend on the host , which provides new insights into how the different vectors may influence DENV evolution . These differences also suggest that viruses transmitted predominantly in an Ae . aegypti-human cycle may produce viruses genetically distinct from those transmitted predominantly in an Ae . albopictus-human cycle . Indeed , it may be a molecular basis for which epidemic emergence is more often associated with Ae . aegypti than Ae . albopictus [46] . Further studies are necessary to determine whether these same residues arise after passing through the midgut barrier and are ultimately present in the saliva of the infected mosquitoes [47 , 48] . Interestingly , our samples did not display the extensive mutations in the 5’ or 3’UTR that have been identified in recent studies [49–51] . We observed only a single position in the 3’UTR to change consensus and only for two of the isolates ( S7 File ) . This consensus change falls within the unstructured region between DB1 and DB2 and is not predicted to have a substantial impact on the overall 3’UTR structures ( S15 File ) . The reasons for the observed stability in this region are unclear . The aforementioned studies were primarily conducted in cell lines with DENV2 and , to a lesser extent , DENV3 [49–51] . Whether the DENV1 serotype is fundamentally different in this regard or whether this can simply be attributed to our limited sample size is an interesting question and deserving of additional studies . The finding of areas within the DENV genome that are constrained in nucleotide variation in the human and mosquito hosts are interesting . These cold-spots are consistent even in two disparate host species and thus suggest that these positions may encode protein-protein interactions that are functionally vital to the DENV lifecycle . The cold-spot at the interface between NS2b and NS3 is particularly interesting . NS3 requires a direct interaction with NS2b as a cofactor for its proteolytic activity . Our findings suggest that if not outright lethal , mutations within this interaction site are likely to cripple the virus . Given that sequence in this region of the genome is highly constrained , a potentially attractive antiviral strategy may be RNA interference ( RNAi ) due to its potential for high specificity to the viral genome [52 , 53] . In this study , we tested three different siRNA’s targeting the E gene , NS3 and NS5 . Although all were able to significantly reduce viral copy number , the siRNA targeting NS5 was not as effective as the other two after the first 24 hours of infection . Given the plasticity observed in this region when these isolates were passaged into Ae . aegypti , this may represent a ‘flexible’ part of the genome and a less than ideal target for this type of antiviral strategy . This may indeed also explain the lack of difference between siRNA that targeted the E gene compared to NS3 as , while there is a degree of variance in the E gene , it is not a ‘flexible’ part of the genome that alters depending on the host species . Indeed , it is possible that sub-therapeutic doses of siRNA against the E gene would be more likely to generate resistant mutants over repeated passages compared to that against the cold spot in NS3 . This could be a useful focus in future investigations . While we have used siRNA in this study , small molecule therapeutics against the cold spot on NS3 would be another option . Binding to either interaction surface should interfere with the catalytic function of the viral protease although our data suggests that the NS3 is a more constant target . A small molecule inhibitor would also have the potential advantage of inexpensive mass production; a distinct advantage for the treatment of affected populations unable to afford more expensive therapies . The changes observed in the PrM and NS5 genes are remarkable as they are highly conserved across most flaviviruses [30 , 54–56] . The Leu → Phe mutation in the prM gene at position 138 occurs in the C-terminal transmembrane domain of the “M” residue that is embedded in the lipid bilayer of the mature virion [54] . The functional significance of this particular amino acid change is not immediately clear . The membrane composition of mosquito cells are substantially different from mammalian membranes , particularly in their cholesterol content [57] . It is conceivable that an alteration at this position may be in response to these differences and plays a role in the infectivity of the virus . The Val → Ala mutation at position 324 in the NS5 gene occurs at the N-terminus of the RdRp domain in a region involved in the binding of β-importin and NS3 [30] . Alteration in the ability of NS5 to interact with these proteins could directly impact the ability of NS5 to shuttle into the nucleus and the ability of the virus to replicate its genome respectively [30 , 58] . The lack of extensive cold-spots in the NS5 gene was surprising to us although this might be due to the fact that we pooled ten mosquitoes inoculated with the same serum together . While this methodology has the advantage of being more rigorous when trying to identify common variants , it may obscure authentic coldspots due the averaging effect of combining individuals with stochastic mutations in these regions . One cold-spot was identified in the fingers subdomain of NS5 though not in the thumb or palm domain , which contains the catalytic active site . The latter has been the focus of attention in anti-dengue drug development [59 , 60] . Furthermore , that we also found variance in two nucleotides in the RdRP domain , one of which changes the protein coding sequence , in DENV that replicated in Ae . aegypti but not Ae . albopictus also raises additional concerns on antiviral drug development efforts that target the RdRP . Most laboratories culture DENV in C6/36 cell line , which is derived from Ae . albopictus . Compounds that show attractive efficacy to DENV cultured in such cells may thus not achieve anticipated efficacy in humans who acquire infection from Ae . aegypti , which is the epidemiologically more important vector . Finally , the analysis we have employed in this study can readily be adapted for other pathogen-host studies affecting the developing world such as influenza , chikungunya and ebola viruses . We suggest that our approach could serve not only to identify areas of constraint in viral genomes but also to monitor the emergence of escape mutants following vaccination or initiation of antiviral therapies [61] .
The samples used in this study were collected under the Early Dengue infection and outcome study ( EDEN ) . This prospective study was approved by the National Healthcare Group Domain Specific Review Board ( DSRB B/05/013 ) and the Institutional Review Boards of the National University of Singapore and DSO National Laboratories . Enrollment of participants into the study was conditional upon written informed consent administered by a designated research nurse . All biological specimens collected for this study were de-identified following collection of demographic and clinical data . Both Ae . aegypti and Ae . albopictus mosquitoes were obtained from a colony at the Duke-NUS Graduate Medical School . The colony was established in 2010 with specimens collected in Ang Mo Kio , Singapore , and infused monthly with field-collected mosquitoes to maintain genetic diversity . Female mosquitoes , three to five days old , were intrathoracically inoculated with 0 . 017 μl of serum from the Early Human sample ( fever day 1–3 ) and the Late Human sample ( taken four days after the initial sample ) for 11 out of 12 patients as previously described [62] . Insufficient sera remained from one patient for the mosquito inoculations . Mosquitoes inoculated with DENV-1 clinical serum were incubated for 10 days , while mosquitoes inoculated with DENV-1 derived from pOEEic infectious clone were incubated for 5 , 10 and 21 days respectively at 28°C and 80% humidity , with access to 10% sucrose and water . Surviving mosquitoes were killed by freezing and examined for the presence of viral antigens in head tissue by direct immunofluorescence assay ( IFA ) . Infected mosquitoes were stored at -80°C until assayed . For each viral sample , including each time point of the infectious clone experiment , 10 infected mosquitoes were pooled and triturated with a pellet pestle ( Sigma Aldrich , St . Louis , MO , USA ) in 250 μl of 1x PBS . DENV RNA was extracted from the sample using TRIzol RNA isolation reagent ( Invitrogen , Carlsbad , CA , USA ) according to the manufacturer's protocol and stored at −80°C until use . A pool of 10 inoculated mosquitoes was triturated in 250ul L-15 ( Gibco , Life Technologies , Carlsbad , CA , USA ) maintenance medium . Total RNA was extracted using TRIzol® ( Life Technologies ) according to the manufacturer’s protocol and stored at -80°C until use . Two separate reverse transcription reactions were carried out using the SuperScrip III First-Strand Synthesis System ( Life Technologies ) according to the manufacturer’s protocol ( 1 ) cDNA was synthesized with random hexamers for downstream amplification of fragments 1 , 2 , 3 and 4 . ( 2 ) cDNA was synthesized with 10μM of reverse primer 10693R for downstream amplification of fragment 5 . cDNA was amplified in 5 fragments by PCR ( S16 File ) . Using the respective primers for each fragment , PCR was carried out using Phusion High-Fidelity PCR Master Mix with HF Buffer ( Thermo Fisher Scientific , Waltham , MA , USA ) . 2μL of cDNA was mixed with 1μL of each primer ( 10μM ) , 25μL of 2X master mix and 21μL of water . The PCR conditions were: 30sec at 98C followed by 40 cycles of PCR at 98C for 10sec , 55C for fragment 1 or 57C for fragments 2 , 3 , 4 and 5 for 20sec , 72C for 2min and final extension at 72C for 10min . The PCR products were run on a 1 . 5% agarose gel . Bands of the correct size were excised and gel purified using Qiagen QiaQuick gel extraction kit ( Qiagen , Valencia , CA , USA ) according to the manufacturer’s protocol . NGS libraries were constructed according to the methods described in Aw et al . [63] . pOEEic was digested with SacII at 37°C for 2 h . Linearized DNA was purified using ultrapure Phenol: Chloroform: Isoamyl Alcohol ( Invitrogen ) according to the manufacturer’s protocol . In vitro transcription of the purified DNA was performed to generate full-length genomic DENV RNA using MEGAscript T7 kit ( Ambion , Life Technologies , Carlsbad , CA , USA ) according to the manufacturer’s protocol . The reaction was spiked with additional rATP after incubation at 37°C for 30 minutes , and further incubated at 37°C for 2 h . 5 μg of RNA was electroporated into approximately 5 . 0×106 BHK cells in a 4 mm cuvette using the Bio-Rad Gene Pulser II with the settings adjusted to 850 V , 25 μF . Each cuvette was subjected to 2 pulses at an interval of 3s . The cells were allowed to recover for 10 mins at 37°C and transferred into 15 ml of pre-warmed culture medium in a T75 flask . Cell culture supernatant was collected 5 days after infection and tested for the presence of infectious DENV-1 using plaque assay . The BHK cells were also scraped off and analyzed for the presence of DENV-1 envelope antigen by indirect fluorescent assay ( IFA ) using the anti-DENV-1 envelope primary mAb HB-47 . The data analysis pipeline used in this study is built upon open-source tools , which are freely available . The bulk of the sequencing analysis was done using the viral analysis pipeline ViPR ( https://github . com/CSB5/vipr ) , which mainly handles mapping of reads and calling of SNVs with LoFreq [29] . For mapping of Illumina paired-end reads to the Sanger sequenced reference genomes we used BWA version 0 . 6 . 2-r126 for all Illumina and version 0 . 5 . 9 for all SOLiD sequencing datasets [64] . LoFreq ( version 0 . 6 . 1 ) was used for SNV calling using default options and regions overlapping primer positions were ignored . Cold-spot analysis was performed as described in [29] . In brief , SNVs predictions from groups of samples are pooled and then scanned for SNV free regions that are larger than expected ( binomial test; Bonferroni corrected p-value < 0 . 05 ) . Reference genomes used for mapping and annotation were specific for each sample and come from the Sanger consensus sequences reported for viruses in Schreiber et al . 2009 [28] . An in-house R script was developed for calculations and visualization [65] . First , the numbers of reads across the entire sequence were extracted from pileup data files and used as default data values . Additional reads for were extracted from . snp format files for all positions where at least one non-consensus nucleotide was present , over-writing the pileup data for that position . Regions where primers bound to amplify the DENV1 genome , corresponding to positions 1–70 , 2065–2084 , 4221–4241 , 6442–6461 , 8519–8540 , 10645–10735 , were excluded from all analysis . As the typical number of reads is ~100 , 000 , any difference between the proportions of non-consensus bases at different time points or in different hosts that is biologically significant is also statistically significant . Selection pressure and genetic drift were measured by calculating the dN/dS ratio , which is the ratio of non-synonymous mutation changes per non-synonymous sites ( dN ) to the synonymous mutation changes per synonymous sites ( dS ) ( S10 File ) . The Mann-Whitney test was then performed on the dN/dS of the samples between a particular protein coding sequence and compared to the rest of the protein coding sequences within an experimental condition ( e . g . the C gene is compared to PrM , E , NS1 , NS2a , NS2b , NS3 , NS4a , NS4b , 2K and NS5 from the early human samples ) ( S11 File ) . The frequency of mutation type was measured by calculating the ratio of transitions ( A ↔ G , C ↔ T ) to transversions ( A↔ C , G ↔ T , G ↔ C , A ↔ T ) for each protein coding sequence ( S12 File ) . Diversity and evenness across the viral polyprotein was calculated using the Shannon diversity index and Shannon equitability measurements respectively and the overall number of mutations across the polyprotein ( either transitions or transversions ) was calculated using a 100bp window ( S13 File ) . We used the Mann-Whitney-Wilcoxon test to assess whether the two samples come from the same population ( S14 File ) . Plasticity at each position-individual-time point was determined by the proportion of the reads that did not agree with the consensus . Three thresholds ( 0% , 1% and 5% ) of plasticity were recorded . For each time point , the number of individuals ( out of 12 in human , 11 in mosquitoes ) with plasticity above that threshold were counted to assess localized reproducibility across samples . If more than a third of the samples for that position-individual-time point exhibited plasticity ( of >0% ) , this was declared significant: with 99 . 6% of positions agreeing with the consensus , this threshold leads to a p-value of < 10−7 . Although this proportion of samples was arbitrarily defined , it nonetheless provides a conservative approach to differentiate stochastic from biologically important variances . Differences between ( logically comparable ) pairs of time points were assessed , for each position-individual-pair , by counting differences , differences of at least 0% , 1% , or 5% , between the proportion of reads of the consensus ( vs . all others ) nucleotide . If more than a third of the samples for that position-individual exhibited a difference between the two time points ( of >0% ) , this was declared significant: with 99 . 2% of positions not exhibiting differences within the two serum samples , this threshold leads to a p-value of < 10−5 . Prevalence of differences in different genes were tested via χ2 tests between all gene pairs , with Bonferoni’s correction for multiple testing . For triples of time points ( e . g . human early , human late , mosquito late ) , we quantified pairs of differences for each position-individual combination . In this analysis , we looked for differences of at least 0 . 1% between the proportions of reads ( of the consensus nucleotide at the first time point ) for each pair of time points , and coded the sign of the differences . If both differences were positive , this position-individual was coded as changing towards consensus; if negative , as changing away from consensus; and if alternating ( positive then negative or vice versa ) , as a reversion ( if either difference were less than 0 . 1% , the position for that individual was ignored ) . The number of individuals with each change was recorded as a measure of reproducibility . Reversions with differences of >5% were stored and plotted individually . Consistent reversions across individuals ( in more than 25% of samples ) were declared significant: with 99 . 6% of positions in Ae . aegypti , and 99 . 5% in Ae . albopictus , not undergoing reversions , this leads to p<10−4 . At this evidence threshold , if throughout the entire sequence reversions occurred by chance alone , we would expect to see 0 . 1 to 0 . 2 locations declared significant , and there is ~15% chance of seeing 1 false positive , and ~1% chance of seeing 2 or more . In addition , we identified positions in which the consensus nucleotide itself changed between time points . To do this , we scanned over positions and noted any in which the dominant nucleotide for any two time points , for any individual , differed . Positions in the list for with several such switches were plotted . | Dengue viruses cause debilitating and potentially life-threatening acute disease throughout the tropical world . While drug development efforts are underway , there are concerns that drug-resistant strains will emerge rapidly . Indeed , many antiviral drugs for other RNA viruses lose efficacy over time as the virus mutates . Here , we sought to determine if there are regions in the dengue virus genome that are constrained in their ability to mutate and could therefore serve as better targets for antiviral drugs . Deep sequencing of the dengue virus 1 genome directly from the blood of twelve dengue patients and from mosquitoes given this blood showed consistent and distinct mutation patterns during infection . Importantly , we identified regions within the viral genome that are resistant to mutation during human and mosquito infection . Collectively , these findings provide fresh insights into potential antiviral targets and could serve as an approach to defining better regions for therapeutic targeting in other RNA viruses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Analysis of Dengue Virus Genetic Diversity during Human and Mosquito Infection Reveals Genetic Constraints |
Recently , a novel approach has been developed to study gene expression in single cells with high time resolution using RNA Fluorescent In Situ Hybridization ( FISH ) . The technique allows individual mRNAs to be counted with high accuracy in wild-type cells , but requires cells to be fixed; thus , each cell provides only a “snapshot” of gene expression . Here we show how and when RNA FISH data on pairs of genes can be used to reconstruct real-time dynamics from a collection of such snapshots . Using maximum-likelihood parameter estimation on synthetically generated , noisy FISH data , we show that dynamical programs of gene expression , such as cycles ( e . g . , the cell cycle ) or switches between discrete states , can be accurately reconstructed . In the limit that mRNAs are produced in short-lived bursts , binary thresholding of the FISH data provides a robust way of reconstructing dynamics . In this regime , prior knowledge of the type of dynamics – cycle versus switch – is generally required and additional constraints , e . g . , from triplet FISH measurements , may also be needed to fully constrain all parameters . As a demonstration , we apply the thresholding method to RNA FISH data obtained from single , unsynchronized cells of Saccharomyces cerevisiae . Our results support the existence of metabolic cycles and provide an estimate of global gene-expression noise . The approach to FISH data presented here can be applied in general to reconstruct dynamics from snapshots of pairs of correlated quantities including , for example , protein concentrations obtained from immunofluorescence assays .
Cells are well known to respond to external conditions by altering their gene expression . In recent years , many examples of altered gene expression programs have been revealed by population level studies , including microarray studies of yeast , mammalian , and bacterial cells . But many cells are also known to alter gene expression is ways that are heterogeneous across a cell population . Examples include the acquisition of competence for DNA uptake [1] , [2] and spore formation [3] in Bacillus subtilis , induction of the lac operon in Escherichia coli depending on “memory” of previous exposure to lactose and the presence of lactose permease [4] , [5] , and the response of Saccharomyces cerevisiae ( budding yeast ) temperature-sensitive mutants to a shift to non-permissive temperature depending on the position of cells in their division cycle [6] , [7] . Heterogeneous changes in gene expression in response to homogeneous external cues may be purely stochastic as in the switch to competence in B . subtilis [1] , [2] , [8] , or may depend on pre-existing non-genetic differences such as the phase of the cell cycle in budding yeast [6] , [7] . Since population level studies are not well suited to reveal heterogenous behavior , how can heterogeneous changes in gene expression be studied and quantified ? Fluorescent reporter proteins have been used successfully to report on expression of a small number of genes either via FACS analysis or fluorescence microscopy . However , the use of fluorescent reporters is generally limited to highly expressed genes , with time resolution severely limited by fluorescent protein maturation and the low turnover rates of the fluorescent marker . Moreover , construction of fluorescent reporters can be laborious and impractical for studies of large-scale transcriptional responses . A promising approach that has recently been used to study gene expression on a cell-by-cell basis is Fluorescence In Situ Hybridization ( FISH ) [9]–[11] . In FISH , fixed cells are exposed to fluorescently labeled probes of specific mRNA transcripts , so that the number of these mRNAs can be counted in each cell by the number of bright spots . Advantages of FISH include: ( 1 ) absolute quantification since the actual number of mRNAs can be counted , ( 2 ) time resolution since there is no delay for reporter maturation , ( 3 ) ability to directly study wild-type cells , and ( 4 ) the ability to probe simultaneously for multiple mRNAs , e . g . by employing probes with different fluorescent spectra [10] , [12] . A significant disadvantage of FISH is the requirement to fix cells . This disadvantage presents a particular challenge when it is the dynamics of gene expression that is of central interest . For example , each individual drawn from an asynchronous yeast population represents a particular moment in the cell division cycle . In essence , the problem we wish to address is how to reconstruct the dynamics of gene expression from what amount to “snapshots” , where each individual cell represents a different point in time . Here , we present an approach to extracting information about the dynamics of gene expression from FISH data by considering correlations of expression between pairs of genes ( cf . Fig . 1 ) . The approach applies even if the dynamics of interest occurs heterogeneously in a population . One class of dynamics we consider are cyclic oscillations of gene expression . Common examples are the cell cycle , circadian oscillations , and metabolic oscillations [13] . Cyclic oscillations of gene expression , such as the cell cycle , have been studied at the population level by synchronizing cells , but for many organisms synchronization is difficult without strongly perturbing the cells . A non-perturbative approach to studying oscillatory gene expression is likely to be of value in these cases . To study metabolic oscillations , cells of the yeast Saccharomyces cerevisiae have been synchronized in chemostats [13] , but those cells demonstrably continue to influence each other via levels of dissolved oxygen and other chemical species . To ascertain if Saccharomyces undergoes metabolic oscillations outside the chemostat , Silverman et al . [14] recently obtained an extensive FISH data set , and argued for the existence of metabolic oscillations based on correlations in gene expression . Using the same data set , we apply our approach to reconstructing oscillatory dynamics , and confirm the existence of metabolic cycles in unsynchronized yeast populations [14] . Our approach can also be applied to transient oscillations in response to external stimulation , such as in the bacterial SOS response to DNA damage [15] or in the analogous eukaryotic p53-Mdm2 system [16] . Another class of dynamics we consider are stochastic switches among different states of gene expression . Examples include persister cells in Escherichia coli [17] , competence [1] , [2] , [8] and swimming/chaining in Bacillus subtilis [8] , the stringent response in mycobacteria [18] , and galactose utilization in Saccharomyces cerevisiae [19] . Specifically , we show how Maximum Likelihood Estimation ( MLE ) [20] can be applied to FISH data obtained for multiple pairs of genes to reconstruct the underlying dynamics of gene expression . MLE consists of finding the set of parameters within a particular family of models for which the observed data is most “likely” . MLE has been applied successfully to biological data analysis in many contexts , from reconstruction of evolutionary trees [21] , [22] to estimation of genetic parameters [23] to understanding the evolution of gene structure [24] . We show using synthetic FISH data that MLE can accurately reconstruct dynamics , even in the presence of substantial noise , provided the number of genes and the number of FISH observations per gene pair are sufficient . Reconstructing gene-expression dynamics is most challenging in the “bursty” regime where mRNAs are often present at very low levels or not at all in the cell , except when transcriptional bursts occur . For this regime , we present a robust approach based on thresholding the FISH data into binary form , followed by MLE analysis . In this case , we show that Principal Component Analysis ( PCA ) of the covariance matrix performs nearly as well as MLE . We suggest that the two-step approach of thresholding followed by MLE or PCA is likely to prove the best practical approach to reconstructing gene-expression dynamics for most real FISH data sets , and we demonstrate this approach using the data set of Silverman et al . [14] . Importantly , the method we present here for inferring intracellular dynamics from data in the form of “snapshots” is quite general , relying only on measurements of pairs of quantities in single cells , with no requirement for exact counts . The method can therefore be applied with little modification in other contexts such as quantitative immunofluorescence or single-cell sequencing studies .
We first consider the continuous regime where many bursts typically contribute to the instantaneous mRNA number . To demonstrate the MLE algorithm , we reconstruct the dynamics of gene expression using synthetic FISH data for which the underlying dynamics is known . We focus on analyzing cyclic dynamics , e . g . the cell cycle or a metabolic cycle; the results can be readily extended to stochastic switches , which are introduced in a later section . We denote the mean expression level of mRNA for gene by , which is taken to be periodic with the same period for genes . For concreteness , we denote the period as , although cannot be inferred from FISH data alone . FISH observations are generated for pairs of genes at randomly chosen times: , where the s reflect fluctuations in mRNA number around the mean , as well as noise in the measurement . is assumed to be a Gaussian random variable of mean zero and standard deviation . We assume that is not a function of the mean expression , but it is straightforward to extend the method to the more general case . ( A natural extension of the model is to consider where characterizes the measurement noise and is the characteristic size of the independent events of mRNA production leading to the total mRNA number . ) We aim at maximizing the likelihood of the observations within a family of harmonic functions of period . Bayes Theorem for the probability of a particular model ( i . e . set of parameters ) given the data states: ( 1 ) We neglect the term , , corresponding to prior knowledge of the parameters , as there is no obvious choice for what such prior knowledge should be; moreover , with sufficient data , including such priors generally has little effect on the results of optimization . The probability of the data is a constant normalization factor , and so does not affect the relative likelihood of models . Therefore the probability of the model given the observations is proportional to the probability of the observations given the model . For each FISH observation , one therefore has: ( 2 ) and for the combined likelihood over all observations , ( 3 ) where the product runs over all FISH observations . In what follows we maximize assuming harmonic oscillations of mRNA levels , ( 4 ) The method can be systematically extended to periodic trajectories that are not simple sine waves by including higher harmonics . It is also straightforward to extend the method to more detailed noise models . For example , non-Gaussian noise can be incorporated by appropriately modifying the Gaussian integrand in Eq . ( 2 ) . Similarly , global transcriptional noise [25] can be modeled in Eq . ( 2 via a single additional random variable multiplying both and . Later , we consider both higher harmonics and global noise in detail for the more physiologically relevant case of bursty mRNA production . We generate synthetic FISH data by first choosing the parameters in Eq . ( 4 ) for the oscillating mRNA levels , and then generating FISH observations based on these parameters . Specifically , we choose random variables uniformly on , for genes . We then define the model parameters in Eq . ( 4 ) as , and . This construction ensures the positivity of the mRNA levels , and also ensures that the genes considered oscillate in time with significant amplitudes . The noise amplitudes are random variables , distributed continuously , . The synthetic FISH data are generated by choosing for each gene pair , random times and random noise values , and constructing . In this way , the synthetic data correspond to a set of independent , pairwise FISH observations . An example is shown in Fig . 2 for . The red ellipse indicates the true mean-mRNA-level trajectory , and the crosses are the randomly generated FISH data points . The blue ellipses correspond to reconstructions of the mean trajectory via maximization of the likelihood in Eq . ( 3 ) . To test the accuracy of reconstruction of mRNA dynamics using our MLE approach , we generated a large number of sets of parameter , and for each parameter set generated synthetic FISH data and then applied MLE to reconstruct the true dynamics . Specifically , for each parameter set defining a trajectory of mean mRNA levels , we maximized the likelihood with respect to . To ensure that we always found the global maximum of the likelihood , the initial guess for the parameters was taken to be the true parameters describing the mean dynamics . ( In Methods , we present a simple algorithm that almost always finds the global likelihood maximum without prior knowledge of the true parameters . However , in Fig . 2 , we chose to present the true MLE optimum as the fundamental limit of reconstruction accuracy , not limited by a particular algorithm . ) As shown in Fig . 2 , with synthetic FISH data for only two genes ( dashed blue ellipse ) the reconstruction is rather poor , in this case mistakenly assigning too large a noise to each gene and missing the phase shift . However , the addition of pairwise information from two more genes to make ( for a total of gene pairs ) is enough to correct these errors and provide a very accurate reconstruction . Since each pairwise data set is independent , the total amount of data grows as . To quantify the accuracy of the MLE algorithm , we computed the reconstruction error , which characterizes how much the reconstructed dynamics varies from the true mRNA dynamics , ( 5 ) where is the MLE reconstructed trajectory for mRNA , and is the ( true ) average number of mRNA over the period . Results for the reconstruction error are shown in Fig . 2B for , , and . Each point is averaged over 20 randomly generated parameter sets . As expected , the results improve with the number of genes and the number of FISH observations per gene pair , but at this noise level the results are already good for and . We now consider the bursty regime where a cell will typically either have few ( or no ) mRNAs of a particular type , or the mRNAs present will come from a single recent burst of transcription . In this limit , the information provided by FISH is essentially binary - either mRNAs for a particular gene are present at significant levels , indicating a recent burst , or they are not . Formally , if a significant number of mRNAs for gene are present , then , otherwise . The optimal threshold to set for the “presence” of mRNAs will depend on burst size and duration , measurement noise , and the total number of FISH observations – see Discussion . FISH data yields an estimate for the mean probability that mRNAs are present above the threshold value for each gene ( in the expression for , the variable reports the absence or presence of mRNA for observation of the pair , and the sum is made on the observations that probe for mRNA of gene . ) The FISH data also yields an estimate for the covariance for each pair of genes . We aim to accurately reconstruct the mRNA dynamics from these quantities and , which capture all the information provided by the binarized FISH data in the bursty regime . However , even with perfect knowledge of mean expression and covariance , the reconstruction of mRNA dynamics has fundamental limitations in this regime . We illustrate by considering both cyclic dynamics and stochastic switches . We denote by the probability that the number of mRNAs of type present at time is larger than some threshold , and we call such an event a burst in what follows . Assuming is any periodic function with period , it can be expanded in harmonics: ( 6 ) with more harmonics generally required to capture more complex oscillation patterns . Note that if is twice the number of harmonics considered , the number of parameters is , the coming from the invariance with respect to the overall phase . For the following discussion it is sufficient to keep only the first two harmonics , shown explicitly in Eq . ( 6 ) . In this case , denoting by the average over a cycle , one finds: ( 7 ) ( 8 ) where and denote the true cycle-averaged mean and covariance , respectively . One immediately sees that the transformation for all leaves both and unchanged . Thus , in this bursty regime , pairwise FISH data alone cannot disentangle different harmonics without prior knowledge . However , any additional constraint , including even a single triplet FISH data set , can readily resolve the ambiguity between harmonics . ( A triplet FISH observation , i . e . simultaneous measurement of three different mRNA types , leads to terms , which do not have the problematic symmetry . ) For simplicity , let us now consider only the lowest harmonic , as in the previous section . We introduce the -dimensional vectors and , defined as and . Each component can vary independently of the others , as there are parameters , and coordinates for the two vectors . Then by inspection the covariance matrix from Eq . ( 7 ) can be written: ( 9 ) which shows that is in general of rank 2 . If the second harmonics are included , is of rank 4 , etc . Note that all symmetric matrices of rank can be written in the form of Eq . ( 9 ) , with eigenvectors , implying that for a covariance matrix of even rank there is always an interpretation of the dynamics in terms of cyclic trajectories . Unfortunately , this interpretation is not unique except for the case of a single harmonic ( ) , which can be seen as follows . The observed mean probabilities for mRNA bursts of each type leads to constraints . The observed covariances lead to an additional constraints . Being a symmetric matrix of rank , the covariance matrix can be defined by this many coefficients , i . e . the number necessary to describe the eigenvectors , enforcing orthogonality among them . The expression for the number of covariance constraints is true for sufficient large , but in general is . ) The total number of constraints provided by FISH is therefore . Thus the number of unconstrained parameters is , which is zero for , but is already 5 for . Hence , for two harmonics at least 5 triplet FISH data sets or other constraints are required to be able to infer all the parameters . An important consideration in analyzing FISH data is that overall transcription rates may vary from cell to cell . Indeed , measurements of gene-expression noise in single yeast cells at the protein level reveal global fluctuations [25] . How can the dynamics of bursty gene expression be reconstructed against the background of these global correlations ? We consider the case of a simple harmonic cycle . The probability that the number of mRNAs of type present at time is larger than some threshold now reads: ( 10 ) where , representing the fluctuating global level of transcription , is a random variable of mean unity and standard deviation . One then obtains for the true cycle-averaged mean and covariance : ( 11 ) ( 12 ) Introducing the definitions , , and , the covariance matrix from Eq . ( 12 ) can now be written: ( 13 ) which shows that is now of rank 3 . If the second harmonics are included , is of rank 5 , etc . The maximum likelihood reconstruction for the model of Eq . ( 10 ) provides an estimate of the level of global transcriptional noise , as we show below for our reconstruction of metabolic cycles in yeast . We now consider a model where the expression pattern can switch stochastically among distinct states , as illustrated in Fig . 3 . We assume all genes of interest switch their expression synchronously , consistent with control by a single transcription factor , and without delays , consistent with state lifetimes long compared to mRNA lifetimes . On average , each state occurs with probability . In state , the true probability that a burst of mRNAs of type is present is denoted . For such models , the number of parameters is therefore , taking into account that . One finds , following simple arithmetic: ( 14 ) ( 15 ) where is the state-averaged burst probability and is the covariance matrix . is a -dimensional vector of components . It is straightforward to see that , which implies that the different vectors are not independent . Together with Eq . ( 15 ) , this dependence implies that the covariance matrix is in general of rank . Thus , pairwise FISH data alone cannot distinguish a 3-state model from simple harmonic dynamics ( or a 5-state model from a cycle including second harmonics , and so on ) . Moreover , even if one assumes that the dynamics is a switch , the parameters cannot be resolved uniquely: the number of constraints set by the measured means and covariances is , so that the number of unconstrained parameters is , which is 1 for a 2-state switch , and 3 for a 3-state switch . The corresponding number of triplet FISH data sets or other constraints are therefore required for parameter inference; however , if this additional data is available , switching parameters can be inferred even in the presence of global noise , as discussed above for the case of a simple cycle . For either cyclic or switching dynamics , maximum likelihood parameter estimation in the regime of bursty mRNA production requires the following steps , ( 1 ) estimating the mean burst probability and covariance from the FISH data , ( 2 ) determining the uncertainty of these estimates , and ( 3 ) obtaining the parameters for which the observed data is most likely . Taking an average over FISH data provides an estimate of the cycle- or state-averaged probability for a burst of mRNAs of type to be present . Specifically , , where reports the absence or presence of mRNA for observation of the pair , and where the sum is made on the observations that probe for mRNA of gene . Similarly , FISH data provide an estimate of the covariance , namely . For a finite number of data points , these estimates will be noisy , i . e . and , where the right hand sides are the exact values . Since coincident bursts of mRNAs of type and will be rare , the covariance estimate from finite FISH data may deviate significantly from the true covariance , and one must allow for this uncertainty in the maximum likelihood calculation . In contrast , one may safely neglect the uncertainty in the FISH estimate of the mean burst probability , both because single mRNA bursts are much more frequent than coincident bursts , and because each mRNA type is probed times more frequently than each pair . In practice , we therefore demand that the MLE parameters yield exactly . To estimate the uncertainty in the covariances , we first note that the true variance in is ( 16 ) where the overbar indicates the cycle or state average . We can then estimate the relevant quantity from the data , , to obtain an estimate for the variance . Using this estimate for , the probability of obtaining a covariance estimate if the true covariance is is given by: ( 17 ) Since the observations for the different mRNA pairs are independent , the likelihood of the observed covariance estimates for a given set of parameters is readily obtained from Eq . ( 17 ) . As discussed above , for bursty mRNA production the means and covariances alone cannot distinguish cyclic from switching dynamics . However , if one has prior evidence that gene expression is cyclic , maximum likelihood estimation can be usefully employed to reconstruct the dynamics . If and for a sufficient data , the algorithm works well , as illustrated in Fig . 4 for harmonic dynamics . A larger data set is needed than in the continuous mRNA regime because observations of coincident bursts are rare . Note that for , the constraints from the observed means and covariances are fewer than the parameters , and reconstruction requires additional constraints , e . g . from triplet FISH data . A stochastic switch between 2 states implies a covariance matrix of rank 1 , and therefore can be distinguished from cyclic dynamics , which leads to a minimum rank of 2 ( unless all the genes are exactly in phase ) . Still , one piece of additional information is required to reconstruct the dynamics . For example , it is sufficient to know the expression level of a single gene in one state . Here , we instead assume that the probability of being in one state is known , and given that constraint we infer all the levels of gene expression from synthetic FISH data . ( Note that FISH data can only reveal the probabilities to be in each state , not the kinetics of switching , e . g . interval durations or branching ratios . ) The MLE algorithm works well , as shown in Fig . 3A , as long as and for sufficient data . To quantify the accuracy of the MLE parameter estimation for switching dynamics , we have plotted in Fig . 3B the reconstruction error which measures the deviation of the reconstructed rates from the true rates ( normalized by the state-to-state variation and weighted by the state probabilities ) and averaged over all measured genes: ( 18 ) where the are the reconstructed rates . In principle , with enough FISH data it should be possible to reconstruct more than just the probability of observing a burst . For example , the entire distribution of mRNAs of each type in each switching state could be obtained using MLE , e . g . via Expectation Maximization ( EM ) [26] , by treating the full distributions rather than just the mean burst probabilities as unknowns . However , the approach proposed above of thresholding and binarizing the data has the advantage of reducing noise , and thereby reducing the required number of FISH observations , while still allowing for inference of the basic gene-expression dynamics . In the regime of bursty mRNA production , all of the information from FISH is contained in the mean burst probabilities and the covariance matrix , suggesting that Principal Component Analysis ( PCA ) could be usefully applied . For example , for a 2-state switch the covariance matrix has rank 1 . Thus , according to Eq . ( 15 ) , performing PCA by diagonalizing directly yields , the vector of differences of burst probabilities between the two states , as the only eigenvector with a non-vanishing eigenvalue . Together with the mean burst probabilities , , this yields full information on the switching dynamics . One caveat is that all the diagonal terms are missing from the estimated covariance matrix , as one cannot obtain an estimate of directly from FISH data . To solve this problem , we initially diagonalize the matrix with a zero diagonal , and obtain the principal eigenvalue and eigenvector . We then approximate the diagonal terms of with the diagonal terms of the rank-1 matrix built using this single eigenvector . We repeat this procedure iteratively to convergence , and take the converged principal eigenvector as an estimate of . This PCA approach performs similarly well to MLE for the case of a 2-state switch , as shown in Fig . 3 . The PCA approach can be easily extended to cases in which has a higher rank , where it also performs well , see Fig . 4 . Of course , like MLE , PCA has the same fundamental limitations discussed above that are inherent to coincidence detection . In practice , elements of the PCA and MLE approaches can be usefully combined . The main utility of PCA lies in diagonalizing to infer its rank . ( The iterative approach to filling in the diagonals of can help refine this procedure . ) From the rank of , one has a direct estimate of the “complexity” of the dynamics . Complexity here means the number of states in a switch model , or the number of harmonics to be considered for cyclic dynamics . This suggests the following heuristic approach to FISH data analysis: First diagonalize . Then isolate a group of eigenvalues that are significantly larger than the rest . Use prior information to select between the different models ( cyclic or switching ) leading to such a rank , and finally compute the model parameters using maximum likelihood estimation . In recent years , McKnight and coworkers demonstrated that the yeast Saccharomyces cerevisiae grown in chemostats can undergo synchronized metabolic oscillations [13] , [27] . As shown in Fig . 5 , the mRNA levels of three clusters of genes – Oxidative , Reductive Building , and Reductive Charging – were found to cycle together , with the expression of each cluster peaking at a different phase of the cycle . These population-level chemostat studies raise the question - is there an intrinsic metabolic cycle in individual cells in unsynchronized cultures ? To address this question , in [14] FISH data were obtained from single , unsynchronized yeast cells . Specifically , correlations of mRNA levels were determined for pairs of genes , each of which cycled in the chemostat . The correlations observed in single cells closely matched those found in the chemostat studies , leading to the conclusion that metabolic oscillations do occur in individual cells in unsynchronized populations as well as in synchronized chemostats . However , in [14] no attempt was made to go beyond correlations to reconstruct the dynamics . Here we use MLE to infer metabolic gene dynamics in unsynchronized populations . Our results support the conclusion of [14] that the gene clusters observed in the chemostat persist in individual cells in unsynchronized cultures . In particular , we find that the genes of the Oxidative cluster oscillate together and so do the genes of the Reductive Building cluster . The situation is still unclear for the Reductive Charging genes , but is likely to be clarified by additional FISH data . To analyze the dynamics , we first binarized the FISH data of [14] as appropriate for bursty gene expression . The data consists of 79 pairwise FISH experiments involving a total of 25 genes . To set an appropriate binary threshold of expression for each gene , we found the median of the mRNA distribution for each gene . Only 7 genes have a median larger than zero ( and in all cases ) , indicating that most genes are indeed bursty – despite the fact that those 25 genes were selected , among other criteria , to have a high expression level [14] . denotes the probability that the number of observed mRNA of gene is strictly larger than and is directly measurable from the data . We found , with the lower range coming from genes for which the median . We assumed that the dynamics is cyclic and considered the expansion of Eq . 6 up to the first harmonic . Such a model has 74 independent parameters for 25 genes . Moreover , the number of data points per pair of genes varies from 175 to 16032 , with only 29 pairs having more than 2000 data points . Thus some of the correlations are well-characterized , but others are not . If only the 29 gene pairs with more than 2000 data points are considered , even a single-harmonic model is under-constrained . To circumvent this problem , we are guided by the observation apparent from Fig . 5 that the gene expression in all clusters becomes much smaller than its mean at some point in time . This suggests a simplified model where the probability of expressing more mRNAs than the median mRNA number for gene cycles as: ( 19 ) Therefore , once is extracted from the data there is a single free parameter per gene , namely its phase . Next , the likelihood of all the observed FISH correlations was maximized with respect to the phases . The global maximum was found by considering various random initial phases , relaxing to a maximum , repeating , and choosing the maximum with the largest likelihood . We consistently found the same maximum after the order of 10 optimization runs . Results for the reconstructed dynamics are shown in Fig . 6 for the 14 most tested genes ( per gene number of observations ) . Genes belonging to each metabolic cluster identified by the chemostat studies are represented by distinct colors as indicated in the legend . The location of the maximum probability for each gene is indicated by an arrow . From the positions of the arrows it is apparent that genes belonging to the Oxidative cluster also cluster in an unsynchronized population , and so do the genes of the Reductive Building cluster . From the existing data we cannot yet conclude whether the Reductive Charging genes also cluster . To quantify our results statistically , we define for each cluster , , the quantity , where is the number of genes in cluster . The , which characterize the average cluster activity , are plotted in Fig . 6 . If the genes belonging to a cluster are perfectly synchronized , i . e . are identical for all , then will reach zero along the cycle . More generally , the lower the minimum of , the more synchronized the cluster is for fixed . We find that the Oxidative and Reductive Building genes are indeed clustered: the probability of finding such low minima for the two corresponding curves would be only and respectively ( when considered together ) if the phases were random . On the other hand , the minimum of the Reductive Charging cluster is comparable to the typical value for random phases . From the chemostat studies [13] , we expect the amplitudes of oscillation of metabolically cycling genes to be large ( 10 fold ) , and so global transcriptional noise ( 2 fold [25] ) should not significantly affect our results . However , to test that our reconstruction of the metabolic cycle is robust with respect to global transcriptional noise , we reconstructed the dynamics allowing for a global correlation among mRNA levels as in Eq . ( 10 ) . Specifically , we extended the model of Eq . ( 19 ) by adding the possibility of a varying global level of transcription : ( 20 ) where is a random variable of mean unity and standard deviation . The results of the reconstruction are essentially identical to those shown in Fig . 6 , where the global level of transcription was assumed to be fixed ( for a comparison see Fig . S1 ) . Moreover , from the reconstruction we infer the amplitude of the global noise of transcription to be ( i . e . 55% ) , which is significant , but considerably smaller than the typical variation during a cycle [13] . In Fig . 6 , the lack of evidence for coherent oscillations of the Reductive Charging genes may reflect a real feature of unsynchronized populations . Alternatively , it may reflect the limited data and/or the simplicity of the model of Eq . ( 19 ) . To investigate the limitations of this model we considered how the pairwise gene covariances it predicts compare with the observed FISH covariances , as shown in Fig . 7 . Even the underlying “true” model should not capture the FISH correlations perfectly , especially since some observations are very noisy due to the limited data . However , some general trends appear . In particular our model in Eq . ( 19 ) systematically underestimates the largest covariances . This may be due to the fact that the single cosine wave that we use to fit the dynamics is less peaked than the typical expression profile observed in the chemostat [13] . Accordingly higher harmonics should be included to obtain a more accurate description of the gene-expression dynamics , an approach that will be achievable once the data set is enlarged to include additional gene pairs .
In general , Maximum Likelihood Estimation ( MLE ) requires finding the set of model parameters for which the observed data are most likely . Finding the global maximum in the space of model parameters can be a challenging task , particularly as there may be many local maxima in which a search algorithm can get stuck . For synthetic FISH data in the regime of continuous mRNA production , we found that such local maxima occurred frequently . ( In contrast , for synthetic FISH data in the bursty regime a simple steepest-descent algorithm invariably found the same maximum , independent of initial conditions . ) To find the global maximum in the continuous regime , we developed a heuristic algorithm that worked very well in practice to reconstruct simple cycles . One approach is to consider various initial parameter values , and to use a steepest-descent algorithm to find the local maximum of the likelihood . Then the global maximum ( with the highest likelihood ) could be chosen among the different solutions . However , in practice this procedure can be very time-consuming if initial conditions are chosen randomly . Here we propose two approaches to first compute estimates of the parameters , and then use these estimates to initiate the optimization protocol . In these two approaches we estimated the parameters as follows: ( 1 ) For the mean expression level we took . ( 2 ) For both the amplitude of oscillations and the noise amplitude we took half the standard deviation of the observations of the corresponding gene . ( 3 ) Empirically we found that the initial choice of phase is critical in determining if the global or only a local maximum is found . Therefore , to accurately estimate the relative phases we introduced the Pearson correlation matrix ( a normalized variant of our covariance matrix ) . This definition implies . yields a rough approximation of , which leads to the following two approximations , the first being extremely crude: Results of optimization using approaches ( a ) and ( b ) to set initial parameter values are shown in Fig . 8A . The results in Fig . 8A are nearly indistinguishable from those obtained using the true parameters as initial conditions , shown in Fig . 8B , demonstrating that the above protocols performs well in identifying the global maximum of the likelihood .
The ability to count mRNA molecules in single cells by Fluorescence In Situ Hybridization ( FISH ) [9]–[11] allows for highly quantitative studies of cell-to-cell variation in gene expression . However , the requirement that cells be fixed before RNA FISH analysis precludes the use of RNA FISH to directly study transcriptional dynamics in single cells . Nevertheless , we have shown here how and when correlations between levels of different mRNAs can be exploited to reconstruct transcriptional dynamics , even if cells are asynchronous . All that is necessary is for FISH data to be obtained simultaneously for pairs of genes ( or in some cases triplets of genes ) a technique that is already well established [10] , [12] . As a practical demonstration , we applied our approach to a large , pairwise FISH data set obtained from a recent study of the yeast Saccharomyces cerevisiae [14] . Our results help confirm the existence of cell-autonomous metabolic cycles in unsynchronized yeast populations [13] . To reconstruct the dynamics of gene expression from FISH data , our approach employs Maximum Likelihood Estimation ( MLE ) [20] to obtain the set of transcriptional parameters most likely to account for the observed data . In the regime of continuous mRNA production , apart from rescaling and inversion of time for cyclic dynamics , there is no intrinsic limit on the accuracy with which transcriptional dynamics can be reconstructed given enough data . In practice , we have shown that MLE applied to simple parameterizations for transcription ( such as the leading harmonics for cyclic dynamics ) allows faithful reconstruction from a moderate number of FISH observations , including noise . On the other hand , the regime where mRNA is produced in shortlived bursts [10] presents additional challenges . In this bursty regime , FISH can at most report coincidences of bursts of different mRNAs , and there are consequently fundamental limits to reconstructing the underlying dynamics . For this bursty regime , successful reconstruction will generally rely on prior knowledge regarding the class of dynamics , e . g . cycle vs . switch , and , even so , will in some cases require additional inputs , such as triplet FISH data . ( In Table S1 , we explicitly quantify the amount of such additional information required for complete dynamical reconstruction . ) In applying our approach , how should one choose among models to reconstruct gene dynamics ? For example , when is it better to use multiple harmonics instead of a single harmonic to model a cycle ? The answer depends on the type of data . We discuss first the regime of continuous mRNA production . For this case , a standard and reliable way to choose among models when fitting data is “leave-one-out” validation , which both rewards a good fit while punishing overfitting . In the leave-one-out approach , a model is selected and its parameters are optimized on the entire data set , but with one data point left out . The resulting parameterized model is then used to fit the neglected data point . The average fitting error , taken over all possible left-out data points , is a robust measure of the quality of the model . Among competing models , the one that minimizes this error can be selected as the better choice . In the regime of continuous mRNA production , leave-one-out validation can be applied within the MLE framework by using the log ( likelihood ) of the left-out data point in place of the fitting error . Among competing models , the one with the largest average log ( likelihood ) is the best choice . In contrast , finding the “best” model for data in the bursty mRNA regime is generally an under-constrained problem . We showed explicitly that for many cases it is impossible in principle to distinguish among different types of models , or even to find a unique best set of parameters for a given model . Intuitively , reduction of bursty FISH data to pairwise covariances means that even as the number of FISH data points approaches infinity , the number of model constraints stays finite . So , for bursty FISH data inference alone cannot guide one in choosing the model , and one must also use common sense . Clearly , prior knowledge of the system under study should be used in selecting a model . In addition , a simple rule is that one should use models that are sufficiently parsimonious in parameters not to have degenerate solutions . For example , in analyzing FISH data on metabolic cycles , we chose the one-harmonic model because there were not enough low-noise covariances to constrain a two-harmonic model . More generally , it is advisable to choose a model with significantly more well-constrained data than parameters . If the model is barely constrained , the peak of likelihood will generally be close to flat in some directions in parameter space and the reconstruction will be poor . Figure 4 illustrates this point: is the minimal number of genes to avoid degeneracy , but it requires 3 times more data per gene ( or twice as many total data points ) to reconstruct as well as for . In practice , one test for the quality of the reconstruction in the bursty regime is to compare the observed covariances to the reconstructed covariances , as shown in Fig . 7 for the case of the yeast metabolic cycle genes . Reconstruction of gene-expression dynamics from FISH data presents multiple practical challenges . One important issue is noise in the measurement of mRNA levels . For the regime of continuous mRNA production , we have shown that sufficient data can compensate for both the noise inherent in gene expression and the noise arising from uncertainty in measurement . For the regime of bursty mRNA production , “binarizing” the data into the presence or absence of a significant number of mRNA molecules substantially reduces the impact of measurement noise . A practical question here is the best threshold to use for binarizing the data . In many cases , the dynamics will be best reconstructed by setting the threshold well above 1 mRNA transcript; for example , in treating the data for metabolic cycles we chose the median expression level for each gene as its threshold . A higher threshold is less sensitive to measurement noise ( fewer false positives ) , and to occasional transcripts produced by promoter leakage ( better identification of true bursts ) , and a higher threshold also allows finer time resolution , as a given burst will remain above threshold for a shorter time ( e . g . preventing blurring of boundaries between switching states ) . However , a higher threshold reduces the number of coinciding bursts in the data , requiring more overall FISH observations . An important related issue is the possibility of correlated noise in the transcription of different genes . An example of such noise is the observed global correlation among transcription rates in yeast [25] . Fortunately , global noise can be readily incorporated within the MLE framework by introducing a single additional variable in the model for gene expression , as in Eq . ( 10 ) . Indeed , our treatment of global noise among genes involved in the yeast metabolic cycle yields an independent , and reasonable , estimate for this noise at 55% of mean expression . ( More complex noise correlations among different genes would require case-by-case analysis . ) False-positive rates and false-negative rates are also both important considerations in analyzing FISH data . These are essentially technical issues beyond the scope of our study , but a few remarks are in order . In Ref . [14] , both false positives and false negatives were reduced by the use of multiple fluorescent probes ( 5 ) for each mRNA . Only high-contrast spots above a fluorescence threshold indicative of multiple bound probes were counted . This threshold was set empirically from the fluorescence distribution of spots outside of cell boundaries , corresponding to single probes . Nevertheless , with any such thresholding method , there will be cases where the “presence” or “absence” of an mRNA is ambiguous , and in the bursty regime such ambiguities can strongly impact the binarization of the data . Fortunately , because MLE is an intrinsically probabilistic approach , ambiguities can be dealt with by treating the two possibilities , present or absent , probabilistically . As in Ref . [14] , by looking at spots outside of cell boundaries , one can obtain the distribution of intensities for spots that are actually noise ( typically single probes that have not been washed away ) , and by looking inside cell boundaries a similar distribution can be obtained for spots that correspond to real mRNAs ( multiple probes ) . Spots inside cells that fall into the region of overlap of these two distributions can then be assigned the corresponding probabilities of being present ( real ) or absent ( noise ) . MLE can then incorporate both possible interpretations of the data , with their appropriate weights , in the data set . A related issue , highlighted by Zenklusen et al . [11] , is the existence of nascent mRNA transcripts at the locus of the gene . In the regime of continuous mRNA production , an estimate of nascent transcript number , possibly non-integer , could simply be added to mRNA counts . In the bursty regime , the existence of such transcripts might well be taken as prima facie evidence for active transcription , and therefore treated as equivalent to the presence of an above-threshold burst . Another practical issue in reconstructing gene-expression dynamics from FISH measurements is that data may come in mixed forms , e . g . pairwise FISH data+triplet FISH data+additional constraints or prior information . Again , MLE is naturally suited to incorporating mixed data types since all sources of information can be combined to produce the overall likelihood of the data given a set of model parameters , including prior information on the model parameters themselves ( cf . Eq . ( 2 ) ) . While these and other practical issues are important to consider , our successful reconstruction of yeast metabolic cycles using the FISH data of Silverman et al . [14] demonstrates that our approach can provide a useful tool for analyzing gene-expression dynamics . In fact , our analysis of this data raises several new questions . First , since our reconstruction was statistically significant for the Oxidative and Reductive Building clusters but not for the Reductive Charging cluster , it is possible that cycles of the latter may be weaker in unsynchronized cultures than in synchronized chemostats . Second , our reconstruction indicates a spread among the oscillatory phases of genes within each cluster – is this spread a consequence of the limited data , or are the oscillation patterns of genes within clusters distinct ? We expect that additional FISH data coupled with MLE analysis will soon provide answers to these questions . The many advantages of FISH – absolute quantification , high time resolution , use of wild-type cells , ability to simultaneously measure multiple mRNA types , and broad application across species from bacteria [28] to yeast [11] , [14] to metazoans [9] , [10] , suggest that FISH will find many uses in future studies of gene expression , including applications beyond those currently demonstrated . For example , FISH can be applied to cells in structured environments such as tissues or biofilms , or even cells in mixed-species consortia . In all of these cases , population level studies of gene expression cannot reveal the important cell-to-cell variations . Of course , FISH is not the only technique that yields quantitative snapshots at the single-cell level . Immunofluorescence and single-cell sequencing also meet the requirements of simultaneous measurements of two or more intracellular factors . We hope that the analysis presented here can facilitate the application of FISH and other single-cell snapshot assays to cases where both cell-to-cell variation and the dynamics of gene expression are of central interest . | Programs of gene expression lie at the heart of how cells regulate their internal processes . Some dynamical gene-expression programs , such as the cell cycle , are well known and studied , others , such as metabolic cycles , have only recently been recognized , and many other dynamical programs including switches are likely to be discovered . Traditional bulk studies typically fail to resolve such cycles or switches , because individual cells are out-of-phase with each other . On the other hand , standard techniques for studying single cells are limited in time resolution and scope . RNA Fluorescent In Situ Hybridization ( FISH ) is a single-cell technique that offers both high time-resolution and precise quantification of mRNA molecules , but requires fixed cells . We have explored how , when , and with what prior information FISH snapshots of pairs of genes can be used to accurately reconstruct gene-expression dynamics . The technique can be readily implemented , and is broadly applicable from bacteria to mammals . We lay out a principled and practical approach to extracting biological information from RNA FISH data to reveal new information about the dynamics of living organisms . | [
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] | [
"computational",
"biology/transcriptional",
"regulation",
"biophysics/theory",
"and",
"simulation",
"computational",
"biology"
] | 2010 | Evaluating Gene Expression Dynamics Using Pairwise RNA FISH Data |
The ongoing global spread of Tomato yellow leaf curl virus ( TYLCV; Genus Begomovirus , Family Geminiviridae ) represents a serious looming threat to tomato production in all temperate parts of the world . Whereas determining where and when TYLCV movements have occurred could help curtail its spread and prevent future movements of related viruses , determining the consequences of past TYLCV movements could reveal the ecological and economic risks associated with similar viral invasions . Towards this end we applied Bayesian phylogeographic inference and recombination analyses to available TYLCV sequences ( including those of 15 new Iranian full TYLCV genomes ) and reconstructed a plausible history of TYLCV's diversification and movements throughout the world . In agreement with historical accounts , our results suggest that the first TYLCVs most probably arose somewhere in the Middle East between the 1930s and 1950s ( with 95% highest probability density intervals 1905–1972 ) and that the global spread of TYLCV only began in the 1980s after the evolution of the TYLCV-Mld and -IL strains . Despite the global distribution of TYLCV we found no convincing evidence anywhere other than the Middle East and the Western Mediterranean of epidemiologically relevant TYLCV variants arising through recombination . Although the region around Iran is both the center of present day TYLCV diversity and the site of the most intensive ongoing TYLCV evolution , the evidence indicates that the region is epidemiologically isolated , which suggests that novel TYLCV variants found there are probably not direct global threats . We instead identify the Mediterranean basin as the main launch-pad of global TYLCV movements .
Tomato yellow leaf curl disease ( TYLCD ) is one of the most devastating emerging diseases of tomato in the warm and temperate regions of the world . It is caused by a complex of at least six virus species in the Begomovirus genus of the Family Geminiviridae [1] , [2] . Tomato yellow leaf curl virus ( TYLCV ) is the most widely distributed and best studied of these species and the begomoviruses responsible for TYLCD are therefore collectively referred to as TYLCV-like viruses . Besides TYLCV , the TYLCV-like viruses include Tomato yellow leaf curl Sudan virus ( TYLCSDV ) , Tomato yellow leaf curl Axarquia virus ( TYLCAxV ) , Tomato yellow leaf curl Malaga virus ( TYLCMLV ) , Tomato yellow leaf curl Sardinia virus ( TYLCSV ) and Tomato yellow leaf curl Mali virus ( TYLCMLV ) [1] , [3] . Although TYLCV-like viruses were first described in the Jordan Valley in Israel during the early 1960s , disease symptoms resembling TYLCD ( stunted tomato plants with downward leaf curling , leaf discoloration and leaf deformation ) had been observed in the Jordan Valley since the late 1920s ( cited in [4] , [5] ) . In Israel during the early 1990s , two begomovirus strains associated with TYLCD infections of different severities were cloned and named Tomato yellow leaf curl virus–Israel ( TYLCV-IL; TYLCV-IL[IL:Reo:86]-X15656 ) and Tomato yellow leaf curl virus–Mild ( TYLCV-Mld; TYLCV-Mld[IL:93] – X76319 ) [5] , [6] . It was subsequently determined that TYLCV-IL was a recombinant of the TYLCV-Mld strain and another begomovirus species related to Tomato leaf curl Karnataka virus ( ToLCKV; [7] ) . TYLCV-IL contains mostly TYLCV-Mld like sequences but the 5′-portion of its rep gene is very ToLCKV-like . Other subsequently characterised TYLCV strains such as the Gezira ( e . g . TYLCV-Gez[SD:96] ) , Iran ( e . g . TYLCV-IR[IR:Ira:98] ) and Oman ( e . g . TYLCV-OM[Om:Alb:05] ) also display evidence of having arisen through unique , albeit similar , inter-species recombination events [8]–[10] . Of all the known TYLCV strains , TYLCV-IL and TYLCV-Mld have the broadest geographical ranges stretching in the Old world from Japan in the east [11] to Spain in the west [12] and the Indian Ocean island of Reunion [13] and Australia [14] in the south . Additionally , TYLCV-IL has apparently jumped at least twice between the Old and New Worlds [15] , [16] and is currently spreading into North and South America [17]–[19] . As the international trafficking of crop varieties is relatively widespread , it is perhaps not surprising that a virus like TYLCV-IL could attain such a global distribution . Nevertheless , amongst the geminiviruses , the TYLCV-IL geographical range is unusually vast . Given that the Mediterranean basin and the Middle East are clearly centers of TYLCV diversity [20] , it is probable that this is where these viruses originate . The region has a climate that favors tomato cultivation and collectively accounts for 30% of global tomato production ( FAOSTAT 2008 ) . It is of some concern therefore that recent reports have indicated a dramatic increase in TYLCD incidence within the region [10] , [21]–[24] . In Iran in particular where the climate has warmed and dried in recent years there has apparently been a steady increase in the incidence of whitefly transmitted geminivirus diseases in tomato crops [9] , [25]–[29] . Considering the high degrees of TYLCV diversity in the Middle East and the amount of inter-strain and inter-species recombination that has been detected between TYLCV and different Middle Eastern begomovirus species [9] , [10] , it is reasonable to suspect that virus evolution within this region has had , and will probably continue to have , a major impact on global TYLCD epidemiology . We therefore isolated and sequenced 15 new Iranian TYLCV isolates which were used along with publicly available sequences both to identify where TYLCV originated , and to retrace the virus' movement patterns around the globe . Together with detailed recombination analysis , we applied a newly developed Bayesian phylogeography method to infer where and when major events in the evolution of TYLCVs have occurred . In congruence with previous assumptions , our analysis clearly indicates both that the emergence and global spread of TYLCV have been extremely rapid , and that the Middle East in general , and the region surrounding Iran in particular , are probably the current and past centers of ongoing TYLCV diversification .
Samples from 27 tomato plants displaying typical TYLCD symptoms ( upward leaf curling , yellowing , distortion , and stunting ) were collected in the major tomato producing regions of Southern Iran ( Kerman , Hormozgan , Bushehr and Fars provinces ) in 2006 and 2007 ( Table S1 ) . Total DNA was extracted from the fresh or dried leaves using High Pure Viral Nucleic Acid Extraction Kit ( Roche , Germany ) according to the method described by the manufacturer . DNA-B and DNAβ molecules that are commonly found within begomovirus infections were tested for using the primer pairs PBL1v2040/PCRc1 [30] , and Beta01/Beta02 [31] . Viral genomes were amplified from total plant DNA extractions using phi29DNA polymerase ( TempliPhi , GE Healthcare , USA ) as previously described [32] , [33] . Amplified genomic concatemers were digested with either XmnI or PstI to yield full length genomes ( ∼2 . 7 kb ) . The linearised fragments were either ligated to PstI digested pGEM 3Zf+ ( Promega Biotech ) or blunt-end ligated to the blunt cloning site of pJET1 . 2 ( CloneJET PCR cloning kit , Fermentas ) . Full genomes were commercially sequenced ( Macrogen Inc . , Korea ) on both strands by primer walking . Sequences were assembled and edited using dnaman ( version 5 . 2 . 9; Lynnon Biosoft ) and MEGA 4 [34] . The 15 new TYLCV genomes were aligned with all full-length begomoviruses , DNA-A and DNA-A-like sequences available in GenBank in July 2009 using POA v2 [35] . This alignment was edited by eye in MEGA 4 [34] with ∼595 poorly aligned alignment columns within the intergenic region being removed from all subsequent analyses ( the resulting alignment is available on request from the authors ) . Maximum likelihood phylogenetic trees were constructed with PHYML [36] with model GTR+G4 ( selected as the best-fitting model by RDP3; [37] and 1000 full maximum likelihood ( ML ) bootstrap iterations . Degrees of sequence identity shared by sequences were calculated using MEGA 4 with pairwise deletion of gaps . Detection of potential recombinant sequences , identification of likely parental sequences , and localisation of possible recombination breakpoints was carried out on using the RDP [38] , GENECONV [39] , BOOTSCAN [40] , MAXIMUM CHI SQUARE [41] , CHIMAERA [37] , SISCAN [42] and 3SEQ [43] recombination detection methods as implemented in RDP3 [37] . The analysis was performed with default settings for the different detection methods and a Bonferroni corrected P-value cut-off of 0 . 05 . Only events detected with two or more methods coupled with significant phylogenetic support were considered credible evidence of recombination . The breakpoint positions and recombinant sequence ( s ) inferred for every detected potential recombination event were manually checked and adjusted where necessary using the extensive phylogenetic and recombination signal analysis features available in RDP3 . The movement patterns of TYLCV over the past century were reconstructed using a recently developed approach that , given a set of sequences sampled from various discreet locations ( such as individual cities , countries or other geographical regions ) over a few decades , models changes in geographical location during the evolution of the sequences [44] . This fully probabilistic approach , implemented in the computer program , BEAST v1 . 5 . 3 [44] , draws on an explicit model describing how , during the evolution of the sampled sequences since their last common ancestor , the unknown geographical locations of ancestral sequences have changed between the known locations of these sampled sequences . In a process that is very similar to that used to infer ancestral nucleotide sequences , the methodology employs continuous-time Markov chain models of discrete state evolution ( meaning that rather than the individual GPS coordinates of each sequence being considered , all the sequences from the same approximate region are assigned the same region state ) to determine the most probable geographical locations of ancestral sequences . Besides inferring where amongst the sampling locations ancestral sequences most likely resided , the method additionally provides a statistically meaningful measure of the over-all confidence that can be associated with movements between any two of these locations . This is achieved by using a so-called Bayesian stochastic search variable ( BSSV ) procedure [44] which is associated with a Bayes factor [45] , [46] test that can be used to identify the best supported movement routes between the various geographical locations considered . Following the results of Duffy and Holmes [47] we assumed a constant population size tree prior and a log-normal relaxed molecular clock for our TYLCV phylogeographic analyses . Individual BEAST runs were performed with 200 million steps in the Markov chain and sampling every 10 , 000 steps to produce a posterior tree distribution containing 20 , 000 genealogies . Similar results allowed us to combine log and tree files using LogCombiner ( available in BEAST package ) . The maximum clade credibility tree ( a point estimate of the tree with the highest cumulative posterior probabilities in the posterior distribution of trees ) was annotated with geographical locations using the software TreeAnnotator ( available in BEAST package ) . We used tools available from http://beast . bio . ed . ac . uk/Google_Earth to produce a graphical animation in key markup language ( kml ) file format of the spatio-temporal movement dynamics of ancestral TYLCV sequences . These kml files , available as Dataset S1 and Dataset S2 , contain information on routes and times of virus movements can be viewed using Google Earth ( available from http://earth . google . com ) . Two temporally structured TYLCV datasets ( sampling dates spanning from 1988 to 2009 ) were analysed ( see Table S1 for details ) . Whereas the first , contained 82 full TYLCV genomes and was called the FG dataset , the second contained 91 ∼940 nt long TYLCV sequences corresponding to genome positions 148–1090 in isolate TYLCV-IL[IL:Reo:86] ( accession number X15656 ) and was called the CP dataset . While the FG dataset contained substantial evidence of inter-species recombination ( particularly in the sequences encoding the complementary sense genes ) , the CP dataset was mostly free of detectable recombination and contained absolutely no evidence of inter-species recombination . Therefore , although it contained fewer phylogenetically informative sites , analyses of the CP dataset were expected to be free of the confounding effects that recombination in the FG dataset might have on estimates of substitution rates and sequence divergence times [48] , [49] . Using the sampling coordinates and a freely available hierarchical clustering method ( called “hclust” ) implemented in R [50] , we were able to optimally define groups of sequences displaying definite geographical clustering . Longitude and latitude coordinates at the centroids of each of the groups thus defined , were used as the discrete sampling locations in our phylogeographic analyses . The sequences in the FG and CP datasets were respectively grouped into seven and nine of these discreet sampling locations ( see Table S1 for details ) . It is important to stress that despite the fact that the dendrogram constructed during the geographical clustering analysis superficially resembles a phylogenetic tree , the groupings depicted by the dendrogram are based entirely on relative geographical proximity and not on relative sequence similarity and as a result the clustering methods could have in no way confounded our subsequent phylogeographic analyses . Whereas for the FG dataset similar numbers of sequences ( between 8–24 ) were sampled from the various locations considered ( the exceptions are Reunion and the Horn of Africa with only 2 and 1 samples respectively ) , there were quite significant sampling biases in the CP dataset with substantially more sequences having been sampled from Iran ( ∼33% ) relative to the other locations considered . We used two separate tests to assess the consequences of such sampling biases on our analyses . In the first test we “equalised” the sample sizes for all locations from which more than eight sequences had been sampled by randomly sub-sampling eight sequences from each of these . For each of ten smaller datasets thus constructed from both of the FG and CP datasets ( the FG-based datasets contained 51 sequences and the CP-based datasets 42 sequences ) we performed the same phylogeographic analyses as those described above . In the second test , the analysis was also carried out as above but the location states of the sequences were randomized using an additional operator in the MCMC procedure ( BEAST can be set up to do this ) . The location state probabilities of the root node determined during these analyses were compared with those determined for the datasets analysed without the location state randomization setting . Based on the dated maximum clade credibility ( MCC ) trees constructed from the temporally structured FG and CP datasets and the parental and recombinant sequences identified in our recombination analyses we could determine the approximate dates when recombination events occurred and pinpoint the geographical locations of the ancestral recombinants . For each detected recombination event we first constructed a neighbour joining tree based on the TYLCV derived sequences found within the recombinant ( using a Jukes Cantor nucleotide substitution model in RDP3 ) . The date ascribed to the corresponding node in the dated MCC tree that marked the branching point of the recombinant sequence ( s ) ( in many cases there were multiple sequences descended from a single ancestral recombinant ) was taken to be the earliest date when the recombination event could have occurred ( with the earlier bound of the associated 95% highest probability density , or HPD , indicating the lowest credible bound of this estimate ) . This “lower” node essentially represents the most recent common ancestor of the recombinant ( s ) with a non-recombinant . In cases where multiple sequences appeared to bare traces of the same ancestral recombination event , the date associated with the MCC tree node representing the last common ancestor of the recombinant sequences was taken as being the latest probable date when the recombination event might have occurred ( with the upper bound of the associated 95% HPD indicating the upper credible bound of this estimate ) . This “upper” node represents the most recent common ancestor of the recombinants . To determine the approximate geographical location of where recombination events might have occurred the inferred geographical locations of sequences at these “lower” and “upper” nodes were assumed to bound the location where the recombination event in question occurred . In cases where only a single sequence carried evidence of a recombination event , the latest date of the recombination event and the upper bound of the 95% HPD of this date were taken as the sampling date of the sequence . In such cases the “upper” bound on the geographical location where the recombination event may have occurred was simply taken to be the sampling location of the recombinant .
We collected samples showing TYLCD symptoms in the provinces of Kerman ( Kahnooj , n = 4; Jiroft , n = 5 Orzuiyeh , n = 1 ) , Fars ( Shiraz , n = 6; Lar , n = 1 ) , Yazd ( Taft , n = 1; Ashkezar , n = 1 ) , Hormozgan ( Roodan , n = 4; Minab , n = 3 ) and Bushehr ( Borazjan , n = 1 ) and cloned and determined full-length DNA-A-like sequence from 15 of these ( Kahnooj , n = 2; Jiroft , n = 4; Orzu'iyeh , n = 1; Shiraz , n = 3; Taft , n = 1; Roodan , n = 1; Minab , n = 2; Borazjan = 1 ) . No DNA-B or Beta molecules were detected in any of the analysed samples . Phylogenetic analysis and pairwise genome-wide similarity comparisons between these 15 new sequences and those deposited in sequence databases ( Figure 1 and Figure S1 ) indicated that five were TYLCV-IL isolates , five were TYLCV-OM isolates , four were TYLCV-Ker isolates and one was an isolate of a potentially new strain that we have tentatively named TYLCV-Bou . TYLCV-Bou represents a new strain based on the currently accepted geminivirus strain demarcation criteria [3] in that it shares 92 . 5–94% identity with TYLCV-Ker isolates ( Figure S1 ) . Different isolates from the individual strain groupings displayed minimal evidence of geographical clustering within Iran ( see Figure S2 ) . It is noteworthy that five of the seven described TYLCV strains are found in Iran . This is a greater number than have been found in any other country ( the next highest is two ) - a fact which marks Iran as probably being close to the global center of TYLCV diversity . As recombination is a major process influencing the evolution of TYLCV and other begomoviruses we analysed 75 TYLCV full length DNA-A-like sequences together with 658 DNA-A and DNA-A-like sequences belonging to other begomoviruses for evidence of ( 1 ) TYLCV sequence fragments being transferred into the genomic backgrounds of other species ( i . e . events with TYLCV donors ) and ( 2 ) the genomic fragments of other species being transferred into mostly TYLCV-like genomic backgrounds ( i . e . events with TYLCV recipients ) . Of the 18 detected recombination events involving TYLCV isolates , 16 were inter-species sequence exchanges ( events 1 to 16 in Table 1 and Figure 1 ) and two were intra-species exchanges ( events 17 and 18 in Table 1 and Figure 1 ) . Only four of the 16 inter-species recombination events involved TYLCVs as donors . The recipient species in these four recombination events were western Mediterranean TYLCSVs ( events 2 , 4 and 5 in Table 1 and Figure 1 ) and TYLCAxV ( event 6 in Table 1 ) isolates . As has been found previously , two of these events ( 2 and 4 in Table 1 ) , both involving TYLCSV as a recipient and TYLCV as a donor , were pivotal in the creation of the TYLCAxV and TYLCMalV species [51] , [52] . In fact , all three of the TYLCAxV isolates examined ( accession numbers AY227892 , EU734831 and EU734832 ) , appear to be independently generated convergent recombinants of TYLCSV and TYLCV-IL , highlighting the possibility that , in the Western Mediterranean at least , such recombinants have a high degree of fitness . The remaining 12 inter-species recombination events involved TYLCVs as recipients of <1000 nucleotide fragments mostly derived from the rep genes of either currently undescribed begomovirus species , or species previously detected only in the Middle East and/or India and Asia . The fact that unique recombination events are detectable within the rep sequences of every TYLCV isolate presents somewhat of a problem when it comes to disentangling the evolutionary origins of the various recombinationally derived fragments within this gene . Specifically , without a provably non-recombinant TYLCV rep gene in hand it is not possible to objectively judge the accuracy of the parental sequence and recombinant designations given in Table 1 and Figure 1 . Put another way , it is possible , if not probable , that some of the parental TYLCV sequences listed in Table 1 are misidentified recombinant sequences and some of the recombinant sequences are misidentified parental sequences . In this regard , parental and recombinant sequence designations for events 7 , 9 , and 11 listed in Table 1 were particularly difficult to interpret . Evidence of these recombination events is found within quite divergent TYLCV lineages implying that they either ( 1 ) predate the divergence of these lineages or ( 2 ) that they are more recent but that the recombinant fragments characterising the events have been propagated by secondary intra-species recombination between the various TYLCV lineages . For example , both the fact that event 7 is found within the TYLCV-Ker , -Mld , -Gez , and -Bou lineages and the evidence of it being overprinted by subsequent recombination events such as 14 in the -Mld lineage , 8 in the –Gez lineage , and 12 in the –Ker ( B ) lineage , implies that it is a reasonably old recombination event . With events 9 and 11 on the other hand , it is plausible that a secondary recombination event carrying a fragment baring traces of both events has been transferred from a TYLCV-IL ( A ) variant into the TYLCV-IL ( C ) variant ( Figure 1 ) . The young age of events 9 and 11 in some of the TYLCV-IL ( A ) isolates is also implied by how closely some of these isolates resemble TYLCV-Mld ( A ) isolates within the portion of their genomes upstream of the event 9 5′-breakpoint . For example , over a stretch of 1640 nucleotides the TYLCV-Mld[ES:Alm:99] isolate , and the TYLCV-IL[ES:Alm:Pep:99] isolate , differ at only two nucleotide positions – implying a very young age for the recombination event in rep that differentiates them . However , over this 1640 nucleotide fragment these two isolates are also much more closely related to one another than either is to any other TYLCV-Mld or TYLCV-IL isolates . This strongly suggests that after the original inter-species recombination event ( s ) that resulted in the differentiation of TYLCV-IL from TYLCV-Mld [7] , the TYLCV-IL fragment containing traces of events 9 , 11 and 10 has , at least once , been transferred back into a TYLCV-Mld isolate ( in this case , one very closely resembling TYLCV-Mld[ES:Alm:99] ) . In recombination analyses such as those which we performed , the resulting recombinants would be virtually indistinguishable from other TYLCV-IL isolates and no recombination would therefore be inferred . The phylogenetic influences of such undetected cyclical recombination events – where parental viruses are recombinants and recombinants converge on parental viruses – are quite clearly depicted in the MCC tree of TYLCV CP sequences presented in Figure 2 . In this tree where the names of IL and Mld isolates are respectively coloured in red and blue , it is immediately obvious that , from the perspective of their CP sequences at least , isolates belonging to each of the strains are more closely related to isolates of the other strain than they are to some isolates of their own strain . This makes it very difficult to phylogenetically determine when recombination events such as those which generated TYLCV-IL from TYLCV-Mld occurred . Given that recombination is known to confound molecular clock analyses [48] , [49] we assembled a mostly recombination-free TYLCV coat protein gene dataset ( called the CP dataset ) . We analysed both this and the full genome ( FG ) TYLCV datasets with BEAST to determine the time and place where TYLCV originated . While the FG analysis indicated that the mean substitution rates during TYLCV evolution was 4 . 5×10−4 subs/site/year ( 95% HPD ranging from ( 2 . 4×10−4 to 6 . 8×10−4 ) , the CP analysis indicated a rate of 7 . 9×10−4 subs/site/year ( 95% HPD ranging from 4 . 9×10−4 to 1 . 1×10−3 ) . These substitution rate estimates are consistent with the previously published tomato infecting begomovirus full genome substitution rate estimate of 2 . 44×10−4 subs/site/year ( 95% HPD ranging from 1 . 3×10−6 to 6 . 1×10−4 [47] ) . Whereas the age of the most recent common TYLCV ancestor was estimated to be 293 years ( 95% HPD 138–515 ) using the FG dataset it was estimated to be only 56 years ( 95% HPD ranging between 35–80 ) using the CP dataset . These contradictory date estimates are almost certainly due to every one of the main TYLCV lineages in the FG dataset being different inter-species recombinants with highly divergent rep genes ( Figure 1 ) . It is expected that with the FG dataset , the much older dates of the last common ancestors of these highly divergent recombinationally acquired rep genes would have legitimately pushed the estimated of the most recent TYLCV common ancestor much deeper into the past [47] , [53] ( i . e . the estimated date is expected to be somewhere between the actual date of the Rep MRCA and the date of the MRCA of the rest of the genome ) . Despite the biasing influence of recombination in the FG dataset on the estimated timing of evolutionary events , both the FG and CP analyses clearly indicated that the most recent common ancestor of the TYLCVs probably resided in the Middle East – either somewhere near Iran ( posterior state probability , or PSP , = 0 . 53 for the FG dataset and 0 . 15 for the CP dataset , Figure S3 ) or somewhere in the Eastern Mediterranean ( PSP = 0 . 13 for the FG dataset and 0 . 48 for the CP dataset , Figure S3 ) . The PSP estimate of 0 . 53 for the FG dataset means that 53% of similarly plausible phylogenetic trees assessed during the analysis are consistent with this ancestral sequence being resident in Iran . Thus 68% of trees assessed during the FG analysis and 61% assessed during the CP analysis are consistent with the most recent common ancestor of the TYLCVs being resident in the Middle East ( i . e . Iran PSP + Eastern Mediterranean PSP ) . These percentages can be considered probability estimates which , although not higher than 95% , indicate that it is more than three times more probable that the most recent common ancestor of the TYLCVs was located near either Iran or in the Eastern Mediterranean than it is that the ancestor was located in the next most probable region ( the Western Mediterranean which has an associated PSP = 0 . 085 for the FG dataset and 0 . 19 for the CP dataset , Figure S3 ) . Importantly , this pattern was recapitulated even in sets of sub-sampled datasets designed to mitigate potential sampling biases in the complete CP and FG datasets . In all ten of the sub-sampled CP and FG datasets the most probable location of the TYLCV MRCA was either the region around Iran ( CP and FG datasets with respective mean PSPs = 0 . 26 and 0 . 25 ) or the Eastern Mediterranean ( CP and FG datasets with respective mean PSPs = 0 . 4 and 0 . 12; Figure S3 ) . Also , when we reran our analyses with the full datasets in such a way that the location state designations of all of the sequences were randomized throughout the MCMC procedure , the maximum PSP achieved at the root node for the most sampled location never exceeded 0 . 18 for the FG dataset and 0 . 22 for the CP dataset – both much lower than the maximum root node PSPs obtained without the location state randomisation setting ( which were 0 . 53 and 0 . 48 for the FG and CP datasets respectively ) . Together these tests indicated that sampling biases had not obviously influenced our identification of the Middle East as the region where TYLCV most probably originated . To pinpoint the source of the TYLCV variants that are spreading throughout the world , we retraced the movement patterns of TYLCVs over the past 50 years . Figure 2 is a phylogenetic depiction of TYLCV movement patterns ( based on the CP dataset MCC tree ) in which the tree branches have been coloured based on the most probable locations of their associated virus lineages such that a colour change between two connected nodes implies a probable migration event . In addition , a plausible spatio-temporal animation of TYLCV movements since the time of the most recent TYLCV common ancestor can be visualised by opening in GoogleEarth ( http://earth . google . com ) the Dataset S1 . kml ( FG dataset ) and Dataset S2 . kml ( CP dataset ) . Figure 3 summarises the results presented in these files . It is important to stress that in these analyses , we only considered the nine and seven discreet locations respectively studied in the CP and FG datasets . It must therefore be borne in mind that the locations indicated for ancestral viruses and the movement patterns inferred from these are simply the most plausible given the studied sampling locations – i . e . that actual locations of ancestral sequences and movement pathways may have included locations outside those that we have considered . Among the locations that we have considered , the FG and CP datasets respectively indicate that the global dispersal of TYLCV has involved at least 15 and 17 discrete migration events . As these viral movements ( or geographical location state transitions ) were inferred from node states of the FG and CP MCC trees , we only summarise the realisations of a potentially rich history of location state transitioning . The reason for this is that the geographical location states mapped to the various tree nodes reflect the starting and ending points of various movements – they do not recapture the potentially long and winding routes taken during these journeys . Both the FG and CP datasets indicate that TYLCVs have moved at least twice from the Eastern Mediterranean to Asia ( Bayes factors , or BF , = 11 . 5 and 3 . 6 where a BF >100 represents decisive support , a BF >10 . 0 represents strong support , a BF >3 . 2 represents substantial support and a BF <3 . 2 is not well supported [44] , three times to the Mediterranean ( BF = 15 . 7 and 1209 ) and once to North America ( BF = 2 . 36 and 11 . 8 ) . The FG analysis also indicated that two independent TYLCV movements have occurred from the Western Mediterranean to Asia ( BF = 3 . 6 ) . Consistent with previous reports [54] , both the FG and CP datasets also indicate two migration events from the Western Mediterranean to the southern Indian Ocean island of Reunion ( BF = 17 . 6 , 729 ) . It is also noteworthy that with the FG dataset two migration events are inferred from the Western Mediterranean to the Eastern Mediterranean ( BF = 15 . 7; although no corresponding migrations were inferred with the CP dataset ) , indicating that TYLCV movements between these regions may be bidirectional . Although the FG analysis indicated that TYLCV probably originated near Iran , this analysis indicated only weak support for three early virus movements out of Iran to the Eastern Mediterranean ( BF = 2 . 64 ) , the Horn of Africa ( BF = 2 . 01 ) and the Western Mediterranean ( BF = 0 . 64 ) . In the CP analysis where the Eastern Mediterranean rather than the Iranian region was identified as the probable origin of TYLCV , three independent , decisively supported ( BF = 265 ) , migration events from the Eastern Mediterranean to Iran were inferred , possibly explaining the broad degree of TYLCV diversity found in the latter region . Finally , our analysis supports the hypotheses that TYLCV-IL has been independently introduced to the New World , once from the region around the Eastern Mediterranean ( BF = 2 . 36 and 11 . 8 for the FG and CP datasets respectively ) and once from Asia ( BF = 13 . 6 and 45 . 7 for the FG and CP datasets respectively; [16] ) . Collectively these data indicate that although the region around Iran is a center of TYLCV diversity and is possibly also the region where the species originated , it has not been the direct source of the TYLCV variants that are currently spreading worldwide . This means that novel pathogenic TYLCV variants that arise in this region will probably be less of a threat to global agriculture than those arising closer to more internationally connected regions such as the Mediterranean basin . Our preliminary recombination analysis indicated that all of the detectable recombinants that discernibly contained TYLCV-like sequences had been sampled in the Mediterranean basin and the Middle East . We suspected that within this region there might be geographical recombination hotspots . By mapping the 18 detected TYLCV recombination events to the FG and CP MCC trees determined during our phylogeography analysis , it was possible for us to approximate the locations where and the times when the recombination events most likely occurred . For each recombination event in each tree this involved identification of the nodes representing ( 1 ) the last common ancestor of the recombinants ( referred to as RecAnc in Table 1 ) and ( 2 ) the last TYLCV ancestor not sharing evidence of the same recombination event ( referred to as Non-RecAnc in Table 1 ) . The dates and locations of the sequences at these two nodes in the MCC trees were assumed to bound the date when , and the location where , the recombination event occurred . Whereas it was possible to use this approach to infer dates and locations for all 18 of the recombination events with the FG dataset , groups of recombinant TYLCV-IL and –Mld sequences sharing evidence of events 7 , 9 , 10 , 11 and 14 were not monophyletic in the CP tree ( probably for reasons explained above in the recombination analysis section; Figure 2 ) , meaning that locations and dates could not be properly inferred for these recombination events using the CP dataset . Despite this , the CP dataset yielded much tighter estimates of recombination dates than the FG dataset , possibly due to its being free of the confounding effects of the inter-species recombination events found in the latter . The FG and CP datasets nevertheless indicated locations where recombination events had occurred that were generally in good agreement with one another ( compare orange and blue bars in Figure S4 ) and recombination date estimates that had broadly overlapping 95% HPDs ( compare orange and blue bars in Figure S5 ) . The exceptions were the five “problematic” events ( 7 , 9 , 10 , 11 and 14 ) mentioned previously . For these the FG and CP datasets yielded support for recombination events having occurred in different locations . For example with events 9 , 10 and 11 the FG dataset indicated that the RecAnc and Non-RecAnc sequences most probably resided near Iran , the CP dataset indicated that the Non-RecAnc sequence most probably resided near the Eastern Mediterranean ( with the location of the RecAnc sequence remaining undetermined for the CP dataset ) . Nevertheless , the clear pattern emerging from these analyses was that all 18 of the detected TYLCV recombination events occurred either in the Western Mediterranean , the Eastern Mediterranean or near Iran . Collectively these geographical locations ( representing 58% of the sequence ) accounted for more than 80% of the posterior probability distribution for every ancestral sequence used to infer the locations of every recombination event . Based on dates inferred from the CP MCC tree , these recombination events were also mostly all quite recent with the oldest ( events 7 and 14 ) having most probably occurred some time after 1964 ( Table 1 and Figure S5 ) . If one discounts the “problematic” recombination events 7 , 9 , 10 , 11 and 14 , the remaining thirteen events have all most probably occurred since 1985 . Nine of these thirteen events most probably occurred near Iran or Israel with both the FG and CP analyses indicating that Iran was the most probable site of eight of them ( supported for all events other than events 1 and 16 by the location state probabilities of all the relevant RecAnc and Non-RecAnc sequences in both the FG and CP datasets ) . Besides being the most probable origin of TYLCV and the center of TYLCV diversity , the Middle East in general , and Iran in particular , is therefore also apparently the region where most of this virus' evolutionary change through recombination has occurred . In this regard it is interesting that recombination events 2 , 4 , 5 and 6 , the only events that almost certainly occurred outside the Middle East , are also the only four involving TYLCV sequences as donors ( i . e . such that a minority of the recombinant's genome consists of TYLCV-like sequences ) . Although this difference between the character of TYLCV recombination events occurring inside and outside the Middle East may be coincidental , it could also be indicative of an important evolutionary trend associated with the migration of viruses into environments different from those in which they evolved . The observed pattern is in fact what one might expect to occur with recombining invasive virus species . For example , it is expected that viruses residing in the locations where they evolved would be well adapted to seasonal changes in the mix of host and vector genotypes that typify their home environments . One might expect both that these adaptations would provide them with a “home environment advantage” over invasive newcomers and that the genetic underpinnings of these adaptations would be distributed throughout their genomes . The invasive newcomers , however , would not be invasive unless they had some specific , especially adaptive genetic trait that provided them with their invasive phenotype . When such indigenous and invasive viruses recombine , the fittest of their offspring would probably be those that incorporate the invader's highly adaptive traits within an indigenous genetic background . Unless TYLCV and TYLCSV only replicate within genetically homogeneous cultivated tomato species and are epidemiologically unaffected by local variations in host species distributions across the Mediterranean and Middle East , it is conceivable that both have an advantage in their respective home environments . The net result may be that in the Middle East when TYLCVs recombine with viruses originating in India or Africa the TYLCVs are the better acceptors whereas in the Western Mediterranean they are better as donors to indigenous viruses like TYLCSV . To retrace the global movements of TYLCV we considered phylogeographic inferences made with both the FG and CP datasets . However , given that the estimated calendar dates of movement events differed between the FG and CP analyses and the probable impact that inter-species recombination has had on evolution rates estimated with the FG dataset , we used results obtained with the mostly recombination-free CP dataset to estimate the timing of key events during the evolution and dissemination of TYLCV ( summarised in Figure 3 ) . It is important to reiterate here that both the age , location and migration route estimates that follow are associated with degrees of uncertainty and that the descriptive history we provide is simply the most plausible given an admittedly sparse TYLCV sequence dataset and imperfect analytical tools . Nevertheless , although the Bayesian analyses underlying the description do not account for important factors such as natural selection , they do provide us with 95% HPD intervals that are an honest expression of the uncertainty surrounding the various date , location and migration route estimates that we infer . At some point between 1937 and 1952 ( 95% HPD 1905–1972 ) , a virus arose somewhere within the Middle East which was the first recognisable TYLCV . By ∼1952 this “first” TYLCV lineage had evolved into the most recent common ancestor of all known contemporary TYLCVs . It is possible , although in no way certain , that this virus was a recombinant that had inherited the majority of its genome from an earlier TYLCV prototype but a large portion of its rep gene and its origin of virion strand replication from some other unknown ( but possibly Asian ) begomovirus species ( see event 7 in Figure 1 , Table 1 and Figures S4 and S5 ) . It is also plausible that the immediate descendants of this virus were responsible for the Middle Eastern TYLCD epidemics of the early 1960s [4] , [5] . There is good evidence from our analysis that during the 1960s these viruses evolved within the Middle East to yield prototypical versions of the TYLCV-Gez strain in the Eastern Mediterranean ( PSP = 0 . 65 ) , by ∼1964 ( 95% HPDs 1948–1978 ) , the TYLCV-Mld strain also in the Eastern Mediterranean ( PSP = 0 . 90 ) by ∼1973 ( 95% HPDs 1963–1982 ) and the TYLCV-Ker strain in the region of Iran ( PSP = 0 . 97 ) by ∼1979 ( 95% HPDs 1964–1992 ) . Later , between 1993 and 2006 ( 95% HPD 1986–2006 ) and also probably in Iran ( PSP = 0 . 98–1 . 00 ) , a recombination event between a TYLCV-Ker variant and CLCuGV ( event 8 in Table 1 and Figure 1 ) created the first member of TYLCV-Bou , the most recently evolved of the seven currently described TYLCV strains . Although both inter- and intra-species recombination events involving early TYLCV variants probably persistently occurred within the broader Middle East during these years , the first of these that would come to largely differentiate the seven current TYLCV strains probably occurred somewhere in this region ( PSP = 0 . 80–0 . 76 ) between 1964 and the mid 1970s ( 95% HPD 1948–1999 ) . This event ( or possibly a series of events ) , traces of which are possibly evident in events 9 , 10 and 11 in Table 1 and Figure 1 , yielded the founder of the IL strain . Similar recombination events between either TYLCV-Mld or -IL ( it is unclear which ) and TolCRV somewhere in the Middle east ( PSP = 0 . 97–1 . 0 ) between 1985 and 1996 ( 95% HPDs 1978– 1996; event 1 in Table 1 and Figure 1 ) and between TYLCV-Mld or -IL and ToLCKV near Iran ( PSP = 0 . 99–1 . 0 ) between 1996 and 2000 ( 95% HPDs 1991–2003; event 3 in Table 1 and Figure 1 ) would respectively yield the first members of what are currently known as the TYLCV-IR and -OM strains . At some point between 1981 and 1989 ( 95% HPDs 1971–1993 ) the world-wide dissemination of TYLCV began when a TYLCV-IL virus ( most likely from the Eastern Mediterranean ) , moved to , and became established within , the Western Mediterranean . This trip was later repeated at least once by a TYLCV-Mld virus between 1990 and 2001 ( 95% HPDs 1982–2003 ) . Although the polarity of the movement is uncertain ( the FG and CP datasets conflict on this point ) , additional movements of IL viruses between the Middle East and the Western Mediterranean also occurred during this period . Viruses within the newly established Western Mediterranean TYLCV-Mld and -IL populations were then transported to Asia between 1989 and 1996 ( 95% HPDs 1983–1996 ) and the Indian Ocean island of Reunion between 1991 and 2002 ( 95% HPDs 1987–2003 ) . At least two other long distance movements of IL viruses to Asia also occurred from the Middle East between 1981 and 1999 ( 95% HPD 1970–1999 ) . Whereas the trans-Atlantic movement of a Middle Eastern TYLCV-IL virus to the New World probably happened between 1992 and 1994 ( 95% HPD 1988–1994 ) – within two years of the first TYLCVs being sampled there [15] – the trans-Pacific transport of an Asian TYLCV-IL virus ( a descendant of one of the lineages introduced from the Middle East ) to North America probably only occurred between 1999 and 2003 ( 95% HPD 1996–2004 ) . We have described how within thirty years of their Middle Eastern origin , both TYLCV-Mld and the TYLCV-IL lineage have ascended to the point where they are today ranked among the greatest biotic threats to tomato production world-wide [55] . This emergence has been so swift that no precise estimates of either their current or projected future economic impacts exist . The epidemiological , evolutionary and ecological impacts of their movements are probably even harder to predict although in this regard patterns seen in the Western Mediterranean where they have spent their greatest time outside the Middle East will possibly prove informative [56]–[58] . For example , given the high frequencies of inter-species TYLCV recombination events that we have mapped to the Middle East , it is perhaps reasonable to expect that , as has happened in the Western Mediterranean [7] , [22] , [51] , TYLCVs introduced to the Americas , the southern Indian Ocean , and Asia will recombine with viruses indigenous to these regions . While it is impossible to predict how evolutionarily productive any such recombination events will be , the possibility remains that TYLCV genetic material within the context of mostly indigenous recombinant begomovirus genomes could shortly begin showing up in Asia , the Indian Ocean islands and the Americas . We envision that tracking the movements of the various TYLCV invasion fronts and monitoring virus sequence data before the fronts hit and in the years thereafter could prove very fruitful in our endeavours to answer some key questions relating to the economic , epidemiological , ecological and evolutionary impacts of such plant virus invasions . | Tomato yellow leaf curl virus ( TYLCV ) poses a serious threat to tomato production throughout the temperate regions of the world . Our analysis , using a suite of bioinformatic tools applied to all publically available TYLCV genome sequences , suggests that the virus probably arose somewhere in the Middle East between the 1930s and 1950s and that its global spread only began in the 1980s after the emergence of two strains – TYLCV-Mld and -IL . In agreement with others , we also find that the highly invasive TYLCV-IL strain has jumped at least twice to the Americas – once from the Mediterranean basin in the early 1990s and once from Asia in the early 2000s . Although our results corroborate historical accounts of TYLCV-like symptoms in tomato crops in the Jordan Valley in the late 1920s , they indicate that the region around Iran is both the current center of TYLCV diversity and is the site where the most intensive ongoing TYLCV evolution is taking place . However , our analysis indicates that this region is epidemiologically isolated suggesting that novel TYLCV variants found there are probably not direct global threats . Moreover , we identify the Mediterranean basin as the main launch-pad of global TYLCV movements . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results/Discussion"
] | [
"computational",
"biology/evolutionary",
"modeling",
"molecular",
"biology/molecular",
"evolution",
"virology/emerging",
"viral",
"diseases",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics"
] | 2010 | The Spread of Tomato Yellow Leaf Curl Virus from the Middle East to the World |
As part of a broader collaborative network of exome sequencing studies , we developed a jointly called data set of 5 , 685 Ashkenazi Jewish exomes . We make publicly available a resource of site and allele frequencies , which should serve as a reference for medical genetics in the Ashkenazim ( hosted in part at https://ibd . broadinstitute . org , also available in gnomAD at http://gnomad . broadinstitute . org ) . We estimate that 34% of protein-coding alleles present in the Ashkenazi Jewish population at frequencies greater than 0 . 2% are significantly more frequent ( mean 15-fold ) than their maximum frequency observed in other reference populations . Arising via a well-described founder effect approximately 30 generations ago , this catalog of enriched alleles can contribute to differences in genetic risk and overall prevalence of diseases between populations . As validation we document 148 AJ enriched protein-altering alleles that overlap with "pathogenic" ClinVar alleles ( table available at https://github . com/macarthur-lab/clinvar/blob/master/output/clinvar . tsv ) , including those that account for 10–100 fold differences in prevalence between AJ and non-AJ populations of some rare diseases , especially recessive conditions , including Gaucher disease ( GBA , p . Asn409Ser , 8-fold enrichment ) ; Canavan disease ( ASPA , p . Glu285Ala , 12-fold enrichment ) ; and Tay-Sachs disease ( HEXA , c . 1421+1G>C , 27-fold enrichment; p . Tyr427IlefsTer5 , 12-fold enrichment ) . We next sought to use this catalog , of well-established relevance to Mendelian disease , to explore Crohn's disease , a common disease with an estimated two to four-fold excess prevalence in AJ . We specifically attempt to evaluate whether strong acting rare alleles , particularly protein-truncating or otherwise large effect-size alleles , enriched by the same founder-effect , contribute excess genetic risk to Crohn's disease in AJ , and find that ten rare genetic risk factors in NOD2 and LRRK2 are enriched in AJ ( p < 0 . 005 ) , including several novel contributing alleles , show evidence of association to CD . Independently , we find that genomewide common variant risk defined by GWAS shows a strong difference between AJ and non-AJ European control population samples ( 0 . 97 s . d . higher , p<10−16 ) . Taken together , the results suggest coordinated selection in AJ population for higher CD risk alleles in general . The results and approach illustrate the value of exome sequencing data in case-control studies along with reference data sets like ExAC ( sites VCF available via FTP at ftp . broadinstitute . org/pub/ExAC_release/release0 . 3/ ) to pinpoint genetic variation that contributes to variable disease predisposition across populations .
Genetic population isolates like the Ashkenazim , Jews who trace their ancestry to eleventh century central European Jewish groups[1] , have previously facilitated the mapping of alleles contributing to human disease predisposition[2–5] . The documented 2–4 fold enrichment of Crohn’s Disease ( CD ) prevalence in the Ashkenazi Jewish ( AJ ) population[6 , 7] motivated the use of exome sequencing and genome-wide array data to evaluate the degree to which bottleneck-enriched protein-altering alleles and unequivocally implicated common variants contribute an excess CD genetic risk to AJ[6] . Despite the progress in mapping genes and alleles for rare diseases with increased prevalence in the AJ population , precise estimates of the risk-allele frequency and the carrier rate in the AJ population have not yet been determined[8] . Through this study , we provide a frequency resource of protein-coding alleles from over 2 , 000 non-CD AJ individuals with low admixture that will serve to improve interpretation of rare disease risk alleles in the AJ population and which we employ to discover new Crohn’s risk alleles by comparison to 1855 AJ Crohn’s cases .
We generated a jointly called whole-exome sequence dataset consisting of 18 , 745 individuals from international Inflammatory Bowel Disease ( IBD ) and non-IBD cohorts[9 , 10] ( S1 Fig ) . Given the increased prevalence of Crohn’s disease in the AJ population , our global sequencing efforts had specifically included 5 , 652 individuals self-reporting as Jewish and , as we aimed to focus on variation observed in the AJ population in comparison to reference populations in ExAC[9 , 11] ( including non-Finnish Europeans ( NFE ) , Latino ( AMR ) , and African/African-American ( AFR ) ) populations , we chose a model-based approach to estimate the ancestry of the study population using ADMIXTURE[12] . To identify AJ individuals and estimate admixture fractions we used a set ( n = 21 , 066 ) of LD-pruned common variants ( MAF>1% , see Supplementary Note for additional details ) filtered for genotype quality ( GQ>20 ) . The 18 , 745 individuals were assigned to four groups ( K = 4 ) using ADMIXTURE ( further described in Supplementary Note , also see S3 Fig ) . One group of 5 , 685 individuals was found consisting mostly ( 84% ) of self-reported AJ individuals , while 3 , 522 of these individuals were further found with high ancestry fraction ( > 0 . 9 ) mapping to this group ( S2 Fig , S1 Table ) . Thus , many self-reported AJ individuals were not included , as they did not have high enough ascertained AJ ancestry fraction . As we were interested in computing an enrichment statistic that would not be affected by possible admixture , we obtained alternate ( non-reference ) allele frequency estimates by restricting the enrichment analysis to the 2 , 178 non-IBD Ashkenazi Jewish individuals that passed QC and relatedness filtering and had AJ ancestry fraction ( genotype ancestry grouping closely with other AJ individuals ) of > 0 . 9 . Our study includes exomes throughout Europe and Israel but the vast majority ( 86% ) of these high ancestry fraction AJ individuals were collected in major US cities including Los Angeles , Boston , Baltimore , and New York ( S2 Table ) . To explore AJ exome population genetics , including proportion of enriched alleles and degree of enrichment , we used the observed alternate allele counts and total number of alleles available from ExAC release 0 . 3 dataset [ntotal = 60 , 706; NFE ( n = 31 , 902; after excluding AJ individuals from ExAC ) , AFR ( n = 5 , 203 ) , and AMR ( n = 5 , 789 ) ] . We focused on protein-coding alleles with estimated allele frequency of at least 0 . 002 and less than . 1 in AJ ( nalleles = 73 , 228; practical cutoff of what could be statistically defined as convincing enrichment , see S4 Fig ) , and applied a one-sided Fisher’s exact test on allele counts ( see Supplementary Note ) , to classify the observed alleles into two groups: “enriched” or “not enriched” . This analysis identified 34% of protein-coding alleles as significantly enriched , with mean 15-fold increased odds of the alternate allele compared to other populations . Different proportions of alleles belong to the enriched group depending on variant annotation: 36% for predicted protein-truncating variants ( PTV ) ; 38% for predicted protein-altering variants ( PRA ) ; and 31% for synonymous variants . The substantially higher PTV+PRA:synonymous ratio observed in the enriched category is consistent with those alleles being drawn randomly from a large pool of much rarer alleles ( where the functional:synonymous ratio is higher[3] ) and abruptly boosted in frequency ( Fig 1 , p < 10−16 across comparisons of PTV and PRA to synonymous variants , two-proportion test , Supplementary Note ) . Since much rarer alleles have a higher probability of being damaging ( e . g . , they have a higher missense/synonymous ratio ) , the advantage to gene mapping arises from the fact that enriched alleles of a certain frequency are more damaging/deleterious on average than non-enriched alleles of the same frequency . Additionally , we may expect that a “depleted” set of alleles arises from the founder effect , but in reality , many of these already rare variants are simply eliminated during the bottleneck . Of course , it is more difficult and less interesting to search for depleted alleles , as their absence provides no opportunity to obtain significant statistics on population enrichment or disease association . We intersected the list of protein-coding alleles identified in the AJ exome sequencing study with alleles reported to be pathogenic with no conflicting evidence ( n = 42 , 226 ) in ClinVar[14] resulting in 148 alleles found both in ClinVar and with p-value less than . 005 of belonging to the AJ enriched set ( S1 Data File ) . In OMIM , 48 of the 148 alleles included documentation of a disease subject with AJ ancestry ( Table 1 ) . This set of enriched alleles includes all of the major AJ mutations for 8 diseases described in the American College of Medical Genetics and Genomics 2008 screening guideline study[15] . In the setting of autosomal recessive disorders these differences in population allele frequencies may contribute a factor proportional to the squared enrichment difference to genetic risk and prevalence between populations ( see Supplementary Note ) . For instance , a 19-fold enriched frameshift indel , p . Tyr427IlefsTer5 , in HEXA , contributes a 361-fold enrichment in genetic risk in AJ to non-AJ population to Tay-Sachs disease . Enrichment in this large adult Ashkenazi exome database reinforces recent publications of founder mutations for rare pediatric disorders including FKTN ( Walker Warburg syndrome ) [16] , CCDC65 ( Primary ciliary dyskinesia ) [17] , TMEM216 ( Joubert syndrome ) [18] , C11orf73 ( Leukoencephalopathy ) [19]; PEX2 ( Zellweger syndrome ) [20] , VPS11 ( Hypomyelination and developmental delay ) [21] and BBS2 ( Bardet-Biedl syndrome ) [22] . While many alleles on this pathogenic list may demonstrate incomplete penetrance ( as in the case of p . V726A in MEFV[23] for Familial Mediterranean fever ) and some may not show recessive inheritance , this resource should provide considerable assistance in gene discovery and clinical genetic screening in AJ ( S2 Data File ) . To assess whether AJ-enriched protein-coding alleles also contribute to the established difference in CD genetic risk we performed case-control association analyses . Since individuals with only partial AJ ancestry will still carry bottleneck-enriched alleles , here we included samples with estimated AJ ancestry fraction > 0 . 4 ( Supplementary Note , S2 Fig ) , resulting in a dataset of 4 , 899 AJ samples ( 1 , 855 Crohn’s disease and 3 , 044 non-IBD ) . To improve ability to detect a true association , we performed a meta-analysis with CD and non-IBD case-control exome sequencing data from two additional ancestry groups: 1 ) non-Finnish European ( NFE ) ( 2 , 296 CD and 2 , 770 non-IBD ) ; and 2 ) Finnish ( FIN ) ( 210 CD and 9 , 930 non-IBD samples ) from a separate callset described in a previous publication[24] for a total of 4 , 361 CD samples and 15 , 744 non-IBD samples . By calling additional non-AJ samples , we hoped to discern which of the AJ-enriched alleles contributed a significant risk factor across all populations . The meta-analysis performed across several populations described should mitigate biases by confirming consistency in effect size across these population groups . Study-specific association analysis was performed with Firth bias-corrected logistic regression[25 , 26] and four principal components as covariates using the software package EPACTS[27] ( S5 Fig ) . We combined association statistics in a meta-analysis framework using the Bayesian models in Band et al . [28] . We used the correlated effects model , obtained a Bayes factor ( BF ) by comparing it with the null model where all the prior weight is on an effect size of zero , reported p-value approximation using the BF as a test statistic , and assessed whether heterogeneity of effects exist across studies for downstream QC ( see Supplementary Note ) . We separately assessed CD associations of enriched protein-altering ( PRA ) and synonymous ( SYN ) alleles in protein-coding genes in CD implicated GWAS loci ( ngwas , pra = 351; ngwas , syn = 167 ) , and outside implicated GWAS loci ( nnon-gwas , pra = 12 , 529; nnon-gwas , syn = 6 , 202 , Fig 2 ) . See Methods and Materials for a description of these loci . We identified ten AJ enriched CD risk alleles ( p<0 . 005 ) : the previously published risk haplotypes in LRRK2 and NOD2 ( LRRK2: p . N2081D; NOD2: p . N852S , p . G908R , p . M863V+p . fs1007insC ) [29 , 30] , in addition to newly implicated alleles ( NOD2: p . A612T , p = 2 . 8x10-9; c . 74-7T>A , p = 1 . 4x10-4; p . L248R , p = 6 . 4x10-4; p . D357A , p = 0 . 0011; LRRK2: p . G2019S , p = 0 . 0014 , a Parkinson’s disease risk allele[31] ) . To assess whether the new NOD2 enriched alleles are conditionally independent of the previously established associated NOD2 alleles we performed conditional haplotype association analysis in PLINK and Bayesian model averaging[32] for variable selection , both of which suggested independent effects for all alleles ( S6 Fig , S3 Table ) . Deviation from additivity can contribute additionally to individual risk but has been difficult to document in complex disease associations with modest ORs . Despite the functional relationship between LRRK2 and NOD2[33] , we do not observe deviation from additivity between LRRK2 and NOD2 ( p = 0 . 273 ) ; that is , the effect of mutations in both LRRK2 and NOD2 is no greater than the sum of their individual effects . We assessed whether composite risk carriers ( carrier of more than one variant allele ) had evidence of deviation from additivity . Deviation from additivity has been reported for p . fs1007insC , p . G908R , and p . R702W in NOD2[34 , 35] . In our AJ exome sequencing data set we estimate a 1-hit effect equal to 1 . 82 ( 95% confidence interval [1 . 59 , 2 . 07] ) and a 2-hit effect equal to 8 . 24 ( 95% confidence interval [6 . 06 , 11 . 21]; we found similar evidence for departure from additivity when restricting the analysis to the newly reported alleles only: p = 0 . 00357 , odds ratio = 7 . 53 ) . We confirmed this finding using the larger non-AJ Crohn’s disease ImmunoChip dataset to provide a more precise estimate of the 1-hit effect ( OR = 2 . 17; 95% confidence interval [2 . 07 , 2 . 27] , S4 Table ) and the non-additive 2-hit effects in NOD2 ( OR = 9 . 93; 95% confidence interval [8 . 88 , 11 . 13] , S5 Table ) . We found no evidence of deviation from additivity for the associated protein-altering alleles in LRRK2 ( p = 0 . 418 ) . Given that enriched genetic variants in NOD2 and LRRK2 contribute to differences in CD risk in AJ population , we next asked whether unequivocally established common variant associations contribute to differences in CD genetic risk . We performed polygenic risk score ( PRS ) analysis using reported effect size estimates from 124 CD alleles including those reported in a previously published study[36] and four variants in IL23R from a recent fine-mapping study[37] , and excluding variants in NOD2 and LRRK2 . We observed an elevated PRS for AJ compared to non-Jewish controls ( 0 . 97 s . d . higher , p<10−16; Fig 3A; number of non-AJ controls = 35 , 007; number of AJ controls = 454 ) , and as expected when performing the PRS analysis using OR calculated from non-Jewish subset of iCHIP data the signal still remains ( p<10−16 , S7 Fig ) . We observed a similar trend for the CD samples ( 0 . 54 s . d . higher; p<10−16; Fig 3B; number of non-AJ CD cases = 20 , 652; number of AJ CD cases = 1 , 938 ) . We demonstrate this is not a systematic property of common risk alleles in AJ by running the same comparison using instead the comparable set of established schizophrenia associated alleles from the Psychiatric Genomics Consortium[38] . To quantify the relative contribution of CD-implicated alleles to the difference in genetic risk between AJ and non-AJ populations we estimated the expected PRS value of an individual and expected difference in PRS between two populations by simply using summary statistics including the frequency of the minor allele in the two populations and the corresponding odds ratio ( Supplementary note , S6 Fig ) . We applied the approach to all CD implicated alleles and observed that variants in GWAS loci annotated as IRGM , LACC1 , NOD2 , MST1 , ATG16L1 , GCKR , NKX2-3 , and LRRK2[36] contribute substantially ( >0 . 01 ) to the increased genetic risk observed in AJ . It is possibly relevant that variants contributing to increased risk in AJ include many autophagy/intracellular defense genes ( IRGM , ATG16L1 , LRRK2 ) , while those contributing to increased risk in non-AJ include many anti-fungal/Th17/ILC3 genes[39] ( IL23R , IL12B , CARD9 , TRAF3IP2 , IL6ST , CEBPB; Fig 3C ) . Both documented variability in the occurrence of CD over time[40 , 41] and substantial uncertainty in reported CD prevalence estimates[42 , 43] impact our ability to precisely estimate the overall contribution of genetics to the established difference in prevalence between populations . To interpret the impact of shifts in genetic risk score on differences in prevalence , we used the logit risk model[35] and evaluated a new estimate of disease probability , pnew , assuming an initial disease probability , p0 , and multiple values for the differences in genetic risk . Assuming log-additive effects , and a logit-risk model , we estimate that the observed differences in genetic risk between the AJ and non-AJ populations contribute an expected 1 . 5-fold increase in disease prevalence in a population with environmental risk factors corresponding to AJ and baseline genetic risk corresponding to non-AJ populations ( S7–S9 Figs ) . To address the extent to which non-additive effects in NOD2 may impact the observed prevalence we assumed 1-hit and 2-hit odds ratios of 2 . 17 and 9 . 93 , respectively . We attribute a 6 . 8% difference in the ratio of estimated disease prevalence in the AJ population to the deviation from additivity , suggesting a small effect on differences in population prevalence ( Supplementary Note ) .
Analyzing data from 5 , 685 Ashkenazi Jewish exomes , we provide a systematic analysis of AJ enriched protein-coding alleles , which may contribute to differences in genetic risk to CD as well as numerous rare diseases , many of which are transmitted via autosomal recessive inheritance . We identified protein-altering alleles in NOD2 and LRRK2 that are conditionally independent and contribute to the excess burden of CD in AJ . We found evidence that common variant risk defined by GWAS shows a strong elevated difference between AJ and non-AJ European population samples ( 0 . 97 s . d . higher in controls , 0 . 54 s . d . higher in cases , p<10−16 in both ) , independent of NOD2 and LRRK2[44] . Highly polygenic diseases are unlikely to have substantially altered incidence as a result of a bottleneck alone—for every enriched variant there are those depleted or lost entirely and population genetics simulations[45] suggest no systematic alteration of overall genetic burden as a function of a bottleneck . Thus , the strong ( approximately 1 . 5-fold , see supplementary note ) difference in Crohn’s incidence in concert with a systematic enrichment of risk-increasing alleles , unlikely to have arisen by chance , suggests non-random selection in the AJ population for higher CD risk alleles . It seems plausible that , rather than ‘selection for Crohn's' per se , this likely suggests a subset of Crohn's risk alleles may contribute to a common biological process ( e . g . , a specific immune response ) or phenotype that was positively selected for in AJ[46–48] . Such weak , widespread ‘polygenic selection’ has previously been observed with respect to height-associated SNPs in Europe[49] , where drift alone could not explain the systematic enrichment of height-increasing alleles in populations of Northern Europe vs . Southern Europe . We found that CD risk alleles that are systematically elevated in AJ are not unusually elevated in another well-established founder population for which we have extensive genotype data ( Finland ) . In Finns , Crohn’s risk alleles were not systematically enriched—they were if anything slightly depleted with 69 risk alleles at higher frequency in Finns than NFE and 79 risk alleles at lower frequency in Finns than NFE . We also demonstrate this is not a systematic property of common risk alleles in AJ by running the same comparison using instead the comparable set of established schizophrenia associated alleles from the Psychiatric Genetics Consortium[50] . We mapped 102 schizophrenia-associated index SNPs to AJ frequency data and again observed no uneven distribution where risk alleles are systematically more or less common . In total , 52 risk alleles were at higher frequency in AJ than NFE and 50 risk alleles were higher frequency in NFE than AJ . This study of CD in the AJ population confirms population-genetic expectations . First , recently bottlenecked populations are uniquely powered to discover alleles with markedly increases in frequency , and , as a consequence , contributors to differences in genetic risk across population groups . Second , while NOD2 and published common variant associations contribute substantially to the genetic risk of CD , other genes with causal alleles that failed to pass through the bottleneck are missed , consistent with predictions from Zuk et al[4] . We provide an exome frequency resource of protein-coding alleles in AJ along with estimates of population-specific enrichment . The sets of enriched alleles should be carefully considered when performing case-control analysis . Population structure can easily lead to false positive associations , especially for low frequency and rare variants , if the AJ:nonAJ ratio is slightly different in cases and controls . Our approach and this resource will likely catalyze our understanding of the medical relevance of enriched alleles in population isolates . Most importantly , the frequency reference provides critical guidance in pinpointing or excluding specific risk factors in individuals in clinical and research settings .
We generated a jointly called dataset consisting of 18 , 745 individuals from international IBD and non-IBD cohorts . Sequencing of these samples was done at Broad Institute . All patients and control subjects provided informed consent . Recruitment protocols and consent forms were approved by Institutional Review Boards at each participating institutions ( Protocol Title: The Broad Institute Study of Inflammatory Bowel Disease Genetics; Protocol Number: 2013P002634 ) . All DNA samples and data in this study were denominalized . For all cohorts , CD was diagnosed according to accepted clinical , endoscopic , radiological and histological findings . G4L WES is a project specific product . It combines the Human WES ( Standard Coverage ) product with an Infinium Genome-Wide Association Study ( GWAS ) array . In addition to the array adding to the genomics data , it also acts as a concordance QC , linking 14 SNPs to the exome data . The processing of the exome includes Sample prep ( Illumina Nextera ) , hybrid capture ( Illumina Rapid Capture Enrichment - 37Mb target ) , sequencing ( Illumina , HiSeq machines , 150bp paired reads ) , Identification QC check , and data storage ( 5 years ) . Our hybrid selection libraries typically meet or exceed 85% of targets at 20x , comparable to ~60x mean coverage . The array consists of a 24-sample Infinium array with ~245 , 000 fixed genome-wide markers , designed by the Broad . On average our genotyping call rates typically exceed 98% . The sequence reads are first mapped using BWA MEM[51] to the GRCh37 reference to produce a file in SAM/BAM format sorted by coordinate . Duplicate reads are marked–these reads are not informative and are not used as additional evidence for or against a putative variant . Next , local realignment is performed around indels . This identifies the most consistent placement of the reads relative to potential indels in order to clean up artifacts introduced in the original mapping step . Finally , base quality scores are recalibrated in order to produce more accurate per-base estimates of error emitted by the sequencing machines . Once the data has been pre-processed as described above , it is put through the variant discovery process , i . e . the identification of sites where the data displays variation relative to the reference genome , and calculation of genotypes for each sample at that site . The variant discovery process is decomposed into separate steps: variant calling ( performed per-sample ) , joint genotyping ( performed per-cohort ) and variant filtering ( also performed per-cohort ) . The first two steps are designed to maximize sensitivity , while the filtering step aims to deliver a level of specificity that can be customized for each project . Variant calling is done by running Genome Analysis Toolkit’s ( GATK ) HaplotypeCaller in gVCF mode on each sample's BAM file ( s ) to create single-sample gVCFs . If there are more than a few hundred samples , batches of ~200 gVCFs are merged hierarchically into a single gVCF to make the next step more tractable . Joint genotyping is then performed on the gVCFs of all available samples together in order to create a set of raw SNP and indel calls . Finally , variant recalibration is performed in order to assign a well-calibrated probability to each variant call in a raw call set , and to apply filters that produce a subset of calls with the desired balance of specificity and sensitivity as described in Rivas et al . ( 2016 ) [24] . Samples with > = 10% contamination are excluded from call sets . Exome samples with less than 40% of targets at 20X coverage are excluded . Variant annotation was performed using the Variant Effect Predictor ( VEP ) [cite PMID: 20562413] version 83 with Gencode v19 on GRCh37 . Loss-of-function ( LoF ) variants were annotated using LOFTEE ( Loss-Of-Function Transcript Effect Estimator , available at https://github . com/konradjk/loftee ) , a plugin to VEP . LOFTEE considers all stop-gained , splice-disrupting , and frameshift variants , and filters out many known false-positive modes , such as variants near the end of transcripts and in non-canonical splice sites , as described in the code documentation . Finnish CD patients were recruited from Helsinki University Hospital and described in more detail previously[52 , 53] . We used the same exome sequencing dataset described in Rivas et al . [24] . We applied additional PC correction in the Finnish identified individuals to remove individuals with membership of Finnish sub-isolate ( Northern Finland ) and excluded based on PC2 0 . 015 ( 853 excluded , 826 controls , 27 IBD ) . We recalculated PCs and included the first four PCs in the association analysis . CD implicated GWAS loci were those loci defined as reaching genome-wide significance in International IBD Genetics Consortium studies ( Jostins , Ripke et al . , Nature 2012 ) and ( Liu et al . , Nature Genetics 2015 ) —Credible sets of SNPs around index associations were defined as in ( Huang et al . , Nature 2017 ) for fine-mapped loci , and for others credible sets were defined as all SNPs with r2 > 0 . 6 to the index variant . Genes within 50 kb of the span of credible set SNPs were considered “implicated’ by GWAS . As the present study aimed to focus on variation observed in Ashkenazi Jewish ( AJ ) population in comparison to reference populations in ExAC including ( non-Finnish Europeans ( NFE ) , Latino ( AMR ) , and African/African-American ( AFR ) ) we chose a model-based approach to estimate the ancestry of the study population using ADMIXTURE[12] . To identify AJ individuals and estimate admixture proportions we included a set ( n = 21 , 066 ) of LD-pruned common variants ( MAF>1% ) after filtering for genotype quality ( GQ>20 ) using the PLINK LD-pruning algorithm , whose description is available at http://pngu . mgh . harvard . edu/~purcell/plink/summary . shtml#prune . For the parameters , we selected a window size of 50 SNPs , a window shift of 5 SNPs at each step , and the variance inflation factor ( VIF ) threshold equal to 2 . The 18 , 745 samples were assigned to four groups ( K = 4 ) , as ancestry was defined as having a single estimated ancestry fraction ≥ 0 . 4 , and remaining three fractions < 0 . 4 ( S2 Fig ) . Individuals mostly representing African/African-American and East-Asian ancestry ( 1 , 267 and 569 individuals respectively ) were discarded from downstream analysis , as well as the 983 admixed individuals with none of the ancestry fractions ≥ 0 . 4 . Thus , a total of 6 , 093 individuals were considered of Ashkenazi Jewish ( AJ ) ancestry , while 9 , 833 were considered to represent Non-Finnish Europeans ( NFE ) . After sample QC and relatedness check , 5 , 685 individuals of Ashkenazi Jewish and 7 , 240 of non-Finnish European ancestry were found with valid IBD case/control status ( S1 Table ) . Individuals with Ulcerative Colitis and unspecified and Indeterminate Colitis were further excluded , resulting in 4 , 899 AJ and 5 , 066 NFE individuals . Prior to enrichment and association analysis , 81 samples ( of total 18 , 745 ) were also filtered due to possible contamination ( heterozygous/homozygous ratio < 1 ) , excess of singletons ( n>2000 ) , deletion/insertion ratio ( >1 . 5 ) and mean genotype quality ( <40 ) . 275 samples were excluded for relatedness ( >0 . 35 cut-off ) . Genotypes with low genotype quality ( <20 ) were filtered , in addition to variants with low call rate ( <80% ) and allele balance deviating from 70:30 ratio for greater than 40% of heterozygous samples if at least 7 heterozygous samples were identified . As we were interested in computing an enrichment statistic that would not be affected by possible admixture , we obtained alternate allele frequency estimates by restricting the enrichment analysis to the 2 , 178 non-IBD Ashkenazi Jewish samples that passed QC and relatedness filtering and had AJ focused ancestry fraction > 0 . 9 ( S1 Fig ) . Principal Component Analysis ( PCA ) was done in each ancestry group using the 21 , 066 variants . Sample QC was done using the Hail software while PCA , differential missingness and sample relatedness analysis was done using PLINK[54] . Hail is an open-source software framework for scalably and flexibly analyzing large-scale genetic data sets ( https://github . com/broadinstitute/hail ) . Allele balance was calculated using PLINK/SEQ ( https://atgu . mgh . harvard . edu/plinkseq/ ) . Assuming log-additive effects in the logit risk model the disease probability for an individual is given as p = ( 1 + exp ( −η ) ) −1 , where η tends towards a normal distribution with parameters μ=log ( p0/ ( 1−p0 ) ) +∑m=1M2fmβm and σ2=2∑m=1Mfm ( 1−fm ) βm2 [35] . Here p0 refers to a baseline disease probability . We can see that μ may be expressed in terms of the expected polygenic risk score , i . e . μ=log ( p0/ ( 1−p0 ) ) +E[PRS] . In the setting where E[PRS]=0 , then E[p]= ( 1+exp ( −log ( p0/ ( 1−p0 ) ) ) ) −1=p0 . To evaluate the impact of a shift in the expected value of polygenic risk score to the expected value of μ we can express the shift as E[Differenceμ]=E[DifferencePRS] . We can compute new values of p for new values of μ to obtain a fold-increase in prevalence for a population that has undergone such a shift . We see that this requires a value to be chosen for p0 and that log ( p0/ ( 1 − p0 ) ) can be represented as a baseline risk score value β0 . To get an estimate of the absolute prevalence of CD in the AJ population , we must choose a baseline β0 , where p0 represents the expected prevalence with zero non-baseline alleles in the population[35] , to which we add a contribution from multiple non-baseline alleles to calculate: 1 ) an individual’s probability of disease , or 2 ) the expected prevalence of the disease in the population . Once we have chosen a value for β0 , we can calculate the ratio of expected prevalence as follows . First , use the means ( μAJ and μNAJ ) and variances ( σAJ2 and σNAJ2 ) of risk scores as calculated above to calculate the probability density function of the disease prevalence . In the case of the AJ population , we have f ( p ) =dndg1σAJϕ ( η−μAJσAJ ) =1σAJp ( 1−p ) ϕ ( 1σAJlog ( p1−p ) −μAJσAJ ) where η is the risk score associated with prevalence p , g is the link function , so p = g ( η ) = ( 1 + e−η ) −1 , and ϕ is the standard normal density function . Next , we integrate to get ∫01p∙f ( p ) dp=E[pAJ] . Finally , we can calculate E[pNAJ] in a similar way , and divide the expected prevalence in the AJ population by that in the non-AJ population to get the prevalence ratio , E[pAJ]/E[pNAJ] . The value of β0 = -20 . 5 was chosen in order to obtain a prevalence in the non-AJ population of ~0 . 5% . At this value of β0 , the ratio of prevalence in the AJ population to that in the non-AJ population was estimated to be 1 . 5 ( E[pAJ] = 0 . 82% , E[pNAJ] = 0 . 55% ) . For different choices of β0 , however , this ratio may vary , as the relationship between probability of disease and risk score is non-linear . S10 Fig shows how the values of the disease prevalence and their ratio vary as β0 is changed . We see that the ratio values range from 1 . 46 to 1 . 52 for different values of β0 with a range of baseline prevalence of . 001 to . 01—the range of prevalence estimates for Crohn’s disease[41 , 43 , 57] . To further understand the effect that choosing a logit-based model had on the results , a comparison of the standard logit and probit models was done using the values inferred from the logit model . No full scale probit modelling was done in this analysis , so the values found with the probit model represent only a close approximation of the expected results . In the logit model for population analysis , we may assume that individual risk scores are chosen from a normal distribution N ( μlogit , σlogit2 ) where μlogit and σlogit represent the mean and standard deviation of the risk scores as defined above . From here , we may calculate the probability density function of probit model risk scores μprobit based on that of logit model risk scores μlogit as f ( ηprobit|μlogit , σlogit ) =f ( ηlogit|μlogit , σlogit ) dηlogit/dηprobit and use this to calculate μprobit and σprobit , the estimated mean and standard deviation of the risk scores in the probit model . Using these values , we obtain a probability distribution for the frequency of disease in the populations using the probit model . While the logit model yielded a prevalence ratio of 1 . 506 , the probit estimation yielded a prevalence ratio of 1 . 5136 , with similar expected prevalence values ( E[pAJ] = 0 . 823% , E[pNAJ] = 0 . 544% ) . These values demonstrate that individual logit and probit analyses would likely give similar results for values of interest . The complete probability densities under the logit and probit models can be seen in S8 Fig . Further , it is interesting to compare the relationship between values of risk scores in the two models . For values of risk scores between -1 and 1 in the logit model , the relationship to those in the probit model is highly linear , with a formula of ηprobit = 0 . 6223 ∙ ηlogit , with r2 = 1 . 0000 . This formula may be used to impute single values in one model or the other assuming that the estimated total risk score is otherwise close to zero , and the imputed value is low . It is worth noting , however , that this does not work for all values of ηlogit , as the relationship between risk score in the logit and probit models deviates from this simple linear model when the risk score values are large . When modeling enrichment , we chose a standard significance cutoff of p < 0 . 05/N for classifying variants as enriched . We noted that the number of variants classified as enriched does not change significantly when the p-value threshold changes . See S11 Fig for more information . | The Ashkenazim are a people with ancestry in northern-European Jewish groups . A founder effect caused a bottleneck in this population approximately one thousand years ago , resulting in a group of enriched alleles in their genetic makeup . A higher documented prevalence of Crohn’s Disease in the Ashkenazim indicates that some enriched alleles may confer risk of having this disease . By studying which genes are enriched , and which of these contribute to Crohn’s Disease risk , we are better able to understand the genetic architecture of the affected population , and of the disease itself . Further , we are able to develop a resource containing tables of significantly enriched alleles that are known or suspected to contribute to other disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"crohn's",
"disease",
"population",
"genetics",
"immunology",
"variant",
"genotypes",
"alleles",
"genetic",
"mapping",
"clinical",
"medicine",
"genetic",
"predisposition",
"gastroenterology",
... | 2018 | Insights into the genetic epidemiology of Crohn's and rare diseases in the Ashkenazi Jewish population |
Tick feeding causes extreme morbidity and mortality to humans through transmission of pathogens and causes severe economic losses to the agricultural industry by reducing livestock yield . Salivary gland secretions are essential for tick feeding and thus , reducing or preventing saliva secretions into the vertebrate host is likely to reduce feeding and hinder pathogen life cycles . Unfortunately , the membrane physiology of tick salivary glands is underexplored and this gap in knowledge limits the development of novel therapeutics for inducing cessation of tick feeding . We studied the influence of inward rectifier potassium ( Kir ) channel subtypes to the functional capacity of the isolated tick salivary gland through the use of a modified Ramsay assay . The secreted saliva was subsequently used for quantification of the elemental composition of the secreted saliva after the glands were exposed to K+ channel modulators as a measure of osmoregulatory capacity . Lastly , changes to blood feeding behavior and mortality were measured with the use of a membrane feeding system . In this study , we characterized the fundamental role of Kir channel subtypes in tick salivary gland function and provide evidence that pharmacological inhibition of these ion channels reduces the secretory activity of the Amblyomma americanum salivary gland . The reduced secretory capacity of the salivary gland was directly correlated with a dramatic reduction of blood ingestion during feeding . Further , exposure to small-molecule modulators of Kir channel subtypes induced mortality to ticks that is likely resultant from an altered osmoregulatory capacity . Our data contribute to understanding of tick salivary gland function and could guide future campaigns aiming to develop chemical or reverse vaccinology technologies to reduce the worldwide burden of tick feeding and tick-vectored pathogens .
The neuroendocrinology and genetic regulation of tick salivary glands has been researched extensively in an effort to identify novel acaricide target sites that can alleviate the burden of tick-borne pathogens [1–10] . Unfortunately , the significant advancements in knowledge relative to tick genomics , tick saliva proteins , and vaccine technologies have translated poorly into successful control efforts . Compounding tick control efforts is the increase in tick-borne bacterial infections . For example , rickettsial diseases are steadily increasing within the Americas [11] and further , recent studies have shown that the most predominant human biting tick , Amblyomma americanum , is capable of acquiring , maintaining , and transmitting Rickettsia rickettsii [12] . The steady increase in tick populations and pathogens [13] , increased vector competency of human biting ticks for rickettsial diseases[12] , and the movement toward an epidemic of tick-transmitted pathogens [14 , 15] highlights the significance of research aimed to identify novel mechanisms of control to curb the health and economic burden of ticks . Arthropod feeding results from the harmonious function of multiple organ systems that include olfactory and gustatory signaling to detect the food source , pharyngeal and cibarial pumps to generate a sucking action to imbibe fluid , and the salivary glands to secrete bioactive proteins that serve a variety of functions [16 , 17] . The documented importance of salivary secretions has stimulated attempts to develop vaccines against proteins in the secreted saliva to control tick [18] and horn fly [19] infestations of cattle and most recently , to prevent successful feeding of mosquitoes [20] . Interestingly , the efforts of previous vaccine development programs have primarily focused on the proteins secreted into the saliva and few efforts [21] have focused on the mechanisms enabling saliva secretion from the salivary gland . The tick salivary gland is multifunctional and performs a key role in two events during blood feeding . First , the tick salivary gland secretes many bioactive proteins that are critical for acquisition of the blood meal [22] and cementing the tick onto the host , suggesting that inhibition of salivary gland function will reduce blood feeding efficacy . Second , the salivary gland is responsible for maintaining a proper salt and water balance during blood feeding . Mammalian blood contains high concentrations of sodium and potassium salts that would be toxic to the tick without coordinated osmoregulatory mechanisms during feeding , which is performed through the salivary gland by returning about 65–70% of the fluid and ion content of the blood meal back into the host [23] . Importantly , failure to osmoregulate would alter the concentration of ions in the hemoceal and result in inhibition of cellular functions , functionally making the blood meal toxic in a similar manner described after inhibition of mosquito Malpighian tubules [24 , 25] . Further , arthropod saliva is the media through which pathogens are horizontally transferred to the vertebrate host . We hypothesize that inducing salivary gland failure will 1 ) prevent feeding , 2 ) prevent osmoregulation during blood feeding to induce acute mortality , and 3 ) prevent pathogen transmission through reduced salivary secretions . However , prior to developing products , it is necessary to understand the mechanisms of ion transport across the membranes of the salivary gland acini and how modulation of these pathways influences saliva and feeding . Potassium ion channels are a fundamental component of cellular physiology since they are responsible for establishing and maintaining the membrane potential of animal cells and serve crucial roles in cellular regulation [26] . In mammalian salivary glands , saliva secretion is associated with a pronounced efflux of K+ ions from the acinar cells , which results from an increase in basolateral membrane permeability to K+ ions [27 , 28] . In particular , inward rectifier potassium ( Kir ) channels and ATP-sensitive Kir ( KATP ) channels are critical for saliva secretion and regulation of the ion concentrations by maintaining the cellular membrane potentials of the mammalian salivary gland acini [29–31] . Although Kir and KATP channels are well characterized in multiple mammalian tissues [32 , 33] and are exploited as drug targets [34] , an understanding of the physiological role , expression patterns , and toxicological potential of these ion channels in ticks , and other arthropod disease vectors , is significantly less developed . Yet , the existing base of work suggests these channels serve significant roles in various arthropod tissues [35–40] and represents a putative insecticide target site [25] , illustrating the need for continued characterization of these channels . Kir channels function as biological diodes due to the unique biophysical property that allows the flow of potassium ions in the inward direction more easily than the outward direction at hyperpolarizing potentials [41] , which allows the cell to return to the resting potential by increasing the intracellular concentration of cations . Kir subunits lack the S4 voltage sensor region that is responsible for gating in all voltage-dependent ion channels; thus , Kir channels are constitutively active if regulator mechanisms , such as sulphonylurea receptors or ATP , are absent . All Kir channels share similar molecular architecture and are tetramers assembled around an aqueous membrane-spanning pore that are gated by polyvalent cations that occlude the pore at cell potentials more positive than the K+ equilibrium potential [42 , 43] . KATP channels are a sub-type of Kir channel superfamily and are octomeric complexes of four pore-forming Kir channel subunits and four regulatory sulfonylurea receptor ( SUR ) subunits [32] . Contrary to “constitutively active” Kir channels , KATP channels are regulated by the ratio of intracellular ATP:MgADP , and thus couple the membrane potential to the metabolic state of the cell [44] . In pancreatic beta-cells , when the ATP:MgADP ratio is elevated , KATP channels close to prevent K+ efflux that results in membrane depolarization and Ca2+-entry through voltage sensitive Ca2+-channels . Conversely , KATP channels open when the nucleotide ratio is decreased , which results in K+ ion efflux that leads to membrane hyperpolarization and cessation of Ca2+-entry . A gap in knowledge regarding the role of K+ ion channels in invertebrate salivary gland function and thus , saliva generation and secretion , has limited the ability to develop products that prevent tick blood feeding . To address this gap in knowledge , our group has aimed to characterize the role of Kir channels in the feeding cascade and salivary gland function of a medically relevant tick , A . americanum . Our previous work [35] and others [45] clearly illustrates the importance of Kir channels for proper salivary gland function and feeding of arthropods , which highlights the potential for these ion channels to be targeted by vaccine or chemical technologies in arthropods of medical significance . Therefore , the overarching goal of this investigation was to leverage multidisciplinary approaches to test the hypothesis that salivary gland function and feeding of A . americanum is reliant upon Kir channels expressed in the salivary gland . Knowledge gained from this study may be used to broadly guide future development of novel synthetic insecticides , RNAi , or transgenic organisms to mitigate human health concerns and curb economic losses that result from tick feeding .
The Kir channel inhibitor VU041 , VU937 ( inactive analog of VU041 ) , VU625 , and VU688 were originally discovered in a high-throughput screen ( HTS ) against the Anopheles gambiae and Aedes aegypti Kir1 channels , respectively [25 , 46] . These molecules were ordered by custom synthesis from Molport ( Rita , Latvia , Europe ) . VU590 and VU608 were initially discovered in a HTS campaign targeting human Kir1 . 1 channels [47] and were purchased from Tocris Bioscience or synthesized by Vanderbilt Center for Neuroscience Drug Discovery , respectively . All KATP channel modulators were purchased from Sigma-Aldrich ( St . Lewis , MO , USA ) . The inactive analog to VU0071063 , termed VU063-I , was generously provided by Dr . Jerod Denton ( Department of Anesthesiology , Vanderbilt University ) and was subsequently purchased by custom synthesis from Molport ( Rita , Latvia , Europe ) . All chemicals were designated to be >98% pure . Hanks Balanced Salt Solution ( HBSS ) with calcium chloride and magnesium chloride was purchased from Life Technologies ( 14025–092 ) and was used in all Fluid secretion ( Ramsay ) assays ( Ramsay , 1954 ) . Dimethyl sulfoxide ( DMSO ) was purchased from Sigma-Aldrich ( St . Louis , MO , USA ) . Clear 100% silicon II Chaulk ( General Electric , M90050 ) was purchased from Ace Hardware and was used for the construction of the feeding membrane . A . americanum adults were purchased from the Oklahoma State University Tick Rearing Facility ( Department of Entomology; Stillwater , OK , USA ) and were declared by the supplier to be free of all known pathogens . Adult ticks were approximately 1 month old at the time of analysis . Prior to allowing the ticks to blood feed , the ticks were stored for 14 days in an incubator at 28° C and 60% RH . Female salivary glands of partially blood fed ticks were dissected under a stereomicroscope in Hanks Buffered Salt Solution ( HBSS ) and the secreted ion and saliva was collected using a modified Ramsay assay [48] . Saliva secretion was stimulated by bath application of 100 μM dopamine HCl . Methods followed those that were previously published [7] . Concentration-response curves ( CRC ) were established with one gland of the tick exposed to dopamine only ( control ) and the second gland treated with the small-molecule modulator prior to dopamine expose , which enabled a paired statistical analysis . The physiology and gene expression profile is different between an unfed , feeding , and fed ticks [49 , 50] , thus all studies were performed on partially fed ticks to normalize gene expression amongst individuals . To obtain partially fed ticks , we adopted an artificial host system that enables blood feeding of multiple ticks on a silicone membrane in lieu of an animal host , which has been previously described [51] and is shown in Fig 1A–1D . Briefly , the membrane consisted of 10 g silicone glue ( Silicone II ) was mixed evenly with 3g silicone oil AR 20 ( Sigma-Aldrich; St . Louis , MO , USA ) and 2g Hexane ( Sigma-Aldrich; St . Louis , MO , USA ) . The silicone mixture was applied onto the lens paper purchased from Tiffen Company LLC ( Hauppauge , NY , USA ) . The mixture was spread evenly using a glass spreader to an approximate thickness of 60–80 μm [51] . The membrane was placed at room temperature for 14–21 days prior to use . The constructed feeding chamber is shown in Fig 1A and 1B . Glass chambers ( 28 mm outer diameter , 2 mm wall thickness , 45 mm cylinder with a 35 mm OD bead that is 12 mm up from the bottom of cylinder ) were ordered from Greatglass ( Wilmington , DE , USA ) and were used to construct the tick-feeding chamber . A membrane that was dried for >14 days were glued on the chamber with the same silicone glue at least one day before loading ticks . Glued chambers were placed into six-well cell culture plate with water to ensure the membrane did not leak . A total of 10 females and 5 males were loaded into one chamber with freshly shaved heifer hair . Cotton balls were inserted into the top of the chamber and approximately 1 cm of head space was maintained for tick feeding ( Fig 1C ) . After testing for leaks , the artificial membrane was immersed in 5 ml defibrinated bovine blood that was poured into a 6-well cell culture plate and tick feeding was allowed to commence immediately upon placement into the chamber ( Fig 1D ) . The blood was purchased from Hemostat Laboratories ( Dixon , CA , USA ) . Gentamycin and ATP were added to the blood at a concentration of 5 μg/ml and 1μM , respectively . The fluophore , rhodamine B ( 100 ppm ) , was included into the blood to ensure that all ticks included in data analysis had punctured the membrane and attempted to feed ( Fig 1E ) . Ticks without fluorescent mouthparts ( Fig 1E , white circle ) were not included in the analysis . Ticks were monitored every 12 hours and individual ticks were tracked over time by numerically numbering each tick ( Fig 1F ) . Feeding chambers were maintained in blood bath at 38°C with a 16 h:8 h light:dark photoperiod . The blood was changed twice a day at 12 h interval for the duration of the experiment . For measurement of secreted fluid , we employed the Ramsay assay that was initially developed to characterize the physiology of insect Malpighian tubules and the role of various membrane bound proteins to urine formation , urine secretion , and osmoregulation [48] . Modifications to enable measurements of secreted saliva of ticks have been previously described and used in this study [6 , 7 , 9] . Partially engorged female ticks ( weighing 10–20 mg , 3 days blood-fed on membrane ) were prepared from the artificial feeding system described above [51] . Salivary glands and the corresponding ductwork were dissected from the partially fed tick and then incubated in HBSS buffer for 1 h before initiating the Ramsay assay . After incubation , a single female gland was immersed in 15 μl HBSS buffer containing a HBSS + DMSO ( vehicle control ) or compound solubilized in HBSS . The main duct was drawn across a narrow grease dam made of high vacuum grease ( Dow Corning Corporation , Midland , MI , USA ) that was approximately 1 mm in height and immobilized on the surface of a petri dish . After drawing the main duct across the grease dam , the entire gland was submerged in HBSS and saliva secretion was stimulated by application of 100 μM dopamine HCl dissolved in HBSS . Basal levels of dopamine mediated saliva secretion measurements were taken in the first 5 minutes after preparation of the fluid secretion assay . A micro-injector ( Nanoliter 2010 , World Precision Instrument , Inc . , Sarasota , FL , USA ) controlled by a micro-syringe pump controller ( Micro4 , World Precision Instrument , Inc . , Sarasota , FL , USA ) was used for withdrawing the secretion formed at the tip of the duct in the heavy mineral oil , and the volume withdrawn was recorded in every 5 min for 30 min . To study the influence of K+ channel modulators ( Fig 2 ) on saliva secretion , the gland was incubated in the compound solution for 30 minutes prior to exposure to dopamine that initiates salivation . Concentration-response curves were performed with one gland of the tick exposed to dopamine only ( control ) and the second gland treated with the Kir channel modulator prior to dopamine exposure , which increased the rigor of the experimental design through paired analysis . Total salivation for each time point of the treated glands was compared to the volume secreted at the same time point for dopamine only treated glands to obtain percent saliva secreted when compared to control . For the ATP challenge experiment , glands were incubated in HBSS buffer containing both 500 μM ATP and 300 μM Pinacidil/1 μM VU063 for 30 min before dopamine exposure . The salivation time course data presented for the CRCs were collected using paired salivary glands where one salivary gland was treated with dopamine only and the other salivary gland was treated with the small-molecule modulator plus dopamine . The ability to pair the glands negated the variability due to factors that influenced individual ticks , such as blood meal size and age . The data points for each time point represent an average where n = 3–5 and the means for each time point of the treated glands was statistically compared to the same time point of the control ( dopamine only treatment ) by a paired t-test . Statistical significance was assessed based on P<0 . 05 . The concentration required to inhibit secretion by 50% ( IC50 ) values were determined through the generation of concentration-response curves that were constructed with 5–6 concentrations . The percent salivation for each concentration of pinacidil/VU063 was determined by the formula: ( secreted volume of gland treated with chemical + dopamine / secreted volume of gland treated with dopamine only ) * 100 . Each comparison was made from paired glands and the data points for each concentration represent the average % salivation of 3–5 paired glands . IC50 values were calculated by nonlinear regression ( variable slope ) using a Hill equation in GraphPad Prism ( GraphPad Software , San Diego , CA , USA ) . For the ATP challenge experiment , bars represent mean ( n = 3–5 ) volume of secreted saliva over a 5 min period from 15–20 minutes while the error bars represent SEM . A one-way ANOVA with a multiple comparisons post-test was performed to determine statistical significance compared to dopamine only treatment . Ticks fed on control blood ( vehicle only ) or treatment blood ( pinacidil or VU0071063 ) were monitored every 12 hours and the time of attachment , the time of detachment , and the time of reattachment was recorded for each tick . The changes in feeding biology was determined by quantifying the number of detachments per tick and the time until the first detachment for the treated groups compared to the non-treated control group . The average detachment per tick per feeding event was determined by counting the number of detachments for each female tick in the feeding chamber until mortality or completion of feeding . Mortality was not considered a detachment . The total number of detachments per tick was summed in each chamber and the detachment rate for 3 feeding chambers were averaged ( n: 30–40 individuals ) . To quantify the volume of imbibed blood , individual unfed ticks were labeled as described before and were weighed to the nearest 0 . 1 mg . Ticks were subsequently removed from the feeding chamber after 1- , 2- , 3- , 4- , 5- , and 6-days of known feeding . Ticks that detached or died prior to the required feeding time were discarded from data analysis . At the predetermined time point , the attached tick was removed from the membrane and immediately weighed to determine the change in mass where an increase of 1 mg in total weight corresponded to 1 μL of blood ingested . Ingestion was corrected for the specific gravity ( 1 . 030 at 37°C ) of bovine blood . Tick mortality was defined as a tick completely non-responsive to mechanical stimuli and only ticks that were previously attached were included in this measurement . The secreted saliva was analyzed via scanning electron microscopy/energy-dispersive X-ray spectroscopy ( SEM-EDS ) with a silicon drift detector ( SDD ) to obtain qualitative elemental composition and concentration of dopamine only- , pinacidil- , and VU0071063- treated glands . A Dual-Beam Focused Ion Beam ( FIB ) SEM equipped with EDS ( EDAX ) at the LSU Socolofsky Shared Instruments Facility ( SIF ) was used to image and analyze each dried saliva spot . The parameters of the image was set to be 10 kV and 5 . 7 nA current . We used EDAX TEAM software to acquire spectra to identify the Na+ , K+ , and Cl- ions based on the characteristic X-ray . Due to the use of a silicon background , silicon was excluded from the analysis and only Na+ , K+ , and Cl- were used for concentration calculations , but the value of silicon was used as an internal control . The concentration of each element was determined by the construction of a standard curve that was developed by spotting NaCl ( in mM: 20 , 50 , 100 , 200 , 400 , 600 , 800 ) and KCl ( in mM: 2 , 4 , 8 , 16 , 32 , 64 , 100 ) onto the silicon substrate with differing concentrations . The standard curves for each ion were generated based on atomic percentage values of each concentration , which were compared to the standard curves to enable calculation of the relative concentration of the elements . The atomic percentages of nine saliva droplets ( 1 nL ) were determined using the SEM-EDS equipment and averaged to obtain a mean percent atomic for each individual time point and treatment group . The mean was compared to the percent atomic of the standard curve that was generated for each element ( Na+ , Cl- , and K+ ) and the concentration of total Na+ , Cl- , and K+ in the whole saliva droplet was determined with the formula for a linear regression ( y = mx + b ) . The molar concentration of the element was multiplied by the dilution factor that stemmed from the reconstitution of the dried saliva and then divided the product by the total volume of saliva secreted . This resulted in the total concentration of each element per nL of secreted saliva .
Here , we used the moderately developed pharmacological library of insect Kir channels [24 , 25 , 46 , 52] to test the influence of Kir channels to the secretory activity of the tick salivary gland . The volume of secreted saliva after exposure to VU041 and VU625 was not significantly ( P>0 . 05 ) different at any time point when compared to dopamine only control glands ( Fig 2A ) . The pharmacological profile of human KATP channels is well characterized [53] and enabled a more complete examination of tick KATP channels with multiple structural scaffolds of inhibitors and activators . The KATP inhibitors tolbutamide and glibenclamide were found to have minimal impact on the volume of secreted saliva at the 10–30 minute time points , but significantly ( P<0 . 01 ) increased secretory activity at 5-minutes by approximately 3-fold when compared to paired control glands ( Fig 2B ) . Select pharmacological activators of KATP channels were shown to nearly abolish the secretory activity of the isolated salivary gland at a discriminatory concentrations ranging from 500 μM to 1 mM ( Fig 2C ) . Pinacidil was shown to prevent nearly all secretory activity at time points ranging from 10 min to 30 minutes with the 5- minute time point producing the largest volume of secreted saliva at an average of 30 ± 12 nL , which was significantly ( P<0 . 05 ) reduced when compared to dopamine-only treated ( control ) glands ( 81 ±11 nL/5 min ) at the same time point ( Fig 2C ) . The second KATP agonist studied , VU0071063 , was shown to have a near identical pattern of inhibition when compared to pinacidil , but at a 3-fold less concentration . VU0071063 treated glands secreted less than 3 nL of saliva at the 10–30 minute time points and only secreted an average ( n = 3 ) of 23 ± 9 nL of saliva whereas the paired control glands secreted 107 ± 21 nL at the 5-min time point , a statistically significant ( P<0 . 001 ) reduction ( Fig 2C ) . Diazoxide , nicorandil , and minoxidil are three structurally diverse activators that have varying specificity for human KATP channel subtypes , but were found to not influence the secretory activity of the isolated tick salivary gland ( Fig 2C ) . In mammalian systems , the off-target effects of pinacidil and VU0071063 are minimal [54] . However , the possibility of modulating the activity of off-target proteins that may alter tick salivary gland activity remains present and therefore , we aimed to ensure the reduced secretory activity was indeed due to modulation of KATP channels . First , we generated concentration response curves to determine the concentration required to inhibit 50% ( IC50 ) of the secretory activity to describe the potency of each molecule and ensure concentration dependency . Pinacidil was shown to be moderately potent at reducing the secretory activity of the A . americanum salivary gland with IC50 values in the mid micromolar range . At 10 minutes , pinacidil was shown to have an IC50 value of 389 μM ( 95% CI: 218–692 μM; Hillslope: -2 . 4; r2: 0 . 89 ) whereas the 20- and 30-minute IC50 values were found to be 1 . 56- and 1 . 50-fold reduced , respectively Fig 3A . The Kir activator VU0071063 was shown to be 169-fold more potent than pinacidil with an IC50 value of 2 . 3 μM ( 95% CI: 1 . 1–4 . 7 μM; Hillslope: -1 . 7; r2: 0 . 85 ) at 10 minutes . An increase in potency of VU0071063 was also observed at 20 and 30 minutes with IC50 values of 4 . 6 μM ( 95% CI 2 . 8–7 . 4 μM; Hillslope: -1 . 9; r2: 0 . 88 ) and 2 . 2 μM ( 95% CI: 1 . 1–4 . 7 μM; Hillslope: -1 . 0; r2: 0 . 9 ) , respectfully ( Fig 3B ) . Representative images of the secretory activity of the isolated salivary gland before stimulation , during stimulation , and after KATP channel exposure is shown in Fig 3C . KATP channels are inhibited by the presence of ATP , which provides an avenue to ensure that the pharmacological activators are modulating KATP channels to induce a physiological response . In theory , the presence of ATP should irreversibly close the KATP channel and prevent pharmacological activation of the channel , thus reducing the potency of pinacidil and/or VU0071063 against the tick salivary gland . To test this hypothesis , the salivary gland was co-treated with ATP and pinacidil or VU0071063 to measure the changes in secretory activity . Exposure to 500 μM ATP did not influence the secretory activity of the isolated gland ( Fig 3D ) when compared to dopamine controls , whereas higher concentrations were shown to reduce the secretory activity . Importantly , the mean ( n = 3–5 ) volume of secreted saliva after the isolated salivary gland was treated with 500 μM ATP and 300 μM pinacidil was found to be 127 ±9 nL / 5min , which was not statistically different from control secretion volumes ( Fig 3D ) . Similarly , the volume of the secreted saliva after the isolated salivary gland was treated with 500 μM ATP and 1 μM VU0071063 was found to be 111 ± 17 nL / 5min , which was significantly increased from the volume secreted after 1 μM VU0071063 exposure , but not statistically different from control secretion volumes ( Fig 3D ) . These data combined with the concentration dependency provide significant support that pinacidil and VU0071063 are indeed modulating the secretory activity through activation of KATP channels . During the development of VU0071063 , it was discovered that removal of the methyl group from the theophylline moiety dramatically decreased potency at the human Kir6 . 2/SUR1 channel ( molecular structure shown in insert of Fig 3E ) . The significant loss of potency of the VU0071063 analog , termed VU063-I , provided an additional mechanism to ensure the reduction of salivary gland secretory activity after VU0071063 exposure was due to KATP channel activation . Indeed , exposure of the isolated tick salivary gland to VU063-I did not alter the secretory activity of tick salivary glands at any time point when compared the secretory activity of paired control glands ( Fig 3E ) . Next , we aimed to translate the in vitro data to a biological response in the live tick and tested the hypothesis that inducing salivary gland failure will prevent tick feeding and blood ingestion . With a membrane feeding system optimized for A . americanum [51] , control and pinacidil treated ticks were shown to imbibe nearly the same volume of blood at days 1 and 2 , with mean ( n = 3–9 ) ingestion of control and treated ticks 0 . 7 ± 0 . 2 μL and 1 . 3 ± 0 . 4 μL , respectively ( Fig 1F ) . At 3 , 4 , 5 , and 6 days of feeding , control ticks were shown to imbibe 1 . 5- , 2 . 2- , 12 . 1- , and 10 . 8-fold more blood when compared to ticks feeding on pinacidil-treated blood , respectively ( Fig 3F ) . The total volume of blood ingestion by ticks exposed to pinacidil was significantly reduced ( P<0 . 05 ) for each time point . Fluorescent images showing the reduced ingestion of blood are shown in ( Fig 3G ) . We expected that VU0071063 would reduce the ability of actively feeding ticks to ingest blood with a greater efficacy than pinacidil , since VU0071063 was found to be significantly more potent at reducing the fluid secretion from the isolated salivary gland . Unfortunately , the increased mortality rate of VU0071063 ( 300 μM ) exposed ticks ( ca . 90% at day 2 ) prevented the ability to accurately determine the ingestion volume beyond day 1 . In addition to reduced blood ingestion , we analyzed the altered feeding behavior of ticks since horizontal transmission of bacterial pathogens does not occur until at least 12 hours of feeding has occurred [55] , suggesting that an interruption of blood feeding ( e . g . detachment ) prior to this time point could occlude pathogen transmission . Individual ticks in the control group were found to detach from the membrane an average of 0 . 23 ± 0 . 1 times during the course of a complete feeding event ( Fig 4A ) . Ticks feeding on pinacidil treated blood increased the number of detachment events per tick to an average of 2 . 1 ± 0 . 4 times per blood feeding event ( Fig 4A ) , a statistically significant ( P < 0 . 001 ) increase when compared to control groups . Similarly , ticks exposed to VU0071063 during blood feeding detached at a significantly ( P<0 . 001 ) greater rate with an average of 1 . 9 ± 0 . 2 detachments ( Fig 4A ) . In addition to the total number of detachments , we analyzed the average time from initial attachment onto the membrane to the first detachment because the time of detachment has direct impact for pathogen transmission . A total of 5% of the control ticks were shown to detach during feeding and the first detachment was found to be 112 ± 18 hours ( Fig 4B ) . Importantly , the time until first detachment was approximately 60–70% of the feeding time required to obtain a complete blood meal ( 168–200 hours or 7–9 days ) with the artificial host membrane feeding system . On the contrary , ticks exposed to pinacidil or VU0071063 were found to feed for an average of 28 ± 3 hours and 15 ± 1 . 5 hours prior to the first detachment , a statistically significant ( P<0 . 001 ) reduction when compared to control ( Fig 4B ) . Elemental analysis of the secreted saliva from isolated salivary glands exposed to pinacidil ( 500 μM ) and VU0071063 ( 5 μM ) significantly altered the concentrations of Na+ , K+ , and Cl- ions , indicating altered osmoregulatory capacity . The relative concentration of Na+ , K+ , and Cl- was statistically ( P<0 . 01 ) increased after exposure to VU0071063 for all time points except 5-minutes ( Fig 5A ) , whereas all time points studied after pinacidil treatment was stastistically significant when compared to control ticks ( Fig 5B ) . A representative crystalized salivary droplet is shown in Fig 5C . These data indicate that KATP channels represent a critical K+ ion conductance pathway in the A . americanum salivary gland and modulation of this pathway prevents fluid secretion ( Figs 2 , 3A and 3B ) , tick blood feeding ( Fig 3F and 3G ) , and altered osmoregulatory capabilities ( Fig 5 ) . The time to effectively kill 50% of ticks ( ET50 ) was found to be significantly different between pinacidil and control treatments with pinacidil ET50 being 1 day whereas the ET50 for control groups was 5 days . At day one , we observed a mean 13 ± 1 . 0% mortality in control groups and a 50 ± 8% mortality at day 1 of pinacidil fed ticks ( Fig 6A ) . Further , mortality exceeded 90% at day 5 in pinacidil treatment groups whereas the same percent mortality was not reached in control groups until 13 days of feeding . Similarly , ticks feeding on VU0071063 ( 300 μM ) treated blood meals yielded mortality that was significantly greater than the mortality observed in control groups and at a faster rate when compared to pinacidil even though it was at a 3-fold lower concentration than pinacidil ( Fig 6A and 6B ) . Statistical significance ( P<0 . 05 ) was observed for mortality at each time point . The ET50 for VU0071063 was found to be less than 12 hours with 71 ± 4% mortality at the 12-hour time point . The ET50 for the control group was found to be 4 . 75 days , which is more than a 9 . 5-fold increase when compared to the ET50 of VU0071063 treated animals . Mortality for VU0071063 treated groups reached 100% 3-days after the start of feeding , whereas control ticks did not reach 75% mortality until day 8 of feeding ( Fig 6B ) . Ticks that fed on pinacidil- or VU0071063- treated blood were lethargic and had uncoordinated movements when they attempted to walk or move their chelicerae . Importantly , injection of pinacidil or VU0071063 into partially fed or non-blood fed adult ticks did not induce mortality or any signs of intoxication .
Tick saliva is critical for feeding of arthropods by carrying proteins that have immunomodulatory , anti-hemostatic , and anesthetic properties [17 , 22 , 56 , 57] . In addition to enabling feeding , arthropod saliva is the media through which pathogens are horizontally transferred to the vertebrate host . Further , the volume of secreted saliva is directly correlated to the severity of infection since most pathogens have an infectious dose and since saliva can affect immune cells to exacerbate infection [58] . This suggests that eliminating or reducing the volume of saliva secreted into the host will prevent feeding and reduce pathogen infectivity . Therefore , the objective of this study was to define the relevance of a conserved K+ ion channel to successful feeding of ticks to provide a platform for future research aimed at developing novel therapeutics to reduce or eliminate the health and economic burden from blood feeding of ticks and other arthropods . Previous work has shown that inward potassium conductance is critical for mammalian salivary gland function [28 , 30 , 31] and may represent a target for inducing arthropod salivary gland failure [35 , 45] . The data presented in this study support the hypothesis that KATP channels in the A . americanum salivary gland are critical for feeding as we provide clear evidence that pharmacological activators of these channels reduce the secretory capacity of the salivary gland and hinder blood feeding . These data suggest that KATP channels regulate the epithelial physiology of the arthropod salivary gland function in a similar manner as Kir channels in mammalian salivary glands [30 , 31] and , from an applied perspective , represent a putative target site for the development of therapeutic agents that induce salivary gland failure to prevent feeding and/or osmoregulation to result in mortality . In mammals , Kir channels represent a critical conductance pathway in multiple tissues , yet these channels have only recently gained traction as a physiologically and toxicologically relevant ion channel in insects [24 , 25 , 52 , 59] . Recent work on insect Kir channels suggests that these channels drive Malpighian tubule function by providing a transport pathway for K+ ions from the hemolymph to the Malpighian tubule lumen [36] . Importantly , in vitro and in vivo results indicate that small-molecule inhibitors specific for insect Kir channels disrupt K+ and fluid secretion at the level of the Malpighian tubules , leading to disruptions of hemolymph K+ and fluid homeostasis of the whole mosquito to result in mortality [24 , 25 , 36] . Mortality was attributed to the altered ion and fluid homeostasis and due to this , it has been proposed that Kir channels expressed in the Malpighian tubules represent a novel mechanism target site for arthropod control . Interestingly , the Malpighian tubules and salivary glands are physiologically related tissues as both tissues are a polarized epithelial tissue [60 , 61] that rely on potassium ion transport across membranes [62–64] to generate isosmotic fluid to form urine or saliva . Furthermore , the Kir-encoding gene of the mosquito Malpighian tubule is analogous to the Kir encoding gene in the salivary gland of Drosophila [65] , which raised the intriguing possibility that Kir channels control the membrane potential and secretory activity of the arthropod salivary gland as it does in the mosquito Malpighian tubules [36] . Indeed , salivary gland specific knockdown of Kir1 mRNA or pharmacological inhibition of Kir1 channels resulted in a significant reduction of sucrose ingestion during D . melanogaster feeding [35] . Further , recent work has shown that inhibition of Kir1 channels in the brown planthopper ( Nilaparvata lugens ) reduces salivary and honeydew secretions [45] . These studies provide proof-of-concept that Kir channels represent an important channel for feeding and proper function of the arthropod salivary gland and represents a pharmacologically tractable target site to induce salivary gland and feeding failure in arthropod vectors . Two classical Kir channel inhibitors , VU041 and VU625 , did not influence the secretory activity of the tick salivary gland , which was surprising since these two molecules are highly potent inhibitors of mosquito Kir1 channels [24 , 25] and because Kir2 . 1 , which is a constitutively active channel similar to insect Kir1 , was identified to be partly responsible for spontaneous fluid secretion in mammalian salivary glands [31] . Although nucleotide or protein differences between mosquito and tick Kir channels may account for reduced efficacy of these molecules , we speculate the lack of activity of classical Kir channel modulators to tick salivation is due to different physiological requirements of blood feeding arthropods when compared to ruminant mammals or fruit flies , such as the requirement for ATP to stimulate blood feeding [66] . Since ATP is required for blood feeding in hematophagous arthropods , we hypothesized that blood feeding arthropods evolved the use of KATP channels instead of classical Kir channels to maintain the membrane potential and regulate the secretory activity of the salivary glands . Indeed , two structurally and mechanistically different activators of KATP channels were shown to reduce the secretory activity of the A . americanum salivary gland in a concentration-dependent manner . Importantly , opposite patterns of salivation were observed with activators and inhibitors at 5 minutes , the inhibition of salivation was concentration dependent , and the reduction of salivation in ticks after exposure to pinacidil/VU0071063 was negated with the application of ATP . These data provide substantial evidence that the reduced salivation is indeed due to modulation of KATP channels and supports the notion that KATP channels are required for salivary gland function and saliva secretion of A . americanum . However , it is important to note that although there is a close correlation between reduced feeding and reduced saliva secretion ( Fig 3 ) , it is conceivable that the reduced blood ingestion could be due to altered Kir/KATP channel function at the phagostimulant sensilla , the feeding pump muscles , the salivary gland , or any combination of these systems . Future work is needed to determine the role of K+ channels in each organ system as well as to functionally couple Kir channels , salivary gland function , and blood feeding . In addition to salivary secretions , the tick salivary gland is responsible for maintaining a proper salt and water balance during blood feeding by returning about 65–70% of the fluid and ion content of the blood meal back into the host to alleviate the burden of the increased salts derived from the blood meal [23] . Kir channels have been shown to be major routes of K+ ion uptake in the mosquito Malpighian tubules , which is the osmoregulatory organ of mosquitoes , and pharmacological inhibition of these channels induced tubule failure and altered osmoregulatory capabilities , which lead to mosquito mortality [24 , 25 , 36] . Accordingly , we hypothesized that pharmacological activators of Kir channels in the tick salivary gland would alter the ion secretion rates and elemental composition of tick saliva . Indeed , elemental analysis of the secreted saliva with SEM-EDS from isolated salivary glands exposed to pinacidil ( 500 μM ) and VU0071063 ( 5 μM ) significantly altered the concentrations of Na+ , K+ , and Cl- ions , indicating altered osmoregulatory capacity . The altered osmoregulatory capacity is important because strict regulation of the ion concentration in the hemolymph and hemoceal is essential for proper function of muscles , nerves , and other physiological systems . The dramatic increase in salt excretion during tick feeding led us to hypothesize that the altered osmoregulatory capacity of ticks when exposed to KATP agonists during blood feeding will lead to mortality in a similar manner to what has been described in mosquitoes [24 , 25] . Indeed , mortality was significantly ( P<0 . 05 ) increased for ticks that fed on pinacidil treated blood meals when compared to vehicle control treatments . The signs of intoxication prior to mortality were reminiscent of neural poisoning since the ticks that fed on chemically treated blood were lethargic and displayed uncoordinated movements . Yet , injection of pinacidil or VU0071063 into partially fed or unfed ticks did not induce toxicity or any signs of intoxication . This difference indicates the mortality from exposure to KATP modulators during blood feeding is likely resultant from reduced ion concentrations in the tick hemolymph stemming from altered osmoregulatory capacity through the salivary gland , similar to mosquito mortality described after Malpighian tubule failure from Kir channel inhibition [25 , 52] . Taken together , the osmoregulatory and toxicity data indicate the possibility of developing novel mechanism acaricides by targeting ion transport and ion conductance pathways in the tick salivary gland that are functionally linked to establishing or maintaining transepithelial ionic balance . In addition to the essential role the tick salivary gland has to blood feeding and osmoregulation , the volume of saliva secreted into the host during feeding is directly correlated to pathogen transmission and disease severity [58 , 67] . Thus , the reduction in secretory activity of the salivary gland after exposure to Kir/KATP modulators led to the logical assumption that horizontal transmission of a pathogen will be negated or reduced when the tick is exposed to Kir/KATP modulators during feeding . The dynamics of pathogen transmission have been studied extensively in ticks; and A . aureolatum was shown to successfully transmit a virulent strain of Rickettsia rickettsii to a vertebrate host after a feeding period of approximately 12–16 hours [55] , suggesting that inducing detachment or inducing mortality within 12–16 hours of feeding would likely prevent pathogen transmission for this vector and pathogen . Importantly , KATP agonists drastically reduced saliva secretions ( Figs 2 , 3A and 3B ) , reduced uninterrupted feeding time to approximately 15 hours ( Fig 4 ) , and dramatically reduced blood ingestion and presumably , the saliva secretions into host ( Fig 3F ) when exposed to pinacidil . Since pathogen acquisition by the arthropod vector is directly correlated to blood intake , it is plausible to suggest that ticks exposed to KATP modulators during feeding will acquire fewer infectious particles since the tick intakes less blood ( Fig 3F ) . Similarly , reduced salivary output is correlated to reduced disease severity because saliva can increase pathogen infection within a vertebrate host through effects on immune cells [58] and since reduced salivation yields less transmission of infectious particles to the host [67] . Further , it would be of great benefit to induce mortality with the same intervention method that is used to interrupt the dynamics of pathogen transmission . Indeed , pinacidil and VU0071063 increased the rate of mortality with VU0071063 inducing 75% mortality at 12 hours of feeding , which is believed to be the minimum feeding time required for bacterial pathogen transmission [55] . Therefore , these data suggest that KATP channel activators are likely to reduce or prevent pathogen transmission from A . Americanum and likely other species of hard ticks . Additional studies analyzing the horizontal transmission and acquisition of bacterial pathogens are required to validate this hypothesis . The data indicate that the pharmacological profile of tick KATP channels is different when compared to human KATP channels , since pinacidil and VU0071063 both prevented salivation , but are selective for different SUR proteins in humans [53] . In mammals , KATP channels consist of multiple heteromeric combinations of Kir6 . x and SURs that result in different pharmacological sensitivities and reflect the various KATP channels in native tissues . For instance , glibenclamide blocks the Kir6 . 2/SUR1 channel and is significantly less potent against Kir6 . 2/SUR2A and Kir6 . 2/SUR2B [68] , whereas tolbutamide inhibits Kir6 . 2/SUR1 currents with high affinity , but does not inhibit Kir6 . 2/SUR2A [68] . Similarly , VU0071063 and diazoxide selectively activate the Kir6 . 2/SUR1 channel , whereas pinacidil preferentially activates Kir6 . 1/SUR2 channels [54] . Yet , in ticks diazoxide did not influence salivation whereas pinacidil and VU0071063 reduced secretory activity . The differences in pharmacological profiles between arthropods and humans suggest a unique heteromeric assembly of Kir and SUR constructs that can be exploited for the development of therapeutic agents that are highly specific for the arthropod KATP channel over the KATP mammalian channel , enabling vaccination or treatment of hosts with chemical modulators . Determining the specific functional role of Kir channels at the cellular level of the salivary gland is important for the subsequent development of products aiming to prevent salivation and feeding and should be a focus of future work . Based on the role of Kir channels in mosquito Malpighian tubules [36 , 38] and in mammalian salivary glands [29 , 31] , we speculate Kir/KATP channels are responsible for maintenance of the acinar membrane physiology that is necessary for function of downstream neuroendocrine systems that regulate fluid secretion across epithelia , such as dopamine . The membrane potential is known to control calcium mobilization , which stimulates prostaglandin E2 and subsequent secretion of bioactive salivary proteins [69] . Therefore , we speculate that Kir/KATP channels indirectly regulate calcium mobilization by regulating the membrane potential of the acini . For instance , KATP channels are open during times of low metabolic activity ( i . e . low ATP/ADP ratio ) , which results in hyperpolarization of the membrane that prevents flux of calcium ions through voltage-gated calcium channels [70] . Therefore , pharmacological activation of KATP channels would lead to hyperpolarization of the salivary gland acini membrane , preventing calcium release and inhibiting gland function , which is supported by our data ( Figs 2 and 3 ) . However , the precise role of Kir channels in salivary gland function and their involvement with maintaining intracellular currents is difficult to ascertain without electrophysiological data describing how these channels interface with other ion channels and neuroendocrine systems . This translational study highlights the critical role KATP channels provide in tick salivary gland function and the data suggest these proteins , and likely other membrane bound proteins linked to salivary gland function , represent novel therapeutic targets for measures that can be applied to reduce tick populations and/or tick vectored pathogens . These data support the notion that inward K+ conductance pathways in the tick salivary gland represent a putative therapeutic target ripe for development . Further , tick Kir/KATP channels appear to have maintained their fundamental role in animal physiology but the different pharmacological profile of KATP modulators between arthropods and humans suggests different protein composition that highlights the likelihood to develop selective agents of control that have low affinity to Kir channels of human , beneficial ( e . g . honey bee ) , or non-target species . The data presented in this study provide insights of broad interest given the importance of insect feeding to the worldwide health and economic burden resulting from arthropod feeding . | Tick feeding results in negative health and economic consequences worldwide and there has been continued interest in the development of products with novel mechanisms of action for control of tick populations . Kir channels have been shown to be a significant ion conductance pathway in arthropods and are critical for proper functioning of multiple biological processes . Previous work on insect Kir channels has focused on their physiological roles in renal system of mosquitoes and the data suggest that these channels represent a viable pathway to induce renal failure that leads to mortality . Based on the functional and cellular similarities of arthropod salivary glands and Malpighian tubules , we hypothesized that Kir channels constitute a critical conductance pathway within arthropod salivary glands and inhibition of this pathway will preclude feeding . Data presented in this study show that pharmacological modulators of Kir channels elicited a significant reduction in the fluid and ion secretory activity of tick salivary glands that resulted in reduced feeding , altered osmoregulation , and lead to mortality . These data could guide the future development of novel acaricides , RNAi , or genetically modified ticks to mitigate health and economic damages resulting from their feeding . Further , these data indicate a conserved function of Kir channels within multiple tissues of taxonomically diverse organisms , such as ticks and humans . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"body",
"fluids",
"ixodes",
"chemical",
"compounds",
"saliva",
"animals",
"organic",
"compounds",
"neuroscience",
"hormones",
"physiological",
"processes",
"salivation",
"amines",
"neurotransmitters... | 2019 | Inward rectifier potassium (Kir) channels mediate salivary gland function and blood feeding in the lone star tick, Amblyomma americanum |
Both endocrine and exocrine pancreatic cells arise from pancreatic-duodenal homeobox 1 ( pdx1 ) -positive progenitors . The molecular mechanisms controlling cell fate determination and subsequent proliferation , however , are poorly understood . Unlike endocrine cells , less is known about exocrine cell specification . We report here the identification and characterization of a novel exocrine cell determinant gene , exocrine differentiation and proliferation factor ( exdpf ) , which is highly expressed in the exocrine cell progenitors and differentiated cells of the developing pancreas in zebrafish . Knockdown of exdpf by antisense morpholino caused loss or significant reduction of exocrine cells due to lineage-specific cell cycle arrest but not apoptosis , whereas the endocrine cell mass appeared normal . Real-time PCR results demonstrated that the cell cycle arrest is mediated by up-regulation of cell cycle inhibitor genes p21Cip , p27Kip , and cyclin G1 in the exdpf morphants . Conversely , overexpression of exdpf resulted in an overgrowth of the exocrine pancreas and a severe reduction of the endocrine cell mass , suggesting an inhibitory role for exdpf in endocrine cell progenitors . We show that exdpf is a direct target gene of pancreas-specific transcription factor 1a ( Ptf1a ) , a transcription factor critical for exocrine formation . Three consensus Ptf1a binding sites have been identified in the exdpf promoter region . Luciferase assay demonstrated that Ptf1a promotes transcription of the exdpf promoter . Furthermore , exdpf expression in the exocrine pancreas was lost in ptf1a morphants , and overexpression of exdpf successfully rescued exocrine formation in ptf1a-deficient embryos . Genetic evidence places expdf downstream of retinoic acid ( RA ) , an instructive signal for pancreas development . Knocking down exdpf by morpholino abolished ectopic carboxypeptidase A ( cpa ) expression induced by RA . On the other hand , exdpf mRNA injection rescued endogenous cpa expression in embryos treated with diethylaminobenzaldehyde , an inhibitor of RA signaling . Moreover , exogenous RA treatment induced anterior ectopic expression of exdpf and trypsin in a similar pattern . Our study provides a new understanding of the molecular mechanisms controlling exocrine cell specification and proliferation by a novel gene , exdpf . Highly conserved in mammals , the expression level of exdpf appears elevated in several human tumors , suggesting a possible role in tumor pathogenesis .
The pancreas is a mixed organ with endocrine and exocrine compartments . The endocrine portion contains four distinct hormone-producing cell types organized into islets of Langerhans . Autoimmune-mediated destruction of endocrine β cells causes type 1 diabetes [1 , 2] . β cell number also gradually declines in type 2 diabetes [2] . The exocrine portion includes acinar cells , which produce digestive enzymes , and duct cells , which form an elaborate duct system that transports these enzymes into the gut . The majority of malignant pancreatic cancers derive from the exocrine portion [3] . Development of all major pancreatic cell types , including endocrine , exocrine , and duct cells , requires the function of the pancreatic-duodenal homeobox 1 ( Pdx1 , also known as Ipf-1 ) gene [4 , 5] . The molecular mechanisms determining early cell fate and the subsequent proliferation of endocrine and exocrine cells are not fully understood . Identification and characterization of novel lineage-specific regulators of exocrine pancreas cell proliferation could shed light on the pathogenesis of pancreatic cancers . Morphogenesis of the pancreas in zebrafish ( Danio rerio ) shares some similarities to that in the mouse . In mice , the pancreas develops from one dorsal and one ventral bud that arise from the posterior foregut [6–8] sequentially . The recognizable dorsal pancreatic bud forms from the prepatterned endoderm at around 22–25 somites ( embryonic day 9 . 5 [E9 . 5] ) and the ventral bud arises slightly later at around 30 somites ( E10 . 25 to E10 . 5 ) . Then the dorsal and ventral buds fuse as a result of gut rotation at E12 . 5 [9] . Different endocrine cell types are specified at different stages . The α and β cells mature at E9 . 5 since glucagon and preproinsulin can be detected by immunohistochemistry [10] , whereas somatostatin can be detected only at E13 . 5 [11 , 12] . Initially , it had been thought that the zebrafish pancreas develops from a single pancreatic anlage that appears at around 15 h postfertilization ( hpf ) [13–15] . This posterodorsal pancreatic anlage gives rise only to endocrine cells . Using a gut:GFP transgenic line , however , Field et al . observed a second anlage ( ventral anlagen ) that arose from the foregut at 34 hpf [16] when exocrine cells begin differentiation . In addition to exocrine cells , this anteroventral anlage also contributes to endocrine cells that are scattered outside of the main islet [16] . The dynamic process of pancreatic development is controlled by extrinsic signals from the adjacent tissues and intrinsic transcription factors . Multiple signals including fibroblast growth factor [17 , 18] , bone morphogenetic protein [19] , Notch [17 , 20–22] , and sonic hedgehog [23] play critical roles for proper pancreas formation . A conserved role of retinoic acid ( RA ) has been reported in many organisms , including zebrafish [24 , 25] , Xenopus [26] , and mouse [27 , 28] . There are conflicting data , however , on the relative effects of RA on endocrine and exocrine pancreas differentiation . In Xenopus , RA treatment promotes endocrine at the expense of exocrine differentiation in the dorsal bud by inhibiting Notch signaling activity [26] . In zebrafish , RA treatment results in anterior expansion of endocrine and exocrine cells [25] . It appears that RA acts directly in the endoderm to induce endocrine pancreatic precursors [29] . In mouse embryonic pancreas cultures , all-trans retinoic acid ( atRA ) inhibits branching morphogenesis and exocrine cell differentiation but accelerates endocrine differentiation , possibly due to increased level of Pdx1 in the endocrine clusters [30] . The differential effects may be explained by the distribution of the RAR and RXR receptors in the developing mouse pancreas [31] . A network of intrinsic transcription factors that act in a cascade fashion to initiate and maintain cell-specific gene expression patterns determines the ultimate lineage-specific cell fate . One of the earliest transcription factors functioning in the developing pancreatic epithelium is PDX1 , which plays an essential role during the early phase of pancreas development . Mice with a targeted mutation in the Pdx1 gene exhibited no development of pancreatic tissue [4] . The agenesis of the pancreas is caused by an early arrest right after initial bud formation [4 , 5] . Furthermore , multiple roles of Pdx1 in cell lineage determination during pancreas formation has been revealed by lineage tracing using a modified version of Cre/lox technology [32] . Cells labeled between E9 . 5 and E11 . 5 give rise to all three pancreatic cell lineages including islet , exocrine acini , and ducts . Conversely , cells labeled at E8 . 5 and E . 12 . 5 or thereafter give rise only to endocrine and acinar cells [32] . These results suggest that temporal regulation of Pdx1 expression is critical for cell fate determination . In addition , several transcription factors have been identified as endocrine specific determinants . Neurogenin 3 ( Ngn3 ) is one of the most important endocrine specific transcription factors [20 , 33] . In contrast to the endocrine lineage , little is known about the mechanisms that control the differentiation of exocrine and ductal lineages . Initially , pancreatic transcription factor 1 , alpha subunit ( Ptf1a ) had been considered an exocrine specific transcription factor since its expression becomes restricted to exocrine cells by E13 . 5 in mice [34] . However , cell lineage tracing experiments revealed that ptf1a-expressing cells give rise to all pancreatic cell types [35] . Here , we provide several lines of evidence indicating that exocrine differentiation and proliferation factor ( exdpf ) , a novel gene identified from zebrafish but highly conserved in mouse and human , is an exocrine cell determinant and required regulator of cell proliferation . The protein encoded by exdpf is a putative signaling molecule and is expressed highly in the exocrine cells during pancreas formation in zebrafish . Knocking down exdpf by antisense morpholino caused significant reduction or loss of expression of exocrine markers . In contrast , overexpression of exdpf resulted in the overgrowth in size of the exocrine pancreas and remarkable decrease of endocrine cells . This result suggests that misexpression of exdpf in the endocrine precursors is able to transform their fate . We further show that the reduction of exocrine cells in exdpf morphants is due to lineage-specific cell cycle arrest . Real-time PCR revealed that the expressions of cell cycle inhibitor genes p21Cip and p27Kip are dramatically increased in the exdpf morphants . To test the effect of RA on exocrine cell differentiation , we performed an epistatic study of exdpf and the RA pathway . The results show that exdpf acts genetically downstream of RA . Exogenous RA treatment induced anterior ectopic exocrine cells via exdpf induction . Moreover , injection of exdpf mRNA partially restored exocrine cell differentiation at the endogenous area in embryos treated with exogenous RA , whereas the expansion of endocrine cells were reduced . Our data establish a critical role of exdpf in exocrine cell fate determination and proliferation . Being a vertebrate-specific gene , the exdpf orthologs are highly conserved from fish to human . A search of National Center for Biotechnology Information ( NCBI ) database revealed that the human exdpf ortholog expression is up-regulated in several human cancers including hepatic , pancreatic , and renal cancers , suggesting that overexpression or mutation in the exdpf gene might be involved in the pathogenesis of cancers .
In our effort to identify pancreas specific genes , we isolated the exdpf gene ( GeneID: 338304 [http://www . ncbi . nlm . nih . gov/sites/entrez] ) from a RNA whole-mount in situ hybridization screen in zebrafish ( our unpublished data ) . A BLAST search of the zebrafish genome revealed a homolog of exdpf named endocrine differentiation and proliferation factor ( endpf , not described here ) . The deduced peptide encoded by the exdpf gene contains 117 amino acids . Multiple sequence alignments using ClustalW showed that the Exdpf protein is specific to vertebrates and highly conserved across the vertebrates including zebrafish , mouse , and human ( Figure S1A ) . The human and mouse orthologs are known as uncharacterized novel open reading frames ( c20orf149 , Gene ID: 79144 and AK154758 ) . Overall , the deduced protein is about 42% identical across different species . The N terminus is highly conserved whereas the C terminus is more diversified , which suggests that subtle functional differences might lie in the C terminus . In addition , synteny analysis showed that a cluster of homologous genes is also conserved between zebrafish and human at the exdpf loci ( Figure S1B ) . We studied the temporal and spatial expression of exdpf by reverse transcriptase PCR ( RT-PCR ) and whole-mount in situ hybridization . RT-PCR results ( Figure S2 ) indicated that the exdpf transcript is maternally deposited since it was detected at the one-cell stage . The amount of exdpf transcript reduced following the one-cell stage , and the lowest level was detected at the shield stage . Then exdpf expression gradually increased from shield stage and a strong level was detected between 1 d postfertilization ( dpf ) and 2 dpf when the exocrine pancreas starts to develop; the highest level of expression was detected between 2 dpf and 5 dpf , the longest time point of this study . This result suggests that zygotic expression of the exdpf gene starts at around the shield stage . We then performed RNA whole-mount in situ hybridization using an exdpf probe to obtain detailed expression analysis of exdpf in the developing pancreas . Double in situ hybridization was performed using either a preproinsulin probe to locate the endocrine β cells or a trypsin probe to mark the exocrine cells in combination with the exdpf probe . Exdpf transcripts were first detected in the developing somites at 8 . 5 hpf ( Figure 1A ) . From the three-somite to 21-somite stage , exdpf was expressed in somites , adaxial cell , slow muscle fiber , and epiphysis ( Figure 1B–1E ) . Interestingly , exdpf started to express in the pancreatic area at 33 hpf ( Figure 1F ) , just before exocrine specification begins . Later in development , the strongest domain of exdpf expression appeared in the pancreas . Cells expressing exdpf ( blue staining in Figure 1F ) were located about one somite anterior to the cluster of preproinsulin-positive cells ( red staining in Figure 1F ) . By 36 hpf , exdpf-expressing cells started to contact the preproinsulin-positive cells ( Figure 1G ) as a result of gut rotation . By 2 dpf , exdpf-expressing cells embraced the cluster of preproinsulin-positive cells ( unpublished data ) . These exdpf-expressing cells continue to grow posteriorly to form a typical pancreas-like shape at 3 dpf ( Figure 1H–1J ) . From 33 hpf to 3 dpf , there was no overlap between exdpf-expressing cells and preproinsulin-positive cells , indicating that exdpf expression is excluded from endocrine cells . Conversely , exdpf transcripts completely overlap with trypsin expression at 3 dpf ( Figure 1J ) . To confirm the exocrine-specific expression of exdpf at 4 dpf , we performed a double in situ hybridization using an exdpf probe and a probe against carboxypeptidase A ( cpa ) , another exocrine marker . As expected , exdpf expression completely overlaps with cpa expression at 4 dpf ( unpublished data ) . Together , these data suggest that exdpf is expressed exclusively in the exocrine cells during pancreas development . Based on the expression pattern of exdpf , we postulated that it is required for exocrine pancreas development . To test this hypothesis , we knocked down exdpf by injection of antisense morpholino oligonucleotides ( MO1exdpf and MO2exdpf ) designed to interfere with its translation . We tested both morpholinos to assure that the phenotypes observed are due to the specific knockdown of exdpf . A morpholino standard control oligonucleotide from Gene Tools was used to inject the control embryos . No specific phenotypes were observed in these control embryos . Double in situ hybridization using a trypsin probe and a preproinsulin probe was performed to assess the effect on exocrine and endocrine development simultaneously . In the control embryos , β cells formed a cluster ( islet ) that was surrounded by exocrine cells at the anterior area of the pancreas ( head ) at 3 dpf ( Figure 2A ) . Both the β cell mass and exocrine mass increased at 5 dpf in the control embryos ( Figure 2B ) . In addition , exocrine cells expanded posteriorly to form a typical pancreas like shape ( Figure 2A and 2B ) . The majority of embryos injected with 2 ng of exdpf morpholino ( MO1exdpf ) exhibited no trypsin expression at 3 dpf ( 86% , n = 50 ) or 5 dpf . However , there were a few embryos ( 14% , n = 50 ) with reduced exdpf function that showed severe reduction of trypsin expression at 3 dpf and the remaining exocrine cells were restricted to the anterior pancreatic area engulfing the endocrine cells ( Figure 2C ) . Furthermore , the exocrine cells failed to grow and expand posteriorly by 5 dpf ( Figure 2D ) . In contrast , the endocrine cells looked largely normal in exdpf morphants ( Figure 2C and 2D ) . Only a small fraction of exdpf morphants ( 10% , n = 50 ) showed scattered preproinsulin cells at 5 dpf ( unpublished data ) . These results indicate that exdpf is required specifically for exocrine cell differentiation and growth . Over 90% of MO2exdpf morphants exhibited similar phenotypes ( Figure S3 ) . To investigate the function of exdpf in the differentiated exocrine cells , we used MO1exdpf for the rest of this study because it gave milder phenotypes . We then studied early exocrine differentiation using a ptf1a probe . The protein encoded by ptf1a is a basic helix-loop-helix ( bHLH ) transcription factor that plays a critical role in exocrine pancreas development . A null mutation of Ptf1a in mouse leads to complete agenesis of exocrine pancreas and spatially disorganized endocrine pancreas [34 , 35] . In zebrafish , ptf1a loss of function by antisense morpholino injection suppresses exocrine markers without affecting endocrine markers and the organization of the main islet [36 , 37] . Since ptf1a can serve as an early marker of exocrine development [36] , we analyzed ptf1a expression in exdpf morphants at 2 dpf and 3 dpf by whole-mount in situ hybridization . In the control embryos , ptf1a-expressing cells formed a loose cluster at 2 dpf ( Figure 2E ) and the expression expanded toward the posterior with exclusion from the endocrine cells at 3 dpf ( Figure 2F ) . Expression of ptf1a was missing in the vast majority of exdpf morphant embryos ( 85% , n = 50 ) . In a small fraction of exdpf morphants ( 15% , n = 50 ) , initial expression of ptf1a appeared normal at 2 dpf ( Figure 2G ) . But the expression of ptf1a failed to expand towards the posterior by 3 dpf ( Figure 2H ) and remained in the same area as in 2 dpf . This result confirms that exdpf is critical for exocrine cell specification and expansion . The exdpf gene is expressed in the developing somites as well as the exocrine pancreas ( Figure 1 ) . To clarify whether the exocrine pancreas defect is due to pleiotropic abnormalities , we used a recently obtained transgenic fish MP760GFP ( our unpublished data ) that expresses green fluorescent protein ( GFP ) in the developing liver , gut , and pancreas . In this transgenic line , pancreatic expression of GFP is restricted in the exocrine portion . A minimal amount of exdpf morpholino was used to achieve the least amount of defects in overall body morphology . At 24 hpf , strong expression of GFP was observed in the presumed pancreatic area in both control embryos and exdpf mRNA injected embryos ( Figure S5A and S5B , arrows ) . By 2 dpf , pancreatic GFP expression became more obvious in the control and exdpf mRNA-injected embryos ( Figure S5A and S5B , arrowheads ) . In contrast , only residual or no pancreatic GFP expression was observed in exdpf morphants ( Figure S5C ) . However , gut GFP expression in exdpf morphants ( Figure S5C ) remained comparable to that in the control embryos ( Figure S5A ) . Exdpf is also expressed in the developing liver ( Figure 1 ) during embryogenesis . To assess whether it is required for liver development , in situ hybridization was performed using a ceruloplasmin ( cp ) probe . At 3 dpf , clear expression of cp was detected in the livers of control embryos injected with standard morpholino control ( Figure S5D ) . No detectable change of cp expression was observed in exdpf morphants ( Figure S5E–S5G ) . This might be due to the redundant function of exdpf homolog endpf since it is also expressed in the developing liver . We further studied the genetic interaction between ptf1a and exdpf . In mild exdpf morphants ( injected with 2 ng of MO1exdpf ) , ptf1a expression was initiated but restricted to the anterior area in 72% of embryos ( Figure 2G , n = 90 ) , indicating that exocrine cell differentiation can start but expansion fails . However , only 15% of embryos ( n = 50 ) still exhibited ptf1a expression when injected with 4 ng of MO1exdpf , suggesting a role for exdpf in exocrine cell differentiation . Knocking down ptf1a by morpholino injection ( Figure 3 ) resulted in agenesis of the exocrine pancreas ( Figure 3C ) ; only about 5% of ptf1a morphants exhibited weak cpa expression ( n = 189 , Table 1 ) . In the control embryos , injection of exdpf mRNA led to a great expansion of the exocrine pancreas ( Figure 3B ) . Interestingly , injection of exdpf mRNA into ptf1a morphants successfully restored expression of exocrine marker cpa to about 70% of embryos ( Figure 3D–3F , n = 168 ) . About 35% of embryos exhibited nearly full restoration of cpa expression in the ptf1a morphants when exdpf mRNA was injected , whereas another 35% exhibited partial restoration ( Table 1 ) . In a reciprocal experiment , injection of ptf1a mRNA into exdpf morphants failed to rescue expression of exocrine markers ( unpublished data ) . Together , these results place exdpf genetically downstream of ptf1a in exocrine development . To test whether ptf1a functions through exdpf in exocrine specification , we performed in situ hybridization analysis of exdpf in ptf1a morphants ( Figure 4 ) . Indeed , exdpf expression in the exocrine pancreas was abolished in such embryos , as expected ( Figure 4B ) . However , the epiphysis expression of exdpf remained unchanged in the ptf1a morphants ( Figure 4B , inset ) , indicating that ptf1a specifically controls pancreatic expression of exdpf . Ptf1 is an unusual heterotrimeric bHLH transcription factor composed of Ptf1a/P48 , a common class A bHLH protein ( such as HEB , E2–2 , E12 , or E47 , also called E-proteins ) , and a third protein that can be either the mammalian Suppressor of Hairless RBP-J or its paralog , RBP-L [38 , 39] . Ptf1 binding sites are bipartite with an E-box ( CANNTG ) and a TC-box ( TTTCCC ) spaced one or two helical turns apart , center to center [39–41] . The heterodimeric subcomplex of Ptf1a and the E-protein binds to the E-box and RBP-J or RBP-L binds to the TC-box [39 , 41] . Binding of the Ptf1 complex to DNA requires both boxes , and the spacing between these elements is critical for Ptf1 binding [42] . Functional binding sites for the Ptf1 complex are present in the 5′ promoter regions of all of the acinar digestive enzyme genes examined [40 , 41] . To determine whether Ptf1a might indeed control the transcription of exdpf , we searched the 5-kb 5′ flanking region and intronic sequences of this gene for potential PTF1-binding sites comprising an E-box and a TC-box spaced one or two helical DNA turns apart . Three potential binding sites were detected . Binding site 1 is about 3 kb upstream of the transcriptional start site ( Figure 4C ) and binding site 2 is around 1 kb upstream of the transcriptional start site ( Figure 4C ) ; binding site 3 is about 500 bp downstream of the transcriptional start site in the first intron ( Figure 4C ) . A 3 . 6-kb exdpf promoter region containing the transcriptional start and three potential PTF1-binding sites increased the activity of the luciferase reporter plasmid 41-fold compared to the promoterless pGL3-Basic in HEK 293 cell lines tested by transfection ( Figure 4D , seq1 ) . While deletion of binding site 1 reduced the activity to 29-fold , a 1 . 5 kb promoter region that retained binding site 2 and binding site 3 still maintained an activity of 26-fold ( Figure 4D , seq 2 and seq 3 ) . However , deletion of binding site 1 and binding site 3 reduced the activity to 8-fold ( Figure 4D , seq4 ) . Not surprisingly , deletion of all three binding sites further reduced the transcription of the reporter gene almost to a basal level ( Figure 4D , seq 5 ) . These results provide strong evidence that Ptf1a can promote the transcription of exdpf , and Ptf1-binding site 1 and binding site 3 are especially critical for the activation of the promoter . Taken together , these data demonstrate that exdpf is a direct target gene of Ptf1a . To assess whether exdpf is sufficient for exocrine specification , we carried out overexpression experiments by injecting synthetic exdpf mRNA into embryos at the one-cell stage ( Figure 5 ) . Elastase A:GFP transgenic fish were used to facilitate the identification of exocrine cells . In this fish , GFP expression is controlled by the elastase A ( elaA ) regulatory sequence which allows exocrine specific GFP expression in larvae and adult [43] . At 3 dpf , the average exocrine cell number in the control embryo is 197 . 8 ± 8 . 2 ( Figure 5C; Table 2 , control , mean ± standard deviation [SD] , n = 5 ) . Overexpression of exdpf caused a mild increase of exocrine mass by about 28% ( Figure 5C , Table 2 , 253 ± 4 . 2 ) . Conversely , knocking down exdpf by morpholino remarkably reduced exocrine cell number to about 10% of control level ( 21 . 6 ± 3 . 4 ) at 3 dpf ( Figure 5C , Table 2 ) . At 5 dpf , exocrine cell number was further increased by 42 . 8% in exdpf mRNA injected embryos ( Figure S6D , 376 . 2 ± 6 . 3 cells per embryo ) compared with that of the control embryo ( Figure S6A , 263 . 4 ± 4 . 4 cells per embryo , n = 5 ) . Interestingly , exocrine cell number in the exdpf morphants did not change much from 3 dpf to 5 dpf ( from 21 . 6 ± 3 . 4 to 22 . 8 ± 3 . 8 ) , indicating that the proliferation rate of the exocrine cells is affected in morphants . Knocking down exdpf gene by antisense morpholino leads to loss or minimal formation of the exocrine pancreas . The strong reduction of the exocrine pancreas could be a result of increased cell death or decreased cell proliferation rate or a combination of both . To address these possibilities , we performed BrdU incorporation experiments to assess cell proliferation . We chose the transgenic line MP760GFP to assess exocrine pancreatic progenitor cell proliferation since GFP expression is observed in the pancreatic area at 33 hpf , 1 h before exocrine cell differentiation ( Song J , unpublished data ) . At this stage , about 80% of exocrine pancreatic progenitors were proliferating in the control embryos injected with control morpholino ( Figure 5A [top] and 5C , Table 2 ) . Injection of exdpf mRNA increased GFP positive cell number by about 20% and 88% of the cells were proliferating ( Figure 5A [center] and 5C , Table 2 ) . In contrast , knocking down exdpf by morpholino caused a severe reduction of the exocrine progenitor cell number by 70% ( Figure 5C , Table 2 ) and only 31% of the cells were proliferating ( Figure 5A [bottom] and 5C , Table 2 ) . These results indicate that exdpf is required for exocrine progenitor specification and proliferation . For differentiated exocrine cells , elastase A:GFP transgenic fish was used to facilitate the identification of exocrine cells . We chose 3 dpf embryos to compare cell proliferating ability because exocrine cells are still undergoing massive proliferation at this stage . At 3 dpf , in the control embryos , the BrdU incorporation rate was about 85% ( Figure 5B [top] and 5C , Table 2 ) . Injection of exdpf mRNA increased the BrdU incorporation rate to about 98% ( Figure 5B [center] and 5C , Table 2 ) . This result suggests that exdpf is sufficient to promote exocrine cell proliferation . In contrast , knocking down exdpf by morpholino significantly reduced the BrdU incorporation rate to about 17% ( Figure 5B [bottom] and 5C , Table 2 ) , indicating that exdpf is also necessary for proliferation of differentiated exocrine cells . To assess whether cell death could contribute to the severe reduction of exocrine cells in the exdpf morphants , we performed TUNEL assay on 5 dpf embryos because exocrine cell number seems unchanged much from 3 dpf to 5 dpf . In the control embryos , a small fraction of exocrine cells were undergoing apoptosis ( Figure S6C [enlargement] and S6J , 6 . 65% ± 2 . 66% , n = 5 ) . Overexpression of exdpf by injecting synthetic mRNA increased exocrine cell number by ∼40% . Only basal level of cell death was detected in these embryos ( Figure S6F and S6J , 6 . 03% ± 2 . 43% , n = 5 ) . Similarly , knocking down exdpf did not significantly increase the ratio of cells undergoing apoptosis ( Figure S6I enlargement and S6J , 7 . 38% ± 3 . 24% , n = 5 ) . This result indicates that reduced exdpf function did not cause massive cell death of the exocrine pancreas . To address the molecular mechanisms causing cell cycle arrest in the exdpf morphants , semi quantitative RT-PCR was performed to examine the expression levels of cell cycle inhibitors cyclin G1 , p21Cip , and p27Kip , as well as G1 to S phase regulator cyclin D1 , at different developmental stages , including 33 hpf ( just before exocrine cell differentiation ) , 2 dpf ( exocrine cells surround the endocrine islet ) , 3 dpf ( exocrine cells expand posteriorly ) , and 5 dpf ( exocrine cells form nice pancreatic shape with head , trunk and tail ) . At 33 hpf , no discernible difference of either p21Cip or p27Kip or cyclin G1 expression was detected in control , exdpf morphants , and exdpf mRNA injected embryos ( Figure 5D ) . From 2 dpf to 5 dpf , a dramatic increase of p21Cip expression ( Figure 5D ) was observed in the exdpf morphants , whereas the expression level remained comparable in the control and exdpf mRNA injected embryos . Similarly , a significant increase of p27Kip expression was also observed in the exdpf morphants , although not as much as p21Cip expression ( Figure 5D , second row ) . In addition , the expression of cyclin G1 also slowly increased in the exdpf morphants from 2 dpf to 5 dpf . Taken together , our results suggest that knocking down exdpf leads to cell cycle arrests through up-regulation of p21Cip , p27Kip , and cyclin G1 . We have shown in our previous result that overexpression of exdpf increased exocrine cell number due to overproliferation of these cells ( Figure 5B , [center] ) . We then confirmed this result by checking the expression level of cyclin D1 through semiquantitative RT-PCR . At 33 hpf ( just before exocrine cell differentiation ) , cyclin D1 expression was slightly increased in the exdpf mRNA injected embryos ( Figure 5D ) . From 2 dpf to 5 dpf , the increase of cyclin D1 expression in the exdpf mRNA injected embryos became stronger . Interestingly , we also detected elevated expression of cyclin D1 in the exdpf morphants from 2 dpf to 5 dpf . We postulate that this type of increase may be caused by exdpf independent mechanisms that control organ size . Due to the strong increase of cell cycle inhibitors in exdpf morphants , an increased level of cyclin D1 is not enough to drive the cell into the proliferating phase . The semi-quantitative RT-PCR results are further confirmed by real-time PCR ( Table 3 ) . Similar up-regulation of cell cycle inhibitors including p21 , p27 , and cyclin G1 were detected in the exdpf morphants ( Table 3 ) . To examine where the increase of p21 expression comes from , we performed in situ hybridization using a p21 probe . At 33 hpf , p21 expression was detected in the developing brain ( Figure S7A and enlargement ) . No p21 expression was observed in the developing pancreatic area . A similar p21 expression pattern was observed in expdf mRNA injected embryos ( Figure S7B and enlargement ) . Conversely , noticeable p21 expression could be seen in the developing pancreatic area in exdpf morphants ( Figure S7C and enlargement ) . Thus , cell proliferation defects observed in the developing exocrine pancreas caused by knocking down exdpf are likely mediated by increased level of p21 expression in the cells . We have shown that exdpf is essential for the exocrine pancreas differentiation and expansion; overexpression of exdpf gene leads to increased exocrine size . This effect is similar to that of RA treatment ( Figure 6 ) . Embryos treated with RA exhibited expanded endocrine pancreas [25] . Exogenous RA treatment also caused ectopic formation of exocrine at positions anterior to the presumptive pancreatic area ( Figure 6B ) . But the exocrine pancreas failed to expand toward the posterior at 5 dpf ( unpublished data; only a small number of embryos survived ) as it did in untreated control embryos . Interestingly , RA treatment of exdpf morphants failed to induce exocrine formation ( Figure 6D ) . This result indicates that exocrine pancreas formation in RA-treated embryos requires exdpf function . In reciprocal experiments , treatment with 10−6 M diethylaminobenzaldehyde ( DEAB ) to block the RA pathway often resulted in no exocrine pancreas formation at 3dpf ( Figure 6F , 85% , n = 181 embryos ) . Only about 15% ( n = 181 ) of DEAB treated embryos exhibited weak expression of an exocrine marker at 3 dpf . This result suggests that RA is required for early exocrine formation during normal pancreas development . To test whether overexpression of exdpf can rescue exocrine pancreas formation in the DEAB treated embryos , we injected synthetic exdpf mRNA into embryos at the 1-cell stage followed by treatment with 10−6 M of DEAB from 9 hpf up to the time of analysis . At 3 dpf , exdpf mRNA injection increased the percentage of embryos exhibiting expression of the exocrine marker cpa to 35% ( Figure 6G and 6I , n = 166 ) . This result suggests that exdpf acts downstream of the RA pathway during normal pancreas development . RA treatment often resulted in bilateral ectopic exocrine formation at the anterior area ( Figure 6N , 83% , n = 191 embryos ) . To test whether exogenous RA treatment induces exdpf expression , in situ hybridization using an exdpf probe was performed . Interestingly , similar bilateral ectopic expression of exdpf was observed in RA treated embryos ( Figure 6K ) . On the other hand , blocking RA synthesis by DEAB completely abolished exdpf expression in the pancreatic area ( Figure 6L ) , whereas the epiphysis expression appeared normal ( Figure 6L , arrowhead ) . These data suggest RA signaling influences exdpf expression in the developing pancreas . Taken together , these results place exdpf genetically downstream of RA in exocrine pancreas development . The expression of the exdpf gene is excluded from the endocrine islet during pancreatic development ( Figure 1I ) . To evaluate the effect of exdpf overexpression on endocrine cell differentiation , we performed in situ hybridization against preproinsulin in 3 dpf embryos ( Figure 7 ) . In the exdpf mRNA injected embryos , preproinsulin expression was dramatically reduced ( Figure 7B ) . Preproinsulin:GFP transgenic fish were injected with exdpf mRNA to quantify the GFP-positive cells . In the control embryos , the average number of GFP positive cells was 45 . 15 ± 7 . 04 ( Figure 7F , mean ± SD , n = 20 ) . As expected , overexpression of exdpf resulted in a significant reduction of GFP positive cell number by about 40%; the average number of GFP-expressing cell was 27 . 55 ± 6 . 02 ( Figure 7F , n = 20 ) . The stage of 3 dpf is relatively late for β cell development . To examine whether the reduction of β cell number at 3 dpf is due to defects in cell proliferation or specification , we quantified preproinsulin:GFP positive cells at 24 hpf ( see Figure S8 for panel of embryos used for quantification ) when the majority of β cells come from newly specified cells . Overexpression of exdpf reduced β cell number by about 43% ( Figure 7E , from 21 . 4 ± 4 . 16 to 9 . 2 ± 4 . 04 cells per embryo ) . However , no significant change in β cell number was observed in the exdpf morphants ( Figure 7E , 19 . 5 ± 3 . 39 cells per embryo ) . These data indicate that overexpression of exdpf inhibits β cell specification and suggests a possible transformation of cell fate in the endocrine pancreatic precursors . Exogenous RA treatment of WT embryos dramatically increased preproinsulin-expressing cell number ( Figure 7C and 7F , 212 . 15 ± 14 . 88 cells per embryo , n = 20 ) compared with that in the DMSO treated control embryos ( Figure 7A and 7F , 45 . 15 ± 7 . 04 , n = 20 ) . Interestingly , overexpression of exdpf in the RA treated embryos inhibited the anterior expansion of the endocrine pancreas ( Figure 7D and 7F , 105 . 2 ± 10 . 88 cells per embryo , n = 20 ) . This result suggests that overexpression of exdpf in the anterior ectopic pancreas induced by RA treatment balanced RA signaling and turned more cells into the exocrine fate .
Our results provide strong evidence that a novel gene exdpf is specifically employed by the zebrafish exocrine progenitors to promote cell differentiation and proliferation . Moreover , epistasis experiments support that exdpf functions downstream of ptf1a in exocrine cell specification . RA also interacts genetically with Exdpf during exocrine formation . How does Exdpf regulate exocrine cell specification ? First , exdpf starts to express in exocrine progenitors at 33 hpf , just before exocrine cell differentiation . Second , exdpf is required for exocrine cell differentiation . Knocking down exdpf by antisense morpholino resulted in the absence of exocrine markers in a majority of injected embryos . The remaining smaller fraction of exdpf morphants exhibited significantly reduced expression of exocrine markers . Third , exdpf is both necessary and sufficient for exocrine cell proliferation . We observed exdpf overexpression increased the proliferation of exocrine cells ( 98% versus 85% in the control embryos ) , whereas knocking down exdpf severely impaired proliferating ability due to the increased level of cell cycle inhibitors p21Cip , p27Kip , and cyclin G1 . Although exdpf is necessary for exocrine cell differentiation , it is not sufficient to induce ectopic formation of the exocrine pancreas . A possible explanation is that exdpf only functions in the committed exocrine progenitors . Our result shows that Exdpf promotes exocrine cell proliferation by regulating cyclin D expression . Exdpf acts genetically downstream of ptf1a in exocrine specification . It is likely that ptf1a controls the expression of exdpf in exocrine progenitors . The expression pattern of exdpf in exocrine pancreas is similar to that of ptf1a [36] . ptf1a is expressed in the exocrine progenitors at 32 hpf . Similarly , ptf1a-expressing cells also surround β cells as a result of gut rotation [36] . In addition , three Ptf1a binding sites have been identified in the promoter region of the exdpf gene ( Figure 4C ) . Using luciferase assay , we were able to show that these Ptf1a binding sites are functional in culture cells . Thus , exdpf is a direct target gene of Ptf1a . Other transcription factors such as Pdx1 might also control exocrine progenitor specific expression of exdpf . During embryogenesis , ptf1a is expressed in the ventral aspect of pdx1-positive domain from 32 hpf to 36 hpf [36] . Moreover , two Pdx1 binding sites have also been identified within exdpf promoter region ( data not shown ) . Knocking down of ptf1a leads to the agenesis of the exocrine pancreas , which is consistent with previous results [36 , 37] . Injection of exdpf mRNA into the ptf1a morphants successfully restored the expression of the exocrine marker . This is likely due to the expansion and differentiation of residual progenitor cells promoted by exdpf mRNA in the Ptf1a morphants . In our rescue experiments , the injected concentration of ptf1a morpholino only created a hypomorph situation and should still contain a residual pool of progenitors . Co-injection of exdpf mRNA promotes proliferation and differentiation of these cells , resulting rescue of the defect . It is also likely that Ptf1a activity recovers following cessation of the morpholino effect , and that this allows for full exocrine differentiation to occur in progenitor cells that have been rescued by exogenous exdpf . Nonetheless , our results provide evidence that Exdpf acts downstream of ptf1a in exocrine formation . In addition , overexpression of exdpf by mRNA injection inhibited endocrine cell fates ( Figure 7B , 7E , and 7F ) . This endocrine repression result is in agreement with a recent report by Dong et al . Using partial loss of function analysis for ptf1a , Dong et al . found that high levels of ptf1a promote exocrine fate whereas low levels promote endocrine fate [44] . It is likely that ptf1a exerts its function through exdpf in this aspect . The exdpf gene encodes a putative signaling molecule containing two SH2 and two SH3 domains as well as other conserved domains ( Figure S4 ) , suggesting that Exdpf may function in response to signals from adjacent mesoderm tissues . Multiple intercellular signals including transforming growth factor beta , Hedgehog , and Notch are critical for the proper specification of endocrine and exocrine cell fates during pancreas development . It is not clear which signal or signals exdpf responds to in order to make the exocrine cell fate decision . Our results support the idea that exdpf promotes exocrine cell proliferation . As a putative signaling molecule , exdpf might be involved in transducing signals from growth factors that are required for exocrine cell proliferation . However , the identities of such growth factors remain elusive . Genetic evidence places exdpf downstream of RA . Exogenous treatment with RA caused anterior expansion of endocrine cells as well as ectopic anterior exocrine cells , which is consistent with previous results [25] . Contradictory results from RA treatment have been reported using different organisms regarding exocrine development . In mouse embryonic culture , RA treatment suppresses exocrine differentiation and branching morphogenesis [30 , 45] . Others reported that atRA treatment leads to endocrine and duct differentiation from the pancreatic bud , but inhibits exocrine differentiation [45] . In Xenopus , exogenous RA treatment causes endocrine expansion in the dorsal bud at the expense of exocrine tissue but stimulates exocrine differentiation in the ventral bud [26] . These contradictory results might derive from different organisms and different concentrations of either atRA or 9cis RA . In our experiment , we find that excessive atRA caused ectopic anterior exocrine formation . Exogenous treatment with RA induced ectopic pancreatic cells including endocrine and exocrine cells . Anterior ectopic formation of the exocrine pancreas by RA treatment requires exdpf function since exdpf morpholino injection blocks ectopic exocrine formation and RA also induces anterior ectopic expression of exdpf . Overexpression of exdpf in wild-type embryos significantly suppresses endocrine cell differentiation , suggesting exdpf transforms the cell fate of pancreatic progenitors ( Figure 7G ) . The balance of RA and overexpressed exdpf in the progenitors results in reduced endocrine cells compared with RA treated wild-type embryos . Pancreatic cancer is one of the leading causes of cancer deaths because it is often highly aggressive and resistant to treatments available at the time of diagnosis [46] . Genetic studies have identified structural mutations in pancreatic cancers; the alterations include the activation of K-Ras proto-oncogene as well as inactivation of tumor suppressor genes such as TP53 or INK4a locus [47–49] . By carefully searching the NCBI database , we found that the human exdpf ortholog is expressed in relatively high levels in multiple tissues including the pancreas , colon , and mammary glands . Interestingly , the EST expression profile also indicates that higher level of exdpf ortholog has been detected in several cancers including pancreatic cancer , breast cancer and kidney cancer ( Figure S9 ) . In addition to structural mutations , many growth factor receptors and their ligands are overexpressed in pancreatic cancers . Since exdpf can promote acinar cell proliferation , it is worth studying the role of this gene in pancreatic cancer . Zebrafish has proven to be a useful model system to study pancreatic cancers . In 2004 , Yang et al . reported that human MYCN caused pancreatic neuroendocrine tumors in transgenic zebrafish that expressed MYCN in β cells , muscle cells and neurons [50] . Recently , transgenic fish that express oncogenic KRASG12V under the ptf1a promoter was generated . In these fish , KRASG12V blocked the differentiation of pancreatic progenitor cells and this undifferentiated progenitor pool lead to invasive pancreatic cancer [51] . These examples demonstrate that the zebrafish model is useful in advancing our understanding of pancreatic cancers . Together , our results reveal a specific requirement for Exdpf in exocrine pancreas formation in zebrafish . The gene exdpf is expressed exclusively in the exocrine progenitors and differentiated exocrine cells . We demonstrate that exdpf is necessary for exocrine cell differentiation . Furthermore , exdpf is both sufficient and necessary for the proliferation of differentiated exocrine cells . We speculate that the study of the function of exdpf in pancreatic cancers could shed light on the pathogenesis of this malignancy .
Zebrafish were raised and kept under standard laboratory conditions at about 28 °C . Embryos were staged according to Kimmel et al . [52] . The elastase A:GFP fish was a gift from Gong's laboratory in Singapore [43] . The wild-type line used was AB . In situ hybridization was performed essentially as previously described [53] . For double in situ hybridization , Fast Red ( Roche ) and NBT/BCIP ( 50 mg/ml; Promega ) were used as alkaline phosphatase substrates . The following probes were used: preproinsulin [54] , trypsin , cpa , ptf1a , p21 , and exdpf . For RA treatment , wild-type zebrafish embryos were incubated with retinoic acid as described [25] . For DEAB treatment , embryos were incubated with 10−6 M DEAB diluted from 10−2 M stock solution in DMSO as described [55] . mRNAs of exdpf and ptf1a were synthesized using T7 or SP6 mMessage mMachine kit ( Ambion ) . mRNA injection was performed as described at the one-cell stage [56] . Sterile water was used for the control experiments . 100 pg of exdpf or ptf1a mRNA was used for all experiments . Antisense morpholino oligos ( MOs ) designed against exdpf: MO1: 5′-GCTGGATGGAATTGCTGCCATTTTC-3′ . MO2: 5′-TCGACCGTGTGGAAGATGGAAAGAT-3′ were obtained from Gene Tools . Ptf1a-MOa , 5′-AGTGTCCATTTTTTGTGCTGTGTTG-3′ was obtained from Open Biosystems and injected into one- to two-cell stage embryos ( for original reference , see [36] ) . A morpholino standard control oligo ( 5′-CCTCTTACCTCAGTTACAATTTATA-3′ ) used as control was obtained from Gene Tools . All morpholinos were prepared and resuspended in 1× Danieau's buffer at 1 ng/nl for injection . Embryos were placed in 10 mM solution of BrdU ( Sigma ) in fish water at 24 hpf and kept in dark at 28 °C for 48 h . Then embryos were fixed in 4% PFA for 2 h at room temperature . After washing ( 5 × 5 min in PBST ) , embryos were incubated in 1 N HCl for 1 h . Then embryos were washed again for 5 × 5 min in PBST and blocked using 5% goat serum for 1 h at room temperature and incubated over night at 4 °C in 1:100 mouse anti-BrdU monoclonal antibody ( Chemicon ) and 1:500 rabbit anti-GFP polyclonal antibody ( Abcam ) . After washing ( 5 × 5 min in PBST ) , embryos were incubated overnight at 4 °C in 1:200 A488 conjugated goat anti-rabbit antibody ( Molecular Probe ) . Embryos were then washed and incubated in 1:200 TRITC conjugated rabbit anti-mouse antibody ( Sigma ) over night at 4 °C . The stained embryos were incubated with DAPI for 30 min and visualized via a Zeiss LSM 510 Laser Confocal Microscope . The terminal deoxynucleotidyl transferase–mediated dUTP-fluoroscein nick-end labeling ( TUNEL ) assay was performed essentially as described previously [57] . Embryos were fixed overnight in 4% PFA in PBS at 4 °C and permeabilized using methanol . Then embryos were rehydrated and washed 5 × 5 min in PBST at room temperature . TUNEL labeling was performed by 1 h incubation at 37 °C in a cell death detection reagent ( in situ cell death Detection Kit-TMR Red , Roche Diagnostics ) . As negative controls , embryos were incubated in the TUNEL label only . For positive controls , embryos were treated with DNaseI for 1 h at 37 °C before TUNEL labeling . After the reaction , embryos were washed 4 × 15 min in PBST at room temperature and stored in PBST at 4 °C in a covered canister . Fluorescence was detected using a Zeiss LSM 510 Laser Scanning Confocal Microscope . Images were captured using a digital camera ( Axiocam ) attached to a compound microscope ( Zeiss: Axioplan 2 or Imager A1 ) and Improvision ( an Openlab program ) or Axiovision . Confocal images were captured using a Zeiss LSM 510 Laser Scanning Confocal Microscope . To quantify elastase A:GFP-positive cell numbers , yolk was removed from the embryo and the rest of the embryo was flattened by cover slide . GFP-positive cells were used to outline exocrine and DAPI staining inside the exocrine area was counted under a compound microscope . Calculations were performed using Microsoft Excel . We report mean and standard deviation of exocrine cell numbers . The probability associated with the Student's t-test ( with two-tailed distribution ) and two samples of unequal variance were also included . RT-PCR was performed for transcript analysis of p21Cip , p27Kip , cyclin D1 , cyclin G1 , and EF1α . Total RNA was extracted from 30 embryos of each group at various developmental stages using Qiagen Rneasy Mini kit ( Qiagen , Valencia , CA ) . The total RNA pellet was resuspended in 30 μl of RNase free water and stored at −80 °C . For reverse transcription , 10 μl of RNA was mixed with 1 μl of random primer and incubated for 5 min at 68 °C and placed on ice . For each reaction , a mix of 4 μl of 5× RT buffer ( 50 mM Tris-HCl , 75 mM KCl , 3 mM MgCl2; Invitrogen , Carlsbad , CA ) , 2 μl of dNTPmix ( 10 mM ) , 2 μl of DTT ( 100 mM ) , and 1 μl of M-MLV reverse transcriptase ( 200 U; Invitrogen , Carlsbad , CA , Canada ) was added and samples were incubated for 1 h at 42 °C followed by 10 min at 68 °C . To test for genomic DNA contamination , equivalent RNA samples were treated in the same manner , except that RNase-free water was added to the reaction instead of the reverse transcriptase . RT-PCR primer sequences and reaction conditions are presented in Table S1 . Real-time PCR was performed using Bio-Rad CYBR Green Supermix on iCycler . For each time point , triplicate was used to determine the ratio of experimental group versus control . Standard curve method was used . p-Values were obtained using the Student's t-test ( with two-tailed distribution ) and two samples of unequal variance . Exdpf promoter sequences were generated by the polymerase chain reaction ( PCR ) with Long Taq PCR Mastermix ( TIANGEN ) from genomic DNA of AB zebrafish . The fragments of the exdpf promoter obtained were inserted into the pGL3-Basic vector ( Promega ) . Zebrafish Ptf1a , human E47 and RBP-J expression constructs were kindly provided by Steven D . Leach [58] , Michael Chin [59] , and S . Diane Hayward [60] , respectively . Human embryonic kidney 293 cells were plated at 70% confluence in 24-well plates and transfected with 0 . 8 μg of DNA per well using Lipofectamine 2000 ( Invitrogen ) following the manufacturer's protocol . A Renilla luciferase control plasmid ( Promega Dual-Luciferase Reporter Assay System ) was included in all transfections to allow normalization for transfection efficiency . Total DNA content per well was made consistent by supplementing with pCDNA3 . 1 ( + ) vector when necessary . Ptf1-activated luciferase activity was defined as the ratio of firefly to Renilla luciferase luminescence and is further normalized against luciferase activity from cells transfected with the pGL3-Basic construct . All values are the means of at least three transfections ± standard error of the mean . | The pancreas is a vital organ comprising endocrine and exocrine components . Both endocrine and exocrine cells derive from a common pool of progenitors present in the gut endoderm during embryogenesis . The molecular mechanisms regulating cell fate decisions and lineage-specific proliferation are not fully understood . In this work , we report the characterization of a novel gene , exocrine differentiation and proliferation factor ( exdpf ) , as a regulator for exocrine cell fate and differentiation/proliferation . We show that it is a direct target of the transcription factor pancreas-specific transcription factor 1a ( Ptf1a ) , which is expressed in progenitors that give rise to all pancreatic cell types . We find that a deficiency of exdpf results in a severe reduction of exocrine size due to defects in cell proliferation . Consistent with this finding , overexpression of exdpf leads to an increase of exocrine size and a decrease of endocrine size , suggesting a possible change in fate of the endocrine progenitors . The human ortholog of exdpf is highly conserved and its expression level appears elevated in several cancers , including hepatic and pancreatic cancers , implying a possible role in pathogenesis of these malignancies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology"
] | 2008 | Exdpf Is a Key Regulator of Exocrine Pancreas Development Controlled by Retinoic Acid and ptf1a in Zebrafish |
Oscillatory activity robustly correlates with task demands during many cognitive tasks . However , not only are the network mechanisms underlying the generation of these rhythms poorly understood , but it is also still unknown to what extent they may play a functional role , as opposed to being a mere epiphenomenon . Here we study the mechanisms underlying the influence of oscillatory drive on network dynamics related to cognitive processing in simple working memory ( WM ) , and memory recall tasks . Specifically , we investigate how the frequency of oscillatory input interacts with the intrinsic dynamics in networks of recurrently coupled spiking neurons to cause changes of state: the neuronal correlates of the corresponding cognitive process . We find that slow oscillations , in the delta and theta band , are effective in activating network states associated with memory recall . On the other hand , faster oscillations , in the beta range , can serve to clear memory states by resonantly driving transient bouts of spike synchrony which destabilize the activity . We leverage a recently derived set of exact mean-field equations for networks of quadratic integrate-and-fire neurons to systematically study the bifurcation structure in the periodically forced spiking network . Interestingly , we find that the oscillatory signals which are most effective in allowing flexible switching between network states are not smooth , pure sinusoids , but rather burst-like , with a sharp onset . We show that such periodic bursts themselves readily arise spontaneously in networks of excitatory and inhibitory neurons , and that the burst frequency can be tuned via changes in tonic drive . Finally , we show that oscillations in the gamma range can actually stabilize WM states which otherwise would not persist .
Oscillations are ubiquitous in neuronal systems and span temporal scales over several orders of magnitude [1] . Some prominent rhythms , such as occipital alpha waves during eye-closure [2] or slow-oscillations during non-REM sleep [3] are indicative of a particular behavioral state . Other rhythms have been specifically shown to correlate with memory demands during working memory tasks , including theta ( 4–8Hz ) [4–7] , alpha/beta ( 8–30Hz ) [8–10] and gamma ( 20–100Hz ) [11–13] . Understanding the physiological origin and functional role of such oscillations is an area of active research . Here we study how oscillatory signals in distinct frequency bands can serve to robustly and flexibly switch between different dynamical states in cortical circuit models of working memory and memory storage and recall . In doing so we characterize the dynamical mechanisms responsible for some of the computational findings in an earlier study [14]; we go beyond that work to include new results on oscillatory control of network states . Specifically , we consider the response of multistable networks of recurrently coupled spiking neurons to external oscillatory drive . We make use of recent theoretical advances in mean-field theory to reduce the spiking networks to a low-dimensional macroscopic description in terms of mean firing rate and membrane potential , which is exact in the limit of large networks [15] . This allows us to perform a systematic and detailed exploration of network states analytically or with numerical bifurcation analysis , which informs us about suitable parameter sets for numerical simulations . The latter serve to give representative examples of the dynamical phenomena investigated here . As a result , we can completely characterize the dynamics of the forced system . Specifically , we consider networks which exhibit multistability in the absence of forcing . Such attracting network states have been proposed as the neural correlate of memory recall [16 , 17] , and as a possible mechanism for sustaining neuronal activity during working memory tasks [18–20] . We find that an external oscillatory drive interacts with such multistable networks in highly nontrivial ways . Low-frequency oscillations are effective in switching on states of elevated activity in simple bistable networks , while in higher dimensional multistable networks they allow for robust switching between stored memory states . Higher frequencies , in the beta range , destabilize WM states through a resonant interaction which recruits spike synchrony . Such oscillatory signals can therefore be used to clear memory buffers . Finally , when networks operate outside the region of multistability , e . g . due to reduced excitability , an oscillatory signal in the gamma range can be used to recover robust memory recall .
Networks of recurrently coupled excitatory neurons can exhibit bistability given sufficiently strong synaptic weights . Such networks act as binary switches: a transient input can cause a transition from a baseline state to a state of elevated activity , or vice-versa . We asked to what extent an oscillatory signal alone could also drive transitions between states in such a network . In particular we were interested in knowing if the directionality of the transition , and hence the final state of the system , could be controlled via the frequency of the oscillatory drive . To investigate this we simulated a network of recurrently coupled excitatory quadratic integrate-and-fire neurons , see Methods for details . Fig 1 shows an illustration of the network dynamics as a function of the stimulus frequency and initial state of the network . In particular , at low frequencies , the oscillations push the system from the state of low activity into the state of high activity , which persists under such forcing , see Fig 1A . As the frequency is increased past a critical value , it is no longer effective in driving a transition , and the network remains in its initial state , see Fig 1B . A further increase then shows the opposite effect: The state of high activity becomes unstable under the forcing , whereas the state of low activity persists , Fig 1C . At large enough frequencies we then observe again that no transitions occur and the initial network state persists , Fig 1D . The results from Fig 1 show that the frequency of an external oscillatory drive can be used to selectively destabilize a given network state . For the parameter values used here , oscillation frequencies in the delta range result in a WM state while frequencies in the beta range force the system to the “ground” state , essentially clearing the WM state , a result seen also in [14] . Oscillations outside these ranges are ineffective in driving transitions . We seek to understand the mechanisms underlying these transitions , and additionally to determine to what extent the precise frequency ranges are influenced by the network parameters . To do this we will take advantage of recent work in which the authors derived a set of simple equations for the mean firing rate and mean membrane potential in a network of recurrently coupled quadratic integrate-and-fire ( QIF ) neurons [15] . In the large-system limit these equations are exact and fluctuations can be neglected . The exact correspondence between the low-dimensional mean-field equations and the original network allows us to use standard dynamical systems techniques to fully characterize the range of dynamical states in the network . The dynamics in networks of recurrently coupled QIF neurons can be described exactly under the assumptions of all-to-all coupling and quenched neuronal variability , i . e . static distributions in cellular or network properties . For the case of a single network of excitatory cells in which the input currents to individual neurons are distributed , the resulting mean-field equations are [15]: τ 2 r ˙ = Δ π + 2 τ v r , τ v ˙ = v 2 + J τ r + η + I ( t ) - π 2 τ 2 r 2 . ( 1 ) Here , r is the network average of the firing rate and v is the network average of the membrane potential , J is the strength of synaptic weights . In the derivation of the mean-field equations each synaptic weight is scaled as 1/N , where N is the system size , leading to an order one contribution to the mean input in the thermodynamic limit , whereas fluctuations vanish . η and Δ are , respectively , the center and width of the static distribution of inputs , which is considered to be Lorentzian . External , time-variant forcing is represented here by I ( t ) . The time constant τ is the membrane time constant of the individual neurons and is set to 20ms throughout . This macroscopic model permits a straightforward investigation of the stationary states in the full network . For sufficiently strong synaptic coupling two stable fixed points co-exist over a range of mean external inputs , see Fig 2A . Linear stability analysis further reveals that the stable high-activity fixed point is a focus for sufficiently high rates , whereas the stable low-activity fixed point is a node [15] . The network therefore shows a damped oscillatory response to external perturbations in the high-activity state . This response reflects transient spike synchrony which decays over time due to the heterogeneity; the characteristic time scale of the desynchronization is in fact proportional to the width of the distribution of input currents Δ [21] . This type of spike synchrony is seen ubiquitously in networks of both heterogeneous and noise-driven spiking neurons operating in the mean-driven regime , in which neurons fire as oscillators [15 , 22 , 23] , and is captured in Eq 1 by the interplay between the mean sub-threshold membrane potential and mean firing rate [24] . We use this macroscopic description to systematically investigate the network response to periodic forcing with amplitude A and frequency f , see Eq 8 . Fig 2B shows a phase diagram of the network dynamics as a function of these two parameters . As in Fig 1 we keep track of the final state of the network as a function of the initial state . For sufficiently slow frequencies and over a range of amplitudes the network is always driven to the high-activity state ( green ) . This region therefore corresponds to recall of the memory state , see Fig 2C ( left ) . For an intermediate range of frequencies a sufficiently strong forcing always drives the network to the low-activity state ( red ) , which corresponds to clearance , Fig 2C ( right ) . The frequency band for clearance is essentially set by the frequency of intrinsic oscillations of the high-activity state , i . e . it is a resonant effect , see Fig 3 . Weak forcing and forcing at very high frequencies fail to drive any transitions , while strong forcing at low enough frequencies can enslave the network dynamics entirely ( orange ) . For the parameter values used here recall occurs for frequencies below about 2Hz and clearance in the range between 10-30Hz . In order to characterize the role of spike synchrony in determining the network response , we derive a reduced firing rate equation with the identical fixed-point structure as in the original , exact mean-field equations Eq 1 , but without the subthreshold dynamics . Specifically , the fixed-point value of the firing rate in Eq 1 can be written as r 0 = Φ ( J τ r 0 + η ) , ( 2 ) where Φ is the steady-state f-I curve , which in the case of Eq ( 1 ) is Φ ( x ) = 1 2 π x + x 2 + Δ 2 . ( 3 ) We use the steady-state f-I curve to construct a heuristic firing rate model given by τ r ˙ = - r + Φ ( J τ r + η + I ( t ) ) , ( 4 ) and investigate its response to periodic forcing I ( t ) . Eq 4 is similar in form to the classic Wilson-Cowan firing rate model for a single population [25] . In this case the high-activity branch of the firing rate is a node , i . e . it no longer shows damped oscillations in response to perturbations , see Fig 2D . Furthermore , the region of “clearance” has completely vanished in the phase diagram in Fig 2E , confirming that in the original network it was due to a resonance reflecting an underlying spike synchrony mechanism . Given the simplicity of the mean-field equations Eq 1 we can calculate the linear response of the system analytically , without the need for extensive numerical simulations . The response of the focus to weak sinusoidal inputs ( linear response ) already shows a clear resonance for the high-activity state ( Fig 3A ) , where the resonant frequency is f r e s = 1 2 π 2 r 0 ( 2 π 2 r 0 - J τ ) , ( 5 ) see Methods . Furthermore , additional , sub-harmonic resonance peaks occur when the forcing is sharply peaked , leading to a broadening of the resonance spectrum ( Fig 3A , right ) ; this effect is due to the presence of many sub-harmonics of the linear resonance in the forcing term itself . Conversely , the node does not show such a resonance , indicating a qualitative difference in the response of the two stable fixed points . However , the switching behaviors seen in Figs 1 and 2 and the corresponding destabilization of network states cannot be attributed to this linear resonance alone—nonlinear effects have to be taken into account . This can be seen by plotting the bifurcation diagram for the response of the network to the forcing for several values of the forcing amplitude . For relatively weak , but finite forcing , the network response consists of a periodic orbit in the vicinity of the corresponding unforced fixed-point , Fig 3B ( top ) . As the forcing amplitude is increased , the resonance peak of the focus moves towards slower frequencies , akin to a softening spring . Then , a pair saddle-node bifurcations leads to a range of frequencies in which three periodic orbits coexist , see Fig 3B middle-right . At large enough amplitudes for the sharply-peaked , non-sinusoidal forcing two additional bifurcations occur which are responsible for the “recall” and “clearance” behaviors respectively , see Fig 3B ( bottom-right ) and Fig 3C . Specifically , for sufficiently large frequencies , the stable periodic orbit due to the low-activity node ( blue line ) coexists with the unstable one due to the saddle-point ( green line ) , and with a third state , emanating from the focus ( red line ) . When the forcing frequency is sufficiently small , only the latter solution persists . This can be understood as quasi-stationary response of the system due to the slow forcing , see the Methods section for details . In other words , the forcing here is slow and large enough to push the system beyond the bistable regime into the basin of attraction of the focus . Therefore , at low frequencies the only solution is the periodic orbit in the vicinity of the high-activity focus , which explains why low frequencies are effective in switching on the high-activity state , i . e . for “recall” . On the other hand , in the range of frequencies over which the network response is resonant , period-doubling bifurcations of the focus lead to a frequency band in which all periodic orbits around the focus are unstable . This is due to the rapid occurrence of further period-doubling bifurcations , leading to the emergence of chaotic responses to the forcing . As we show in the Methods section , there exist narrow frequency bands in which these chaotic responses are stable , but a numerical investigation of these shows that they quickly become unstable as the frequency of the forcing is changed . Therefore , the periodic orbit in the vicinity of the low-activity node is the only stable solution . Frequencies in this range are therefore effective in switching off the high-activity state , i . e . for “clearance” . As we show in Fig 3D , the loci of bifurcations that periodic orbits around the focus undergo , explain well the parameter range in the ( A , f ) -plane in which clearance is observed , i . e . the red area in Fig 2B . A single bistable network of neurons serves as a canonical illustration of a memory circuit . However , such a network can only store a single bit of information; actual memory circuits must be capable of storing more information . In terms of neuronal architecture this can be achieved by having a network which is comprised of several or many neuronal clusters [16 , 17 , 26] . We asked to what extent the frequency-selective switching behavior seen in a single bistable network could also be found in a clustered network . We look first at a simple , two-cluster network and then the more general case of a higher-dimensional multi-clustered network . Thus far we have treated oscillations as an extrinsic effect , i . e . we are agnostic as to their origin . To be effective for flexible control of memory states , the oscillatory forcing we have considered here must fulfill two requirements: First , it must have a broad range of possible frequencies , and secondly , it must have a burst-like shape . Here we show that a simple circuit comprised of interacting excitatory and inhibitory populations can satisfy both these requirements . Specifically , we construct a network of QIF neurons consisting of an E-I circuit which spontaneously oscillates , and drives a downstream population of E cells , which itself is bistable , see Fig 6 . Using the corresponding mean-field equations for the E-I circuit , we found a broad region of oscillatory states of the E-I network as a function of the mean external drive to the E and I populations , ηe and ηi respectively , see the phase diagram Fig 6B . By adjusting the external drive to the E and I populations alone we can tune the output frequency over an order of magnitude . This allows us to selectively switch the downstream network on and off , as shown in Fig 6C . Outside the region of bistability ( or multistability in the case of clustered networks ) , neuronal networks will relax to a single stationary state in the response to a transient input . Here we show that this need not be the case if the network activity is subjected to ongoing oscillatory modulation . As an illustration we take a single population of excitatory neurons with strong recurrent excitation , but insufficient tonic drive to place it in the region of bistability . As a result , the response of the network to a transient excitatory stimulus decays to baseline , as seen in Fig 7A ( top ) . However , in the presence of an oscillatory input in the gamma range , which itself only very weakly modulates the network activity ( Fig 7A middle ) , the transient input now switches the network to an activated state with prominent gamma modulation Fig 7A ( bottom ) . Once the oscillations cease ( green arrow ) the activated state vanishes . This phenomenon depends crucially on the presence of the spike-synchrony mechanism underlying the damped oscillatory response of the high-activity focus discussed earlier . Specifically , for the parameter values used in Fig 7 the only fixed-point solution which exists is the low-activity node . Nonetheless , oscillatory forcing at sufficiently high firing rates can still recruit and resonate with the damped oscillatory interaction between the mean firing rate and mean membrane potential in the network . The resulting resonant frequency can no longer be associated with the linear response of the focus as it is a fully nonlinear network property . The phase diagram Fig 7B shows the regions of bistability given an oscillatory forcing , for different forcing amplitudes . For zero amplitude the curve corresponds to the saddle-node ( SN ) bifurcation of the unforced system ( horizontal black line ) . Note that only sufficiently high frequencies allow for bistability given tonic inputs which place the network below the SN . Furthermore , there is a clear resonance in the range of 60–90Hz for these parameter values . As the forcing frequency f → ∞ the curves converge to the SN line of the unforced system . This is because the forcing we use has zero-mean and hence , given the low-pass filter property of neuronal networks , has no effect on the network dynamics at high frequencies .
In this article we have studied the role of oscillations in switching or maintaining specific brain states . Specifically , we identified distinct frequency bands: delta , beta , and gamma with specific functional roles . This finding is especially intriguing given that the networks we study are relatively simple . Connectivity is all-to-all and neurons are exclusively excitatory . For the multi-population networks , interactions between populations are assumed to be mediated by fast inhibition , leading to a winner-take-all behavior . Furthermore , synaptic transmission is considered to be instantaneous , with the only relevant time scale being the membrane time constant ( τ = 20ms ) . The susceptibility of the networks to forcing of distinct frequencies therefore does not depend on the presence of multiple time scales associated with intrinsic currents , synaptic kinetics or sub-classes of inhibitory cells . Rather , the key dynamic factors are: bistability or multistability due to recurrent excitatory reverberation , and transient spike synchrony in response to external drive . Given this , we expect to see the same phenomenology in more biophysically realistic networks as long as there is bistability and external noise sources are not too strong . Additionally , none of the mechanisms we study depend crucially on the specific choice of neuron model , at least for type I spiking models . The “switching-on” at low frequencies depends only on the presence of a saddle-node bifurcation , which is ubiquitous in networks of spiking neurons in the bistable regime . Similarly , the “switching-off” or “clearance” depends only on recruiting spike synchrony , which occurs readily in both integrate-and-fire models as well as conductance-based spiking models [24] . In fact , in the mean-driven regime spiking networks in general robustly exhibit a resonance to oscillatory inputs , which reflects the underlying synchrony mechanism [20 , 23 , 33] . In the region of bistability , low frequencies are effective in pushing the network into a high-activity state; for not too large amplitudes the network remains in the activated state on the downswing of the input . The cut-off frequency for this “recall” signal is determined by the escape time of the network from the vicinity of the saddle-node bifurcation in the low-activity state , and here is a few Hertz , see Fig 2A . In multi-stable networks , this same mechanism allows for robust switching between distinct memory states . On the other hand , frequencies in the beta range are effective in switching off the high-activity state by resonantly driving bouts of spike synchrony . The precise frequency range depends on network parameters , see Fig 8 . In both cases the relevant frequency ranges scale with the membrane time constant of the neurons . Therefore , e . g . choosing a time constant τ = 10ms will simply stretch the x-axis of the phase diagram in Fig 2B by a factor of two . Finally , we showed that forcing in the gamma range can allow for robust working memory states which otherwise do not exist , i . e . the system sits outside the region of bistability with oscillatory forcing . This mechanism once again depends on resonantly recruiting spike synchrony . We find that non-sinusoidal , burst-like drive is most effective in switching the network state , see Fig 3A and 3B . In fact , this is precisely the type of oscillation which readily emerges in a simple E-I network . Furthermore , the oscillation frequency can be modulated over a wide range through changes in the tonic drive to the E-I circuit alone , see Fig 6 . This means that the state of downstream memory networks can be flexibly controlled via an E-I circuit through global changes in excitability alone . While here we have considered networks in which intrinsic oscillatory activity is due to transient spike synchronization , spiking networks can also generate oscillatory activity due to E-I and I-I loops , which can occur in the absence of strong spike synchrony . For example , networks of coupled excitatory ( E ) and inhibitory ( I ) spiking neurons readily generate oscillations via a Hopf bifurcation when excitation is sufficiently strong and fast [34 , 35] . The E-I loop , and in particular the ratio of E to I time constants , largely sets the frequency of these oscillations , which tend to lie in the gamma range ( 30–100Hz ) . On the other hand , the I-I loop itself can underlie the generation of fast oscillations ( >100Hz ) , the frequency of which is set by the inhibitory synaptic delay [35 , 36] . Both the E-I and I-I loops contribute to the population frequency in E-I networks , with the E-I loop dominating when recurrent excitation is strong . Resonant responses to periodic stimuli due to the E-I loop in neuronal circuits have been studied in firing rate models [37–39] as well as in networks of LIF neurons [33] . Damped oscillatory activity due to the E-I loop can also arise in the high-activity state of the bistable regime of E-I networks [40] . In this scenario external periodic drive could also be used for “clearance” of the activated state by resonating with the E-I loop . While the phenomenology of this resonance would be similar to the resonance we have considered in this manuscript , the mechanism is nonetheless distinct as it does not involve spike synchrony . On the other hand , spike synchrony does robustly lead to resonances in E-I networks , as measured for example by the linear response [23] . In principle both resonances could be present in the bistable regime of E-I networks , allowing for an even more complex response to oscillatory input than we have studied here . Current non-invasive brain-stimulation techniques , such as repetitive transcranial magnetic stimulation [41] , or transcranial alternating current stimulation [42] , apply transient oscillatory signals to large parts of the brain . Our study may be useful to investigate the impact of such signals on the dynamics of neuronal mass models and the psychological and behavioral effects of neuromodulation . Our results could also be of relevance for investigating the use of deep-brain stimulation to treat Parkinson’s disease [43] and ( pharmacologically ) treatment-resistant depression [44] . Although the model used here describes networks of spiking neurons with instantaneous synapses , future studies could also incorporate synaptic dynamics with appropriate time scales for excitatory and inhibitory transmission [24] . These time scales can be influenced by drugs , or ( pathological ) changes in neurotransmitters . The framework developed here may therefore serve as a tool to study the cause of functional deficiencies in synapse-related conditions , so-called “synaptopathies” [45 , 46] .
Neural mass and neural field models are an important tool for understanding macroscopic neuronal dynamics . Classical models include the Wilson-Cowan model [25 , 47] or the Amari model [48 , 49] . However , such macroscopic models of brain activity often pose a stark simplification of the actual dynamics , and often miss important features from the spiking dynamics , such as spike synchronization . Recently , there have been advances in linking the microscopic and macroscopic dynamics of networks of spiking neurons [15 , 23 , 50–57] . We consider a neural mass model that was recently derived from networks of all-to-all coupled quadratic integrate-and-fire neurons in the thermodynamic limit [15] , see Eq ( 1 ) . To simplify the mathematical treatment , we divide t by τ which represents the case of time being measured in units of τ , thus eliminating τ from the equations: r ˙ = Δ π + 2 v r , v ˙ = v 2 + J r + η + I ( t ) - π 2 r 2 . ( 6 ) Here , r represents the ensemble average of the firing rate of neurons , and v represents the ensemble average of the membrane potential . The parameters η and Δ represent the center and witdh of the Lorentzian distribution of time-invariant input currents into the neuronal ensemble , and J is the coupling constant between neurons . Time-varying external inputs are given by I ( t ) . The original model ( 1 ) can then be recovered by t → τt , r → r/τ . As we set τ = 20ms , r = 1 here corresponds to a firing rate of r = 50Hz in the full model . Here , we consider I ( t ) to be T-periodic , i . e . I ( t + T ) = I ( t ) . We distinguish between two types of input: sinusoidal input , I ( t ) = A sin ( 2 π f t ) , ( 7 ) and non-sinusoidal input , I ( t ) = A ( γ sin ( π f t ) n - 1 ) , ( 8 ) where we take n = 20 for the simulations presented in this paper . The parameter A represents the amplitude of the forcing . The constant γ is chosen such that ∫ 0 T I ( t ) d t = 0 . We choose this type of zero-mean forcing to avoid any changes in network excitability which a tonic DC-offset might cause . In other words , the input models a reorganization of afferent spikes into periodic volleys without adding any additional spikes . In the non-sinusoidal case the spikes are more synchronized than in the sinusoidal case . We compare the full model equations with its equivalent heuristic firing rate equation , which preserves the fixed point structure but reduces the dynamical behavior . This is done by considering stationary solutions given by 0 = Δ π + 2 v r , 0 = v 2 + J r + η - π 2 r 2 . ( 9 ) Solving these equations for r is equivalent to solving Eq 2 . Thus , the reduced heuristic firing rate equations can be expressed by r ˙ = - r + Φ ( J r + η ) , ( 10 ) where the f-I function Φ ( Jr + η ) is given by Eq 3 . Ignoring transient dynamics , the response of the model equations to the external input I ( t ) is T-periodic as well , at least in the limit of small amplitudes A ( an exception are period-doubled solutions , which are a nonlinear phenomenon only relevant at larger A ) . In this case the corresponding Fourier spectra of the firing rate r ( t ) and of the membrane potential v ( t ) are discrete: r ( t ) = r 0 + ( r 1 e i ω t + r 2 e 2 i ω t + … + c . c . ) , v ( t ) = v 0 + ( v 1 e i ω t + v 2 e 2 i ω t + … + c . c . ) ( 11 ) For brevity of exposition we use here the angular frequency ω = 2πf instead of the ordinary frequency f . This approach describes the projection of solutions of r and v from a continuous space R onto a discrete function space V , with orthogonal basis functions einωt , n ∈ Z . The same Fourier decomposition applies to the input current I ( t ) : I ( t ) = I 0 + ( I 1 e i ω t + I 2 e 2 i ω t + ⋯ + c . c . ) ( 12 ) To determine the linear response of the model equations [58] , we first carry out a Fourier decomposition of the system linearized around the fixed points given by r0 and v0: i n ω r n = 2 v 0 r n + 2 r 0 v n , i n ω v n = J r n + I n + 2 v 0 v n - 2 π 2 r 0 r n . ( 13 ) Solving this set of linear equations , we obtain r n = 2 r 0 I n Ω n - 1 , v n = ( i n ω - 2 v 0 ) I n Ω n - 1 , ( 14 ) with Ω n = ( 2 v 0 - i n ω ) 2 + ω 0 2 , ( 15 ) where ω0 is the ( angular ) resonant frequency: ω 0 2 = - 2 r 0 ( J - 2 π 2 r 0 ) . ( 16 ) The resonant frequency is state-dependent and changes with model parameters . Reintroducing the time scale τ , perturbations of the upper branch solution resonate at a frequency ω r e s = 2 r 0 ( 2 π 2 r 0 - J τ ) , ( 17 ) where r0 is the value of the steady-state firing rate . This is true as long as the argument of the square root is positive . Therefore as the firing rate decreases along the upper branch , for decreasing external input , the frequency decreases to zero at which point the focus becomes a node . This point occurs before the saddle-node is reached unless Δ = 0 in which case it exactly coincides with the saddle-node . Fig 8 shows how the linear resonant frequency of the stable focus in the bistable regime of a network of excitatory QIF neurons varies as a function of the mean external input η and the strength of synaptic coupling J . Recall is not possible to the left of the red curve given the nonlinear forcing used here . This line is determined by setting Amin = Amax , see Eqs ( 26 ) and ( 27 ) further below . The time-dependent linear response of the firing rate and the membrane potential is now given by r ( t ) = ∑ n = 1 ∞ 2 r 0 I n Ω n - 1 e i n ω t + c . c . , v ( t ) = ∑ n = 1 ∞ ( i n ω - 2 v 0 ) I n Ω n - 1 e i n ω t + c . c . ( 18 ) From this , we can derive the amplitude of the linear response of the firing rate , r l i n ( ω ) = ( max t r ( t ) - min t r ( t ) ) / 2 , ( 19 ) and analogously of the membrane potential . Alternatively , one can derive the time-averaged linear response ( “power” ) of the system: R 2 ( ω ) = 1 T ∫ 0 T r ( t , ω ) 2 d t = 8 r 0 2 ∑ n = 1 ∞ | I n | 2 | Ω n | - 2 , ( 20 ) V 2 ( ω ) = 1 T ∫ 0 T v ( t , ω ) 2 d t = 2 ∑ n = 1 ∞ ( n 2 ω 2 + 4 v 0 2 ) | I n | 2 | Ω n | - 2 . ( 21 ) Here we have made use of the orthogonality of the basis functions , and the fact that T = 2π/ω . In order to exhaustively and accurately trace the bifurcations that occur in the model equations , we make use of AUTO 07p [59] . Since this software is designed to deal with autonomous systems , we recast the ( non-autonomous ) model Eq ( 1 ) into a set of autonomous equations: r ˙ = Δ π + 2 v r , v ˙ = v 2 + J r + η + A I ( x ( t ) ) - π 2 r 2 , x ˙ = x + ω y - ( x 2 + y 2 ) x , y ˙ = y - ω x - ( x 2 + y 2 ) y . ( 22 ) The last two equations create the periodic stimulus x ( t ) = sin ( ωt ) in the model equations . We distinguish the sinusoidal case , I ( x ( t ) ) = x ( t ) , ( 23 ) and the non-sinusoidal case I ( x ( t ) ) = γ x ( t ) 20 - 1 . ( 24 ) Continuation of the forced system is performed by starting from a known fixed point ( r0 , v0 ) at A = 0 , and continuing solutions by increasing A up to the desired value . We use the L2-norm as a scalar measure to represent periodic solutions: L 2 ( r ) = 1 T ∫ 0 T r ( t ) 2 d t . ( 25 ) Where we perform this one-parameter continuation , we represent solution branches by plotting the L2-norm against the parameter that is being varied . Where we perform two-parameter continuation , we plot the loci of bifurcations against the two parameters being varied . Here we illustrate in greater detail the mechanisms underlying the “switching on” of activated network states ( or simple switching between attractors in the case of a multi-stable network ) at low frequencies , and the “switching off” of activated states at frequencies in the beta range . A natural extension of the single-population model is to consider a network of neural masses: r ˙ n = Δ π + 2 v n r n , v ˙ n = v n 2 + J ∑ m = 1 N A n m r m + η + I ( t ) + σ ξ n ( t ) - π 2 r n 2 , ( 28 ) where the adjacency matrix A with entries Anm determines the connectivity structure between neural masses . The term σξn ( t ) describes an additional noise input , where σ is the noise amplitude , and ξn ( t ) is the random variable . In this paper we consider two scenarios , the first of which is two neural populations with recurrent excitation and mutual inhibition . The adjacency matrix of such a network is given by A = ( J e J i J i J e ) , ( 29 ) where Ji < 0 < Je . In the second scenario , we examine the dynamics within a Hopfield network [16] . Rather than creating the network through learning algorithms , we build the network as follows . First , we choose the patterns that the network should encode and write them into an array U . Each column of this array represents one pattern , where we put 1 for populations that are active in this pattern , and 0 otherwise . As a result , the array U has the size N × Npat , where Npat is the number of patterns encoded , and N is the network size . Each pattern consists of Np active populations . The adjacency matrix of a network that encodes these patterns can then be constructed as follows [17] , A = ( U - p ) × ( U - p ) T - Q , ( 30 ) with p = Np/N . The entries of A are capped at a maximum value of ( 1 − p ) 2 − Q to account for saturation effects in synaptic plasticity . Otherwise , the strength of connections in the network would steadily increase as patterns are added . The offset Q introduces global inhibition that stabilizes the encoded patterns . We set Q = 0 . 2 . Each population is subjected to an independent Ornstein-Uhlenbeck process ξn ( t ) to break the symmetry of the networks . The Ornstein-Uhlenbeck process is implemented as Langevin equation: τ ξ ˙ n = - ξ n + ζ n ( t ) , ( 31 ) where ζn ( t ) are independent Gaussian white noise sources , 〈ζn ( t ) ζm ( t − s ) 〉 = δ ( s ) δmn , and τ is the characteristic time scale , which we set to τ = 20ms . To create a network that generates oscillations , we consider a network of an excitatory population interacting with an inhibitory one: r ˙ e = Δ π + 2 v e r e , v ˙ e = v e 2 + J e r e + J i r i + η e - π 2 r e 2 , r ˙ i = Δ π + 2 v i r i , v ˙ i = v i 2 + J e r e + J i r i + η i - π 2 r i 2 . ( 32 ) For simplicity , we choose Je = −Ji = J . The two populations differ in terms of the means of their tonic input currents , ηe and ηi . We vary these two parameters to identify the regime where stable oscillations exists , and to change the frequency of these oscillations . In the network model , the high-activity branch of solutions in the bistable regime exhibits damped oscillations . Periodic external drive can resonate with these intrinsic oscillations , leading to destabilizing period-doubling bifurcations as seen in the previous section . Here we show that this mechanism is present in the simplest possible model exhibiting a saddle-node bifurcation and for which the upper branch becomes a focus: x ˙ = y , ( 33 ) y ˙ = μ - x 2 - a x y + I ( t ) . ( 34 ) This model is a particular unfolding of the so-called Takens-Bogdanov normal form [60] , for which there is no Hopf bifurcation , which is the relevant case for our network model . It is easily shown that a saddle-node bifurcation occurs in these equations at μ = 0 and that the fixed point solutions are x 0 = ± μ and y0 = 0 for μ > 0 , see Fig 11A . Furthermore , the solution x 0 = - μ is a saddle , and x 0 = μ is a stable focus for which the frequency goes to zero smoothly as μ → 0 . Fig 11B shows that in the forced system there is a range of frequencies for which there is no stable solution; in the normal form equation the solution diverges while in the network model the system settles to a periodic orbit in the vicinity of the low-activity state . The instability is due to a series of period-doubling bifurcations as in the full system . Furthermore , comparison of the phase diagram of the normal form equation with that of the full system shows they are qualitative similar , see Fig 11C . This indicates that the nonlinear resonance seen in the network of QIF neurons is a generic feature of any system with a stable focus in the vicinity of a saddle-node bifurcation . | Oscillations are ubiquitous in the brain and often correlate with distinct cognitive tasks . Nonetheless their role in shaping network dynamics , and hence in driving behavior during such tasks is poorly understood . Here we provide a comprehensive study of the effect of periodic drive on neuronal networks exhibiting multistability , which has been invoked as a possible circuit mechanism underlying the storage of memory states . We find that oscillatory drive in low frequency bands leads to robust switching between stored patterns in a Hopfield-like model , while oscillations in the beta band suppress sustained activity altogether . Furthermore , inputs in the gamma band can lead to the creation of working-memory states , which otherwise do not exist in the absence of oscillatory drive . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"resonance",
"frequency",
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"membrane",
"potential",
"electrophysiology",
"neuroscience",
"learning",
"and",
"memory",
"cognitive",
"neuroscience",
"cognition",
"memory",
"computer",
"and... | 2018 | Network mechanisms underlying the role of oscillations in cognitive tasks |
Schistosomiasis is a helminthic disease that affects more than 200 million people . An effective vaccine would be a major step towards eliminating the disease . Studies suggest that T follicular helper ( Tfh ) cells provide help to B cells to generate the long-term humoral immunity , which would be a crucial component of successful vaccines . Thus , understanding the biological characteristics of Tfh cells in patients with schistosomiasis , which has never been explored , is essential for vaccine design . In this study , we investigated the biological characteristics of peripheral memory Tfh cells in schistosomiasis patients by flow cytometry . Our data showed that the frequencies of total and activated peripheral memory Tfh cells in patients were significantly increased during Schistosoma japonicum infection . Moreover , Tfh2 cells , which were reported to be a specific subpopulation to facilitate the generation of protective antibodies , were increased more greatly than other subpopulations of total peripheral memory Tfh cells in patients with schistosomiasis japonica . More importantly , our result showed significant correlations of the percentage of Tfh2 cells with both the frequency of plasma cells and the level of IgG antibody . In addition , our results showed that the percentage of T follicular regulatory ( Tfr ) cells was also increased in patients with schistosomiasis . Our report is the first characterization of peripheral memory Tfh cells in schistosomasis patients , which not only provides potential targets to improve immune response to vaccination , but also is important for the development of vaccination strategies to control schistosomiasis .
Schistosomiasis remains a major public health problem in many developing countries . Estimates place the current number of infections at approximately 200 million people , with another 600 million considered at risk [1] . Although praziquantel remains highly effective in schistosomiasis treatment , it provides only short-term protection and does not block disease transmission or reinfection [2] . Furthermore , drug resistance and decreased susceptibility to praziquantel may occur with long-term use of the drug [3] . Thus , an effective vaccine against schistosome infection would be a major step towards eliminating this devastating and widespread tropical parasitic disease . An effective anti-schistosome vaccine would immensely reduce the morbidity associated with schistosomiasis through induced immune responses leading to decrease in parasite load and reduced egg production [4 , 5] . The antibody dependent cell mediated cytotoxicity ( ADCC ) of effector immune cells such as eosinophils and macrophages has been suggested as one of the most important mechanisms of anti-schistosome vaccine-mediated protection [6–8] . Thus , the generation of long-term humoral immunity is a crucial component of successful vaccines . Interactions between T cells and B cells in germinal centers ( GCs ) are reported to be required for the generation of long-term humoral immunity [9] . Recent studies reveal that in GCs , a specialized subset of CD4+ T cells called T follicular helper ( Tfh ) cells , provide help to B cells to undergo proliferation , isotype switching and somatic hypermutation , resulting in long-lasting antibody ( Ab ) responses [10–12] . Thus , understanding the biological characteristics of Tfh cells in schistosomiasis patients is one of central issues to develop the vaccination strategies to control schistosomiasis . In this study , we for the first time explored the characteristics of peripheral memory Tfh cells in patients with schistosomiasis japonica , which provides a better understanding of the role of Tfh cells in schistosomiasis and contributes to the development of the future vaccination strategies in schistosomiasis .
Ethical clearance for this study was obtained from the Institutional Review Board of Nanjing Medical University , Nanjing , China ( Permit Number: 2014NMUIEC001 ) . The aims and objectives of the study were explained to each participant and written informed consent was obtained . All personal identifiers of the study notes and tapes were kept confidential and destroyed once the study was completed . The study was conducted on a total of 100 subjects , and all subjects were from a village in Chizhou City , Anhui province . The subjects included 50 healthy adult controls , 50 patients with schistosomiasis japonica by egg detection using the Kato-Katz method with duplicate examination of 3 consecutive stool specimens obtained from each individual [13] . The healthy controls did not display a history , laboratory or clinical signs of schistosomal infection , did not suffer from coinfections with HBV or HCV , and did not use medication two weeks before blood collection . Human peripheral blood mononuclear cells ( PBMCs ) were collected into sodium heparin tubes ( BD Biosciences , San Diego , CA ) and purified by Ficoll-paque plus ( GE healthcare , Sweden ) density gradient centrifugation . Cells recovered from the gradient interface were washed twice , and stained for 30 min at 4°C with the following antibodies: CD3-FITC ( clone HIT3a ) , CD4-Percp-Cy5 . 5 ( clone RPA-T4 ) , CD45RA-APC-H7 ( clone HI100 ) , CXCR5-Alexa Fluor 647 ( clone RF8B2 ) , PD-1-PE-Cy7 ( clone EH12 . 1 ) , ICOS-PE ( clone DX29 ) , CCR6-PE ( clone 11A9 ) , CXCR3-PE-Cy7 ( clone 1C6 ) , CD27-APC-H7 ( clone M-T271 ) , CD38-PE ( clone HIT2 ) , CD86-PE-Cy7 ( clone 2331 ) , CD19-APC ( clone HIB19 ) , all from BD Biosciences . In brief , for total or activated peripheral memory Tfh surface marker analysis , cells were incubated with CD3-FITC , CD4-Percp-Cy5 . 5 , CD45RA-APC-H7 , CXCR5-Alexa Fluor 647 , PD-1-PE-Cy7 , ICOS-PE . For Th1 , Th2 , or Th17 surface marker analysis , cells were incubated with CD3-FITC , CD4-Percp-Cy5 . 5 , CCR6-PE , and CXCR3-PE-Cy7 . For Tfh1 , Tfh2 , or Tfh17 surface marker analysis , cells were incubated with CD3-FITC , CD4-Percp-Cy5 . 5 , CD45RA-APC-H7 , CXCR5-Alexa Fluor 647 , CCR6-PE , and CXCR3-PE-Cy7 . For circulating B cell surface marker analysis , cells were stained with CD3-FITC , CD4-Percp-Cy5 . 5 , CD19-APC , CD27-APC-H7 , CD38-PE , and CD86-PE-Cy7 . The samples were fixed with 1% paraformaldehyde/PBS . Cells acquisition was performed using a FACSVerse cytometer ( Lasers: 488 and 633; Mirrors: 507 LP , 560 LP , 665 LP , 752 LP , 660/10 , and 752 LP; Filters: 488/15 , 527/32 , 568/42 , 700/54 , 783/56 , 660/10 and 783/56 , BD Biosciences ) . Data were analyzed with FlowJo ( Tree Star , version 10 . 0 . 7 ) . To evaluate the percentages of GATA-3+ Tfh and Tfr cells , PBMCs were stained with CD3-FITC , CD4-Percp-Cy5 . 5 , CD45RA-APC-H7 , CXCR5-Alexa Fluor 647 . Then , the cells were further intracellular stained with GATA-3-PE ( clone L50-823 , BD Biosciences ) or FOXP3-PE ( clone PCH101 , eBioscience , San Diego , CA ) after they were permeabilized with cold Fix/Perm Buffer ( eBioscience ) . The samples were fixed with 1% paraformaldehyde/PBS . Cells acquisition was performed using a FACSVerse cytometer ( BD Biosciences ) . Data were analyzed with FlowJo ( Tree Star , version 10 . 0 . 7 ) . To quantify the total serum IgG and IgE levels two commercial kit ( Bethyl , Texas , USA ) with established protocols from the manufacturer was used . Briefly , 96-well plates ( Nunc MaxiSorp ) were coated with 1 μg/well of capture antibody for IgE ( catalog number A80-108A , Bethyl ) or IgG ( catalog number A80-104A , Bethyl ) in Coating Buffer ( 0 . 05M carbonate-bicarbonate , pH 9 . 6 ) for 1 h at 25°C and blocked for 30 min with 200 μL/well Blocking Solution ( 50 mM Tris , 0 . 14 M NaCl , 1% BSA , pH 8 . 0 ) . Between each step , the plates were washed 5 times with Wash Solution ( 50 mM Tris , 0 . 14 M NaCl , 0 . 05% Tween 20 , pH 8 . 0 ) . The serum from each patient was diluted 1:2 or 1:20 , 000 for IgE or IgG in Sample/Conjugate Diluent Solution ( 50 mM Tris , 0 . 14M NaCl , 0 . 05% Tween 20 , 1% BSA ) , and 100 μL/well was added to the plates . Known concentration of purified human IgE ( catalog number RC80-108-6 , Bethyl ) or IgG ( catalog number RS10-110-4 , Bethyl ) was added to each plate to obtain a standard curve . Serum samples and standards were incubated for 1 h at 25°C . IgE or IgG bound to the plates was detected by the addition of HRP-anti-human IgE ( catalog number A80-108P , Bethyl ) or HRP-anti-human IgG ( catalog number A80-104P , Bethyl ) at 1:75 , 000 or 1:200 , 000 dilution in Sample/Conjugate Diluent Solution , followed by the addition of Substrate Solution ( catalog number E102 , Bethyl ) . After 15 min , the reaction was stopped with 100 μL of 0 . 18M sulfuric acid solution . Absorbance was measured at 450 nm using an automated ELISA reader ( BioTek Synergy HT , Texas ) . For each patient , the amount of total IgE or IgG was quantified in triplicate . All data were analyzed using SPSS software ( IBM , version 22 ) . Significant Differences between specimens were determined by using Student’s t test or Mann-Whitney U test . Correlations were determined by Spearman’s ranking . The differences at p<0 . 05 were considered to be statistically significant .
Both our previous study [14] and other literature [15] described the increased frequency of Tfh cells in mice with schistosome infection . However , whether Tfh cells increase in percentage in schistosomiasis patients remains unknown . To study the biological characteristics of Tfh cells in patients with schistosomiasis , a total of 50 patients and 50 healthy controls were recruited . There was no statistically significant difference in the distribution of age or gender between patients and healthy controls ( Table 1 ) . The frequency of CD4+ T cells among total lymphocytes was comparable between patients and healthy controls , although the percentage of T cells was slightly lower in patients with schistosomiasis ( Fig 1B and 1C ) . Next , we compared the frequencies of total peripheral memory Tfh cells ( CXCR5+CD45RA-CD4+ T cells ) [16] and activated peripheral memory Tfh cells ( PD-1+ICOS+Tfh cells ) [17 , 18] among CD4+ T or total peripheral memory Tfh cells in 50 patients with schistosomiasis japonica to 50 healthy controls . Results showed the increased frequencies of total and activated peripheral memory Tfh cells in patients with schistosomiasis ( Fig 1A and 1D–1H ) . In addition , we found that almost all of the CXCR5+CD45RA-CD4+ T cells expressed high level of GATA-3 ( Fig 1I ) . Evidence supports that peripheral memory Tfh cells in human can be subdivided into three major subsets with distinguished biological functions according to expression of CXCR3 and CCR6 , Tfh1 ( CXCR3+CCR6-CD45RA-CXCR5+CD3+CD4+ cells ) , Tfh2 ( CXCR3-CCR6-CD45RA-CXCR5+CD3+CD4+ cells ) , and Tfh17 ( CXCR3-CCR6+CD45RA-CXCR5+CD3+CD4+ cells ) [19] . We then determined the distribution of peripheral memory Tfh-cell subsets in healthy controls and schistosomiasis patients . Results showed that Tfh2 cells were a predominant subset of peripheral memory Tfh cells , and accounted for more than 50% of total peripheral memory Tfh cells in patients with schistosomiasis ( Fig 2A and 2B ) . In addition , the percentage of total Th2 cells , which include Tfh2 cells , was greater in schistosomiasis patients than those in healthy controls ( Fig 2C ) . Furthermore , we found that the percentage of Tfh2 cells , rather than that of Tfh17 or Tfh1 , is significantly increased in schistosomiasis patients ( Fig 2D–2G ) . Given that Tfr cells were identified as a Treg cell subset specialized for suppressing B and Tfh cells [20–22] , we next investigated the biological characteristics of circulating Tfr cells in schistosomiasis patients . Results showed that most of circulating Tfr are CD45RA- cells ( Fig 2H ) , which is consistent with the previous observation that circulating Tfr cells have memory-like properties [22] . Furthermore , we found that the percentage of circulating Tfr cells within CD4+ T cells ( Fig 2I and 2J ) or total Tfh cells ( Fig 2K ) was significantly increased in schistosomiasis patients . Given the roles of Tfh cells in providing help to B cells , we next characterized the frequencies of different subsets of B cells by flow cytometry analysis . As shown in Fig 3 , the percentages of CD27+CD19+CD3-CD4- memory B cells [23–25] , CD86+CD19+CD3-CD4- activated B cells [26–28] , and CD38++CD19+CD3-CD4- plasma cells [29–31] in schistosomiasis patients were significantly greater than those in the HCs , although the percentage of total B cells was slightly decreased in patients with schistosomiasis . In contrast , the percentage of CD27-CD19+CD3-CD4- naïve B cells [32] in patients was less than that in the HCs ( Fig 3C ) . Furthermore , the percentage of Tfh2 cells was moderately correlated with the percentage of plasma cells ( rs = . 362 , p = . 01 ) in schistosomiasis patients , and its correlation with the percentage of activated B cells ( rs = . 211 , p = . 141 ) followed the same trend but it did not reach statistical significance ( Fig 3I and 3J ) . Given that Tfh2 cells are efficient at promoting IgG and IgE secretion [33] , we next determined whether the frequency of Tfh2 cells was associated with the levels of total IgG and IgE antibodies in schistosomiasis patients . Results showed the increased concentrations of total IgG [HC vs . Sj: mean = 359 . 7 , 95% confidence interval ( 95% CI ) = 343 . 6–375 . 7 vs . mean = 417 . 9 , 95% CI = 398 . 2–437 . 5; p<0 . 001] and IgE antibodies ( HC vs . Sj: mean = 10 . 6 , 95% CI = 9 . 5–11 . 7 vs . mean = 38 . 6 , 95%CI = 30 . 5–46 . 7; p<0 . 001 ) in schistosomiasis patients ( Fig 4A and 4C ) . Furthermore , a striking correlation between the percentage of Tfh2 cells and the level of total IgG ( rs = . 425 , p = . 002 ) was observed in schistosomiasis patients ( Fig 4B ) . Although there was a tendency of correlation between the percentage of Tfh2 and the level of total IgE ( rs = . 173 , p = . 229 ) in schistosomiasis patients , it did not reach statistical significance ( Fig 4D ) .
Prevention and control of schistosomiasis demands an effective vaccine . T follicular helper cells have a pivotal role in the generation of the long-term humoral immunity and are proved to be one of crucial contributors of successful vaccines . However , the lack of knowledge about Tfh cells in schistosomiasis patients limits the ability to develop successful anti-schistosome vaccinations . Here , we characterized the distribution of peripheral memory Tfh cells in patients with schistosomiasis japonica . Our study significantly extends our understanding of Tfh cells in patients with schistosomiasis , which is helpful for vaccine design for the prevention of schistosome infection . Although the phenotypes of Bona fide Tfh cells in GCs are easy to be analyzed by flow cytometry , it is not only difficult to get lymph nodes from schistosomiasis patients only , but from humans in general . Fortunately , studies in humans showed that peripheral memory Tfh cells share functional properties with bona fide Tfh cells in secondary lymphoid organs [17 , 18 , 33–36] , indicating the analysis of peripheral memory Tfh cells by flow cytometry is an alternative approach to study the biological characteristics of bona fide Tfh cells in human . Here , for the first time we revealed that the percentages of total and activated peripheral memory Tfh cells were significantly increased in schistosomiasis patients . These findings are in accordance with our previous observation that Tfh cells are substantially increased in schistosome-infected mice [14] . Human peripheral memory Tfh cells can be divided into three major subsets with distinguished functions according to the analysis of CXCR3 and CCR6 expression , including in CXCR3+CCR6- Tfh1 cells , CXCR3-CCR6- Tfh2 cells and CXCR3-CCR6+ Tfh17 cells [19] . Results from staphylococcal enterotoxin B in vitro coculture experiment suggested that human blood Tfh2 and Tfh17 , but not Tfh1 , cells can help naive B cells to produce immunoglobulins via producing interleukin-21 [33] . More specifically , it suggests that Tfh2 cells are considered to be efficient at promoting IgG and IgE secretion , whereas Tfh17 cells promote IgG and IgA secretion [33] . Our data showed that the significant increase of Tfh2 cells is a major contributor to the increased frequency of total peripheral memory Tfh cells in patients with schistosomiasis japonica . These findings nicely connect to our observation that the levels of memory B cells , activated B cells , plasma cells as well as total IgG and IgE responses were considerably increased in patients with schistosomiasis . More importantly , both IgG and IgE antibodies have been reported to be essential components of protective immunity , and involved in the ADCC of eosinophils and macrophages against schistosome larvae [37–39] , which is considered as one of the most important means of anti-schistosome vaccine-mediated protection [6–8] . Thus , Tfh2 cells , a predominant subset of peripheral memory Tfh cells in schistosomiasis patients , might be considered as a potential target to improve IgG and IgE responses to vaccination . However , Tfh2 cells secrete Th2 cytokines , i . e . , IL-4 , IL-5 , and IL-13 [33] . Given that the prolonged excessive production of Th2 cytokines contributes to the development of hepatic fibrosis and chronic morbidity in schistosomiasis [40] and the progression of Th2-mediated pathology in some diseases , such as asthma and other infectious diseases caused by extracellular parasites [41] , it is very important for us to consider adverse effect of anti-schistosome vaccines on triggering Th2-mediated inflammation responses , particularly on liver pathology in patients with a history of infection with schistosoma japonica , when we manipulate Tfh2 cells to enhance IgG and IgE responses to vaccination . Strikingly , we found that the frequency of circulating Tfr cells , which play a crucial role in GC responses by limiting Tfh and GC B cell numbers as well as plasma cells differentiation [20] , was significantly increased in schistosomiasis patients . In summary , our study for the first time described the distribution of peripheral memory Tfh cells , circulating Tfr cells , and B cells in patients with schistosomiasis japonica , which provides us a better understanding of the biological characteristics of these cells in patients with schistosomiasis . | Schistosomiasis affects more than 200 million people worldwide and causes more than 280 , 000 deaths per year . Current control strategies are based on chemotherapy , but recurrent reinfection of people living in endemic areas makes researchers search for an effective vaccine to provide long-term protection against schistosomiasis . The generation of long-lived high-affinity antibodies after vaccination is a pivotal step for anti-schistosome vaccine to eliminate schistosomiasis . Considering it is well-known that Tfh cells are specialized effector CD4+ T cells that provide help for germinal center ( GC ) formation and induce GC B cells to develop protective antibody responses , understanding the biology of Tfh cells in schistosomiasis patients is fundamental for vaccine strategy development . Here , for the first time , we documented increased frequencies of total and activated peripheral memory Tfh cells in schistosomiasis patients . Furthermore , we showed that Tfh2 cells were a major contributor to increased frequency of peripheral memory Tfh cells in patients with schistosomiasis japonica . More importantly , we found the significant correlations of the percentage of Tfh2 cells with both the frequency of plasma cells and the level of total IgG antibody in schistosomiasis patients . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Distribution of Peripheral Memory T Follicular Helper Cells in Patients with Schistosomiasis Japonica |
Permethrin is the active component of topical creams widely used to treat human scabies . Recent evidence has demonstrated that scabies mites are becoming increasingly tolerant to topical permethrin and oral ivermectin . An effective approach to manage pesticide resistance is the addition of synergists to counteract metabolic resistance . Synergists are also useful for laboratory investigation of resistance mechanisms through their ability to inhibit specific metabolic pathways . To determine the role of metabolic degradation as a mechanism for acaricide resistance in scabies mites , PBO ( piperonyl butoxide ) , DEF ( S , S , S-tributyl phosphorotrithioate ) and DEM ( diethyl maleate ) were first tested for synergistic activity with permethrin in a bioassay of mite killing . Then , to investigate the relative role of specific metabolic pathways inhibited by these synergists , enzyme assays were developed to measure esterase , glutathione S-transferase ( GST ) and cytochrome P450 monooxygenase ( cytochrome P450 ) activity in mite extracts . A statistically significant difference in median survival time of permethrin-resistant Sarcoptes scabiei variety canis was noted when any of the three synergists were used in combination with permethrin compared to median survival time of mites exposed to permethrin alone ( p<0 . 0001 ) . Incubation of mite homogenates with DEF showed inhibition of esterase activity ( 37% ) ; inhibition of GST activity ( 73% ) with DEM and inhibition of cytochrome P450 monooxygenase activity ( 81% ) with PBO . A 7-fold increase in esterase activity , a 4-fold increase in GST activity and a 2-fold increase in cytochrome P450 monooxygenase activity were observed in resistant mites compared to sensitive mites . These findings indicate the potential utility of synergists in reversing resistance to pyrethroid-based acaricides and suggest a significant role of metabolic mechanisms in mediating pyrethroid resistance in scabies mites .
Scabies is an infectious skin disease caused by the ectoparasite , Sarcoptes scabiei . The mite lives in the skin of hosts where it passes through a series of life stages ( eggs , larva , protonymph , tritonymph , and adult ) . A severe manifestation of the disease i . e . , crusted scabies can occur which may predispose to streptococcal pyoderma and the subsequent development of acute post streptococcal glomerulonephritis , acute rheumatic fever and rheumatic heart disease [1] . Treatment generally entails the application of topical creams for classical scabies , while oral ivermectin is recommended for crusted scabies [1] . Permethrin , used at a concentration of 5% , is the active component of topical creams commonly used to treat the disease . Since its introduction in Australia in 1994 for the treatment of scabies , it has been widely used in endemic communities in mass treatment programs [2] . Recent evidence from a prospective study of in vitro acaricide sensitivity has demonstrated increased tolerance to permethrin of scabies mites collected from indigenous communities across northern Australia ( unpublished data ) . Pyrethroids constitute one of the most important classes of insecticide , accounting for 17% of the world insecticide market [3] . Their intensive use worldwide over the last 30 years has led to the development of resistance in many arthropods , such that resistance now constitutes a serious threat to many programmes for control of pests and ectoparasites in agriculture , veterinary and human practice . Various resistance mechanisms such as behavioural ( avoidance of treated surfaces ) , physiological ( reduced penetration and/or increased excretion of insecticides ) , reduced sensitivity of target site ( by target alteration ) and metabolic degradation ( by hyperproduction of enzymes ) mediate pesticide resistance of arthropods . Target site insensitivity and metabolic degradation have been demonstrated to play major roles in conferring resistance to pyrethroids in some arthropods . In the German cockroach , Blattella germanica , pyrethroid resistance has been demonstrated to be mediated by multiple mechanisms including target site insensitivity , decreased cuticular penetration , and enhanced metabolism by monooxygenases and esterases [4] . The same is true for cattle ticks , a class of arthropods more closely related to scabies mite , where a point mutation in the kdr gene and metabolic degradation by esterases have both been identified as mediators of pyrethroid resistance [5] . Recently , pyrethroid resistance in Italian strains of peach potato aphid , Myzus persicae has been shown to be conferred by both mutations in the kdr gene and hyperproduction of esterases [6] . An effective way to manage pesticide resistance is the coformulation of synergist with the pesticide to counteract metabolic resistance . Synergists act by blocking metabolic pathways that would otherwise break down pesticides , thus restoring susceptibility to the agent . Piperonyl Butoxide ( PBO ) is co-formulated with insecticides such as carbaryl , methomyl , fenvalerate , permethrin , parathion , malathion and dimethoate; S , S , S-tributyl phosphorotrithioate ( DEF ) is co-formulated with malathion and permethrin; and Diethyl maleate ( DEM ) is co-formulated with parathion , malathion and dimethoate [7] . For example , PBO has been used as a synergist to pyrethrins in commercially available treatment for head lice ( caused by Pediculus humanus capitis ) to counteract resistance . Examples of over-the –counter treatment products for headlice are: Pronto ( 33% pyrethrins and 4% PBO ) ( Del Laboratories , Uniondale , NY ) , RID ( 33% pyrethrins and 4% PBO ) ( Bayer , Morristown , NJ ) , A-200 ( 33% pyrethrins and 4% PBO ) ( Hogil Pharmaceutical Corp . , Purchase , NY ) available in the US and BanLice Mousse ( pyrethrin 1 . 65 mg and 16 . 5 mg PBO ) ( Pfizer ) available in Australia [8] . In addition , insecticide synergists have been used to investigate resistance mechanisms and to identify the specific metabolic pathways targeted . Tolerance of honey bees , Apis mellifera , to pyrethroids is largely reversed by PBO and at low levels by DEF , suggesting a significant role of cytochrome P450s and a lesser role of esterases in detoxification of the chemical [9] . However , the metabolic routes blocked by synergists are not yet fully understood and maybe dependent on the species of arthropods . For example , although initial evidence in the 1970's suggested that PBO acts as a specific inhibitor of monooxygenases ( P450s ) , recently , esterases of pyrethroid resistant strains of the cotton bollworm , ( Helicoverpa armigera ) were shown to be inhibited by PBO [10] . In a study conducted with insecticide resistant strain of the German cockroach , Blatella germanica , PBO and DEF were tested as synergists to propoxur ( a carbamate insecticide ) . Both were shown to inhibit cytochrome P450 monooxygenases at different levels [11] , thus , suggesting that this enzyme family is primarily responsible for the metabolism of the pesticide . Previously , we identified a single nucleotide polymorphism ( SNP ) in the kdr gene associated with permethrin resistance in a population of scabies mites [12] . In this study , the same strain of scabies mites , Sarcoptes scabiei var canis was used to investigate the role of metabolic degradation as a contributory mechanism to observed permethrin resistance . PBO , DEF and DEM were tested as synergists to permethrin to determine if susceptibility to the acaricide can be restored , and to explore their activity in inhibiting specific metabolic pathways .
Live scabies mites ( Sarcoptes scabiei ) were collected from two sources: permethrin naïve mites ( Sarcoptes scabiei var . suis ) were collected from a colony maintained on pigs ( Kelly , A . unpublished ) while permethrin resistant mites ( Sarcoptes scabiei var . canis ) that had been maintained under permethrin treatment for many years were collected from rabbits [12] , [13] . The establishment of mite colony on pigs was approved by the Animal Ethics Committee of the Department of Primary Industries and Fisheries , Queensland adhering to the Institution's guidelines for animal research . Rabbits were maintained under a protocol approved by the Wright State University Laboratory Animal Care and Use Committee in adherence to institutional guidelines for animal husbandry . Mites were used live immediately in bioassays or collected and stored at −80°C until use in enzyme inhibition assays . Synergists used in the bioassays and biochemical assays were: Piperonyl Butoxide ( PBO- Sigma , Milwaukee , WI ) ; Diethyl Maleate ( DEM- Sigma , Milwaukee , WI ) ; and S , S , S tributylphosphorotrithioate ( DEF- ChemServices West Chester , PA ) . All other chemicals and reagents used in enzyme assays were purchased from Sigma . Reagents used in the bioassays were: 5% Permethrin ( Elimite Cream- DSM Pharmaceuticals Inc , Greenville , NC ) ; Benzyl Benzoate ( Ascabiol Lotion- Aventis Pharma Pty Ltd , Lane Cove , NSW ) ; and mineral oil ( Sigma ) . In preliminary experiments , a small number of mites ( 30/synergist ) were exposed to each synergist alone ( serially diluted starting from 300 mM ) to determine maximum concentration that the mites could tolerate and remain alive for 24 hours ( data not shown ) . The result of this preliminary experiment led us to select a concentration ( 30 mM ) of each synergist for use in subsequent bioassays . Acaricide bioassays were performed on live scabies mites as previously described [12] with some modifications . Briefly , acaricide and synergists were spread thinly on plastic petri dishes and then groups of female scabies mites were placed in the dish . Mites were exposed to 5% Permethrin alone , 30 mM of each synergist alone ( PBO , DEF and DEM ) , 30 mM of each synergist alone ( PBO , DEF or DEM ) plus 5% Permethrin , Benzyl Benzoate ( positive control acaricide ) and mineral oil ( negative control ) for 24 hours . Petri dishes containing the mites were held in a 28°C incubator at 100% relative humidity . The mites were held at 28°C because lower temperature and higher relative humidity have been found to favour survival of mites outside the host [14] . Individual mites were observed at hourly intervals for 24 hours under a microscope . The time of death , defined as complete absence of movement and a cessation of peristalsis of the gut was recorded . A total of 100 mites in each treatment group were assayed in each set of synergist bioassay . Survival times of mites in acaricide and controls were analysed using Survival Analysis in Graph Pad Prism 4 . Kaplan-Meier survival curves of mites in each treatment group were generated and statistical significance of differences in survival curves determined by logrank test . Esterase activity was determined using alpha-napthyl acetate as the substrate which forms a diazo-dye complex with Fast Blue RR Salt . Colorimetric formation of the complex is directly proportional to esterase activity and is measured continuously in a kinetic assay as described by Gunning et al , 1996 [15] with some modifications: To mite homogenate ( with and without inhibitor ) set up in triplicate , 200 ul of substrate ( 6 mg of Fast Blue dissolved in 10 ml of 0 . 2M phosphate buffer pH 6 . 0+100 ul of 100 mM alpha-naphthyl acetate in acetone ) was added . Change in absorbance was measured at 450 nm at 14 second intervals for 10 minutes in a kinetic microplate reader ( POLARstar Optima , BMG LabTech ) . As an assay control , esterase from rabbit liver , diluted 1∶10 ( ( 0 . 10 mg/ml ) and 1∶100 ( 0 . 01 mg/ml ) was included in each experiment . GST activity was measured by using monochlorobimane ( MCB ) as the substrate which forms a stable fluorescent conjugate with reduced glutathione . The enzymatic conversion of MCB to bimane-glutathione is measured in an endpoint fluorescence assay as described by Nauen and Stumpf , 2002 [16] with some modifications: To mite homogenate ( with and without inhibitor ) set up in triplicate , 200 ul of substrate ( 3 mg reduced glutathione dissolved in 10 ml of 0 . 05M Tris-HCl buffer pH 7 . 5+500 ul of 1 . 5 mg/ml of monochlorobimane ( MCB ) in methanol ) was added . Mite homogenate and substrate were incubated for 20 minutes at room temperature . Fluorescence intensity was read on a microplate reader ( Polarstar ) with 465 Emission and 390 Excitation filters . As an assay control , GST from equine liver , diluted 1∶10 ( 0 . 10 mg/ml ) and 1∶100 ( 0 . 01 mg/ml ) was included in each experiment . Cytochrome P450 activity was measured by using 7-ethoxy coumarin as the substrate which results in a fluorescent 7-hydroxycoumarin . The O-deethylation of 7-EC is measured in an endpoint fluorescent assay as described by Stumpf and Nauen , 2001 [17] with some modifications: To mite homogenate ( with and without inhibitor ) set up in triplicate , 40 ul of 0 . 1M sodium phosphate buffer pH 7 . 6 , 2 ul of 20 mM 7-EC in acetone and 10 ul of 10 mM Aqueous NADPH was added . The enzyme and substrate reaction mixture was incubated for 30 mins at 30°C with shaking . Excess NADPH was removed by adding 10 ul of 100 mM oxidized glutathione and 13 ul of 10 U/ul of glutathione reductase and the reaction incubated for 10 mins at room temperature . Fluorescence intensity was read immediately after incubation using 465 Emission and 390 Excitation filters in a microplate reader ( Polarstar ) . As an assay control , human Cytochrome P450 isoform diluted 1∶10 ( 0 . 10 mg/ml ) and 1∶100 ( 0 . 01 mg/ml ) was included in each experiment .
Survival time of both permethrin susceptible and permethrin resistant mites over the 24-hour exposure to acaricide and acaricide+/−synergist are shown in Figure 1A–C and median survival times are listed in Table 1 . Both resistant and sensitive mites ( 100% ) died within minutes of contact with benzyl benzoate ( positive control acaricide ) while 98% of mites exposed to mineral oil ( negative control ) and 97% of mites exposed to synergists alone remained alive throughout the 24-hour observation period . A significant reduction in survival time of resistant mites was observed ( p<0 . 0001 ) when exposed to permethrin in combination with PBO ( median survival = 4 hours ) as compared to survival time when exposed to permethrin alone ( median survival = 15 hours ) ( Fig . 1A ) . A similar result was observed when resistant mites were exposed to permethrin in combination with DEF ( Fig . 1B ) , with a median survival reduced from 15 to 6 hours ( p<0 . 0001 ) . Similarly , the median survival of the resistant mites was significantly reduced to 3 hours ( p<0 . 0001 ) when DEM was used as a synergist ( Fig . 1C ) . In contrast , median survival time of permethrin susceptible pig scabies mites exposed to permethrin and permethrin in combination with PBO , DEF and DEM did not show significant differences ( p = 0 . 128; p = 0 . 465; p = 0 . 220 , respectively ) . Of note , the median survival ( 4 hours ) of resistant mites in permethrin+PBO was identical to the median survival of sensitive mites exposed to permethrin alone suggesting complete reversal of the resistance phenotype using PBO as synergist . To investigate possible metabolic pathways inhibited by synergists , enzyme inhibition assays were performed . Homogenate supernatant from resistant mites were incubated with the synergists ( PBO , DEF , DEM ) acting as inhibitor and enzyme levels were compared with and without inhibitors . Homogenate supernatants from susceptible mites were also incubated with the synergists for comparison . Enzyme inhibition experiments were repeated several times and showed consistent results . Significant differences in enzyme activities with and without inhibitors were determined by t test . Esterase activity in resistant scabies mites was significantly inhibited by DEF ( 36%; p<0 . 0001 ) , and to lesser degree by PBO ( 16%; p = 0 . 006 ) . Inhibition of esterase activity in resistant mites by DEM was unremarkable ( 0 . 6%; p = 0 . 507 ) . A decrease in esterase activity was likewise observed in susceptible mites with PBO , DEF and DEM . However , the decrease was not statistically significant ( p = 0 . 337; p = 0 . 280; p = 0 . 443 , respectively ) ( Fig . 2A ) . Varying concentrations of DEF did not alter esterase activity in resistant mites ( Fig . 3 ) . Esterase activity levels in resistant mites was 7-fold ( p<0 . 0001 ) higher than in sensitive mites . GST levels of resistant scabies mites were significantly inhibited by DEM ( 73%; p<0 . 0001 ) and to lesser degree by PBO ( 18 . 0%; p = 0 . 017 ) and DEF ( 17 . 5%; p = 0 . 018 ) . PBO , DEF and DEM decreased GST activity in sensitive mites but the decrease was not statistically significant ( p = 0 . 219; p = 0 . 094; p = 0 . 065 , respectively ) ( Fig . 2B ) . A dose-dependent inhibition of GST activity by DEM was also observed in the resistant mites with maximum effect at 30 mM concentration . GST activity was observed to return to uninhibited level at 4 mM concentration of DEM ( Fig . 4 ) . A 4-fold ( p<0 . 0001 ) increase in GST levels was observed in resistant mites compared to sensitive mites . Cytochrome P450 monooxygenase activity in resistant mites was significantly inhibited by PBO ( 81%; p<0 . 0001 ) , DEF ( 65%; p = 0 . 0073 ) and DEM ( 49%; p = 0 . 0003 ) at 3mM concentration . In sensitive mites , CYP450 monooxygenase activity is significantly inhibited by DEM ( 34%; p<0 . 02 ) ; by DEF ( 24%; p = 0 . 03 ) and PBO ( 21%; p = 0 . 04 ) at the same concentration of inhibitor ( Fig . 2C ) . Results of CYP450 inhibition in resistant and sensitive mites are unremarkable at 30 mM concentration of inhibitor . A 2 . 4-fold ( p = 0 . 0002 ) increase in cytochrome P450 monoxygenase activity was observed in resistant mites compared to sensitive mites .
Detoxifying enzymes and target alteration are equally important mechanisms of insecticide degradation in ticks and mites . While a point mutation in the Vssc gene has been associated with pyrethroid resistance in two strains of cattle ticks , Boophilus microplus , high levels of esterase activity have been observed in another two strains in the absence of this mutation [5] . Resistance to tau-fluvalinate in the honeybee mite , Varroa destructor has been associated with sodium channel insensitivity as well as elevated cytochrome P450s and high levels of esterases [18] , [19] . We have previously addressed target alteration as a mechanism of pyrethroid resistance in scabies mites . A non-synonymous SNP was identified in the kdr gene associated with permethrin resistance , in the population of Sarcoptes scabiei var canis mites studied here and developed a highly sensitive high resolution melt ( HRM ) assay for its detection [12] . In this study , further investigation was done to verify if the resistance phenotype may also be mediated by detoxifying enzymes , hypothesizing that multiple mechanisms may be present as had been reported in the cattle ticks and honeybee mites . To better understand metabolic mechanisms of pyrethroid resistance in Sarcoptes scabiei mites , commonly used synergists PBO , DEF and DEM were used in acaricide bioassays . Enzyme assays were performed using three biochemical markers such as esterases , GSTs and cytochrome P450s linked to insecticide resistance to further define specific metabolic pathways inhibited in a population of resistant scabies mites . In this study , significantly higher levels of esterase , GST and cytochrome P450 monooxygenase activity in resistant mites were observed when compared to sensitive mites suggesting metabolic degradation as a mechanism of pyrethroid resistance in this population of scabies mites . It was also observed that PBO , DEF and DEM when used as synergists to permethrin in bioassays , significantly decreased the survival times of resistant mites . These findings are in agreement with results of enzyme inhibition experiments that showed inhibition of cytochrome P450 monooxygenase activity ( 81% ) by PBO , inhibition of GST activity ( 73% ) by DEM and inhibition of . esterase activity ( 36% ) by DEF when homogenate supernatant of resistant mites were incubated with the synergists . These results suggest possible roles of GSTs , esterases and cytochrome P450 monooxygenases in detoxifying permethrin in this population of resistant mites . In initial assays of cytochrome P450 we failed to detect enzyme activity in both sensitive and resistant mites using homogenates spun at high speed . However , in subsequent assays using mite homogenate spun at a much lower speed ( 1000×g ) , positive results were observed [20] . This change in methodology is based on the hypothesis that mites maybe similar to insects in that monooxygenase activity is primarily present in the microsomes; released out of the endoplasmic reticulum by mechanical homogenisation [21] and lower centrifugation allows the microsomes to stay in solution to be used in the assay . The cytochrome levels detected in resistant mites was significantly inhibited by all three synergists with the greatest inhibition achieved by PBO at a much lower concentration . These findings suggest oxidative metabolism catalysed by P450's may also be an important route for the detoxification of pyrethroids in this population of resistant mites . Overall , these findings have demonstrated potential utility of synergists in reversing resistance to acaricide and validated metabolic mechanisms of pyrethroid resistance in scabies mites . The results of this study also support our hypothesis that permethrin resistance in this population of scabies mites may be mediated by multiple mechanisms including target alteration , as demonstrated in our previous study [12] , and by metabolic degradation , as demonstrated in this study . The addition of a synergist to topical creams containing permethrin as the active ingredient may result in better clinical management of acaricide resistant scabies . Improved management of permethrin resistant head lice infestations has already been achieved by adding PBO as a synergist to pyrethrin based head lice shampoos and are now available as over the counter products [8] . This work now leads us to consider undertaking in-vitro susceptibility and pesticide synergist studies on clinical isolates from scabies-endemic communities in remote Aboriginal communities in Northern Australia where mass drug administration is underway . These studies will determine the potential role of metabolic degradation as a mechanism of permethrin resistance in human scabies mites . | Synergists are commonly used in combination with pesticides to suppress metabolism-based resistance and increase the efficacy of the agents . They are also useful as tools for laboratory investigation of specific resistance mechanisms based on their ability to inhibit specific metabolic pathways . To determine the role of metabolic degradation as a mechanism for acaricide resistance in human scabies , PBO ( piperonyl butoxide ) , DEF ( S , S , S-tributyl phosphorotrithioate ) and DEM ( diethyl maleate ) were used with permethrin as synergists in a bioassay of mite killing . A statistically significant difference in survival time of permethrin-resistant Sarcoptes scabiei variety canis was noted when any of the three synergists were used in combination with permethrin compared to survival time of mites exposed to permethrin alone ( p<0 . 0001 ) . These results indicate the potential utility of synergists in reversing tolerance to pyrethroid-based acaricides ( i . e . the addition of synergists to permethrin-containing topical acaricide cream commonly used to treat scabies ) . To further verify specific metabolic pathways being inhibited by these synergists , enzyme assays were developed to measure esterase , glutathione S-transferase ( GST ) and cytochrome P450 monooxygenase activity in scabies mites . Results of in vitro enzyme inhibition experiments showed lower levels of esterase activity with DEF; lower levels of GST activity with DEM and lower levels of cytochrome monooxygenase activity with PBO . These findings indicate a metabolic mechanism as mediating pyrethroid resistance in scabies mites . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"chemistry/biochemistry",
"dermatology/skin",
"infections"
] | 2009 | The Effect of Insecticide Synergists on the Response of Scabies Mites to Pyrethroid Acaricides |
Inferring the combinatorial regulatory code of transcription factors ( TFs ) from genome-wide TF binding profiles is challenging . A major reason is that TF binding profiles significantly overlap and are therefore highly correlated . Clustered occurrence of multiple TFs at genomic sites may arise from chromatin accessibility and local cooperation between TFs , or binding sites may simply appear clustered if the profiles are generated from diverse cell populations . Overlaps in TF binding profiles may also result from measurements taken at closely related time intervals . It is thus of great interest to distinguish TFs that directly regulate gene expression from those that are indirectly associated with gene expression . Graphical models , in particular Bayesian networks , provide a powerful mathematical framework to infer different types of dependencies . However , existing methods do not perform well when the features ( here: TF binding profiles ) are highly correlated , when their association with the biological outcome is weak , and when the sample size is small . Here , we develop a novel computational method , the Neighbourhood Consistent PC ( NCPC ) algorithms , which deal with these scenarios much more effectively than existing methods do . We further present a novel graphical representation , the Direct Dependence Graph ( DDGraph ) , to better display the complex interactions among variables . NCPC and DDGraph can also be applied to other problems involving highly correlated biological features . Both methods are implemented in the R package ddgraph , available as part of Bioconductor ( http://bioconductor . org/packages/2 . 11/bioc/html/ddgraph . html ) . Applied to real data , our method identified TFs that specify different classes of cis-regulatory modules ( CRMs ) in Drosophila mesoderm differentiation . Our analysis also found depletion of the early transcription factor Twist binding at the CRMs regulating expression in visceral and somatic muscle cells at later stages , which suggests a CRM-specific repression mechanism that so far has not been characterised for this class of mesodermal CRMs .
A major area in genome research is understanding how the regulatory information is encoded . Work over the past few decades has resulted in the notion of a combinatorial regulatory code: the concerted binding of a context-specific set of transcription factors ( TFs ) to regulatory sequences , which is crucial for proper gene expression . Studies of a handful of single genes and their few well-characterised enhancers prevailed in the early days ( see [1] for review ) . The traditionally experimental dissection of enhancers allowed the placing of TFs within a regulatory hierarchy . A canonical example of this traditional dissection is the identification of the various stripe enhancers of the Drosophila even-skipped gene that respond to different TFs involved in early patterning ( see [2] , [3] for review ) . With the advent of genome-wide detection methods , hundreds of genome-wide TF binding and histone modification profiles have been generated [4]–[6] with the aim of deciphering the combinatorial regulatory code at the global level . Whereas the inference of the regulatory code may greatly benefit from having additional data , such as the expression patterns of the genes of interest under mutant conditions , it is often difficult to collect at the genome level . In the absence of such additional data , a typical strategy is to assume that correlation in TF binding indicates functional interaction between TFs , and to perform correlation-based analyses , such as enrichment analysis ( see [7] for a review of strategies in analysing multiple TF binding profiles ) . However , recent studies provide evidence for so-called “hotspots” to which many interacting or non-interacting TFs may bind [6] , [8] , [9] , which leads to high correlations among binding profiles of both functionally “relevant” and functionally “irrelevant” TFs . It remains a significant challenge to distinguish relevant and important TFs from the others in the understanding of the combinatorial regulatory code . Similar to the gene regulation problem described above , many other biological problems involve highly-correlated features and high correlation does not necessarily indicate functional relevance . Machine learning approaches , especially classification methods , have been developed to use the measurements of these features ( or “explanatory variables” ) to predict biological outcomes ( or “target variables” ) , e . g . using core promoter DNA motifs to predict transcription start site locations [10] or using DNA motifs and transcript structures to predict splicing patterns [11] . Although these approaches may produce robust predictions , they do not distinguish which features directly or indirectly influence the biological outcome . Other machine-learning approaches such as standard feature selection methods ( see [12] for review ) are also not appropriate for this kind of inference in the general case [13] , [14] . In contrast , graphical models ( GM ) [15] encompass a broad class of tools that infer the joint probability distribution of the variables in the network ( or graph ) , and distinguish direct from indirect interactions under broad assumptions . Graphical models achieve this distinction through the notion of conditional independence , which is explained in the Results section . Bayesian networks , also known as Directed Acyclic Graphs ( DAGs ) , are a type of graphical model that further permit the interpretation of causality of the inferred interactions . Two concepts are particularly important in the theory of Bayesian networks: the causal neighbourhood and the Markov blanket . Specifically , if there is a directed edge from variable A to the target variable T in the network , then variable A is defined as the causal parent of T . If the directed edge goes from T to A , then A is the causal child of T . The causal neighbourhood of the target variable consists of the causal parents and causal children of the target variable . It is thus the set of variables that are most “causally immediate” for the target variable . The Markov blanket of the target variable T contains its causal neighbourhood as well as other causal parents of T's causal children ( these other causal parents are T's causal spouses ) . From the information-theoretical perspective , the Markov blanket contains all the information about the target variable [15] , [16] . In terms of statistical inference , existing algorithms for inferring Bayesian networks can be broadly classified into constraint-based , score-based and hybrid algorithms [17] . Constraint-based algorithms perform statistical tests for conditional independence , whereas score-based algorithms estimate the most ( or highly ) likely joint distribution of the variables in the network . Hybrid algorithms are a combination of the other two , initialising a score-based search with a network inferred by a constraint-based algorithm . In this paper we develop a novel constraint-based graphical model method , the Neighbourhood Consistent PC ( NCPC ) algorithms , to infer the causal neighbourhood and the Markov blanket of a target variable . Through synthetic data , we demonstrate that our algorithm has superior performance to existing algorithms when the variables are highly correlated , the data of the target variable is sparse , and the coupling of the target variable and other variables is weak . We also develop a novel graphical representation , the Direct Dependence Graph ( DDGraph ) , which can represent the dependence patterns inferred from the NCPC algorithms . This representation is broader than the common representation in DAGs , and is useful for exploratory analyses of NCPC results . In particular , the DDGraph shows the conditional independencies in the data even if the underlying network is cyclic or non-faithful to a DAG . Both NCPC and DDGraph are implemented in the R package ddgraph , which is part of Bioconductor ( http://bioconductor . org/packages/2 . 11/bioc/html/ddgraph . html ) . Applying our algorithm to genome-wide TF profiles and expression profiles of cis-regulatory modules ( CRMs ) published in [18] provides novel insight into the transcriptional regulation during mesoderm differentiation in Drosophila embryonic development . We identify not only known TFs that are relevant for specific CRM classes , but also a potentially CRM-specific repression mechanism that has not been suggested before . Although we focus on gene regulation in our paper , our algorithm is applicable to other scenarios discussed earlier that involve highly correlated biological features .
We illustrate the concepts of direct and indirect dependencies in terms of the combinatorial binding code of transcription factors . Our aim is to identify transcription factors that directly influence the regulatory output of a set of CRMs . Consider the following example . Transcription factor A binds to the CRM of a number of genes and thus directly regulates these target genes , whereas transcription factor B binds to several CRMs where A also binds ( perhaps because of chromatin structure ) , but does not regulate the target genes of A . Therefore , A and B have overlapping binding profiles , and both appear to be associated with gene expression changes . However , the apparent effect of B can be explained away by the effect of A . This means that , if we divide the CRMs into those bound by A and those not bound by A , the binding of B is not associated with gene expression changes in either group . Mathematically speaking , B and the genes are conditionally independent given A , suggesting that the effect of B is at most indirect . In contrast , if we divide the CRMs into those bound by B and those not bound by B , the binding of A is still associated with gene expression changes in either or both groups . Mathematically speaking , A and the genes are dependent given B , suggesting that the effect of A is direct . Detecting conditional independence is thus central in separating direct from indirect effect [15] . Incidentally , when we consider all the CRMs together , both A and B can be associated with ( or equivalently , marginally dependent with ) the genes . Below we formally define the types of statistical dependencies our NCPC algorithm and its extension detect . We use Xi , a binary vector , to represent the binding states of the i-th TF at a set of CRMs . We use T , also a binary vector , to represent the expression states of the genes with which the CRMs are associated . We denote the set of all m TF binding profiles as , such that . As mentioned in Introduction , T is the target variable or outcome , and the Xs are the explanatory variables or features . Consistent with standard notation , we use symbol to represent “marginally independent” , and symbol to represent “marginally dependent” . We also use symbol | to represent “conditioning on” . Bold capital letter S indicates a subset of , whereas S ( Xi ) indicates a subset of that does not include Xi , i . e . , . Here we present two versions of the Neighbourhood Consistent PC ( NCPC ) algorithm , which are based on the PC algorithm [15] . Similar to the PC algorithm ( see Supplementary Text ) , our algorithms perform a series of statistical tests on each explanatory variable to select variables in direct , conditional and indirect dependencies to target T . More importantly , our algorithms detect these dependencies even when the explanatory variables X are highly correlated among themselves . For example , consider the case where two highly correlated variables Xi and Xj both have direct or conditional dependence with the target variable T . However , when testing the null hypothesis of Xi ( or Xj ) and T being independent given Xj ( or Xi ) for data with a finite sample size , we may not reject this null hypothesis for a given confidence level . Thus , both Xi and Xj may be discarded during the selection procedure . Indeed , the original PC algorithm discards such variables , leading to a low accuracy rate in these scenarios ( see Section “Comparison with other algorithms on synthetic data” ) . To account for potential correlation among variables X , our NCPC algorithms specifically check for and retain pairs of variables with the two patterns described below . These patterns depend on the type I error rate α of the statistical test used in the algorithm . NCPC and NCPC* output labels for the explanatory variables X . These labels are the inferred types of dependence , namely “direct” , “indirect” and “joint” , as defined in Definitions 1–3 , and the candidate dependency patterns , namely “conditional” and “conditional joint” , as described in Candidate Patterns 1–2 . To visualise the inferred dependencies between multiple explanatory variables and the target variable , and especially to represent Candidate Patterns 1–2 , we develop a novel graphical representation: the Direct Dependence Graph ( DDGraph ) . DDGraphs use both directed edges ( ending in dots ) and undirected edges to capture a multitude of dependency patterns with respect to the target variable T ( see Figure 1 for the graphical vocabulary ) . For example , directed edge Xi –• Xj represents that Xj is conditionally independent of T given Xi . Solid undirected edge Xi–Xj represents that Xi and Xj are both dependent given T and marginally dependent . Dashed undirected edge Xi - - Xj represents that Xi and Xj are conditionally independent given T . Additionally , black edges indicate dependence patterns that are mathematically consistent , and grey edges indicate the dependence patterns that are inconsistent ( e . g . edges in Candidate Patterns 1 and 2 ) . A DDGraph and a DAG with the same dependence patterns around the target variable T is shown in Figure 2A . In a DDGraph , variables connected to the target variable T with an undirected edge are in the causal neighbourhood of T , and variables reachable from T by traversing only undirected edges are in the Markov blanket of T ( Figure 2A ) . These variables are also easily recognizable with their oval shapes , whereas variables in indirect dependence with T have a rectangular shape . By contrast , the causal neighborhood and Markov blanket in a DAG have to be inferred from the direction of the edges ( Figure 2A ) . A DDGraph also represents joint and conditional joint dependency patterns , which are mathematically inconsistent and thus impossible to represent with DAGs ( Figure 2B ) . Indeed , DAGs , as well as other factorization-based graphs , such as factor graphs [23] , that represent a factorization of a joint probability distribution cannot represent these inconsistent dependency patterns . We generated synthetic data based on the 15 correlated TF binding profiles in [18] . See Materials and Methods for details on data generation . The target variable T , which is a binary vector that contains the expression states of a set of CRMs , is sparse: similar to the real data , only around 10% of CRMs show class-specific expression . We generated data for three sample sizes: 300 , 500 and 1000; the sample size in the data of [18] is 310 . In addition , we simulated a causal neighborhood of two variables ( X1 , X2 ) for T , and these causal neighbors are weakly correlated with T ( correlation 0 . 17–0 . 25 ) . We simulated data with four levels of correlation between the two causal neighbors: no correlation ( 0 ) , weak correlation ( 0 . 25 ) , strong correlation ( 0 . 50; similar to the average correlation of 0 . 46 we found in the data from [18] ) , and very strong correlation ( 0 . 75 ) . We further introduced a third variable ( X3 ) as the confounding variable in the network and generated correlated data for two realistic scenarios: With these synthetic data , we focus on the performance of separating direct from indirect dependence and detecting the causal neighbourhood . We applied our NCPC and NCPC* algorithms , at an α level of 0 . 05 , to these data . Of the constraint-based algorithms , multiple testing correction has been mathematically and empirically demonstrated only for the PC algorithm [21] , [24] , [25] , therefore , for fair comparison we applied all algorithms , including NCPC/NCPC* , without any multiple testing correction . To investigate the effectiveness of identifying pairs of variables in Candidate Patterns 1 and 2 ( see Section “Neighbourhood Consistent PC algorithms” ) , we applied the NCPC algorithm in two ways: detecting variables only in direct dependence with the target variable , and in addition detecting pairs of variables in joint dependence ( Candidate Pattern 1 ) . Similarly , we applied the NCPC* algorithm in two ways: detecting variables only in direct and conditional dependence with the target variable as well as pairs of variables in joint dependence , and detecting , in addition , pairs of variables in conditional joint dependence ( Candidate Pattern 2 ) . For comparison , we also applied the following algorithms to the synthetic data: the original PC algorithm [15]; score-based algorithms that infer the whole network , such as Hill-climbing with BIC penalization [26] or with a Dirichlet prior ( BDe penalization [27] ) ; other constraint-based algorithms that infer the local structure , such as IAMB [28] , FastIAMB [29] , InterIAMB [29] and MMPC [30]; as well as a hybrid algorithm MMHC [30] . We measured the proportion of correct predictions from these algorithms over 1000 data sets generated for each combination of the sample size and correlation in either of the two scenarios . A prediction is correct when only the two causal neighbors and no other variables are identified . These prediction rates for the “Time” scenario are summarized in Figure 4 . The prediction rates for the “Hidden” scenario are similar and are summarized in Supplementary Figure S1 in Text S1 . Identifying variables in direct dependence and in joint dependence , the NCPC algorithm ( “NCPC dir+jnt” ) , has the highest ( accounting for variation in simulated data ) rate of correct predictions amongst all the algorithms in all the cases examined here , except in the biggest dataset with 0 correlation . This superior performance is particularly notable when the correlation between the variables is high and the dataset is small . By including the variable pairs in joint dependence , “NCPC dir+jnt” achieves better performance “NCPC dir” because this inclusion drastically improves recall ( corresponding to low false negative rates ) , especially when the sample size is not large , although the inclusion lowers precision ( corresponding to high false positive rates ) slightly ( see rates of precision and recall defined in Materials and Methods and computed in Supplementary Figures S2 and S3 in Text S1 ) . The comparison of the two implementations of the NCPC algorithm provides some empirical evidence for including pairs of variables at least in Candidate Pattern 1 as candidates for direct dependence . With the sample size as large as 1000 , the data are informative enough for “NCPC dir” to perform similarly or even slightly better than “NCPC dir+jnt” . The performance of the two implementations of the NCPC* algorithm , however , is worse than the NCPC algorithm in most cases . This is likely because in order to identify the Markov blanket , which is larger than the causal neighbourhood , the NCPC* algorithm sacrifices the false positive rates more to gain even lower false negative rates . At different levels of correlation , the NCPC and NCPC* algorithms both have more stable precision and recall rates than other algorithms ( Supplementary Figures S2 and S3 in Text S1 ) . This may explain why the NCPC and NCPC* algorithms ( four implementations ) perform better than all the other algorithms . Increasing the sample size improves the prediction for most algorithms , as we expected . However , when the correlation in the data is 0 . 75 , the NCPC and NCPC* algorithms have lower rates of correct predictions for data with a sample size of 500 than for data with a sample size of 300 . This may be due to the α level chosen for the statistical test before running the algorithm , especially when the P-values obtained by the NCPC and NCPC* algorithms are close to the value of α . A more stringent α level such as 0 . 01 leads to improved performance ( Supplementary Figures S4 and S5 in Text S1 ) . This highlights the importance of choosing an appropriate α value , and suggests re-running the algorithm with a different α level if the P-values obtained are close to the initial α value . We recommend the user to inspect the P-values of key conditional independence tests that give rise to the DDGraph and to change the α value accordingly . Zinzen et al . [18] published an in vivo ChIP-chip temporal binding profiles of key transcription factors that are involved in mesoderm development in fly embryos , as well as the CRM Activity Database ( CAD ) , the largest such database thus far , which contains tissue-specific temporal expression patterns driven by these CRMs . The five key TFs , Twist ( Twi ) , Myocyte enhancer factor 2 ( Mef2 ) , Tinman ( Tin ) , Bagpipe ( Bap ) , and Biniou ( Bin ) , were each measured in some or all of five developmental stages , producing 15 correlated TF binding profiles ( Figure 5 ) . Zinzen et al . further focused on 310 CRMs from the CAD that have both TF binding and expression data , and classified these CRMs into five classes based on their tissue-specific expression patterns: mesodermal ( Meso ) , mesodermal and somatic muscle ( Meso&SM ) , visceral muscle ( VM ) , visceral and somatic muscle ( VM&SM ) and somatic muscle ( SM ) . Here we applied the NCPC and NCPC* algorithms to the same 310 CRMs with the 15 TF binding profiles . The advantage of this dataset is that any computational predictions can be benchmarked against a wealth of previously established biological results . At an α level of 0 . 05 , we identified expression class-specific causal neighbourhoods using NCPC ( Figure 6 and Supplementary Figure S6 in Text S1 ) . The Markov blankets identified by applying NCPC* ( Supplementary Figure S7 in Text S1 ) are similar to their corresponding causal neighbourhoods . We discusss the biological implications of our inference in the next section . We also applied other algorithms benchmarked in the previous section to this data set . Hill-climbing with BIC identified a smaller but overlapping set of variables ( Supplementary Figure S8 in Text S1 ) , consistent with our results on synthetic data that this algorithm has higher precision but a lower recall rate than our NCPC algorithms . Hillclimbing with BDe identified a bigger but overlapping set of variables ( Supplementary Figure S9 in Text S1 ) , also consistent with our results on synthetic data that this algorithm has lower precision than but a similar recall rate to our NCPC algorithms . The IAMB family of methods found either a smaller but overlapping set of variables , or no variables ( Supplementary Figures S10 , S11 and S12 in Text S1 ) . The original PC algorithm performed similarly to the IAMB methods ( Supplementary Figure S13 in Text S1 ) . MMHC produced similar results to those from the ordinary hill-climbing method ( Supplementary Figures S14 and S15 in Text S1 ) . We applied our method on the dataset of early mesoderm development in the Drosophila embryo [18] . The five transcription factors Twi , Tin , Mef2 , Bin and Bap have been previously implicated in mesoderm development of the fly . Among the five TFs we analysed here , Twi , together with another TF Snail , is the earliest marker of mesoderm and is required for mesoderm formation [31] . Tin , a direct target of Twi , is crucial for the differentiation of heart , somatic and visceral mesoderm and is present also in dorsal somatic muscle precursor cells [32] . Mef2 , crucial for early muscle differentiation , is present in both visceral and somatic muscle [33]–[35] . Activated by Tin , Bap specifies cells that become the visceral muscle [36] , [37] . Finally , Bin is expressed only in visceral muscle cells and is crucial for their differentiation [38] . After identifying the causal neighbourhood , we further examined which specific TF combinations are enriched or depleted in each of the five expression classes , compared with the rest of the 310 CRMs analysed here ( Figure 7 ) . Most of these TF combinations have been established in single-gene studies: In addition to previously established regulatory principles outlined above , the genome-wide statistics also suggest a thus far uncharacterized mechanism of prevention of early Twi binding at 2–4 h of embryogenesis for the class of CRMs active in visceral and somatic muscle ( VM&SM ) at 8–12 h of development . This suggests that these CRMs are selectively shut off during early embryogensis , but are bound later on by tissue-specific transcription factors: Twi 2–4 h is identified to also have direct dependence with this VM&SM CRM class ( Figure 6 ) , and , interestingly , it is the lack of Twi binding at 2–4 h that is significantly associated with the activity of this CRM class . Note that this observation is consistent with the negative correlations between the binding profiles ( over all 310 CRMs ) of Twi 2–4 h and both Bin and Mef2 at later stages ( Figure 5 ) . However , it is unclear how depletion of Twi at an earlier stage leads to activity of these CRMs several hours later . One plausible biological explanation is that these CRMs may be silenced during early embryogenesis ( for example , the chromatin they are located in is inaccessible during this stage ) , and be bound by tissue-specific TFs , such as Bin and Mef2 , later . Activation of the CRMs in this class may require concerted efforts , which may be specific to this CRM class , and which may involve remodelling of chromatin or inhibition of early Twi binding . It is also unclear whether additional transcription factors or chromatin remodelling factors are involved in the activation of this CRM class .
In this paper we present a novel graphical model-based method that distinguishes direct from indirect dependencies between explanatory variables ( or features ) and the target variable . Our NCPC and NCPC* algorithms work particularly well in cases of highly correlated features and of sparse or weak signals , as seen in comparison with other algorithms on synthetic data . We applied our algorithms to data published in [18] , which consist of the 15 transcription-factor binding profiles over 310 CRMs in Drosophila during mesoderm development . Our analysis identified known combinations of TFs associated with expression of different CRM classes . Our analysis also suggests an uncharacterized repression mechanism: depletion of Twist binding at 2–4 h plus presence of tissue-specific factors Mef2 and Bin indicates activity of the CRMs in the visceral and somatic muscle development , through CRM silencing in early embryogenesis and/or chromatin remodelling . Additional TFs may be involved in mesodermal development , and our algorithms can be easily applied to newly available data [40] to improve the local network structures we identified here . Our NCPC algorithms assume no hidden variables in the Markov blanket of the target variable . This assumption is frequently not met in reality; for example , in the case of the transcriptional regulation , a number of relevant TFs might not have been measured . In that case , a seemingly irrelevant TF might be inferred as a causal neighbour if it is correlated with the unmeasured relevant TF ( e . g . due to open chromatin structure ) . Such a TF would be a “proxy” for the binding of the relevant TF . Our NCPC algorithms also assume no feedback loops in the Markov blanket of the target variable . This may not be the case in a real biological system . However , if time course data are available and informative enough such that the underlying Markov blanket is acyclic at each time point , then our NCPC algorithms can still be applied ( similar to the way we re-analysed the fly mesoderm development data ) to identify causal neighbours . Transcriptional responses are typically slow ( on the order of minutes [41] ) which allows for the data to be collected as time series so that the next time point is a product of the previous time point and thus the dynamics made acyclic in time . The statistical tests our algorithms perform for the variables in these systems tend to be highly dependent . It is still a challenge to control the false discovery rate for highly dependent tests . We implemented the multiple testing procedure of [21] for controlling the false discovery rate ( see Supplementary Text for detail ) . However , we found that this procedure can be overly conservative and can lead to loss of statistical power , for example even at 0 . 3 FDR the somatic muscle ( SM ) class has no causal neighbours ( data not shown ) although in-vivo validation found a weak but predictive signal [18] . Further development in controlling the FDR for dependent tests in network inference is needed . The NCPC algorithms infer the causal neighbourhood and do not optimise the prediction accuracy of the target variable . Hence , we do not expect these algorithms to be an optimal feature selection procedure for classification . Nonetheless , the NCPC algorithms may in principle be used for feature selection to improve prediction accuracy , for example , by using cross-validation to choose a P-value threshold that minimises the cross-validation error . Directly incorporating the dependence structure in a classifier is still challenging , since it is difficult to robustly estimate higher-order conditional probabilities from small datasets ( a Naive Bayesian Classifier has been used in practice; see [42] ) . A wealth of genome-wide data have been and are currently produced , featuring binding sites of transcription factors , chromatin marks and RNA levels [6] , [9] , [43] . Our NCPC algorithms can be applied to tackle more effectively the high correlations that have been noted among these features [44] and uncover the underlying combinatorial code specific to a set of regulatory sequences of interest . However , before the NCPC algorithm can be used on genomewide data , technical artefacts ( e . g . systematic biases in reporter assays or tested enhancers ) need to removed and biases in the data accounted or corrected for , otherwise they might lead to spurious associations [45] , [46] . Although we have focused on TF binding and CRM activity in this paper , our NCPC algorithms are applicable to other biological problems involving possible highly correlated features . For instance , high-throughput imaging of knock-down strains can produce large sets of highly correlated visual features describing cell shape [47]–[49] . Our NCPC algorithms can be applied to explore the relationships between these visual features and the genes knocked down , or between these features and characteristics ( e . g . , elongation ) of the cells involved . Similarly , genome-wide RNAi screens with multiple classes of phenotypic readout [50] , [51] might produce features ( phenotypes ) that are highly correlated , in addition to features of gene functional and spatial/temporal annotation . In the ideal case , we can find out if a phenotype is a consequence of another phenotype or any of the gene features . Dissecting direct and indirect effects in these highly correlated datasets would provide further valuable insight into the underlying biological mechanisms . A unified interface to all causal neighbourhood/Markov blanket methods benchmarked in this paper , including the NCPC/NCPC* algorithms and the DDGraph representation , is available as the R package ddgraph , which is part of Bioconductor ( http://bioconductor . org/packages/2 . 11/bioc/html/ddgraph . html ) .
We used the data from Supplementary Figure 8 of [18] . These data include 5 TFs previously implicated in development of mesoderm during D . melanogaster embryogenesis: Twist ( Twi ) , Tinman ( Tin ) , Myocite enhancing factor 2 ( Mef2 ) , Biniou ( Bin ) and Bagpipe ( Bap ) . Their binary occupancy at 310 CRMs were measured in some or all of 5 stages , leading to 15 binding profiles . Their pairwise correlations are displayed in Figure 5 . The data also contain the in vivo-tested expression patterns of the 310 CRMs . Most of these ( 210 ) did not show expression in the mesoderm , but showed expression in other tissues during embryogenesis . Out of the 100 that did show mesodermal expression , they were classified in 6 broad categories based on expression in specific tissues: Mesodermal ( Meso ) , Mesodermal and Somatic Muscle ( Meso&SM ) , Visceral Muscle ( VM ) , Visceral and Somatic Muscle ( VM&SM ) , Somatic Muscle ( SM ) and Cardiac Muscle ( CM ) . We focused on the first five in our analysis , like in the original paper . To construct the synthetic dataset we used Hill-climbing with BIC to infer a Bayesian network from the real biological dataset ( [18]; see the previous section ) . We estimated the mean number of causal parents per node to be roughly 1 . 5 and the maximum to be 2 . We therefore assumed a binomial distribution for the number of causal parents . We used a beta distribution to generate the probabilities in the conditional probability table associated with each node . With these distributions we generated a network structure that had both marginal probabilities and pairwise correlations similar to the real data . We used this network structure to generate binary data for 15 nodes in the network , which is the number of TF binding profiles in the real data . The target variable is generated separately using a noisy AND function . To generate the CRM class target variables we considered a causal neighbourhood of size 2 and used a noisy AND function , representing the simplest combinatorial code of 2 TFs . The noise in the AND function is incorporated into both the inputs and the output of the function . The noise in the inputs models the activity of other TFs , which might , for example , inhibit the CRM activity in the presence of the TF , or activate the CRM in the absence of the TF . The noise in the output models the noise in the reporter assay used to find the activity of a CRM . Let F ( RA , RB ) be a boolean AND function with two inputs . Thus F ( RA , RB ) = 1 only if RA = RB = 1 . Further , let A and B denote the real functional binding profiles of two TFs that constitute the combinatorial code . The noise at the input of the boolean AND function can be modelled by “readout” probabilities: output = F ( RA , RB ) · P ( RA|A ) · P ( RB|B ) . If we assume that the conditional probabilities have the same distribution for A and B: P ( RA|A ) = P ( RB|B ) , then we just need to specify two readout probabilities . We set these to be P ( RA = 1|A = 1 ) = 0 . 5 and P ( RA = 1|A = 0 ) = 0 . 1 . At the output of boolean AND function , we use a false positive rate of 0 . 01 and false negative rate of 0 . 2 . This parameter setting results in 10% of the CRMs being active , similar to the Zinzen et al . data . Furthermore , the data generated for these CRMs from the noisy AND function is weakly correlated ( correlation between 0 . 17 and 0 . 25 ) with A and B . This level of correlation is also similar to the observed correlations between the CRM classes and TF binding profiles in the real data . To incorporate the two scenarios “Time” and “Hidden” described in the main text , we randomly chose three variables in each simulated network . We then rewired these three variables to match each scenario . For the “Time” scenario we allowed for the first variable to have causal parents as in the unmodified network , while variables two and three have causal parents only from the scenario . However , they retained the original causal children of the unmodified network . This ensured that we can fully control the correlation between these three variables , but also leave it as much as possible in the context of rest of the network . In the “Hidden” scenario , we generated an additional hidden variable and made it a causal parent for the three variables in the scenario . Now the three variables only retained their original causal children , but not their causal parents . To generate the binary profile of the target variable , we applied the noisy AND function as before . The hill-climbing and IAMB algorithms were applied using the bnlearn R package , and PC algorithm was applied using the pcalg R package . Both can be accessed using a unified interface in our R package ddgraph . For NCPC and NCPC* we used the Monte-Carlo chi-square test , while for the IAMB algorithms we used the Mutual Information test recommended by the authors [28] , but with Monte Carlo-calculated P-values due to small sample sizes . We compared these two tests in a simple case of two variables and found that the Monte-Carlo chisquare test was slightly better than the Monte-Carlo Mutual Information test . However , their differences were not noticeable when applied to our synthetic data . For MMHC we use the default constraint-based algorithm ( MMPC ) . To assess the performance of the algorithms , we defined a prediction as correct if there are no false positive and no false negatives . The accuracy was measured by the prediction rate , which was the proportion of correct predictions over all the synthetic networks . We also defined precision as TP/ ( TP+FP ) , where TP is the number of true positives , and FP is the number of false positives . Additionally , we defined recall as TP/ ( TP+FN ) where FN is the number of false negatives . Rates of precision and recall were also averaged over all the synthetic networks . As the size of the conditioning set increases , the power of the test decreases . To increase power , we limited the total count l of datapoints per conditioning set to 10 . Our NCPC and NCPC* algorithms performed the test if this requirement was met and considered the variables to be dependent otherwise . Alternatively , one may constrain the size of the conditioning set . Since our data are binary , we set the maximal size of the conditioning set k to , where Tmin is the smaller of the number of ones and the number of zeros in T . We found that these two rules performed similarly on our binary data . The second rule , however , may also be applied to continuous features with a binary target variable . For n TFs , each of which is either present or not at a CRM , we performed Fisher's exact test to test whether a combination of presence and absence of these TFs is statistically significantly associated with a CRM class . This test essentially compares the frequencies of the combination within this CRM class and across the other four classes . We applied the Benjamini-Hochberg correction [19] , which adjusts the P-values to control the False Discovery Rate ( FDR ) , and retained those combinations with adjusted P-values smaller than 0 . 15 . | Transcription factors ( TFs ) are proteins that bind to DNA and regulate gene expression . Recent technological advances make it possible to map TF binding patterns across the whole genome . Multiple single-gene studies showed that combinatorial binding of multiple transcription factors determines the gene transcriptional output . A common naive assumption is that correlated binding profiles may indicate combinatorial binding . However , it has been found that many TFs bind to distinct hotspots whose role is currently unclear . It is thus of great interest to find transcription factor combinations whose correlated binding is causally most immediate to gene expression . Building upon theories of statistical dependence and causality , we develop novel graphical modelbased algorithms that handle highly correlated transcription factor binding profiles more efficiently and reliably than existing algorithms do . These algorithms can also be applied to other biological areas involving highly correlated variables , such as the analysis of high-throughput gene knock-down experiments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"algorithms",
"genome",
"analysis",
"tools",
"computer",
"science",
"genomics",
"genetic",
"networks",
"functional",
"genomics",
"regulatory",
"networks",
"biology",
"computational",
"biology",
"computing",
"methods",
"computer",
"inferencing",
"genetics",
"and",
"genomic... | 2012 | A Graphical Modelling Approach to the Dissection of Highly Correlated Transcription Factor Binding Site Profiles |
Human cytomegalovirus ( HCMV ) is an important , ubiquitous pathogen that causes severe clinical disease in immunocompromised individuals , such as organ transplant recipients and infants infected in utero . Antiviral chemotherapy remains problematic due to toxicity of the available compounds and the emergence of viruses resistant to available antiviral therapies . Antiviral antibodies could represent a valuable alternative strategy to limit the clinical consequences of viral disease in patients . The envelope glycoprotein B ( gB ) of HCMV is a major antigen for the induction of virus neutralizing antibodies . However , the role of anti-gB antibodies in the course of the infection in-vivo remains unknown . We have used a murine CMV ( MCMV ) model to generate and study a number of anti-gB monoclonal antibodies ( mAbs ) with differing virus-neutralizing capacities . The mAbs were found to bind to similar antigenic structures on MCMV gB that are represented in HCMV gB . When mAbs were used in immunodeficient RAG-/- hosts to limit an ongoing infection we observed a reduction in viral load both with mAbs having potent neutralizing capacity in-vitro as well as mAbs classified as non-neutralizing . In a therapeutic setting , neutralizing mAbs showed a greater capacity to reduce the viral burden compared to non-neutralizing antibodies . Efficacy was correlated with sustained concentration of virus neutralizing mAbs in-vivo rather than their in-vitro neutralizing capacity . Combinations of neutralizing mAbs further augmented the antiviral effect and were found to be as potent in protection as polyvalent serum from immune animals . Prophylactic administration of mAbs before infection was also protective and both neutralizing and non-neutralizing mAbs were equally effective in preventing lethal infection of immunodeficient mice . In summary , our data argue that therapeutic application of potently neutralizing mAbs against gB represent a strategy to modify the outcome of CMV infection in immunodeficient hosts . When present before infection , both neutralizing and non-neutralizing anti-gB exhibited protective capacity .
Human cytomegalovirus ( HCMV ) is an important and ubiquitous human pathogen that is found throughout all geographic locations and socioeconomic groups . Initial infection with HCMV is followed by life-long persistence characterized by episodes of periodic reactivation . While most infections are subclinical in the immunocompetent host , HCMV can cause severe disease and death in immunocompromised patients and newborns infected in utero . As such HCMV is the most frequent viral cause of congenital infection and affects 0 . 5–2% of all live births worldwide [1 , 2] . It is the leading infectious cause of childhood sensorineural hearing loss and an important cause of mental retardation [3] . In addition , HCMV is a major cause of morbidity and mortality in recipients of solid organ or stem cell transplants in both the early and late transplant period and is thought to contribute to graft dysfunction leading to graft loss late after transplantation and to overall decreased long term survival in transplant recipients [4 , 5] . Prevention of end-organ disease and treatment of clinical disease in transplant patients has been achieved using antiviral chemotherapy , although toxicity associated with these compounds and emergence of viruses resistant to currently available antiviral therapies continue to represent a challenge in the clinical care of these patients [6] . In congenitally infected infants , treatment with antivirals has shown some benefit in the most severely affected infants , but the relative benefit of this treatment and the considerable short term and unknown long term toxicity of these agents has resulted in very restricted recommendations for their use [7] . Therefore , prophylactic vaccination has been long argued to be the preferred approach for prevention of HCMV infection and disease in risk groups . However , a prophylactic vaccine remains elusive [8] not least because the nature of the protective immunity to HCMV is far from understood . In general , induction of virus-neutralizing antibodies has been shown to represent a correlate of protection for most effective antiviral vaccines and HCMV will likely be no exception [9] . Thus , identification of major targets for the neutralizing antibody response and characterization of the mode of protection by these antibodies will represent a major step towards development of an effective anti-HCMV vaccine . Within the envelope of HCMV two proteins or protein complexes have been identified as being the most important targets for the neutralizing antibody response: glycoprotein ( g ) B and the gH-containing complexes ( gH/gL/gO and gH/gL/UL128/UL130/UL131A ) [10] . The gH containing protein complexes have recently received attention as potential vaccines since the UL128/UL130/UL131A components of the gH-pentamer complex induce extremely potent neutralizing antibodies during infection [11] . However , these antibodies have restricted activity in that they inhibit infection in-vitro of endothelial cells , some epithelial cells and primary cytotrophoblasts but are completely ineffective in preventing infection of fibroblasts and have significantly reduced neutralizing activity against other cell types including trophoblast progenitor cells [11–13] . While vaccination with pentameric complex has repeatedly demonstrated induction of neutralizing antibodies in preclinical models , the protective efficacy of antibodies directed at the pentameric complex in humans remains to be shown [14 , 15] . Glycoprotein B represents the virion fusion protein for herpesviruses including HCMV [16] . It is essential for the infection of all types of target cells . As such gB remains as an attractive target for inclusion in a human vaccine and has been a major focus of experimental vaccination strategies . In fact , an efficacy study of adjuvanted recombinant gB vaccine ( gB/MF59 ) in postpartum , HCMV-seronegative women suggested an efficacy of approximately 50% protection from acquisition of HCMV infection [17] . Another phase II study in solid organ transplant recipients using the same vaccine showed 50% efficacy in controlling viremia in high-risk patients [18] . In addition , vaccination studies with gB in rhesus macaques and subsequent RhCMV challenge showed significantly reduced RhCMV DNA in plasma [19] . Mucosal immunization with a replication-deficient adenovirus vector expressing murine cytomegalovirus glycoprotein B induced mucosal and systemic immunity [20] . Finally , a number of studies using the guinea pig model demonstrated that congenital infection and mortality in pups was reduced following gB DNA or recombinant protein subunit vaccination strategies [21–23] . The reason ( s ) for the limited protection in the human gB-vaccine trials are unknown and could be many fold . One reason , among others , could be the induction of an unfavorable ratio of neutralizing antibodies versus non-neutralizing antibodies , which in some cases may even be competitive binders . Available data suggest that vaccination with the gB/MF59 vaccine induced high titers of binding antibodies but more limited titers of neutralizing antibodies [24 , 25] . Early studies have shown that passive transfer of immune serum obtained from mice infected with MCMV protected recipient mice from a lethal challenge with homologous viruses [26 , 27] . Using mAbs whose target proteins were not definitively identified , Farrell and Shellam observed some protection in immunocompetent mice [28] . In addition , Jonjic and co-workers demonstrated that B cells , and thus most probably antibodies , were critically involved in restricting dissemination of reactivated virus thus limiting recurrent infection [29] . Our own previous studies have provided evidence that antibodies can provide protection from MCMV-induced pathology in the brain of infected newborn mice [30] . In addition , we could show that in immunodeficient RAG-/- mice , adoptive transfer of memory B-cells or immune serum reduced viral load in organs even when administered 3 days after infection thus exhibiting therapeutic potential [31] . In these studies with polyclonal sera from immune donors , the specificity of protective antibodies and mechanism of protection were not defined . We have initiated studies to define the role of various anti-gB antibodies for protection in the murine CMV model . As in the case of HCMV , gB represents a dominant antibody target during MCMV infection . A panel of monoclonal antibodies against gB was generated , characterized in-vitro and their antiviral capacity in-vivo was investigated . Our data indicated that therapeutic application of neutralizing anti-gB antibodies has a greater potential to limit virus dissemination than non-neutralizing antibodies . When given prophylactically both neutralizing and non-neutralizing antibodies showed similar protective capacity . In-vitro , neutralizing anti-gB antibodies exhibited greater activity in limiting viral cell-to-cell spread which may represent one mechanism for their enhanced protective effect in-vivo .
In order to analyze the potential protective capacity of anti-gB antibodies in-vivo , we generated a number of gB-specific mAbs . As the primary aim in this experiment was to isolate mAbs with different antiviral capacity in-vitro , we used two experimental screening strategies: This approach resulted in the isolation of a number of gB-specific mAbs with different antiviral activities in-vitro ( Table 1 ) . mAbs that were identified by screening with virus neutralization assays had 50% inhibitory concentration ( IC50 ) values of 1–4 μg/ml . These IC50 values were similar to human mAbs directed against HCMV gB [11 , 32 , 33] . Most mAbs that were identified from binding assay utilizing transiently expressed gB had IC50s of >20 μg/ml . These mAbs were classified as non-neutralizing . To obtain more information on the binding region within gB , the mAbs were characterized with respect to binding to different antigenic regions on MCMV gB . As a blueprint for generating gB fragments that potentially could harbor antibody binding sites , we followed a strategy that was used to identify conformational epitopes on HCMV gB [34 , 35] . First , a 3D model of MCMV gB was generated based on the homology to HCMV gB and the crystal structure of HCMV gB [36] ( Fig 1A–1C ) . On the amino acid level HCMV gB and MCMV gB show >50% homology thus allowing for the generation of a 3D model of MCMV gB . Structural domains ( Dom ) that are conserved between the gBs of human herpesviruses were also identified in MCMV gB i . e . DomI-V ( Fig 1A–1C ) . The model also revealed overall structural similarity of the antigenic domains ( AD-1 , AD-4 , AD-5 ) which were previously identified in HCMV gB ( Fig 1A–1C ) [32 , 34] . To determine the mAb binding structures , the protein regions corresponding to AD-1 , AD-2 , AD-4 and AD-5 of HCMV as well as larger fragments of gB were generated by transient expression in mammalian cells and tested for mAb binding in indirect immunofluorescence analysis [32 , 34 , 35] . While no mAbs reacting with AD-2 or AD-5 were isolated , mAb 5F12 bound to AD-1 and mAbs 1F11 , 10H10 and M11 bound to AD-4 ( Fig 1D ) . In addition , binding of mAb 97 . 3 required expression of the NH-terminal part ( residues 1–488 ) of gB while 18A5 required expression of the complete gB protein for recognition . Thus , our panel of mAbs consisted of antibodies with different neutralizing capacity binding to different regions of MCMV gB . In a first series of in-vivo experiments , we tested mAbs individually for protection from MCMV infection in a therapeutic setting . In these experiments , we used a protocol that was similar to our previous studies which demonstrated protection of immunodeficient mice from the lethal course of MCMV infection by adoptive transfer of polyclonal sera from MCMV-immune donor animals [31] . RAG-/- mice , which do not harbor functional B and T cells , were infected with 105 pfu of MCMV157luc and three days later were treated with mAb or immune serum . On day 10 post infection ( p . i . ) mice were sacrificed and viral load in organs was quantified . Compared to controls , the viral load was reduced in all organs of animals treated with a neutralizing mAb and this effect reached statistical significance for all organs assayed with the exception of the lung ( Fig 2 ) . However , the reduction in viral load that was achieved with mAbs was less pronounced when compared to the effect observed when infected mice were treated with polyclonal serum from MCMV-immune donor animals ( supplemental S1 Fig ) . Interestingly , mAb treatment was able to significantly reduce viral load in salivary glands , the major organ linked to horizontal viral transmission in rodents . Treatment with mAbs 97 . 3 or M11 resulted in greater reduction in viral load than mAbs 27 . 7 and 1F11 despite comparable neutralizing activity when assayed in-vitro ( Table 1 ) ( Fig 2 ) . A potential explanation for the different in-vivo activity of 97 . 3 and M11 versus 27 . 7 and 1F11 was found when sera from the mAb-treated animals were analyzed for in-vitro neutralization one and four days after in-vivo administration . One day after mAb injection , sera from most animals resulted in 100% neutralization at a dilution of 1:10 with the exception of animals treated with antibody 1F11 where complete neutralization was reached in serum from only two animals ( Fig 3 ) . On day four after antibody transfer , 100% in-vitro neutralization was still achieved with sera from animals treated with mAb 97 . 3 or M11 ( Fig 3 ) . In contrast , sera from animals treated with mAbs 27 . 7 or 1F11 showed lower in-vitro neutralization titers which exceeded 50% only in a single animal . The remaining sera had neutralization titers of well below 50% at a dilution of 1:10 ( Fig 3 ) . Thus , the most likely explanation for the differences in in-vivo protection between 97 . 3 and M11 compared to 27 . 7 . and 1F11 was reduced antibody concentration in the serum of infected animals treated with mAbs 27 . 7 and 1F11 . Consistent with this explanation was the finding of a statistically significant negative correlation between the decay in neutralizing activity in serum and the reduction in virus titer in organs ( supplemental S2 Fig ) . In-vitro neutralization titers of sera from animals treated with immune serum also declined between days 1–4 but exceeded 50% at a dilution of 1:10 in three of four animals . In the next series of experiments , the non-neutralizing mAbs were tested for their antiviral activity in-vivo . A reduction in viral load 10 days p . i . was also observed following administration of non-neutralizing anti-gB antibodies ( Fig 4 ) . Although the reduction in viral load reached statistical significance , the overall antiviral effect of non-neutralizing mAbs was less pronounced compared to neutralizing mAbs . The extent of reduction in viral load was in the range of 1–10 fold ( supplemental S3 Fig ) compared to 10–100 fold following administration of the most potent neutralizing mAbs ( supplemental S1 Fig ) . This was particularly apparent for viral load in the salivary glands where only a single non-neutralizing mAb ( 20H7 ) was capable of significantly reducing the viral load when compared to the activity of the control immune serum ( Fig 4 ) . We also tested if treatment with combinations of mAbs could result in increased in-vivo protective activity . To exclude application of mAbs which competed for binding to the same epitope , the mAbs were first tested in competition assays and only mAbs were combined which did not exhibit competitive binding in-vitro ( supplemental S4 Fig ) . Animals received either a combination of the two most effective neutralizing mAbs ( 97 . 3 and M11 ) or the two non-neutralizing mAbs ( 18A5 and 20H7 ) or a combination of all four mAbs . The total amount of IgG that was injected into the animals was 250 μg/animal in all combinations . As can be seen in Fig 5A , the combination of mAbs 97 . 3 and M11 was equally effective in reducing the viral load when compared to the activity of polyclonal immune serum . The relative reduction in viral load of mAb combinations was comparable to immune serum and superior to mAb monotherapy ( supplemental S5 Fig ) . In contrast , the combination of non-neutralizing mAbs was less effective in reducing viral burden , although it reached statistical significance when compared to the control IgG2a ( Fig 5A ) . The simultaneous application of neutralizing and non-neutralizing mAbs resulted in intermediate protection as could be expected from the reduction of the amount of neutralizing antibody that was transferred in this four antibody combination . The MCMV157luc recombinant virus that was used in the above experiments was a derivative of the original BAC construct as described by Messerle et al . [37] . The resulting virus , however , has been shown to carry a mutation in the gene coding for MCK2 which results in altered cell tropism and in-vivo dissemination of the virus especially with respect to dissemination to the salivary glands [38] . To explore whether the presence of an intact MCK2 gene would influence the outcome of mAb protection experiments , a new virus was constructed with a repaired MCK2 gene ( termed MCMVlucMCK2 ) . This virus was then compared to the MCMV157luc virus in protection experiments . Importantly , the combination of neutralizing mAbs 97 . 3 plus M11 showed similar protection capacity against the virus carrying an intact MCK2 gene compared to the MCK2-mutated virus . Dissemination of MCMVlucMCK2 to the salivary glands was comparably reduced for both recombinant viruses indicating that the repaired genotype ( MCK+ ) did not influence antibody susceptibility of the virus to neutralizing activity of antibodies during dissemination ( Fig 5B ) . To define potential mechanism ( s ) of protection of these antibodies , we performed mAb-mediated plaque inhibition assays in-vitro . This assay was selected as surrogate activity for the effect of mAb on viral cell-to-cell spread in-vivo . We used a recombinant virus MCMVC3X-gfp , which is a derivative or the original BAC described by Messerle et al . [37] and which expressed the green fluorescent protein ( GFP ) under control of the HCMV IE enhancer , thus enabling direct visualization of live infected cells [39] . Murine fibroblasts were infected and mAb or serum was added 4 h later and plaque development was monitored between day 3 and day 7 p . i . in live cells . In the absence of antibody only single GFP-expressing cells were observed on day 3 p . i . , whereas plaque formation was clearly visible on day 5 p . i . and plaques containing large numbers of fluorescing cells were formed by day 7 p . i . ( Fig 6 ) . The addition of non-neutralizing mAbs had little , if any , effect on plaque formation as illustrated by findings from assays using mAbs 20H7 or 18A5 ( Fig 6 ) . In the panel of neutralizing mAbs , M11 completely prevented plaque formation while the remaining mAbs clearly reduced the number of infected cells per plaque , including mAb 97 . 3 ( Fig 6 ) . Polyclonal immune serum had an intermediate effect with respect to numbers of infected cells within plaques whereas serum from naïve mice failed to limit plaque formation in this assay . We next tested the protective capacity of the mAb combinations when transferred before infection . RAG-/- mice were given 250 μg total IgG one day before infection with 104 pfu MCMV157luc . On days 10 , 17 and 24 p . i . blood was taken and viral DNA was quantified by quantitative real time PCR ( qPCR ) . In PBS treated mice DNA copies increased from about 500 copies at day 10 to >40 000 copies per μg of total DNA on day 24 , indicating an ongoing viremia ( Fig 7 ) . In immune serum-treated mice , viral DNA was detectable with low ( 2–10 ) copy numbers in some animals . Application of the mAb combinations , either neutralizing or non-neutralizing , resulted in the clearance of viral DNA from the blood at any time point p . i . ( Fig 7A ) . We next repeated this experiment utilizing the MCK2+ virus ( MCMVlucMCK2 ) to determine if the presence of the MCK2 viral gene could result in altered viral replication and pathogenesis in infected mice that would in turn alter the antiviral activity of mAbs administered in a prophylactic protocol . Infection with the MCK2+ virus resulted in higher viral DNA copies in the blood compared to the MCK2- virus . The increased copy number of the MCK+ virus in the blood of the infected animals was not unexpected as MCK2 has been shown to increase the number of infected leukocytes and facilitate the recruitment and infection of monocyte/macrophages [40–42] . Importantly , however , the viral load in animals following application of neutralizing or non-neutralizing mAbs was similar to the activity of immune serum , again indicating that both classes of antibodies have similar antiviral activity when given prophylactically . ( Fig 7B ) In a second experimental approach , survival was monitored following infection with the original MCMV157luc virus and prophylactic administration of a different set of anti-gB mAbs , namely the therapeutically less potent neutralizing mAb 1F11 or the non-neutralizing mAb 5F12 . Mice treated with control IgG2a succumbed to the infection by day 40 p . i . ( supplemental S6 Fig ) . In contrast , about 80% of mice receiving gB-specific mAbs or immune serum survived the infection until day 100 . Again , the non-neutralizing mAb 5F12 was as effective as the neutralizing mAb 1F11 in prolonging survival of infected mice when given prophylactically . Taken together , these data indicate that the presence of antibody before the infection can result in significant protective capacity largely independent of the in-vitro neutralizing activity of the antibody .
We have previously shown that serum from MCMV-infected animals can protect B and T cell deficient hosts from lethal MCMV infection , thus demonstrating that passively acquired antibodies present in immune serum are protective in-vivo in the absence of de-novo adaptive responses to MCMV [31] . The antiviral antibody responses induced during MCMV infection is complex and includes antibody responses against a large number of virion structural and virus-encoded non-structural proteins . Only some of these responses have measurable in-vitro and/or in-vivo antiviral effector activity with the most well studied antiviral responses being directed against envelope glycoproteins . Following both MCMV and HCMV infections , anti-gB , gH/gL , and gM/gN antibody responses can be demonstrated and antibodies against gB , gH , or gN have been shown to neutralize virus in-vitro [10] . In the case of HCMV , antibodies against a pentameric complex consisting of gH/gL/UL128/UL130-131a have been shown to exhibit potent in-vitro neutralizing activity in assays utilizing epithelial and endothelial cell targets and a limited number of other cell types , but not in more permissive cell types such as fibroblasts [11–13] . Although we have isolated anti-gH and anti-gN virus neutralizing antibodies from MCMV infected mice , the goal of the current study was to evaluate gB-specific mAbs for their protective potential to model the role of gB antibodies during HCMV infections since gB represents a dominant antigen for the induction of antibodies during HCMV infection . In addition , we studied the in-vivo activities of both neutralizing as well as non-neutralizing anti-gB antibodies as both types of antibodies are induced during HCMV infection [32] . Antibodies selected for study included two groups of mAbs that were classified in-vitro as neutralizing or non-neutralizing as antibodies exhibiting these in-vitro activities were representative as a sample of anti-gB responses in MCMV infected mice . This different antiviral activity is also consistent with data from studies that have described antiviral neutralizing and non-neutralizing anti-gB antibodies against HCMV [32] . In addition , the antigenic regions that were recognized by the MCMV gB-specific mAbs were similar to those recognized by anti-HCMV gB mAbs , providing further support for the relevance of results from studies of the protective activities of anti-gB antibody responses in MCMV infected mice as a model of the role of anti-gB antibodies in HCMV infections [32 , 43] . In experiments designed to determine the therapeutic potential ( treatment ) of the anti-gB mAbs , we used a rigorous protocol in which gB mAbs were given three days after an i . p . infection of immunodeficient RAG-/- mice with 105 pfu of MCMV157luc . The quantity of antibody that was transferred per animal roughly translates to 10 mg/kg , a concentration of antibody that has been has been used clinically in humans [44 , 45] . In RAG-/- mice , in addition to the B and T cell deficiency , virus control by NK cells was largely eliminated secondary to the deletion of the m157 gene that encodes a NK activating ligand [46] . Within the experimental time frame of 10 days following infection , MCMV had disseminated to all organs including the salivary glands , a privileged anatomical site of cytomegalovirus immune evasion and persistence which has relevance for horizontal transmission of CMVs [47] . Monotherapy with neutralizing mAbs showed significant reduction in viral load in all tested organs with exception of the lungs . Interestingly , the viral load in salivary glands was also reduced significantly by passively transferred virus neutralizing antibodies , indicating inhibition of virus dissemination during secondary viremia [48] . The reduction in viral load was correlated quantitatively with the concentration of mAbs that was maintained in the serum of treated animals . Animals that received mAbs that exhibited an accelerated in-vivo decline in the serum concentration , such as 27 . 7 and 1F11 , were poorly protected as compared to animals treated with antibodies that maintained sustained mAb concentration . We can only speculate on the reasons for the differences in the serum decay of the different mAbs in-vivo since this process is complex , but obvious mechanisms could include off-target binding and/or immune complex formation resulting in increased rates of IgG elimination [49] . Significant differences in in-vivo half-life of transfused IgG has also be observed in other systems [50] . Regardless of the mechanism ( s ) of increased mAb elimination from the serum of MCMV infected mice , our findings indicated that mAb monotherapy that resulted in sustained high titers of circulating MCMV neutralizing antibodies in infected mice resulted in significant protection from lethal MCMV infection in immunodeficient mice . Importantly , this finding established an in-vivo property of antiviral antibodies that were required for optimal protection from viral dissemination and argued that simple classification of antibodies as neutralizing and non-neutralizing in in-vitro assays may not be fully predictive of in-vivo antiviral activity . Results from studies utilizing non-neutralizing antibodies argued that there was not an absolute requirement for potent in-vitro neutralizing activity in order to achieve protection in-vivo . Although the reduction in viral load following administration of non-neutralizing mAbs was less than that seen following transfer of neutralizing mAbs , the reduction in viral load was still statistically significant for antibodies such as 18A5 and 20H7 ( compare supplemental S1 and S3 Figs ) . Similar to the results following treatment with neutralizing mAbs , we observed differential activities of individual mAbs with respect to reduction of viral load in mAbs treated animals . Whether the differential protection was also based on variation of in-vivo serum concentration of the individual mAbs could not be determined as we have no quantitative assay to specifically quantify the concentration of these mAbs in serum . In any case , non-neutralizing mAbs provided remarkable protection in-vivo in the absence of appreciable in-vitro neutralizing activity . The finding that non-neutralizing antibodies can provide protection from viral infections has also been observed previously [51] and has recently become a major area of interest in studies of protective antibodies against influenza virus and HIV infections [52 , 53] . Effector functions associated with protection by our panel of mAbs are likely to be complex and remain undetermined at this time . Interaction of the antibody Fc-fragment with Fc-receptor bearing cells , resulting in antibody-dependent cell mediated cytotoxicity ( ADCC ) or antibody-dependent cellular phagocytosis ( ADCP ) may represent important mechanisms ( reviewed in [54] ) . However , affinity of individual Fc-receptor molecules for the different IgG subclasses molecules has been shown to be different . Thus , for our panel of mAbs representing different IgG subclasses the contribution of Fc-receptor interactions could vary . For example , 20H7 , being an IgG3 , will not be bound by any of the known Fc-receptors [55] . Moreover , CMVs also express virally encoded Fc-receptor molecules that potentially add even more complexity to the contribution of Fc-receptor dependent antiviral antibody activity to the interpretation of our findings ( reviewed in [56] ) . Finally , in addition to Fc-receptor binding , complement activation in-vivo may also be operative for some of the mAbs in our panel . However , complement binding and activation is also different for individual IgG subclasses ( reviewed in [57] ) . Moreover , herpesviruses including CMVs have developed a number of strategies to evade complement mediated functions . Among those are incorporation of complement control proteins in the virion particle or inhibition of complement-mediated lysis [58 , 59] . Thus , additional studies will be required to elucidate the mechanism ( s ) by which the individual mAbs represented in our panel will provide protection . A potentially important mechanism of virus dissemination in-vivo is by cell-to-cell spread , a mechanism that could be limited by antiviral antibodies . Herpesviruses , including CMVs , are believed to have the capacity to spread to contiguous cells without having to transit via the extracellular space [60] . In the case of CMV infections , in-vivo cell-to-cell spread is considered to be a major route of viral dissemination . However , the routes and molecular mechanisms through which CMVs spread from cell-to-cell in-vivo remain poorly defined [61 , 62] . Regardless of the mechanism ( s ) of cell-to-cell spread , in-vitro studies using plaque formation/expansion has been suggested to be a surrogate for cell-to-cell spread . Our data indicate that cell-to-cell spread of MCMV can be inhibited by a subset of antiviral antibodies from a group of mAbs with comparable neutralizing activity when measured in-vitro with cell-free virus , suggesting that the requirements of mAb activity for inhibition of cell-to-cell spread could be distinct from those characteristics of mAbs required for neutralization of cell free virus infection . Consistent with additional effector functions of virus neutralizing mAbs was the finding that a combination of M11 and 97 . 3 was more potent in in-vivo protection than either antibody transferred individually . However , synergism in in-vitro virus neutralization between these two antibodies could not be demonstrated ( supplemental S7 Fig ) . Also , the serum neutralizing capacity when both antibodies were given in combination as well as the concentration in-vivo of neutralizing activity did not differ from that observed in mice treated with an individual antibody ( supplemental S7 Fig ) . However , the finding that they have different inhibitory activity potency when compared in an in-vitro plaque formation inhibition assay indicated that their mode of action in-vivo , apart from neutralization of free virus , could be different and potentially was linked to the increased protective activity of the combination of mAbs . Finally our findings also suggest that inhibition of plaque formation in-vitro may represent a more informative assay for prediction of in-vivo protection than neutralizing capacity determined by an in-vitro assay using cell free virus . Perhaps one of the most interesting and unexpected findings in this study was that for the panel of mAbs tested there was no difference in the in-vivo antiviral activity between non-neutralizing and neutralizing mAbs when these mAbs were used in a prophylactic protocol . Moreover , monotherapy with single antibodies from these mAb combinations also provided significant protection from the lethal course of the infection in RAG-/- mice . Although these results will require further studies to define mechanisms of protection provided by neutralizing and non-neutralizing antibodies in this model system , there are several possible explanations that could account for these results . Importantly , there is a fundamental difference in antibody-virus interaction between a primary virus inoculation and an established infection when mAbs are used in a therapeutic protocol . In our experiments , the virus was inoculated intraperitoneally and free virus spreads within the first hours via a haematogenous route to the spleen and liver [63] . Thus , disseminating virus comes into contact with circulating antiviral IgG . As both , neutralizing and non-neutralizing mAbs can bind to free virions , antibody effector functions mediated via the Fc could prevent infection of the first cellular target of infection . As the antibody coated virus is transported to structures such as the spleen which are rich in cells carrying Fc-receptors it could be eliminated via Fc-mediated effector functions . Whether an antibody has neutralizing function or not could be less critical in terms of its protective activities during these early events of infection . Virus neutralizing activity could play a more important role in control of virus spread at later time points during infection i . e . during cell-to-cell spread in infected tissues and/or secondary viremia . Vaccination trials in humans using an adjuvanted gB have provided conflicting evidence of protection from community acquisition of HCMV [17] . In the initial report , three doses of the gB vaccine limited acquisition of HCMV in a group of women and although differences between vaccine recipients and placebo controls were observed , the statistical difference between the two groups was not robust [17] . In a follow-up vaccine trial in adolescent females , there was no statistically significant difference in acquisition of HCMV between gB vaccine and placebo recipients [64] . In contrast , our findings demonstrated that anti-gB antibodies have potent protection capacity in-vivo , particularly when used as prophylaxis to limit infection . There are several explanations that could account for this difference: Taken together , our study has provided convincing evidence that a subset of antibodies directed against gB of MCMV raised either following infection or after immunization with intact virions can provide significant protection in-vivo when transferred into MCMV infected mice either in a prophylactic or treatment protocol . Potent in-vitro neutralizing activity seems not to be an absolute prerequisite for this effect . Whether antibodies with these specificities and activities can be generated following vaccination will be investigated in future studies .
RAG-/- mice were obtained from in-house breeding based on mice from Charles River and maintained under specific pathogen-free conditions . In experiments involving therapeutic application of mAbs , mice were infected with 1 x 105 plaque forming units ( pfu ) of MCMV157luc or MCMVlucMCK2 by intraperitoneal ( i . p . ) infection . In experiments involving prophylactic application of antibodies , 104 pfu MCMV157luc or MCMVlucMCK2 was used . In-vivo bioluminescence imaging was done exactly as described [31] . All experiments were conducted in accordance with institutional guidelines for animal care and use . The experiments were approved by the Regierung von Mittelfranken ( Government of Frankonia ) approval 54–2532 . 1-57/12 and adhered to the EEC Council Directive 2010/63/EU . The animal facility of the University of Erlangen-Nürnberg is approved by the Dept . of Health & Human Services , USA , approval number A5903-01 . Mouse embryonic fibroblasts ( MEF ) and ST-2 cells were cultured in DMEM medium ( Life Technologies , Germany ) supplemented with 10% fetal calf serum ( FCS ) ( Sigma-Aldrich , Germany ) , glutamine ( 100 mg/ml ) , and gentamicin ( 350 mg/ml ) . All virus strains were derived from the original MCMV BAC as described by Messerle [37] . MCMV157luc was propagated and purified as described [31] . The MCK2 mutation in MCMV157luc was repaired as reported by Jordan et al . [38] resulting in virus MCMVlucMCK2 . Virus titer was determined by end-point titration using indirect immunofluorescence . Briefly , serial dilutions of viral preparations were used to infect MEF that had been seeded in 96-well plates ( 12 000 cells/well ) . Two days later , cells were fixed with ethanol and infected cells were stained and quantified using the monoclonal antibody Croma101 , which is specific for the viral immediate early protein 1 of MCMV . MCMV-C3X-gfp which expresses the green fluorescent protein ( GFP ) under control of the HCMV immediate early promoter was a kind gift from M . Messerle , Hannover , Germany [39] . Immortalized antibody-producing B-cell lines were generated from the spleens of infected or immunized donor mice by conventional hybridoma technology . Briefly , C57BL/6 mice were i . p . infected with 1x106 pfu MCMV157luc and spleen cells were harvested 4–6 weeks after infection . One week before harvest of the spleen , the animals were boosted with 5 μg of UV-inactivated MCMV157luc virions . Following this protocol the Mabs 1F11 , 27 . 7 , M11 and 97 . 3 were obtained from three different fusions . Mabs 18A5 , 20H7 were isolated from Balb/c mice that were treated with an identical protocol . Mabs 5F12 and 10H10 were obtained following immunization of a Balb/c mouse ( three times 5 μg each of UV-inactivated MCMV157luc virions i . p . and intervals between injections of at least 4 weeks . ) Three to four days after the last immunization , spleens were removed and splenic cells ( 100–200 x106 cells ) of the donor mouse were fused with 50–100 x 106 SP2 . 0 cells . Cells were seeded in 96 F-bottom cell culture microplates in 150 μl medium per well . 8–10 days later , supernatants were tested for neutralization ( from mice that were infected ) or virion binding antibodies in an ELISA ( from mice that were immunized with UV-inactivated virions ) . Clones of interest were subcloned using a Beckman Coulter MoFlo cell high-speed sorter® . Following additional rounds of subcloning , hybridoma supernatants were characterized by ELISA , indirect immunofluorescence using transiently expressed gB and neutralization . mAbs were purified in-house by protein A chromatography or prepared and purified by BioXCell ( USA ) . IgG subtypes were determined using a mouse immunoglobulin panel ( Southern Biotech , Germany , Cat . No:5300–01 ) and an IgG2c isotype control antibody ( GeneTex , USA , Cat . No:GTX35043 ) as coating reagents in standard ELISA assays . Standard IgGs were coated at a concentration of 100ng/well in 96 well plates and compared to undiluted samples from hybridoma supernatants . Assays were developed using a Southern Biotech SBA Clonotyping System-HRP ( Cat . No . :5300–05 ) according to the manufacturer´s suggestion complemented by Goat Anti-Mouse IgG2c HRP to detect IgG2c ( Southern Biotech Cat . No . :1079–05 ) . Biotinylation of purified mAbs ( 500 μg each ) was carried out using the EZ Link® Sulfo-NHS Biotinylation Kit ( Thermo Fisher Scientific , Germany ) according to the manufacturer’s instructions . Biotinylation was confirmed in ELISA assays using the mAbs as coating reagent and HRP conjugated Streptavidin as detecting reagent . Organs were harvested and snap frozen in liquid nitrogen . For determination of virus titer , organs were thawed and homogenized in Glo Lysis Buffer ( Promega , Germany ) using a Precellys 24 homogenizer ( Peqlab Biotechnologie , Germany ) . Homogenates were centrifuged at 4°C for 10min at 16000xg and protein concentration was determined in the supernatant using a BCA Protein Assay Kit ( Perbio Science , Germany ) . 30 μl Glo lysis buffer containing 30 μg protein of lysates were transferred into white 96 well LIA-plates ( Greiner Bio-one , Germany ) . 50 μl assay buffer ( 15 mM KH2PO4 , 25 mM glycylglycine , 1 M MgSO4 , 0 . 5 M EGTA , 5 mM ATP , 1 mM DTT ) per well was added . Injection of 50 μl D-luciferin- ( P . J . K . , Germany ) solution per well ( In 25 mM glycylglycine , 1 M MgSO4 , 0 , 5 M EGTA , 2 mM DTT and 0 , 05 mM D-Luciferin ) and detection of chemiluminescence was performed by a Centro LB 960 Luminometer ( Berthold Technologies , Germany ) . MicroWin2000 Software ( Mikrotek Laborsysteme , Germany ) was used for analysis . To measure virus copy numbers in peripheral blood , DNA was isolated from 200 μl of EDTA-blood using the QIAamp DNA blood kit according to the manufacturer’s protocol ( QIAgen , Germany ) . For MCMV-specific qPCR , 50 ng of the isolated DNA was subjected to a 20-μl reaction mixture containing 10 μl 2x TaqMan PCR Mastermix ( Applied Biosystems , Germany ) , 10 μM probe and 5 μM of each primer . Primers and probe for the detection of MCMV were based on the MCMV ie1/4 exon 4 sequence ( forward primer: 5′-TGCCATACTGCCAGCTGAGA-3′; reverse primer: 5′-GGCTTCATGATCCACCCTGTT-3′; and probe: 5′-CTGGCATCCAGGAAAGGCTTGGTG-3′ ) . For in-vitro neutralization , serial dilutions of sera or monoclonal antibody were incubated with 1200 pfu MCMV157luc for 1h . The mixture was added to ST-2 cells that were seeded at a density of 1 . 2x104 the day before in 96-well plates . Following incubation for 4h the culture medium was changed and infection continued for 48hrs . Thereafter cells were lysed in 100 μl Glo lysis buffer and 30 μl were used to measure luciferase activity as described above . Sera were not heat inactivated and no exogenous complement was added . To express MCMV gB , the coding sequence from orf m55 strain Smith ( GenBank accession number NC_004065 . 1 ) was inserted into pcDNA3 . The plasmids encoding fragments of gB were constructed by inserting the appropriate DNA fragment into the vector pcUL132-sig-HA . This pcDNA3 . 1 ( Invitrogen , Germany ) based plasmid contains the coding sequence of the HCMV gpUL132 authentic signal sequence aa 1–27 , followed by the coding sequence for the HA-epitope YPYDVPDYA [67] . Cos7 cells ( 5x104 per well ) grown in 24-well plates on 15-mm glass coverslips were transfected with 0 . 8 μg plasmid DNA using Lipofectamine ( Invitrogen , Germany ) . 48 hours after transfection , cells were fixed and permeabilized with ice cold methanol . Primary antibodies were then added for 45 min at 37°C . Unbound primary antibody was removed by three PBS washing steps . Binding of the primary antibody was detected with the appropriate FITC-conjugated secondary antibody ( fluorescein isothiocyanate ) ( Dako , Germany ) ( 45 min at 37°C ) . Counterstaining of cell nuclei was done with DAPI ( 4' , 6-diamidino-2-phenylindole ) . Images were collected using a Zeiss Axioplan 2 fluorescence microscope fitted with a Visitron Systems charge-coupled device camera ( Puchheim , Germany ) . Images were processed using MetaView software and Adobe Photoshop . MEF were seeded in 96-well plates ( Ibidi , Germany ) at a density of 4x104 per well . 24 h later cells were infected with 2000 pfu/well of MCMV C3X-gfp using centrifugal enhancement ( 5min , 500xg ) to enable synchronous infection . 4h later the inoculum was removed and cells were incubated in 300 μl/well medium containing 20 μg/ml mAb or serum at a dilution of 1:100 . Infection was documented using a Leica DMI 6000B microscope starting at day 3 post infection . Magnification: 200fold . Filter: excitation 488 nm , emission 509 , exposure times: 40 ms , picture size: 640 μm x 478 μm . The model of the MCMV gB structure was generated by standard homology modelling procedures using the program MODELLER 9 . 10 [68] based on a sequence alignment with the template structure of HCMV gB ( PDB code: 5CXF ) . Two loop regions ( Asn409-Gln410 and Lys435-Val475 of HCMV gB ) were not resolved in the reference structure and were therefore not modelled . Images were generated with VMD 1 . 9 . 1 [69] . Statistical analysis was performed by one way ANOVA using Bonferroni´s multiple comparison test using GraphPad Prism ( version 6; GraphPad Software , USA ) . | Human cytomegalovirus ( HCMV ) is a major global health concern and a vaccine to prevent HCMV disease is a widely recognized medical need . However , no vaccine has been licensed to date . A major obstacle for the development of a vaccine is a lack of knowledge of the nature and specificities of protective responses that should be induced by the vaccine . HCMV is a complex virus containing numerous antigens within the viral envelope that could be targets for protective antibodies . Glycoprotein B ( gB ) is an important target for neutralizing antibodies and hence an interesting molecule for intervention strategies such as vaccination or passive immunotherapy . We have used the murine model system of CMV ( MCMV ) to explore the potential of gB-specific antibodies in immunotherapy or prophylaxis . Our results show that anti-gB antibodies can protect immunodeficient hosts from the lethal course of the infection . When used as therapy for established infection , both neutralizing as well as non-neutralizing antibodies showed significant protection with neutralizing antibodies being superior . Among the neutralizing antibodies , protection correlated with sustained in-vivo neutralizing activity rather than with the magnitude of the in-vitro neutralizing titer . Interestingly , both neutralizing and non-neutralizing antibodies showed comparable protection when given prophylactically i . e . one day before infection with MCMV . Thus , our data indicate that in-vitro neutralizing capacity of CMV-specific antibodies may not be reflective of antibody effector functions that provide protection in-vivo . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"serum",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"immunology",
"microbiology",
"cytomegalovirus",
"infection",
"vi... | 2017 | Protective capacity of neutralizing and non-neutralizing antibodies against glycoprotein B of cytomegalovirus |
Accurate timing of action potentials is required for neurons in auditory brainstem nuclei to encode the frequency and phase of incoming sound stimuli . Many such neurons express “high threshold” Kv3-family channels that are required for firing at high rates ( >∼200 Hz ) . Kv3 channels are expressed in gradients along the medial-lateral tonotopic axis of the nuclei . Numerical simulations of auditory brainstem neurons were used to calculate the input-output relations of ensembles of 1–50 neurons , stimulated at rates between 100–1500 Hz . Individual neurons with different levels of potassium currents differ in their ability to follow specific rates of stimulation but all perform poorly when the stimulus rate is greater than the maximal firing rate of the neurons . The temporal accuracy of the combined synaptic output of an ensemble is , however , enhanced by the presence of gradients in Kv3 channel levels over that measured when neurons express uniform levels of channels . Surprisingly , at high rates of stimulation , temporal accuracy is also enhanced by the occurrence of random spontaneous activity , such as is normally observed in the absence of sound stimulation . For any pattern of stimulation , however , greatest accuracy is observed when , in the presence of spontaneous activity , the levels of potassium conductance in all of the neurons is adjusted to that found in the subset of neurons that respond better than their neighbors . This optimization of response by adjusting the K+ conductance occurs for stimulus patterns containing either single and or multiple frequencies in the phase-locking range . The findings suggest that gradients of channel expression are required for normal auditory processing and that changes in levels of potassium currents across the nuclei , by mechanisms such as protein phosphorylation and rapid changes in channel synthesis , adapt the nuclei to the ongoing auditory environment .
Transfer of information in the auditory system occurs at rates that are faster than can be accomplished by single neurons . Accurate responses to auditory stimuli frequently require the discrimination of events that are separated in time by only tens of microseconds , and the ability to follow changes in the stimulus at rates of up to several thousand Hertz . In contrast , neuronal action potentials typically have durations close to a millisecond and the most rapidly firing neurons can generate action potentials at only several hundred Hertz . Experimental clues on how the central nervous system copes with high rates of information transfer have come from studies of auditory brainstem neurons , such as those of the anteroventral cochlear nucleus ( AVCN ) and the medial nucleus of the trapezoid body ( MNTB ) [1] , [2] , [3] , [4] , [5] . Among other targets , these neurons project to the lateral and medial superior olives ( LSO and MSO ) , where information that is relayed through the nuclei is used to detect microsecond differences in the timing of auditory stimuli , as well as small differences in their amplitude , to calculate the positions of auditory stimuli in space [1] , [6] . Neurons in the AVCN and MNTB lock their action potentials very precisely in time to the phase of sounds at frequencies up to ∼3000 Hz [7] , [8] , [9] , [10] , [11] and/or lock their action potentials to the temporal envelope of higher frequency sounds that are amplitude-modulated at frequencies up to ∼3000 Hz [12] . Nevertheless , individual neurons in these nuclei have maximal firing rates of only a few hundred Hz [13] , [14] , [15] . In order to follow higher rates of auditory signals , these phase-locking neurons respond selectively to only a subset of the stimuli . For example , in response to a 600 Hz auditory stimulus , a neuron that fires at a maximal rate of 350 Hz may lock its action potentials to every other cycle of the stimulus , effectively firing at only 300 Hz . The limitations of the intrinsic excitability of neurons raise some fundamental questions . For example , how can a neuron differentiate between two patterns of synaptic inputs , one at 600 and the other at 1200 Hz if it fires at exactly 300 Hz in response to each of these stimuli ? How can small differences in frequency be detected in the face of the high degree of spontaneous activity that occurs even in the absence of sound stimulation ? In the MNTB , the phase-locking principal neurons fire in vivo at rates between 10 to 300 Hz even in silence [7] , [10] . This spontaneous activity is believed to be generated by spontaneous transmitter release from sensory hair cells in the cochlea [16] , [17] , [18] , [19] . The resolution of some of these questions is that a sound stimulus in encoded in the activity of an ensemble of neurons rather than by single neurons . The accuracy of the output of an ensemble is , however , dependent on the number and characteristics of its individual neurons . A variety of modeling studies , as well as experimental studies , have documented that the intrinsic electrical properties of neurons in nuclei such as the AVCN and MNTB play a key role in determining their responses to synaptic inputs triggered by auditory stimuli [20] , [21] , [22] , [23] , [24] , [25] . These intrinsic electrical properties are , however , not uniform but vary across the lateral-to-medial tonotopic axis of these nuclei . For example , levels of the voltage-dependent “high-threshold” K+ channel Kv3 . 1 are low in the lateral , low-frequency part of the MNTB and high in the medial , high-frequency aspect of the nucleus [26] , [27] , [28] , [29] , [30] , [31] . In contrast , some other “lower-threshold” channels such as , Kv1 . 3 [32] , Na+-activated K+ channels [33] , and an intermediate voltage-activated K+ current [26] are expressed in an opposite gradient with highest levels in lateral MNTB neurons . This manuscript describes numerical simulations of the firing patterns of ensembles of neurons with firing properties based on those of phase-locking auditory brainstem neurons . It is demonstrated that , in the absence of spontaneous activity , the accuracy of timing of outputs is enhanced in ensembles that have gradients of high-threshold K+ channels over those that have uniform levels of these channels . Paradoxically , accuracy of phase-locking is also substantially enhanced by random spontaneous activity in the input to the neurons , even when the neurons have uniform electrical properties . For any stimulus pattern in the presence of spontaneous activity , greatest temporal accuracy is obtained when levels of the high-threshold K+ current in all neurons are adjusted to a specific uniform level . This specific level corresponds to the level found in those neurons in the gradient that initially respond to the stimulus pattern more accurately than their neighbors . The present simulations provide a potential biological explanation for the existence of gradients of ion channels across neurons of a single nucleus . They also suggest that the relatively rapid modulation of K+ channels within gradients reflect an adaptation that allows the auditory system to adjust the processing of timing information to different auditory environments . Modulation of K+ channels has been found to occur in response to changes in ongoing auditory activity [30] , [34] , and is mediated by mechanisms such as channel phosphorylation [34] , [35] , [36] and synthesis of new channel proteins [29] , [37] , [38] .
To examine the accuracy with which phase-locking neurons are capable of transmitting timing information , simulations of the pattern of firing of model neurons with ionic currents based on those recorded in MNTB neurons in brain slices were first carried out [39] . The models that describe the patterns of firing of the presynaptic neurons used in this study have been described in detail previously [34] , [36] , [40] , [41] , [42] . Individual model neurons ( Input cells ) were stimulated by brief current pulses ( 250 µs ) at rates of >100 Hz ( Input , Fig . 1A ) . When the upstroke of action potentials crossed 0 mV , axonal propagation that triggered the presynaptic release of neurotransmitter was assumed to occur . A second trace , representing the postsynaptic effects of the stimulus train , was then calculated . Postsynaptic currents , with a decay time constant of 2 msec , were triggered by each presynaptic action potential ( Output , Fig . 1A ) . For convenience , calculations were made for excitatory postsynaptic currents , such as those evoked in AVCN neurons by inputs from the auditory nerve , or in MNTB neurons from the AVCN input . Exactly the same general principles would apply , however , to inhibitory postsynaptic currents , such as those evoked in superior olivary neurons by inputs from the MNTB . To calculate the timing accuracy of the postsynaptic responses with respect to the stimulus train , the strength of a phase vector that evaluates how accurately the evoked postsynaptic responses are locked to individual stimuli during a 100 msec train was calculated . For each postsynaptic trace the height and time of the maximal peak in current that occurred between one stimulus pulse and the next was determined . The strength of the phase vector was then calculated as described previously [34] , [41] . The strength of this phase vector was then multiplied by a factor that reflects the proportion of stimuli that evoked a postsynaptic response ( see Methods ) [29] . For a neuron with one postsynaptic output this factor is the same as entrainment , i . e . it reflects the proportion of input stimuli that triggered action potentials in the input cells . This adjusted parameter was termed PE ( for Phase/Entrainment ) . For example if only 50% of stimuli evoke responses , the maximal value of PE is 0 . 5 . At lower stimulus rates , the neurons are capable of firing an action potential in response to each stimulus . This is seen in the left panels of Fig . 1B , which shows the presynaptic action potentials and postsynaptic currents at the onset of a 300 Hz stimulus train . If each action potential and postsynaptic response has the same latency with respect to its stimulus pulse the strength of the vector PE can come close to the theoretical maximum value of one . The exact timing of the action potentials , however , depends on the level of high-threshold Kv3 current ( KHT ) in the presynaptic neuron . As has been found previously in pharmacological , genetic knockout and modeling studies , changes in KHT current produce only minor changes in action potential width but have a major impact on the response to repetitive stimulation [34] , [36] . Fig . 1B compares the responses for a low conductance ( gKHT = 0 . 1 µS ) and a high conductance ( gKHT = 0 . 7 µS ) of KHT in the presynaptic neuron . The low KHT conductance produces action potentials that have a uniform delay for each stimulus pulse . In contrast , the timing of the action potentials in the neuron with high KHT is progressively delayed , such that the latency from the onset of the stimulus pulse to the postsynaptic response is clearly different for the third stimulus compared to the first . Such delays have been termed “errors” in timing [34] and are caused by the increased relatively refractory period that results from high levels of voltage-dependent K+ conductance . A different situation occurs when the input neurons are stimulated at higher rates ( 900 Hz in Fig . 1B ) . As is also found in pharmacological and genetic knockout studies [36] , neurons with low KHT are incapable of generating more than a single action potential at the onset of the train . Conversely , higher levels of KHT allow a neuron to maintain firing throughout the train of rapid input stimuli . This is because the stronger repolarization with high KHT levels allows sodium channels to recover from inactivation more rapidly , and thus maintains firing during the synaptic barrage . Fig . 1C displays plots of the vector strength PE as a function of stimulus frequency for these two different levels of KHT . Thus high levels of KHT allow neurons to fire at high rates , but introduce timing errors such as those seen at 300 Hz in Fig . 1A , while neurons with low KHT levels preserve timing accuracy but are incapable of firing at high rates [34] , [36] . Fig . 2 shows plots of PE against levels of KHT between 0 . 1 and 0 . 7 µS for stimulation rates of 200–1500 Hz . At each stimulus rate , a red circle indicates the level of KHT that provides the optimal PE vector strength for that rate of stimulation . At low rates of stimulation ( 200–400 Hz ) the ability of neurons to follow the stimulus with high temporal accuracy is maximal if the neuron has a low level of KHT . As stated above , this is because the relative refractory period after an action potential , which delays the timing of a subsequent action potential , is minimized at low levels of KHT [34] . In contrast , at high rates of stimulation ( 1000–1500 Hz ) , neurons with low levels of KHT are incapable of following the stimulus , and high levels of KHT improve the value of PE . At high rates of stimulation , however , individual neurons are incapable of responding to every stimulus , and the maximal value of PE is very low compared to that achieved with low stimulus rates . For intermediate rates of stimulation , optimal PE vector strength occurs in neurons that have intermediate levels of KHT . In animals with normal hearing , levels of high threshold Kv3 channel proteins in auditory brainstem nuclei are not uniform . They are low in the lateral , low-frequency part of nuclei such as the MNTB and AVCN , and progressively increase towards the medial , high-frequency aspect of these nuclei [26] , [27] , [28] , [29] , [30] , [31] . Fig . 3A shows how the relative levels of Kv3 . 1 change along this axis of the MNTB in published measurements made on normal-hearing rodents and compares these to data from studies in which Kv3 . 1 levels become uniform across the nucleus , either because of the onset of partial hearing loss [31] or in an animal model of Fragile-X Syndrome [29] . It also shows the relative gradients of KHT that were used in simulations . To test the effects of such gradients on temporal accuracy , an ensemble of 50 neurons in which the levels of KHT vary systematically across the ensemble was simulated . For each ensemble , the PE vector strength was first calculated for the postsynaptic output of each individual neuron , and then again calculated for linear combinations of the postsynaptic output of groups of 2 , 3 , 4 , …50 neighboring neurons ( Fig . 3B ) . An important aspect of the interpretation of the parameter PE for the output of an ensemble is that it no longer reflects the amount of entrainment of individual input neurons by the stimulus input . For example for a stimulus that is greater than the maximal firing rate of a single neuron , the value of PE can still be equal to 1 if the combined postsynaptic outputs of the ensemble are of equal amplitude and delay for every stimulus pulse that is applied , even if each individual neuron fires only in response to a fraction of the input stimulus pulses . To portray the effect of combining the outputs of multiple neurons from the ensemble of 50 cells , a color code was used to represent the strength of the adjusted phase vector ( PE ) in which brown corresponds to random firing ( V = 0 ) and dark blue to perfect phase-locking and entrainment in the output ( PE = 1 ) ( Fig . 4 ) . The left edge of the rectangle represents 50 points corresponding to PE for the output of individual neurons . The right edge is a single color that represents the unique value of PE when the output of all 50 neurons is combined equally to generate a single postsynaptic trace ( as in Fig . 3B right ) . Intermediate points along the x-axis of the rectangle represent the values of PE of linear combinations of 2–49 neighboring neurons along the gradient . The top set of panels of Fig . 4 show the result obtained for an ensemble with a uniform distribution of Kv3 . 1 , stimulated at 700 Hz . If the intrinsic electrical properties of neurons in an ensemble are all identical , and all of the neurons receive the same stimulus , then , as in this case , the combined output of the ensemble is identical to that for a single neuron . The lower set of panels in Fig . 4 shows results for ensembles with the four gradients depicted in Fig . 2C . Also shown graphically are the mean values of PE for the combined outputs of 1–50 neurons ( right panels ) . It is evident that combining the outputs of multiple neurons in a gradient can slightly improve the value of PE over that produced by an ensemble of uniform neurons . This is because the intrinsic properties of the neurons vary across the ensemble and the timing of action potentials triggered by the stimulus train is different for each neuron , as illustrated in the raster plots at the left of Fig . 4 . In the combined postsynaptic output , there will be a summation of postsynaptic currents that are triggered by action potentials locked to individual stimulus pulses . Postsynaptic currents resulting from errors in action potential timing can be averaged out in the combined output . Thus the combined output of the ensemble can be enhanced in accuracy over the response of any one cell . There is a high degree of spontaneous activity in the auditory nerve and in neurons of auditory brainstem nuclei even in the absence of sounds . This spontaneous discharge rate typically varies from ∼10 Hz to over 200 Hz and is believed to be triggered by spontaneous transmitter release from cochlear inner hair cells [7] , [10] , [16] , [17] . As is illustrated schematically in Fig . 5A–C , the occurrence of random spontaneous activity might be expected to shift the pattern of response to an ongoing stimulus . For example , the occurrence of a spontaneous action potential immediately before an incoming stimulus would be expected to delay the response of the neuron to the next stimulus . This would be expected to improve the value of PE by desynchronizing responses during high rates of stimulation so that cycle-skipping is no longer synchronized across the population . Such spontaneous activity would also be expected to randomize the occurrence of stimulus-evoked errors in timing . Fig . 5D , which shows the effect of introducing spontaneous activity into ensembles with 1 , 2 , 5 or 10 input neurons , demonstrates that such improvement does in fact occur . Spontaneous activity was applied to individual presynaptic neurons , which all had the same value of KHT ( 0 . 58 µS ) , as brief current pulses identical to those used for the stimulus train but with a random distribution in time ( and with a mean frequency of 100 Hz ) . In the presence of spontaneous activity , the output from the ensemble of 10 neurons has peaks of similar amplitude and delay in response to each of the stimulus pulses that were applied at 700 Hz . In contrast , with no spontaneous activity , the output of the 10 neuron ensemble is identical to that of a single neuron with a KHT value of 0 . 58 µS . In this case the timing of peaks in the postsynaptic output is limited by the firing rate of the individual neurons . A more complete test of the effects of random spontaneous activity was carried out using the ensemble of 50 neurons with uniform electrical properties first shown at the top of Fig . 4 . As expected , when random patterns of stimulation were presented to the model ensembles of neurons in absence of a stimulus , calculations of vector strength with respect to any stimulus produced values close to zero ( Fig . 6 , top panels ) . When random spontaneous activity at mean rates of 10–200 Hz was combined with a coherent 700 Hz stimulus , however , spontaneous activity greatly enhanced the temporal accuracy of the output of the ensemble ( Fig . 6 , lower panels ) . Moreover , increases in the overall rate of spontaneous activity systematically increased the overall phase vector strength of the output . When , however , the rate of spontaneous activity was increased up to and past the rate of the stimulus itself , the steepness of the relation between PE and the number of convergent output cells ( N ) was progressively decreased , reflecting degradation of the signal ( data not shown ) . Both spontaneous activity and the existence of gradients in potassium channels produce diversity in the firing patterns of individual neurons . To determine if these mechanisms occlude each other , tests were made of the effects of introducing random spontaneous activity into ensembles with gradients . In all cases , the phase vector strength of the combined synaptic outputs was improved by combining spontaneous activity with the gradient over that with either spontaneous activity or a gradient alone . This is illustrated in the top panels of Figs . 7A and 8A in which spontaneous activity ( 100 Hz ) was added to an ensemble with the gradient first shown in the bottom panel of Fig . 4 and the ensemble was stimulated at 700 Hz or 900 Hz respectively . In the presence of spontaneous activity , values of PE could be very substantially improved by further adjustments of high-threshold K+ current . Paradoxically , this occurs by “flattening” out the original gradient . Fig . 2 showed that , for any specific stimulus rate in the absence of spontaneous activity , the temporal accuracy of the output of single neurons depends on their level of KHT . The same is true for stimulation in the presence of spontaneous activity . For stimulation at 700 Hz , the greatest values of PE in the output of single neurons were achieved with high values of KHT ( Fig . 7B ) . The overall fidelity of the output could then be substantially improved by adjusting levels of KHT in all neurons in the ensemble to those found in the neurons with the best individual responses . This is shown on the right side of Fig . 7A . The lower two sets of panels show that the accuracy of the output of the ensemble is increased by fixing KHT in all the neurons at 0 . 58 or 0 . 7 µS , values at the high end of the gradient . Note that traces of the postsynaptic output for KHT = 0 . 58 µS were shown in Fig . 5D . Conversely fixing KHT in all neurons at a low value ( 0 . 2 µS ) reduces temporal accuracy relative to the ensemble with full gradient . This is also shown in Fig . 7C , which plots the value of PE as a function of the number of converging inputs ( N ) for each of the ensembles . Simulations were carried out for a wide variety of parameters and stimulus rates . As expected from Fig . 2 , in the presence of spontaneous activity , values of PE were enhanced by adjusting levels of KHT in all neurons to low values for low rates of stimulation . Conversely , for high rates of stimulation , PE was increased over that in a full gradient by adjusting KHT to a high level in all neurons . Moreover , for all stimuli at high rates , the accuracy of the output of the ensemble was greater than that of individual neurons . At stimulus rates at which individual neurons responded optimally with intermediate values of KHT , the output of the ensemble was enhanced by adjusting KHT to this intermediate value in all neurons . This is illustrated in Fig . 8 for stimulation of the same ensemble at 900 Hz . In either the presence or absence of spontaneous activity , the value of PE in individual neurons with values of KHT at 0 . 4 µS is greater than in neighboring neurons with lower or higher KHT ( Figs . 2 , 8B ) . At low levels of KHT ( 0 . 2 µS ) , neurons are incapable of firing more than a few action potentials at the onset of the train , while higher values produce errors of timing ( Fig . 8A ) . Adjustment of KHT in all neurons to the intermediate conductance of 0 . 4 µS produces optimal values of PE in the output of the ensemble . In the simulations described above , the input stimuli were applied at a single rate between 100 and 1500 Hz . Auditory stimuli , however , may generally contain multiple frequency components in the phase-locking range . These components may represent either the sound frequency itself or modulation of higher frequency sounds at rates that permit locking to the envelope of the stimulus . A variety of stimulus patterns containing multiple rates were therefore then tested . Fig . 9A shows the results of stimulating the ensemble with a repeated pattern that consists of two stimuli with a 1400 Hz inter-stimulus interval followed by a pause with in inter-stimulus interval corresponding to 700 Hz . This amounts to a two-stimulus 1400 Hz burst applied repeatedly at 466 . 6 Hz . Even with multiple low frequency components in the stimulus train , there exist clear differences in the ability of individual neurons in a gradient to lock to the stimulus train , either in the presence or absence of spontaneous activity ( Fig . 9B , C ) . Again , in the presence of spontaneous activity , adjustment of KHT in all the neurons to the optimal value found among the individual neurons ( corresponding in this case to an intermediate value of 0 . 348 µS ) substantially enhances PE for the output of the ensemble ( Fig . 9D ) .
The microsecond time scale and precision with which the auditory system operates implies that even simple aspects of an auditory stimulus , such as its temporal envelope or its location in space cannot be encoded in the activity of single neurons but must be distributed across an ensemble of neurons . The simulations presented here indicate that the accuracy that is required for estimates of timing of incoming stimuli is improved either by an orderly gradation of intrinsic excitability within the ensemble , or by random spontaneous activity . Maximal accuracy of the ensemble occurs , however , in the presence of spontaneous activity when K+ conductance within an ensemble of neurons can be adjusted to that present in individual neurons that initially respond with highest degree of phase-locking . Random spontaneous activity plays two distinct roles in improving the temporal accuracy of an ensemble . The first is to increase the entrainment of the output of the entire ensemble . Consider , for example , an ensemble that is stimulated at 600 Hz but in which the individual neurons respond at 300 Hz with perfect phase-locking . In the absence of spontaneous activity , the value of Phase/Entrainment parameter PE will be 0 . 5 , and the output of the ensemble will be a perfectly timed 300 Hz train of postsynaptic potentials . If , as was shown in Fig . 5D , spontaneous activity randomizes the timing of the onset of firing in individual neurons , the combined postsynaptic output will be a 600 Hz train of postsynaptic potentials and the PE value will be closer to 1 . 0 . The second effect of spontaneous activity is to randomize the timing of “errors” such as those shown in Fig . 1B for high KHT levels , thereby decreasing their impact on the output of the ensemble . For spontaneous activity to be effective , however , it must be random across the ensemble , with little or no correlation between individual units . A significant degree of correlation of the spontaneous activity among individual neurons is likely to be interpreted as a sensory signal , and could potentially contribute to the condition of tinnitus in humans . The high rates of spontaneous activity in the auditory nerve and in auditory brainstem neurons in the absence of sounds are known to be generated by spontaneous transmitter release from inner hair cells [16] , [17] , [18] , [19] . For this mechanism to be effective would therefore require a relatively low degree of convergence of inputs from individual hair cells to individual brainstem neurons that constitute the presynaptic neurons in these simulations . The contributions of a gradient of KHT , and the effect of modulation of levels of this current to timing accuracy are distinct from those of spontaneous activity . The level of KHT in individual neurons determines both the number of errors of timing that result from the relative refractory period that follow each action potential ( e . g . Fig . 1B left ) and whether or not the neurons are able to respond to stimulation at different rates ( Fig . 1B right ) . For a given stimulus pattern applied to an ensemble with a KHT gradient , some neurons will respond with greater temporal accuracy than others . In addition , the diversity of intrinsic excitability in a gradient disperses the timing of the onset of firing in individual neurons , aiding the entrainment of the output of the full ensemble . In general , therefore , even in the absence of spontaneous activity the output of an ensemble with a KHT gradient will be improved over that of one with a randomly-selected uniform level of KHT . A central conclusion of this work is , however , that rapid adjustment of the K+ conductance in those neurons with a suboptimal KHT to that of the optimal value for a specific stimulus pattern will , in the presence of spontaneous activity , provide the most accurate temporal representation of the input . This can result in some “flattening” of the gradient in the face of a fixed maintained auditory stimulus , as has been observed experimentally ( see Fig . 4 of reference [29] ) . Of course a subsequent change in the stimulus pattern may require a further change in the value of KHT . In this study only gradients of the high threshold Kv3 . 1-like KHT potassium current were considered . There is an abundant experimental evidence for the existence of gradients in this channel [26] , [31] , [43] and auditory stimulation has been demonstrated to produce very rapid changes in Kv3 . 1 current levels [34] , [44] . In addition , longer-term changes in auditory stimulation produce long-term changes in levels of Kv3 . 1 protein and KHT currents in individual neurons and adjust the overall tonotopic gradient of channel expression [27] , [29] , [30] , [31] , [38] , [45] . Gradients of expression along the tonotopic medial-to-lateral axis of auditory brainstem nuclei have also been described for a variety of other ion channels in many species , and are not restricted to animals with specific ranges of hearing frequency [9] , [26] , [28] , [31] , [34] , [46] . In the MNTB of mammals , these include the low-threshold Kv1 . 1-like , Kv1 . 3 and KNa currents [26] , [32] , [33] . Gradients also exist in the lateral superior olive , a target of MNTB axons [47] . Similar gradients have been characterized using both electrophysiological and molecular approaches in two different auditory nuclei of chickens and barn owls [46] , [48] . Such tonotopic gradients are not confined to ion channels but have also been described for neurotransmitter receptors , synaptic proteins and signaling molecules [23] , [49] , [50] , [51] . Gradients and adjustments in other currents , such as low-threshold K+ current , can also have a similar effect on accuracy of timing as changes in KHT ( data not shown ) . Thus it is likely that changes in the gradient of other parameters that influence transmission through brainstem nuclei , such as the differences in size of somata , dendrites or axons or in neurotransmitter receptors would also contribute in much the same way to enhanced processing . The finding of gradients of expression of ion channels in the central auditory system was preceded by findings of similar gradients in sensory hair cells in the cochlea [52] , [53] , [54] , [55] , [56] . In some lower vertebrates , these differences in levels of ion channels in different hair cells serve to tune the electrical responses of the cells to specific sound frequencies [57] , [58] , [59] , [60] , [61] , [62] . As in the central nervous system , gradients in the periphery are not confined to ion channels but can be found for other signaling molecules such as calcium-binding proteins and are manifest in subtle differences in the structural properties of synapses along the tonotopic axis [54] , [63] , [64] . To measure the accuracy of timing of the outputs of ensemble of neurons , the present simulations calculated the summed outputs as combined postsynaptic currents , as would be recorded experimentally in the voltage-clamp mode . This allowed timing information in the output to be evaluated in a way that is uncontaminated by intrinsic conductances in the postsynaptic cell . How this information is then used and shaped by intrinsic conductances in the postsynaptic cells then clearly depends on biological role of the specific circuit . Phase-locking neurons participate in circuits that encode both the frequency of incoming sound stimuli and their amplitude . Extraction of timing information would require the postsynaptic neurons to respond with high temporal fidelity to individual peaks in the postsynaptic potentials , as occurs in the AVCN and MNTB . Determination of amplitude , as occurs in the LSO , is likely to optimal when each input pulse in a train generates a postsynaptic current of equal size . Just as for timing information , this corresponds to the highest possible value for the PE parameter in the summed postsynaptic output . Thus the simulations are likely to be equally relevant to processing of timing and amplitude information . The present findings raise the question of which specific cellular and system-level mechanisms lead to the adjustment in levels of KHT in response to different patterns of stimulation . Normal physiological changes in the auditory environment produce rapid ( seconds to minutes ) changes in the state of phosphorylation of a serine residue in the cytoplasmic C-terminus of the Kv3 . 1 channel protein [34] . Specifically , the channel is normally phosphorylated in quiet environments but becomes dephosphorylated in louder environments . Such dephosphorylation results in an increase in Kv3 . 1 current , allowing neurons to fire at higher rates [34] , [36] . Interestingly , such phosphorylation is also organized tonotopically such that , in resting neurons , a greater proportion of Kv3 . 1 channel is phosphorylated at the medial low-frequency aspect of the MNTB [34] . Changes in phosphorylation state appear to occur coherently across large parts of the MNTB [34] , and may be mediated through cell-cell communication by messengers such as nitric oxide [44] . Thus acute changes in phosphorylation may be a key mechanism that adjusts KHT currents to optimize temporal accuracy . Longer-lasting changes in the auditory environment produce changes in the levels of Kv3 . 1 protein in neurons , most likely by increasing the rate of synthesis of new channel subunits [29] , [30] , [45] . The subset of neurons in which synthesis is enhanced depends on the frequency of the auditory stimulus that is used , and this has been shown to produce a clear change in the overall tonotopic distribution of Kv3 . 1 [30] . These findings suggest that , by altering the fine structure of tonotopic gradients [65] , the central auditory system may be able to adjust the timing and accuracy of processing in brainstem nuclei . Perhaps in concert with feedback from higher centers , this may allow fine discriminations of patterns of inputs in a given auditory environment [66] .
Simulations of individual presynaptic neurons in an ensemble of 50 neurons were carried out using a model similar to that used previously to describe the firing patterns of MNTB neurons [34] , [36] , [40] , [41] , [67] . Responses were simulated by integration of the equation:where INa represents Na+ current , IKHT and IKL represent components of voltage-dependent K+ currents . ILeak is the leak current . Individual neurons in an ensemble were stimulated by applying step currents Iext ( t ) ( 0 . 2 msec , 1 . 5 nA ) at rates of 100–1500 Hz , in the presence or absence of randomly timed stimuli with the same parameters . The capacitance C of each model neuron was 0 . 01 nF . Equations for INa , IKHT , IKL and IL were identical to those in Macica et al . ( 2003 ) , and are based on fits to currents in MNTB neurons . Specifically , The evolution of the variables m , h , n , l and r were given by equations of the formwhereand j = m , h , n , l , r . Kinetic parameters for the evolution of the variables m and h were gNa = 0·5 µS , kαm = 76 . 4 ms−1 , ηαm = 0 . 037 mV−1 , kβm = 6 . 93 ms−1 , ηβm = −0 . 043 mV−1 , and kαh = 0 . 000135 ms−1 , ηαh = −0 . 1216 mV−1 , kβh = 2 . 0 ms−1 and ηβh = 0 . 0384 mV−1 . For the KHT ( Kv3 . 1 ) current , values of gKHT were varied as described in the text , with kαn = 0 . 2719 msec−1 , ηαn = 0 . 04 mV−1 , kβn = 0 . 1974 msec−1 ηβn = 0 mV−1 , kαp = 0 . 00713 ms−1 , ηαp = −0 . 1942 mV−1 , kβp = 0 . 0935 ms−1 and ηβp = 0 . 0058 mV−1 . For the fixed “low threshold” KL potassium current , gKL = 0 . 02 µS , kαl = 1 . 2 ms−1 , ηαl = 0 . 03512 mV−1 , kβl = 0 . 2248 ms−1 , ηβl = −0 . 0319 mV−1 , kαr = 0 . 0438 ms−1 , ηαr = −0 . 0053 mV−1 , kβr = 0 . 0562 ms−1 and ηβr = −0 . 0047 mV−1 . The equations for Ntot ( usually 50 ) individual neurons in an ensemble were integrated numerically for a period of 120 ms , and stimulation with external current pulses Iext ( t ) was carried out for 100 msec beginning 10 ms after the onset of integration . For each integrated trace , the times of the occurrence of the peaks of the action potentials , tN , P , were first calculated , where N is the index for the number of the neuron in the linear array of Ntot neurons and P is the index for consecutive action potentials in the trace . The time of each action potential was defined as the time when the upstroke of the action potential crossed 0 mV . For each neuron , these were used to generate a set of step functions F1 , N , F2 , N…FX ( N ) , N , corresponding to the times of the peaks of action potentials , where X ( N ) is the total number of action potentials evoked in neuron N . ThusThese were in turn used to generate a normalized postsynaptic current trace S ( Nstart , Nend , t ) for the combined output of a group of neurons , where Nstart and Nend denote the positions of the starting and ending neurons in the linear array for which the combined postsynaptic trace was calculated: For the simulations τ was set at 2 . 0 ms . To calculate phase vector strength , the times of occurrences of peaks in each postsynaptic current trace S ( Nstart , Nend , t ) were first determined . These were used to calculate a set of delays Δ ( Nstart , Nend , u ) from the onset of each stimulus applied to the ensemble to the time of the largest peak detected in the postsynaptic trace during the subsequent inter-stimulus interval , where u is the index of each consecutive stimulus applied to the ensemble during the train . An initial phase vector strength V' was next calculated for the output of each linear subset of the ensemble:where Z is the inter-stimulus interval , umax is the number of stimuli that evoked a postsynaptic peak during inter-stimulus interval and , is the mean delay for these peaks . The amplitudes of the postsynaptic current peaks A ( Nstart , Nend , u ) for each of the stimuli in the train , as well as the maximal Smax ( Nstart , Nend ) and minimum Smin ( Nstart , Nend ) values of current were then calculated . Using the number of total stimuli that were applied to the ensemble during the train Utot , an adjusted phase vector PE that reflects the proportion of stimuli that evoked a postsynaptic response was then calculated . Thuswhere | In order to detect the nature and location of a sound stimulus , neurons in the central auditory system have to fire at very high rates with extreme temporal precision . Specifically , they have to be able to follow changes in an auditory stimulus at rates of up to 2000 Hz or more and to lock their action potentials to the stimuli with a precision of only a few microseconds . An individual neuron , however , cannot fire at such high rates , and the intrinsic electrical properties of neurons , such as the relative refractory period that follows each action potential , severely limits accuracy of timing at high rates . The intrinsic excitability of neurons is governed by the potassium channels that they express . It has been found in auditory brainstem nuclei that there exist gradients of these channels such that each neuron typically has a different number of channels than its neighbors . In this study , computational models based on measurements in auditory neurons demonstrate that , in the presence of random spontaneous activity such as is normally observed in auditory neurons , rapid adjustments of levels of potassium current within neurons along the gradient are required to allow the ensemble to transmit accurate timing information . The findings suggest that regulation of potassium channels within gradients is an integral component of auditory processing . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"biology"
] | 2012 | Gradients and Modulation of K+ Channels Optimize Temporal Accuracy in Networks of Auditory Neurons |
Ammonia is a major signal that regulates nitrogen fixation in most diazotrophs . Regulation of nitrogen fixation by ammonia in the Gram-negative diazotrophs is well-characterized . In these bacteria , this regulation occurs mainly at the level of nif ( nitrogen fixation ) gene transcription , which requires a nif-specific activator , NifA . Although Gram-positive and diazotrophic Paenibacilli have been extensively used as a bacterial fertilizer in agriculture , how nitrogen fixation is regulated in response to nitrogen availability in these bacteria remains unclear . An indigenous GlnR and GlnR/TnrA-binding sites in the promoter region of the nif cluster are conserved in these strains , indicating the role of GlnR as a regulator of nitrogen fixation . In this study , we for the first time reveal that GlnR of Paenibacillus polymyxa WLY78 is essentially required for nif gene transcription under nitrogen limitation , whereas both GlnR and glutamine synthetase ( GS ) encoded by glnA within glnRA operon are required for repressing nif expression under excess nitrogen . Dimerization of GlnR is necessary for binding of GlnR to DNA . GlnR in P . polymyxa WLY78 exists in a mixture of dimers and monomers . The C-terminal region of GlnR monomer is an autoinhibitory domain that prevents GlnR from binding DNA . Two GlnR-biding sites flank the -35/-10 regions of the nif promoter of the nif operon ( nifBHDKENXhesAnifV ) . The GlnR-binding site Ⅰ ( located upstream of -35/-10 regions of the nif promoter ) is specially required for activating nif transcription , while GlnR-binding siteⅡ ( located downstream of -35/-10 regions of the nif promoter ) is for repressing nif expression . Under nitrogen limitation , GlnR dimer binds to GlnR-binding siteⅠ in a weak and transient association way and then activates nif transcription . During excess nitrogen , glutamine binds to and feedback inhibits GS by forming the complex FBI-GS . The FBI-GS interacts with the C-terminal domain of GlnR and stabilizes the binding affinity of GlnR to GlnR-binding site Ⅱ and thus represses nif transcription .
Biological nitrogen fixation , the conversion of atmospheric N2 to ammonia ( NH3 ) , is carried out by a specialized group of prokaryotes and plays an important role in world agriculture [1] . Yet the great demands for nitrogen in modern agriculture far outstrip this source of fixed nitrogen , and chemical nitrogen ( N ) fertilizer is used extensively in agriculture . Overuses of N fertilizer in many parts of the world have led to soil , water , and air pollution [2] . Ammonia is a major signal that regulates nitrogen fixation in most diazotrophs [3 , 4] . Regulation of nitrogen fixation in the Gram-negative diazotrophs is well-characterized . In these bacteria , this regulation occurs mainly at the level of nif gene transcription , which requires a nif-specific activator , NifA [5] . NifA acts as an enhancer binding protein ( EBP ) that recognizes sequences ( TGT-N10-ACA ) , located upstream of the -24/-12 region of the promoters controlled by RNA polymerase containing the alternative σ54 factor [3 , 6–8] . Paenibacillus is a large genus of Gram-positive , facultative anaerobic , endospore-forming bacteria . The genus Paenibacillus currently comprises more than 150 named species , approximately 20 of which have nitrogen fixation ability , including eight novel species described by our laboratory [9] . Diazotrophic Paenibacilli has been extensively used as a bacterial fertilizer in agriculture [10] . However , the regulation mechanism of nitrogen fixation in response to nitrogen availability in Paenibacilli is not clarified , partially due to hardness in genetic transformation of these bacteria . Our recent studies by comparative genomic sequence analysis have revealed that a minimal and compact nif cluster comprising nine genes ( nifB nifH nifD nifK nifE nifN nifX hesA nifV ) encoding Mo-nitrogenase is conserved in 15 N2-fixing Paenibacillus strains [11] . Phylogeny analysis suggests that the ancestral Paenibacillus did not fix nitrogen . The N2-fixing Paenibacillus strains were generated by acquiring the nif cluster via horizontal gene transfer ( HGT ) from a source related to Frankia [11] . The 9 genes ( nifBHDKENXhesAnifV ) within the nif cluster are organized as an operon under control of a σA ( σ70 ) -dependent promoter located in front of nifB gene [12] . A global transcriptional profiling analysis revealed that nif gene transcription in P . polymyxa WLY78 was strongly regulated by ammonium and oxygen [13] . However , unlike Gram-negative diazotrophs , diazotrophic Paenibacilli have no nifA gene encoding transcriptional activator NifA and no NifA-binding site in the nif promoter region . But a glnR gene and GlnR/TnrA-binding sites in the promoter region of the nif operon are conserved in the 15 diazotrophic Paenibacillus strains by comparative genomics analyses [11] , indicating the role of GlnR as a regulator of nitrogen fixation . GlnR is a central regulator of nitrogen metabolism in the class Bacilli , and the glnR gene in the diazotrophic Paenibacilli is not associated with the transferred nif gene cluster , indicating that Paenibacillus GlnR is indigenous . The recent studies with Surface Plasmon Resonance ( SPR ) experiments have demonstrated that GS stabilizes the binding of GlnR to nitrogen fixation gene operators in Paenibacillus riograndensis SBR5 [14] . However , these studies did not fully investigate the regulatory mechanism of GlnR in nitrogen fixation . GlnR and TnrA are the two transcriptional regulators for the regulation of nitrogen metabolism in the Gram-positive model organism Bacillus subtilis [15 , 16] . They were previously recognized as the members of the MerR family regulators according to their common winged-HTH ( helix-turn-helix ) domains [17] . However , the recent studies have revealed that TnrA and GlnR are a new family of dimeric DNA-binding proteins with C-terminal , flexible , effector-binding sensors that modulate their dimerization that represents a separate branch of the MerR family proteins [18] . TnrA/GlnR form weak dimers by hydrophobic residues located on its winged-HTH and residues in its N-terminal helix [18] , whereas MerR proteins form tight dimers via their extended C-terminal coiled coils [19] . Both of GlnR and TnrA proteins of B . subtilis have a high sequence similarity at their N terminal domains and bind a common consensus sequence ( 5’-TGTNAN7TNACA-3’ ) , but the C terminal domains of these proteins differ completely [20–24] . GlnR of B . subtilis generally acts as a repressor repressing gene or operons required for ammonium assimilation like the glnRA operon , tnrA and ureABC ( the urease gene cluster ) under nitrogen-excess condition [15 , 16 , 25] . In contrast , TnrA serves in most cases as an activator , for instance activating ammonia transport ( nrgAB = amtBglnK ) , ureABC , nitrate and nitrite reduction ( nasABCDEF ) and its own gene ( tnrA ) [23] , whereas in a few cases , it acts like GlnR as a repressor repressing alsT ( encoding an H+/Na+ amino acid symporter ) [26] , gltAB ( encoding glutamate synthase ) [27 , 28] and ilvBHC-leuABCD ( encoding branched-chain amino acid biosynthesis proteins ) [29] . During excess nitrogen , glutamine ( Gln ) binds to and feedback inhibits glutamine synthetase ( GS , the product of glnA ) ) by forming the complex FBI-GS . Formation of the feedback-inhibited GS ( FBI-GS ) signals the presence of excess nitrogen and transmits that signal by interacting with and affecting the DNA-binding and transcription programs of both GlnR and TnrA . Under nitrogen limitation , the C-terminal region of GlnR folds back and forms an autoinhibitory helix that prevents dimer formation and thus inhibits DNA binding [18 , 20–22] . Under excess nitrogen , FBI-GS functions as a chaperone by a transient interaction with the GlnR autoinhibitory domain and relieves autoinhibition , shifting the equilibrium from the inhibited form to the DNA-binding active form and thus turning on GlnR repression [18 , 20–22] . In contrast , FBI-GS forms a stable complex with TnrA , inhibiting its DNA-binding function under excess nitrogen , whereas TnrA is released from FBI-GS , allowing TnrA dimerization and activation of its transcription program under nitrogen limitatiom . GlnK appears to play an ancillary role in TnrA dimerization by acting as a templating agent for TnrA [25–30] . In this study , we fully investigate the regulation mechanisms of nitrogen fixation in P . polymyxa WLY78 by using comprehensive molecular methods . We reveal that during nitrogen limitation , GlnR binds to GlnR-binding site Ⅰ located upstream of -35/-10 regions of nif promoter of nif operon ( nifBHDKENXhesAnifV ) in a weak and transient association way and then activates nif transcription . During excess nitrogen , glutamine ( Gln ) binds to and feedback inhibits glutamine synthetase ( GS ) by forming the complex FBI-GS . FBI-GS interacts with C-terminal domain of GlnR and stabilizes the binding of GlnR to site Ⅱ located downstream of nifB transcription start codon and thus represses nif transcription . GS encoded by glnA within glnRA operon is involved in regulation of nif transcription . Also , overexpression of glnR and mutagenesis of glnA or GlnR-binding site Ⅱ led to constitutive nitrogen fixation in the absence or presence of ammonia . Our study not only reveals the novel regulation mechanisms of nif gene expression in Paenibacilli , but also provides insight into dual active and repressive functions of GlnR .
The genome of P . polymyxa WLY78 contains a glnR gene and two paralogs of glnA , but it lacks a tnrA gene [11] . We found that of the two glnA genes , one was linked to glnR as a dicistronic glnRA operon and the other ( here designated as glnA1 ) was elsewhere in the genome . The current analysis by using BLAST alignment showed that GS and GS1 proteins encoded by the glnA and glnA1 genes had 39% identity . To elucidate the function of GlnR in nitrogen fixation of Paenibacillus , we constructed an in-frame deletion mutant ΔglnR , a complemention strain ( ΔglnR/glnR ) for the mutated glnR and an overexpression strain ( WT/glnR ) , as described in S1 Fig . In comparison with wild-type P . polymyxa WLY78 which exhibited the highest nitrogenase activity in the absence of NH4+ and no activity in the presence of more than 5 mM NH4+ , activity in ΔglnR mutant was at basal constitutive levels under all conditions ( Fig 1A ) . Deletion of glnR resulted to nearly loss of activity , indicating that GlnR is essentially required for nitrogen fixation under nitrogen limitation . Somewhat higher activity was observed in the ΔglnR mutant at high ammonia than in the ammonia-repressed wild-type strain . Complementation of ΔglnR with a single copy of glnR integrated on the amyE site of its genome restored nitrogenase activity to the wild-type level in complemented strain ( ΔglnR/glnR ) , suggesting that change of nitrogenase activity was due solely to deletion of glnR . Overexpression of glnR by introduction of glnR carried on multicopy vector pHY300PLK into wild-type strain led to enhancement of activity in the presence of NH4+ . The ΔglnR and the wild-type strains exhibited similar growth phenotypes on minimal media with glutamine , glutamate and ammonium as sole nitrogen sources ( S2 Fig ) . Taken together , these results indicate for the first time that GlnR positively regulates nitrogen fixation under nitrogen-limited condition . To examine the effect of glnR mutation on the transcription of nif genes in P . polymyxa WLY78 , the transcription levels of nifH , nifD and nifK were determined by qRT-PCR . As shown in Fig 1B , the transcription levels of the nifHDK in wild-type strain exhibited more than 1000-fold of increase under nitrogen-limited condition ( 2 mM glutamate as sole nitrogen ) compared to nitrogen-excess condition ( 2 mM glutamate + 100 mM NH4+ ) . However , the nifHDK genes in ΔglnR mutant were expressed constitutively under both conditions at very low level which was approximately 2 . 7% of that observed in wild-type strain under nitrogen-limited condition ( Fig 1B ) . These results are consistent with nitrogenase activity in this ΔglnR mutant , indicating that GlnR activates nif transcription under nitrogen-limited condition . To further examine the effect of GlnR on regulation of nif expression , a transcriptional lacZ fusion to nif promoter region was constructed and then this Pnif-lacZ fusion was introduced into wild-type and ΔglnR mutant , respectively . As shown in Fig 1C , the β-galactosidase levels produced by Pnif-lacZ fusion in wild-type strain were 5000-fold higher in nitrogen-limited condition than in nitrogen-excess condition . However , the β-galactosidase levels produced by Pnif-lacZ fusion in ΔglnR mutant were similar in both conditions . The data are consistent with the above described qRT-PCR results and nitrogenase activities . Furthermore , qRT-PCR analysis demonstrated that glnR transcription was highly induced under nitrogen-limited condition compared to under nitrogen-excess condition ( Fig 1D ) , suggesting that glnR expression itselfis nitrogen-dependent . Also , the transcription profiles of glnR and nifH were similar under nitrogen limitation ( Fig 1E ) . The current results are consistent with our previous global transcriptional profiling analysis that the expressions of glnR and nif genes were significantly up-regulated when P . polymyxa WLY78 was grown in N2-fixing condition ( without O2 and NH4+ ) compared to non-N2-fixing condition ( air and 100 mM NH4+ ) [13] . These results indicate that the expressions of glnR and nif genes are highly coordinated . To examine the role of GS proteins encoded by glnA and glnA1 in regulation of nitrogen fixation , a series of in-frame-deletion mutants , including ΔglnA1 , ΔglnA and ΔglnRA mutants , and their complementary strains ΔglnA/glnA and ΔglnRA/glnRA were constructed as described in S1 Fig . We found that nitrogenase activities were similar in ΔglnA1 mutant and wild-type strain under both nitrogen-limited and -excess conditions , suggesting that glnA1 is not involved in regulation of nitrogen fixation ( Fig 2A ) . However , nitrogenase activity in ΔglnA mutant was produced constitutively at modest level under both nitrogen-limited and -excess conditions . Complementation of ΔglnA with glnA gene ( complementary strain ΔglnA/glnA ) restored the nitrogenase activity to the wild-type level ( basal nitrogenase activity ) under nitrogen-excess condition and to 80% of wild-type level ( high nitrogenase activity ) under nitrogen-limited condition ( Fig 2A ) . Nitrogenase activity in ΔglnRA double mutant was almost abolished just as observed in ΔglnR single mutant . Complementation study showed that glnRA could partially restored the activity of ΔglnRA double mutant ( Fig 2A ) , suggesting that the role of GS is dependent on GlnR . These results indicate that both GS and GlnR are required for the repression of nitrogen fixation under nitrogen-excess condition . qRT-PCR analysis showed that the transcription levels of nifH gene were similar in both ΔglnA1 and wild-type strains under both nitrogen-limited and -excess conditions , in agreement with nitrogenase activity in these strains and suggesting that GS encoded by glnA1 is not involved in regulation of nif gene expression ( Fig 2B ) . In contrast , the nifH gene in ΔglnA mutant was transcribed constitutively at modest level under both nitrogen-limited and -excess conditions , in agreement with nitrogenase activity in ΔglnA mutant . Transcription levels of the nifH gene in ΔglnRA double mutant were at basal low level under both nitrogen-limited and -excess conditions , in agreement with nitrogenase activity in this strain . These data suggest that GlnR and GS encoded by glnA within glnRA operon are responsible for negative regulation of nif gene expression according to nitrogen availability . GlnR protein of B . subtilis has a high sequence similarity at the N terminus with TnrA , but the C-terminal signal transduction domain of GlnR is sequentially distinct from TnrA and contains an extra 15 residues [20 , 23 , 24] . Here , sequence alignments showed that the 137-residue GlnR protein of P . polymyxa WLY78 exhibited 54% and 40% identity with GlnR and TnrA of B . subtilis 168 , respectively ( S3 Fig ) . GlnR , GS and GS1 from P . polymyxa WLY78 with His6-tag at the N-terminus were overexpressed and purified in Escherichia coli , respectively . Also , GlnRΔ25 , a truncated GlnR with a deletion of the last 25 C-terminal codons ( aa 113–137 ) was overexpressed and purified in E . coli . Of these purified proteins , GlnR was further evaluated by size-exclusion chromatography analysis . The GlnR protein was eluted as a broad peak with two maxima . Judged from the elution positions of marker proteins , this profile could reflect the coexistence of His6-GlnR monomers and dimers in non-instantaneous equilibrium ( sequence-deduced masses of His6-GlnR monomers and dimers are 19 . 6 kDa and 39 . 2 kDa a respectively ) ( S4A Fig ) . SDS-PAGE revealed for the two maxima the same band with the expected mass for His6-GlnR ( S4B Fig ) . Our results are different from some reports that P . riograndensis SBR5 GlnR is mainly the dimeric form [14] and B . subtilis GlnR is mainly monomeric form [20] . Then , the interaction of GlnR with GS proteins was evaluated by surface plasmon resonance ( SPR ) assay . His6-tagged GlnR was immobilized on a Ni-nitrilotriacetic acid-activated chip sensor surface . Then different concentrations of GS and FBI-GS ( GS and glutamine ) were loaded onto the GlnR chip surface . In the absence of glutamine , only a weak interaction between GlnR and GS was observed , and the GlnR-GS complex also dissociated quickly even when the concentration of GS was increased from 200 nM to 3 . 2 mM ( Fig 3A ) . In contrast , in the presence of glutamine , there was still strong interaction between GlnR and FBI-GS even when the concentration of GS protein was decreased from 200 nM to 6 . 25 nM ( Fig 3B ) . These results indicated that GlnR of P . polymyxa WLY78 interacted with the feedback inhibited GS form ( FBI-GS ) . Our results are consistent with the reports that GS , in its feedback inhibited form , interacts with GlnR of P . riograndensis SBR5 [14] and GlnR and TnrA of B . subtilis [18 , 20 , 21] . However , nearly no interaction between GlnRΔ25 and FBI-GS was observed ( Fig 3C ) , suggesting that the C-terminal domain is required for interaction between GlnR and FBI-GS . In contrast , only a basal weak interaction between GlnR and GS1 was detected whether the feedback inhibitor glutamine was present or not ( Fig 3C ) , in agreement with the above-described results that mutation of glnA1 did not affect regulation of nif transcription and nitrogenase activity . We predicted that the promoter region of nif operon of P . polymyxa WLY78 contained two GlnR-binding sites: GlnR-binding site Ⅰ and GlnR-binding site Ⅱ ( Fig 4A and S5 Fig ) by using MEME/MAST software [31] . The two sites were 118 bp separate . Site Ⅰ was located 58 bp upstream of -35 regions of nif promoter , and site Ⅱ was seated 24 bp downstream of the nifB transcription start site . The binding motifs of the two sites resembled the common consensus sequences ( 5’-TGTNAN7TNACA-3’ ) of the GlnR/TnrA-binding site [15 , 32 , 33] . To determine whether the two GlnR-binding sites are direct targets of GlnR , the in vitro and in vivo binding of GlnR protein to the two GlnR-binding sites were performed by using electrophoretic mobility shift assays ( EMSA ) , surface plasmon resonance ( SPR ) spectroscopy and chromatin immunoprecipitation-quantitative PCR ( ChIP-qPCR ) . EMSA experiments revealed that in vitro GlnR bound to the two sites ( Fig 4B ) . Addition of both GS and glutamine ( the feedback-inhibited GS ) enhanced the binding affinity of GlnR to the two sites ( Fig 4B ) . Also , the addition of both GS and glutamine did not change the band positions of the DNA-GlnR protein complex , supporting that FBI-GS did not directly bind to DNA and it functioned as a chaperon to activate the DNA-binding activity of GlnR [20 , 21] . Then , ChIP-qPCR experiments were performed to investigate the in vivo binding of GlnR to the two GlnR-binding sites . GlnR polyclonal antibody was used to measure binding of GlnR to its target and qRT-PCR with primers corresponding to the GlnR-binding site Ⅰ and site Ⅱ was performed . As shown in Fig 4C , GlnR bound to the both sites under both nitrogen limitation and nitrogen excess conditions , but the binding levels of GlnR to both sites were much higher under excess nitrogen than under nitrogen limitation . These findings agree with the results obtained by EMSA . Also , Fig 4C shows that the binding level of GlnR to site Ⅱ was higher than to site Ⅰ under both conditions . Furthermore , we tested the in vitro affinity of GlnR to the two GlnR-binding sites by SPR spectroscopy . This SPR assay demonstrated that GlnR alone could specifically bind to the two GlnR-binding sites , but the affinity of GlnR for site Ⅱ was much stronger than for site Ⅰ . Regardless of the interaction intensity with each site , in the absence of glutamine , GlnR-DNA binding was transient and unstable due to quick dissociation ( Fig 4D ) . However , addition of both GS and glutamine ( the feedback-inhibited GS ) greatly stabilized the DNA-protein complex . Especially , FBI-GS significantly stabilized the DNA ( site Ⅱ ) -protein complex , consistent with the classical function of GlnR as a repressor [15 , 16] . The affinity of the truncated GlnR ( GlnRΔ25 ) protein to the two GlnR-binding sites was also investigated by SPR assay . As shown in Fig 4E , in comparison with wild-type GlnR , GlnRΔ25 protein had higher affinity for both sites . The addition of FBI-GS greatly stabilized the DNA-GlnR complex , but it did not have obvious effect on the DNA-GlnRΔ25 complex . Our results indicate that the C-terminal region of P . polymyxa GlnR is an autoinhibitory domain that inhibits DNA-binding ability of GlnR and that the C-terminal domain is also required for the interaction between FBI-GS and GlnR , consistent with the observations in B . subtilis GlnR [20 , 21] . To clarify the affinity of GlnR to both sites , quantitative evaluation was carried out with SPR . A double-stranded DNA oligomer that contained the sequence of site Ⅰ or site Ⅱ was fixed onto the chip as described in Material and Methods . Different concentrations of GlnR protein were loaded onto the DNA chip surface . As shown in Fig 5A and 5B , there was no binding signal in the absence of GlnR and the binding signals became strong with the increase of concentrations of GlnR protein . When the concentration of GlnR was increased to 500 nm and 1000 nM , an obvious binding of GlnR to site Ⅰ was found , but it dissociated quickly , indicating that the binding of GlnR to site Ⅰ is transient and unstable . In contrast , the binding of GlnR to site Ⅱ was stronger and it dissociated slowly , indicating that the binding of GlnR to site Ⅱ is stronger than to site Ⅰ due to slow dissociation . The corresponding KA and KD values for site Ⅰ were calculated to be 1 . 09×106 and 9 . 16×10−7 , respectively . Whereas the KA and KD values for site Ⅱ were 1 . 86×107 and 5 . 37×10−8 , respectively ( Fig 5C ) . The values of KA for site Ⅱ was consistently higher than that for site Ⅰ , and the values of KD for site Ⅱ was much lower than that for site Ⅰ , indicating that affinity of site Ⅱ for GlnR is higher than site Ⅰ . Our current results are different from those obtained in P . riograndensis SBR5 where GlnR bound to the two GlnR-binding sites [PnifM ( 1 ) and PnifM ( 2 ) ] of the main nif gene cluster at similar levels and whose GlnR affinity for site Ⅰ was slightly higher than for site Ⅱ [14] . Since GlnR was positively and negatively involved in the regulation of nitrogen fixation according to nitrogen availability , both increase and decrease of nitrogenase activity could be expected through mutations of the two GlnR-binding sites in the nif promoter region . Thus , the site-specific mutagenesis of the two GlnR-binding sites was performed . As shown in Fig 6A , the consensus sequence TGACGT in site Ⅰ region was replaced with a restriction site of Kpn Ⅰ ( GGTACC ) via homologous recombination , generating mutant MPnif1 . The consensus motif ATAACG in site Ⅱ was replaced by a restriction site of Cla Ⅰ ( ATCGAT ) , which generated the mutant MPnif2 . A double mutant MPnif3 with mutations of both GlnR-binding sites was generated . Also , the mutant MPnif97 with deletion of site Ⅰ was also constructed . EMSA confirmed that GlnR did not bind to the mutated sites ( Fig 6B ) . In comparision with wild-type strain , only basal nitrogenase activity was observed in both mutants MPnif1 and MPnif97 under both nitrogen-limited and -excess conditions ( Fig 6C ) , suggesting that site Ⅰ is essentially required for nitrogen fixation . The data are consistent with nitrogenase activity in ΔglnR mutant , indicating that site Ⅰ is the target of GlnR . In contrast , nitrogenase activity in mutant MPnif2 was derepressed partially under nitrogen-excess condition , suggesting that site Ⅱ is involved in repressing nitrogen fixation . The data are consistent with nitrogenase activity in ΔglnA mutant , suggesting that site Ⅱ is the target of GS encoded by glnA . Nitrogenase activity in the double mutant MPnif3 was nearly abolished under both nitrogen-limited and excess conditions , in agreement with nitrogenase activities in ΔglnRA double mutant . The nif gene transcription levels determined by qRT-PCR ( Fig 6D ) were consistent with the nitrogenase activities in mutants MPnif97 , MPnif1 , MPnif2 and MPnif3 . Taken together , these results indicate that GlnR binds to site Ⅰ to activate nif expression under nitrogen-limited condition and binds to site Ⅱ to repress nif transcription under nitrogen-excess condition . FBI-GS is involved in repressing nif transcription by its interaction with GlnR under excess nitrogen .
GlnR is a global transcription regulator of nitrogen metabolisms found extensively in Bacillus and other Gram-positive bacteria . It generally acts as a repressor repressing the transcription of glnRA operon , tnrA and ureABC in B . subtilis under excess nitrogen [15 , 16 , 25] . TnrA is another transcription regulator of nitrogen metabolisms found mainly in Bacillus and it serves in most cases as an activator under nitrogen limitation [26] . In the present work , we reveal that P . polymyxa GlnR simultaneously acts as an activator and a repressor for nitrogen fixation by binding to different loci of the single nif promoter region according to nitrogen availability . GS is ne cessarily required for nif repression mediated by GlnR . In this study , two GlnR-binding sites flanking the -35/-10 regions of the promoter of nif operon in P . polymyxa WLY78 is predicted by software and then confirmed by in vitro EMAS and SPR experiments and by in vivo ChIP-qPCR . The two sites are 118 bp separated . Site Ⅰ is located 58 bp upstream of -35 region of nif promoter , and site Ⅱ is seated 24 bp downstream of the nifB transcription start site . The location of site Ⅰ is an indicative of activation site , since regulator , such as TnrA , bound at this position most likely activates gene transcription [23 , 25 , 26] . Site Ⅱ located just downstream of promoter is an indicative of repression site , since regulator bound at this site will represses gene transcription by sterically hindering RNA extension [34] . The binding motif ( 5’-TGTAAGGGAATATAACG-3’ ) of site Ⅱ possesses the common consensus sequence ( 5’-TGTNAN7TNACA-3’ ) of the GlnR-binding site ( S5 Fig ) , while the consensus sequences ( 5’-CGATATATTACTTGACG-3’ ) of site Ⅰ fit the TnrA-specific motif ( 5’-NGNNAN7TNACN-3’ ) which clearly lacks the conserved A and T at the 3′ and 5′ end [15 , 32 , 33] . Our studies of deletion or mutagenesis of GlnR-binding site Ⅰ , site Ⅱ and both sites demonstrated that site Ⅰ is responsible for activating nif expression and site Ⅱ is required for repressing nif transcription , in agreement with the locations of the two sites . Two GlnR-binding sites flanking the -35/-10 nif promoter region were also found in P . riograndensis SBR5 [14] , but they exhibit a little difference with those of P . polymyxa WLY78 in the precise consensus sequences and locations . As shown in S6 Fig , the site Ⅱ in P . polymyxa WLY78 is located 16 bp upstream of ATG ( translation start site ) , while the OA-nifB ( site Ⅱ ) in P . riograndensis SBR5 is located 60 bp upstream of ATG . A common consensus sequence ( TGTNAN7TNACA ) of GlnR-binding motif in the class Bacillus ( S5 Fig ) is more conserved in the two GlnR-binding sites of P . riograndensis SBR5 than in those of P . polymyxa WLY78 . SPR assay demonstrated that GlnR-binding site Ⅰ of P . riograndensis SBR5 displayed higher affinity for GlnR , whereas the second site had lower affinity and dissociated faster [14] . In contrast , GlnR-binding site Ⅰ of P . polymyxa WLY78 exhibited lower affinity for GlnR and dissociated faster , while site Ⅱ displayed higher affinity due to slow dissociation , especially in the presence of FBI-GS . Based on the two binding sites in the nif promoter region of P . riograndensis SBR5 , a DNA-looping model that represents a strong and strict regulation for nif genes was proposed [14] . In this model , DNA loop formation was induced by two GlnR dimers bound to both GlnR-binding sites and bridged by feedback-inhibited GS . However , our results from deletion and complementation analyses of glnR , glnA and glnRA and from mutation analyses of GlnR-binding sites did not support the DNA-looping model . Our data demonstrate evidently that GlnR bound to site Ⅰ in a weak and transient way and then activated nif gene transcription under nitrogen limitation , and the FBI-GS stabilized the binding affinity of GlnR to binding site Ⅱ and the strong binding of GlnR to site Ⅱ repressed nif gene transcription by interfering RNA extension under nitrogen-excess condition . Our studies that mutation of the 4–5 nucleotides in the half-sequences within the GlnR-binding site Ⅰ ( ACGATATATTACTTGACGT ) or site Ⅱ ( ATGTAAGGGAATATAACGT ) resulted to no binding of GlnR are consistent with that 4 nucleotides in each operator half-site of the DNA consensus sequence ( TGTNAN7TNACA ) were required for GlnR/TnrA specific DNA binding [18 , 23 , 26 , 35] . Our data also support that TnrA/GlnR form a weak symmetric dimer by binding their palindromic cognate sites [18 , 21] . Thus we think that it is unlikely for GlnR to be a tetramer formed by interacting between two dimers bound on the two GlnR-binding sites . There are also two GlnR-binding sites in the promoter region of B . subtilis glnRA , one of which lies immediately upstream of the -35 promoter element and the other site overlaps the -35 region [36] . It was previously reported that GlnR bond to these sites in a cooperative manner , and both sites were required for full repression of B . subtilis glnRA [37] . Mutation of GlnR-binding site Ⅱ made the mutant MPnif2 have nitrogenase activities and express nif genes under both nitrogen-limited and -excess conditions ( Fig 6C and 6D ) , supporting that GlnR bound to site Ⅰ to activate nif gene transcription under nitrogen limitation . However , the nitrogenase activity and nif gene transcription in mutant MPnif2 did not reach similar levels under both nitrogen-limited and -excess conditions . We deduce that perhaps the mutation of site Ⅱ made the mutant MPnif2 have more FBI-GS proteins to strength the binding of GlnR to site Ⅰ and then interfere nif gene transcription under nitrogen excess . However , under normal physiological condition , since there are two GlnR-binding sites and the affinity of GlnR for site Ⅱ was much stronger than for site Ⅰ , repression of nif gene transcription was mediated by site Ⅱ . Whether site Ⅰ , together with site Ⅱ was involved in repressing nif gene transcription under nitrogen-rich condition needs to be determined in the future . Our study by deletion , complementation and overexpression of glnR , glnA and glnRA and by mutagenesis or deletion of GlnR-binding sites reveals that GlnR bound to GlnR-binding site Ⅰ and activated nif transcription under nitrogen limitation , and GlnR bound to GlnR-binding site Ⅱ and repressed nif transcription under excess nitrogen . The novel , dual positive and negative regulatory mechanism is for the first time reported in nitrogen fixation . Although dual function of GlnR in Streptomyces hygroscopicus var . jinggangensis 5008 was reported [38] , Streptomyces GlnR is an OmpR-like response regulator which does not display any similarity to the Paenibacillus/Bacillus GlnR regulator belonging to the MerR family [39] . Although GlnR protein alone could bind to the two sites in nif promoter of P . polymyxa WLY78 , this binding was transient and unstable . We deduce that the transient GlnR-DNA interaction is sufficient for GlnR to act as an active regulator . Interestingly , under nitrogen-limited condition , TnrA ( but not GlnR ) of B . subtilis is further stabilized by an interaction with GlnK [40 , 41] . It also reported that GlnR protein exhibited an increased affinity for the glnRA operon promoter when bound to GlnK in Streptococcus mutans [42] . Whether GlnR dimer is stabilized by GlnK in P . polymyxa WLY78 under nitrogen-limited condition needs to be investigated in the future . It is well characterized that the C-terminal domain of B . subtilis GlnR protein is sequentially distinct from TnrA and contains an extra 15 residues [23] . This region acts as an autoinhibitory domain that prevents GlnR dimerization and thus inhibits DNA binding [20–22] . FBI-GS acts as a chaperone to stabilize dimerization and subsequent DNA binding of GlnR [20–22] . In this study , the P . polymyxa GlnRΔ25 , a truncated GlnR with a deletion of the last 25 C-terminal codons was overexpressed and purified in E . coli . SPR analyses show that the interaction between GlnRΔ25 and GS is greatly decreased compared to wild-type GlnR . Also , GlnRΔ25 had higher binding affinity to both GlnR-binding sites than wild-type GlnR ( Fig 4E ) . The addition of FBI-GS greatly enhanced the DNA-binding affinity of wild-type GlnR , but it did not obviously increase the DNA-binding affinity of GlnRΔ25 protein . FBI-GS also stabilized the DNA-GlnR complex , but it had no effect on the DNA-GlnRΔ25 complex . These results reveal that the C-terminal region of P . polymyxa GlnR is an autoinhibitory domain and it is also involved in the interaction between FBI-GS and GlnR , in agreement with the results obtained in B . subtilis GlnR . Our results demonstrate that FBI-GS stabilizes the binding of GlnR to the two site , especially site Ⅱ , consistent with the observations in B . subtilis GlnR [20 , 21] . We deduce that the monomers in the mixture of dimers and monomers of P . polymyxa GlnR protein were shifted to dimers by the interaction of FBI-GS with GlnR under excess nitrogen . Consequently , the strong binding of GlnR to site Ⅱ led to the represstion of nif transcription by interfering RNA extension . This mode of repressing nif gene transcription of P . polymyxa during excess nitrogen is a classical function of GlnR regulator found extensively in B . subtilis and some other Gram-positive bacteria . Although the activity of TnrA is also controlled by FBI-GS , the mechanisms of regulation are different between TnrA and GlnR . Under excess nitrogen , FBI-GS forms a stable complex with TnrA , which inhibits its DNA-binding activity [20 , 24] . TnrA and GlnR are generally recognized as the members of the MerR family regulators according to their common winged-HTH ( helix-turn-helix ) domains . However , TnrA and GlnR may regulate transcription using molecular mechanisms distinct from MerR proteins . MerR proteins activate transcription by distorting and realigning DNA promoters with nonoptimal spacing between the -10 and -35 boxes [43] . Unlike MerR members , the promoters bound by TnrA and GlnR are optimally arranged and a 17-bp inverted repeat sequences with the consensus TGTNAN7TNACA constitutes the minimal binding site for these proteins [15 , 32] . It was previously suggested that TnrA functions primarily as an activator by binding operator DNA sites and recruiting RNA polymerase ( RNAP ) [24 , 26] , whereas GlnR does not bind RNAP and hence functions as a repressor . The recent study on structures has revealed that GlnR induces bend and conformational changes in the DNA similar to those in TnrA [18] , supporting our results that GlnR functions as an activator just as TnrA does under nitrogen limitation . The wild-type P . polymyxa WLY78 has the highest nitrogenase activity in the absence of NH4+ and has no activity in the presence of more than 5 mM NH4+ ( Fig 1A ) . Deletion of glnR leads to loss of both nitrogenase activity and nif gene transcription under nitrogen limitation , suggesting that GlnR is essential required for activating nif gene expression . Deletion of glnA makes the ΔglnA mutant have both nitrogenase activity and nif gene transcription under both nitrogen-limited and -excess conditions , suggesting that GS encoded by glnA is involved in repressing nif gene transcription under nitrogen-excess condition . Mutation of GlnR-binding site Ⅰ results to loss of both nitrogenase activity and nif transcription under nitrogen limitation , suggesting that site Ⅰ is responsible for activating nif gene transcription . However , mutation of GlnR-binding site Ⅱ makes the mutant MPnif2 have both nitrogenase activities and nif gene transcriptions under both condition , consistent with the nitrogenase and nif gene ranscription in ΔglnA mutant . Our study with SPR also demonstrates that the affinity of GlnR for site Ⅱ is stronger than for site Ⅰ . Under nitrogen-excess condition , glutamine is synthesized and it feedbacks the GS , yielding FBI-GS , EMSA , SPR and Chip-PCR reveal that the presence of FBI-GS ( GS and glutamine ) greatly stabilizes the GlnR-DNA complex and decreases the dissociation of GlnR from binding site Ⅱ . According to our results , we proposed a regulatory model of GlnR involved in nitrogen fixation in P . polymyxa WLY78 ( Fig 7 ) . GlnR exists in a mixture of dimers and monomers . Monomer of GlnR is an autoinhibitory form whose C-terminal region folds back and inhibits dimer formation . Under nitrogen-limited condition , GlnR dimer binds to site Ⅰ in a weak and transient association way and then activates nif expression ( Fig 7A ) . Although GlnR also sequentially or simultaneously binds to site Ⅱ , binding of GlnR to this site does not repress nif transcription due to GlnR having only a weak and transient association with DNA during this condition . Also , the large amounts of GlnR produced under this condition enable nif transcription to carry on , since expression of glnR itself is nitrogen-dependent . Under nitrogen-excess condition ( Fig 7B ) , glutamine is in excess and it binds to and feedback inhibits GS by forming the complex FBI-GS . The FBI-GS interacts with the C-terminal tail of GlnR and relieves autoinhibition , shifting the monomer to the DNA-binding active form . The FBI-GS further stabilizes the binding affinity of GlnR to both sites , especially site Ⅱ . The stable binding of GlnR to site Ⅱ blocks the RNA extension and thus represses nif transcription . In conclusion , our combined data reveal a novel molecular regulatory mechanism of nitrogen fixation in P . polymyxa WLY78 . GlnR binds to site Ⅰ to activate nif gene transcription under nitrogen-limited condition , and it binds to site Ⅱ to repress nif gene transcription under nitrogen-excess condition . The activity of GlnR is controlled by GS in response to nitrogen availability .
Bacterial strains and plasmids used in this study are summarized in S1 Table . P . polymyxa strains were grown in nitrogen-limited medium ( 2 mM glutamate ) or nitrogen-excess medium ( 2 mM glutamate +100 mM NH4+ ) under anaerobic condition [12] . For assays of nitrogenase activity , β-galactosidase assays and nif expression , P . polymyxa strains were grown in nitrogen-limited medium or nitrogen-excess medium under anaerobic condition . Escherichia coli strains JM109 and BL21 ( DE3 ) were used as routine cloning and protein expression hosts , respectively . Thermo-sensitive vector pRN5101 [44] was used for gene disruption in P . polymyxa . Shuttle vector pHY300PLK was used for complementation experiment and transcriptional fusion construction . pET-28b ( + ) ( Novagen ) was used for expressing recombinant His6-tagged protein in E . coli . When appropriate , antibiotics were added in the following concentrations: 100 μg/ml ampicillin , 25 μg/ml chloramphenicol , 12 . 5 μg/ml tetracycline , 50 μg/ml kanamycin , and 5μg/ml erythromycin for maintenance of plasmids . The four in-frame-deletion mutants: ΔglnR , ΔglnA1 , ΔglnA and ΔglnRA , were constructed by a homologous recombination method . The upstream ( ca . 1 kb ) and downstream fragments ( ca . 0 . 5 kb ) flanking the coding region of glnR , glnA1 , glnA and glnRA were PCR amplified from the genomic DNA of P . polymyxa WLY78 , respectively . The primers used for these PCR amplifications were listed in S2 Table . The two fragments flanking each coding region of glnR , glnA1 , glnA and glnRA were then fused with SalⅠ/BamHⅠ digested pRN5101 vector using Gibson assembly master mix ( New England Biolabs ) , generating the four recombinant plasmids pRDglnR , pRDglnA1 , pRDglnA and pRDglnRA , respectively . Then , each of these recombinant plasmids was transformed into P . polymyxa WLY78 as described by [45] , and the single-crossover transformants were selected for erythromycin resistance ( Emr ) . Subsequently , marker-free deletion mutants ( the double-crossover transformants ) ΔglnR , ΔglnA1 , ΔglnA and ΔglnRA were selected from the initial Emr transformants after several rounds of nonselective growth at 39°C and confirmed by PCR amplification and sequencing analysis . Complementation for ΔglnR , ΔglnA and ΔglnRA was performed . For complementation of ΔglnR , the glnR gene and its promoter was inserted into the amyE site on genome of ΔglnR strain . To do this , two fragments: an 1161 bp DNA fragment and an 1017 bp fragment flanking the amyE gene , were PCR amplified from the genomic DNA of P . polymyxa WLY78 , respectively . An 803 bp DNA fragment carrying the glnR ORF ( 414 bp ) and its own promoter ( 389 bp ) was also PCR amplified . Then , three fragments and the vector pRN5101 digested with BamHⅠ and HindⅢ were fused together using Gibson assembly master mix , generating the recombinant plasmid pRCglnR . The recombinant plasmid pRCglnR was transformed into the cells of ΔglnR strain and then double-crossover transformants were selected after several rounds of growth at 39°C . Finally , the complementation strain CglnR which contains an 803 bp DNA fragment carrying glnR ORF and its promoter integrated on the amyE site was obtained and confirmed by PCR and DNA sequencing . For complementation of ΔglnRA mutant , a 2116 bp DNA fragment containing the complete glnRA operon and its own promoter was PCR amplified from the genomic DNA of P . polymyxa WLY78 . For complementation of ΔglnA , a 1419 bp DNA fragment containing the coding region of glnA and a 280 bp promoter region of glnRA operon were PCR amplified , respectively , and then the two fragments were fused together using Gibson assembly master mix . These fragments were digested with BamHⅠ/SalⅠ , and ligated into vector pHY300PLK , generating glnA-complemented vector pHYglnA and glnRA-complemented vector pHYglnRA , respectively . Each of these recombinant plasmids was correspondingly transformed into ΔglnA and ΔglnRA mutants , and tetracycline-resistant ( Tetr ) transformants were selected and confirmed by PCR and sequencing . The strain WT/glnR in which glnR is overexpressed was also constructed . An 803 bp DNA fragment carrying the glnR ORF ( 414 bp ) and its own promoter ( 389 bp ) was PCR amplified and then ligated to multicopy vector pHY300PLK and then transformed to P . polymyxa WLY78 , generating the glnR overexpression strain . The primers used here are listed in S2 Table . Four mutants with deletion or mutagenesis of the GlnR-binding site ( s ) were performed via homologous recombination . A 313 bp nif promoter region ( from -253 to +60 relative to the nifB transcription start codon ) containing both of the GlnR-binding sites Ⅰ and Ⅱ was used as a target for mutation . Thus , three 313-bp DNA fragments PnifM1 , PnifM2 and PnifM3 ( S3 Table ) were synthesized based on the sequences of nif promoter region . Notably , PnifM1 contains the mutated GlnR-binding site Ⅰ where the last six base pairs TGACGT within the 19-bp consensus sequences ( ACGATATATTACT TGACGT ) of the GlnR-binding site Ⅰ were replaced by a restriction site of KpnⅠ ( GGTACC ) . PnifM2 carries the mutated GlnR-binding site Ⅱ where the last six base pairs ATAACG within the 19-bp consensus sequences ( ATGTAAGGGAAT ATAACG ) was replaced by a restriction site of ClaⅠ ( ATCGAT ) . PnifM3 contains both of mutated site Ⅰ and site Ⅱ where the consensus motifs TGACGT and ATAACG were simultaneously replaced by restriction sites KpnⅠ and ClaⅠ . The three DNA fragments PnifM1 , PnifM2 and PnifM3 were then cloned to plasmid pUC19 , respectively . Then , each of the three fragments PnifM1 , PnifM2 and PnifM3 ( S3 Table ) was PCR amplified from the recombinant plasmids . Two homologous arms ( 1205 bp and 1101 bp ) flanking the 313 bp region in nif promoter were amplified from the genomic DNA of P . polymyxa WLY78 using the primers MPnif1/MPnif2 and primers MPnif5/MPnif6 ( S4 Table ) , respectively . Each of the two arms contains ca . 20 bp overlap with the above-described 313 bp DNA fragments ( PnifM1 , PnifM2 and PnifM3 ) . Then , the two arms and the DNA fragments PnifM1 , PnifM2 and PnifM3 were assembled to the BamHⅠ/HindⅢ digested plasmid vector pRN5101 , yielding the recombinant plasmids pRMP1 , pRMP2 , pRMP3 . Each of these recombinant plasmids was introduced into P . polymyxa WLY78 by transformation . The single-crossover transformants were selected for erythromycin resistance ( Emr ) . Subsequently , the double-crossover transformants were selected from the initial Erythromycin resistance transformants after several rounds of nonselective growth at 39°C . These mutants were confirmed by PCR amplification using the primers and subsequent digestion with KpnⅠ or ClaⅠ and then by DNA sequencing . The mutant with deletion of site Ⅰ was also constructed as follows . A 1418 bp DNA upstream fragment and an 1101 bp downstream fragment were PCR amplified from the genomic DNA of P . polymyxa WLY78 . The two fragments were assembled to vector pRN5101 , yielding the recombinant plasmid pRMP100 and then the plasmid was transformed into P . polymyxa WLY78 . The mutant with deletion of 213 bp fragment ( from -40 bp to -253 bp relative to the nifB transcription start codon ) containing GlnR-binding site Ⅰ was obtained as described above . A 313 bp of the native nif promoter ( Pnif ) ( from -253 to +60 relative to the nifB transcription start codon ) containing both of the GlnR-binding sites Ⅰ and Ⅱ was amplified from the genomic DNA of P . polymyxa WLY78 using primers LPnif1 and LPnif2 ( S5 Table ) . The lacZ coding region was PCR amplified with primers LPnif3 and LPnif4 from the plasmid pPR9TT . The two PCR-amplified fragments were fused together with vector pHY300PLK and then it was transformed into P . polymyxa WLY78 . The glnR , truncated glnR ( GlnRΔ25 , for C-terminal deletion of GlnR , removing the last 25 amino acid residues ) , glnA within glnRA operon and glnA1 were PCR amplified from the genomic DNA of P . polymyxa WLY78 , respectively . These PCR products were cloned into pET-28b ( + ) ( Novagen ) to construct tagged proteins with His-tag at the N-terminus and then transformed into E . coli BL21 ( DE3 ) . The recombinant E . coli strains were cultivated at 37°C in LB broth supplemented with 50 μg/ml kanamycin until mid-log phase , when 0 . 2 mM IPTG was added and incubation continued at 20°C for 8 hours . Cells were collected and disrupted in a lysis buffer ( 50 mM NaH2PO4 , 300 mM NaCl , 10 mM Imidazole ) by sonication on ice . Recombinant His6-tagged proteins in the supernatant were purified on Ni2-NTA resin ( Qiagen , Germany ) according to the manufacturer’s protocol . Fractions eluted with 250 mM imidazole were dialyzed into storage buffer ( 10 mM Tris-HCl pH7 . 5 , 1 mM EDTA , 80 mM NaCl , 20% ( v/v ) glycerol ) for antibody production or binding buffer ( 20 mM HEPES pH 7 . 6 , 1 mM EDTA , 10 mM ( NH4 ) 2SO4 , 1 mM DTT , 0 . 2% Tween 20 , 30 mM KCl ) for electrophoretic mobility shift assays ( EMSA ) and HBS-Mg buffer ( 10 mM HEPES pH 7 . 4 , 300 mM NaCl , 3 mM MgCl2 , and 0 . 005% Nonidet P-40 ) for surface plasmon resonance spectroscopy ( SPR ) . Purified His-GlnR was used to raise polyclonal rabbit antibody ( Beijing Protein Innovation ) and for size-exclusion chromatography . Primers used here are listed in S5 Table . Transcription levels of genes were compared among P . polymyxa WLY78 strain and ΔglnR , ΔglnA and ΔglnRA mutants by quantitative real-time RT-PCR ( qRT-PCR ) analysis . At each experimental time point , 50 ml of culture were harvested and rapidly frozen under liquid nitrogen . Total RNAs were extracted with RNAiso Plus ( Takara , Japan ) according to the manufacturer’s protocol . Remove of genome DNA and synthesis of cDNA were performed using PrimeScript RT reagent Kit with gDNA Eraser ( Takara , Japan ) . qRT-PCR was performed on Applied Biosystems 7500 Real-Time System ( Life Technologies ) and detected by the SYBR Green detection system with the following program: 95°C for 15 min , 1 cycle; 95°C for 10 s and 65°C for 30 s , 40 cycles . Primers used for qRT-PCR are listed in S6 Table . The relative expression level was calculated using ΔΔCt method . 16S rRNA was set as internal control and the expression levels of genes in WT strain under nitrogen-excess condition were arbitrarily set to 1 . 0 . Each experiment was performed in triplicate . EMSAs were performed as described previously using a DIG Gel Shift Kit ( 2nd Generation; Roche , USA ) [12] . The promoter fragments of nif operon were synthesized by Sangon Biotech Co . , Ltd ( Shanghai ) . Two DNA fragments corresponding to the sequences of the first strand and the complementary DNA strand were synthesized . The two strands were annealed and then labeled at the 30 end with digoxigenin ( DIG ) using terminal transferase , and used as probes in EMSAs . Each binding reaction ( 20 μl ) consisted of 1 μg poly [d ( A-T ) ] , 0 . 3 nM labelled probe , and various concentrations of purified His6-GlnR in the binding buffer . Reaction mixtures were incubated for 30 min at 25°C , analyzed by electrophoresis using native 5% polyacrylamide gel run at 4°C with 0 . 5×TBE as running buffer , and electrophoretically transferred to a positively charged nylon membrane ( GE healthcare , UK ) . Labelled DNAs were detected by chemiluminescence according to the manufacturer’s instructions , and recorded on X-ray film . The primers used here are listed in S6 Table . Chromatin immunoprecipitation-quantitative PCR ( ChIP-qPCR ) was performed as described by [46] . 100 ml of culture of WT or ΔglnR grown in nitrogen-limited or -excess media at 30°C were harvested and immersed in cross-linked buffer ( 0 . 4 M sucrose , 1 mM EDTA , 10 mM Tris-Cl , pH 8 . 0 ) with 1% formaldehyde and 1% PMSF for 20 min at 28°C . Cross-linking was stopped by addition of glycine ( final concentration 125 mM ) and incubation for another 5 min . After cross-linking , cells were sonicated to break chromosomal DNA into 200–500 bp fragments . Supernatant containing 2 mg total protein was diluted in 1 ml lysis buffer . 5 μl GlnR polyclonal antibody was added into precleared supernatant and incubated overnight at 4°C . Purified immunoprecipitated DNA was resuspended in 120 μl double-distilled water . 2 μl DNA was used for qPCR , using the primer pairs listed in S6 Table . Relative levels of GlnR-coprecipitated DNAs were determined by comparison with negative controls . SPR experiments [47] were carried out using Biacore 3000 SPR sensor ( Biacore AB , Uppsala , Sweden ) . All assays were carried out at 25°C . HBS buffer supplied with 3 mM MgCl2 ( 10 mM HEPES pH7 . 4 , 300 mM NaCl , 3 mM MgCl2-6H2O , 0 . 2 mM EDTA , and 0 . 005% Tween-20 ) was used as the running buffer . Protein-DNA interaction assays were performed with Sensor Chip SA . First , a biotinylated single-stranded DNA capture linker ( biotin-GCAGGAGGACGTAGGGTAGG ) was irreversibly bound to the chip . DNA oligomer used for SPR assays ( S7 Table ) were designed and synthesized based on nif promoter region harboring GlnR-binding sites and containing a single-stranded overhang complementary to the linker . Then a partially double-stranded DNA oligomer that contained the GlnR-binding site Ⅰ or site Ⅱ in the double-stranded region with a single-stranded overhang complementary to the capture linker was fixed onto the chip , reaching a signal of 250 RU . Control DNA was fixed onto flow cell 1 ( FC1 ) , and DNA containing GlnR binding sites were fixed onto flow cell 2 and 3 ( FC2 , FC3 ) . GlnR with or without 25 nM GS protein was injected at a flow rate of 30 ul/min . Protein-Protein interaction assays were performed with Sensor Chip CM5 . GlnR was immobilized via amine groups onto all four flow cells , receiving a signal of 1000 RU . Purified GS with or without 1 mM glutamine were injected separately at a flow rate of 30 μl/min . Purified His6-GlnR from E . coli was used for size-exclusion chromatography . Analytical size-exclusion chromatography was carried out on an Akta purifier system equipped with a Superdex 200 column 10/300 ( geometric column volume of 24 mL GE Healthcare ) . The running buffer contains 50 mM Tris-HCl ( pH 7 . 4 ) and 300 mM NaCl . His6-tagged GlnR was diluted on running buffer to reach a concentration of 2 mg/ml . 1 mL filtered and centrifuged sample was injected at a flow rate of 0 . 3 ml/min . Purified His6-GlnR from E . coli was used for size-exclusion chromatography . Analytical size-exclusion chromatography was carried out on an Akta purifier system equipped with a Superdex 200 column 10/300 ( geometric column volume of 24 mL GE Healthcare ) . The running buffer contains 50 mM Tris-HCl ( pH 7 . 4 ) and 300 mM NaCl . His6-GlnR was diluted on running buffer to reach a concentration of 2 mg/ml . 1 mL filtered and centrifuged sample was injected at a flow rate of 0 . 3 ml/min . The apparent molecular weights of proteins were estimated after calibration of the column with standard proteins: thyroglobulin ( 670 kDa ) , globulin ( 158 kDa ) , ovalbumin ( 44 kDa ) , myoglobin ( 17 kDa ) , vitamin B12 ( 1 . 35 kDa ) ( Bio-Rad gel filtration standard ) . Acetylene reduction assays were performed as described previously to measure nitrogenase activity [12] . P . polymyxa WLY78 and its mutant strains were grown in 5 ml of LD media ( supplemented with antibiotics ) in 50 ml flasks shaken at 250 rpm for 16 h at 30°C . The cultures were collected by centrifugation , washed three times with sterilized water and then resuspended in nitrogen-deficient medium containing 2 mM glutamate containing 2 mM glutamate plus 0–100 mM NH4Cl as nitrogen source under anaerobic condition to a final OD600 of 0 . 2–0 . 4 . Here , nitrogen-deficient medium containing 2 mM glutamate as nitrogen source and nitrogen-excess medium containing 2 mM glutamate and 100 mM NH4Cl as nitrogen source are generally used . Then , 1 ml of the culture was transferred to a 25-ml test tube and the test tube was sealed with robber stopper . The headspace in the tube was then evacuated and replaced with argon gas . After incubating the cultures for 6–8 h at 30°C with shaking at 250 rpm , C2H2 ( 10% of the headspace volume ) was injected into the test tubes . After incubating the cultures for a further 3 h , 100 ml of culture was withdrawn through the rubber stopper with a gas tight syringe and manually injected into a HP6890 gas chromatograph to quantify ethylene ( C2H4 ) production . The nitrogenae activity was expressed in nmol C2H4/mg protein/hr . All treatments were in three replicates and all the experiments were repeated three or more times . β-galactosidase activity was assayed according to the method described by [48] . Each experiment was performed in quintuplicate . | GlnR is a global transcription regulator of nitrogen metabolism in Bacillus and other Gram-positive bacteria . GlnR generally functions as repressor and inhibits gene transcription under excess nitrogen . Our study for the first time reveals that GlnR simultaneously acted as an activator and a repressor for nitrogen fixation of Paenibacillus by binding to different loci of the single nif promoter region according to nitrogen availability . In excess glutamine , the feedback inhibited form of glutamine synthetase ( GS ) encoded by glnA within glnRA operon directly interacts with the C-terminal domain of GlnR and then controls the GlnR activity . Also , overexpression of glnR or deletion of glnA or mutagenesis of GlnR-binding site Ⅱ led to constitutive nif expression in the absence or presence of high ( 100 mM ) concentration of ammonia . This work represents the first instance of a dual positive and negative regulatory mechanism of nitrogen fixation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"bacillus",
"microbiology",
"dna-binding",
"proteins",
"plant",
"physiology",
"organic",
"compounds",
"dna",
"transcription",... | 2018 | Positive and negative regulation of transferred nif genes mediated by indigenous GlnR in Gram-positive Paenibacillus polymyxa |
Diseases caused by Aedes-borne viruses , such as dengue , Zika , chikungunya , and yellow fever , are emerging and reemerging globally . The causes are multifactorial and include global trade , international travel , urbanisation , water storage practices , lack of resources for intervention , and an inadequate evidence base for the public health impact of Aedes control tools . National authorities need comprehensive evidence-based guidance on how and when to implement Aedes control measures tailored to local entomological and epidemiological conditions . This review is one of a series being conducted by the Worldwide Insecticide resistance Network ( WIN ) . It describes a framework for implementing Integrated Aedes Management ( IAM ) to improve control of diseases caused by Aedes-borne viruses based on available evidence . IAM consists of a portfolio of operational actions and priorities for the control of Aedes-borne viruses that are tailored to different epidemiological and entomological risk scenarios . The framework has 4 activity pillars: ( i ) integrated vector and disease surveillance , ( ii ) vector control , ( iii ) community mobilisation , and ( iv ) intra- and intersectoral collaboration as well as 4 supporting activities: ( i ) capacity building , ( ii ) research , ( iii ) advocacy , and ( iv ) policies and laws . IAM supports implementation of the World Health Organisation Global Vector Control Response ( WHO GVCR ) and provides a comprehensive framework for health authorities to devise and deliver sustainable , effective , integrated , community-based , locally adapted vector control strategies in order to reduce the burden of Aedes-transmitted arboviruses . The success of IAM requires strong commitment and leadership from governments to maintain proactive disease prevention programs and preparedness for rapid responses to outbreaks .
IAM proposes to tailor vector control responses according to the following 5 scenarios based on the local stage of Aedes distribution and level of virus transmission risk: Scenario 1 ( or S1 ) , no Aedes present ( and no transmission ) ; S2 , Aedes locally established and no transmission; S3 , Aedes widely established and sporadic transmission; S4 , Aedes widely established and endemic transmission; and S5 , Aedes widely established and epidemic transmission ( Fig 1; Box 1 ) . Risk scenarios are not fixed in space ( at country , province , or district levels ) or time and are likely to evolve based on updates in entomological and epidemiological risk assessment . The IAM aims to provide ‘graduated’ responses according to the risk level , but the switch from one scenario to another does not systematically follow a ‘linear transition’ . For example , Key West , Florida , transitioned directly from S3 to S5 during the 2017’s Zika outbreak . Similarly , La Reunion Island and Italy switched from S3 to S5 during the chikungunya outbreak in 2007 . Note that S4 ( ‘endemic transmission’ ) will typically be applied to viruses that have been established in a given location of some time , e . g . , endemic dengue . A novel introduction and spread of a new arbovirus ( e . g . , Zika or chikungunya ) or a new dengue serotype can rapidly produce a transition to an outbreak ( S5 ) . Integrated surveillance is an ongoing systematic collection , recording , analysis , interpretation , and dissemination of data to aid control efforts for initiating suitable public health interventions for prevention and control , including the M&E of the implemented control measures [12] . Collecting and using data in this way is intended to support assessment of risk for introduction and spread of vectors and viruses and to monitor and evaluate the control efforts followed by adjustments over time if necessary in accordance with predefined indicators ( Table 1 and S1 Table ) . Entomological and epidemiological surveillance data should be promptly integrated , which will require efficient collaboration between vector-control and public health programs , and made available on a shared , easily accessible platform ( e . g . , the DHIS2 , https://www . dhis2 . org/overview ) . Surveillance of Aedes vectors is important for identifying changes in geographic distribution , to obtain relative measurements of variation in vector density over time , to facilitate appropriate and timely decisions regarding interventions , and to assess the entomological impact of mosquito control programs to see whether the intervention had the expected effect on the target mosquito population [13] . For routine surveillance , entomological measurements have to be done in the same location ( sentinel sites ) at regular time intervals in order to establish a baseline to follow variation over time ( seasonal dynamics ) . The frequency of data collection should be based on programme capacity and the need to generate reliable data in an appropriate format . Given that the geographic distribution of Aedes is increasing globally , a systematic surveillance for Aedes in every country is needed . Surveillance at points and/or routes of entry , such as sea ports , airports , and land country borders is important for early detection of the introduction of invasive Aedes species ( S1 , Table 1 ) . If mosquitoes are introduced into suitable habitats , they may become established locally ( S2 ) or more widely ( S3 , S4 , S5 ) . At this stage , surveillance consists of monitoring the spread of the mosquitoes ( e . g . , using ovitrap networks ) in order to identify areas and/or periods of high transmission risk based on vector infestation , e . g . , mapping seasonal dynamics and disease hot-spot identification [14] . The presence and abundance of Aedes species are estimated from measures of different entomological indices ( e . g . , larvae , pupae , adult ) , each with their strengths and weaknesses , as summarised in S2 Table . It should be mentioned , however , that cross-sectional studies have failed to find good correlations between entomological indices and episodes of dengue [15] , and no larval entomological thresholds have proven effective in predicting Aedes-borne virus epidemics [16] . This can be explained by the fact that dengue virus transmission is complex and varies through time and space , and the relationship between vector density and risk of human infection is not static nor adequately characterised through periodic entomological surveillance [15] . New technologies ( i . e . , geo-informatics tools , remote sensing , and mathematical and simulation models ) can be helpful in mapping the spatial distribution of vectors and/or in predicting their spread and seasonal dynamics using climatic ( e . g . , temperature , rainfall ) , social ( e . g . , rent value or education level ) , demographic ( e . g . , population density or distance to urban habitats ) , and landscape ( e . g . , vegetation cover or type of urbanisation ) variables and can be less expensive than field surveillance [17] . A good understanding of the typology and productivity of habitats suitable for mosquito larval development ( S3 , S4 ) is essential in order to target larval control operations . Due to the possibility of introduction or selection of resistant individuals , insecticide resistance should be monitored regularly , preferably during nonepidemic periods ( S3 and S4 ) to guide the choice of insecticides used for mosquito control . A combination of biological , biochemical , and molecular tools can be used to measure the frequency , intensity , and mechanisms of insecticide resistance in natural populations . Each resistance testing tool has its own advantages and weakness [18] . In areas with virus transmission ( S3 , S4 , S5 ) , if resources permit , it can be helpful to screen vectors for virus infection in order to confirm the role played by suspected species in local transmission , to monitor spatial and temporal patterns in virus transmission dynamics , and to evaluate interventions . Because virus detection in mosquitoes is costly and time-consuming and finding infected mosquitoes in natural populations is often challenging , it is not regularly done for surveillance of Aedes-transmitted viruses [19] . In the field of Aedes-borne viral diseases , the objectives of epidemiological ( or human and possibly animal ) surveillance are to ( 1 ) evaluate potential public health threats , carrying out risk assessments and detecting outbreaks early; ( 2 ) select and evaluate the effectiveness of control activities; and ( 3 ) monitor trends in public health burden to obtain data for assessing the social and economic impact on the affected community ( Table 1 ) [20] . The threat to public health is assessed by identifying recent introductions of a virus ( S3 ) , monitoring travellers returning from areas where target viruses circulate ( S3 , S5 ) , and mapping local spread of the virus ( S4 , S5 ) . This kind of surveillance system requires robust indicators and action plans to be defined in order to stratify epidemiological risk and guide decision-making to facilitate switching from one epidemic scenario to another [21] . Epidemiological data play a key role for guiding and prioritising vector control responses , which will be graduated according to transmission risk ( Fig 1 ) . Firstly , attention should be focussed on timely detection of imported viraemic people ( S3 ) , followed by identifying hot-spots of virus transmission ( S4 , S5 ) . Monitoring geographical and temporal trends in human cases is common to all scenarios . The main differences lie in the targets of surveillance—trends in virus importation ( mainly for S3 and S5 ) , trends in circulating strains or serotypes , disease incidence , morbidity and mortality ( S4 and S5 ) . Evidence of non−vector-borne transmission ( e . g . , sexual transmission and blood transfusions ) should be investigated where appropriate ( e . g . , Zika infections ) as well as potential sylvatic transmission of the viruses , e . g . , yellow fever . In geographical areas where local virus circulation is already established ( S4 , S5 ) , a significant proportion of all suspected cases should be confirmed by specific laboratory diagnostics , depending on local resources [13] . Epidemiological surveillance is most often based on a combination of active and passive surveillance in order to reconcile cost , sensitivity , response time , and geographical coverage . Depending on the purpose of the epidemiological surveillance system and the risk scenario , certain considerations—such as available resources ( e . g . , human and diagnostic capacities ) or strength of the healthcare system ( e . g . , public and/or private and accessibility ) , and sensitivity or response capacity—will be critical for guiding stakeholders in their choice of design of the overall disease management system ( Table 1 ) . Epidemiological data , however , have several limitations . The most notable challenges are a high proportion of people with asymptomatic and/or mild infections , or differential diagnosis with low specificity of symptoms; a broad range of disease manifestations , from no detectable illness to death; lack of standardisation in case definitions; limited or low diagnostic capacity; underreporting; and variation in treatment-seeking behaviour by infected people [22] . Active case detection in the surroundings of a person with a confirmed infection may help identify additional cases or clusters , which often go unreported or undiagnosed . Where there is an epidemic alert , passive surveillance can be enhanced to reduce delays in reporting cases or to extend the area of surveillance . In areas at risk of sporadic transmission ( S3 ) , healthcare personnel usually have to report cases of imported and autochthonous arboviral disease , such as dengue , chikungunya , or Zika , to public health authorities . Increasing awareness among clinicians and travellers returning from endemic areas combined with good laboratory capacity has greatly improved case reporting . Laboratory-based surveillance has been shown to play a role in monitoring the introduction of a novel dengue virus serotype , a switch of virus strains between vector species and cocirculation of different arboviruses [22] . In nonendemic and/or nonepidemic areas ( S2 , S3 ) , surveillance can target imported cases because these represent the main threat for introduction into immunologically naive populations . This can be achieved by health professionals notifying the relevant authorities of suspected or confirmed imported cases [23] or by fever screening travellers at points of entry [24] . Evidence-based risk assessment should be carried out within all risk scenarios ( S1−S5 ) and should be conducted by national and international health ( and environmental ) agencies . The assessment should form the basis for developing guidelines for the actions needed to keep risk to a minimum . To our knowledge , there is no global framework for conducting risk assessment for Aedes-borne diseases , but several regional documents have been drawn up ( see Table 1 ) . Regular M&E of the delivery of dengue prevention and control services and of the impact of interventions ( this one being a critical one ) are important IAM activities in all scenarios . Suitable indicators for measuring the progress of implementation ( e . g . , intervention coverage ) and the outputs and outcomes ( e . g . , reductions in vector density or disease ) should be identified [13] . Vector control efforts need to be sustained over time , which requires well-structured administration , coordination with the public health programme that is diagnosing cases , political will , skilled staff , funding , and , crucially , community engagement and mobilisation from the outset [4] . Vector control can be undertaken either as a ‘routine’ activity ( i . e . , a preemptive sequence of actions regularly carried out ) or as an ‘emergency’ measure ( i . e . , a response to an excess of vectors and/or an unusual increase in the human disease incidence calling for immediate action ) . Both types of measures should be prepared for , but vector control is most cost-effective if it is ‘proactive’ ( preventive ) rather than implemented ‘in response mode’ ( after the start of epidemic ) [25] . Because programmes move from areas prone to virus introduction ( S3 ) to endemic-epidemic ( S4−S5 ) scenarios , a shift in the allocation of resources from ‘reactive’ to ‘proactive’ vector control should be considered . Targeting immature mosquitoes has been a prevalent paradigm for Aedes control , but far more attention should be directed at methods targeting both larvae and adults to maximise impact on adult Aedes density , longevity , and role in virus transmission [9] . Strategies and interventions should be adapted to local vector ecology and available resources , guided by results from operational research and subject to routine M&E ( see the ‘Risk assessment and M&E outcomes’ section ) . In order to prolong the life span of existing insecticides , it is imperative that noninsecticide-based tools are used whenever possible . When chemicals have to be deployed , they should be used rationally and preferably not as ‘monotherapies’ [10 , 26] . Several insecticide resistance monitoring strategies exist in vector control , which are based on rotations of insecticides , mixtures of unrelated insecticides , use of interventions in combination , and mosaic spraying . Resistance management is not a ‘stand-alone’ strategy and should be implemented in the broad context of integrated vector management and be carefully monitored and evaluated . Activities to control transmission should target homes and outdoor areas in their immediate vicinity ( i . e . , in the place of residence as well as in neighbouring houses ) . Treating nonresidential areas—i . e . , places where human–vector contact occurs during the daytime , such as schools , hospitals and workplaces , especially their surroundings , such as outside lunch gathering areas—can provide measurable impacts [27] . Restricting control to residences within a certain radius of a case’s home is not as effective as uniform treatment of broad geographic areas . By the time a case is detected , human movement has taken the virus beyond a radius of 100 m to 200 m [27 , 28] . The evidence base for the public health value of Aedes vector control is unfortunately weak . There are little data demonstrating reduction in human infection or disease for many tools currently in widespread use [29] . Epidemiological outcomes are needed to demonstrate the public health benefit of a vector control intervention and are the basis of evidence-based policy [30] . In order to provide more robust guidance on preferred Aedes vector control tools and those that should be avoided , we summarised existing evidence based on recent systematic reviews . We categorised the hierarchy of evidence according to whether there was epidemiological or entomological evidence and by study design , with randomised controlled trials providing the highest quality of evidence and nonrandomised or observational studies providing the lowest quality evidence [30] . Results and specified details for different interventions are shown in S2 Table . The strengths and limitations of current adult mosquito control methods and the strength of evidence for their entomological and epidemiological effects are summarised in Table 2 . Despite widespread use , there is limited entomological and epidemiological evidence for ultra-low volume ( ULV ) space spraying [31 , 32] . In the case of virus transmission ( S3 , S4 , S5 ) , ‘peridomestic’ or ‘perifocal’ space spray treatments with insecticides can be carried out in and around households where human infection is suspected or has been reported . Space sprays can also be adequate in specific situations , for example , ( i ) to prevent local establishment of invasive mosquito species , such as A . albopictus ( small area < 25 km2 , S2 ) , ( ii ) to halt an incipient outbreak ( S3 ) , and ( iii ) to curtail an ongoing epidemic and/or endemic situation ( S4 and S5 ) . Different treatment methods ( house-to-house application using portable equipment , vehicle-mounted fogging , and cold or thermal fogging ) are available , but they must be tailored to the risk scenario , the area to be covered , accessibility , and the Aedes species . Indoor space spray ( ISS ) should be distinguished from outdoor applications . Because A . aegypti tends to be endophilic and endophagic [33] , only in cases in which A . albopictus ( or A . aegypti ) populations are primarily outdoors are outdoor applications likely to be effective . Outdoor space spraying ( OSS ) and outdoor residual spraying ( ORS ) on vegetation have been used for controlling the exophilic species A . albopictus [34] , with some entomological evidence of efficacy ( Table 2 , S2 Table ) . The efficacy of ORS and its impact on the environment is still controversial , facing the same challenges as described above for ULV space spraying . Indoor residual spraying ( IRS ) , in particular targeted IRS ( TIRS , in which IRS is performed on exposed low walls , under furniture and inside closets ) has not been widely used for Aedes control so far , although it may be a promising tool for controlling Aedes-borne arboviral transmission ( S3 and S4 ) in areas where the endophilic mosquito A . aegypti is responsible for transmission [35] . The costs , human resources , and logistics needed for suitable coverage with IRS may represent a challenge for their rapid and broad-scale deployment during outbreaks . The use of contact tracing technologies to deploy IRS and/or TIRS can be an option to overcome those limitations [35] . In S4 and/or S5 , local health authorities can promote or subsidise the use of insecticide-treated materials ( e . g . , insecticide house screening and treated curtains ) that have been proven effective ( S2 Table ) and , in emergency situations ( S5 ) , promote the use of topical repellents that provide protection against mosquito bites ( Table 2 , Fig 1 ) . Finally , epidemiological and entomological evidence exist with regards to the mass deployment of gravid oviposition traps to reduce Aedes mosquito density ( S2 Table ) that can be a low-cost , community-based , and sustainable participation complementary strategy ( Table 2 ) [36] . Furthermore , there is currently considerable innovation in vector control for prevention of Aedes-borne viral disease [37] . Novel approaches—such as the sterile insect technique , Wolbachia , genetically modified mosquitoes , removal trapping , and spatial repellents—gather relevant entomological data , and many are engaged in well-designed field trials that will generate the epidemiologic data necessary to develop public health policy for their deployment . Methods of larval control and their strengths , limitations , and evidence base are described in Table 2 . The aim of targeting mosquitoes at immature life stages ( i . e . , larvae and pupae ) is to reduce adult Aedes emergence and to reduce adult population densities . Control may be intensified during the early mosquito season but necessary year-round in tropical regions and requires high coverage because there may be sufficient temporary larval habitats to maintain high mosquito adult densities and virus transmission . Larval control ( i . e . , environmental management , source reduction , larviciding , or biological control [community-based and/or top-down approaches] ) is more effective when it is consistent and routine ( S4 ) rather than in a periodic emergency response ( S5 ) . Larval control needs to be sustained in order to reduce the size of the adult vector population and to keep the population density below certain , currently still undefined thresholds to minimise the risk of virus transmission . Source reduction has been , and continues to be , a key component of dengue , Zika , and chikungunya control programmes [9 , 38] . It should primarily target artificial containers in private and public spaces , although some natural containers , such as bamboo and bromeliads , can also harbour Aedes larvae . Larvicides are generally long-lasting and moderately costly . Unfortunately , the evidence supporting larviciding as part of control programmes is mixed , with some studies showing a beneficial effect of pyryproxifen as part of a community-based strategy reducing dengue rates and entomological indices [39] , whereas others such as Temephos or Bacillus thuriengensis ( Bti ) do not have strong evidence in the review of evidence ( S2 Table and Table 1 ) . Biological control methods ( fish , copepods , and others ) are relatively acceptable and can be used for treatment of large and permanent breeding sites , but existing evidence is inadequate for assessing the impact of this strategy for dengue control ( S2 Table ) . Community-based source reduction ( as clean-up campaigns or the use of water container covers ) can reduce Aedes vector populations and is supported by epidemiological cluster randomised controlled trials results of lower infection with dengue virus in children and fewer reports of dengue illness from a trial in Mexico and Nicaragua ( Table 2 and S2 Table ) [40 , 41] . Human behaviour is the common denominator of all Aedes-borne virus epidemic risk scenarios and therefore of prevention and control strategies . Social mobilisation is a key factor in the success of Aedes mosquito control strategies and in preventing outbreaks . There is evidence ( S2 Table ) that community participation is effective in reducing larval indices and disease prevalence [40 , 41 , 42 , 43] . Community participation and education ( e . g . , door-to-door visits , workshops , and webinars ) can inform the population on how to reduce Aedes populations by emptying or eliminating nonpermanent water containers and covering permanent water storage containers with untreated or insecticide-treated covers . Other actions can be carried out by health education programmes , such as distribution of printed materials , educational meetings , involvement of local opinion leaders , sensitisation at schools , and the use of mass media ( radio , television , newspapers , leaflets , posters ) [42] . Health education efforts should be carried out routinely and intensified before peak periods of virus transmission ( S4 and S5 ) . Sex education is also important for Zika prevention because the virus can be transmitted sexually . WHO recommends the use of communication for behavioural impact ( COMBI ) , an approach that integrates behavioural and social communication to reduce risk and prevent disease . COMBI is used in an increasing number of countries for dengue control [44] , and a toolkit has been developed to deliver more effective outbreak response measures [45] . In practice , education and communication strategies are often implemented too late , that is , after the outbreak has begun to decline ( S4 and S5 ) . Social communication is more likely to be successful when information is disseminated early , which means before the introduction of vectors or virus ( S1 ) , when transmission has recently been established ( S2 ) , or before transmission has peaked ( S1 , S2 , S3 and S4 ) . Activities to promote behavioural change should produce measurable and visible results and should be monitored using appropriate indicators [44] . It is important to note that social mobilisation is not a ‘stand-alone’ strategy and that community-based control campaigns are carried out in combination with other vector control interventions [29 , 41 , 42 , 43] . Aedes control cannot be successful without effective and sustained intra- and intersectoral collaboration [4 , 13 , 46] . Within the health sector , Aedes control should not be the responsibility of a single department . Interagency collaboration is fundamental for a successful programme . The vector control unit should , therefore , establish strong links with other vector-borne disease programmes ( e . g . , malaria vector control ) , epidemiological surveillance , clinical diagnosis and management , vaccine delivery ( when appropriate ) , maternal and child health ( e . g . , integrated management of childhood illnesses ) , health education , veterinary surveillance , and environmental health [46] . Intersectoral actions for vector control ( Table 3 ) should be guided by site-specific knowledge of larval aquatic habitats , locations where risk of infection is highest ( i . e . , inside homes for endophilic and endophagic A . aegypti or outside homes for exophilic and exophagic A . albopictus ) , and current and historical hot-spots of reported illness . The specific actions that can be taken in collaboration with other sectoral actors will depend on the setting and feasibility . Outside the public health sector , collaborations should be forged with , for example , ministries of education , environment , water , and urban planning and housing [5] , and with the private sector , nongovernmental organisations ( NGOs ) , and town councils ( Table 3 ) . For example , provision of reliable piped water should be encouraged to prevent storage of water in containers in and around the home , which can harbour Aedes . Solid waste management should be improved to remove rubbish from the peridomestic environment , which can accumulate water and provide habitats for Aedes to lay their eggs . A pilot study of recycling of plastic materials in Merida , Mexico , was able to reduce entomological indices by incentivised recycling . Bonus points were given for large volumes of reusable materials in exchange for commodities and targeted at the most at-risk neighbourhoods [47] . The scheme was organised by the local government through the Health Services and in coordination with the Ministries of Social Development , Urban and Environmental Development , and Education . The Ministry of Housing and Infrastructure can develop and enforce housing and building codes , for example , to mandate installation of screened doors and windows on properties and rainwater runoff control for new housing developments as well as prohibit construction of open groundwater wells . The Ministries of Education and Health can work together to disseminate behaviour change communication on prevention of Aedes population and disease proliferation ( see the ‘Vector control’ section ) . Information on prevention of Aedes-borne diseases should be integrated into school curricula for long-term sustainability . NGOs , including community groups , such as neighbourhood women’s , religious , environmental , and social action organisations , should be engaged . Local community groups can be involved in promoting and implementing environmental management as well as delivering behaviour change communication . NGOs can be influential in mobilising communities and encouraging acceptance of routine and outbreak vector control methods ( S5 ) . The private sector can be engaged . For example , commercial companies can support recycling , e . g . , disposal and recycling of discarded tyres . In Brazil , a partnership between the Ministry of the Environment and the National Association of the Tyre Industry encourages consumers to return used tyres to collection points at which point they are used as alternative fuel or recycled into flooring and other products [13] . Private health facilities can be incorporated into the epidemiological surveillance system . Academic and research institutions can cooperate with the Ministry of Health to train personnel and carry out surveillance through the sharing of facilities , i . e . , entomology laboratories , insectaries , and human resources . Development projects and commercial agriculture can undertake assessments of the health impact of Aedes-borne diseases and implement mitigation strategies . NGOs , UN agencies , and bilateral or multilateral donors can be engaged to implement control measures to prevent virus transmission in conflict zones . All these activities should be coordinated through an interministerial steering committee with broad representation that seeks regular input from nonministerial stakeholders , such as NGOs , research and educational establishments , community organisations , and the private sector [13] . The committee should have clearly defined terms of reference and meet regularly , not just during outbreaks . Working in an integrated fashion has the potential to increase efficiency and public health impact more than narrowly focused , uncoordinated actions from the health sector alone . Additional important complementary activities for achieving effective vector control and Aedes-borne disease prevention include capacity building , advocacy , policies and laws , and research and innovation ( Fig 1 ) . Supporting activities are briefly summarised in S3 Table .
During the past 50 years , the world has seen the emergence and dramatic spread of Aedes-transmitted arboviral diseases . Social , environmental , and demographic changes have facilitated the proliferation of existing transmission systems and the spread of viruses and vectors into new ecological settings [4 , 46] . Notable deficiencies in the planning and implementation of vector control programmes were reported and include the following: The call for a global response and preparedness for vector borne diseases ( i . e . , GVCR ) is , therefore , timely [11] , and implementing and sustaining integrated surveillance and locally adapted Aedes control measures should be a priority [50] . Unfortunately , there have been very few well-conducted epidemiological field trials of Aedes vector control ( S2 Table ) , which means that prioritisation of control strategies is difficult . Limited epidemiological evidence supports the deployment of community-based source reduction , ISS , TIRS , and house screening and larviciding , but evidence is lacking for most other vector control tools ( S2 Table , Table 2 ) . Promising new interventions targeting adults—such as Wolbachia for population replacement and/or population suppression , genetically modified mosquitoes , sterile insect techniques , community-based mass trap deployment , spatial repellents , and attractive targeted sugar baits—are currently being considered for dengue prevention , but their public health efficiency in field trials ( i . e . , effectiveness trials ) has yet to be determined , and large-scale roll-out ( effectiveness studies ) of these programmes will take years [37 , 51] . There are no magic bullets for Aedes-borne diseases control , including vaccines . The desire to find easy or rapid solutions has been tried for decades without success and is likely to continue to lead to disappointment . The most practical and productive path forward is to intensify surveillance and to strengthen the evidence base so that , when needed , ‘a box’ of effective tools can be deployed in an integrated manner that takes into account the local situation and available resources . The IAM provides a technical guidance on how and when to implement integrated management for Aedes , tailored to location-specific entomological and epidemiological risk scenarios based on current evidence . To be effective and sustainable , IAM must be fed with robust entomological and epidemiological data collection at different time and spatial scales ( country , district , city , and neighbourhood ) and in particular for local scenarios in endemic and epidemic settings with randomised controlled trials of combinations of several tools , to guide-decision making for Aedes-borne disease control . The IAM needs be supported by human and financial resources and must be carefully monitored and evaluated . Increased funding is crucial given the growing threat to public health and the need for evidence that innovation makes disease prevention more effective . We hope that the measures outlined here will help promote and implement the WHO GVCR and will be used to guide actions that improve Aedes-borne disease prevention and outbreak response . The ultimate goal is to use available resources as effectively and efficiently as possible to safeguard global health . | Aedes aegypti and A . albopictus are mosquito species that thrive in towns and cities and can transmit viruses to humans that cause diseases , such as dengue , Zika , chikungunya , and yellow fever . The geographic range of human infection with these viruses is rapidly expanding globally . Even when preventative or therapeutic treatments are available to fight these diseases , controlling the mosquito vector will remain an important control option . We therefore developed a framework called IAM that offers decision-making guidance based on available evidence of effective control of Aedes at different levels of infestation and virus transmission risk . Our work aims to strengthen the capacity of countries at risk of and/or affected by these diseases and vectors so they will be better prepared for existing and emerging Aedes-borne disease threats . | [
"Abstract",
"Introduction",
"Conclusion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"animals",
"review",
"infectious",
"disease",
"control",
"insect",
"vectors",
"zoology",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"agrochemicals",
"epidemiology",
"disease",
"vectors",
... | 2018 | Integrated Aedes management for the control of Aedes-borne diseases |
RNA secondary structure plays a central role in the replication and metabolism of all RNA viruses , including retroviruses like HIV-1 . However , structures with known function represent only a fraction of the secondary structure reported for HIV-1NL4-3 . One tool to assess the importance of RNA structures is to examine their conservation over evolutionary time . To this end , we used SHAPE to model the secondary structure of a second primate lentiviral genome , SIVmac239 , which shares only 50% sequence identity at the nucleotide level with HIV-1NL4-3 . Only about half of the paired nucleotides are paired in both genomic RNAs and , across the genome , just 71 base pairs form with the same pairing partner in both genomes . On average the RNA secondary structure is thus evolving at a much faster rate than the sequence . Structure at the Gag-Pro-Pol frameshift site is maintained but in a significantly altered form , while the impact of selection for maintaining a protein binding interaction can be seen in the conservation of pairing partners in the small RRE stems where Rev binds . Structures that are conserved between SIVmac239 and HIV-1NL4-3 also occur at the 5′ polyadenylation sequence , in the plus strand primer sites , PPT and cPPT , and in the stem-loop structure that includes the first splice acceptor site . The two genomes are adenosine-rich and cytidine-poor . The structured regions are enriched in guanosines , while unpaired regions are enriched in adenosines , and functionaly important structures have stronger base pairing than nonconserved structures . We conclude that much of the secondary structure is the result of fortuitous pairing in a metastable state that reforms during sequence evolution . However , secondary structure elements with important function are stabilized by higher guanosine content that allows regions of structure to persist as sequence evolution proceeds , and , within the confines of selective pressure , allows structures to evolve .
RNA secondary structures play fundamental roles in the replication of all positive-strand RNA viruses . Because of their small genomes ( which are largely devoted to encoding viral proteins ) , these viruses use available sequence space highly efficiently . The genomic RNA of viruses forms structures necessary for multiple replicative functions . For example , internal ribosome entry site elements interact with the cellular translation initiation machinery , diverse structural signals direct packaging of viral RNA into viral particles , and RNA structure can provide control signals for differential viral gene expression . The human immunodeficiency virus type 1 ( HIV-1 ) is no exception , and well-characterized RNA structures within the coding domains of the genome play critical roles in regulation of replication . These include a structure in the env gene , the Rev Response Element ( RRE ) , that binds the viral protein Rev leading to the transport of unspliced and singly-spliced viral mRNA out of the nucleus [1] , [2] , and a hairpin structure preceded by a poly ( U ) slippery sequence that mediates a frameshift during synthesis of the Gag-Pro-Pol polyprotein [3] , [4] . The untranslated regions ( UTRs ) of HIV-1 and simian immunodeficiency virus ( SIV ) contain the TAR hairpin , which recruits the Tat protein to modulate transcription [5] , [6] ( reviewed in [7] ) and other stem-loop structures that are important for dimer initiation ( DIS ) [8] , splicing [9]–[11] , and viral RNA packaging [12]–[14] ( reviewed in [15] ) . Several lines of evidence emphasize that the HIV-1 genome contains extensive additional RNA secondary structures whose functional roles are not yet fully understood [16]–[18] . The structures of large RNAs , like viral RNA genomes , are too complex to be predicted with confidence from first principles or thermodynamic-based algorithms alone . Useful working models can often be obtained when additional information is available to restrain the number of possible secondary structure elements . Two such approaches are to compare evolutionarily related sequences to identify RNA motifs that co-vary to preserve base pairs , and to experimentally probe the RNA structure with chemicals or nucleases to infer the presence of paired versus unpaired regions . In the selective 2′-hydroxyl acylation analyzed by primer extension ( SHAPE ) chemical probing approach , nucleotide reactivities show a strong inverse correlation with the probability that a nucleotide is base-paired . SHAPE-informed prediction of RNA folding has been used to develop secondary structure models for diverse RNAs [19]–[23] including the full-length genomic RNA structure of HIV-1NL4-3 [17] . This HIV-1NL4-3 model shows a very strong correlation between regions that can be targeted by siRNAs to inhibit viral replication [18] and regions that are predicted to be single-stranded , suggesting that many global structural features are likely correct . One approach to evaluating the broader significance of these structures is to examine the conservation of these structures in a related virus . To this end , we analyzed the secondary structure of the genomic RNA of a second primate lentivirus , SIVmac239 , a representative of the SIVsm/HIV-2 lineage of primate lentiviruses . HIV-2 evolved from a different primate reservoir than HIV-1 . HIV-2 arose from an SIV that infects the sooty mangabey ( Cercocebus atys ) , and SIVsm has also infected rhesus macaques in primate centers to cause an AIDS-like illness . SIVmac239 [24] now serves as a prototype reference sequence for comparative analysis of the HIV-2/SIVsm lineage [25] . SIVmac239 has a larger evolutionary distance from HIV-1NL4-3 than the more similar SIVcpz or other HIV-1 clades , and conservation of structures between SIVmac239 and HIV-1NL4-3 represents an especially stringent test for functional relevance . In this analysis , we describe areas where RNA structure is maintained between SIVmac239 and HIV-1NL4-3 , where it is divergent , and outline possible mechanisms for understanding the interplay between rapid sequence evolution and maintenance of function of RNA structural motifs .
To develop an experimentally based secondary structure model for the genomic RNA structure of SIVmac239 ( GenBank accession M33262 ) , we used a strategy similar to that used to develop a model for the secondary structure of genomic HIV-1NL4-3 RNA [17] . Viral RNA was purified from SIVmac239 particles and derivatized with the SHAPE reagent 1-methyl-7-nitroisatoic anhydride ( 1M7 ) under physiologically relevant conditions to discriminate between single-stranded ( generally reactive ) positions versus nucleotides constrained by base pairing or other interactions ( and therefore unreactive ) [21] , [26]–[28] . The derivatized positions were identified as terminations to DNA synthesis by reverse transcriptase ( primers listed in Table S1 ) . SHAPE reactivities were measured for 9 , 605 nucleotides , 99 . 6% of the genome . These data were used as pseudo-free energy change constraints to constrain RNA secondary structure prediction . In the secondary structure model for the SIVmac239 RNA genome ( Figure 1 and Figure S1 ) , 4 , 970 nucleotides were predicted to be base-paired ( 51 . 5% ) , whereas 4 , 676 nucleotides were modeled as single-stranded ( 48 . 5% ) . Highly structured regions in an RNA can be inferred in a model-free way by identifying regions with low overall median SHAPE reactivities . Many areas of the SIVmac239 RNA genome have low median SHAPE reactivity ( defined as less than 0 . 4 on a scale from 0 to ∼1 . 5 ) over a 75 nucleotide window , and these correspond to regions of structure with both known and unknown function . To isolate regions of low reactivity to a more select group , we assigned numerical values to the areas that have median SHAPE reactivities below 0 . 3 ( labeled in Figures 1 and 2 ) . The lowest median SHAPE reactivity values occurred at the 5′ and 3′ ends of the genome . The highly structured 5′ region extends until nucleotide 539 ( Figures 1A and 2B , motif 1 ) , and the structured 3′ region begins at position 9462 at the start of the 3′ TAR structure within the terminal repeat ( R ) regions ( Figures 1B and 2B , motif 15 ) . In addition , the Gag-Pro-Pol frameshift ( G-P-P FS ) element and the Rev Response Element ( RRE ) are highly structured ( Figures 1B and 2B ) . Similarly , when we used RNA Decoder ( a program that predicts evolutionarily conserved RNA secondary structure in the context of the protein-coding sequence of an RNA [29] ) with an HIV-2/SIVsm sequence dataset to infer conserved regions of secondary structure , we also found that the 5′ and 3′ UTRs and the RRE showed the strongest signal for conservation of structure ( Figure S2 ) . Using the RNA Decoder approach we conclude that some features of secondary structure are slightly conserved within the coding region , but none at the level of the RRE . Other regions have low median SHAPE reactivities , yet currently unknown RNA functions ( Figures 1 and 2B , motifs 2–14 ) . The secondary structure profile presented here is likely not the only structure that is possible for the SIVmac239 RNA to form . Instead , it is the lowest free energy structure predicted , given the SHAPE reactivity constraints . The potential for alternative or multiple structures can be partially inferred in the reactive nucleotides that are predicted to be paired and the unreactive nucleotides that are predicted to be single-stranded ( Figure 1 ) . Even so , over a 75 nucleotide window , the fraction of nucleotides whose pairing prediction is in concordance with the SHAPE reactivity ( assuming nucleotides with SHAPE reactivity values below 0 . 4 are likely to be paired or those with reactivity values equal to or above 0 . 4 are preferentially unpaired ) averaged 74% across the entire SIVmac239 genome ( Figure 2C ) , with certain regions of the genome ( including the 5′-UTR , the Gag-Pro-Pol frameshift , the RRE , the first splice acceptor , and the polypurine tracts ) having an even higher concordance . It is possible that some regions fold to unique structure while others exist in multiple structures , with the SHAPE-informed structure being one of them and the reactivity representing the average of two or more states . In the following sections , we examine structure in regions with low SHAPE reactivity and high pairing concordance to infer biological importance based on their conservation with HIV-1NL4-3 , first on a global scale and then in detail . SIVmac239 is the second full-length genomic RNA of a primate lentivirus to be evaluated by SHAPE-directed modeling; the first was that of HIV-1NL4-3 [17] . Comparison of the structural models of these two distantly related retroviral RNA genomes should reveal conserved structural elements . Visually , the patterns of 1M7 reactivity in the 5′ noncoding region , the Gag-Pro-Pol frameshift site , and the RRE — all regions with well-established conserved functions — are similar for SIVmac239 and HIV-1NL4-3 RNAs ( Figure 2B ) . A bootstrapping analysis ( see Methods ) showed that the measured SHAPE profiles across both genomes were significantly more similar than expected by chance ( 10 , 000 trials , p<0 . 0001 ) . Thus , in a broad view , there appears to be a propensity to conserve the overall level of local RNA structure across similar regions for these two genomes . The RNA folding algorithm employed for structure prediction includes a pseudo-free energy change term to account for the SHAPE reactivity ( Text S1 ) . Newly optimized parameters for calculating the pseudo-free energy term [30] were used to predict the current SIVmac239 structural model and to revise the secondary structure model for HIV-1NL4-3 based on the original reactivity data [17] ( Figure S3 ) . These new parameters resulted in changes in the organization of some structures , especially those with higher average SHAPE reactivity , but the locations of highly structured regions did not change significantly . We compared the codon-aligned sequences of SIVmac239 and HIV-1NL4-3 for equivalency in terms of base pairing; these two genomes share only 50% identity at the nucleotide level . We found that roughly half of the nucleotides predicted to be base-paired in the HIV-1NL4-3 sequence were also paired in the SIVmac239 sequence; conversely , only half of the nucleotides predicted to be single-stranded in the HIV-1NL4-3 sequence were also single-stranded in the SIVmac239 sequence ( Table 1 ) . In spite of the limited conservation of paired bases , we did observe numerous regions in similar locations in the two genomes with low SHAPE reactivity ( defined as median reactivity below 0 . 4 over a 75 nucleotide window ) . These areas ( Figure 2B , gray dashes ) largely fold into structures of unknown function . None of these structures of unknown function have conserved base pairs , and in only a few examples are there even small hairpins that share pairing partners within 40 nucleotides in both alignments . As a more stringent definition of structural conservation , we tallied the number of base pairs in both genomes where both nucleotide positions of the base-pairing partners were maintained , using the protein codon sequence to establish the alignment . Only 58 codon-aligned base pairs were fully conserved between the two genomes within the 8 , 738 nucleotides of the SIVmac239 coding region; an additional 13 base pairs were conserved in the 5′-UTRs . Overall , only 71 base pairs , 3% of the total base pairs , were precisely conserved in these two primate lentiviral genomes ( Figure 1 , blue bars in paired regions ) . Overall , regions that share a low SHAPE reactivity are areas of base pairing in both viruses , but the exact structures are not conserved between the two genomes . We aligned the 5′-UTRs of each virus both using the structures as predicted by SHAPE-constrained modeling and by identifying the positions of the functionally conserved TAR region , 5′ polyadenylation [poly ( A ) ] signal , primer binding site ( PBS ) , major splice donor ( SD1 ) sequence , dimerization initiation sequence ( DIS ) , and the Psi packaging element ( Figure 3 ) . Each of these functional elements folds into a similar structure in both viral RNA genomes even though only 13 of the ∼180 predicted base pairs have the same two pairing partners in both 5′-UTR regions ( Figure 3 , emphasized with heavy blue lines to indicate conserved pairing partners ) . Additional conserved structures include a stem immediately 5′ of the PBS , and the SL1 ( DIS ) , SL2 ( SD1 ) , and SL3 helices . The stem containing the gag start site ( Figure 3 , MA start ) , termed U5-AUG and previously identified by phylogenetic analysis and structure probing [21] , [31] , [32] , accounts for six of the identical pairing partners between SIVmac239 and HIV-1NL4-3 . This interaction has also recently been visualized by NMR analysis of the dimerized and packaged form of HIV-1 RNA [33] . There are structural differences in the TAR motif , which features three stem-loops in SIVmac239 but only a single stem in HIV-1 , as previously described [34]–[36] . Notably , the stems in SIVmac239 5′-UTR are generally longer with more base pairs than the equivalent structures in HIV-1NL4-3 ( Figure 3 ) . This has also been shown for HIV-2 leader RNA in an analysis of the first 560 nucleotides [37] . Because of the sequence redundancies at each end of the retroviral genome , a poly ( A ) signal ( AAUAAA ) occurs at each end . The virus prevents use of the signal at the 5′ end to avoid producing truncated RNA transcripts . One way this is accomplished is through a stable hairpin directly 3′ of TAR , which contains the 5′ poly ( A ) signal [38] . This hairpin structure is similar in both genomes , with unreactive loop nucleotides in the same area in SIVmac239 as in HIV-1NL4-3 . In the 5′ poly ( A ) region of HIV-1 , in vitro analysis suggests formation of a pseudoknot [39] . The SIVmac239 SHAPE reactivity is also low in the region in gag that corresponds , based on the codon alignment , to one of the pseudoknot stems in HIV-1NL4-3 . It is likely that a pseudoknot forms in this region of the SIVmac239 RNA as well ( Figure 3 ) . Nucleotide sequence conservation of the poly ( A ) stem-loop was previously noted for HIV-1 , SIV , and HIV-2 sequences [39]; structural similarity in the MA region , based on SHAPE analysis , provides an independent line of evidence supporting this pseudoknot structure ( Figure 3 ) . We directly tested the long-range pseudoknot interaction between the stem loop that forms near the 5′ poly ( A ) signal and the complementary sequence at the beginning of the MA coding domain using a locked nucleic acid ( LNA ) oligonucleotide to anneal to and sequester pairing partners in the folded SIVmac239 RNA . An LNA was designed to bind to the 3′ side ( in the MA coding region ) of the predicted pairing interaction and annealed to the RNA . We then probed the surrounding area ( nts 1–882 ) for changes in SHAPE reactivity , excluding nucleotides bound by the LNA . The putative pairing partner GCUGCC in the poly ( A ) stem-loop showed a clear overall increase in reactivity after seclusion of the downstream pairing partner ( Figure 4 ) . A few additional nucleotides showed slight reactivity changes ( data not shown ) , which likely reflect RNA structural shifts due to the disruption of the pairing interaction . Furthermore , an analysis performed using an RNA spanning nucleotides 1–560 of the 5′ region of HIV-2 , which has about 88% nucleotide identity to SIVmac239 over these sequences , provides SHAPE data of the poly ( A ) hairpin when its potential pseudoknot pairing partner within the MA-coding region has been deleted [37] . The HIV-2 data for the truncated RNA show significantly higher reactivity over the 5′ half of the predicted pseudoknot ( GCUGCC ) as compared to the full-length SIVmac239 RNA analyzed here with the downstream pairing partner intact . Together , these results support formation of a pseudoknot in the loop region of the 5′-UTR poly ( A ) stem , similar to that observed in the same area in the HIV-1 genomic RNA [39] . In sum , although only 13 base pairs are formally identical , the 5′-UTR structure is highly conserved between the SIVmac239 and HIV-1NL4-3 RNA genomes both at the level of overall structure ( Figure 2B ) and in terms of the local architecture of multiple functional elements ( Figure 3 ) . Overall , this analysis emphasizes the striking capacity of RNA to form conserved functional structures despite a very low level of absolutely conserved base pairing . The primate lentiviruses generate more of the Gag polyprotein than the Gag-Pro-Pol polyprotein by controlling expression using minus-one frameshifting to join translation in the Gag reading frame to translation in the Pro-Pol reading frame . Frameshifting occurs at a “slippery sequence” , a poly ( U ) region , and is facilitated by a downstream structure that stalls the ribosome [3] , [4] . The RNA structure in the region of the frameshift site is similar in SIVmac239 and HIV-1NL4-3 . In both cases , the poly ( U ) slippery sequence is part of a stem , although the poly ( U ) region is not paired in the SIVmac239 structure . In addition , there is a second stem just downstream of the U stretch; the frameshift stem is further from the poly ( U ) sequence in SIVmac239 than it is in HIV-1NL4-3 ( Figure 5A ) . Despite the fact that this region carries out a conserved and essential function in retrovirus replication , the organization of these stems is different in the two genomes and they have no shared base pairs ( Figure 5A ) . We attempted to define the pathway through which these structures evolved by examining the sequences surrounding the poly ( U ) slippery sequence . To facilitate this analysis , we included a sequence from SIVagm ( GenBank accession M30931 ) , which is distantly related to both SIVmac239 and HIV-1NL4-3 . All three sequences aligned well at the protein and nucleotide levels , both upstream and through the poly ( U ) slippery sequence ( Figure 5B , gray boxes ) . However , the alignment is lost three nucleotides 3′ of the poly ( U ) sequence . The sequence becomes re-aligned at the conserved protease processing site ( KPRNFP ) , lost , then aligns again at the conserved PTAPP motif in the Gag p6 coding domain ( Figure 5B , 3′-most gray box ) . One explanation for the abrupt loss of sequence alignment in this region is that the frameshift hairpin is , itself , mutagenic , consistent with the idea that structure in the RNA would induce pausing of reverse transcriptase and thereby enhance recombination and mutation during viral DNA synthesis [40] , [41] . Both frameshift hairpins are predicted to be among the most stable in their respective genomes . The frameshift hairpin ranks in the top 5% in its calculated stability among all hairpins in the SIVmac239 genome; the relevance of this calculation is supported by the low reactivities of paired residues in this hairpin relative to other hairpins ( data not shown ) . A structure stable enough to stall the ribosome may also induce pausing of reverse transcriptase during DNA synthesis , ultimately promoting template strand exchange ( recombination ) [42] , which can be mutagenic when pairing at the 3′ end of the cDNA is imprecise . We thus hypothesize that the rapid evolution of structure in the frameshift region is due to the mutagenic effect of the structure itself . The RRE includes binding sites that mediate oligomerization of the Rev protein; oligomerized Rev mediates export of unspliced and singly-spliced viral RNA from the nucleus [43] . The sequence is conserved in many primate lentiviral genomes [44] . The predicted RRE structure ( Figure 1 ) consists of a long , irregular stem I helix terminated by a set of small hairpins including the IIb stem , previously described as the primary Rev binding site [45] , [46] , and the auxiliary hairpins ( stems III , IV , and V ) that facilitate Rev multimerization [47] . The RRE structures for HIV-1 and SIV have been previously evaluated [48] , here we analyze the SHAPE-constrained structures from both genomes as they relate to one another . Twenty-nine of the 71 base pairs conserved between SIVmac239 and HIV-1NL4-3 occur in the small terminal hairpins in the RRE ( Figure 6A , blue bars ) ; these nucleotides are 78% conserved at the sequence level . By contrast , the long stem is mostly devoid of conserved pairing partners , and there is a shift by one nucleotide in the codon-aligned pairing ( Figure 6A ) . In an effort to propose a mechanism by which this shifted pairing in stem I may have occurred , we mapped the variant nucleotides in stem I of SIVmac239 onto the structure of HIV-1NL4-3 . There are 18 nucleotide differences total ( 36% of the 50 nucleotides that compose the SIVmac239 stem I ) , eight nucleotides in HIV-1 would break a base pair that exists in SIVmac239 , and six nucleotides in SIVmac239 would do so for HIV-1NL4-3 . Seven nucleotide differences in the stem I region account for the five amino acid differences in this region , and of these , four have no effect on pairing partners in the stem ( Figure 6A ) . We suspect , therefore , that the shift in the pairing register is a product of multiple synonymous nucleotide changes that occurred over time and not due to one or two specific nucleotide mutations . We postulate that the viruses have accomplished structural divergence without introducing a deleterious intermediate through the placement of bulges within stem I . Furthermore , when we used the codon alignment ( Figure 6B ) to force superposition of the equivalent SIVmac239 pairing partners onto the RRE structure of HIV-1NL4-3 , there was a significant reduction ( 30% ) in the number of base pairs formed in stem I ( Figure 6A , right structure ) . Thus , conservation of specific base pairs is limited to four of the small hairpins that serve as protein interaction regions , whereas the long stem is comprised of pairs that are shifted by one nucleotide in the two RNA genomes ( Figure 6 ) . We infer that neither the sequence nor the exact base-pairing partners of the long stem I are critical , although its presence and long length are conserved . In contrast , the strong conservation of pairing partners in the small , Rev-binding stems reflects the specificity of the Rev-RNA interaction as the source of the selective pressure that conserves both nucleotide identity and pairing partners . Retroviruses use diverse splice donor ( SD ) and splice acceptor ( SA ) sites to generate numerous spliced mRNAs which direct synthesis of the small regulatory proteins and the Env protein while retaining some unspliced RNA for both translation of Gag and Gag-Pro-Pol and for packaging of the full-length genome into new virions [10] . Splicing to generate these mRNAs is highly regulated . This regulation takes place at both the sequence level and at the RNA structure level . Five of the base pairs that are precisely conserved between SIVmac239 and HIV-1NL4-3 are in the stem of the hairpin structure that contains the first splice acceptor site ( SA1 ) ( Figure 7 ) , which generates the transcript that codes for the viral Vif protein [49] . We have termed this conserved structure SLSA1 ( stem-loop at splice acceptor 1 ) . Most of the other splice acceptor regions of SIVmac239 ( SA2–SA8 ) downstream of SA1 are part of short hairpins as well , with the exception of SA4 ( Figure 1 ) . Each of these short hairpins has low median SHAPE reactivity ( most are below 0 . 4 ) ; however , only SLSA1 is precisely conserved between SIVmac239 and HIV-1NL4-3 with several of the same pairing partners . The viral splice acceptors have weak splicing sequences to allow balanced usage of each with the major splice donor SD1 [50] . The relative strengths of HIV-1 splice acceptor sequences have been previously analyzed , and splicing is most efficient to SA1 of all the splice acceptor sites [51] . We hypothesize that the purpose of the conserved SLSA1 pairing is to regulate splicing to this site and to allow balanced use of this site and the other downstream splice acceptor sites . We tested the importance of the SLSA1 structure by making mutations that disrupt pairing interactions in this motif and then monitored the effect of these mutations on viral mRNA splicing patterns in HIV-1NL4-3 . Mutations to SLSA1 incorporated the following constraints: ( i ) to avoid disrupting described splicing enhancer elements within the region downstream of the SA1 site [51] , [52] , ( ii ) to maintain the gag-pro-pol coding sequence , and ( iii ) to use alternative codons that also occur nearby in the sequence as not to require rare or non-viral codon usage ( Figure 8B ) . Although limited in mutation choices , the resulting mutant virus ( HIV-1SLSA1m ) has two single nucleotide substitutions that disrupt the pairing interactions at SLSA1 . We infected CEMx174 cell cultures for three days and examined the splicing profile of each viral mRNA pool . The spliced RNA sequences were converted to cDNAs using primers that were placed either after the SD4 or SA7 elements to monitor either the incompletely spliced 4 kb class or the completely spliced 1 . 8 kb class of viral mRNAs , respectively ( Figure 8A ) . These same primers were then used in PCR that included a unique Nar I site at the 5′ end , upstream of SD1 . Each pool of products was cleaved with Nar I and end-labeled , imparting single 5′ labels on each amplicon , indepenent of RNA length . These products were resolved by acrylamide gel electrophoresis and their identity inferred by comparison with a previous assignment [10] , although not all products were observed ( Figure 8C ) . This analysis did not resolve the product for the singly spliced Vif mRNA ( D1A1 ) ; however , we did not see a significant difference in the ratio of products of D1/A1 and D1/A2 for Vpr mRNA , suggesting that the mutations did not have a major effect on the use of A1 . The mutations did have a significant effect in suppressing the use of SA3 , an acceptor that is used to generate a series of Tat mRNAs in the 1 . 8 kb class ( and their incompletely spliced equivalents in the 4 kb class ) ( Figure 8C ) . This effect was specific to SA3 as the downstream splice acceptors for Rev mRNA ( SA4a-c ) were not affected . Given the complexity of the splicing patterns it will be important to improve the confidence of these assignments , for example using a deep sequencing approach [53] . However , disruption of the conserved SLSA1 RNA structure shifts the balance of spliced mRNA products , although by a mechanism that emphasizes the complexity of splicing regulation and a potential role for long-range RNA structural interactions in this regulation . Retroviruses prime plus-strand DNA synthesis from polypurine tracts ( PPT ) that are derived from viral RNA during minus-strand DNA synthesis [54] . These primers are resistant to degradation by RNase H , and their specificity as second-strand primers is enhanced by the viral NC protein [55] . Primate lentiviruses prime from two regions , one near the center of the genome ( cPPT ) and one just upstream of the U3 sequence near the 3′ end of the genome ( PPT ) [56] . We observed a common structural motif in these polypurine tracts in SIVmac239 and in HIV-1NL4-3 . The cPPT and PPT motifs in both viruses contain a 5′ A-rich single-stranded region followed by a 3′ G-rich base-paired region ( Figure 9A ) that have strikingly similar patterns of SHAPE reactivity ( Figure 9B ) . Since the PPT and cPPT function as second-strand primers while hybridized with the first/minus strand of viral DNA , it is unlikely that RNA secondary structure per se modulates the function of these sequences as plus strand primers . These patterns drew our attention to the possibility that guanosine and adenosine play very different roles in defining secondary structure more globally across the genome , and that these roles might be reflected in the structures of the PPTs as a byproduct of their high G content in the context of a purine-rich run . This caused us to consider the role of base composition in defining secondary structure more broadly across the genome . The base compositions of both the SIVmac239 and HIV-1NL4-3 RNA genomes are dominated by adenosine ( 34% ) ; the percentage of cytidine is low , around 17% , and the percentages of guanosine and uridine are each about 25% ( Figure 9C ) . Regions with large numbers of base pairs must have approximately the same number of pyrimidines and purines . In structured regions with known function , including the 5′-UTR and the RRE in both SIVmac239 and HIV-1NL4-3 , the average base composition is 25% A , 29% G , 22% C , and 22% U ( Figure 9D ) . The higher percentage of guanosines compared to adenosines in these base-paired regions has the effect of concentrating adenosines in unpaired regions , where adenosines represent fully half of the nucleotides . Only about 30% of adenosines are base-paired , whereas approximately 60–70% of guanosines , cytidines , and uridines are base-paired in the two lentivirus secondary structure models ( Figure 9C ) . This trend toward favoring unpaired adenosines but paired guanosines , cytidines , and uridines is also observed in highly structured regions of RNAs , including bacterial ribosomal RNAs [57] . Given the A-rich primate lentiviral genome , the overall effect is to create A-rich single-stranded regions . We considered the possibility that the greater stability of the G-C base pair relative to A-U or G-U base pairs might define structures that are conserved between the two genomes . We compared the base compositions of structures within SIVmac239 with known function ( 5′-UTR , Gag-Pro-Pol frameshift stem , and the RRE ) to regions of extensive structure but unknown function . The conserved structures with known function have a higher average guanosine content and a significantly higher ( p = 0 . 04 ) percentage of G-C pairs ( 57 . 9% ) than structured regions of the genome that are not conserved ( 49 . 5% ) ( Table S2 ) . Moreover , the SHAPE reactivity was higher for both adenosine and guanosine residues in regions of non-conserved structures ( data not shown ) , suggesting that adenosine-rich structures may facilitate RNA unfolding and refolding or may allow for multiple conformations as compared to guanosine-rich structures in functional regions . We also considered the possibility that clustering of guanosine nucleotides may reflect restricted codon usage for functionally conserved amino acids in the translation product . This model , however , is not supported given that the fraction of codons whose first two positions contain an A and those that contain a G is similar across the SIVmac239 genome and not statistically different within the RRE or Gag-Pro-Pol frameshift . The variation in G or A abundance thus reflects changes that occur at wobble positions and is independent of the amino acid sequence of the translation product . We therefore hypothesize that selection to maintain functional secondary structures in primate lentiviruses such as SIVmac239 and HIV-1NL4-3 has resulted in regions defined by clusters of guanosines within an otherwise adenosine-rich genome . In an initial attempt to use clusters of guanosines to reveal regions that may be under selection to maintain conserved structures , we scanned a set of aligned primate lentiviral genomes using a sliding window to record a “G minus A” score ( Figure 10 ) . As noted above , high guanosine content is easily seen in the 5′ and 3′ UTRs , the Gag-Pro-Pol frameshift site , and the RRE across seven selected genomes ( Figure 10 , vertical gray bars ) . In addition , the guanosine cluster at SA1 is readily apparent in these comparisons and is preceded by a dip in guanosine abundance near the cPPT . For the SIVmac239 RNA , we did not find strong evidence for conservation of structure at the boundaries of regions of the RNA that encode protein folding domains ( data not shown ) , a feature reported for the HIV-1NL4-3 genome [17] . Consistent with this observation , there is no conserved pattern of guanosine clustering at these boundaries , perhaps with the exception of the region encoding the signal peptide ( SP ) of Env ( Figure 10 , vertical yellow bars ) . We suggest that the central region of primate lentiviral genomes , including the region around the cPPT and SA1 , and perhaps the region near the 5′ end of the env coding domain , represent new functional structures that are conserved in location if not in detail within the primate lentivirus lineage . Ultimately , HIV-1 and SIVsm/HIV-2 genomic RNAs accomplish many of the same functions in the context of viral replication . There is abundant evidence that RNA structure is either critical for or directly modulates diverse functions including viral DNA synthesis , RNA splicing , genome packaging , and mediation of interactions with both viral and cellular proteins . We sought to identify functionally important RNA secondary structures in the primate lentiviral genome by comparison of SHAPE-directed nucleotide-resolution structure probing information and by developing structural models of representative HIV-1 and SIVsm genomes . We developed a secondary structure model for SIVmac239 and compared it to a modestly revised structural model for the HIV-1NL4-3 genome [17] . One paradigm for assessment of conserved function is the ribosomal RNAs where strong base-pairing patterns are highly conserved despite large sequence variations over the course of evolution . The HIV-1NL4-3 and SIVmac239 genomes share about 50% sequence identity and feature a similar fraction of base-pairing ( 60% ) . However , in contrast to ribosomal RNA-like behavior , only about 3% of predicted base pairs were with identical partners within the coding region of each genome . Almost one-half of these identical pairs were clustered in the Rev binding domain of the RRE . Thus , there has been massive reorganization of the patterns of RNA secondary structure between these two genomes suggesting a selection environment that is very different from that experienced by ribosomal RNA . Even within regions of highly conserved function , there were large differences in sequence and pairing partners . Dramatic differences in the structure of TAR , longer stems in the 5′-UTR of SIVmac239 relative to that of HIV-1NL4-3 , different pairing partners and poor sequence alignment in the Gag-Pro-Pol frameshift stem , and a one-base shift in the alignment in the large RRE stem all point to large-scale remodeling of these domains . Stable secondary structures can promote recombination during retroviral DNA synthesis [40] , [41] and certain stable structural elements appear to be mutagenic . In regions like the Gag-Pro-Pol frameshift stem , selective pressure does not maintain a particular set of base pairs , but rather ensures that a sufficiently stable structure exists . In contrast , regions involved in RNA-protein interactions , such as the Rev oligomerization domain , displayed a higher level of conservation than other regions of the genome , indicating that selective pressure maintains a particular structure for interaction with protein . In regions with conserved secondary structures , we observed significantly higher guanosine content compared to the overall base composition of the genome . Higher guanosine levels may serve to stabilize functionally critical structures . Lentiviral genomes are adenosine-rich , and the resulting less stable secondary structures may exist in alternative states , even if they are drawn as a single representative structure in our models . We propose that scanning for guanosine-rich regions in these and other adenosine-rich retroviral genomes may facilitate identification of important structural domains . One source of the selective pressure to maintain an adenosine-rich genome is the action of APOBEC3-G and -F , enzymes that deaminate cytidines on the DNA minus strand during viral DNA synthesis giving rise to G-to-A transitions on the plus strand [58] . Although these lentiviral genomes are adenosine-rich , they are not guanosine-poor ( approximately 25% G content ) but rather cytidine-poor ( at 17% C content ) . Thus mechanisms must be in place to retain guanosines in regions of functional RNA secondary structures . Focusing on the clustering of guanosines across the primate lentiviral lineage has allowed us to suggest several small areas in the genome that should be explored as possible structural elements that contribute function in the viral replication cycle . Our analysis shows that within the relatively short evolutionary distance between the HIV-2/SIVsm lineage and HIV-1 , conserved secondary structures at the individual base pair level occur at the ends of the RNA and in the RRE . This observation is consistent with the interpretation that , for most of the lentiviral genome , there is little selective pressure to maintain specific pairing interactions . This contrasts the evolutionary pressure on ribosomal RNA . The sequences of 16S rRNA are less than 50% conserved when eubacterial , archaebacterial , and eukaryotic RNAs are compared , but their structures have been maintained through evolution by mutations that compensate for sequence changes that directly affect base pairing ( reviewed in [59] ) . In strong contrast , the lentiviral RRE structure , particularly in stem I , did not evolve through base changes that maintained pairing . We previously examined a model where a higher rate of transition versus transversion mutations exists in paired regions of many RNA structures ( presumably to maintain pairing partners ) , but found that the HIV-1 RRE was an exception as its mutation pattern did not fit this model [60] . The relatively low conservation of base pairs between SIVmac239 and HIV-1NL4-3 is consistent with this observation since very few pairing partners are maintained . Clearly the constraints on rRNA evolution , with its many bound proteins , are distinct from the dramatically less rigid constraints on these viral RNA genomes . We propose that the lentiviral genomic structure is evolving in the context of two significant mutagens . First , APOBEC3-G and -F indirectly mutate guanosines to adenosines , which weakens stability of structural motifs . Second , the structural motifs themselves are mutagenic during DNA synthesis . The effect of these mutagens is filtered by selective pressure to maintain useful structural motifs . The majority of these genomes is thus depleted of both guanosines and strong secondary structure and , in this way , has evolved to be less susceptible to these mutagens . In contrast , the RNA regions that form essential structures are enriched in guanosine residues that provide stability , which also has the effect of ensuring some structure is retained in the face of these mutagens which impart selective pressure . Since the submission of our manuscript , the following publications have appeared . Van der Kuyl et al . [61] show that retroviruses are strongly adenosine-rich and cytidine-poor , in agreement with our conclusions . Van Hemert et al . [62] show that the abundant adenosines seem to accumulate in single-stranded regions and not in base-paired regions of the HIV-1NL4-3 genome , in agreement with our conclusions .
An infectious clone of SIVmac239 ( GenBank accession M33262 ) was a gift from Dr . Ronald Desrosiers ( New England Regional Primate Center , Harvard Medical School ) [63] . SIVmac239 was used to infect SupT1 CCR5 CL . 30 cells ( a gift from Dr . James Hoxie , University of Pennsylvania ) ; these cells are a non-Hodgkin's T cell lymphoma cell line ( a modified version of the SupT1 cell line ) [64] . The virus produced was purified as described [65] . Viral genomic RNA was gently extracted from purified SIVmac239 viral particles as described [17] in a manner that avoided denaturation of RNA secondary or tertiary structure . Viral RNA was extracted using phenol/chloroform after lysis and treatment with Proteinase K . No heating steps , chelating agents , or chemical denaturants were used in the purification . The final RNA product was precipitated in 70% ( v/v ) ethanol with 300 mM NaCl and stored at −80°C until use . RNA was treated as described [17] . Briefly , the precipitated RNA was collected by centrifugation and the ethanol removed . Each pellet , containing 62 pmol of SIVmac239 genomic RNA , was individually resuspended in 620 µl of 50 mM HEPES ( pH 8 . 0 ) , 200 mM potassium acetate ( pH 8 . 0 ) , 3 mM MgCl2 and incubated at 22°C for 10 min then at 37°C for 15 min . Aliquots of 32 µl of 45 mM 1M7 in dimethyl sulfoxide ( DMSO ) [27] or DMSO alone were warmed at 37°C for 30 sec , then 288 µl of the RNA solution was added to each and incubated at 37°C for 5 min . RNA was recovered by adding 32 µl of 50 mM EDTA ( pH 8 . 0 ) and precipitated with ethanol . An oligonucleotide consisting of DNA and locked nucleic acid ( LNA ) with the sequence 5′-TTGCTGCCCA-3′ was designed to be complementary to the sequence to be tested for a pseudoknot interaction at the 5′ poly ( A ) loop . The LNA was added in 5-fold excess to the target RNA with 50 mM HEPES ( pH 8 . 0 ) , 200 mM potassium acetate ( pH 8 . 0 ) , 3 mM MgCl2 . RNA modification with 1M7 was then performed as described above . Each primer contained a 5′ six carbon linker terminated with an amino group ( IDT ) ; a total of 38 primers were used ( Table S1 ) . The primers were tethered to 5-FAM or 6-JOE fluorophores ( AnaSpec ) using N-hydroxysuccinimide chemistry . Purified primers were spectrophotometrically determined to have at least 82% labeling efficiency , with most labeled to greater than 95% , as determined by the [dye]/[DNA] ratio . Both the ( + ) and ( − ) 1M7 reagent reactions were subjected to reverse transcription with FAM-labeled primers using SuperScript III Reverse Transcriptase ( Invitrogen ) . A sequencing length ladder was generated using the JOE-labeled primers and termination with a dideoxynucleotide . After cDNA synthesis , the reverse transcription reaction products were combined with their corresponding JOE-labeled sequencing reactions , the latter performed using plasmids containing SIVmac239 sequences , p239SpSp5′ , and p239SpE3′ ( obtained through the AIDS Research and Reference Reagent Program , Division of AIDS , NIAID , NIH from Dr . Ronald Desrosiers [63] . Primer extension products were resolved by length using an Applied Biosystems AB3130 capillary electrophoresis instrument . ShapeFinder software was used to convert the raw capillary electrophoresis electropherograms of fluorescence intensity to normalized SHAPE reactivities [21] , [22] , [28] . Data were processed as described [17] . The SIVmac239 sequence ( 9646 nucleotides ) with the addition of a poly ( A ) tail consisting of 10 adenosines was folded using the RNAstructure algorithm [66] , [67] . SHAPE reactivities were incorporated into the thermodynamic folding algorithm to constrain secondary structure . Due to computational restrictions in the folding algorithm ( caused by the length of the genomes ) , folding was accomplished in large overlapping pieces consisting of at least two-thirds of the entire genome . The structure of the whole genome was generated by combining the separately folded pieces over regions with identical structures . Multiple analyses with varying lengths gave consistent structures . Although we are unable to identify pseudoknots de novo with the current algorithm , the model includes the pseudoknot at the 5′ poly ( A ) stem that we predict based on low SHAPE reactivity of nucleotides in loop regions and by sequence alignment with HIV-1NL4-3 . Recently updated folding parameters m = 1 . 9 and b = −0 . 7 [30] were used to generate a new version of the HIV-1NL4-3 RNA structure ( Supporting Material ) , which was used for these analyses . The viral plasmid pNL4-3 was acquired from the National Institutes of Health ADS Research and Reference Reagent Program . For site-directed mutagenesis of pNL4-3 , fragments digested with PflMI and AgeI ( New England Biolabs ) were inserted into vector pT7Blue ( Novagen ) . Mutagenesis primers 5′-GAGATCCAGTATGGAAAGGTCCAGCAAAGCTCCTC-3′ and 5′-GCTTTGCTGGACCTTTCCATACTGGATCTCTGCTG-3′ were used in accordance with the previously described mutagenesis protocol [68] . The resulting plasmid and pNL4-3 were then digested with PflMI and AgeI ( New England Biolabs ) and ligated with T4 DNA Ligase ( New England Biolabs ) to create the plasmid pSLSA1m , which was sequenced to confirm the presence of the given mutation . A total of 2 µg of pSLSA1m or pNL4-3 and was used to produce mutant and wild-type infectious virus ( HIV-1wt and HIV-1SLSA1m , respectively ) by transfection into 3×105 293T cells in a volume of 2 ml DMEM following the FuGENE ( Promega ) protocol . After 48 hrs , supernatant from the cells was centrifuged , transferred to 1 ml aliquots , then stored at −80°C . One aliquot per virus was used in a viral infectivity assay [69] to determine infectious units per ml of supernatant . To obtain viral mRNA , 1×106 cells were infected with 100 µl virus ( either HIV-1wt or HIV-1SLSA1m ) supernatant in a volume of 0 . 5 ml for 2 hrs at 37°C before being brought to a final volume of 10 ml and incubated at 37°C for three days . Cells were centrifuged and supernatant was removed . The cell pellet was homogenized through a QiaShredder column ( Qiagen ) and total mRNA was purified by the RNeasy Mini Prep Plus kit ( Qiagen ) according to the manufacturer's protocol . Using viral mRNA isolated from three flasks of cells infected with HIV-1wt and three flasks of cells infected with HIV-1SLSA1m , we digested each sample with RQ1 DNase ( Promega ) for 2 hrs at 37°C and purified them again using the same RNeasy Mini Prep kit ( Qiagen ) . We then performed One-Step RT-PCR ( Qiagen ) following the manufacturer's protocol and using primers 5′-AGTCAGTGTGGAAAATCTCTAGCAGTGG-3′ and either 5′-CCGCAGATCGTCCCAGATAAG-3′ ( 1 . 8 kb class ) or 5′-CTATGATTACTATGGACCACAC-3′ ( 4 kb class ) in a volume of 25 µl . Each was then digested at a unique restriction site with NarI ( New England Biolabs ) for 2 hr at 37°C and labeled with 0 . 78 µCi 32P-αdCTP ( Perkin-Elmer ) using Klenow fragment ( New England Biolabs ) . Each was then purified through a PCR Purification column ( Qiagen ) and eluted with 30 µl elution buffer ( Qiagen ) . A sample of 10 µl of each was run on a 6% polyacrylamide gel . Controls included mock-infected cultures and cDNA amplification reactions without reverse transcriptase . SIVmac239 and HIV-1NL4-3 sequences were aligned at the codon level using the Los Alamos lentivirus compendium ( www . hiv . lanl . gov ) . Protein start and end positions , known RNA structures , and known protein functional regions were taken into consideration as well as conserved amino acids . Deletions and insertions were incorporated into the sequence alignment where appropriate . A Matlab 7 . 8 ( R2009a ) script was used to compare the actual average absolute difference in SHAPE reactivity value across all aligned SIVmac239 and HIV-1NL4-3 genome positions . We randomized the position assignments for reactivity values in the SIVmac239 genome to make a distribution of average absolute differences , then repeated this 100 , 000 times to generate a random distribution curve and plotted the observed average on this curve . We employed a two-tailed Fishers exact test to compare the GC content to AU and GU content of structures with known and unknown function in SIVmac239 . We used a two-tailed Fishers exact test to compare codon composition in the Gag-Pro-Pol frameshift and RRE regions to the rest of the coding region of SIVmac239 . The number of adenosines or guanosines in non-wobble positions were counted and compared in both the regions of known function and the rest of the genome . The secondary structures of the RNA models were organized using xrna ( http://rna . ucsc . edu/rnacenter/xrna ) . Structure predictions using RNA-Decoder [29] were performed as previously described [17] with the following modifications . The input alignment was a reduced version of the HIV-2 web alignment available from the Los Alamos lentivirus compendium ( www . hiv . lanl . gov ) . Codon positions in overlapping regions were designated according to the reading frame of the first member of the following pairs: gag-pro , pol-vif , vif-vpx , vpr-tat1 , tat1-rev1 , env-tat2 , env-rev2 , env-nef . The alignment was scanned using separate phylogenetic trees for the upstream and downstream sections , which were generated by Tree-Puzzle [70] using the GTR+γ ( 4 ) model , 10 , 000 puzzling steps , “accurate” parameter estimation , and other default settings . The tree for the first half of the genome was built on the third codon positions of the gag , pro , pol , and vif genes and the 5′ non-coding region , and the downstream tree was inferred from the third positions of the vpx , vpr , tat1 , env , and nef genes and the 3′ non-coding region . Trees are available from the authors on request . The following GenBank accession numbers refer to the given sequences mentioned in the text: SIVmac239 , M33262; HIV-1NL4-3 , AF324493; SIVL'hoest , AF075269; SIVsyk173 , L06042; SIVagm , M30931; HIV2BEN , M30502; HIV-1 MVP , L20571 . | We have taken advantage of the rapid evolution of primate lentiviruses to assess the conservation of secondary structure in the viral RNA genome . We determined the structure of the SIVmac239 RNA genome to allow a detailed comparison with the previously determined structure of the HIV-1NL4-3 genome . In comparing the two structures , we find very few conserved base pairs with the same pairing partners , indicating that RNA structure is evolving even faster than the sequence . This suggests that most of the genome is in a metastable state that refolds during sequence evolution . Specific areas of structure that are required for function are maintained by the clustering of guanosines in the otherwise adenosine-rich genome , although the precise organization of the structure evolves . The strong effect of selection on maintainence of protein recognition sites can be seen in the conservation of pairing partners within the Rev binding sites in the RRE RNA . We propose that the more stable elements of RNA structure that are needed for function are susceptible to mutation during viral DNA synthesis . This causes the structures to evolve rapidly , yet still within the constricts of selective pressure , allowing maintenance of function . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"biochemistry",
"rna",
"viral",
"evolution",
"rna",
"structure",
"infectious",
"diseases",
"nucleic",
"acids",
"virology",
"hiv",
"biology",
"microbiology",
"viral",
"diseases"
] | 2013 | Comparison of SIV and HIV-1 Genomic RNA Structures Reveals Impact of Sequence Evolution on Conserved and Non-Conserved Structural Motifs |
Empirical evidence suggests the incentive value of an option is affected by other options available during choice and by options presented in the past . These contextual effects are hard to reconcile with classical theories and have inspired accounts where contextual influences play a crucial role . However , each account only addresses one or the other of the empirical findings and a unifying perspective has been elusive . Here , we offer a unifying theory of context effects on incentive value attribution and choice based on normative Bayesian principles . This formulation assumes that incentive value corresponds to a precision-weighted prediction error , where predictions are based upon expectations about reward . We show that this scheme explains a wide range of contextual effects , such as those elicited by other options available during choice ( or within-choice context effects ) . These include both conditions in which choice requires an integration of multiple attributes and conditions where a multi-attribute integration is not necessary . Moreover , the same scheme explains context effects elicited by options presented in the past or between-choice context effects . Our formulation encompasses a wide range of contextual influences ( comprising both within- and between-choice effects ) by calling on Bayesian principles , without invoking ad-hoc assumptions . This helps clarify the contextual nature of incentive value and choice behaviour and may offer insights into psychopathologies characterized by dysfunctional decision-making , such as addiction and pathological gambling .
Standard theories of decision-making assume that the incentive value of an option should be independent of options presented in the past and options available during choice [1–4] . These theories are fundamentally challenged by empirical evidence showing that expectations ( derived from past experience ) about upcoming options change value attribution and choice behaviour [5–14] . For example , in a series of recent experiments from our lab [8–10] , participants made choices in blocks ( i . e . contexts ) associated with one of two distinct , but partially overlapping , reward distributions . Participants’ choices were consistent with attributing a larger incentive value to rewards ( common to both contexts ) in blocks associated with low compared to high average reward . In other words , the incentive value of a reward increased when the average was lower . In addition to the average reward of a context , evidence from a similar task indicated that reward variance within a given context also exerts an influence on incentive value [11] . These findings highlight contextual effects exerted by expectations about options ( induced , for example , by options available during previous choices ) ; namely , between-choice contextual effects . In addition , the empirical literature has highlighted contextual influences elicited by options available during choice; namely , within-choice context effects [6 , 15–20] . Standard theories of decision-making assume that the incentive value of an option should be independent of other options available during choice [1–4] . This implies that the choice proportion between two options , comprising a more valuable and a less valuable option , should be unaffected by the introduction of a third [2] . However , a recent study [6] has shown that this choice proportion follows a U-shape function , which diminishes as the value of a third option approaches the value of the target options–and starts increasing thereafter ( Fig 1A ) . This is hard to reconcile with standard theories and represents a form of within-choice context effect , whereby the value of an option is affected by other options available during choice . In this task , it is unnecessary to compare options across different attributes ( single-attribute decisions; [6] ) . However , other forms of within-choice context effect have been observed when options are defined by the same set of attributes that have to be traded of against each other ( multiattribute decisions; [15–20] . For example , consider a binary choice between a high-quality and expensive car A versus a low-quality and cheap car B ( Fig 1B ) . Imagine the values of the attributes are such that an agent is indifferent about the two options ( i . e . , the higher price of car A is exactly compensated by its quality ) , resulting in the same probability of choosing options A and B . What happens if a third option is also available ? Standard models ( based on the assumption that values are independent of other options ) predict that the choice probability difference will remain zero , independent of a third option . However , empirical data highlight a so-called similarity effect [20–23] , whereby preference for an option over a second option–which is equally preferable during binary decisions–increases if a third option is available that is similar to the second option ( Fig 1B ) . In our example , the choice probability difference between car A and car B will be positive when a third low-quality and cheap ( similar to car B ) car C is also available . A form of influence called the attraction effect [15 , 24–26] has also been found with the availability of a third option that is characterized by a low score for one attribute and an intermediate score for the other ( Fig 1B ) . The presence of such a third option favours the option with a high score for the attribute for which the third option has an intermediate score . In our example , the choice probability difference between car A and car B will be positive when a third medium-quality and expensive car D is also available . Finally , empirical data are consistent with a so-called compromise effect [17 , 25 , 27] . This applies when the choice set includes two options scoring high in one attribute and low in another plus a third option characterized by intermediate scores for both attributes . While the three options are equally preferred ( i . e . , are chosen an equal amount of times ) if presented in pairs during binary choices , when they are all available , a preference for the option characterized by intermediate scores is observed ( Fig 1B ) . For instance , although during binary choices an average-price and average-quality car E is not preferred over car A or over car B , car E will be favoured when presented together with both car A and car B . Several explanations have been proposed to account for contextual effects on incentive value and choice , with most models focusing on within-choice context effects during multiattribute decisions [16–20 , 27 , 28] . Other theories have been proposed to explain between-choice context effects [29–31] , and disregard within-choice effects . We are aware of a single attempt to encompass both between-choice and within-choice effects , though restricted to non-multiattribute decisions for the latter type of effects [6] . Whether models developed to explain a certain class of context effects generalise to other effects remains unclear–and a unifying account encompassing all known context effects is lacking . Developing a parsimonious account would represent an important theoretical advance , as it would explain diverse empirical phenomena with the same underlying principles . The goal of the present paper is to describe a unifying theory , referred to as Bayesian model of context sensitive value ( BCV ) that explains between-context and within-context effects , in single and multiattribute decisions . This theory represents a generalization of a recent model developed to explain between-choice contextual effects [11] . The key idea is that agents build a generative model of reward within a context and , every time a new reward or option is presented , use Bayesian inference to invert this model to form a posterior belief about the underlying reward distribution . Incentive value is computed during this belief update and corresponds to a precision-weighted reward prediction error . The advantage of this theory relies on its grounding upon simple normative principles of Bayesian statistics . In addition , the model can explain between-choice context effects [8–11] and makes specific predictions that have been confirmed empirically . In brief , BCV proposes that the incentive value of a stimulus ( or option ) corresponds to the change in reward expected ( in any given context ) when the stimulus is presented . This makes precise predictions about choice under ideal ( Bayesian ) observer assumptions ( with a minimal number of free parameters ) . Crucially , predictions include specific forms of context effects , and raise a question of whether these predicted effects are consistent with empirical findings . In this paper , we applied BCV to multi-alternative choice ( considering both single and multiattribute decisions ) and ask whether the model predicts the context-effects found empirically . We first present a theoretical extension of BCV applicable to decisions in which multiple options are available and can be characterized by multiple attributes . We show that predictions derived from the model are remarkably similar to empirical findings on within-choice contextual effects , both during non-multiattribute and multiattribute decisions . We next review BCV in relation to between-choice context effects and describe how the model can also explain these empirical findings . On this basis , we offer the model as a principled description of between and within-choice context effects .
The idea behind BCV is to establish a link between theories of value and normative accounts of brain functioning based on Bayesian statistics [32–37] . The Bayesian brain framework rests on the idea that an agent builds a model of the processes generating sensory cues . The generative model comprises a set of random variables ( i . e . , hidden states or causes of sensory outcomes ) and their causal links ( i . e . , probabilistic contingencies ) . The variables can be separated into hidden and observable variables , the former representing the latent causes of observations , and the latter representing sensory evidence or cues . Sensory evidence conveyed by observable variables is combined with prior beliefs about hidden causes to produce a posterior belief about the causes of observations . The application of this logic has proved effective in explaining several empirical phenomena in perception [32–37] . For instance , psychophysical data indicate that human perception depends on integrating different perceptual modalities ( e . g . , visual and tactile ) in a manner consistent with Bayesian principles [38] , where evidence is weighted by the precision of sensory information . Furthermore , process theories that mediate Bayesian inference ( e . g . , predictive coding ) have a large explanatory scope in terms of neuroanatomy and physiology [39] . Inspired by a recent framework that conceptualises planning and choice as active inference [40–45] , our core proposal is that Bayesian inference drives the attribution of incentive value to reward , and this in turn determines choice . In a previous work , we have developed a version of BCV applicable to conditions where past options elicit context effects by shaping expectancies before a reward is presented ( [11]; see below ) . However , our previous formulation did not consider conditions where multiple options ( potentially characterized by multiple attributes ) are available . Here we generalize BCV to encompass conditions in which multiple options are available and options can be characterized by multiple attributes . We define a multi-attribute option un ( e . g . , car A or car B ) as a contract that yields reward amount Ri , n relative to each attribute i ( e . g . , price or quality ) : un={Ri , n}i=1 , … , InwithRi , n∈R ( 1 ) An option set u is the set of options currently available: Setu={un}n=1 , … , Nwithunanoption∀n ( 2 ) The expected value ( EV ) of an option un corresponds to: Run=∑iRi , n ( 3 ) For example , the total reward for car A is equal to the reward associated with price plus the reward associated with quality . BCV assumes that an agent builds a generative model of the reward amounts Ri , n ( Fig 2 ) . Specifically , an agent believes that , for each attribute i , reward amounts Ri , n across options are sampled from the same population . To distinguish among attributes , we assume that an agent believes that an independent population of reward amounts is associated to each attribute . For example , if two attributes characterize options , two independent populations of reward amounts are considered by the agent ( Fig 2 ) . Formally , for each attribute i , the average of the population of the reward amounts Ri , n is represented by a random variable Ci , which is assumed to be sampled from a Gaussian distribution with prior mean μCi and uncertainty ( variance ) σCi2: Ci∼N ( μCi , σCi2 ) ( 4 ) The agent assumes that μCi and σCi2 are known but that Ci is not directly observable and therefore needs to be inferred from observing the different instances of reward amounts Ri , n of options for the attribute i . This is realized in the generative model by treating Ci as a hidden cause of Gaussian variables Ri , n with mean Ci and uncertainty σRi2: Ri , n∼N ( Ci , σRi2 ) ( 5 ) On the basis of the generative model , for each attribute i , the agent can estimate C^i=N ( μ^Ci|Ri , σ^Ci|Ri2 ) , namely the posterior belief about the variable Ci ( i . e . , the average reward amount relative to the attribute i; the hat symbol indicates estimates of unknown quantities ) , given the observation of all reward amounts of all options available for the attribute i , represented by the set Ri . In other words , an agent assumes that there is an average reward for each attribute which is unknown but can be estimated based on the reward amounts . According to Bayes’ rule , the posterior belief of Ci can be calculated by considering the associated Ri , n sequentially in any order . We propose such sequential belief updating for BCV , even if options ( and the associated reward amounts ) are presented simultaneously , and we assume that the order of options considered is random ( with potentially different orders for different attributes ) . For example , when three options characterized by two attributes are available ( represented by R1 , 1 , R1 , 2 and R1 , 3 for attribute one , and R2 , 1 , R2 , 2 and R2 , 3 for attribute two ) , inference can involve computing , in order , P ( C1|R1 , 1 ) , P ( C1|R1 , 1 , R1 , 3 ) and P ( C1|R1 , 1 , R1 , 2 R1 , 3 ) for attribute one , and P ( C2|R2 , 3 ) , P ( C2|R2 , 1 , R2 , 3 ) and P ( C2|R2 , 1 , R2 , 2 R2 , 3 ) for attribute two . In the example above , an agent may consider first car A and next car B when estimating the average reward for price , and first car B and next car A when estimating the average reward for quality . The rationale behind sequential belief updating is that the brain is equipped with a limited computational capacity , which precludes the instantaneous ( and parallel ) evidence accumulation , and hence requires the processing of one option after another . A similar evidence accumulation process is implicit in some theories of perceptual and value-based decision-making ( e . g . , [16 , 46 , 47] ) . Below , we will show that this evidence accumulation , in the form of sequential Bayesian belief updating , endows agents with the right sort of sensitivity to context . Formally , if Ri , n is the reward amount considered first during belief updating , in relation to attribute i , the posterior mean μ^Ci|Ri , n is [48]: μ^Ci|Ri , n=μCi+σCi2σCi2+σRi2 ( Ri , n−μCi ) ( 6 ) The posterior uncertainty σ^Ci|Ri , n2 is: σ^Ci|Ri , n2=σCi2−σCi2σCi2+σRi2σCi2 ( 7 ) The crucial proposal we advance is that the incentive value Vi ( Ri , n ) –attributed to a reward amount Ri , n in relation to the attribute i and associated with option un–is central to belief updating ( see Eq 6 ) and corresponds to a precision-weighted prediction error [49]; namely , to the difference between Ri , n and the prior mean μCi , multiplied by a gain term which depends on the uncertainty of that attribute σRi2 and the prior uncertainty σCi2: Vi ( Ri , n ) =σCi2σCi2+σRi2 ( Ri , n−μCi ) ⇒μ^Ci|Ri , n=μCi+Vi ( Ri , n ) ( 8 ) Within BCV , incentive value imbues reward and associated options with behavioural relevance , by favouring either approach to ( for positive incentive values ) or avoidance of ( for negative incentive values ) these reward amounts and options . This implies two fundamental forms of contextual normalization . First , a subtractive normalization is exerted when μCi is different from zero . For example , if we assign positive and negative numbers to rewards ( i . e . , Ri , n > 0 ) and punishments ( i . e . , Ri , n < 0 ) respectively , their corresponding incentive values will change in sign , depending on whether punishment ( i . e . , μCi < 0 ) or reward ( i . e . , μCi > 0 ) is expected a priori . Small rewards may appear as losses in contexts where large rewards are expected . Second , a divisive normalization depends on considering the gain term σCi2σCi2+σRi2 . This implies that the positive and negative value of profits ( i . e . , Ri , n > μCi ) and losses ( i . e . , Ri , n < μCi ) are magnified by a large gain term , when we have precise beliefs about the average reward of the population . Sequential Bayesian belief updating means that inference proceeds by considering one reward amount at a time . If Ri , n is considered at step t+1 and Ri , t is a set containing all reward amounts already seen up until step t for attribute i , then a posterior mean μCi|Ri , t , Ri , n is obtained at step t+1 equivalent to ( Bishop , 2006 ) : μ^Ci|Ri , t , Ri , n=μ^Ci|Ri , t+σ^Ci|Ri , t2σ^Ci|Ri , t2+σRi2 ( Ri , n−μ^Ci|Ri , t ) ( 9 ) Implying a value for the reward amount Ri , n: Vi ( Ri , n ) =σ^Ci|Ri , t2σ^Ci|Ri , t2+σRi2 ( Ri , n−μ^Ci|Ri , t ) ( 10 ) For each attribute i , incentive values are accumulated in memory until inference is completed ( i . e . , all reward amounts have been considered ) . We can assume that inference proceeds in sequence or in parallel across attributes; however , this has no impact on incentive values , as the agent believes that attributes are associated with independent reward populations ( formally: P ( C1 , C2 , … , CI|R ) = P ( C1|R1 ) , P ( C2|R2 ) , … , P ( CI|RI ) ) . When all attributes for an option un have been considered , we assume that the incentive value of the option corresponds to the sum of the incentive values of associated reward amounts: V ( un ) =∑i=1Vi ( Ri , n ) ( 11 ) Inference proceeds until , for all attributes i , the posterior expectation about rewards μ^Ci|Ri is evaluated and , at this point , a choice is realized following a softmax rule based on the incentive values of the available options [2] . In summary , BCV is based on the following assumptions: Below these assumptions are discussed in detail . Assumptions ( i ) , ( iii ) , and ( iv ) are implicit in adopting a Bayesian scheme . Assumption ( vii ) is based on a standard approach in which incentive values are summed and a softmax choice rule is adopted . Assumption ( ii ) captures the notion of multiple attributes , in other words it enables an agent to link rewards to their attributes . Assumption ( vi ) ( sequential belief updating or evidence accumulation ) reflects the real world constraint that people have to evaluate available options and rewards one by one . In other words , agents cannot magically and instantaneously assimilate all the options on offer–they have to accumulate evidence for the underlying payoffs by evaluating each in turn . This notion plays a central role since we will see that context effects emerge because a reward is contextualized by previous rewards encountered during inference . This underlies assumption ( v ) that associates incentive value with a precision-weighted prediction error–a central construct in Bayesian inference . Heuristically , this scheme implies that an option is more likely to be selected if it increases expectations of reward , and will be avoided if it decreases expectations . In other words , an option is more likely to be selected if it suggests the situation is better than indicated by options considered previously during belief updating . Note that a Bayesian perspective may suggest that incentive value corresponds to a posterior belief–rather than a precision-weighted prediction error . As an example , this would imply that the value of the same dish will be perceived as ‘higher’ in a ‘better’ restaurant . However , empirical data are consistent with the opposite notion that ( adopting the same example ) the value of the same dish is perceived as lower in a better restaurant [6–14] . This evidence motivated our proposal that incentive value corresponds to a precision-weighted prediction error , and not to a posterior belief . In sum , BCV provides a principled explanation for how Bayesian inference , assigning a key role to prior expectation and uncertainty , might underlie value computation and choice . The key role of uncertainty is reflected in the precision-weighting of prediction errors . The hypothesis we entertain here is that the mechanisms postulated by BCV may be general and explain multiple forms of context effects . We have previously applied BCV to explain between-choice context effects; namely , those elicited by options presented in the past [11] . Here , we explore the possibility of applying the same model to within-choice effects , which arise when multiple options are available . In what follows , we will consider single and multiattribute choices under this Bayesian formalism . Here , we apply BCV to explain within-choice contextual influences during non-multiattribute decisions . These comprise choices in which trading-off different attribute is not required , as for instance when options are defined by a single attribute . Consider first how different prior expectations μC ( i . e . , the prior expectation over the average reward of the attribute ) and reward uncertainty σR2 affect the choice between two options characterized by a single attribute ( Fig 3 ) . We can examine the predicted proportion of choosing a better option ( associated with high reward RH ) compared to a worse option ( associated with low reward RL ) , as a function of prior expectation μC and reward uncertainty σR2 . Classical theories predict a flat function because they do not model an influence of the prior mean μC and reward uncertainty σR2 [1–4] . In contrast , BCV predicts bell-shape functions over prior expectations , that peak at the prior mean of μC = ( RH–RL ) /2 ( Fig 3 ) . In this setting , the reward uncertainty σR2 determines the width of the function ( larger uncertainties produce narrower functions ) . Below , we analyse conditions where more than two options are available–and within-choice context effects come into play . Classical decision-making models predict that , during choice , the choice ratio between two options should not be affected by the reward associated with a third option [1–4] . However , a recent study has challenged this hypothesis , highlighting within-choice context effects [6] . Adopting a choice task in which three options were available during choice , this study showed that the choice proportion between a more valuable and a less valuable target option diminished as a third option value increased towards the value of the target options ( Fig 1A ) . After this point , the choice proportion started increasing ( Fig 1A ) . Here , we examine the implications of applying BCV in this scenario . Fig 4 illustrates the predictions of BCV of the ratio of choices of the two target options ( a better target option RH and a worse target option RL ) as a function of the reward of a third option R3 and as a function of the agent’s prior belief about the average option reward μC and about the reward uncertainty σR2 . This figure shows that all these variables exert an influence . First , for certain values of reward uncertainty σR2 and prior mean μC , the reward of a third option R3 influences the choice proportion between the two target options according to a U-shape function , in a way that is consistent with empirical findings ( Fig 1A ) . Second , the impact exerted by the reward of a third option R3 decreases as the reward uncertainty σR2 increases . In other words , within-choice context effects emerge only with small reward uncertainty σR2 . This can be explained by the fact that a small σR2 magnifies reward prediction error ( RPE ) , enabling contextual effects to emerge . Third , when the reward uncertainty σR2 is sufficiently small , the prior mean μC comes into play . Overall , a larger prior mean μC increases the choice proportion between the two target options ( independently of the reward of a third option R3 ) . Furthermore , the prior mean μC exerts a modulatory influence on the effect of the reward of a third option R3 , as the effect exerted by R3 is enhanced with a larger prior mean μC . Note that context effects exerted by R3 are obtained with μC = 0 , which can be considered a default value for this parameter . Collectively , these simulations provide proof of principle that BCV can explain within-choice contextual effects in single-attribute decisions that are remarkably similar to those seen in empirical studies [6] . In what follows , we now extend the explanatory scope of BCV to multiattribute problems . Empirical studies of multi-attribute decisions have highlighted three forms of effects , including the similarity [20–23] , attraction [15 , 24–27] , and compromise effect [17 , 25 , 27] . Here , we apply BCV to multi-attribute decisions and ask whether the predictions that emerge from the model reproduce the context effects found empirically . To this aim , we consider two options ( e . g . , the two cars A and B described above ) defined by two attributes ( e . g . , price p and quality q ) . Considering the reward amounts of car A , we assign Rp , A = 1 to price ( low scores indicate high price ) and Rq , A = 10 to quality . Conversely , when considering the reward amounts of car B , we assign Rp , B = 10 to price and Rq , B = 1 to quality . We now consider the choice probability difference between option A and option B as a function of the reward amounts Rp , K and Rq , K of a third option K . Empirical evidence is hard to reconcile with standard models of choice , which predict that the choice probability difference between option A and option B should not depend on the value of a third option K . Fig 5A summarises the empirical findings by plotting the probability of choosing A minus the probability of choosing B as a function of the attributes of a third option K . This graph shows conditions in which the choice probability difference is bigger or smaller than zero , illustrating both a similarity and an attraction effect . Specifically , a similarity effect favours option A when option K is good in price and bad in quality ( top-left of the graph ) , and favours option B when option K is bad in price and good in quality ( bottom-right of the graph ) . An attraction effect favours option A when option K is bad in price and has an average quality ( bottom-middle of the graph ) , and favours option B when option K has an average price and is bad in quality ( middle-left of the graph ) . We can now apply BCV to model choices in this scenario by analysing the influence on the choice probability ( difference between option A and option B ) of the prior mean μC ( we use an equal prior mean for both attributes price and quality; formally: μCp = μCq ) , the reward uncertainty σR2 ( we use an equal reward uncertainty for both attributes price and quality; formally: σRp2=σRq2 ) , and the reward amounts Rp , K and Rq , K , associated with price and quality respectively , of option K . Fig 5B illustrates the choice probability difference ( between option A and option B ) with prior mean μC = 0 and reward uncertainty σR2=0 . 1 . Focusing on areas of the graph where a similarity effect can be tested ( i . e . , top-left and bottom-right ) , we see that the similarity effect is reproduced by BCV . Moreover , focusing on areas of the graphs where an attraction effect can be tested ( i . e . , bottom-middle and middle-left ) , we can see that this effect can also be explained by BCV . Collectively , these simulations provide proof of principle that , for some sets of values of the prior mean μC and of the reward uncertainty σR2 , BCV explains both a similarity and an attraction effect . Note that these effects are obtained with μC = 0 , which can be considered a default value for this parameter . Fig 6A and 6B examines the effects of adopting other values of the prior mean μC ( fixing the reward uncertainty σR2 to 0 . 1 ) in this scenario . This figure shows that an attraction effect is obtained when the prior mean μC is smaller ( μC = −2 in our simulation ) , but no similarity effect emerges . Conversely , a similarity effect is evident when the prior mean μC is larger ( μC = 2 in our simulation ) , but the attraction effect vanishes . Fig 6C and 6D illustrates the choice probability difference ( between option A and option B ) for different values of the reward uncertainty σR2 ( the prior mean μC was fixed to zero ) . We can see that both similarity and attraction effects are not detectable when reward uncertainty σR2 is high . For smaller values of uncertainty , a similarity effect emerges but there is no attraction effect . Both effects can be obtained only when the reward uncertainty σR2 is sufficiently low ( Fig 5B ) . This highlights the role of the reward uncertainty σR2 in determining the degree of contextual effects . In summary , our analyses show that , when simulating multi-attribute decisions with BCV , similarity and attraction effects emerge for appropriate values of the prior mean μC and the reward uncertainty σR2 . The first parameter regulates the balance of the two effects , as an attraction effect ( but no similarity effect ) is obtained when the prior mean μC is small , while a similarity effect ( but no attraction effect ) is obtained when the prior mean μC is large . Both effects emerge for intermediate values of the prior mean μC , including a prior mean μC = 0 , which is a default value for this parameter . The reward uncertainty σR2 plays a key role too , because context effects vanish when this parameter is high . Decreasing levels of reward uncertainty σR2 reveal a similarity effect first and then an attraction effect . These results indicate that the similarity and attraction effects arise naturally from BCV , without any ad-hoc assumptions–and under natural values of model parameters ( prior mean μC reward uncertainty σR2 ) . A compromise effect [17 , 25 , 27] has been observed when the choice set includes two options scoring high in one attribute and low in another , in addition to a third option with intermediate scores for both attributes . Crucially , the three options are equally preferred ( i . e . , are chosen an equal amount of times ) if presented in pairs during binary choices . However , when they are available altogether , a preference for the option characterized by intermediate scores is seen . We model this scenario by manipulating the distance between attributes for two options A and B , namely assigning Rp , A = 5 − d and Rq , A = 5 + d for option A , and Rp , B = 5 + d and Rq , B = 5 − d for option B , where the proximity parameter d varies ( across simulations ) from zero to four . To represent the option with intermediate scores for both the two attributes , we assign Rp , K = 5 and Rq , K = 5 . Fig 7A , 7B and 7C shows the prediction of BCV using these settings during binary choices between option K and option A , using different parameters for the prior mean μC and the reward uncertainty σR2 . The results indicate that the choice probability difference is always zero , irrespective of the values of the proximity parameter d or the parameters of the model ( prior mean μC and reward uncertainty σR2 ) . Fig 7D , 7E and 7F shows the choice probability difference between option K and option A , when option B is also available . For certain values of the parameters ( prior mean μC and reward uncertainty σR2 ) , this difference is zero with d = 0 and increases with the proximity parameter d . This effect disappears when reward uncertainty σR2 is too large or when the prior mean μC is too small . Overall , these results show that the compromise effect emerges naturally from BCV , without any ad-hoc assumptions and under default values of the parameters ( prior mean μC and reward uncertainty σR2 ) . Interestingly , these simulations predict a correlation between the compromise effect and the proximity parameter d , reflecting differences between the intermediate and extreme options . This phenomenon is predicted by another model of the compromise effect [19] but remains to be validated empirically . In summary , these simulations provide proof of principle that BCV predicts within-choice contextual effects during multiattribute decisions that are remarkably similar to those seen in empirical studies . In other words , the similarity , attraction and compromise effects seen empirically are all emergent properties of BCV . In the next section , we turn from within choice effects and consider between-choice context effects . To characterize between-choice context-effects [11] , BCV uses the same generative model as above , characterized by a prior belief μC ( here we consider only options defined by a single attribute ) over reward ( with uncertainty σC2 ) and by an observation of reward amount R ( with uncertainty σR2 ) . Here , the generative model is extended to include a Gaussian observation variable O that reflects contextual information provided before an option is presented ( Fig 8A ) . This depends on the hidden cause C and is endowed with uncertainty σO2 ( as for the reward amount ) : O∼N ( C , σO2 ) ( 12 ) As above , we assume that an agent infers the posterior expected reward of options afforded by a given context , based on the reward amount but also now on contextual information ( i . e . , μ^C|O , R ) . Since the latter is provided before the option , we assume that the agent infers μ^C|0 first and then μ^C|O , R , when the option is presented . Assuming a prior mean equal to zero μC = 0 , then: μ^C|0=σC2σC2+σO2O ( 13 ) And the posterior uncertainty: σ^C|O2=σC2−σC2σC2+σO2σC2 ( 14 ) The mean of the posterior distribution P ( C|O , R ) corresponds to: μ^C|O , R=μ^C|0+σ^C|O2σ^C|O2+σR2 ( R−μ^C|0 ) ( 15 ) Implying the following incentive value for the option: V ( R ) =σ^C|O2σ^C|O2+σ^R2 ( R−μ^C|0 ) ( 16 ) This shows that , other things being equal , information about context ( reflected in the value of O ) induces subtractive value normalization . For instance , when contextual cues O supports a larger reward , μ^C|0 will be larger and hence the reward prediction error ( i . e . , R−μ^C|0 ) will be smaller . An extension of this generative model is illustrated in Fig 8B , where contexts are organized hierarchically . Combining the influence of reward expectancies within a hierarchy allows the generative model to explain the impact of context at multiple levels . For instance , the value attributed to a certain dish may depend on the reward distribution associated with a restaurant ( a more specific context ) , integrated with the reward distribution associated with a city ( a more general context ) . In detail , a higher-level prior belief about the average reward amount of options ( e . g . , at the level of the neighbourhood ) is represented by a Gaussian distribution with mean μHC equal to zero and uncertainty σHC2 , from which a value HC is sampled . Contextual information about HC is provided and represented by HO that is sampled from a Gaussian distribution with mean HC and uncertainty σHO2 . A lower-level belief about the average reward amount of options ( e . g . , the restaurant ) is represented by a ( Gaussian ) distribution with mean HC and uncertainty σLC2 , from which a value LC is sampled . Contextual information about LC is provided and represented by LO , which is sampled from a Gaussian distribution with mean LC and uncertainty σLO2 . A reward is obtained and sampled from a Gaussian distribution with mean LC and uncertainty σR2 . We propose that agents infer the posterior expectation μ^LC|HO , LO , R P ( LC|HO , LO , R ) sequentially by estimating μ^HC|HO , μ^LC|HO , μ^HC|HO , LO and finally μ^LC|HO , LO , R . This produces an equation for incentive value with the following form ( see Materials and Methods for derivation ) : V ( R ) =K ( R−τLOLO−τHOHO ) ( 17 ) Three normalization factors are implicit here . The first ( τLOLO ) is a subtractive normalization factor proportional to the value LO observed at the low contextual level . The second ( τHOHO ) is a subtractive normalization factor proportional to the value HO observed at the high contextual level . The terms τ represent gain-dependent effects and describe the relative precision of information conveyed by the low-level ( τLO ) and high-level ( τHO ) observations . Finally , a third factor ( K ) implements divisive normalization and depends on a gain term which includes reward uncertainty ( see Materials and Methods for details ) . In recent studies [8–11] , we have investigated the nature of contextual influence on incentive value that depends on reward expectations established before choice presentation ( between-choice effects ) . In these studies , we have used a simple decision-making task , where participants had to repeatedly choose between a sure monetary reward and a fifty-fifty gamble . These options comprised double the sure monetary reward and a zero outcome , ensuring that the two options had equivalent expected reward or value ( EV ) . Across blocks , we manipulated the distribution of EVs , such that these distributions overlapped . We analysed choice behaviour with EVs common to both contexts to examine whether incentive value attributed to the objective EV changed according to BCV predictions . In one experiment ( Fig 9A and 9B; [8 , 9] ) , in different blocks , the sure monetary gain was drawn from one of two distinct , but partially overlapping , distributions of rewards ( low-average and high-average context ) . Choice behaviour was consistent with attributing a larger incentive value to common EVs in the low average compared to high-average context . This and similar evidence [5–14 , 50] suggests that incentive values are , to some extent , rescaled to the average reward expected in a given context , such that they increase ( resp . decrease ) with smaller ( resp . larger ) average reward expectations . These data fit within predictions of BCV . In addition , BCV postulates a between-choice influence of expected reward variance on incentive values ( Fig 9C and 9D ) . In a recent study [11] , we used the same gambling task described above and manipulated contextual variance on two levels; one associated with blocks where two target trial EVs were presented ( low-variance context ) , and another with blocks where the same two target trial EVs plus a larger and a smaller EV were presented ( high-variance context ) . Crucially , this ensured that the two contexts had equivalent average reward but different variance . BCV predicts that the incentive value of the smaller target trial EV will be lower in the low-variance compared to the high-variance context , and the incentive value of the larger target trial EV will be higher in the low-variance compared to the high-variance context . In other words , BCV predicts a larger value difference between the two target trial EV in the low compared to high-variance context . This derives from the gain term , which depends on contextual reward variance . Specifically , low variance magnifies the reward prediction error and hence further reduces the value of rewards that are lower than expected and enhances the value of rewards that are larger than expected . We have previously provided data that are consistent with this prediction [11] . This latter study supports the hypothesis that between-choice reward variance influences incentive value consistent with BCV . In the same study ( Fig 9E and 9F; [11] ) , we also reported that between-choice context effects can be expressed at different hierarchical levels , in line with predictions of BCV . Participants played a computer-based task , where two decks of cards ( representing a low-level context ) appeared . Each card was associated with a monetary reward , and decks contained cards with different average rewards . A card was drawn from a selected deck and participants had to choose between half of the card reward for sure and a gamble between the full reward and a zero outcome , each with 50% chance . Two sets of decks ( representing a high-level context ) alternated in a pseudo-random way . The empirical data showed that the lowest incentive values were attributed when both high-value decks and deck-sets were simultaneously presented , while the highest incentive values were attributed when low-value decks and deck-sets were simultaneously presented . Intermediate incentive values were attributed when decks and deck-sets had one high value and the other low value . Collectively , these empirical studies provide evidence consistent with between-choice contextual effects on incentive value that depends on beliefs about the average reward and variance expected across choices at multiple hierarchical levels . Furthermore , the empirical findings endorse the predictions derived from BCV .
We have shown that the principles underlying BCV can explain a wide range of empirical findings on the context sensitivity of value-based choice . Several previous accounts have focused on a single context effect , especially during multiattribute decisions . Some models have been developed explicitly for explaining the similarity effect [20 , 53–55] , other models for explaining the attraction effect [56 , 57] , and other models for the compromise effect [27] . However , a shortcoming of these models is their inability to explain all three effects within a single formal framework . More recently , adopting connectionist architectures , the multi-alternative decision field theory [16 , 58 , 59] and the leaky competing accumulator [19 , 59 , 60] have been able to reproduce all three effects ( see also [61] ) . The first model [16 , 58] is based on a process modelling attentional switches across attributes and a comparator mechanism which , for the attribute under attention , computes the difference between the reward of each option and the mean reward across options . The second model [19 , 59 , 60] is similar , except that the comparator applies a non-linear asymmetric ( loss-averse ) value function to the difference . Although these models fit remarkably with empirical literature and shed light on the neural mechanisms underlying choice , we argue that BCV presents several advantages . First , it is based on normative principles of Bayesian inference . This constrains the model in terms of empirical predictions . In other words , the similarity , attraction and compromise effect are implicit in the way the model works . In fact , these effects arise when defaults parameters are used . Second , BCV is a more parsimonious model; as the number of free parameters is much lower ( essentially , the prior mean and the reward uncertainty ) . Third , without any further assumptions , BCV applies to a wider range of phenomena including single-attribute decisions and also accounts for between-context effects . Overall , while previous connectionist models are informative especially at the implementation level , BCV helps clarify context sensitivity at the algorithmic and computational level . The concept of wealth in expected utility theory [3] and status quo in prospect theory [62] have been recently re-casted in terms of average expected reward [29] . This formulation opens the possibility of context effects dependent on changes in reward expectation . In line with this view , empirical evidence indicates a between-choice context effect that depends on the average contextual reward ( as for example inferred from past choices ) , consisting in attributing larger incentive values in contexts characterized by lower reward . A similar idea has inspired decision by sampling theory [14 , 31] , which evokes a few basic cognitive processes to explain choice behaviour . According to this model , each choice option elicits retrieval from memory ( in the form of random sampling ) of stimuli encountered in the past , especially those associated with the current context . A set of binary comparisons follows between the option and the samples , and the number of comparisons in which the option is favoured over each sample is recorded . This number corresponds to the incentive value of the option and is computed for all options available , hence determining their relative preference . Since samples are drawn from memory , they depend on past experience and therefore reflect the distribution of options and outcomes characterizing the environment of an agent . This model can account for an attribution of larger incentive value to the same reward in contexts where lower compared to higher reward is expected before options are provided . This effect is explained by a decreased likelihood , in the former compared to the latter context , of sampling stimuli from memory that are preferred to rewards common to both contexts ( assuming a recency effect in memory sampling; [14 , 31] . BCV extends these views by appealing explicitly to Bayesian principles ( i . e . Bayesian belief updating and evidence accumulation ) , with implications for empirical predictions . For instance , contrary to BCV and empirical findings , it remains unclear whether these previous models can account for between-choice contextual influence of reward variance or any within-choice contextual effects . Divisive normalization theory [6 , 63–68] has been proposed recently to explain both between-choice and within-choice contextual effects during single attribute decisions . Divisive normalisation was first proposed in the sensory domain to explain phenomena such as neural adaptation within the retina to stimuli of varying intensity [63] . There is evidence that similar principles can explain higher-order cognitive processes , such as selective attention and perceptual decision-making [63 , 69] . Recently , divisive normalisation has been extended to contextual adaptation effects in value-guided choice [6] , and proposes that incentive value corresponds to the reward divided by the average reward of past or current choices . This can explain contextual influences elicited both within-choice effects during non-multiattribute decisions and between-choice effects that depend on the average contextual reward . Though this scheme relies on a normalization scheme similar to BCV , different empirical predictions arise . It remains unclear whether this divisive normalization scheme is able to explain between-choice effects deriving from reward variance , and can explain data on multi-attribute choices . In addition , BCV , but not divisive normalization theory , is based on normative principles of Bayesian statistics . However , an attractive aspect of divisive normalization theory is the explicit connection with mechanisms characterizing biological neural processes [63] . A similar connection can be motivated for BCV , given several proposals showing how Bayesian inference ( the framework of BCV ) is compatible with neuronal processes [49 , 70 , 71] . The manner in which BCV conceptualizes incentive value is similar to recent economic models that postulate incentive value is adapted to the statistics of the expected reward distribution [29 , 30] . These theories can be broadly classified into those based on subtractive normalization , which assume that incentive value corresponds to the reward minus a reference value [29] , and those based on divisive normalization , assuming that incentive value corresponds to the reward divided ( or multiplied ) by the range of an expected distribution of rewards [30] . An important difference between BCV and these theories is the derivation of the former but not the latter from normative assumptions of Bayesian inference . From Bayesian belief updating , BCV derives the proposition that incentive value corresponds to precision-weighted prediction error , hence implying both a subtractive normalization to the expected reward and a divisive normalization with respect to the reward uncertainty . Importantly , these predictions are not ad hoc but derive from Bayesian assumptions , distinguish BCV from other models , and have been recently supported empirically [11] . In addition , while these recent economic models focus on between-choice context effects , BCV is more general as it can reproduce within-choice effects in both single and multiattribute decisions . Like BCV , a recent proposal has interpreted multi-attribute within-choice effects based on the notion that perception of reward is stochastic [72] . The idea is that , for each attribute , an agent forms noisy observations of reward amounts and of the ordinal positions of the reward amounts . Multi-attribute effects can then be obtained by integrating these two observations [72] . Though there are analogies between BCV and the model of Howes et al . [72] , we emphasize several important differences . First , the latter does not employ a Bayesian framework , since it is not based on integrating prior beliefs and observations , nor it is based on optimal weighting of different sources of information ( as in multi-sensory integration ) . Second , the model of Howes et al . [72] has been applied to aspects of multi-attribute effects ( such as the impact on reaction times ) , which remain to be explored with BCV . On the other hand , the model of Howes et al . , [72] remains to be explored in relation to within-choice effects involving a single attribute and in relation to between-choice effects . Specific empirical predictions can be derived from BCV , and here we highlight some of these . Standard economic theories assume that choice should be independent of whether options are presented simultaneously or sequentially . However , the latter case remains largely to be investigated . BCV may inspire this investigation , as it predicts that a higher value will be attributed to an option after presentation of lower value options . This because BCV proposes a sequential belief updating in which options considered so far contextualize the option observed now . Other predictions involve interactions regarding between- and within-choice effects . For example , consider the example above in which an agent usually evaluates equally car A ( expensive and high quality ) and car B ( cheap and low quality ) . One may design an experiment where participants are first exposed to a set of cars having a fixed level of quality and varying on price . BCV predicts that this manipulation would determine a lower reward uncertainty for quality compared to price . In other words , quality would become more salient than price , predicting a preference for car A over car B . In addition , BCV predicts other forms of interactions regarding between- and within-choice effects dependent on manipulations of the reward uncertainty and the prior mean ( see above ) , which also remain to be explored empirically . Finally , BCV may be relevant for research on the neural underpinnings of decision-making . A main aspect of this theory is the idea that incentive value corresponds to a precision-weighted reward prediction error . Interestingly , reward prediction error is reflected in activity of brain regions involved in reward processing [73] . BCV raises the possibility that a stimulus which elicits a stronger prediction error response in the brain will be attributed a higher incentive value . There are shortcomings to BCV , though we argue that the same framework may be fruitfully used to address some of these shortcomings . A shortcoming of our current formulation assumes that model parameters are given . In reality , these parameters need to be learned in the first place . Questions about the mechanisms that might underpin learning of generative models adopted for Bayesian inference are still largely open , though substantial contributions exist , particularly in the context of structure learning [74–80] . A second shortcoming is that here we have assumed that choices occur after inference has considered all observations . An important extension of BCV is a consideration that action tendencies actually develop during evidence accumulation , and this speaks to models of choice that focus on action dynamics , sequential policy optimisation and reaction times [16 , 46 , 47] . Another important extension of BCV would be to generalize to domains outside incentive value computation . Context effects similar to those observed in value-based decision-making have been reported in many other conditions during perception and judgement [81–84] . Notably , multi-attribute context effects have been recently shown outside incentive value computation [85 , 86] , suggesting that they may derive from a general way in which the brain works [61] . We offer BCV as a unifying theory of contextual effects during choice behaviour based on Bayesian normative principles . BCV predictions are in line with available empirical evidence about context sensitivity seen empirically both within and between-choice . These different effects are explained using the same simple set of principles , invoking minimal assumptions . We argue that strengths of this model are its foundation on normative principles , simplicity , the link with other influential models of brain function , and the ability to explain a wide range of empirical data . This theory may help clarify the nature of incentive value attribution and choice behaviour . This is particularly prescient when trying to understand ecological phenomena and psychopathologies characterized by dysfunctional choice , such as addiction .
Here we derive Eq 17 from the generative model shown in Fig 9B . A higher-level contextual variable ( e . g . , a neighbourhood containing several restaurants ) is represented by a Gaussian distribution with mean μHC equal to zero and uncertainty σHC2 , from which a value HC is sampled . Sensory evidence about HC is provided and represented by HO which is sampled from a Gaussian distribution with mean HC and uncertainty σHO2 . A lower-level contextual variable ( e . g . , one of the restaurants ) is represented by a ( Gaussian ) distribution with mean HC and uncertainty σLC2 , from which a value LC is sampled . Sensory evidence about LC is provided and represented by LO , which is sampled from a Gaussian distribution with mean LC and uncertainty σLO2 . A reward is obtained and sampled from a Gaussian distribution with mean LC and uncertainty σR2 . The posterior distribution P ( LC|HO , LO , R ) can be inferred sequentially in the order P ( HC|HO ) , P ( LC|HO ) , P ( LC|HO , LO ) , and P ( LC|HO , LO , R ) . The posterior mean of P ( HC|HO ) is: μ^HC|H0=σHC2σHC2+σHO2HO ( 18 ) And the posterior uncertainty: σ^HC|H02=σHC2−σHC2σHC2+σHO2σHC2 ( 19 ) The posterior mean of P ( LC|HO ) is equal to μ^HC|H0 ( μ^LC|H0=μ^HC|H0 ) , while the posterior uncertainty is: σ^LC|H02=σ^HC|H02+σLC2 ( 20 ) The posterior mean of P ( LC|HO , PO ) is: μ^LC|HO , LO=μ^LC|H0+σ^LC|HO2σ^LC|HO2+σLO2 ( LO−μ^LC|H0 ) ( 21 ) And the posterior uncertainty: σ^LC|H0 , LO2=σ^LC|H02−σ^LC|HO2σ^LC|HO2+σLO2σ^LC|H02 ( 22 ) The posterior mean of P ( LC|HO , LO , R ) is: μ^LC|HO , LO , R=μ^LC|HO , LO+σ^LC|HO , LO2σ^LC|HO , LO2+σR2 ( R−μ^LC|HO , LO ) ( 23 ) Finally , with few rearrangements , we obtain the following incentive value for a reward offer: V ( R ) =σ^LC|HO , LO2σ^LC|HO , LO2+σR2 ( R−σ^LC|HO2σ^LC|HO2+σLO2LO−σLO2σ^LC|HO2+σLO2σHC2σHC2+σHO2HO ) ( 24 ) This equation implements three normalization factors: ( i ) a subtractive normalization factor ( σ^LC|HO2σ^LC|HO2+σLO2LO ) proportional to the value LO observed at the low contextual level , ( ii ) a subtractive normalization factor ( σLO2σ^LC|HO2+σLO2σHC2σHC2+σHO2HO ) proportional to the value HO observed at the high contextual level , ( iii ) a divisive normalization factor ( σ^LC|HO , LO2σ^LC|HO , LO2+σR2 ) that captures the weighting dependent on the ( relative ) reward uncertainty . If we define the three factors as τLO and τHO and K respectively , we obtain Eq 17 . | Research has shown that decision-making is dramatically influenced by context . Two types of influence have been identified , one dependent on options presented in the past ( between-choice effects ) and the other dependent on options currently available ( within-choice effects ) . Whether these two types of effects arise from similar mechanisms remain unclear . Here we offer a theory based on Bayesian inference which provides a unifying explanation of both between and within-choice context effect . The core idea of the theory is that the value of an option corresponds to a precision-weighted prediction error , where predictions are based upon expectations about reward . An important feature of the theory is that it is based on minimal assumptions derived from Bayesian principles . This helps clarify the contextual nature of incentive value and choice behaviour and may offer insights into psychopathologies characterized by dysfunctional decision-making , such as addiction and pathological gambling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"decision",
"theory",
"medicine",
"and",
"health",
"sciences",
"decision",
"making",
"applied",
"mathematics",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"cognitive",
"psychology",
"mathematics",
"statistics",
"(mathematics)",
"cognition",
"memory"... | 2017 | A unifying Bayesian account of contextual effects in value-based choice |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.